Essays on Argumentation in Antiquity 3030708160, 9783030708160

Citation preview

Argumentation Library

Joseph Andrew Bjelde David Merry Christopher Roser   Editors

Essays on Argumentation in Antiquity

Argumentation Library Volume 39

Series Editor Frans H. van Eemeren, University of Amsterdam, Amsterdam, The Netherlands Editorial Board Fernando Leal Carretero, University of Guadalajara, Guadalajara, Jalisco, Mexico Maurice A Finocchiaro, Department of Philosophy, University of Nevada at Las Vegas, Las Vegas, NV, USA Bart Garssen, Faculty of Humanities, TAR, University of Amsterdam, Amsterdam, Noord-Holland, The Netherlands Sally Jackson, Communication, University of Illinois Urbana-Champaign, Urbana, IL, USA Wu Peng, School of Foreign Languages, Jiangsu University, Zhenjiang, China Sara Rubinelli, University of Luzern, Nottwil, Luzern, Switzerland Takeshi Suzuki, School of Information and Communication, Meiji University, Chiyoda-ku, Tokyo, Japan Cristián Santibañez Yañez, Faculdad de Psicologia, University of Concepción, Concepción, Chile David Zarefsky, School of Communication, Northwestern University, Evanston, IL, USA Sara Greco, IALS, Università della Svizzera Italiana, Lugano, Ticino, Switzerland

Since 1986 Springer, formerly Kluwer Academic Publishers, publishes the international interdisciplinary journal Argumentation. This journal is a medium for distributing contributions to the study of argumentation from all schools of thought. From a journal that published guest-edited issues devoted to specific themes, Argumentation has developed into a regular journal providing a platform for discussing all theoretical aspects of argumentative discourse. Since 1999 the journal has an accompanying book series consisting of volumes containing substantial contributions to the study of argumentation. The Argumentation Library aims to be a high quality book series consisting of monographs and edited volumes. It publishes texts offering important theoretical insights in certain major characteristics of argumentative discourse in order to inform the international community of argumentation theorists of recent developments in the field. The insights concerned may pertain to the process of argumentation but also to aspects of argumentative texts resulting from this process. This means that books will be published not only on various types of argumentative procedures, but also on the features of enthymematic argumentation, argumentation structures, argumentation schemes and fallacies. Contributions to the series can be made by scholars from a broad variety of disciplines, ranging from law to history, from linguistics to theology, and from science to sociology. In particular, contributions are invited from argumentation theorists with a background in informal or formal logic, modern or classical rhetoric, and discourse analysis or speech communication. A prerequisite in all cases is that the contribution involved is original and provides the forum of argumentation theorists with an exemplary specimen of advanced scholarship. The Argumentation Library should enrich the study of argumentation with insights that enhance its quality and constitute a fruitful starting point for further research and application. All proposals will be carefully taken into consideration by the editors. They are to be submitted in fourfold. If the prospects for including a certain project in the series are realistic, the author(s) will be invited to send at least three representative chapters of their manuscript for review to the editors. In case the manuscript is then judged eligible for publication, the complete manuscript will be reviewed by outside expert referees. Only then a final decision can be taken concerning publication. This book series is indexed in SCOPUS. Authors interested in submitting a proposal or completed manuscript can contact either [email protected] or the Series Editor.

More information about this series at http://www.springer.com/series/5642

Joseph Andrew Bjelde · David Merry · Christopher Roser Editors

Essays on Argumentation in Antiquity

Editors Joseph Andrew Bjelde Institut für Philosophie Humboldt-Universität zu Berlin Berlin, Germany

David Merry Institut für Philosophie Humboldt-Universität zu Berlin Berlin, Germany

Christopher Roser Institut für Philosophie Humboldt-Universität zu Berlin Berlin, Germany

ISSN 1566-7650 ISSN 2215-1907 (electronic) Argumentation Library ISBN 978-3-030-70816-0 ISBN 978-3-030-70817-7 (eBook) https://doi.org/10.1007/978-3-030-70817-7 © Springer Nature Switzerland AG 2021 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Introduction

Ancient thought about argumentation is strikingly different from most contemporary thought about it. Consider the central issue of what makes for good arguments. According to a conservative contemporary view, good arguments must be deductively valid. The standard contemporary view is a bit broader. It would fault this conservative view of good argumentation for requiring deductive validity, and would seek to extend the conservative view by generalizing in the right way from deductive validity. That generalization might take one of several forms: perhaps the inclusion of defeasible argumentation, or of a Bayesian theory of inductive inference. The standard view is not universal, of course. In the wake of Toulmin’s sustained criticism of it ([5]), argumentation theorists inspired by the ancients have developed alternatives to it. But the standard view remains widespread. We suspect that this is in part because philosophers fear that any further broadening of the conception of argumentation and its uses will lead us rapidly away from the domain of philosophy and the pursuit of truth, and into sales, self-help, or politics. For such readers, we hope this collection, by showcasing the philosophical richness of ancient thought about argumentation, will help usher in a less constricted view about argumentation. Readers who are instead inclined to doubt the standard view will discover ample support for their doubts. For ancient thought fairly bristles with heresies against the standard contemporary view. Plato’s Socrates often demands rather more of an argument than merely that its premises are true and its conclusion follows deductively. When he offers a protreptic argument, he hopes that it will actually convince his interlocutors to change their whole outlook—including their desires—for the better. In contrast, Aristotle in the Topics demands rather less. Because dialectical argumentation is in part a competition to prove who is the more skilled arguer, downright fallacious arguments have their place.1 And when the point is to persuade an opponent, it can be reasonable to argue from false premises, so long as one’s opponent believes them. Other rules of dialectical argument are even more foreign. Arguments must be kept relatively short because of time limits on dialectical debates. And an answerer must accept an 1 Topics

I.18.108a26–31; Topics II.3.110a24–28 for Aristotle encouraging use of equivocation in dialectic. v

vi

Introduction

epagoge if they are confronted with a series of examples in which the generalization holds, and they cannot think of an exception. If the point of argumentation is to carry truth, or justification, from the premises to the conclusion by way of a secure inference, then this rule is no good. But dialectic often had a different and equally noble goal: showing someone they did not know nearly as much about a topic as they thought they did. With this goal in mind, the rule seems much better. And that goal itself is recognizably—perhaps paradigmatically—philosophical. Persuasion also plays a strikingly different role in ancient conceptions of argumentation than it does on the standard contemporary view. For that contemporary view relegates persuasion to the realm of psychologists, evangelists, salespeople, and politicians—anyone but philosophers. The job of a philosopher, as that of a physicist, is to discover the truth, and whether anybody believes a word they say is not their concern. Readers who share this view are often surprised to find Socrates in the Gorgias agreeing with his interlocutors—the rhetors—in aiming not only at truth but at persuasion (though they differ on whom to persuade; cf. Gorgias 471c-2b). Indeed, according to one reading, the Gorgias and Euthydemus are exploring how the right, philosophical, kinds of persuasion differ from the wrong, sophistical or rhetorical, kinds. Authors writing before Aristotle may have struggled to share the standard contemporary view, as they did not yet have the central notion of formal logical validity. Having invented syllogistic, Aristotle was in a position to formulate the thought that deductively valid syllogisms are the central cases of good argument. But then it is all the more striking that Aristotle does not seem to adopt that standard contemporary view. Even setting aside the Topics, consider what Aristotle has to say about argumentation in the Rhetoric, or the role of endoxa elsewhere in Aristotle’s philosophical works (cf. [2]), or the recent interpretive furor over his notion of epagoge (cf. [3]). In each of these cases, Aristotle incorporates the earlier ideas without giving any indication that they are to be reduced to or understood in terms of his syllogistic. We are not advocating going as far as Toulmin [5] did in rejecting the standard contemporary view in favour of an empirical study of the arguments actually made. Much less are we suggesting that, if contemporary philosophers only knew how to read it, they would find the ultimate truth about argumentation in Plato’s Euthydemus or Aristotle’s Topics. Nonetheless, we do think that the standard contemporary view asks too much of argumentation in some ways, and too little in others. And we also think that the rich array of views we find discussed in antiquity offers some helpful insights. Argumentation theory has already benefited considerably from many such insights, but our understanding of ancient thought on argumentation remains limited, and much of what we have to learn from ancient thinkers is still inaccessible to contemporary readers. The essays in this volume go some distance towards improving the situation; we hope that they will also encourage further improvement. But the study of ancient thought on argumentation is not primarily an exercise in excavating correctives to the standard contemporary view. It is also our best hope for learning how to engage with ancient philosophers on their own terms. Very often, in reconstructing an argument in Plato or Aristotle we wonder whether it commits a fallacy, e.g. begging the question. As historians, it is better practice to see whether

Introduction

vii

the argument meets Plato’s or Aristotle’s understanding of begging the question than the understanding in the latest edition of Argumentation. Even more often, we find an argument that is very difficult to reconstruct. For instance, perhaps the author left a number of premises either tacit, or the Greek or Latin we have is opaque, ambiguous, riddled with lacunae, or different in different manuscripts. In that case, we must make educated guesses as to what these premises were. Understanding the author’s thought about argumentation helps us navigate these difficult straits. In reconstructing Aristotle’s critiques of Plato’s theory of forms, for example, we need to know whether we should fill the gaps with premises Aristotle would have believed, Plato would have believed, or both. A passage in the Topics suggests that, in this task, we should opt for Platonic doctrines, even ones Aristotle himself rejected. But the context of the passage is difficult, and considerable care is required in its interpretation.2 Similarly, interpreters not infrequently suggest that Plato is intentionally making Socrates argue fallaciously. If Plato’s critique of sophistic leaves room for nonsophistical uses of fallacy, then this might confirm their suspicions. That confirmation would in turn have serious ramifications for how we applied the principle of charity to Socratic arguments. For a final example, we are often left wondering just what kind of support a given premise is supposed to provide for a given conclusion—for instance, whether the argument is an instance of inference to the best explanation. A better understanding of ancient thought on argumentation would help answer such questions. For interpreting Aristotle, a close familiarity with the argument forms in the central books of the Topics would no doubt be a great asset in accurately identifying forms of argument. But reading these books is a confusing and frustrating exercise, for which extant scholarly resources remain deeply inadequate. The essays in this volume go some distance also towards improving our understanding of how ancient philosophers conceived of their own argumentation. Many of the essays in this volume grew out of talks presented at a conference ‘Argumentation in Classical Antiquity: dialectic, rhetoric, & other domains’, which took place at Humboldt Universität zu Berlin in June 2016. In selecting essays from that conference and beyond, we have preferred contributions which are of broader interest, rather than broader coverage of thinkers or issues. That preference has also driven the organization of the collection into three sections: ‘Before Plato’, ‘Plato’, and ‘After Plato’, which reaches as far as contemporary argumentation theory. Clearly, Plato takes centre stage in this volume. But, in our view, this is a natural casting, since Plato occupies a central position for studying argumentation in antiquity. Moreover, Plato’s treatment of argumentation is deeply engaged with earlier traditions, such as the rhetorical tradition which Gorgias and Isocrates represent, and was also especially influential for later philosophical treatments, such as Aristotle’s. A scholar of Aristotle’s rhetoric can perhaps do without a particularly good understanding of Gorgias, and a scholar of Gorgias may not need to take Aristotle into account. But both need to take Plato into account, and scholars of Plato can profit a great deal from reading both Aristotle and Gorgias. So a scholar looking to get

2 Topics

VIII.5.195b16–33; cf. Theaetetus 166a—b.

viii

Introduction

a sense of the overall development of ancient thought about argumentation will do well to see Plato as standing at the centre of it. Two further reasons for the centrality of Plato have to do with his use of dialogue form. Inspired in part by Vlastos’s analysis of the Socratic elenchus and Kahn’s concentration on dramatic contexts, scholars increasingly see arguments in Plato not, as the standard contemporary view would, as merely free-floating propositions bound together by an x-to-1 inferential relation, but rather as attempts of one character to turn another towards the truth or away from ignorance through persuasion (cf. [1] and [4]). As a result, the inadequacy of the standard contemporary view for capturing what is philosophically interesting in Plato’s arguments has long been relatively widely appreciated, and existing scholarship on argumentation in Plato is correspondingly richer than elsewhere in ancient philosophy. Second, Plato’s interlocutors often offer rhetorical and dialectical arguments, about which Aristotle gives advice in the Topics and Rhetoric. So investigating Plato’s dialogues occasions fruitful questions about whether the interlocutors argue as Aristotle would have advised them to. Seeing that they do may sometimes be key to understanding an otherwise difficult passage in Plato, and may by the same token illuminate Aristotle’s advice. Several of the essays in this volume move us further in these directions. More details on individual contributions are below.

Before Plato For earlier thinkers, where we lack any record of reflection about argumentation, a careful examination of the structure of their argumentation is called for. This is especially the case for Parmenides’ poem, which is often cited as the first example of explicit philosophical or deductive argumentation. So, in Truth Attending Persuasion, Stephen White details Parmenides’ debts to traditional styles of argument, as well as his pioneering argumentative innovations. In particular, White argues that the overall structure of the poem is elenctic, in the sense that it supports its conclusions by showing how their denials lead to contradictions. Under that novel structure, however, lie traditional and informal techniques that are best characterized as rhetorical strategies. Lisa Indraccolo’s ‘Argumentation and Persuasion in Classical Chinese Literature’ reveals that similar issues were important in China during the Warring States period (475–221 BCE). Indraccolo describes and analyses the texts of ‘wandering persuaders’ who acted as advisors and diplomats to local rulers. Even at the time, some of them—especially those associated with the ‘school of names’—were harshly criticized for their use of words. As no handbook or manual on classical Chinese rhetoric has been transmitted, scholars of Chinese rhetoric in this period must extrapolate the key techniques from argumentative texts. Indraccolo’s chapter discusses a range of techniques drawn from a representative selection of dialogues from Classical Chinese pre-Imperial and early imperial texts (4th century BCE–2nd century CE). It focuses on the use of authoritative quotes, historical allusions, rhetorical devices,

Introduction

ix

tropes and arguments for discussion, and transition words. It further shows how these techniques are deployed in a relatively consistent pattern across different texts, indicating a well-established set of compositional rules and conventions. Indraccolo’s chapter makes these issues accessible to a non-Sinological audience. Aside from the intrinsic interest of the material, Indraccolo’s chapter suggests avenues for comparison and contrast with what was occurring in the western canon at the time. A full understanding of thought on argumentation in ancient Greece involves situating it in its historical global context, and showing how it is distinctive and how it is similar to other traditions. A history of thought about argumentation that focuses solely on its origins within ancient Greece is strikingly incomplete. Our hope is that Indraccolo’s chapter will facilitate the cross-disciplinary work required to provide a more global perspective on the issues discussed in this volume. Wolfgang Mann’s essay considers Gorgias’ evaluation of argumentation (logos) in the Defense of Palamedes and the Encomium of Helen. He shows that Gorgias actually argues for the shortcomings of argumentation in his speeches. This shows that Gorgias had a much more sceptical view of the strength of argumentation than Plato attributes to him in the Gorgias. While Plato’s character Gorgias in the Gorgias suggests that argumentation is enormously powerful, the real Gorgias emphasizes its weakness. Mathieu Marion explores how ideas of dialectic were formed. In particular, he argues it arose as a specialized form of dialogue for purposes of inquiry and refutation. In so doing, he elucidates the relation between dialectic, as a specialized form of argumentation, and other forms of argumentation in Plato’s dialogues, by using early classifications of the dialogues in terms of their purposes and modern argumentation theory.

Plato Plato’s own reflection on argumentation is as enigmatic as it is seminal. The contributions here focus on how Plato understands different kinds of argumentation and how he incorporates influences from Socrates, rhetors, and sophists. Christopher Roser considers how Plato distinguished dialectical from rhetorical uses of arguments. He argues that none of the seven different attempts to understand the distinction provides a general criterion of distinction. This shows not only that the difference between them is more tacit than has often been thought, but also that the common specifications of dialectical argumentation do not provide a sharp definition of it. However, he argues that we should not conclude from this that there is no significant difference between these two forms of argumentation. Instead, he proposes that Plato distinguishes the two by specifying paradigmatic differences between them. A different kind of sophistic influence in Plato is traced in David Crane’s essay, Platonic βραχ υλoγ ι´α and Aristotle on Say-What-You-Believe. Crane uses Aristotle’s remarks about sophistic uses of say-what-you-believe to bring out a richer

x

Introduction

understanding of Socrates’ argumentation in Plato’s Gorgias. In particular, he argues that say-what-you-believe is not a purely Socratic requirement, but a sophistic one—though Socrates’ use of the requirement distinguishes him from the sophists. Benoît Castelnérac considers how Plato distinguishes two modes of public argumentation by contrasting Socrates with his rhetorical interlocutor in the Gorgias. The one type of argumentation is oriented towards the earnest search for truth, and the other one is driven by the love of victory. With this distinction in hand, he rejects recent attempts to identify these two modes with two methods, the elenctical and rhetorical method. The value of argumentation lies in the personal motives, not in the method. Chris Tindale explores the relationship between dialectic and rhetoric. He first argues against the common view that Plato evaluated rhetoric negatively, and proposes instead that the relation between dialectic and rhetoric in the later Plato can be understood by attributing to each different argumentative roles—to the former a refutative role, to the latter a persuasive role. Holly Moore’s ‘Platonic Epag¯og¯e and the “Purification” of the Method of Collection’ argues that the step of ‘purification’ in Plato’s method of collection corresponds to what Aristotle would call epag¯og¯e, and so his theory of purification can be understood as a theory of epag¯og¯e. This is a striking conclusion, as Plato is generally thought to have no theory of epag¯og¯e. Moore argues that this is because Robinson wrongly assumed that Platonic epag¯og¯e would have to be empirical. Drawing on descriptions of purification in the Sophist and the Statesman, Moore shows how Platonic purification involves moving from the many to one, whether this involves moving from sensory particulars, or from less to more general intelligibles, matching the description of dialectic in the Republic and the Philebus.

After Plato Aristotle’s Topics, as a kind of training manual for aspiring dialecticians, offers a rich source for Aristotelian thought about dialectic. But how relevant the Topics is for understanding Aristotle’s thought depends heavily on two things. First, it depends on the dating of the work. If the Topics is an early work, written while Aristotle was in Plato’s Academy, then it would not necessarily tell us anything about Aristotle’s mature thought. But that developmental hypothesis is necessarily entangled with views about the philosophical content of the Topics. For instance, if the Topics describes positions incompatible with views in the Posterior Analytics or other mature works, that would support developmental hypotheses. One especially thorny point of interpretation on which the Topics has been thought to be incompatible with Aristotle’s mature works is the epistemological and philosophical roles foreseen for dialectic in Topics I.2. In Endoxa and Epistemology in Aristotle’s Topics, Joseph Bjelde argues that dialectical skill is epistemologically useful for coming to grasp definitions, on the basis of a much-discussed passage of Topics I.2—against Robin Smith’s influential criticisms. But, he argues, endoxa have no special status which

Introduction

xi

could help explain that fact. So, he concludes, the best explanation of how dialectical skill plays this role is that, at least in the Topics, Aristotle has a Platonic view on which dialectical testing satisfies a necessary condition on the grasp of first principles which Aristotle will call nous in Posterior Analytics II.19: it makes them more familiar. Andrew Aberdein asks whether Aristotle was a virtue argumentation theorist, and argues for a qualified positive answer. In particular, he suggests that the role played by the phronimos in Aristotle’s rhetoric suggests that phronesis is necessary for doing non-deductive reasoning well. If Aristotle is right about that, he argues, then other kinds of virtue theory—virtue ethics and virtue epistemology, especially—will in an important sense depend on virtue argumentation theory. David Merry argues that Seneca’s separation of argumentation from inference offers intuitionists a potential response to Walter Sinnott-Armstrong’s argument from disagreement. It does so by eliminating biases, which it does by means of an informal study of argumentative strategies aimed at enabling clear moral judgement. Douglas Walton shows how deeply indebted contemporary theory of argumentation is to Aristotle’s Topics. He argues that argumentation theorists have underestimated how closely the structure of Aristotelian dialectical argument resembles recent approaches. This paper shows how Aristotelian dialectical argument can be modelled using ten features, and underscores its importance as a precursor for contemporary argumentation theory. As this brief overview makes clear, the essays in this volume cover a broad span of antiquity, from Parmenides to Seneca, and take a range of different approaches, from close analysis of passages to consideration of major thematic issues. Collectively, these essays show that ancient thought on argumentation is surprisingly extensive and interconnected in striking ways. Our hope is that they also make it clear how much work remains to be done on understanding ancient argumentation, including work on thinkers and issues underrepresented in this volume, and how that work might be fruitfully undertaken.

References [1] [2] [3] [4] [5]

Blondell, R. (2012). The play of characters in Plato’s Dialogues. Cambridge University Press. Frede, D. (2012). The Endoxon Mystique: What Endoxa are and what they are not. Oxford Studies in Ancient Philosophy, 43, 185–215. McCaskey, J. P. (2007). Freeing Aristotelian Epagoge from “Prior Analytics” II.23. Apeiron, 40(4): 345–374. Rowe, C. J. (2007), Plato and the art of philosophical writing, Cambridge University Press. Toulmin, S. (1958). The Uses of Argument. Cambridge University Press.

Contents

1

Truth Attending Persuasion: Forms of Argumentation in Parmenides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Stephen White

1

2

Argumentation and Persuasion in Classical Chinese Literature . . . . Lisa Indraccolo

21

3

Gorgias and the Weakness of Logos . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Wolfgang-Rainer Mann

49

4

Plato’s Dialogues: Dialectic, Orality and Character . . . . . . . . . . . . . . . Mathieu Marion

69

5

What is the Difference Between the Dialectical and Rhetorical Use of Arguments? A Paradigm-Based Approach to Plato’s Socrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Christopher Roser

99

6

Platonic βραχυλoγ…α and Aristotle on Say-What-You-Believe . . . . . 137 David Crane

7

Rhetoric, Dialectic and Shame in Plato’s Gorgias . . . . . . . . . . . . . . . . . 157 Benoît Castelnérac

8

The Prospects for Rhetoric in the Late Plato . . . . . . . . . . . . . . . . . . . . . 173 Christopher W. Tindale

9

Does Plato Have a Theory of Induction? Epag¯og¯e and the Method of Collection “Purified” of the Senses . . . . . . . . . . . . 185 Holly Moore

10 Endoxa and Epistemology in Aristotle’s Topics . . . . . . . . . . . . . . . . . . . 201 Joseph Bjelde 11 Was Aristotle a Virtue Argumentation Theorist? . . . . . . . . . . . . . . . . . 215 Andrew Aberdein xiii

xiv

Contents

12 Seneca’s Argumentation and Moral Intuitionism . . . . . . . . . . . . . . . . . 231 David Merry 13 Aristotelian Dialectic, Argumentation Theory and Artificial Intelligence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245 Douglas Walton

Chapter 1

Truth Attending Persuasion: Forms of Argumentation in Parmenides Stephen White

Abstract Parmenides marks a watershed in the history of argumentation, presenting the earliest surviving sequence of recognizably deductive reasoning in the Greek tradition. This chapter focuses on the central section of his poem (fr. 8 DK) and examines the form of its argumentation: its use of indirect proof, the articulation of its reasoning, and the role necessity plays in it.

The poem of Parmenides is a watershed in the history of argumentation, introducing and deploying at considerable length and for the first time on record forms of reasoning Aristotle reasonably calls deductive (“syllogistic”: Physics 1.2– 3), including many credible candidates for valid arguments. The poem does this, famously, in the service of claims so profound in their implications, and so tantalizing in their obscurity, that readers ever since have puzzled over its meaning and import. Fascination with the mysteries of “being” that occupy most of the poem’s surviving lines has accordingly diverted attention from the way in which those mysteries are revealed. At the close of the past century David Sedley could still complain that “too little has been said about his detailed arguments” [19, p. 113]. The intervening years have seen major strides in redressing the imbalance, notably by [11], [16], and [23].1 Yet their contributions still focus on the substance of the arguments, the claims advanced and their significance, attending to matters of form and formulation mainly as guides to reconstruction or interpretive constraints. What remains to be done is to isolate and assess the form of the poem’s argumentation, the various ways it articulates its arguments, enunciates its premises, supports its conclusions, in short the architecture of its reasoning. That is the project to which I aim to contribute here by offering a fresh assessment of the poem’s formal achievement. Sidestepping 1 Thus [1, pp. 177–178] introduces his trailblazing analysis with the verdict that “in B8 we have a deduction far more complex and far more self-conscious than anything the Presocratics have yet offered us.” Cf. [8] and [9, pp. 92–94].

S. White (B) University of Texas at Austin, Austin, TX, USA e-mail: [email protected]

© Springer Nature Switzerland AG 2021 J. A. Bjelde et al. (eds.), Essays on Argumentation in Antiquity, Argumentation Library 39, https://doi.org/10.1007/978-3-030-70817-7_1

1

2

S. White

mysteries of metaphysics so far as possible, I examine how the poem frames its arguments in order to highlight both its pioneering innovations and some of the more traditional strategies it deploys. My analysis focuses on the long sequence of argument in fr. 8 DK, and from three different angles. The first argues that the entire sequence is fundamentally elenctic. In line with an injunction issued immediately beforehand—“decide by logos the highly contentious elenchus | spoken by me” (fr. 7.5–6)—a series of four arguments each supports a distinct conclusion by showing how denying it entails contradiction. The second then shows how the sequence also deploys informal styles of argument, notably rhetorical questions and a pervasive reliance on unstated premises characteristic of Aristotelian enthymeme. The third takes a closer look at the role assigned to necessity throughout the sequence, as marked especially by various modal terms and potential optatives.

1.1 Road of Truth The poem opens with the strange epiphany of an unnamed goddess, who promptly distinguishes in epistemic terms two domains, which she characterizes as exhaustive ´ (παντα fr. 1.28): one of “persuasive truth” (singular ¢ληθε´ιης), hitherto accessible only to immortals, and the other of “mortals’ beliefs” (plural δ´oξας), hitherto the only domain accessible to us (1.28–32). The poem then proceeds in three stages: a preliminary survey of these two domains along with a third, which is dismissed as “wholly uninformative” (frs. 2–7); then a dense chain of argument circumscribing the domain of persuasive truth (fr. 8.1–49); and finally an extended account of the domain of mortal beliefs (fr. 8.50–61 and frs. 9–19). The middle section on truth may be complete, the introductory section nearly so, but only shards of the third survive: roughly fifty lines each. The key passages for our topic are frs. 2 and 8. The introductory section is a minefield of controversy, rife with disputes of all sorts, textual, syntactic, semantic, logical, and more.2 Fortunately, most of these we can set aside, thanks to our focus on argumentation. The only crucial point appears in fr. 2, where the goddess characterizes the domain she will examine more fully in fr. 8: a domain accessible only via the first of two roads she declares can yield a rarefied form of Noesis that alone unites truth and persuasion (fr. 2). Come, I shall tell and you bring back the tale you hear, The only roads of inquiry there are for Noesis3 : 2 Stretching

back to Parmenides’ own day: see [15].

3 Taking the infinitive νoÁσαι as final, not epexegetic: two roads “leading to Noesis”, not “as objects

of Noesis”; see [16, pp. 63–73] at p. 71, citing [20, §1969] and Empedocles fr. 3.12 π´oρoς ™στ`ι νoÁσαι; as often, purpose shades into result, cf. [20, §§2004–5 and §2011a]. The goddess introduces a third road in fr. 6, so these first two cannot be the only roads “conceivable”, though they may be

1 Truth Attending Persuasion: Forms of Argumentation in Parmenides

3

One road, that it is and is not not to be, Is the path of persuasion, for it attends truth; The other, that it is not and is ordained not to be, That very one I declare to you is a trail wholly uninformative; For nor could you know anything that is not—it can’t be done— Nor could you declare it.4

Immense effort has been devoted to determining what force the several uses of “is” (™στι and its cognates) have here and elsewhere in the poem; which if any uses have an implied subject or complement; and what those subjects or complements might be in each case. Happily we can sidestep those disputes. All that matters here is a simple point of grammar: lines 3 and 5, which introduce “the only roads of inquiry for Noesis” (as I shall call it to defer questions about what kind of cognitive state is at issue), each characterize these roads with a conjunction; and unless the second conjunct in each case is redundant, the item or items on each road must meet not one but two conditions. The first road—call it Road 1—involves in some as yet unspecified way, as itinerary or destination, objects or objective, or whatever else, only “is” and admits no corresponding negations, no “is not”. Likewise the second road—call it Road 2—has two parallel but opposite constraints: it involves only “is not” and admits no corresponding (double) negations, no “is” at all. What exactly the second conjunct in either case adds—importing modal force of some sort, for example, or simply glossing the first clause—need not trouble us here.5 All I insist on is that the second clauses are universalizing, so that Road 1 bans any and all denial of anything affirmed on it; and Road 2 conversely bans any affirmation entirely, ruling out the affirmation of anything at all.6 Put another way, if a potential item for Noesis—whether an entity A, or its existing, or its being F, or anything else—is to be both affirmed and denied in any way—if A is F in one way or respect but not another, for example—then it does not figure on either road. Only if the item is only to be affirmed, in every way or respect, does it figure on Road 1; and only if there is nothing to be affirmed at all, in any way or respect, does it figure on Road 2. Crucially, although the initial clauses by themselves form a pair of contradictories, both exclusive and exhaustive, the second clauses make the two Roads only contraries, exclusive but no longer exhaustive, by opening up an otherwise excluded middle—call it Road 3—for anything affirmed in some ways or respects but denied in others. Moreover, what Road 1 excludes is simply any denial of anything affirmed. Only affirmations that hold universally can be found on this road. So far as we know at this point, the same goes for affirmations that contain some form of negation, the only ones offering Noesis, as opposed to merely true belief; see [16, pp. 51–105]. The aorist also favors some sort of cognitive success rather than mere endeavor, hence knowledge or insight achieved rather than ongoing episodes of thinking about something. 4 All translations are mine; they aim to be literal and neutral, but much remains controversial. 5 For [16, pp. 83–85], both import necessity; cf. [23, pp. 19–21]. For [12, pp. 71–72], both “function epexegetically” to limit predications to what anything “really is”. 6 These formulations remain seriously indeterminate, especially for Road 2; but this indeterminacy is not something we need resolve here.

4

S. White

either implicit or explicit denials. Provided what is denied only “is”—provided what is denied “is” in every way and respect, the way every diameter always and only is both incommensurable and not commensurable, for example—then there will be room for it too on Road 1. In any case, this requirement of universality, which I call the Governing Constraint (GC) because it figures repeatedly and decisively in the argument of fr. 8 as the goddess deduces a set of “signs” for Road 1, stipulates that this Road involves only being: what is and only is—in whatever way we are to construe “is” for any of its subjects or complements. The sequel in fr. 2 goes on to declare Road 2 “wholly uninformative” and thereby, despite its initial introduction as a viable path for Noesis (2.2), ineligible for any ` ν γνo´ιης 2.7; cf. ¢ν´oητoν 8.17). After elaborating her “knowing” at all (oÜτε γαρ critique of Road 2 and raising related problems for Road 3—the intermediate domain of what both is and is not—in frs. 6–7, the goddess returns to Road 1 as the sole remaining option for Noesis; and after first directing the youth who is her auditor and emissary, and thereby us as her external audience, to weigh her words critically (fr. 7.5–6), she launches her account of Road 1 (fr. 8.1–4)7 : Decide by logos the highly contentious elenchus Spoken by me. Only one tale of a road still remains, that it is; and on this road there are signs, very many: that in being it is ungenerated and unperishing, whole and monogenous and unshaken and unwavering.8

´ These “signs” (σηματα) are then taken up in order over the course of the next 45 lines, which present a series of arguments for each of them in turn.9 The dominant strategy throughout is elenctic, as the goddess herself announces; and as we shall see, 7 Immediately, if Sextus, Adv. Math. 7.111 is accurate in citing fr. 7.2–6 continuously with 8.1–2; otherwise very shortly thereafter. The main objection to following Sextus is lexical: metrically the continuation requires μ´oνoς (so Sextus) but Parmenides elsewhere uses epic μoàνoς (2.2 and probably 8.4); see [16, pp. 381–382]. A distinct question is whether the injunction to “decide by logos” applies to the impending arguments of fr. 8 or only the preceding critique in frs. 2–7, as “elenchus spoken by me” (aorist ·ηθšντα) has led some to conclude; so [16, pp. 108–109]. But aorist participles are often prospective (“after it has or will have been spoken”); and since any injunction is inherently prospective, and in this case a call for critical reasoning that requires meticulous analysis, listeners will reasonably expect the entire injunction, and thus the range of the elenchus, to extend in both directions, covering both frs. 2–7 and fr. 8, perhaps also the sequel extending through frs. 9–19 as well; cf. [6]. 8 The list is complicated by both textual and interpretive problems, including variants for every term in 8.4. I translate the eclectic text in [2], except for the last term, where I adopt “unwavering” (¢ταλαντoν, ´ cf. Empedocles fr. 17.19 DK), over the widely preferred “complete” or “perfect” (τšλειoν or τελεστ´oν); so [1, p. 176, n. 6], [19, p. 118], cf. [17]. Given the disputed sense of μoυνoγενšς (uniform or unique?) I simply transliterate; cf. [3, p. 71]. In 8.3 I take ™´oν as conditional, modifying the understood subject of ™στιν, i.e. any item on Road 1, which must adhere to GC as announced in fr. 2.3; cf. [7, p. 433] “insofar as it is”; alternatively, it could be pronominal, as in [2] and [16, p. 139, n. 3]. 9 The list presents a “programme” for what follows; see [13, pp. 76–77], [1, pp. 179–180], [16, pp. 137–140]. Their articulation correlates four signs in 8.3–4 with four stages of argument; but I see no need to join [5] and [16] in transposing 8.34–41 after 8.42–52. [11] assigns fourteen terms to

1 Truth Attending Persuasion: Forms of Argumentation in Parmenides

5

she repeatedly argues that any denial of these attributes is incompatible with GC (the Governing Constraint) or otherwise leads to contradiction. In effect, she proceeds by considering but rejecting a series of familiar and basic features of the world we perceive around us: its temporally and spatially limited, variously differentiated, continually shifting and changing contents, be they objects, situations, facts, attitudes, or anything else. Accordingly her language is thick with negation, privation, and contrariety. Thus, for the first sign, which she initially formulates as a correlative pair of privatives “ungenerated and unperishing” (8.3), she argues for ruling out their contraries “generation” and “perishing” for anything on Road 1. Consider her opening moves (fr. 8.5–11): Not (P1) once was it nor (P2) will it be, since (Q) now it is all at once One continually. For (R1) what birth will you seek for it? (R2) How and whence increased? (R3) Not out of not being shall I Allow you to declare or Noein; for not for declaring or Noesis Is it that it is not. (R4) What ordinance would have roused it Later or before to develop, starting from nothing? (C) In this way is it ordained either entirely to be or not.10

The first line translates the two contraries into verbs: if any being ever “perishes”, then it stops being, and so there is a time when it “was” and no longer is; and if any being is ever “generated”, then there is a time when it “will be” and is not yet.11 The goddess will reject both options on the ground that they violate GC (8.8–9 and 8.15–18). But to show how they violate it, she appeals first to Q: anything on Road 1 “is now all at once one continually” (8.5–6a). Formally this opening move involves a simple application of modus tollens: if anything admits either (P1) perishing or (P2) six groups and analyzes the arguments accordingly; but his grouping disrupts the clear articulation of fr. 8’s argument into four parallel stages; see [16, pp. 143–144]. 10 Steps in the argument are labeled systematically both for simpler reference in what follows, and to indicate corresponding factors in the arguments for each of the four “signs”. Thus, “P” designates the initial step in each argument, each of which introduces the attribute in question for that “sign”; “Q” designates the initial rejoinder to each P (all marked by “since”); “R” indicates a step in the subsequent reasoning; and “C” designates a conclusion of that reasoning as indicated explicitly by a verbal connective. As we shall see, it is not always clear how these steps are related to one another, whether for example C is to be seen as a basis for Q, or equivalent to it or to P; or whether a particular R is to be seen as a basis for C or for Q, or for another R. But the pattern is remarkably consistent, and plainly significant; cf. [16, pp. 143–144] calling P the “demonstrandum”, Q its “subsidiary programme” (an initial “rationale” indicating the grounds for P to be supplied by R), and C the “principal conclusion” (restating P). 11 Taking temporal πoτε with both verbs (obviously not with initial oÙδš: “not ever”) and restrictively: not “was once upon a time (and may still be)” or “will be (as it always has been)” but rather “sometime was (and no longer is)” and “sometime will be (but not yet)” as in 8.20; cf. [16, pp. 141–142]. So construed the argument is framed by a fourfold disjunction that is exclusive and exhaustive: beings on Road 1 have either a period of being before not being (perishing), or one after not being (generated), or only being, or only not being; but the fourth has already been ruled out in frs. 2–7; and appeal here to GC (the third option) now rules out the first two, thereby removing “is not” for all time and leaving the third option, only being.

6

S. White

generation, then it is not the case that “it is now all at once one continually” (not Q); but Q; so nothing on Road 1 can admit either perishing or generation (neither P1 or P2); and so neither option has any place there.12 But to establish Q, and thereby show how it rules out P1 and P2, she now appeals to GC, arguing that allowing “is not” for any time either past or future contravenes its requirement for “is” to hold universally. Parmenides has her proceed in two stages, targeting first generation (8.6–11), then perishing (8.12–15). Starting over with generation (P2), the argument follows the same elenctic pattern, now more elaborately. It again has two main targets, each framed by a rhetorical question challenging the target thesis P2. The goddess first asks (R1) “what birth” or origin (τ´ινα γšνναν) anything on Road 1 will have (“will you seek” echoes the epistemological orientation of 2.2); and supposing it has one, she then asks, in abrupt asyndeton reflecting the conceptual linkage, (R2) “how and whence” (πÍ π´oθεν) this “birth” might occur? I call these questions “rhetorical” because they presume a negative answer without stating one explicitly.13 Instead, the goddess appeals in R3 to GC as the basis for the answer she leaves implicit. Again she relies on modus tollens, twice over now: if (P2) something is generated, (R1) it must have some birth or origin (8.6); but if so, (R2) it must come “somehow from something” (8.7); but it doesn’t (answer to R2 as implied by GC); so it has no birth (answer to R1 as implied by GC); so contra P2, it is not generated. Nor is that all. Lest we wonder why we should answer R2 negatively here—after all, plenty of things do come “somehow from something”—she goes on to address its two questions separately and in reverse order, first (R2b) “whence” and then (R2a) “how”. So the twofold second question R2 is best taken as explicating the first: any (R1) birth for anything presupposes (R2a) some way or explanation for its birth (whether the Dative indicates manner, means, cause, or some other factor), and also (R2b) some source of its birth. The strategy is thus to show that any positive answer to either part of R2 violates GC; and so, given R1’s dependency on R2—no birth or origin without a way and a source—there can be no birth or origin, at least not for anything on Road 1. The goddess addresses the possibility of a source first (8.7b–9a). If something has a birth, presumably (R2b) it has a source (π´oθεν 8.7a); but if so, it comes “out of (its) not being” (™κ μη` ™´oντoς 8.7b); but given GC, (R3) “I will not allow you to say out of not being” (8.7b–8a); and so it has no source, and hence no “birth” either. She then turns to the other term in R2, arguing now that any “birth” would require further specification and explanation. If something has a birth, (R2a) it must do so “some way” (πÍ 8.7a); but if so, then (R4) “some need roused it later or sooner 12 Use of ™πε´ι marks the clause as affirming the antecedent of an implied conditional. Stoic logic later

called this a “paraconditional” (Diog. Laert. 7.71), presumably because it employs a conditional as an abbreviated argument (or strictly speaking, conflates the two, hence “para”), viz. their first indemonstrable, modus ponens. 13 Formally, a rhetorical question Q challenges a claim P by assuming “if P, then Q” and inviting the rejection of Q; cf. Aristotle, Rhet. 3.18. So here the question R1 “what birth will you find?” assumes that anything generated (P) has a birth (Q) and invites the denial of any birth.

1 Truth Attending Persuasion: Forms of Argumentation in Parmenides

7

to develop, starting from nothing” (8.9–10)14 ; but that—any way it is not—has just been disallowed by R3; and if not R4, then not R2a, and so not R1: it has no birth or origin. Both arguments articulate a series of presuppositions for generation and then rely on repeated applications of modus tollens to reject them all. At the pivotal point in the first argument, where the goddess forbids any “declaration or Noesis” of anything on Road 1 not being (R3: 8.7b–8a), she appeals explicitly to GC, in nearly identical ´ as the basis for this step: “for not for declaring or terms but marked explicitly by γαρ Noesis | is it that it is not” (8.8b–9a).15 And so, concluding on that basis—namely, R1-4—that generation is inconsistent with GC on multiple counts, she states her initial result: “In this way (oÛτως), it is ordained either entirely to be or (entirely) not (to be)” (8.11). The advent of necessity announced here by “ordained” (χρεων ´ ™στιν) is both apt and sharply circumscribed. It is apt because the elenchus has now shown that anything that only is, since its not being is now ruled out for any prior time, thereby must be for all prior time, and so it must be ungenerated. And this necessity is circumscribed by “in this way” (initial position, answering to initial “somehow” in 8.7) because the argument so far has established necessity in only this one dimension: for all prior time.16 It remains to be seen whether it holds also for all subsequent time (no perishing), everywhere (no division), and so on for each of the other signs. Only beings that also meet all those further conditions on what “is and only is” (2.3) are on Road 1 and available for Noesis. The first sign receives the most extensive treatment (seventeen lines), presumably in part to model at the outset of this revolutionary discourse the novel forms of argumentation integral to Road 1 and characteristic of Noesis as here envisioned.17 Indeed, the argument continues for another eleven lines, invoking GC again even more emphatically in 8.15–18. But I’ll bypass its further intricacies in order to indicate how the argumentation for the other signs follows largely the same elenctic strategy.

14 Or taking the several terms of R4 in turn: if somehow, then “later or sooner” (8.10a); and if so, then “some need roused” it to do so then (8.9b); and if so, then it “starts from nothing” (8.10b), which contravenes R3. 15 Adverbial Óπως (8.9) is commonly read as a conjunction introducing indirect discourse: “it is not to be declared that (it) is not”; cf. the formulation in fr. 2.3. But as an adverb of manner it yields GC: “not to be declared is any way in which (it) is not.” 16 Demonstrative oÛτως, commonly assigned inferential force here (“thus”), is better construed adverbially, modifying “is entirely” to specify the way in which any item on Road 1 “is entirely”— namely for all prior time, since ungenerated; so [1, p. 188], cf. [11, pp. 194–195]. Then modal χρεων ´ serves to flag the conclusion as logically necessary; on this key term, see [12, pp. 277–278], and for its logical role here, sec. 1.3 below. My translation reflects its association in the poem with divine norms (esp. fr. 1.28, cf. 1.32, 2.5, 8.54); how this relates specifically to logical necessity is a complex question I address elsewhere. 17 The disproportionate length might also reflect the stock of reasons available, sheer virtuosity, or any number of other factors; cf. Gorgias’ expansive treatment of the power of logos (8–15) alongside the other options on his agenda in the Encomium of Helen.

8

S. White

The second sign the goddess demonstrates is a total absence of any division (fr. 8.22–5).18 Nor (P) is it divided, since (Q) it is all alike: (R1) Nor is it any more this way that might restrain it from holding together, (R2) Nor is it any inferior, but (R3) it all is full of being. Thereby (C) it’s all held together, for (R4) being is next to being.

As with the first sign, the argument is again thoroughly elenctic, first announcing “divided” as its target for elimination, then showing how division violates the constraints of Road 1. We’ll look more closely at the argument in the next section. Here I’ll just highlight how it follows a very similar pattern. It first denies the opposite of the focal attribute in P, again on the basis of an initial rejoinder Q introduced by “since” (8.22 as in 8.5, cf. 8.27, 8.42). Then R1-2 deny a pair of contraries associated with its opposite (8.23–4 as in 8.6–7); and they do so by appeal to GC (R3) in the guise of “all full of being” (8.24 as in 8.7-9). A conclusion C finally formulates the attribute in question positively as “all held together” to affirm the opposite of the initial target (8.25 as in 8.11). By articulating distinct ways in which something might be “divided”—“somehow more” or “somehow lesser” or “not holding together” (8.23–4a)—and then rejecting them all on the basis of GC (8.24b), the goddess establishes the absence of division, at least for all three ways considered: any being on Road 1 is not more or lesser or untogether but “all alike” (8.22) and “all held together” (8.25). If the three alternatives she considers are exhaustive, her conclusion is secure universally. The third sign, initially characterized as “unshaken” (8.4), she now reintroduces in directly privative terms as not only “unchanging” (or “unmoving” 8.26) but, deploying paired contraries again, “unstarting, unceasing” (8.27), all again supported by an initial ™πε´ι-clause, which now invokes the first sign and its corresponding pair of contraries (fr. 8.26–33): But (P) unchanging in the limits of great bindings It is unstarting, unceasing, since (Q) generation and perishing Were banished far away, pushed away by true conviction. (R1) Staying the same in the same it lies by itself And (R2) this way steadfast it stays there; for (R3) mighty Necessity Holds it in its limit’s bindings, which restrains it all around. (C) Wherefore Right it is for what is to be not unfinished; For (R4) it is not needy, and being (so) it would need everything.19

In this case the contrary attribute targeted for elimination is embedded in the privative term used to designate the sign itself, and there is no independent mention of “change” (or motion). But change (strictly, not-P) is plainly the target of the 18 The privative “undivided” does not occur in extant lines; the initial list apparently uses two terms

to specify it (8.3): both “whole” (cf. 8.38) and “monogenous” (cf. n. 8 above). 19 For the text of fr. 8.33, excising μη ` before ™`oν δ’, see [16, p. 384].

1 Truth Attending Persuasion: Forms of Argumentation in Parmenides

9

argument that follows. Already the initial rejoinder Q in 8.27-8 relies on a missing premise to the effect that “if something changes, then it either starts or ceases” to be some way; but since any beings on Road 1 are “unstarting and unceasing” (8.27), they do not change. Again in the central argument R1-3, the inference relies on a similar unstated assumption articulating in positive terms three ways of not changing: if something changes, it is not “the same remaining in the same the same way” (8.29); but beings on Road 1 are the same in all three of these ways; so they do not change. Likewise for a final argument in R4: if something changes, it is needy; but anything on Road 1 “is not needy” (8.33), so it does not change. Three applications of modus tollens thus eliminate the target opposed to the third sign or attribute of anything available for Noesis. The argument for the fourth sign—being “perfected” or “fulfilled”—again follows the same elenctic strategy of articulating a series of opposed attributes, then ruling them all out, this time apparently on the basis of previous results already derived by GC but without any direct appeal to GC (8.42-9). But since (Q) it has a last limit, (P) it is perfected On all sides like the mass of a well-rounded ball, From the middle equal everyway; for (R1) not any greater Nor any smaller is it ordained to be either here or there. For (R2) nor is it not being, which might stop it reaching Its like, nor is it being in any way that it might be More of it here and less there, since (R3) it is all inviolate. For (R4) equal to itself on all sides it attains alike its limits.

The form of the argument, as marked explicitly by inferential particles, runs as follows: Since Q, P; for R1; for R2, since R3; for R4. How the several premises relate to one another—whether R1 supports Q or P, for example, or only some of the multiple predicates in P, and whether R2 supports all of R1 or only one of its conjuncts—is a vexed question to which I’ll return later. To recognize its elenctic form, however, it is enough to note how the bulk of the argument in R1-3 again articulates and rejects a series of possibilities opposed to the attribute in question: neither greater or smaller in R1 (8.44-5); then neither not being or being in R2 (8.46-8), with another negated disjunction embedded: neither more or less; and finally not “inviolate” in R3 (8.48). If these alternatives are also exhaustive, then their elimination secures the fourth sign and shows that beings on Road 1 must be “perfected” (8.42). But to determine whether they are exhaustive, or are at least meant to be so, would take us into semantics, and a host of contested questions outside our scope. The heavy reliance on indirect argument we’ve surveyed is hardly breaking news. But why these arguments should count as indirect proofs is not a question much discussed. And yet, if they are elenctic in the form familiar from Plato’s Socratic dialogues, then they neither prove or refute anything. For as [22] famously argued, Socratic elenchus is in the first instance only a test for consistency, showing at best only that a claim in question is inconsistent with a given set of other claims. Failing such a test does not by itself refute the claim in question, or establish the truth or

10

S. White

falsity of any claims in the set. To achieve those stronger results, an elenchus must also be sound, deploying exclusively true premises; only in such a case does a valid elenchus yield a proof or refutation of any of the claims involved. That is a burden here shouldered in part by GC, which underwrites crucial steps in each stage of the goddess’s “contentious elenchus” by demarcating the special domain to which it applies.

1.2 Missing Premises It is unlikely to have escaped notice that all four arguments, as I’ve reconstructed them, rely on steps not explicitly stated in the text. Or to put the same point a little differently, the goddess again and again appears (in my reconstructions at least) to rely on missing premises. To address this worry we shall look more closely at the treatment of the second sign, the shortest of the lot but sufficient to illustrate the problem and suggest a resolution. The argument is dense, packing eight clauses into its four hexameters. A schematic formulation can indicate the surface structure and frame the questions to address (fr. 8.22–5): Not P, since Q; R1, and R2, and R3; hence C, since R4.

So construed, it becomes clear that the argument proceeds mainly backwards, first stating its conclusion, then supplying reasons for it, and similarly again in the closing line. But the logical relation between the first line (8.22) and the sequel (8.23–5) is unclear on three counts. First, what are R1-3 meant to support or explain: the denial of P, the initial ground Q, the conclusion C, or some combination?20 Second, what relation does C have to P and Q, and to R1-3? And third, what role does R4 play, not only for C but for the preceding claims (R1-3, Q, and P)? To start with R1-3, the first two are clearly paired not only by parallel phrasing and metrical position, but also semantically, each isolating one of a pair of opposites “more” and “lesser”.21 As comparatives, moreover, the pair forms an exclusive disjunction that if joined by the corresponding intermediate term would be exhaustive: either more or less or equal.22 An attractive possibility is that R3 (“all full”) fills exactly that gap. If so, then the two central lines (8.23–4) amount to a simple 20 Initial oÙδš in 8.23 (repeated in 8.24) supplies the requisite syntactic connective but leaves unstated any logical connection R1-2 has to P or Q in 8.22. 21 Why χειρ´ oτερoν rather than metrically equivalent βαι´oτερoν (same position in 8.45) or the more common ¼σσων (with μαλλoν ˜ in 8.48)? Perhaps to signal a wider scope for this sign, including qualitative “differentiation” as well as spatial “division”; cf. later use of the term “division” for methods of conceptual analysis, notably in Plato’s Eleatic dialogues. 22 The goddess follows the same strategy in frs. 2–7: fr. 2 announces a pair of contrary options in parallel phrasing (Roads 1 and 2), one of which it promptly rejects; fr. 6 then introduces a third option (Road 3) that occupies the space between the initial pair; its elimination then leaves only the first, if the three alternatives are exhaustive—as they seem to be: being in every respect, not at all, and in only some respects.

1 Truth Attending Persuasion: Forms of Argumentation in Parmenides

11

argument by elimination, ruling out two of the three options to leave the third. And they do so by assuming the full disjunction, which is left unstated. But what basis is given for the elimination? To reject the first option, the goddess casts R1 as a conditional with counterfactual force: if (a) any item on Road 1 were more in any way, then (b) that would prevent it from holding together. But the initial “nor” negates the entire conditional, which implies, contra (b), that nothing does prevent that; and so, contra (a), the item isn’t more in any way. R1 thus amounts to another application of modus tollens: If (a) some being is “more in some way”, then (b) that prevents it from holding together; but (R1b) nothing prevents it; so (R1a) it is not more at all. As it stands, that leaves R2 in the next line unsupported. But the parallel phrasing suggests that it too relies on R1b for a parallel conditional in a parallel modus tollens: If (a) some being were “less in some way”, (b) that would prevent it holding together; but nothing does so; so (R2) it is not less in any way. In that case, R1b rules out both R1 and R2, on the ground that “holding together” is incompatible with any difference in either “more” or “less”—whether degree or intensity or anything else. And that leaves only the medial option expressed in R3, which it appears the goddess considers logically equivalent to the denial of the possibility envisioned in R1b of not holding together. So much for the argument’s structure. Why should denying R1b support the elimination of R1 and R2? Not only does the goddess take “holding together” to rule out being “more or less”; evidently she also takes it to be equivalent to R3 “being all full of being” (8.24). How that equivalence might be plausibly interpreted is a question we can defer, thanks to our focus on argumentation.23 But the recurrence of “held together” in the following line (8.25) raises more serious formal questions. There C deploys the same term as R1b, apparently affirmed now as an inference (“thereby”) from either R3 alone or the full argument R1-3; but since both rest on R1b, that threatens to make the entire sequence circular. One simple remedy is semantic: if “holding together” is used in different senses in R1 and C, circularity vanishes—leaving a fallacy of equivocation instead. Or perhaps R1-3 are not meant to be exhaustive after all; but removing the argument by elimination leaves nothing in its place. Should we then bite the bullet and accept circularity? Here some might appeal to an isolated remark by the goddess recorded elsewhere, that “it is equal for me | wherever I start, for there I shall arrive back again” (fr. 5). But that smacks of desperation, and a more charitable and credible solution may emerge if we pursue my other two questions about the role of C and of R4. Our problem arises from the assumption that C is supposed to follow directly from R1-3. But in that case, we might wonder why the goddess appends R4, which is ´ as somehow supporting C. Taking 8.25 by itself, R4 looks plainly marked (by γαρ) like a distinct point drawing a connection between its two predicates, in effect another implied conditional: if (R4) A is “next to” B, then (C) A and B are “held together” or “continuous”. As we’ve seen, the goddess frequently relies on unstated conditionals. The novelty here is her use of modus ponens, affirming the antecedent R4 to infer C: 23 Whether literally, for example, by spatial continuity or volumetric repletion, or by some more abstract sort of unity or uniformity (cf. n. 21 above).

12

S. White

since being is “next to” being, being does “hold together” continuously. In this case, moreover, the implied conditional expresses what was later considered an analytical connection: juxtaposition implies continuity; or in the concrete terms the goddess uses here, if A is “next to” B, then they “hold together” as connected parts of one and the same continuous whole.24 Turning to C, supposing it is meant to follow only from R4 and not R1-3, what role does it play in the larger argument? Is it meant to support Q, and thereby rule out P, or rather to rule out P directly? If the latter, then Q ends up otiose, either a simple paraphrase of C or distinct but logically equivalent; and it’s not at all clear that “holding together” (or continuity) either entails or is equivalent to “likeness”—even allowing that in archaic usage “likeness” may convey similarity in either magnitude or quality.25 Alternatively, if C is meant to support Q, and only thereby the elimination of P, then it still seems otiose, equivalent now to R1-3; nor is it any more clear how continuity might ground likeness directly. Have we evaded circularity only to reach a dead end? Fortunately another option remains. Suppose we turn the alleged circularity on its head, so that C grounds the crucial step in R1-3, namely, the initial appeal to continuity in R1b. In that case the argument runs as follows (schematized to highlight the predicates involved): 1. 2. 3. 4. 5.

If Next, then Continuous; Next; so Continuous. (since R4, C) If More or Less, then not Continuous; (1); so neither More or Less. (if C, R1-2) Either More or Less or Full; (2); so Full. (if R1-2, R3) If Full, then Like; (3); so Like. (if R3, Q) If Like, then not Divided; (4); so not Divided. (since Q, not P)

Or taking the sequence in reverse as the goddess does in presenting them: Q (8.22) first specifies the way in which anything on Road 1 must be undivided, namely, by being “all alike”—where likeness may include both quantity and quality. Then R13 (8.23–4) specify in turn the way in which anything must be “all alike”, namely by being “all full”; and as R1b shows, that in turn depends on being “continuous”; which finally hinges on R4 (8.25), being “all next to”—where “next” may range across categories, including location, quantity, quality, and others. The result is a set of “notionally equivalent” factors or features. But that result is achieved and supported by a chain of deductions in a series of five simple syllogisms, deploying modus ponens, then tollens, then disjunctive elimination, then modus ponens twice again, as articulated in steps 1–5 above.

24 Present πελαζει ´ is commonly taken transiently as “draws near”; but the stative sense of its cognate

adverb πšλας is plainly more apt for Road 1; cf. “neighbors” in [1, p. 211]. In epic usage, the verb indicates close or immediate proximity and appears almost always in the aorist, implying arrival; the only present indicates abstract inherence (Athena inflicting pains on Ares in Iliad 5.766). If the present has similar force here, where an aorist would undercut the very argument against change in which it appears, then the conditional “if R4, C” has solid footing. 25 Proclus, On Euclid Elements 1.5 citing Eudemus (fr. 140 Stork) on Thales (A20 DK); fr. 8 has both „σoπαλšς (8.44) and σoν (8.49) alongside Ðμ´oν (8.47).

1 Truth Attending Persuasion: Forms of Argumentation in Parmenides

13

So runs the argument for the second sign; or rather, so goes a reconstruction that renders it valid by supplying the requisite missing premises. For at every step, the initial premise is left unstated. In the first and last steps (in the last and first lines), the presence of an inferential (or explanatory) particle mitigates the omission: the use of ´ in step 1 (8.25) and ™πε´ι in step 5 (8.22). In another, parallelism and connective γαρ particles perform the same function: paired correlatives with δš (and oÙδš twice) in step 3 (8.23–4). And the potential relative clause in step 2 implies a conditional logically equivalent to the requisite premise there. Only in one case is no explicit basis given for the initial premise: the connection between being “full” (R3) and being “alike” (Q) in step 4, where the minor premise enters well after the inference it supports, separated by three intervening clauses that follow it in asyndeton (between 8.22 and 8.23–4). By the strict rules of Aristotelian or Stoic syllogistic, the omissions represent logical flaws. Elsewhere, however, the omission of implied premises (at least plainly implicit ones) is standard, and not only in rhetoric, where Aristotle makes the practice characteristic of enthymemes. It is pervasive in Plato’s Socratic dialogues, and likewise his most Eleatic dialogue, the Parmenides, especially its tortuous second half. At the dawn of deductive argument as we know it, it is not the omission of obvious steps, least of all conditionals, but their inclusion that would be surprising. There may be another, more pointed rationale too. Recall the goddess’ injunction to use reasoning in assessing her discourse (7.5): we’re directed to think for ourselves, not to listen raptly as she spells out every connection fully and explicitly step by step. After all, she offers to enable her listeners to think like herself and other gods: not to swallow whole whatever she says as self-evident revelation, but to engage in a revolutionary way of thinking never before articulated by mortals. Two more observations. First, being more or less in any way (R1-2) implies not being the same in every way. But Road 1 rules that out entirely: there is no room for anything that “is not” the way it is. So if the goddess relies on GC in R1-2, or if she appeals to it implicitly in R1b—if not “holding together” amounts to or implies some way of not being—then the argument as reconstructed here turns yet again on GC. On the other hand, it is tempting to read R3 itself as a paraphrase of GC: for anything to “be full of (its) being” might serve to indicate in vivid terms that it only is, and in every way. In that case GC figures explicitly in the argument, but at the price now of depending on other premises, four in fact: R4, C, and R1-2 (in logical order). Alternatively, and this brings me to my second point, Parmenides might mean to present two related but distinct arguments here: one based on GC as formulated in R3, which grounds Q by eliminating R1-2 directly; and then a second, based on R4, which rules out P via C rather than Q. That would put us back where we started, with a certain redundancy: two separate arguments against P, each now operating independently, via Q or C. Moreover, if R4 also invokes GC as I suggested R3 does, then both arguments are firmly grounded. But then this alternative is back on even footing with my initial reconstruction in steps 1–5; since it too depends ultimately on R4, it too is now securely grounded.

14

S. White

1.3 Signs of Necessity The argumentation of B8 is predominately indirect and elenctic, and it trades heavily on unstated premises, especially implied conditionals. In fact a paucity of explicit conditionals invites the question whether the few we do find might carry special significance. At only one point in all 150 extant lines does the standard conditional marker “if” appear: in the closing argument for the first sign, twice in a single line, both with indicative (aorist and future) in mixed simple conditions: “for if it came about, it is not, nor is it if it ever is going to be” (8.20). Another marker, however, appears eight times in the extant lines, all but once in the deductive sequence of 8.5–49: at least one potential optative figures in the arguments for each of the four signs.26 In every case, moreover, the implicit conditionals in which these optatives occur are negated either explicitly or implicitly as rhetorical questions. In other words, every occurrence takes the form of a negated possibility: for example, in R1 of the argument for the second sign, “nor is it any more any way that might restrain it from holding together” (8.23). As we saw, that comes to something like this: it is not the case that if it is more some way, it would hold together. The main point can be expressed in three steps as an application of modus tollens. But while that preserves universal quantification for the negated conditional (“there is no way”), it fails to capture the special force of the optative, which adds a modal dimension to these otherwise strictly factual claims. As things stand, then, the goddess appears to deny the very possibility of any way any being on Road 1 either could be more or less in some respect or could fail to hold together. Or conversely, what she affirms is that these beings are necessarily the way they are. But where is this necessity supposed to come from? Does she offer any basis for it, or simply conjure it out of thin air? Another look at the sign terminology points to a solution. For the second sign, scholars disagree on whether to read the verbal adjective διαιρετ´oν as “divisible” or simply “divided.”27 Although later usage might favor modal force, given that participles are available for a strictly descriptive sense, two factors tell otherwise here. One is simply Parmenides’ own usage. In the initial list of signs (8.3–4) only one of five terms is a verbal adjective (¢γšνητoν); and it never appears again in the extant verses, not even in the argument for the first sign, where we find only the corresponding verb (four times: present and aorist infinitives in 13, aorist indicative and optative in 8.19) and two nouns (8.6 and 8.21).28 Not only is every other sign 26 The

single exception proves the rule: in 2.7, in what [23] calls the “governing deduction” that eliminates Road 2 and thereby underpins Road 1 and GC. 27 [13] and [2] assign it modal force, and [11] does so tentatively; [1], [18], [16], and [23] settle for description. 28 Likewise for its correlative ¢νωλεθρoν ´ (8.3): present infinitive (8.14), and noun (8.21, cf. 27). I count eleven verbal adjectives in the extant lines: six are plainly descriptive (πλαγκτ`oν 6.6, ¢γšνητoν 8.3, ¥παυστoν alongside ¥ναρχoν 8.27, ¢τελετητoν ´ 8.32, ¢κρητoιo ´ 12.1, πoλυπλαγκτων ´ 16.1), one probably so (¥πυστoς 8.21, alongside ¢πšσβεσται), and four probably modal (¢νυστ´oν 2.7, ϕατ`oν and νoητ´oν 8.8, ¢ν´oητoν 8.17). The first seven all refer to items on Road 1, and the four modal instances all occur in epistemic contexts describing what mortals can know or tell. The modality in question thus appears to be narrow and precise: twice deontic

1 Truth Attending Persuasion: Forms of Argumentation in Parmenides

15

introduced in descriptive terms; only the third (¢τρεμšς) is ever characterized in ´ similar terms (¢κ´ινητoν 8.26). In fact, the fourth (¢ταλαντoν or alternatives; cf. n. 8 above) is designated without a whiff of modal force simply by a perfect participle (τετελεσμšνoν 8.42). Moreover, and this is the other factor, potential optatives typically signify specifically epistemic possibility, marking the uncertainty of a prospect considered unlikely but possible (hence the standard English label “future less vivid”): what might happen, so far as we now know, but marked as a remote possibility. For that very reason, however, the denial of a potential optative carries special epistemic weight, since it rules out even remote possibilities.29 Moreover, its counterfactual force is strictly extensional, limited to universality: nothing of the sort will occur.30 In fact, quantification is marked explicitly in nearly every step of the argument for the second sign: explicitly “all” in three steps (Q, R3, C), and explicitly “none” in two others (R1-2). Only R4 has no explicit quantifier, and like R3 it invokes being itself, in fact twice over. This double usage, I suggest, is again designed to evoke GC: just as “being is all full of being” in R3 is tantamount to the complete elimination of “is not”, so is “being is next to being” in R4. Both factors—use of descriptive rather than modal terms, and negated potential optatives—are also more in line with GC and the thrust of the larger argument, which rules out a series of ordinary attributes universally: first no generation or perishing at any time, then no division or differentiation in any way, no change or motion in any respect, and finally no irregularity at all. But since this universality applies only to ascription of these attributes, not to any attribute in its own right, the terms of the conclusions remain descriptive too. At least, that is all that the argumentation can establish on its own. To be sure, if the premises do have independent modal force, as some may, then any necessity they have might also be preserved by valid inferences and transmitted to the conclusions. But that raises questions beyond our scope here. The crucial point for present purposes is simply that any necessity in the conclusions depends in the first instance on the form of argumentation, namely, on the necessity of logical consequence. In that respect, at least, the arguments of the goddess are designed to preserve whatever modal status her premises may have, and the only explicit necessity ascribed to any of her conclusions is logical.

(in a single prohibition: “not to be said or pondered”, following “I won’t allow” in 8.7) and twice predictive (failure on Road 2). 29 “Denial” here requires external negation like 8.23 (“it’s not the case it might”), not internal negation like “it might not occur.” 30 Contrast the narrow scope of “contrary-to-fact” claims about particular situations present or past, expressed in the indicative with (¥ν or κε), and the broad scope and counterfactual force of generalizations expressed with subjunctive antecedents (and ¥ν or κε; or optative with past tenses). In this light, the total absence of subjunctive conditionals (the only extant subjunctive, in 8.61, is purposive) may itself reflect an extensional focus.

16

S. White

Consider the language of the argument for the third sign, where the goddess twice invokes personified modalities: Necessity in R3 and Themis or Right in C (8.30–2).31 And (R2) this way steadfast it stays there; for (R3) mighty Necessity Holds it in the bindings of its boundary, which constrains it all around. (C) Wherefore Right it is for what is to be not unfinished.

Here “mighty Necessity” represents an embodiment of GC (cf. “ordained” 11, “Jus´ tice” 14, “necessity” 16 for the first sign) and serves as the basis for R2 in a γαρ-clause: anything bound within a boundary (R3) is held “steadfast” where it is (R2); and what imposes the binding is GC, the “necessity” on Road 1 of having only “is” and no “is not” in any way or respect, as stipulated by the goddess initially (2.3). The other modality figured here, by contrast, represents inferential necessity itself, as indicated by the connective “wherefore” (oÛνεκεν) that announces C as the conclusion of this stretch of argument. Any beings on Road 1 (here τ`o ™´oν), whatever else they may be, must be “not unfinished”, not simply because of the Necessity invoked in the preceding lines, but in the first instance because of the several factors articulated in the surrounding argument that make it “Right” for any being on Road 1 to be so. Only in conjunction with those premises does the Necessity of GC asserted in R3 provide a basis for the inferential “wherefore” that introduces Themis in 32 as the logical personification that makes the conclusion C “Right”. The treatment of the fourth sign shows the same pairing of logical consequence and universality still more clearly, asserting necessity once and invoking potential optatives twice. The train of thought is convoluted, and controversial, though not in ways that affect the present issue.32 The sequence of steps runs as follows, again highlighting logical connectives (8.42–9): 1. 2. 3. 4. 5. 6.

Since (Q) it has a “last boundary”, (P) it is Perfected “like a well-rounded ball.” (8.42–3) For (R1) it must not be either Greater at all or Smaller at all. (8.44–5) For (R2) neither is it not some way that would stop it being Alike; (8.46–7) nor is it some way that it would be More or Less than it is, (8.47–8) since (R3) it is all inviolate. (8.48) For (R4) anything Equal on all sides is Alike in its boundaries. (8.49)

As with the second sign, here again the argument as a whole proceeds in reverse: Q provides a preliminary ground for the attribute affirmed in P, which the sequence R1-4 then underwrites by ruling out a series of contrary alternatives. The precise logical connections among R1-4 are obscure and disputed but need not concern us

31 Personified in the first instance as dramatic figures or verbal agents, as Necessity here, and as the goddess Themis or “Right” in fr. 1.28; and even where “Right” is syntactically impersonal, as here, the name has special resonance, evoking the distinctive authority of Themis. 32 See [11, pp. 211–212]; cf. [1, pp. 202–204] and [16, pp. 157–158], though I cannot follow [16] in taking R3 as the ground for R4 by accepting asyndeton with ™πε´ι in 8.48 (contrast the initial αÙταρ ` ™πε´ι in 8.42).

1 Truth Attending Persuasion: Forms of Argumentation in Parmenides

17

here.33 What deserves emphasis is the explicit invocation of the special modal force found in GC by the reference in R1 to what is “ordained” (step 2, cf. 8.11), which R2 then backs up in steps 3 and 4 with a pair of negated optatives for another pair of contrary alternatives (embedding still another correlative pair in 4). Both optatives assert universality, denying there is any way in which anything on Road 1 fails to be whatever way it is, either by failing to be that way, or by being more or less that way. But why is R1 explicitly assigned the modality of being “ordained”? If either it or its underwriters in R2 invoke an earlier result from Sign 2, as they appear to do (8.44– 5 and 8.48 echo “neither more nor lesser” from 8.23–4), then far from conjuring necessity out of aether, the goddess simply indicates the status of R1 as already “proven”: designating it a valid consequence of her argument for the second sign. In that case, here again there is no whiff of necessity in the predicates themselves, only in her argumentation and its twofold reliance on GC and the logical necessity of the consequences she draws from it. I have argued that Parmenides attributes necessity to two related but distinct factors in the goddess’ arguments in fr. 8: first to GC itself, and then as she proceeds, to its consequences as established by her novel form of reasoning. The latter depend on a kind of logical necessity, the necessity of consequence. Any additional necessity in the consequents she derives must depend on the predicates her arguments deploy, or GC itself, or both, which raise issues beyond our present focus on argumentation. To be sure, both the soundness of the poem’s arguments and the modal status of their conclusions depend on the truth and modality of their premises. But what deserves emphasis here is that their validity depends strictly on adherence to canons of logical truth and deductive reasoning—a revolutionary standard that the poem proudly celebrates at pivotal points by invoking the divine names of Necessity, Right, Justice, and what the goddess herself declares “ordained” (χρεων). ´ This extraordinary power to transmit and preserve truth through a train of inferences from premises to conclusions, rare as examples must have been in the era of its discovery, would readily inspire intimations of divinity. Its associated powers of persuasion— of securing conviction—were evidently recognized too. No sooner does the goddess announce the stringent requirements of GC to characterize Road 1 (fr. 2.3) than she promptly proclaims this “the path of Persuasion, for it [sc. she] attends truth” (2.4, cf. 1.29).34 Deductive argument can certainly be convincing, even cogent or compelling. Yet as noted previously in connection with indirect proof (sec. 1), valid arguments ensure only logical consistency; actual proof requires true premises as well. Otherwise there is no guarantee—no logical necessity—that the converse will hold for the path of persuasion, with truth attending it in turn. key source of obscurity is the scope of the connectives ™πε´ι and γαρ ´ in R3-4: does R3 support all of R2, or only step 4, or only one of its contrary options; and does R4 support all of R3, or only step 4, or simply R2 or R1 by recapitulating intervening results? The ramification of R2 into three terms also obscures how its pair of negated conjuncts (steps 3 and 4) is meant to support R1, singly or jointly. 34 Reading dative ¢ληθε´ι; see [16, p. 380]. The subject of the verb “attends” is plainly Persuasion, personified as the goddess Peitho in the role of a divine attendant; cf. Hesiod, Works and Days 73 and Theogony 79–80, and see [12, pp. 136–163]. 33 A

18

S. White

The poem’s own solution to this impasse hinges on GC, which stipulates at the outset that anything under discussion on Road 1 not only is but only is (whatever way it is). Given that guarantee, then truth is sure to attend persuasion—provided the mode of persuasion employed is equally sure in turn to preserve truth. And that is effectively what the goddess promises in the proclamation she adds in 2.4 as a pendant to GC in 2.3, that “persuasion attends truth.” The two lines thus articulate a complementary pair of claims that frame the agenda for fr. 8 by specifying both the domain that Road 1 travels and its requisite mode of transit: first a realm of truth occupied by what only is—forever, undivided, unchanging, and perfect, as it turns out—and then a form of reasoning called “the path of persuasion” that ensures the preservation of truth as its constant “attendant.” The restriction to what is and its attendant requirement of persuasive reasoning together mark the two factors required for sound as well as valid argument, and their union ensures that Road 1 thereby remains also a “path of truth” from start to finish, wherever its travelers may stop. The first factor naturally receives pride of place, introduced first, the basis for distinguishing other roads, the main object of investigation in fr. 8, and arguably its primary objective. The other factor receives only passing mention and no sustained discussion anywhere in the surviving lines. Accordingly, the image of persuasion as an “attendant” of truth privileges the latter. Yet the same image aptly characterizes this special mode of persuasion in terms that foreshadow its pervasive presence in the goddess’ discussion of Road 1 in fr. 8, “attending” her argument throughout as its logical form. Moreover, we are clearly expected to register the force of this presence, as her invocation of various divine names for fundamental norms and constraints intermittently reminds us. Even as she continually directs our attention to the “signs” of what only is, the vehicle of her argument exerts its influence. My title, by inverting her account of its role, highlights the extraordinary ambition of the endeavor she proposes: the project of combining GC with valid reasoning to achieve nothing but truth. It is all the more remarkable, then, that the poem nonetheless prefers to call Road 1 “the path of persuasion”: highlighting neither truth nor being as its starting point or domain or destination, but the mode of transit its investigation requires. Whether by humility, prescience, or otherwise, the poem’s formulation may be more apt. Whatever else the dense chain of arguments in fr. 8 may be saying, its single most salient feature is surely the sheer power of the kind of argumentation it deploys, the likes of which mortal ears had never previously heard.35

35 I

thank the participants and organizers of the Berlin conference for lively discussion, likewise everyone at the 12th UNAM-UT Encuentro, Raymundo Morado for an illuminating response, Jonathan Beere for helpful comments, and especially Matt Evans for invigorating criticism and the stimulus of his own work on Parmenides.

1 Truth Attending Persuasion: Forms of Argumentation in Parmenides

19

Works Cited 1. Barnes, J. (1982). The presocratic philosophers. Routledge and Kegan Paul. 2. Coxon, A. H. (2009). The fragments of Parmenides: A critical text with introduction and translation, the ancient testimonia and a commentary, revised ed. R. McKirahan. Las Vegas: Parmenides Publishing (original edition, Van Gorcum, 1986). 3. Curd, P. (1998). The legacy of Parmenides. Princeton University Press. 4. Curd, P., & Graham, D. (Eds.). (2008). Oxford handbook of presocratic philosophy. Oxford University Press. 5. Ebert, T. (1989). Wo beginnt der Weg der Doxa? Eine Textumstellung im Fragment 8 des Parmenides. Phronesis, 34, 121–138. 6. Lesher, J. (1984). Parmenides’ critique of thinking: The poludêris elenchos of fragment 7. Oxford Studies in Ancient Philosophy, 2, 1–30. 7. Lesher, J. (2008). The humanizing of knowledge in presocratic thought. In [4], 458–84. 8. Lloyd, G. E. R. (1990). Demystifying mentalities. Cambridge University Press. 9. Lloyd, G. E. R. (2003). Demonstration and the idea of science. In J. Brunschwig, G. E. R. Lloyd, & P. Pellegrin (Eds.), The Greek pursuit of knowledge (C. Porter et al., Trans.), 92–117. Harvard University Press. 10. Long, A. A. (Ed.). (1999). Cambridge companion to early greek philosophy. Cambridge University Press. 11. McKirahan, R. (2008). Signs and arguments in Parmenides B8. In [4], 189–229. 12. Mourelatos, A. P. D. (1970). The route of Parmenides. Yale University Press; reprinted with additions Parmenides Publishing, 2008. 13. Owen, G. E. L. (1960). Eleatic questions. Classical Quarterly, 10: 84–102; reprinted in Owen, Logic, science and dialectic: Collected papers in Greek philosophy (Cornell University Press, 1986), 3–26. 14. Palmer, J. (1999). Plato’s reception of parmenides. Oxford University Press. 15. Palmer, J. (2008). Classical representations and uses of the presocratics. In [4], 530–554. 16. Palmer, J. (2009). Parmenides and presocratic philosophy. Oxford University Press. 17. Reilly, T. J. (1976). Parmenides, fragment 8.4: A correction. Archiv für Geschichte der Philosophie, 58, 57. 18. Schofield, M. (1983). Parmenides of Elea. In G. S. Kirk, J. E. Raven, & M. Schofield (Eds.), The presocratic philosophers, 239–262. Cambridge University Press. 19. Sedley, D. S. (1999). Parmenides and Melissus. In [10], 113–133. 20. Smyth, H. W. (1956). Greek grammar, revised ed. G. Messing. Harvard University Press. 21. Stork, P. (forthcoming). Eudemus of Rhodes: The sources, text and translation. In T. Dorandi & E. Spinelli (Eds.), Eudemus of Rhodes: Text, translation, and discussion. Routledge. 22. Vlastos, G. (1983). The Socratic Elenchus. Oxford studies in ancient philosophy 1: 27–58; revised in Vlastos, Socratic studies (Cambridge University Press, 1994), 1–37. 23. Wedin, M. (2014). Parmenides’ grand deduction: A logical reconstruction of the way of truth. Oxford University Press.

Chapter 2

Argumentation and Persuasion in Classical Chinese Literature Lisa Indraccolo

This contribution aims at providing an overview of the characteristics of the two main Classical Chinese rhetorical techniques of “argumentation” (biàn 辯) and “persuasion” (shuì 說),1 which, according to the received literature, were practiced and performed in polemical debates by early Chinese “wandering persuaders” (yóushuì 游說) allegedly active during the Warring States period (475–221 B.C.). In my inquiry, I will first briefly illustrate the historical and socio-cultural background that led to the emergence of the phenomenon of wandering persuaders and the concomitant flourishing of rhetoric in late pre-imperial and early imperial China (ca. 4th c. B.C.–2nd c. A.D.). I will then proceed to analyze in detail the two rhetorical techniques that were mostly employed by these persuaders in politico-philosophical debates, or at least, what we could tentatively identify as the Classical Chinese counterparts of these techniques, making them accessible to non-sinologists. I will examine the main formal, structural, and grammatical characteristics of selected cases of “argumentation” and “persuasion” in representative examples of dialogues 1 *I would like to thank Wolfgang Behr and James Weaver for their insightful comments on previous

drafts of this article. This article is based on the paper I presented at the conference “Argumentation in Classical Antiquity: dialectic, rhetoric & other domains,” Humboldt University of Berlin, 23–25 June 2016. For an introduction to early Chinese rhetoric and the techniques of “argumentation” and “persuasion” in early Chinese politico-philosophical discourse, see [75]; [20] and [21]; [81]; [119]; [37]; [41] and [42]; [89]; [90]; [48] and [49]; [31]; [52]; [60]. 2 While the study of transition terms is a rather well attested practice in Classical and Biblical studies, a systematic and exhaustive analysis and classification of this kind of markers in early Chinese literature still remains a desideratum. A few promising articles have been published especially on this specific use of the particle fú 夫, which is otherwise commonly employed as a topic marker at the very beginning of a sentence, see [32]; [141]. On this topic, see also [39, esp. pp. 33, 39, 44], [2, p. 237], and [106]. L. Indraccolo (B) Institute of Humanities, Tallinn University, Tallinn, Estonia e-mail: [email protected] © Springer Nature Switzerland AG 2021 J. A. Bjelde et al. (eds.), Essays on Argumentation in Antiquity, Argumentation Library 39, https://doi.org/10.1007/978-3-030-70817-7_2

21

22

L. Indraccolo

taken from pre-imperial and early imperial received literature in Classical Chinese. Particular attention will be paid to the use of intertextual references and authoritative quotes drawn from the shared cultural lore and from those rather loose sources that were later systematized in the Confucian Canon; (pseudo-)historical allusions; different kinds of structural and rhetorical devices (i.e. more or less complex cases of parallelism, cross-references, analogies, metaphors etc.); tropes and arguments for discussion; and transition terms or “linkers.”2 The timeframe of the present study corresponds by and large to the last centuries of the Zh¯ou 周 Dynasty (ca. 1045–221 B.C.). The period of reign of the Zh¯ou ruling house is conventionally divided into two main subperiods according to the location of the capital, the Western Zh¯ou period (ca. 1045–771 B.C.) and the Eastern Zh¯ou period (771–221 B.C.). The transition between the Western and Eastern Zh¯ou periods is clear-cut. It is marked by the sack of the Western Zh¯ou capital Hào 鎬 at the hands of a “barbarian tribe” from the northwestern steppe, the Quˇanróng 犬戎, and the consequent loss of a large chunk of territory, which eventually forced the Zh¯ou to move the capital eastwards. The Eastern Zh¯ou period is further divided into the Spring and Autumn period (Ch¯unqi¯u 春秋, 770–476 B.C.) and the Warring States period (Zhànguó 戰國, 475–221 B.C.). In this paper, I discuss phenomena that directly affect or have emerged in this latter period. As the name suggests, the Warring States period is a time of political instability and ubiquitous warfare. It is characterized by the fragmentation of the central political power, as the legitimate Zh¯ou Dynasty is gradually divested of its political and military authority by local rulers that were previously in a vassal-like relationship to the ruling house. These rebellious local rulers or “feudal”3 lords are mostly linked to the Zh¯ou ruling house by ties of blood, and reign over a constellation of satellite states that were formerly under the direct political, religious, and cultural influence of the Zh¯ou kingdom. The preeminence of the Zh¯ou lineage had primarily been determined by the royal privilege of having a preferential access to communication with the divine and spiritual realms of existence, and especially with the supreme deity (‘Heaven,’ Ti¯an 天) and the ancestral spirits. Such communication channels were strictly regulated through a complex system of rites and sacrifices.4 Despite the challenging socio-political and historical circumstances, the Warring States period can be acknowledged as the epoch of the maximum flourishing of early Chinese politico-philosophical debate, and it is traditionally considered as the Golden Age of Classical Chinese rhetoric.5 The Warring States period is not only an age of 3 The

use of the term “feudal” applied to the ancient Chinese Warring States period should be understood in a fairly loose manner. It has known a progressive decline—and for very well-founded reasons—in most recent sinological studies, as it is considered somewhat improper and culturally patronizing in the light of the substantial differences existing between the early Chinese and the Western Medieval historical contexts. For a more transparent and updated historical overview of the period, see for instance [87]; [146]. 4 For a detailed account of the Zh¯ ou ritual, sacrificial and kinship system, see for instance [123]; [147]; [6]; [140]. On rites and state organization, see [30]. 5 For a problematization of the legitimacy and the limits of the use of the term “rhetoric” as applied to a non-Western context in general, and to the Chinese specific case in particular, see [55] and [61].

2 Argumentation and Persuasion in Classical Chinese Literature

23

political and social turmoil. It is also a time of intense intellectual activity. While the Zh¯ou kingdom inexorably falls apart, new more or less powerful hegemonic strongholds are established at the local level. Concomitantly, increased physical mobility6 across borders leads to an unprecedented exchange of ideas—an almost unique conjuncture in the history of Chinese culture. The situation is fueled by the development of patronage, exerted by local rulers over their courtiers, and the consequent spread of the phenomenon of “wandering persuaders.” Local sovereigns are, blatantly put, usurpers—and they are fully aware of their questionable status. As such, they are constantly looking for potential sources of legitimization to consolidate their authority and power. At the same time, they are eager to acquire any form of knowledge or technique that might contribute to strengthen their newly established, and still somewhat fragile independent kingdoms, constantly threatened by attempts at conquest and annexation coming from belligerent neighboring states. In order to weave their own glorious narrative, local rulers engage in the practice of patronage, which has two main functions. On the one hand, it is a matter of prestige to attract and be able to support a large number of guests and retainers. On the other, although these people are not necessarily involved directly in the government of the state, the most capable are often employed as officials, sent as envoys on diplomatic missions, or act as advisers to the throne. The local ruler’s interest in wandering persuaders is thus strongly motivated by political pragmatism: sovereigns welcome anyone who can bring useful expertise in governmental practice and who provides efficacious advice and strategies to preserve and enlarge a polity’s domains. In the received literature, wandering persuaders7 are described as independent political advisers and diplomats who travel from court to court to offer their services to local rulers. They are depicted as shrewd and exceptionally skilled debaters, known especially for their mastery of two main rhetorical techniques, “argumentation” (biàn) and “persuasion” (shuì). When talking about these wandering persuaders, a distinction must be drawn between historical figures of actual persuaders that occasionally feature in the received literature, and the vast majority of pseudo-historical or plainly fictional characters. Especially the relatively straightforward, at times downright irreverent way in which they address their rulers provides grounds for doubt.8 The figure of the wandering persuader as “the Warring States topos of the itinerant philosopher intent on persuading a prince to accept his teaching”9 certainly has some historical basis.10 However, it is highly likely that the numerous dialogues that persuaders engage in with local rulers in the received literature, though possibly 6 See

the still relevant [50]. See also [111] and [112]. p. 126]. Oliver [108, p. 84] classifies wandering persuaders into three categories: storytellers, professional persuaders, and diplomatic agents. On this topic, see also [89]; [81, p. 52]; [127, p. 20]. 8 As Bruce E. Brooks and A. Taeko Brooks point out, it is highly unlikely that such strong claims as, for instance, Mencius makes in certain passages of the eponymous text, could have been possible at all in real life in front of a ruler [13, p. 277, footnote 2], at least not without risking one’s own head to be cut off. 9 [75, p. 12]. 10 Kai Vogelsang [139] has recently provided an interpretation of the role of the minister/advisor in this kind of dialogues in terms of “Politikberatung” (political consultation). 7 [75,

24

L. Indraccolo

based at least in part on descriptions of actual debates, are in most cases examples of a conventional literary pretext, used to impart a moral teaching or to convey a certain political or ideological agenda. As such, the figure of the wandering persuader might tentatively be defined as the most basic Classical Chinese “rhetorical trope.” Together with local rulers, wandering persuaders are the main protagonists of those texts that might tentatively be classified as argumentative in the Classical Chinese tradition, mostly—but not exclusively—belonging to so-called “Masters literature” (zˇısh¯u 子書). “Masters literature” is a bibliographical category that labels fairly heterogeneous, accretional11 compilations of different kinds of textual materials, including mostly clusters of anecdotes and short stories; polemico-philosophical dialogues; more or less structurally complex treatises; and collections of sayings and maxims. These compilations are in most cases immediately recognizable because they are typically named after eponymous persuaders or “masters” of thought, though exceptions to this rule do apply. A typical example would be the Mèngzˇı 孟子, that is Master 12 (= zˇı 子) Meng, best-known in the Western world by its latinized title Mencius. Similarly, other “Masters texts” that will be cited below include, for instance, Master Han Fei (Hán F¯eizˇı 韓非子), Master Xun (Xúnzˇı 荀子), Master Gongsun Long (G¯ongs¯un Lóngzˇı 公孫龍子), just to name a few examples. All these compilations invariably have an intricate and controversial textual history.13 These texts were collated, heavily edited, and in many cases even compiled (at least in some of their parts) during the early imperial period, i.e. roughly between the Hàn 漢 Dynasty (206 B.C.–220 A.D.) and the early Chinese Medieval period (220–589 A.D.). The term “Masters Literature” was initially created ad hoc as a convenient bibliographic category within a much broader and more complex taxonomic system, established by the scholar and librarian Liú Xiàng 劉向 (77–6 B.C.) together with his son Liú X¯ın 劉歆 (50 B.C.–23 A.D.) and an équipe of other literati. This editorial team was officially appointed in 26 B.C. by Emperor Chéng 成 (51–7 B.C., r. 33–7 B.C.) of the Western Hàn Dynasty (206 B.C.–220 A.D.) to systematize 11 Early Chinese manuscript texts were characterized by a rather fluid nature. Before the spread of silk and the invention of paper as material carriers in the early imperial period, manuscript texts were composed of more or less bulky bundles of bamboo or wooden slips held together by strings of different materials. Such bundles could comprise one individual text, sections or chapters of a longer text, or, at times, more than one text that had been copied together on the same bundle. Slips could vary greatly in size and number, and so did the number of graphs per slip. Fairly consistent conventions regulating the length of strips according to the content and genre of the written text were established only later in the imperial period, see [136] and [35]. Due to the malleability of the material carrier and these “modular” structural characteristics, it was relatively easy to modify or replace individual slips, or to replace, add and insert new “chapters” within a larger textual unit. As a result, texts underwent a gradual process of growth and “accretion” that lasted several centuries and included also later editorial additions and integrations during the major enterprise of the reorganization of the holdings of the imperial library during the Hàn 漢 Dynasty (206 B.C.–220 A.D.). For these reasons, scholars talk about the “composite nature” of early Chinese texts. See [11] and [12]. See also [95]; [120]; [77]; [98]. 12 The conventional translation “Master” is problematic to a certain extent. For an alternative interpretation of the term as a designation for social rank (“squire,” in the original German text “Juncker”), especially with reference to the Mèngzˇı, see Gassmann’s recent publication [38, pp. 153–155]. 13 [10]; [65]; [120].

2 Argumentation and Persuasion in Classical Chinese Literature

25

and catalogue the holdings of the Imperial Library. It can be assumed with a degree of certainty that at least the core of some “Masters texts” dates back to an earlier period. Still, it is almost impossible to ascertain the extent to which editors intruded into these texts and which sections were actually emended. Consequently, the attribution of even parts of these texts to Warring States persuader figures, let alone “authors” is a convenient literary convention rather than a truthful representation of the contemporary modalities of text production.14 This is especially true before the background of the substantial lack of a proper notion of authorship15 i.e. of a clearly identifiable authorial voice before the late Warring States-early Hàn period. Hence, these works should not be considered as a genuine representation of Warring States society, nor do they necessarily provide reliable accounts of contemporary historical events. They should be treated as an early imperial cultural “product” that mostly served the political agenda of the central government by promoting certain specific values or readings of the past, and contributed to the establishment of a shared, consistent cultural memory, thereby strengthening its cultural and political authority.16 Nevertheless, despite these evident shortcomings, “Masters texts” still provide invaluable information about the practice and the performance of rhetoric in early China.17 The reason why we have to resort to and rely on this kind of problematic sources in studying early Chinese rhetoric is that there is no contemporary textual evidence for the transmission and tradition of rhetoric in ancient China. No proper systematic handbook or didactic material has been handed down. Of course, this does not preclude the possibility that such kind of materials did exist and have simply not been transmitted, either as the result of a deliberate editorial/authorial choice, or because of external circumstances such as the physical preservation of manuscript texts. Another potentially problematic issue is often identified in the fact that an autochthonous Chinese term usually translated as “rhetoric” (xi¯ucí 修辭, literally “polished words”) is attested for the first time only in a much later, Medieval work, Liú Xié’s 劉勰 (c. 465–c. 22) Wénx¯ın di¯aolóng 文心雕龍 (The Literary Mind and the Carving of Dragons), the first extant work of literary criticism in the history of Chinese literature. Xi¯ucí is also the term that is still conventionally employed to translate “rhetoric” into Modern Chinese today. However, despite the absence of an unequivocal technical term that might be used to express a semantic field equivalent to the Western concept of “rhetoric” for earlier periods, the terms for “argumentation” (biàn) and “persuasion” (shuì) are occasionally used as a compound to designate the “rhetorical arts” or “rhetorical activity” in its broadest sense.18 Rhetoric as biànshuì 辯說 (literally “argumentation and persuasion”) seemingly identifies an acquired 14 On

text production in pre-imperial China, see [95] and [96]; [65]; [82]; [85]. the issue of authorship and the concept of “composite authorship” in early China, see [103]; [64] and [65]; [5]; [125]. 16 See [95] and [96]; [73] and [74]; [113]. 17 On the value of and the contribution provided by “Masters texts” in this kind of inquiry, see [28]; see also [135]. 18 See [61, esp. pp. 906–908]; a list of occurrences in early Chinese received literature can be found at p. 906, footnote 56; see also [75, p. 122]: “The Warring States topos of the itinerant 15 On

26

L. Indraccolo

competency, the mastery of a set of pragmatic skills related to the effective use of language. It thus echoes the Western Classical tradition of rhetoric as a pragmatic practice or téchn¯e quite closely.19 As will be shown in more detail below, the practice and the performance of the different though complementary20 techniques of “argumentation” and “persuasion” are widely attested in Classical Chinese received literature. These two techniques were predominantly—though not exclusively—employed in a political context. As it has already been remarked, persuaders made a living mostly by acting as officials, advisors, or diplomats at the service of local rulers. Accordingly, most cases of “argumentation” and “persuasion” were performed at court and, in certain cases, even for the entertainment of the court. We can identify three main “polemical scenarios” in which cases of “argumentation” and “persuasion” are staged21 : (1) the one-on-one debate (occasionally performed also as a form of court entertainment) in the case of “argumentation.” This kind of debate usually takes place at court, involves two or occasionally more characters, and assumes the form of a dialectical skirmish; (2) the address to a sovereign or, more generally, to a superior in rank, in the case of persuasion. A persuader, typically an official or an envoy, engages in an impassioned plea to persuade their superior to assume a certain desirable behavior, or to undertake a course of action that is considered advisable under the current circumstances. It is usually addressed to a superior in rank, though occasional cases of top-down persuasions or persuasions involving people of equal status do also exist. (3) the teaching scene. The teaching scene implies the presence of one or more disciples or students gathered around a teacher who imparts an oral teaching.22 This scenario can provide the setting for teaching and practicing both rhetorical techniques. It emerges quite clearly from this description already that there is a fundamental discriminant factor that can help disambiguate cases of “argumentation” and “persuasion.” These two

philosopher intent on persuading a prince to accept his teaching was related not just to persuasion but to disputation as well.” 19 For a detailed analysis of the use of the terminology for “rhetoric” in early China and a comparison between the two Chinese terms for rhetoric, see [55] and especially [61, pp. 906–908]. 20 As Kroll points out, “the arts of “disputation” (pien) [biàn 辯] and “persuasion” (shui) [shuì 說], were ascribed to the same person as mutually connected skills” [75, p. 126]. 21 My interpretation is partially indebted to Lu Xing’s threefold classification [89, pp. 64–65] of what I define here as “polemical scenarios.” Lu claims that “the first (setting for rhetorical activity) was the political realm where you shui [游說, persuaders] or bian shi [辯士, debaters] served as advisers to the kings. […] [T]the second arena for rhetorical activities was the realm of education where intellectuals taught their students various subjects using a lecture and discursive format.” Lu postulates also a third possible scenario, the (Jìxià 稷下) “Academy.” However, increasing doubts have been raised in the past years about the possibility that a structured and organized institution recognizable as a proper “Academy” might have actually existed in the pre-imperial period. Therefore, I prefer to leave this latter potential setting aside. For a recent account of the alleged activities of the Jìxià Academy, see [145]. 22 On the teaching scene or “scene of instruction,” see [144]; see also [28, Chapter 2].

2 Argumentation and Persuasion in Classical Chinese Literature

27

techniques can be preliminarily distinguished according to the kind of relationship— equality or subordination—existing between the two (or more) characters involved in the dialogic action. Let us now proceed to analyze in more detail the main features of these two techniques, starting with “persuasion.” The term we translate as “persuasion” (shuì) identifies a rhetorical technique that is usually employed in at times highly structured dialogues, which might be further embedded into broader anecdotal narratives and/or used as illustrative examples. Shuì typically assumes the form of a plea and is usually addressed to a superior in rank. It is aimed at convincing the persuadee to agree on some key ethical or political issue and to provide an appropriate response or reaction, or to assume a certain desirable (though not necessarily “moral”) behavior. The persuasive force of such an attempt is typically enhanced through the widespread use of contrastive examples, where by “contrastive examples” I mean the juxtaposition of a positive (typically analeptic) versus a negative example that can range from a one-liner to a brief anecdote. In most cases, the positive example is drawn from a remote and idealized past, while the negative one refers to contemporary times and denounces a situation of moral decline due to a process of progressive deterioration of values and customs. This kind of plea expresses either a thinly veiled criticism or an advice, which can be either provided spontaneously or solicited at the very beginning of the exchange. Criticism is usually directed against a type of impractical, morally inacceptable, or generically detrimental political course of action that the persuadee has pursued or intends to pursue. Advice can pertain to the public or the private sphere. In the former case, advice is targeted at political practice: either internal affairs, in particular handling the issue of manipulative people and sycophants scheming at the ruler’s back inside the court, or matters of foreign policy and diplomatic relations with other states. When it touches upon the private sphere, advice turns into a moral teaching imparted to the persuadee, often illustrated with colorful analogies and metaphors drawn from the natural world and the animal realm, or the domain of human activities, such as manual work and craftsmanship.23 As we are explicitly told in two “Masters texts,” in Chapter 12 “The Difficulties of Persuasion” (‘Shuìnán’ 說難) of Master Han Fei (Hán F¯eizˇı 韓非子) and in Chapter 14 “On Attracting Scholars” (‘Zhìshì’ 致士) of Master Xun (Xúnzˇı 荀子),24 to successfully achieve their communicative goal, persuaders have to be sensitive and receptive to the persuadee’s moods and state of mind, to read him and understand his psychology so that they can turn the latter’s innermost desires to their 23 There are several publications available that focus on individual examples of these two kinds of metaphors. As an in-depth analysis of this topic goes beyond the scope of the present article, I am quoting here only a selection of recent studies for further reference. For a theoretical treatment of the use of metaphors in early Chinese philosophy, see [128] and [129]; on the use of metaphors specifically in the “Masters text” Zhu¯angzˇı 莊子, see [143], [16], and [70]. On natural metaphors, and especially the predominant water metaphors, see for instance [1]; [91]; [17]. On the relatively still understudied animal metaphors, see [132]; [114]; [92]; and [58]. On tool metaphors, see [138]; [25]; [47]; [27], to quote just a few. 24 [71, 14/5/52, p. 209].

28

L. Indraccolo

own advantage.25 In order to do so, the persuader must, as Mary Garrett puts it, “appeal to emotions, to enlightened self-interest, and to idiosyncratic desires”26 by purposely approaching the topics that are more likely to stir the desired reaction in the dialogic counterpart. It is especially crucial to do so in a timely and convincing way.27 Hence, a case of “persuasion” is always circumstantial and contextual, as persuaders have to attune their speech and tailor it to the audience, to the specific occasion, and to the environment in which the persuasive speech is delivered. In Classical Chinese received literature, “persuasion” overall has a positive connotation. It is considered as a healthy, pragmatic means of indirectly criticizing or admonishing, guiding, and positively influencing a ruler’s (or a superior’s) actions and behaviors for their own good and, most importantly, for the sake of the security and stability of the state. Instances of shuì are fairly elaborate dialogues conceived primarily as written texts with an explicit persuasive intent. As mentioned previously, such dialogues are sometimes embedded into detailed and lengthy narratives. Such narrative frameworks might or might not be directly related to the content of the “persuasion” per se. This is because cases of “persuasion” can also be used as illustrative examples in an oblique, allusive way, and included into a broader speech that might be a case of “persuasion” as well, in a matryoshka-like structure. Independent, self-standing cases of “persuasion” are sometimes framed by an introductory narrative and a final concluding remark, ranging in length from a one-liner to a short paragraph. These provide information on the extratextual context of the “persuasion” and describe the setting and the occasion of the speech, and its subsequent outcomes, respectively. Final remarks alternatively might cite a wisdom saying drawn from the shared cultural lore imparting a moral teaching.28 Also, brief narrative junctions act as “fictionalizing elements” that are used to build up the narrative flesh of the story by establishing coherent conceptual bridges among the different sections of the dialogues. These “fictionalizing elements” are usually brief sentences that help sketch a more detailed account of the characters’ psychology or connect different passages in a more fluid manner by bridging the gap between the characters’ lines. However, it must be noted that this kind of “fictionalizing elements” are not an exclusive feature of cases of “persuasion,” as they can also be found in cases of “argumentation.” Cases of “persuasion” are usually staged at court. They can involve two or more characters, and are generally “hierarchical,” as they typically describe a bottom-up process characterized by a situation of social disparity between the persuader and the persuadee. Cases of “persuasion” are always situational, and stress is put on

25 On

the psychological dimension of persuasion, see [37]; [36]; [46]; and [122]. p. 112]. 27 “Motivated by the self-interested desire for profit, you shui employed utilitarian appeals in persuading their audiences. In speeches they made to the kings of the various states, for example, they would spell out the material gain, in terms of land, beautiful women, horses, food, and clothing. The ruler could expect benefits he would gain by adopting the proposed plan” [89, p. 119]. See also [41, p. 5] and [42]; [63, p. 230]. 28 On framing devices in cases of “persuasion,” see [20] and [21]; [63]; [89]; [41] and [42]. 26 [37,

2 Argumentation and Persuasion in Classical Chinese Literature

29

the audience.29 In the majority of cases, the persuader steals the scene and engages in a rather long monologue. The persuadee, on the contrary, might be tentatively defined as an anti-dialogic counterpart. They listen silently, almost without reacting until the end of the speech, or they reply rather briefly, asking further questions and clarifications. However, their reactions are always functional to the development of the speech. They are formulated in a way that triggers the persuader’s reply, giving him the chance to proceed to the following step in the reasoning. Typical cases of “persuasion” are, for instance, the numerous dialogues between ministers/persuaders and kings included in the Zhànguó cè 戰國策 (Strategies of the Warring States) and in the Mencius. The Zhànguó cè is an early imperial example of historical fiction. It is a collection of dialogues, anecdotes, and speeches involving rulers and ministers that seemingly date back to the Warring States period, but that have been arbitrarily selected, compiled and arranged into a collection only at a later stage by Liú Xiàng during the Hàn period. Such selection is supposedly representative of one of the six trends of thought of the Warring States period,30 the so-called zònghéng 縱橫 (Vertical and Horizontal Alliances, also translated as “school31 of diplomacy”). Cases of “persuasion” in the Zhànguó cè often look like proper narrative anecdotes rather than speeches. A fairly large amount of preliminary background information is provided in the introductory narrative about the setting, the extra-textual context, and the political situation in which the characters find themselves to act and interact. Extensive narrative junctions also link up the different sections of the dialogues, 29 On the role of the audience and its influence on how a “persuasion” is developed and delivered, see [26, p. 96]; [37]; [89]; [81]. 30 The term ji¯ a 家 (literally, “household,” “family,” but also “expert, person with an expertise in a certain field”) has long been used rather inappropriately to translate the names of six philosophical trends of thoughts or groups of experts of the Warring States period as “schools” of thought. These were first identified and described by the Grand Historian and Astronomer S¯ımˇa Tán 司馬談 (c. 165–c. 110 B.C.) in his treatise “The Cardinal Principles of the Six Trends of Thought” (‘Liùji¯a zh¯ı yàozhˇı’ 六家之要旨), included in Chapter 130 of his son S¯ımˇa Qi¯an’s 司馬遷 (c. 145–c. 86 B.C.) historiographical work Shˇıjì 史記 (Records of the Grand Historian). These six trends of thought are the Cosmologists (y¯ın-yáng 陰陽), the Confucians (Rú 儒), the Mohists (Mò 墨), the Logicians or experts on names (míng 名), the Legalists or experts on laws and standards (fˇa 法), and the Daoists (dàodé 道德). However, this taxonomy should rather be considered as a bibliographical classification rather than any meaningful attempt at categorizing more or less loosely organized philosophical groups active during the Warring States period. Actually, these labels were created to classify “Masters text” into different subcategories during the process of systematization and reorganization of the holdings of the Imperial Library (see pp. 4–5 above). Recent archaeological manuscript findings over the past forty years have radically altered our views on early Chinese culture and society, urging the need to rethink traditional interpretive categories. According to these newly excavated manuscripts, ideas and philosophical concepts seems to have been circulating in a much more fluid fashion than it was once believed, allowing several cases of transversal “contamination,” see [95] and [96]; [28]; [73]; [69]. In consideration of this, the existence of clear-cut boundaries between allegedly antagonistic, organized entities that might be identified as “schools” has been largely disproved, with the exception of only two cases, the Confucians and the Mohists, who apparently managed to establish proper structured teaching institutions. 31 On the misuse of the term “school” as applied to the Chinese pre-imperial period and other possible, more suitable interpretations of the term ji¯a 家, see [110]; [121]; [103]; [22]; [23]; [130].

30

L. Indraccolo

which are closely intertwined with the narrative “flesh” of the story. In these instances of “persuasion,” narrative action is decidedly predominant.32 Cases of “persuasion” in the Zhànguó cè have been extensively studied before especially by James I. Crump.33 Therefore, also in consideration of the average length of the episodes included in the Zhànguó cè, I will focus here on a particular kind of persuasive dialogue included in the Mencius that has not been thoroughly studied yet from the point of view of its formal and structural characteristics. The Mencius is a “Masters text.” It is the eponymous didactic collection of anecdotes and dialogues traditionally attributed to the Confucian thinker Mèngzˇı 孟子 (372–289 B.C.)—or Mencius in its latinized form, the core parts of which seemingly date back to the Warring States period.34 The received text is composed of seven chapters, each divided into two parts (commonly marked as A and B or shàng 上 and xià 下 respectively). According to the tradition, the text originally would have comprised eleven chapters altogether, seven “inner” chapters and four “outer” chapters. However, the outer chapters would have been considered spurious and excised by the Eastern Hàn Dynasty scholar Zhào Qí 趙岐 (d. A.D. 201) when he edited the Mèngzˇı adding also his own commentary to the text. According to the biography of Mencius recorded in Chapter 74 (Mèngzˇı Xún Q¯ıng lièzhuàn 孟子荀卿列傳) of the Shˇıjì 史記 (Records of the Grand Historian), Mencius would have written the eponymous text with the help of his disciples. However, the soundness and the reliability of this statement are dubious, and Mencius’ alleged authorship of the text, or even of parts of it, is a highly disputed and controversial issue.35 In fact, we might say that the text in its present arrangement is most probably to be considered a Hàn product. The work is well-known especially for the dialogues, whether staged and fictional or not, that Mencius undertakes with different local rulers in the attempt to pursue his own ethico-political agenda. In this enterprise, he suggests concrete means to improve the living conditions of the local population and promotes the enactment of an ethical government that should finally be up to the standards set by the example of the legendary Sage Kings of Old,36 a mythicized archaic epoch in which cultural and moral refinement had allegedly reached its apex and its most accomplished form. I will focus here especially on this cluster of “court speeches” preserved in the Mencius to exemplify the characteristics of a case of “persuasion,” where by “court speeches” I mean dialogues that Mencius holds at court with local rulers and sovereigns of 32 About

the structure of speeches in the Zhànguó cè, see [89, pp. 116–126, in particular 124]; [41] and [42]; [20] and [21]. 33 See [20] and [21]. See also [117, pp. 117–122]; [41] and [42]. 34 See [14], and in particular [13]; [78]; and [116]. 35 [13]. 36 The so-called S¯ an Huáng Wˇu Dì 三皇五帝, the Three Sovereigns (usually Fúx¯ı 伏羲, Nüˇ w¯a 女媧 and Shénnóng 神農, but there are other possible variants of this triad) and Five Thearchs (Huángdì 黃帝, Zhu¯an X¯u 顓頊, Yáo 堯, Shùn 舜, and Yˇu 禹). According to the tradition, these are a set of eight legendary pre-dynastic kingly figures, endowed with superhuman powers. They are considered epitomes of moral rectitude and superior knowledge, and, in some cases, they also act as cultural heroes. In virtue of their superior moral qualities and discernment, they are capable of enacting and ideal government. See [93]; [62]; [9]; [15]; [79]; [8]; [83] and [84]; [80].

2 Argumentation and Persuasion in Classical Chinese Literature

31

different states. As I will try to show in detail, these “court speeches” are not merely generic dialogues, as they have typically been classified in the secondary literature on the topic,37 but are rather to be considered examples of a specific type of dialogue, that is “persuasion.” “Court speeches” in the Mencius are rather sketchy, especially with regard to the background and the setting of the dialogic action. They present a quite regular and predictable structure, and focus mainly on the dialogic exchange between the two characters involved. A “court speech” is typically constituted by two main parts: a relatively short introductory section that includes one or more question-and-answer exchanges, and a second, longer, more structured section that I will call a “concluding monologue.” The lines of the two characters are easily identifiable, as in most cases they are clearly marked by the use of the verb “to say” (yu¯e 曰) that typically introduces direct speech, or by other verbs of saying (i.e. “to reply” duì 對, “to inform” gào 告, “to say” wèi 謂 and “to ask” wèn 問). In the introductory section, the ruler typically opens the discussion by posing a generic question on a topic related to government or asking for advice about the suitable behavior to be undertaken in a usually rather specific, concrete situation. The ruler’s questions—both at the very beginning and in the thrust and counterthrust that follows—can occasionally be formulated in a humble and respectful tone and introduced by the verb qˇıng 請 (“to please, to invite, to ask”), though in most cases the ruler addresses Mencius in a much more straightforward way. When an example used as argumentative topic for discussion is introduced in this introductory section, whether at the very beginning or in the middle of the dialogic exchange, it is occasionally cast in interrogative form and closed by the conventional rhetorical formula (yˇouzh¯u 有諸) “has there been such a case?” The latter section that I defined as “concluding monologue” can be identified as the proper act of “persuasion.” Here the persuader invites the ruler to reflect on his words, to trust them, or to adopt a particular kind of behavior—invariably moral, in the Mencius. This is sometimes marked explicitly by a sentence that uses the verb qˇıng 請 (“to please,” “to invite,” “to ask”). In this latter section, Mencius finally engages in a large-scale persuading effort. He presents a more elaborated argument to promote his cause and treats ethical issues that are anticipated, sometimes just as catchwords in the previous exchanges, in more detail. This second section is also the part where the highest number of structural and rhetorical devices are employed. Repetitions, parallel sentences, contrastive examples, historical allusions, literary quotations and intertextual cross-references among other rhetorical devices usually frame the whole textual block. In this section, a pivotal turning point or a contrastive situation is often introduced and marked by the adverb j¯ın 今, “now,” which establishes a strong contrast between an ideal and idealized past and the contemporary situation that witnesses the degeneration of traditional moral values and customs.38 Apart from the common causal adverb “therefore/this is the reason why” gù 故, the systematic 37 See 38 On

for instance [45]; [100]; [126]; [76]; [51]; [28]. transition terms, see footnote 2 at p. 1.

32

L. Indraccolo

analysis of “court speeches” preserved in the Mencius39 has also highlighted the use of a few other rhetorically-charged transition terms or “linkers”40 : • fú 夫 (“hence”, “as a consequence”), sometimes used in rhetorical questions to underline a fact that is inferred from the previous reasoning, adding another step to the argument that derives logically from the previous one41 ; • qiˇe 且 (“moreover”), used to reinforce a point or, more often, to cite an additional example to further improve the persuasiveness of an argument; • yì 亦 can assume two different meanings and has two different rhetorical uses, respectively: (1) (“perhaps”, “if”, “might”) dubitative, it can be employed by both characters in a dialogue to express their perplexity, or to cast doubt on whether a certain behavior or situation is morally acceptable; it is also employed to soften a critical statement and make it more digestible for the ruler, or to suggest in a rather subtle way that a certain situation “is not the case;” (2) (“also”, “alike”), used to establish an equal relationship between two characters, or to equate two situations; • gài 蓋 (“indeed”, “to be sure”), used to reinforce an argument or to emphasize a statement, it is usually employed to underline the fact that an action is an exemplary case of a certain kind of behavior, or that someone or something belongs to a specific category. The “concluding monologue” usually ends with a final moral teaching that can assume the form of a rhetorical question, a wisdom saying, or an authoritative quote taken from the Confucian canonical texts,42 the Odes (Sh¯ı 詩) or the Documents (Sh¯u 書), followed by Mencius’ explanation.43 Let us now have a closer 39 The present analysis has been carried out on a cluster of dialogues in the Mencius (“court speeches”) that present the same stylistic and structural characteristics. It is therefore not necessarily indicative of the use of transition terms at large in the whole text. 40 An interesting term of comparison is the use of similar transition terms is the later highly formalized and stylized rhetoric of the eight-legged essay during the Míng 明 (1368–1644) and Q¯ıng 清 (1644–1911) dynasties. This suggests a fairly high degree of consistency and continuity in the use of certain rhetorical structures across the ages, hinting at a much more solid foundation of such structures in the literary tradition also for the earlier periods than it is usually acknowledged. See [29, in particular p. 398]. 41 For an updated analysis of fú as a transition term, see [141]. 42 I am talking here of canonical texts because the quotations employed in this cluster of dialogues are all known and can be easily tracked in the Five Classics (Wˇuj¯ıng 五經), and specifically in the Classic of Odes (Sh¯ıj¯ıng 詩經) and the Classic of Documents (Sh¯uj¯ıng 書經). These quotations are evidently part of an already codified and well-established corpus of literature, contrary to what happens with similar intertextual references that were still circulating rather fluidly in the Warring States period [124], as recent archaeological manuscript findings prove. See [95] and [97]; [120]; [66], [67], and [68]; [73]. This seems to support a later composition or at least editing of the Mencius during the late Western Hàn or Eastern Hàn period, when the gradual process of canonization of the Confucian Classics started to take place. See [101]; [102]; [104]; [105] and [107]; [86]. 43 In the received literature, the use of quotations taken from the Confucian Canon as a framing or internal structuring device is a distinctive feature that is often used especially to stress the words of those characters that are more closely associated with the Confucian tradition, though the texts in which they appear are not necessarily nor exclusively included in Confucian texts.

2 Argumentation and Persuasion in Classical Chinese Literature

33

look at an example of “court speech” to illustrate how this whole fairly complex rhetorical structure is enacted in the text. 1B3 齊宣王問曰: 「交鄰國有道乎?」 King Xu¯an of Qí asked:

“Is there a way to maintain a good relationship with the neighboring states?”

孟子對曰:

「有。惟仁者為能以大事小, 是故湯事葛, 文王事昆 夷; 惟智者為能以小事大, 故大王事獯鬻, 句踐事吳 。以大事小者, 樂天者也; 以小事大者, 畏天者也。樂 《詩》 云: 『畏天之威, 天者保天下, 畏天者保其國。 于時保之。』」

Mencius replied:

“There is. Only a benevolent man is capable to serve a greater [state] with his smaller [state]: this is the reason why T¯ang served Gé and King Wén served the K¯un barbarians. Only a wise man is capable of serving a greater state with a smaller state: this is the reason why King Tài served the X¯unyù and G¯oujiàn served Wú. He who serves a smaller state with his greater state is someone who delights Heaven. He who serves a greater state with his smaller state is someone who is in awe of Heaven. He who delights Heaven protects All-under-Heaven [i.e. the whole world, ti¯anxià 天下]. He who is in awe of Heaven protects his own state. An Ode says: ‘Being in awe of Heaven’s awesomeness, we will protect it throughout time’.44 ”

王曰:

「大哉言矣! 寡人有疾, 寡人好勇。」

The King said:

“Great words indeed! But I, your unworthy king, have a weakness, I am fond of valor.”

對曰:

「王請無好小勇。夫撫劍疾視曰, 『彼惡敢當我哉 』!此匹夫之勇, 敵一人者也。王請大之! 《詩》 云: 『 王赫斯怒, 爰整其旅, 以遏徂莒, 以篤周祜, 以對于 《 天下。』此文王之勇也。文王一怒而安天下之民。 書》 曰: 『天降下民, 作之君, 作之師。惟曰其助上 帝, 寵之四方。有罪無罪, 惟我在, 天下曷敢有越厥 志?』一人衡行於天下, 武王恥之。此武王之勇也。 而武王亦一怒而安天下之民。今王亦一怒而安天 下之民, 民惟恐王之不好勇也。」

44 Ode

n. 272 我將 Wˇo ji¯ang.

34

L. Indraccolo

Mencius replied:

“You, my Majesty, are kindly invited not to be fond of small valor. When a man brandishes his sword and says with a fierce look ‘How dare you oppose me?,’ this is the valor of a common fellow, who can stand up to only one adversary. You, my Majesty are kindly invited to make it something greater! An Ode45 says: ‘The King blazed with rage, and then he marshaled his troops, to stop the enemies marching toward Jˇu, to consolidate the blessings of Zh¯ou, to reciprocate towards All-under-Heaven.’ This was the valor of King Wén. And King Wén in one outburst of wrath appeased the people of All-under-Heaven. In the Documents46 it is said: ‘Heaven made the lowly people descend down [to earth], made him their ruler, made him their master, so that it would be said that they assisted the Lord on High, bestow his favor to the four quarters [of the world]. It is the guilty as well as the innocent that are my concern, who in All-under-Heaven would dare to step over my will?’ If there was somebody creating disorder in the world, King Wˇu would be ashamed about this. This was the valor of King Wˇu. And King Wˇu, also in one outburst of wrath, appeased the people of All-under-Heaven. Now if Your Majesty, also in one outburst of wrath, appeases the people in All-underHeaven, the people will only fear that Your Majesty is not fond of valor.”47

The King opens the discussion posing a brief question on a concrete political issue. Mencius’ answer evokes two different types of moral behavior, exemplified by two pairs of corresponding historical exempla represented by four figures of kings from the past (T¯ang and King Wén, King Tài and G¯oujiàn). The persuader describes the positive consequences that such behavior might have on All-under-Heaven, trying to bring to the king’s attention a concrete advantage he could gain by spontaneously assuming moral behavior. The whole answer is cast in parallel sentences with partial repetitions and is closed by an authoritative quote taken from the Classic of Poetry that aims at sanctioning the validity and the authority of the argument. The King replies rather briefly, admitting his weakness for valor. He thereby provides Mencius with the necessary cue from where to start to engage in his concluding “monologue,” the proper act of “persuasion.” Mencius’ intention to persuade the ruler is made more evident through the explicit use of the verb “to please, to invite, to ask” (qˇıng 請) as he exhorts the ruler twice. First, Mencius n. 241 皇矣 Huáng yˇı. quotation is drawn from Tàishì shàng 泰誓上, with slightly different wording. 47 Translations in this article are mine. 45 Ode 46 The

2 Argumentation and Persuasion in Classical Chinese Literature

35

tries to dissuade the king from being fond of “small valor.” Then he wholeheartedly encourages the ruler to maximize the kind of valor he is fond of and turn it into something greater. It is quite evident that Mencius is well aware that the king’s conduct is far from exemplary, and that it is definitely not up to the moral standards set by the legendary kings he is talking about. Nevertheless, Mencius tries to flatter the ruler so as to convince him that he is also capable of greater endeavors. The main body of the “concluding monologue” is tripartite, and each part corresponds to one among three different kinds of valor: the valor of a petty common fellow, and the valor of two exemplary kings of the past, King Wén and King Wˇu, who are personifications par excellence of civic and martial virtue, respectively. The last two positive examples are both introduced by a quote from authoritative sources (the Classic of Odes and the Classic of Documents) and closed by two parallel and almost identical sentences, in which only the name of the king changes. A pivotal turning point that sets apart this first section from the final maxim in the conclusion is marked by the adverb “now” j¯ın 今, which abruptly takes us back to reality and to the contemporary situation of moral derangement, epitomized by the far less than exemplary figure of King Xu¯an. Let us now move on to analyze the rhetorical technique of “argumentation” (biàn 辯), the dialectical counterpart of “persuasion” (shuì). The term that denotes “argumentation” (biàn) literally means “to distinguish, to discriminate, to argue.”48 Accordingly, a case of “argumentation” typically assumes the form of a dispute or a dialectical skirmish among peers. More precisely, as Mary Garrett defines it, “bian is disagreement over theses that are contradictories […].”49 Over time, “argumentation” assumes a strong negative connotation, and becomes a synonym for hair-splitting argument. This rhetorical technique is often harshly criticized as an artful or deceitful type of reasoning that is employed in a deliberately misleading fashion, with no other intention than outtalking an opponent at any cost. Accordingly, in several “Masters texts,” to indulge in “argumentation” is often dismissed as a perverse method, unworthy of a true scholar. The main criticism directed against the use of “argumentation” is the apparent disinterest in achieving any superior ethical goal, and the lack of scruples with which persuaders employing this technique are ready to support even the most incredible propositions for the sake of winning an argument. The received literature claims that similar stratagems are only used by the shrewdest persuaders to manipulate and deceive, or for the sheer pleasure of arguing for arguing’s sake. One of the preferred—though definitely not the only—target of such forms of criticism is a group of (pseudo)historical characters, the so-called “experts on names” (míngji¯a 名家) or Logicians, commonly known

48 In

Old Chinese [i.e. the varieties of Chinese spoken in China approximately until the foundation of the Qín 秦 Dynasty (221–207 B.C.), see [3, pp. 1–8], 辯 *brenP (“to debate,” “to dispute,” “to argue”) is phonologically identical with 辨 *brenP (“to distinguish,” “to discriminate”), hence the interchangeability of the two graphs and the broader semantic spectrum of the term. For the phonological reconstruction, see [3, p. 229]. See also [4]. For a definition of biàn that keeps into consideration both aspects and nuances, see [89, p. 88]. See also [37]. 49 [37, p. 107].

36

L. Indraccolo

also as “Chinese Sophists.”50 The most representative figure among this group of persuaders is G¯ongs¯un Lóng 公孫龍 (trad. c. 320–250 B.C.), especially famous for his paradoxical claim that “white horse is not horse” (bái mˇa f¯ei mˇa 白馬非馬), which will be discussed in more detail below. Two of the most tenacious detractors of the technique of “argumentation” are, for instance, the “Masters texts” Hán F¯eizˇı (Master Han Fei) and Xúnzˇı (Master Xun). The Hán F¯eizˇı repeatedly complains about the pernicious influence exercised by shrewd and devious disputers (biànzhˇe 辯者 or biànshì 辯士) crowding the court. By dwelling on and indulging in “thorny talk of the kind of the ‘white horse argument’” 「 ( 棘刺白馬之說」in Ch. 32 “Outer Congeries of Sayings, Upper Left Series” ‘Wài chˇushu¯o zuˇo shàng’ 外儲說左上), these disputers succeed in captivating the ruler’s attention and distract him from far more urgent government issues. The Xúnzˇı equally gives vent to a notorious and widespread Confucian reproach against flowery and ornate discourse, labeled as an insincere and deceitful way of expressing oneself. However, a clarification must be made here. Even those received texts associated with the teachings of masters considered epitomes of virtue, such as, for instance, the Mencius itself, are known to make widespread use of both the techniques of “argumentation” and “persuasion.” As Jurij L. Kroll has underlined, “it appears that […] Hsun Tzu [i.e. Xúnzˇı], and, evidently, Meng Tzu [i.e. Mèngzˇı, Mencius], among others, happened both to persuade and to dispute. The arts of ‘disputation’ (pien) [i.e. biàn] and ‘persuasion’ (shui) [i.e. shuì] were ascribed to the same person as mutually connected skills.”51 Therefore, such vehement accusations of immorality should not be interpreted literally as has long been the case, but rather be considered as examples of the technique of the “rhetoric of anti-rhetoric,”52 a literary cliché that was apparently largely employed by all the persuaders and philosophical “masters” of the time to belittle adversaries in the political arena.53 In Classical Chinese literature, cases of “argumentation” are heated and fast-paced dialogical exchanges between two (or occasionally more) characters of equal social status. As a general rule, the two characters involved can be identified as a persuader and his virtual opponent or preferential “sparring partner.” In a case of “argumentation,” the persuader often supports a disconcerting, unusual or subversive view on reality, while his opponent represents common thought. While the range of topics addressed in cases of “persuasion” is usually of a rather pragmatic nature and is

50 [33]. To my knowledge, Alfred Forke is the first to have employed the label of “Chinese Sophists”

to described the Logicians. p. 126]. 52 Paolo Valesio has defined this strategy as “a sophisticated rhetorical mechanism […] employed to convey a message that attacks ‘rhetoric.’ […] This rhetorical structure manifests itself both at the semantic level (in the ideological contrast that is set up) and at the level of the syntactico-semantic (or synctatico-lexical) form of the message: that is respectively both at the level of the topic and at the figurative level […],” in which “praise of plain speech is couched in a rhetorically sophisticated language” [137, pp. 41–42]. 53 See [56]. 51 [75,

2 Argumentation and Persuasion in Classical Chinese Literature

37

closely related to politics or ethics, “argumentation” can also deal with highly speculative and abstract theoretical problems. However, as will be shown, such abstract speculations are only apparently detached from reality. They are more than pure logical divertissements for their own sake. At times they may entail a deeper, hidden meaning, and allude to concrete situations in an oblique way. In particular, according to the rich and varied repository of anecdotes preserved in the received literature, the whimsical language jokes and paradoxes54 that are characteristic of this technique are often used in an allusive, oblique way to solve delicate diplomatic and political issues. Cases of “argumentation” are usually rather sketchy, although they can occasionally be framed by an introductory narrative that contextualizes the dialogue within a fictitious debating scene. The most common rhetorical devices employed in the construction of an argument in a case of “argumentation” are the sorites or chainreasoning, the syllogism, and the reductio ad absurdum. Argumentative force is enhanced with the aid of illustrative examples, and of a rich repository of analogies and metaphors drawn mostly from the semantic fields of handicraft and nature. In a case of “argumentation,” the main goal of a persuader is to outtalk their opponent by forcing them into open contradiction, thus leaving them speechless and thereby winning the debate. To achieve this goal, persuaders employ any means at their disposal to disorient their adversary. They dazzle their opponent with stringent logic, and deliberately confuse them with language jokes or paradoxical statements. Eventually, the persuader turns the opponent’s own argument against them by twisting their words and subverting their meaning. The opponent, frustrated and confused, fails to notice the potential flaws or fallacies in the persuader’s reasoning, eventually falling into open contradiction. Typical examples of “argumentation” are, for instance, the dialogues preserved in the G¯ongs¯un Lóngzˇı (Master Gongsun Long). The G¯ongs¯un Lóngzˇı is a collection of six fairly independent texts attributed to the eponymous dialectician,55 whose core supposedly dates back to the Warring States period. The collection is infamous for its highly cryptic content and the pervasive use of paradoxical statements and language jokes. This work and its alleged “author” have been harshly criticized even by contemporaries for the aggressive use of “twisted words” and, especially, the apparent lack of any superior ethical goal in his philosophical agenda. The cases of “argumentation” included in this collection provide a significant example, as they are deprived of any fictionalizing element, to the point that not even the names of the characters involved in the dialogues are mentioned. As such, they provide a purely theoretical framework for discussion. This peculiar characteristic suggests that these texts most probably were not meant to be read alone in silence without an external teaching aid, or “enjoyed” as written texts per se. Though there is no way to determine what the actual goal behind the compilation of this collection 54 For a recent, detailed overview and tentative explanation of most famous paradoxes preserved in the early Chinese literary tradition, see [57]. See also [131]; [133]; [117] and [118]; [48]; [24]; [16]; and [34]. 55 On the fictionality of these figures, see [23]; [130]; [110].

38

L. Indraccolo

was, we might tentatively hypothesize that this kind of material could have been used as written drafts, teaching materials or aide-mémoires to improve the study and the memorization of a stock of sample arguments, and to foster their subsequent original re-elaboration and adaptation to different contexts.56 To illustrate the main characteristics of the rhetorical technique of “argumentation,” I am taking here as a case study what is surely the most famous paradoxical argument in the history of Chinese literature, the so-called “white horse argument.”57 The discussion of this argument is preserved in the first chapter of the G¯ongs¯un Lóngzˇı, entitled ‘Disquisition on White and Horse’ (‘Báimˇa lùn’ 白馬論). A complete and thorough study of the whole chapter is beyond the scope of this article. Consequently, I am presenting here selected passages in an abridged form in order to provide an accurate overview of the way in which the argument is constructed and developed throughout the text.58 Apart from being a rather witty and refined example of logical exercise, the argument also entails a deeper meaning that has often been too easily dismissed without any further investigation. At a closer look, the debate can be metaphorically interpreted as an exhortation to re-establish meaningful, univocal correspondences between names and the corresponding realities they designate, contextualizing itself within the then contemporary debate on the need for a “rectification of names” (zhèngmíng 正名).59 Read in these terms, the whole argument might allude to the disruption of the fundamental hierarchical social relationships that should frame, inform and order civil society, conveying an implicit underlying criticism directed towards the contemporary situation of social and political disorder, and the decay of traditional moral values.60 The “white horse” argument is a repartee that includes seven exchanges61 between an anonymous persuader and their opponent. This kind of question-and-answer structure is repeated consistently throughout the dialogue. All the exchanges can be easily identified, as they are marked by the speech act verb yu¯e 曰. The whole argument revolves around the issue of whether the apparently paradoxical statement made by the persuader at the very beginning that a white horse would not be a horse is admissible or not. O(曰)62 : 「白馬非馬可乎。」 56 See

[61].

57 The Dàozàng 道藏 (Daoist Canon) version is largely considered the most reliable existing edition

of the G¯ongs¯un Lóngzˇı, excerpts from the text included in this contribution are taken from this version. On different editions of this text and the alleged superiority of the Dàozàng edition, see [53] and [54]. For an updated analysis and a partial translation of the “white horse argument,” see [59]. For a complete analysis, see [53]. 58 For a detailed study and translation of the text, see the still actual [44]; [94]; [72]; [109]. 59 On the theory of the “rectification of names,” see for instance [88]. 60 See [59, esp. pp. 7–8]. 61 [44, p. 169]. A summary of the key points made by the persuader in each exchange is given by Cheng Chung-ying [19, pp. 150–151). 62 This is the only line in which the verb yu¯ e is missing. Since all the other lines are consistently introduced by yu¯e, we can add it here with confidence.

2 Argumentation and Persuasion in Classical Chinese Literature

39

P曰:

「可。」

O曰:

「何哉。」

P曰:

「馬者所以命形也。白者所以命色也。命色者非命形也故曰白馬非 馬。」 “To say that ‘white horse is not horse,’ is it admissible?” “It is admissible.” “How is that?” “‘Horse’ is what denotes shape, ‘white’ is what denotes color. What denotes color is not what denotes shape.63 Therefore I say that ‘white horse’ is not ‘horse’.”

O: P: O: P:

The dialogue opens with the opponent’s rhetorical question that doubts the admissibility of what in his view is evidently an outrageous statement, contrary to good sense. The whole dialogue is structured in this way, with the persuader arguing from the starting point of a topic introduced by the opponent’s questions or at times scandalized replies. The core issue of the whole dialogue finds its expression here, in the very first exchange. These first few lines introduce the persuader’s notoriously puzzling statement—that it is admissible to claim that a white horse is not a horse under certain conditions—and reveals on which grounds he is able to support such a bizarre position. O曰: 「有白馬不可謂無馬也。不可謂無馬者非馬也。有白馬為有馬白之 非馬何也。」 P曰: 「求馬黃黑馬皆可致。求白馬黃黑馬不可致。使白馬乃馬也是所求 一也。所求一者白者不異馬也。所求不異如黃黑馬有可有不可何也。 可與不可其相非明。故黃黑馬一也而可以應有馬, 而不可以應有白馬 。是白馬之非馬審矣。」 O: “There being a white horse, one cannot say that there is no horse. Of what one cannot say there not being a horse, you say it is not horse. There being a white horse is deemed as there being a horse, how can it be that a white one is not a horse? P: “If you were looking for a horse, both a dun horse or a black horse could be offered to you. If you were looking for a white horse, a dun horse or a black horse could not be offered to you. If you now consider a white horse as a horse, what you were looking for would be one and the same thing. If what you were looking for was one and the same thing, ‘white’ would not be different from ‘horse.’ If what you were looking for was one and the same thing, it would be like the case of the dun horse and the black horse that are admissible in one instance and inadmissible in the other instance, how could it be possible? It is clear that admissible and inadmissible mutually exclude each other. Therefore, a dun horse and a black horse are the same in the sense that they can correspond to there 63 See

[134, p. 26].

40

L. Indraccolo

being a horse, but they cannot correspond to there being a white horse. Indeed, this proves that ‘white horse’ is not ‘horse.’ The opponent rigorously applies deductive reasoning to the case in point, and claims that a white horse must necessarily be a horse. This is because if there were a real white horse just in front of the two of them, then it would be clearly impossible to deny that a white horse is a horse. This is a matter of fact, the evidence of which nobody can deny—or at least the opponent believes so. The persuader is perfectly aware that the opponent is right. However, he ingeniously shifts the topic and takes a long logical detour, forcing the opponent to admit that looking for a horse with the hide of a specific color is not exactly the same as looking for a generic horse. If somebody was looking for a horse with the hide of a specific color, only a horse with the hide of that specific color would suit their needs. Thus, the opponent is eventually forced to fall into self-contradiction. O曰: 「以馬之有色為非馬。天下非有無色之馬也。天下無馬可乎。」 P曰: 「馬固有色故有白馬。 使馬無色有馬如64 已耳安取白馬。 故白者非 馬也。白馬者馬與白也。馬與白{非}65 馬也。故曰白馬非馬也。」 O: “You deem a horse of a certain color not to be a horse, but there are no colorless horses in the world. Might it be that there are no horses at all in the world then? P: “Horses necessarily have colors, that is why there can be white horses. If horses did not have colors, there would be just plain horses as such, and that would be it. How could we pick out white horses? Therefore, a white one is not the same as ‘horse.’ ‘White horse’ is made of ‘horse’ and ‘white’ combined. ‘Horse’ and ‘white’ combined is not ‘horse.’ Therefore I say that ‘white horse’ is not ‘horse.’” The opponent fails to notice where the persuader’s line of argument employed in the preceding exchange is faulty and misleading, losing the opportunity to produce an effective counterargument. The opponent accuses the persuader of claiming that only colorless horses could exist in the world. However, the persuader candidly replies that of course horses necessarily have a color. If horses did not have any color, it would be impossible to talk about white horses at all! As Christoph Harbsmeier points out, “here the sophist uses the rather subtle argumentative device of showing that the conclusion he wishes to demonstrate is actually implicit in what his opponent is saying.”66 O曰: 「馬未與白為馬。白未與馬為白。合馬與白復名白馬是相與以不相 與為名 。未可。故曰白馬非馬未可。」 P曰: 「以有白馬為有馬。謂有白馬為有黃馬可乎。」 O曰: 「未可。」 如 = ér 而. [142, p. 12, footnote 2]; [134, p. 26]; [44, p. 188]. Qián Mù [115, p. 49] and [18, p. 65], I add the negative copula f¯ei 非 [“not to be (the same as)”] that is used as negation for nominal sentences before “horse” mˇa 馬. It is highly likely that the sentence is corrupted and that a few characters are missing here. See [44, p. 188], [72, p. 33, footnote 3]. 66 [48, p. 307]. 64 rú

65 Following

2 Argumentation and Persuasion in Classical Chinese Literature

P曰:

O:

P: O: P:

41

「以有馬為異有黃馬是異黃馬於馬也。異黃馬於馬是以黃馬為非馬 。以黃馬為非馬而以白馬為有馬。此飛者入池而棺槨異處。此天下 之悖言亂辭也。」 “You deem ‘horse’ not yet67 combined with ‘white’ to be ‘horse,’ and ‘white’ not yet combined with ‘horse’ to be ‘white.’ To call by the compound ‘white horse’68 the union of ‘horse’ and ‘white’ combined is to call them by the names they had when uncombined when they are combined: it is inadmissible. Therefore, I say that ‘white horse is not horse’ is inadmissible.” “If there being a white horse is deemed there being a horse, is it admissible to say that there being a white horse is deemed there being a dun horse?” “It is inadmissible.” “To deem there being a horse different from there being a dun horse is to differentiate ‘dun horse’ from ‘horse.’ To differentiate ‘dun horse’ from ‘horse’ is to deem ‘dun horse’ not to be ‘horse.’ To deem ‘dun horse’ not to be ‘horse’ and at the same time to deem ‘white horse’ to be ‘horse’! This is like if flying animals were diving in a pond and the inner and outer coffins were swapped places. These are the most contradictory and incoherent words in the world!”

The persuader distracts the opponent by changing the topic once again. Thereby, he succeeds in making the opponent admit that eventually a white horse is not exactly the same as a horse of another color. Finally, the persuader turns the opponent’s answer to his own advantage, twisting his words and using them against him. The persuader’s deceptive argument is structured as a chain-reasoning, and resorts to the technique of reductio ad absurdum. The persuader concludes with two hyperbolic metaphors69 to underline the absurdity of the situation and the vicious circle in which the opponent just got himself stuck into. After this analysis we can conclude that even if no proper handbook nor list of rhetorical techniques and devices have been handed down in the Classical Chinese tradition, it is still possible to demonstrate that rhetoric—or at least certain forms of “rhetoric”70 —was largely practiced and played a key role in both governmental activity and in the intellectual sphere in Early China. There is undoubtedly abundant evidence of the practice and the performance of this type of “rhetoric” in Classical Chinese received literature, as numerous relevant examples show.71 In particular, the two techniques of “argumentation” and “persuasion” among others are well 67 [44,

p. 188]. compound the name ‘white horse’ for horse and white joined together is to give them when combined their names when uncombined” [43, p. 146]. 69 Christoph Harbsmeier remarks the importance of this only apparently frivolous “triumphantly flowery rhetorical picture,” suggesting that “the stylistic flourish is rather important, because it shows that the dialogue as we have it is not entirely thought of as an algebraic disputation but as related to a real situation at court” [48, p. 309]. 70 [61]. 71 [99]. 68 “To

42

L. Indraccolo

attested throughout the early Chinese period. They are explicitly acknowledged as complementary techniques that have to be fully mastered by “wandering persuaders,” and represent a fundamental acquired skill in their repertoire.72 There is no proper technical description of the main features of these techniques in currently available sources. This is rather unsurprising and does not necessarily speak to a lesser degree of importance attributed to these techniques, nor is it indicative of a relatively “low” status of “rhetoric” in early Chinese intellectual culture. As it is often the case in Classical Chinese literature, a detailed and systematic treatment of a certain topic as we would expect it in the Western tradition is the exception rather than the rule, especially for the earlier periods. Key information typically has to be extrapolated and deduced indirectly through the close scrutiny of texts. Accordingly, through a close analysis of rhetorical pieces preserved in pre-imperial and early imperial received texts, we can assess that it is indeed possible to identify a set of fairly consistent structural features, and a rich repository of rhetorical and stylistic devices used in the make-up of early Chinese texts in which the techniques of “argumentation” and “persuasion” are employed. On this basis, it can be concluded that these texts are not a casual array of disparate and incoherent materials, but rather the result of careful authorial/editorial choices and that they abide by a well-established set of compositional rules and conventions.73

Bibliography 1. Allan, S. (1997). The way of water—The sprouts of virtue. State University of New York Press. 2. Baxter, W. H. (1998). Situating the language of the Lao-tzu: The probable date of the Tao-teching. In L. Kohn & M. LaFargue (Eds.), Lao-tzu and the Tao-te-ching (pp. 231–253). State University of New York Press. 3. Baxter, W. H., & Sagart, L. (2014a). Old Chinese: A new reconstruction. Oxford University Press. 4. Baxter, W. H., & Sagart, L. (2014b), Baxter-Sagart old Chinese reconstruction version 1.1. http://ocbaxtersagart.lsait.lsa.umich.edu/BaxterSagartOCbyMandarinMC2014-09-20.pdf. 5. Beecroft, A. (2010). Authorship and cultural identity in early Greece and China: Patterns of literary circulation. Cambridge University Press. 6. Behr, W. (2008). Reimende Bronzeinschriften und die Entstehung der chinesischen Endreimdichtung. Projekt Verlag. 7. Behr, W., & Gentz, J. (Eds.). (2005). Komposition und Konnotation – Figuren der Kunstprosa im alten China, special issue, Bochumer Jahrbuch für ostasiatische Forschung, 29. 8. Birrell, A. M. (1993). Chinese mythology: An introduction. The Johns Hopkins University Press. 9. Bodde, D. (1961). Myths of ancient China. In S. N. Kramer (Ed.), Mythologies of the ancient world (pp. 369–408). Doubleday & Co. 10. Boltz, W. G. (1997). Manuscripts with transmitted counterparts. In E. L. Shaughnessy (Ed.), New sources of early Chinese history: An introduction to the reading of inscriptions and manuscripts (pp. 253–283). Society for the Study of Early China and the Institute of East Asian Studies, University of California. 72 [75]. 73 [120].

2 Argumentation and Persuasion in Classical Chinese Literature

43

11. Boltz, W. G. (1999). The fourth-century B.C. Guodiann manuscripts from Chuu and the composition of the Laotzyy. Journal of the American Oriental Society, 119, 590–608. 12. Boltz, W. G. (2005). The composite nature of early Chinese texts. In M. Kern (Ed.), Text and ritual in early China (pp. 50–78). University of Washington Press. 13. Brooks, E. B., & A. Brooks, A. T. (1998). The original Analects: Sayings of Confucius and his successors. Columbia University Press. 14. Chan, A. K.-L. (Ed.). (2002). Mencius—Contexts and interpretations. University of Hawai’i Press. 15. Chang, K.-C. (1976). A classification of Shang and Chou myths. In K.-C. Chang, Early Chinese civilization: Anthropological perspectives (pp. 149–173). Harvard University Press. 16. Chen, D. (2005). Three meta-questions in epistemology: Rethinking some metaphors in the Zhuangzi. Journal of Chinese Philosophy, 32(3), 493–507. 17. Chen, G.-M., & Holt, R. G. (2002). Persuasion through the water metaphor in the Dao De Jing. Intercultural Communication Studies, 11, 153–171. 18. Chén, Z. 陳柱 (1937). G¯ongs¯un Lóngzˇı jíjiˇe 公孫龍子集解 [Collected explanations of Master Gongsun Long]. Sh¯angwù Yìnsh¯uguˇan. 19. Cheng, C.-Y. (1997). Philosophical significance of Gongsun Long: A new interpretation of theory of zhi as meaning and reference. Journal of Chinese Philosophy, 24(2), 139–178. 20. Crump, J. I., Jr. (1964). 戰國策 Intrigues: Studies of the Chan-kuo Ts’e. The University of Michigan Press. 21. Crump, J. I., Jr. (1979). Chan-kuo Ts’e. Chinese Materials Center Inc. 22. Csikszentmihalyi, M. (2002). Traditional taxonomies and revealed texts in the Han. In L. Kohn & H. D. Roth (Eds.), Daoist identity—History, lineage, and ritual (pp. 81–101). University of Hawai’i Press. 23. Csikszentmihalyi, M., & Nylan, M. (2003). Constructing lineages and inventing traditions through exemplary figures in early China. T’oung Pao, 83, 59–99. 24. De Reu, W. (2006). Right words seem wrong: Neglected paradoxes in early Chinese philosophical texts. Philosophy East and West, 56(2), 281–300. 25. De Reu, W. (2010). How to throw a pot: The centrality of the potter’s wheel in the Zhuangzi. Asian Philosophy, 20(1), 43–66. 26. Defoort, C. (2005–06). Argumentation and persuasion in ancient Chinese texts—Introduction. Oriens Extremus, 45, 91–98. 27. Defoort, C. (2015). Heavy and light body parts: The weighing metaphor in early Chinese dialogues. Early China, 38(3), 55–77. 28. Denecke, W. (2010). The dynamics of Masters literature—Early Chinese thought from Confucius to Han Feizi. Harvard University Press. 29. Elman, B. A. (2000). A cultural history of civil examinations in late imperial China. University of California Press. 30. Elman, B. A., & Kern, M. (Eds.) (2010). Statecraft and classical learning—The Rituals of Zhou in east Asian history. E.J. Brill. 31. van Els, P., Graziani, R., Pines, Y., & Sabattini, E. (Eds.). (2012). Political rhetoric in early China/Rhétorique et politique en Chine ancienne, special Issue, Extrême Orient-Extrême Occident, 34. Presses Universitaires de Vincennes. 32. van Ess, H. (2005–06). Argument and persuasion in the first chapter of the Huainan zi and its use of particles. Oriens Extremus, 45, 250–270. 33. Forke, A. (1901–02). The Chinese sophists. Journal of the China Branch of the Royal Asiatic Society, 34, 1–100. 34. Fraser, C. (2020). Paradoxes in the school of names. In Y.-M. Fung (Ed.), Dao companion to Chinese philosophy of logic (pp. 285–307). Springer. 35. Galambos, I. (2006). Ortography of early Chinese writing: Evidence from newly excavated manuscripts. Department of East Asian Studies, Eötvös Lórand University. 36. Galvany, A. (2012). Sly mouths and silver tongues: The dynamics of psychological persuasion in ancient China. In P. van Els, R. Graziani, Y. Pines, & E. Sabattini (Eds.), Political rhetoric in early China/Rhétorique et politique en Chine ancienne, special issue, Extrême Orient-Extrême Occident, 34 (pp. 15–40). Presses Universitaires de Vincennes.

44

L. Indraccolo

37. Garrett, M. M. (1993). Classical Chinese conceptions of argumentation and persuasion. Argumentation and Advocacy, 29(3), 105–115. 38. Gassmann, R. H. (2016). Menzius. Eine kritische Rekonstruktion mit kommentierter Übersetzung. Welten Ostasiens, 22 (1–3). de Gruyter. 39. Gentz, J. (2005). Zwischen den Argumenten lesen. Doppelt gerichtete Parallelismen zwischen Argumenten als zentrale Thesen in frühen chinesischen Texten. In W. Behr & J. Gentz (Eds.), Komposition und Konnotation—Figuren der Kunstprosa im alten China, special issue, Bochumer Jahrbuch für ostasiatische Forschung, 29, 35–56. 40. Gentz, J., & Meyer, D. (Eds.) (2015). Literary forms of argument in early China. E.J. Brill. 41. Goldin, P. R. (1993). Miching mallecho: The Zhanguoce and classical rhetoric. Sino-Platonic Papers, 41, 1–27. 42. Goldin, P. R. (2005). Rhetoric and machination in Stratagems of the Warring States. In P. R. Goldin, After Confucius—Studies in early Chinese philosophy (pp. 6–89). University of Hawai’i Press. 43. Graham, A. C. (1964–65). Two dialogues in the Kung-sun Lung Tzu: ‘White horse’ and ‘Left and right.’ Asia Major, 11, 128–152. 44. Graham, A. C. (1986). Studies in Chinese philosophy and philosophical literature. The Institute of East Asian Philosophies, National University of Singapore. 45. Graham, A. C. (1989). Disputers of the Tao: Philosophical argument in ancient China. Open Court. 46. Graziani, R. (2012). Rhetoric that kills, rhetoric that heals. In P. van Els, R. Graziani, Y. Pines, & E. Sabattini (Eds.), Political rhetoric in early China/Rhétorique et politique en Chine ancienne, special issue, Extrême Orient-Extrême Occident, 34 (pp. 41–78). Presses Universitaires de Vincennes. 47. Graziani, R. (2014). Of words and swords: Therapeutic imagination in action—A study of chapter 30 of the Zhuangzi ‘Shuo jian’ 說劍. Philosophy East and West, 64(2), 375–403. 48. Harbsmeier, C. (1998). Language and logic. In J. Needham (Ed.), Science and civilization in China (Vol. 7). Cambridge University Press. 49. Harbsmeier, C. (2002). The rhetoric of premodern prose style. In V. Mair (Ed.), The Columbia history of Chinese literature (pp. 881–908). Columbia University Press. 50. Hsu, C.-Y. (1965). Ancient China in transition: An analysis of social mobility, 722–222 B.C. Stanford University Press. 51. Huang, C.-C., Paul, G., & Roetz, H. (Eds.). (2008). The Book of Mencius and its reception in China and beyond. Harassowitz Verlag. 52. Hunter, M. (2013). The difficulty with the ‘Difficulties of persuasion’ (“Shuinan” 說難). In P. R. Goldin (Ed.), Dao companion to the philosophy of Han Fei. Springer. 53. Indraccolo, L. (2010). Gongsun Long and the Gongsun Longzi—Authorship and textual variation in a multilayered text. Ph.D. Dissertation, Ca’ Foscari University, Venice. Open access publication: http://dspace.unive.it/handle/10579/922. 54. Indraccolo, L. (2013). Textual criticism of the Gongsun Longzi: The Daozang and Shuofu editions. T’oung Pao, 99(4–5), 249–267. 55. Indraccolo, L. (2014). What is rhetoric anyway? Briared in words in ancient China. Asiatische Studien/Études asiatiques, 68(1), 331–341. 56. Indraccolo, L. (2015). Words not to be spoken, measures not to be taken—The ‘rhetoric of antirhetoric’ in early Chinese literature. Paper presented at the 26th Conference of the German Association for Chinese Studies (Deutsche Vereinigung für China Studien) (under review for publication). 57. Indraccolo, L. (2016a). The ‘white horse’, the ‘three-legged chicken’ and other paradoxes in classical Chinese literature. Antiquorum Philosophia, 10, 67–88. 58. Indraccolo, L. (2016b). Be smart, be like a larva! Insects and invertebrates in classical Chinese analogical thinking. Paper presented at the Joint East Asia Studies Conference 2016, SOAS, London, unpublished. 59. Indraccolo, L. (2017). The white horse argument. Philosophy Compass, 12(10), 11 pp. https:// doi.org/10.1111/phc3.12434.

2 Argumentation and Persuasion in Classical Chinese Literature

45

60. Indraccolo, L. (2020). Argumentation (biàn 辯). In Y.-M. Fung (Ed.), Dao companion to Chinese philosophy of logic (pp. 171–180). Springer. 61. Indraccolo, L., & Behr, W. (2014). Introduction. In L. Indraccolo & W. Behr (Eds.), Masters of disguise? Conceptions and misconceptions of ‘rhetoric’ in Chinese antiquity” special issue. Asiatische Studien/Études asiatiques, 68(4), 889–913. 62. Karlgren, B. (1946). Legends and cults in ancient China. Bulletin of the Museum of Far Eastern Antiquities, 18, 199–365. 63. Kern, M. (2000). ‘Persuasion’ or ‘treatise’? The prose genre shui 說 and shuo 說 in the light of the Guwenci leizhuan of 1779. In L. Bieg, E. von Mende, & M. Siebert (Eds.), Ad Seres et Tungusos – Festschrift für Martin Gimm zu seinem 65. Geburtstag am 25. Mai 1995 (pp. 221–245). Harassowitz Verlag. 64. Kern, M. (2002). Methodological reflections on the analysis of the textual variants and the modes of manuscript production in early China. Journal of East Asian Archaeology, 4(1–4), 143–181. 65. Kern, M. (Ed.). (2005). Text and ritual in early China. University of Washington Press. 66. Kern, M. (2005). The Odes in excavated manuscripts. In M. Kern (Ed.), Text and ritual in early China (pp. 149–193). University of Washington Press. 67. Kern, M. (2005). Quotation and the Confucian canon in early Chinese manuscripts: The case of ‘Zi yi’ (Black robes). Asiatische Studien/Études asiatiques, 59(1), 293–332. 68. Kern, M. (2015). Speaking of poetry: Pattern and argument in the ‘Kongzi Shilun.’ In J. Gentz & D. Meyer (Eds.), Literary forms of argument in early China (pp. 175–200). E.J. Brill. 69. Kern, M., & Meyer, D. (Eds.). (2017). Origins of Chinese political philosophy—Studies in the composition and thought of the Shangshu (Classic of Documents). E.J. Brill. 70. Kim, C. C. (2006). Zhuangzi and the nature of metaphor. Philosophy East and West, 56(3), 370–391. 71. Knoblock, J. (1990). Xunzi: A translation and study of the complete works—Vol II, Books 7–16. Stanford University Press. 72. Kou, P.-K. I. (1953). Deux sophistes chinois – Houi Che et Kong-Souen Long. Bibliothèque de l’Institut des Hautes Etudes Chinoises, Imprimerie Nationale. 73. Krijgsman, R. (2014). Travelling sayings as carriers of philosophical debate: From the intertextuality of the *Yucong 語叢 to the dynamics of cultural memory and authorship in early China. Asiatische Studien/Études asiatiques, 68(1), 83–116. 74. Krijgsman, R. (2016). The textualization of cultural memory in early Chinese manuscripts. Ph.D. Dissertation, Oxford University. 75. Kroll, J. L. (1985–86). Disputation in ancient Chinese culture. Early China, 11–12, 119–145. 76. LaFargue, M. (2003). More ‘Mencius-on-human-nature’ discussions: What are they about?— Review article of Adam Kam-leung Chan, editor, Mencius: Contexts and interpretations. Honolulu, HI: University of Hawai’i Press, 2002. Chinese Review International, 10(1), 1–28. 77. Lai, G. (2017). Textual fluidity and fixity in early Chinese manuscript culture. Chinese Studies in History, 50(3), 172–184. 78. Lau, D. C. (Trans. and Ed.). (2004 [1970]). Mencius. Penguin Books Limited. 79. Le Blanc, C. (1985–86). A re-examination of the myth of Huang-ti. Journal of Chinese Religions, 13–14, 45–63. 80. Le Blanc, C., & Mathieu, R. (Eds.). (2007). Approches critiques de la mythologie chinoise. Les Presses de l’Université de Montréal. 81. Levi, J. (1992). L’art de la persuasion à l’epoque des Royaumes Combattants (Ve –IIIe Siècles a.v. J.C.). Extrême Orient-Extrême Occident, 14, 49–89. 82. Lewis, M. E. (1999). Writing and authority in early China. State University of New York Press. 83. Lewis, M. E. (2006). The flood myths of early China. State University of New York Press. 84. Lewis, M. E. (2009). The mythology of early China. In J. Lagerwey & M. Kalinowski (Eds.), Early Chinese religion. Part one: Shang through Han (1250 BC–220 AD) (pp. 543–594). E.J. Brill.

46

L. Indraccolo

85. Li, F., & Prager Branner, D. (Eds.). (2011). Writing and literacy in early China—Studies from the Columbia early China seminar. University of Washington Press. 86. Loewe, M. (2012). ‘Confucian’ values and practices in Han China. T’oung Pao, 98(1–3), 1–30. 87. Loewe, M., & Shaughnessy, E. L. (Eds.). (1999). The Cambridge history of ancient China— From the origins of civilizations to 221 B.C. Cambridge University Press. 88. Loy, H.-C. (2014). Language and ethics in the Analects. In A. Olberding (Ed.), Dao companion to the Analects (pp. 137–158). Springer. 89. Lu, X. (1998). Rhetoric in ancient China, fifth to third century B.C.E.—A comparison with classical Greek rhetoric. University of South Carolina Press. 90. Lu, X., & Frank, D. A. (1993). On the study of ancient Chinese rhetoric/bian 辩. Western Journal of Communication, 57(4), 445–463. 91. Ma, R. (2000). Water-related figurative language in the rhetoric of the Mencius. In A. Gonzales & D. V. Tanno (Eds.), Rhetoric in intercultural contexts (pp. 119–129). Sage. 92. Major, J. S. (2014). Animals and animal metaphors in Huainanzi. In S. A. Queen & M. J. Puett (Eds.), The Huainanzi and textual production in early China (pp. 133–151). E.J. Brill. 93. Maspero, H. (1924). Légendes mythologiques dans le Chou King. Journal Asiatique, 204, 1–100. 94. Mei, Y.-P. (1953). The Kung-sun Lung Tzu with a translation into English. Harvard Journal of Asiatic Studies, 16(3–4), 404–437. 95. Meyer, D. (2011). Philosophy on bamboo: Text and the production of meaning in early China. E.J. Brill. 96. Meyer, D. (2017). Recontextualization and memory production: Debates on rulership as reconstructed from ‘Gu ming’ 顧命 (Testimonial charge). In M. Kern & D. Meyer (Eds.), Origins of Chinese political philosophy: Studies in the composition and thought of the Classic of Documents (pp. 106–145). E.J. Brill. 97. Meyer, D. (Forthcoming). Documentation and argument in early China: The Shàngsh¯u 尚書 (Venerated Scriptures) and the “Sh¯u” traditions. de Gruyter. 98. Meyer, D. & Indraccolo, L. (2021). Sentences, paragraphs, and sections. In J. W. Chen, A. Detwyler, X. Liu, C. Nugent, & B. Rusk (Eds.), Literary information in China—A history (pp. 113–118). Columbia University Press. 99. Mittler, B., & Wuthenow, A.-B. (2008). Rhetoric and stylistics in East Asia. In U. Fix, A. Gardt, & J. Knape (Eds.), Rhetorik und Stylistik, Ein internationales Handbuch historischer und systematischer Forschung 1 (pp. 2027–2039). de Gruyter. 100. Nivison, D. S. (1996). The ways of Confucianism—Investigations in Chinese philosophy. Open Court. 101. Nylan, M. (1997). Han classicists writing in dialogue about their own tradition. Philosophy East and West, 47(2), 133–188. 102. Nylan, M. (1999). A problematic model—The Han ‘orthodox synthesis,’ then and now”. In K.W. Chow, O.-C. Ng, & J. B. Henderson (Eds.), Imagining boundaries—Changing Confucian doctrines, texts, and hermeneutics (pp. 17–56). State University of New York Press. 103. Nylan, M. (2000). Textual authority in pre-Han and Han. Early China, 25, 205–258. 104. Nylan, M. (2001). The five “Confucian” Classics. Yale University Press. 105. Nylan, M. (2009). Classics without canonization: Learning and authority in Qin and Han. In J. Lagerwey & M. Kalinowski (Eds.), Early Chinese religion—Part one: Shang through Han (1250 BC–220 AD) (pp. 722–776). E.J. Brill. 106. Nylan, M. (2014). Logical connectives in the Huainanzi. In S. A. Queen & M. J. Puett (Eds.), The Huainanzi and textual production in early China (pp. 225–266). E.J. Brill. 107. Nylan, M. (2016). Han views of the Qin legacy and late Western Han ‘classical turn.’ Bulletin of the Museum of Far Eastern Antiquities, 79–80, 51–98. 108. Oliver, R. T. (1971). Communication and culture in ancient India and China. Syracuse University Press. 109. Perleberg, M. (1952). The works of Kung-sun Lung-tzu. LTD.

2 Argumentation and Persuasion in Classical Chinese Literature

47

110. Petersen, J. Ø. (1995). Which books did the first emperor of Ch’in burn? On the meaning of pai chia in early Chinese sources. Monumenta Serica, 43, 1–52. 111. Pines, Y. (2009). Envisioning eternal empire – Chinese political thought of the Warring States era. University of Hawai’i Press. 112. Pines, Y. (2012). The everlasting empire—The political culture of ancient China and its imperial legacy. Princeton University Press. 113. Pines, Y., Goldin, P. R., & Kern, M. (Eds.). (2015). Ideology of power and power of ideology in early China. E.J. Brill. 114. Pitner, M. G. (2012). The wasp and the worm: Getting inside the body of the sage. In B. Gardenour & M. Tadd (Eds.), Parasites, worms, and the human body in religion and culture (pp. 41–64). Peter Lang. 115. Qián, M. 錢穆 (1934). Huì Sh¯ı G¯ongs¯un Lóng 惠施公孫龍 [Hui Shi and Gongsun Long]. Shànghˇai Sh¯udiàn. 116. Qián, M. 錢穆 [1986 (1935)]. Xi¯an Qín zh¯uzˇı jìnián 先秦諸子繫年 [Dates of Pre-Qin philosophers]. D¯ongdà Ch¯ubˇan. 117. Raphals, L. (1992). Knowing words—Wisdom and cunning in the classical traditions of China and Greece. Cornell University Press. 118. Raphals, L. (1998). On Hui Shi. In R. T. Ames (Ed.), Wandering at ease in the Zhuangzi (pp. 143–161). State University of New York Press. 119. Reding, J.-P. (1985). Les fondements philosophiques de la réthorique chez les sophistes grecs et chez les sophistes chinois. Peter Lang. 120. Richter, M. L. (2013). The embodied text—Establishing textual identity in early Chinese manuscripts. E.J. Brill. 121. Ryden, E. (1996). Was Confucius a Confucian? Confusion over the use of the term ‘school’ in Chinese philosophy. Early China News, 9(5–9), 28–29. 122. Schaberg, D. (2016). The ruling mind: Persuasion and the origins of Chinese psychology. In P. Varsano, (Ed.), The rhetoric of hiddenness in traditional Chinese culture. State University of New York Press. 123. Schwartz, B. I. (1985). The world of thought in ancient China. Harvard University Press. 124. Schwermann, C. (2005). Collage-Technik als Kompositionsprinzip klassischer chinesischer Prosa: Der Aufbau des Kapitels ‚T¯ang wèn‘ (Die Fragen des T¯ang) im Liè zˇı. In W. Behr & J. Gentz (Eds.), Komposition und Konnotation – Figuren der Kunstprosa im alten China, special issue, Bochumer Jahrbuch für ostasiatische Forschung, 29, 125–160. 125. Schwermann, C., & Steineck, R. C. (Eds.). (2014). That wonderful composite called author— Authorship in East Asian literatures from the beginning to the seventeenth century. E.J. Brill. 126. Shun, K.-L. (2000). Mencius and early Chinese thought. Stanford University Press. 127. Sivin, N. (1995). The myth of the naturalists. In N. Sivin (Ed.), Medicine, philosophy and religion in ancient China—Researches and reflections (pp. 1–33). Variorum, Ashgate Publishing. 128. Slingerland, E. (2003). Effortless action: Wu-wei as conceptual metaphor and spiritual ideal in early China. Oxford University Press. 129. Slingerland, E. (2011). Metaphor and meaning in early China. Dao, 10, 1–30. 130. Smith, K. (2003). Sima Tan and the invention of Daoism, ‘Legalism’ et cetera. The Journal of Asian Studies, 62(1), 129–156. 131. Solomon, B. (1969). The assumptions of Hui-tzu. Monumenta Serica, 28, 1–40. 132. Sterckx, R. (2002). The animal and the daemon in early China. State University of New York Press. 133. Stevenson, F. W. (1991). South has (no) limits: Relative and absolute meaning in Hui Shih’s ten points. Tamkang Review, 21(4), 325–346. 134. Tán, J. 譚戒甫 [2006 (1956)]. G¯ongs¯un Lóngzˇı xíngmíng f¯aw¯ei 公孫龍子形名發微 [Subtleties about the theory of “forms and names” in the Gongsun Longzi]. Wˇuhàn Dàxué Ch¯ubˇanshè. 135. Tian, X. (2006). The twilight of the Masters: Masters literature (zishu) in early Medieval China. Journal of the American Oriental Society, 126(4), 465–486.

48

L. Indraccolo

136. Tsien, T.-H. (1962). Written on bamboo and silk: The beginnings of Chinese books and inscriptions. University of Chicago Press. 137. Valesio, P. (1980). Novantiqua—Rhetorics as a contemporary theory. Indiana University Press. 138. Vankeerberghen, G. (2005–06). Choosing balance: Weighing (‘quan’ 權) as a metaphor for action in early Chinese texts. Early China, 30, 47–89. 139. Vogelsang, K. (2019). The smart, the stupid, and the enlightened—On the origin of political consultation in ancient China. Paper presented at the 21st International Conference on Chinese Philosophy, University of Berne, unpublished. 140. Vogt, P. N. (2012). Between kin and kings: Social aspects of Western Zhou ritual. Ph.D. Dissertation, Columbia University. 141. Wagner, R. G. (2015). A building block of Chinese argumentation: Initial fu 夫 as a phrase status marker. In J. Gentz & D. Meyer (Eds.), Literary forms of argument in early China (pp. 37–66). E.J. Brill. 142. Wáng, G. 王琯 [1992 (1988)]. G¯ongs¯un Lóngzˇı xuánjiˇe 公孫龍子懸解 [Explanations of unresolved issues in the Gongsun Longzi]. Zh¯onghuá Sh¯ujú. 143. Wang, Y. (2004). The strategies of ‘goblet words’: Indirect communication in the Zhuangzi. Journal of Chinese Philosophy, 31(2), 195–218. 144. Weingarten, O. (2014). The sage as teacher and source of knowledge: Editorial strategies and formulaic utterances in Confucius dialogues. Asiatische Studien/Ètudes asiatiques, 68(4), 1175–1223. 145. Weingarten, O. (2015). Debates around Jixia: Argument and intertextuality in Warring States writings associated with Qi. Journal of the American Oriental Society, 135(2), 283–307. 146. Wilkinson, E. (2018). Chinese history—A new manual. Fifth Edition. Harvard-Yenching Institute Monograph Series. Harvard University Asia Center. Harvard University Press. 147. Zh¯u, F. 朱凤瀚 (2004). Sh¯ang Zh¯ou ji¯azú xíngtài yánji¯u 商周家族形態研究 [A study on the structure of the clans in the Sh¯ang and Zh¯ou dynasties]. Ti¯anj¯ın Gˇují Ch¯ubˇanshè.

Chapter 3

Gorgias and the Weakness of Logos Wolfgang-Rainer Mann

Abstract After briefly considering Plato’s objections to rhetoric—it disregards the truth, aiming only to persuade, and it manipulates our emotions rather than instructing us—I turn to the historical Gorgias. The ‘Encomium of Helen’ ascribes to logos (speech) virtually all-powerful capacities for persuasion, seduction, and even bewitchment. Here Gorgias (seemingly) celebrates the very things Plato rejects. Yet in the ‘Defense of Palamedes’ considerable anxieties about whether logos actually does possess such (persuasive) strength are voiced: the weakness, not the power, of logos comes to occupy center stage. Gorgias’ own views thus are likely more complex than Plato allows.

3.1 Some Preliminaries Before turning to Gorgias himself, a few words about Plato, who is implicitly committed to a kind of demarcation project, a way of distinguishing (genuine) philosophy and its (legitimate) form of discourse from other enterprises which, in unfavorable circumstances, or in the presence of the unwary, could pass for philosophy. Most obviously and most famously, Plato seeks to distinguish philosophy from what he

W.-R. Mann (B) Department of Philosophy, Columbia University, New York, NY, USA e-mail: [email protected] © Springer Nature Switzerland AG 2021 J. A. Bjelde et al. (eds.), Essays on Argumentation in Antiquity, Argumentation Library 39, https://doi.org/10.1007/978-3-030-70817-7_3

49

50

W.-R. Mann

calls rhetoric.1 Here it can seem as if an easily identifiable formal feature is a sufficient indicator of the difference: rhetoric, as Plato conceives it, involves speaking at length. Plato’s Socrates, by contrast, especially in those moments when he seeks to steer or cajole his interlocutors towards his own preferred mode of discourse, stresses brevity. The difference, however, is not only a matter of length. It also concerns whether one party engages in what amounts to a kind of monologue (with another party playing the role of audience), or whether two parties instead engage in dialogue, in back-and-forth exchanges, in particular, exchanges consisting of questions and answers. Now, in principle, nothing appears to preclude monologue-like brevity: short quips or gnomic utterances—things like ‘Know thyself’ or ‘Nothing to excess’—clearly fit the bill.2 Likewise, there seems to be nothing to rule out ex ante the possibility of dialogue-like, yet lengthy, exchanges: paired speeches, where the speech that follows engages with, and responds to, the one that precedes it, and in the limit case, seeks to deny or rebut what had been said in the first speech. There would thus be, as it were, a built-in opposition (sometimes greater, sometimes lesser) between speeches so related. One could use the verb antileg¯o to label speaking in this way, and the noun antilogia for an occasion where such speeches are being given—as it is, say, used by, Herodotus to refer to a trial (9.88), or by Thucydides to refer to occasions which could be, but should not be, mistaken for trials (e.g. at 1.73).

Schiappa, in [25], and T. Cole, in [5], go so far as to argue that the very word ῥητορική is a Platonic coinage. On the origins of the term, see also Schiappa [26], 14–29, and [27], 39–58. N. O’Sullivan, in [23], and G. Pendrick, in [24], by contrast, argue that the term must have been in use earlier. This issue is in principle separable from one with which it is sometimes coupled, the question of whether the earliest so-called handbooks (τέχναι) of rhetoric were, roughly speaking, theoretical works, consisting of, among other things, precepts about how speeches should be structured, or were rather collections of something like sample speeches. (Cf. Aristotle, SE 183b36–184a1, with Solmsen’s emendation, in his [31], 214, n. 3, on Gorgias’s practice of ‘handing out [sc. to his students] speeches (λόγοι ῥητορικοί) to be learned by heart (ἐκμανθάνειν)’.) To the extent that there is a consensus on the matter, it has been shifting more towards the second picture. See also S. Usher [34], 1–6, for a brief, sensible discussion, with extensive references to earlier literature, of the earliest practitioners of oratory/rhetoric. Usher there also provides a helpful introduction to the roles of Corax and Tisias, two somewhat shadowy (for us) figures, who in antiquity were frequently credited with having ‘pioneer[ed] . . instruction [in] oratory’ ([34], 2, n. 4)—unless the origin of artful speaking was projected all the way back to Homer; see, e.g. G. Kennedy [12], A. J. Karp [11], and P. Toohey [33], esp. 153–162. Resolving these issues luckily does not matter a great deal for my present project, since all parties will agree that Gorgias offers us exempla of Kunstprosa, i.e. a very deliberately and artfully wrought language (at times shading into something artificial and mannered), and so counts as producing speeches of the kind Plato wishes to label as ‘rhetoric’, irrespective of whether he, Gorgias, in addition had theoretical underpinnings for his practice, or would want to use the word ῥητορική to speak of either his practice or his theorizing (if any). 2 See, for example, the discussion of the ‘wisdom of the Spartans’ at Protagoras 342 A–343 B, and the references there to seemingly any random Spartan’s ability to fire off, pretty much at will, a ‘memorable, brief saying’ (ῥῆμα ἄξιον λόγου βραχύ) (342 E 2), and to the Spartans’ fondness for expressing their wisdom using ‘brief sayings [that are] worthy of being remembered’ (ῥήματα βραχέα ἀξιομνημόνευτα) (343 A 7–8)—all of which is summed up as ‘a kind of Laconic brevity’ (βραχυλογία τις Λακωνική) (343 B 5). 1 E.

3 Gorgias and the Weakness of Logos

51

Though there is room for this broader picture, Plato in fact seems to be working with a single contrast between lengthy speeches, addressed to a largely passive audience by a single party, often in a more or less public setting, and brief questionand-answer exchanges involving, in the first instance, two parties, each of which plays an active role: the one, the questioner, puts fairly brief questions to the other, the answerer, who in turn responds to the questions via brief answers (often and paradigmatically, simple ‘yes’s’ or ‘no’s’).3 Such question-and-answer exchanges are for their part typically depicted as taking place in more or less private settings. (Cf. e.g. Protagoras 329 A–B, Gorgias 448 D and 449 B, or Sophist 268 B.) In addition, however, Plato is intent on distinguishing philosophy and its privileged form of discourse from something that to all appearances is (formally) almost the same as it, perhaps even is actually identical to it. For in the Euthydemus, Dionysodorus and his brother use a technique of argumentation (labeled sophia eristik¯e: eristic skill, or eristic expertise, 279 B 9–10) that shares strikingly many features with Socrates’ own preferred method of argumentation. The two brothers do not go in for lengthy speeches; rather, they start by putting a choice of two alternatives to their interlocutors, e.g. ‘Are those who learn the knowledgeable ones, or the ignorant ones?’ And once an interlocutor commits himself to one of the alternatives, they follow up with a series of yes/no questions, with the aim of inducing the person answering to say something that conflicts with his initial answer, or to deny something that follows from it. More succinctly: the eristic of these sophists can look uncomfortably similar to the dialectic of Socrates.4 Now here there will not be available any simple, formal or quasi-formal marker, to allow us to distinguish the two. Though the matter would require considerable argument—which I shall not seek to provide here—I believe it is difficult, perhaps impossible, to arrive at a fully satisfying way of characterizing the (supposed) difference between dialectic and eristic, between (supposedly) legitimate and illegitimate uses of the method of question and answer. Be that as it may, it looks as if, irrespective of whether Plato can characterize the differences in suitable ways, he is committed to demarcating philosophy from two heterogeneous alternatives—rhetoric and eristic—which at first glance appear to have nothing in common besides the features of not being (what he calls) philosophy, and of not using (what he considers) its proper method of discourse, dialectic. Such a conclusion would be too hasty. For in the dialogues Plato will have the partisans of rhetoric and of eristic each say that their preferred way of speaking can 3 The

larger question of which sorts of discourse were recognized (already in the mid- to late fifth century) as making for something like distinct genres of discourse is treated well by H. Ll. Hudson-Williams, in [10]. 4 See A. Nehamas [22]. G. E. R. Lloyd’s [16], Ch. 2 (‘Dialectic and Demonstration’), 59–125, contains much useful discussion, especially also of relevant antecedents of the method of question and answer in the Hippocratic Corpus. W. Müri’s [21] remains fundamental, not least for laying out clearly that Plato uses a variety of expressions as de facto synonyms for διαλεκτική, including διαλέγεσθαι, ἐπίστασθαι λόγον τε δοῦναι καὶ δέξασθαι, and ἐρωτᾶν τε καὶ ἀποκρίνεσθαι (as well as variants of all of these). I offer a brief discussion of the expression λόγον διδόναι in my [19]. On the proximity of Socratic dialectic to sophistic eristic, see also R. Barney’s [1], esp. at 81–82.

52

W.-R. Mann

be successful wholly independently of whether the person speaking is speaking the truth. Indeed, Plato has these partisans praising their respective modes of discourse precisely because they can disregard the truth. Thus in the Euthydemus, Socrates reports: [The two brothers] have now mastered the one form of fighting they had previously left untried; as a result, not a single man can stand up to them, the two of them have become so awesome (δεινώ) in fighting with words (ἐν τοῖς λόγοις μάχεσθαι) and in refuting whatever may be said, no matter whether it is true or false. (272 A–B, trans. R. Kent Sprague in [6], with changes)

And once the question-and-answer rounds with Cleinias begin (he is the young man, as it were the ‘victim’, on whom the brothers will demonstrate their wisdom), Dionysodorus cheerfully tells Socrates that ‘whichever way the boy answers, he’ll be refuted’ (275 E 4–6; cf. 276 E 1–2). The subsequent course of the eristic exchanges confirms this: first Cleinias answers that those with expertise (οἱ σοφοί) are the ones who learn, and is refuted (276 A–B); Dionysodorus immediately goes on to show that the opposite answer—that it is the ignorant ones (οἱ ἀμαθεῖς) who learn—equally leads to trouble (276 B–D). The exercise is repeated with the question, ‘Do those who learn, learn the things they know, or the things they do not know?’ Here, again, the brothers are able to lead Cleinias into trouble as he opts for each of the alternatives in turn (276 E–C).5 Correspondingly, in the Gorgias, a core claim made on behalf of rhetoric is that the accomplished speaker will be able to persuade an audience of a position on matters concerning which he himself has no expertise, even on a matter where a relevant expert is present and is actually arguing the opposite position. Thus a speaker who has mastered the art of rhetoric will be able to persuade an audience about, say, which medical treatment is appropriate, even if an expert doctor argues, on the basis of his mastery of medicine, that a wholly different course of treatment is actually required (see 456 B–C; cf. 459 A). Plato’s Gorgias goes so far as to boast that someone who has mastered rhetoric will be able to have ‘the doctor as his slave’—and so too, analogously, have as his slaves experts in any other arts (452 E 5). Perhaps this point about slavery actually goes back to the historical Gorgias; at any rate, it gets repeated in Plato’s much later Philebus (see 58 A 8–B 3).6 In response to this striking claim about rhetoric’s power, Socrates proposes that rhetoric really is a ‘crafter of persuasion’ (πειθοῦς δημιουργός) (453 A 1), a proposal 5I

offer a more detailed discussion of eristic and dialectic in the Euthydemus in my [20]. Barney also holds that the claim about rehtoric’s power to enslave may very well go back to Gorgias himself, see [2], at 21–22. And in the Philebus, too, rhetoric is contrasted with dialectic (see 57 E ff.), though now ‘the power of dialectic’ (ἡ τοῦ διαλέγεσθαι δύναμις) (57 E 7) is said to be ‘concerned with being, and with what is really, and with what is naturally in every way eternally the same’ (58 A 2–3). It would take us too far afield to consider how well this characterization of dialectic fits with the picture of dialectic as the method of question and answer. Note, however, that Rep. 7, 534 B-E already suggests that there will be a close connection between the kind of over-arching knowledge of reality the dialcetician is said to possess and his ‘passing through [sc. unscathed] all refutations, as if in a battle’ (καὶ ὥσπερ ἐν μάχῃ διὰ πάντων ἐλέγχων διεξιών) (534 C 1–2). 6 R.

3 Gorgias and the Weakness of Logos

53

Gorgias readily accepts. Socrates goes on to distinguish between learning (μάθησις) and having learned (μεμαθηκέναι) on the one hand, and pistis (conviction) and having being convinced (πεπιστευκέναι) on the other (454 C–D); and he suggests that while there can be both true and false pistis, there cannot be both true and false knowledge (ἐπιστήμη) (454 D 5–8). This difference, Socrates says, shows that pistis on the one hand, and knowledge and learning on the other, really are different. Given this difference, into which category does rhetoric belong? Well, we of course know the answer: Gorgias holds that rhetoric is a producer of pistis, not of learning, which is to say that it can be persuasive without teaching. In other words, the pistis produced can be true, but it need not be so. We also know that Socrates will go on to argue that this power of rhetoric—the ability to persuade an audience of something untrue—depends on the ignorance of the audience in question. Let us set that, and the subsequent discussion, aside. For my purposes here, I simply want to stress that Plato characterizes rhetoric as involving speeches, aimed at producing conviction or persuasion, independently of whether the matter the of which the audience becomes convinced-and-persuaded is true. And he has the fifth century’s most famous practitioner of rhetoric enthusiastically endorsing precisely this claim. Something analogous is true of the Phaedrus, which many interpreters maintain offers a more nuanced critique of rhetoric. The discussion of proper speeches and speech-writing takes its departure from this exchange between Socrates and Phaedrus: Socrates: Phaedrus: Socrates: Phaedrus:

Well, then, we ought to examine the topic we proposed just now: When is a speech well (καλῶς) written and delivered, and when is it not? Plainly. Won’t someone who is to speak well and nobly (εὖ καὶ καλῶς) have to have in mind the truth about the subject he is going to discuss? What I have actually heard about this, Socrates, my friend, is that it is not necessary for the person who is going to be a public speaker (ῥήτωρ) to learn what is really just, but only what will seem just to the crowd who will act as judges (ἀλλὰ τὰ δόξαντ̓ ἂν πλήθει οἵπερ δικάσουσιν). Nor again what is really good or noble, but only what will seem so (ἀλλ̓ ὅσα δόξει). For that is what persuasion proceeds from (τὸ ἐκ γὰρ τούτων εἶναι τὸ πείθειν), not truth. (259 E 1–260 A 4; trans. Nehamas and Woodruff in [6], with minor changes).

Now, part of what Socrates does in the Phaedrus is to argue that being maximally effective in speaking, irrespective of whether one is concerned with arriving at the truth and persuading an audience of something true, will require that the speaker knows the truth (about the matters he is speaking of). In fact, if speakers want to be maximally effective at misleading an audience and persuading it of something false, they will themselves still need to know the truth (about the matter under discussion). Thus to put Socrates’ claim in a perhaps overly pointed way: the indifference towards,

54

W.-R. Mann

or the disregard of, the truth requires precise knowledge of, and close attention to, the very truth being disregarded in order to be genuinely effective. We can, once more, leave to one side further moves Plato makes, and subsequent conclusions he has Socrates draw. What matters here is that, again, the accomplished rhetorician is being presented as someone who aims at persuading an audience, with no concern about whether what he says is true or false. We now can see clearly, at least in outline, what the two enterprises, or modes of discourse, from which Plato seeks to distinguish philosophy and dialectic, have in common: it is just this capacity—or so their practitioners want to claim—to be successful, independently of whether what is being said is true or not. Skilled questioners like Euthydemus and his brother will be able to confound an answerer, irrespective of whether he gives a true, or a false, answer. For his part, an accomplished rh¯et¯or like Gorgias will be able to persuade an audience, irrespective of whether what he says is true, or not. And while Socrates seeks to qualify the latter claim by adding that this persuasion will only be possible if the audience consists of non-experts (about the matter under discussion), in reality this qualification is not very restrictive, since most people will not, by Socrates’ lights, count as expert (about anything) in the relevant way. In short, Plato looks to be ascribing vast power to rhetoric. This, of course, is a large part of the reason why he regards it as so dangerous and thus as needing to be opposed vigorously. ∗ ∗ ∗ But what I want to focus on here is the extent to which the historical Gorgias’s own claims, or at any rate, things that are ‘said’ in two of his surviving logoi, ‘speeches’, fit in with Plato’s characterization of rhetoric and its power.7 To anticipate the rest of this chapter: in the ‘Encomium of Helen’, Gorgias puts forward a picture of the power of logos that is wholly in line with what Plato says; perhaps, it even has the logos of rhetoric being more powerful than Plato allows. By contrast, in the ‘Defense of Palamedes’—where Gorgias is not speaking on his own behalf, but is taking on the persona of Palamedes—considerable anxieties about the possible weakness of logoi are voiced. These anxieties appear as ‘bookends’, enclosing the central arguments, and thus serve to undermine their persuasiveness.

7I

set aside the short piece entitled, ‘On Not-Being, or Concerning Nature’. There are challenging questions about the work, including philological ones about the relation of the two transmitted versions (one preserved in Sextus Empiricus, the other in the pseudo-Aristotelian treatise, On Melissus, Xenophanes, and Gorgias) to any would-be Gorgianic original. Addressing these questions is beyond the scope of this chapter.

3 Gorgias and the Weakness of Logos

55

3.2 Gorgias and the Power of logos: The ‘Encomium of Helen’ Leaving aside any complications its last sentence, specifically its very last word, may give rise to—here Gorgias looks to be saying that the entire piece has been some kind of joke or game (παίγνιον)—the official, stated goal of the ‘Encomium of Helen’ is to praise Helen. As, however, already §2 makes clear, a key component of that praise will be devoted to absolving Helen from blame (for having eloped to Troy with Paris, and thus for being responsible for the Trojan war together with all its calamitous consequences).8 The bulk of the speech, beginning in §6, in fact consists of seeking to free her from blame. The structure of the ‘defense’ involves four disjuncts (which are implicitly taken to partition the logical space of possibilities exhaustively):9 it is likely that either Fate, necessity, and the decision of the gods were the cause of Helen’s ‘doing what she did’ (viz. ‘leaving’ her husband, Menelaos, and ‘going to’ Troy), or she was taken by force, or she was possessed by er¯os, or she was seduced by logoi. (In keeping with the exposition I offer, I have switched the order of the last two from how they appear in the actual text.) In each case, Gorgias argues, were it the correct account (of what happened), she would not be responsible and, hence, would not be to blame. Since, or so Gorgias implies (vide §6, init. and §20), there are no additional possibilities as to why Helen did what she did, i.e. as to what was responsible for her winding up at Troy with Paris, she is in fact not responsible for what happened. In §§6–7, the first two options are quickly disposed of: in both cases Helen proves weaker and is overcome by someone, or something, stronger. Obviously so, if the gods are the cause of what happened. And in the case of force, Gorgias emphasizes (§7) that the force in question would be a matter of Helen having been raped (ὑβρισθεῖσα) and abducted (ἁρπασθεῖσα)— thus how could the response that is reasonably called for not be pity rather than condemnation (πῶς οὐκ ἂν εἰκότως ἐλεηθείη μᾶλλον ἢκακολογηθείη)? In either case, if we were (still) to hold her responsible, this would be a matter of, as we now say, ‘blaming the victim’, i.e. blaming her for something for which she clearly is not responsible. In §§15–19, Gorgias considers er¯os and other ‘forces’ which effect their work via sight (διὰ τῆς ὄψεως). The soul receives impressions by seeing things; and these impressions can produce reactions in those who have them: thus a person can be stricken with fear and panic at the sight of weapons of war, made of, presumably, gleaming bronze and steel, and thus flee. Indeed, Gorgias suggests, there is a 8 R. Barney, in [2], offers an excellent entrée to the larger philosophical issues at stake in the arguments about assigning, or rather, not assigning, responsibility to Helen for her actions. 9 See also [2], 5–7, on the overall structure of the argument.

56

W.-R. Mann

widespread, natural or close to natural, quasi-causal process that leads from (seeing) frightening sights (φοβερά) to fear (ὁ φόβος), which in turn extinguishes and drives out thought (τὸ νόημα). More generally, sight engraves in the mind (ἐν τῶι φρονήματι) images of the things seen. The images of frightening things linger (sc. and thus play ongoing roles in the person’s conduct and life) (§17). We might surmise that they linger because they engraved firmly or deeply upon the mind—although Gorgias does not say this expressis verbis. In a transition that is not as clear as we might want it to be, Gorgias goes on to say that certain combinations of colors and figures can, for their part, delight the sense of vision, and make for a pleasant sight for the eyes. Such pleasant sights can, in turn, produce longing and er¯os. And if the eye of Helen, pleased at (the sight of) the body of Paris, transmitted to her soul eager desire and the striving for er¯os,10 this should not be surprising (§19). The train of thought is not entirely clear, but it seems that we are to think of the process as analogous to the one that led from seeing frightening things, to fear, and then to action based on fear. That is, seeing pleasing things leads to pleasure—which somehow is, or results in, desire—and then to action based on that pleasure and desire. Gorgias ends this section by saying that if er¯os is a god, with divine power, then how could Helen have refused him? But if it is a human illness (νόσημα) or an oversight (ἀγνόημα) of the soul, her elopement should be regarded not as a going astray (ἁμαρτήμα), but as a misfortune (ἀτύχημα), for which Helen, again, is not to blame. ∗ ∗ ∗ The real heart of the ‘Helen’, however, comprises §§8–14, devoted to the power of logos.11 Our stretch of text opens as follows: But if it was logos who persuaded (πείσας) her and deceived (ἀπατήσας) her soul — not even in reply to this is it difficult to argue ‘for the defense’ and to acquit her of blame/responsibility (αἰτία), as follows: logos is a mighty potentate (λόγος δυνάστης μέγας ἐστίν), who with a tiny and unseen body accomplishes feats most divine. For he has the power (δύναται) to stop fear and to banish grief and to effect joy and to enhance pity. That these things are so, I will [now] show.

10 The words are προθυμία καὶ ἅμιλλα ἔρωτος. Older translations here speak of the ‘contest of love’.

But this, it seems to me, does not quite make sense. LSJ, s.v. ἅμιλλα suggest ‘eager desire’; but this is perhaps too weak and also does not make for a sufficient difference from προθυμία (which would by itself naturally be rendered as ‘eager desire’). Thus I offer ‘striving for er¯os’ as a paraphrase. In their recent Loeb translation [14], Laks and Most offer ‘an eagerness and a striving for love’ for προθυμία καὶ ἅμιλλα ἔρωτος. 11 Cf. Barney, who observes: ‘. . . his discussion of logos here is outsize and extravagant and clearly meant to stand out: the Helen as a whole may be just an elaborate frame for what Gorgias wants to say here about speech’ ([2], 9).

3 Gorgias and the Weakness of Logos

57

Before proceeding, three brief observations. First, when Peith¯o is personified in the tradition, she is often as much a goddess of seduction as of persuasion, that is, a goddess of persuasion in erotic settings.12 Of course, here there is no talk of peith¯o directly, but of logos as being the ‘agent’ who persuaded (ὁ πείσας) Helen; yet given the erotically charged context, thoughts of seduction should very much be kept in mind. Secondly, and closely related to the first point, logos is identified with deception or cheating (ἀπάτη). Thirdly, Gorgias is emphatic about the strength or power of logos here: not only does δύναται (‘he has the power’) pick up on δυνάστης μέγας (‘mighty potentate’), but by omitting both a connective particle (e.g. a γάρ) and the definite article for the noun, he makes the claim about logos seem especially strong and vivid, an almost gnomic pronouncement. Via what medium is the strength of logos transmitted? Not through sight, but through the sense of hearing: the first example he provides is poetry, which Gorgias characterizes as ‘logos with meter’ (§9), and which brings to those listening to it ‘the trembling of extreme terror, tear-flooding pity, and grief-loving longing (πόθος)’.13 A listener’s soul experiences these emotions, which arise from and are about the actions and bodies of others, in good and bad fortunes, through logoi (διὰ τῶν λόγων ἔπαθεν ἡ ψυχή): here, the aural effects of heard logoi are presented as the counterparts to the visual effects of sights seen. But Gorgias immediately moves on to ascribe still greater power to logoi, likening them to divinely inspired incantations (ἔνθεοι ἐπωιδαί)—conveyed via logoi—which lead towards pleasure and away from pain, for, by mixing with doxa in the soul, the power of incantation (ἡ δύναμις τῆς ἐπιωιδῆς) bewitches, persuades, and transforms the soul with witchcraft (γοητεία) (§10). In the next sentence, Gorgias pairs witchcraft with magic: we thus have moved on from the realm of natural- and causal-processes to that of super-natural ones. It is true that Gorgias seemingly returns to the more ordinary world in §11, focusing on the idea that logos is often deceptive (here I largely follow MacDowell’s translation, in [18]):

12 Much of the early evidence for this, including material from the realm of the visual arts, is well discussed by R. G. A. Buxton, in [3]; see esp. Ch. 2, 29–65. On p. 31, commenting on our lines, he writes: ‘. . . peitho is the seductive persuasion which may have induced Helen to go off with Paris. Yet it also the power used and the effect produced by oratory in contexts which we would regard as non-erotic—but to the Greeks all peitho was “seductive”. Peitho is a continuum within which divine and secular, erotic and non-erotic come together’. 13 I am following N. Worman’s translation here, except in rendering πόθος as ‘longing’ rather than ‘desire’; see [37], 161.

58

W.-R. Mann How many men have persuaded and do persuade how many, on how many subjects, by fabricating false speech!14 For if everyone, on every subject, had memory of the past and < understanding > of the present and advance knowledge of the future, logos would not be equally < powerful >15 ; but as it is, neither remembering a past event nor investigating a present one nor prophesying a future one is easy — so that on most matters most people make doxa their adviser; but doxa, being slippery and insecure (σφαλερὰ καὶ ἀβέβαιος), brings slippery and insecure successes to those making use of it.

And in §12, Gorgias argues that logos, on account of this weakness of doxa, can be, and is, powerful—indeed, is just as powerful as necessity or compulsion (ἀνάγκη). Thus ‘the persuader (ὁ πείσας), because he compelled (ἀναγκάσας), did wrong; but the persuaded woman (ἡ πεισθεῖσα), because she was compelled (ἀναγκαθεῖσα) by his logos, is wrongly spoken ill of’. Yet after this intrusion of rationalistic thinking, Gorgias returns at the end (of the sections on logos) once more to the world of witchcraft and magic: The power of logos stands in the same relation to the order (τάξις) of the soul as the order (τάξις) of drugs stands to the nature of bodies: just as different drugs expel different humors from the body — some putting an end to disease, but some to life — so too some logoi cause grief, some delight, some fear, and some give the hearers confidence, but some drug and bewitch the soul with evil persuasion (οἱ δὲ πειθοῖ τινι κακῆι τὴν ψυχὴν ἐφαρμάκευσαν καὶ ἐξεγοήτευσαν). (§14)

From the perspective of the ‘Defense of Helen’, logos looks to be powerful indeed: taken at face value, Gorgias’s argument suggests that anytime someone was persuaded to do something (paradigmatically, something bad or wrong), that person, qua having been persuaded, cannot rightly be held responsible (i.e. accountable) for what was done.16 14 So

D. M. MacDowell, in [18], 24; cf. 13–15, and 38, ad loc. He is arguing against the traditional view which takes ὅσοι to mean ‘all’, and which would thus have us render the sentence as: ‘All who have persuaded and do persuade people of things do so by framing a false logos’. Such a view is adopted e.g. by W. J. Verdenius [35], who argues that for Gorgias, all logos is false and deceptive. Laks and Most, in [14], in effect construe the sentence in a similar fashion, rendering it as: ‘Whoever has persuaded, and also persuades, whomever about whatever [scil. does so] by fabricating a false discourse (logos)’. One might think that the very first sentence of the ‘Helen’ is sufficient for refuting so universal a claim, for there Gorgias says ‘truth is the adornment (κόσμος) of logos in the same way as the excellence of its men is an adornment of a city; beauty, that of a body; wisdom, that of a soul; and virtue, that of an action’. As we shall see, the final part of the ‘Palamedes’ would also make no sense if true, persuasive logoi were in principle impossible. 15 Laks and Most [14], following Diels Kranz, print: οὐκ ἂν ὁμοίως ὅμοιος ἦν ὁ λόγος, which translators render as ‘speech would not be similarly similar’. MacDowell, [18], p. 38, argues that this ‘cannot be right’, among other reasons, because ‘“similarly similar” makes no sense’. He thinks that ὅμοιος is a scribal error (‘derived from the previous ὁμοίως’), and it must have ‘ousted a word meaning “effective” or “powerful”’. This is a fairly drastic intervention in the text, but I am inclined to agree that ‘similarly similar’ fails to make good sense. 16 Barney, in [2], helpfully distinguishes between a ‘broad’ claim that might seem to underly the ‘Helen’, namely that ‘nobody is ever morally responsible for anything at all’ (16, 17), and a ‘narrow’ claim, where the actions under considersation are, like Helen’s, ‘irrational, self-destructive, and/or wrong’ (17–20). It may be worth recalling that in ascribing responsbility, the Greeks were often assigning, or looking to assign, blame. Consider, for example, the opening of Herodotus, where he

3 Gorgias and the Weakness of Logos

59

3.3 Gorgias and the Weakness of Logoi: The ‘Defense of Palamedes’ Before turning to the text, two points about matters which Gorgias’s speech may appear to leave open: first, Palamedes in no way alludes to what was standardly taken as the motive for Odysseus’s accusation of treason and his hatred towards him— namely that Palamedes had exposed Odysseus’s sham insanity and thus had ‘forced’ him to join the expedition against Troy. The story is that when Paris’s abduction of Helen came to light, Palamedes and Nestor were sent to recruit for the war the various Greek kings, who, at the wedding of Menelaos and Helen, had sworn an oath to defend their marriage. When Palamedes arrived in Ithaca, Odysseus, who had been tipped-off about Palamedes’ imminent arrival and who wanted to avoid joining the expedition, was plowing his fields with an ox and a horse yoked together, and (in some versions) sowing salt in the plowed furrows. (This behavior was supposed to show that he was mad and thus unfit for military service.) Palamedes, suspecting that Odysseus was ‘faking it’, grabbed the infant Telemachus—in some versions, actually snatching him from Penelope’s lap—and placed him directly in the path of the plow. Odysseus stopped plowing, rescued his son, and so revealed that he was not in fact insane. (This story obviously relies on the thought that if Odysseus really had been insane, he would have continued plowing, thereby killing Telemachos.) Having been tricked by someone even trickier than himself and forced to participate in the war, Odysseus bore a grudge against Palamedes and sought to destroy him, when the opportunity to do so arose. The things Gorgias has Palamedes say in his speech are not incompatible with this ‘back story’, but they do not at all rely on it. There is a second matter the speech may be leaving open: Gorgias’s Palamedes does not describe the would-be evidence Odysseus furnished of his, Palamedes’, supposed treason. In many versions of the story, Odysseus forges a letter, or two letters: one from Priam, the king of the Trojans, to Palamedes, offering him gold for betraying the Greeks; another from Palamedes to Priam, accepting the offer. Odysseus also, through a complicated ruse involving the Greeks temporarily moving their camp, hides the specified amount of gold under Palamedes’ tent. The letters are ‘discovered’, Palamedes is charged with treason and the gold underneath his tent is ‘found’, and thereupon he is immediately sentenced to death by Agamemnon, the overall commander of the Greek armies (see also nn. 20 and 31). On these matters, the situation may be a bit more complicated, as we shall shortly see. But here, too, Gorgias’s Palamedes does not directly refer to the supposed seeks to determine ‘on account of what cause (δι᾽ ἣν αἰτίην)’ the Greeks and the foreigners ‘entered into war against each other’ (1.1.0), and then quickly moves from talking of which parties were originally ‘responsible (αἰτίους)’ for the emnity (1.1.1) to talking of who committed the ‘first wrong (τῶν ἀδικημάτων πρῶτον)’ (1.2.1) (cf. all the uses of αἰτία, rather, αἰτίη, and αἴτιος in 1.1 through 1.5). If Gorgias is concerned with responsibility in such a narrower sense, then the upshot of the ‘Helen’ would be that no one is ever to blame for conduct that, viewed more ‘naïvely’, calls for disapprobation.

60

W.-R. Mann

evidence (the letters or the gold). One way, in particular, in which this lack of specificity about the evidence plays a role in the speech, is that it gives Gorgias the ‘possibility space’ for having Palamedes explore a series of appeals to likelihood. These, as has often been remarked, form the core of Palamedes’ substantive arguments, though I shall not be discussing them here. ∗ ∗ ∗ With these observations in place, I should like now to turn to some sections of the speech itself. As just mentioned, the central idea in Palamedes’ defense is that no account of how the treason is supposed to have been carried out is plausible; indeed, each imagined alternative proves to be implausible, because, on inspection, it is revealed to be impracticable. In addition, Palamedes asserts, when it comes to his (supposed) motives, none of the possibilities makes sense, either. As one commentator remarks, echoing §5 and §21,17 the main defense divides into two major parts: ‘if I had wanted to betray the Greeks . . . I would not have been able to do so (§§6–12)’ [sc. none of the ways of doing would be at all likely to prove successful], ‘and if I had been able to do so, I would not have wanted to do it (§§13– 21)’ [sc. none of the reasons why I might have wanted to do so is at all a likely one to ascribe to a sensible person, and to me in particular].18 Other parts of the speech are concerned to cast doubt on Odysseus (either on his motives or on his competence) (§5, init., and §§22–27),19 or they seek to remind the Greeks of Palamedes’ past life and the benefits he conferred to all (§28 ff., esp. §30). Here it is important to recall two things: first, in antiquity Palamedes was taken to be a ‘culture hero’ and was credited with having invented various clever and highly useful things, including counting and currency, techniques pertaining to military organization and logistics, and the alphabet, or least some of the letters; secondly, as noted above, stories around the lead-up to the Trojan War point to a bitter rivalry

17 At the end of §5, Palamedes says, ‘. . . for neither if I had wished to would I have been able to undertake actions of such a kind, nor if had been able would I have wished to do so’; the end of §21 picks this up, ‘That I would not < have been able to, nor > have wished, to betray Hellas has been proved by what I just said (διὰ τῶν προειρημένων δέδεικται)’. The supplement is due to Keil, who is obviously relying on the final sentence of §5. 18 So D. G. Spatharas, in his [32], here at 395. 19 Gorgias is presumably counting on his intended audience’s (and Palamedes’ fictional audience’s) familiarity with the ‘facts’ (i) that Odysseus knows how to say ‘many false things that are like the truth’ (Od. 19. 203), and (ii) that he boasts of being famous among men ‘for all manner of trickery (δόλοι)’ (9. 19–20). On Odysseus’s association with δόλος and the role this plays in his relation to Palamedes, see M. Christopoulos [4].

3 Gorgias and the Weakness of Logos

61

between Odysseus and Palamedes.20 But this rivalry is wholly suppressed in the Homeric epics themselves.21 Finally, Palamedes closes his speech, in a way that Plato may well have looked to in the Apology, by warning the Greek leaders that if they convict him unjustly, they will have to live with the reputation of having committed injustice in condemning an innocent man to death (§§ 33–36, esp. 36).22 But the central arguments of the speech are flanked by two cautionary remarks which contrast sharply with the kinds of things Gorgias says in the ‘Encomium of Helen’. And it is these I want to focus on. The first four sections of the speech lay out the stark alternatives: everyone must die, so the decision or judgement is not about death, it rather concerns whether Palamedes will die justly or unjustly (§1); and either Odysseus is making his accusation out of good will (εὔνοια) towards Greece, or with envy, malicious conspiring (κακοτεχνία), and all-around πανουργία (§3). (A brief aside: our dictionaries define πανουργία as ‘knavery’, a word no longer used in everyday language. The basic idea seems to be that the πάνουργος person will, as we say, ‘stop at nothing’, that is, he is literally ready to do anything—this sense then also allows for the less negative connotations the word has on occasion, see LSJ, s.v. It is also worth recalling that over the course of the fifth century Odysseus comes

20 For evidence from the Epic Cycle, see M. L. West [36], 102–103, and 123–125; see also Philostratus, Heroicus, 33.1–34.7. Such evidence as we have suggests that Palamedes was a popular figure for the tragedians of the fifth century; see for example G. Koniaris [13], esp. 87–89, and nn. 13–16. R. Scodel, in her [29], argues against Koniaris that Euripides’ three plays of 415 (i.e. the Alexander, Palamedes, and Troades) were composed to make up a trilogy of related works. But see two reviews of Scodel [29], namely [30] and [8], for scepticism on the issue. This matters, because if the tragedies are not related in this way, we cannot use the extant Troades as a basis for speculation about the lost Palamedes. What is clear, however, is that Euripides is relying on a version of the story that is consistent with Gorgias’s speech. (The relative dates cannot be established securely, so no inference about the supposed direction of influence is reliable.) — Scodel summarizes the plot of the Palamedes as follows: ‘Odysseus forged or caused to be forged a letter purporting to be from Priam to Palamedes in which Palamedes was promised a certain sum of gold in return for betraying the Greek camp. This letter was given to a Phrygian prisoner, whom Odysseus had killed near the camp; through an elaborate stratagem he secretly buried the amount of gold named in the letter beneath Palamedes’ tent. The central action of the play was the trial of Palamedes. The corpse having been found and the letter brought to Agamemnon, the king investigated the matter. Palamedes defended himself against Odysseus’s accusations, and perhaps himself suggested that his tent be searched for the gold. When the gold was found, he was condemned, but his brother Oiax lamented his murder and wrote his story on oars which he cast into the sea in the hope that one of them might reach his father, Nauplius. The ending of the play is uncertain, but there cannot have been any real resolution’ ([28], 407). 21 Philostratus, Heroicus, 43.15–16, has Homer conjuring Odysseus from the dead and begging him to exclude Palamedes from his narrative: ‘Do not bring Palamedes to Ilion, don’t treat him as a soldier, nor say that he was wise! Other poets will say these things, but because they have not been said by you, they will not seem convincing (πιθανὰ δὲ οὐ δόξει μὴ σοἰ εἰρημένα)’. In return for this favor, the shade of Odysseus provides Homer with material for his epics (see all of 43.10–16). 22 Scholars have long argued that there are significant parallels between the Apology and the ‘Palamedes’; for a close reading of the texts that does so with exceptional care and also provides copious references to earlier literature, see J. Coulter’s important [7].

62

W.-R. Mann

increasingly to be thought of, and presented as, a πάνουργος-type.23 Thus Gorgias could well be relying on his audience’s being familiar with claims of this kind about Odysseus.) To return to the text: after having made these observations about the predicament he faces and about the kind of person Odysseus may be, Palamedes, rather than either launching into his defense, or into a direct attack on Odysseus, concedes that he does not quite know where to begin, what to say first, or where to turn for his defense (§4, 4–5).24 Here, in his explanation for why he is unable to proceed straightaway, we find his first cautionary comment: For an unsupported25 accusation makes for visible shock,26 and on account of that shock, I am necessarily at a loss in my logos, unless I learn something [by drawing] from the truth itself and from the [constraint of] my present necessity, coming upon teachers more dangerous than helpful.27 αἰτία γὰρ ἀνεπίδεικτος ἔκπληξιν ἐμφανῆ ἐμποιεῖ, διὰ δὲ τὴν ἔκπληξιν ἀπορεῖν ἀνάγκη τῶι λόγωι, ἂν μή τι παρ’ αὐτῆς τῆς ἀληθείας καὶ τῆς παρούσης ἀνάγκης μάθω, διδασκάλων ἐπικινδυνοτέρων ἢ ποριμωτέρων τυχών.

The sentence is difficult, but it seems to be saying that a true logos will not, simply in and of itself, be persuasive. And this is so, despite the fact (or so Palamedes asserts much later, in §33) that it is right for him (to seek) to escape the accusation of treason by means of ‘the clearest justice, expounding the truth’. Now, why exactly will a true logos not be persuasive? It seems to me that there are two reasons which Palamedes adduces, and a third one which emerges near the end of the whole speech. First, there is the general point (made in §33, to which I shall turn more closely shortly), that if the ‘truth of actions’ (ἡ ἀλήθεια τῶν ἔργων) were clear through logoi, then making decisions (of the kind the Greeks are facing) would be easy. This is not unrelated to claims about the difference between knowledge and opinion Palamedes makes, in particular, the claim that he knows the truth (of what happened, or rather, of what did not happen), while the others do not. Thus those who are judging him must rely on opinion (doxa), which, as Palamedes says in his remarks directed at Odysseus, is a most untrustworthy thing (ἀπιστότατον πρᾶγμα) (§24). (Compare the characterization of doxa as ‘slippery and insecure’ in the ‘Helen’.) Secondly, and perhaps more interestingly, in his comments to Odysseus, he says that he and Odysseus face a fundamental asymmetry, which is revealed when it 23 See

Worman [37], 173–176. the artful use of the three interrogative words in quick succession, each introducing its own brief question: πόθεν, τί, and ποῖ. 25 It is tempting to hear ἐπίδειξις in ἀνεπίδεικτος, in which case the meaning might rather be that the accusation has not been demonstrated (i.e. it is as if ἀνεπίδεικτος were standing in for ἀναπόδεικτος). 26 See Worman [37], 174–175, for a discussion of ἔκπληξις (and related terms) in earlier and contemporary literature, including the report (Diodorus Siculus, 12.53.2–5) that Gorgias himself ‘struck (ἐξέπληξε) with the strangeness of his usage’ his Athenian audience when he appeared in Athens in 427. Cf. also ἐκπλάγεντες at ‘Encomium of Helen’, §16. Laks and Most [14] have ‘a manifest consternation’ for ἔκπληξιν ἐμφανῆ; but this seems too weak. 27 πόριμος, viz. ποριμώτερος, is pretty clearly picking up ἀπορεῖν, so the wished-for teachers would, ideally, provide the resources for overcoming the (sense of a) lack of resources of which Palamedes speaks. 24 Note

3 Gorgias and the Weakness of Logos

63

comes to the all-important matter of witnesses. Palamedes and Odysseus (and, one assumes, the Greeks judging the case) are agreed that if there is a witness to testify (to the events the accused is charged with), the accusation will be more credible (πιστότερον) (§22, fin.) on account of having been so witnessed. Thus, Palamedes maintains, Odysseus ought to produce witnesses to what Palamedes (allegedly) did. And he imagines Odysseus maintaining, by a kind of parity-of-reasoning consideration, that he, Palamedes, ought for his part to produce a witness or witnesses to what did not happen, to what (he says) he did not do (§23). But, Palamedes continues, there is, and can be, no parity here, or in any analogous situation: ‘Things which did not happen cannot at all be attested by witnesses (μαρτυρηθῆναι), whereas concerning what did happen, not only is it not impossible, but easy, and not only easy, but < necessary > [sc. to be attested by witnesses]’ (ibid.). In effect, Palamedes is, locally, looking to shift the burden of argument: it is really Odysseus who needs to show that Palamedes committed treason, not Palamedes who needs to show that he did not. But in the larger context of the logos as a whole, the (imagined) impossibility of showing that what did not happen did not happen is perceived by Palamedes as an enormous burden; in fact, it may be an insurmountable obstacle. Thirdly, Palamedes does not appeal to the emotions—indeed, he insists that doing so would be suitable only if he were being judged by a ‘mob’ (§33)—and thus he does not even attempt to do the kinds of things Gorgias, in the ‘Encomium of Helen’, tells us logos can do: sway the thinking of those addressed by changing their feelings or, in the most extreme case, by bewitching them. Palamedes’ second set of cautionary comments comes as part of his address to the Greeks, his jurors, to whom he says he is submitting the record of his life (see §28), much in the way in which a public official would be submitting the records of monies spent and received, while in office, to an audit (the εὔθυνα; see also LSJ, s.v.). Let us consider a substantial part of this entire section: (33) As for what remains, my logos concerns you and is directed towards you; when I’ve said this, I will conclude my defense. Appeals to pity, entreaties, and the supplication of friends may be useful when the trial is before a mob; but before you, who both are and are reputed to be the foremost of the Greeks, there is no need for me to persuade (πείθειν) you with the help of friends, or entreaties, or appeals to pity; rather, I need to free myself of (διαφύγειν) this accusation with the clearest justice, by teaching you the truth (διδάξαντα τἀληθές), not by deceiving you. (34) And you [sc. for your part] must not pay more attention to the logoi than to the actions (τὰ ἔργα), and not prefer the accusations over the rebuttals, nor consider a short time to be a wiser judge than a long one, nor take slander (τὴν διαβολήν) to be more trustworthy (πιστοτέραν) than tested-experience (τῆς πείρας). Concerning all matters, it is the part of good men to take great care not to go wrong, and still more so in irreversible matters than reversible ones: for these are under the control of those with foresight (προνοήσασι), but are irremediable by those with hindsight (μετανοήσασι); and the situation is of this sort, whenever men judge a man in a death-penalty case — which is what you have before you here. (35) Now if, through logoi, the truth of actions became clear to an audience in unblemished form (διὰ τῶν λόγων τὴν ἀλήθειαν τῶν ἔργων καθαράν γε γενέσθαι τοῖς ἀκούουσι φανεράν),28 the judgement — on the basis of what was said [sc. by me] — would already be easy (εὔπορος). But since this is not how the matter stands, take care 28 See D. M. MacDowell [17], at 123, for a defense of the manuscript reading (i.e. for not following

Diels either in changing γε to τε, or in inserting καί before φανεράν). He proposes that the γε be

64

W.-R. Mann for my person (τὸ σῶμα τοὐμόν), take more time, and render your judgement accompanied by the truth.29

The underlying thought, it seems, is that if the audience (the jurors) had direct access to the ‘truth of actions’, it (they) could simply see what happened, and so render the correct verdict, acquitting Palamedes. In an important respect, this recalls the sentiment expressed in the ‘Helen’, in §11—‘if everyone, on every subject, had memory of the past and < understanding > of the present and advance knowledge of the future’, then there would in effect be no need for logos (of the kind Gorgias is interested in), and certainly not so great a need. But here, unlike in the ‘Helen’, there is not only no guarantee that logoi will be persuasive, but there is not even the expectation that they will be so—as a matter of fact, Palamedes goes so far as to make a part of his logos the injunction to his hearers that they not pay attention to logoi! This would be a distinctly perverse move, if he had any real confidence in the power of his own speech. After all, he had stated, at the end of §21, that the case for his innocence was proved by what he had said (διὰ τῶν προειρημένων δέδεικται)—i.e. by the arguments based on likelihood. Thus if he believed that those earlier logoi had been persuasive, saying not to pay attention to them now would amount to undermining his case foolishly; but if he instead believed that his earlier logoi had been unpersuasive, why hold that this logos, now, will be any different? It might seem that I am making too much of this, since Palamedes may seem to be doing nothing more than stating a commonplace along the lines of ‘actions speak louder than words’. Yet Palamedes, in a way, had said things like that before. Most conspicuously, he did so just a bit earlier, in §30, where he recounts the benefits he conferred on the world: I might say — and in saying it, I would neither be lying nor be refuted — that not only am I blameless, but I am also a great benefactor (μέγας εὐεργέτης) to all of you: to the Greeks and all human beings, not only those who are [sc. alive] now, but also those who will be in future. For who [sc. else] would have made human life well-provided for from having been without provision, well-ordered from having been disordered (τίς γὰρ ἂν ἐποίησε τὸν ἀνθρώπειον βίον πόριμον ἐξ ἀπόρου καὶ κεκοσμημένον ἐξ ἀκόσμου), by inventing military positions (τάξεις πολεμικαί),30 making for the greatest advantages [sc. in war]; written laws, the guardians of justice; letters, the instrument of memory; measures and weights, the convenient intermediaries of commerce; number, the guardian of possessions; signal-fires, the strongest and fastest messengers; and [the game of] checkers, a harmless pastime of leisure?

understood as both emphasizing and limiting: ‘“The truth of the affair—I mean the whole truth”’. I have italicized ‘unblemished’ (καθαράν) in the translation to bring out this sort of emphasis. Another possibility might be simply to delete γε, as an instance of dittography. Laks and Most [14] follow Diels and render the words as: ‘Well, if it were possible for the truth about actions to become pure < and > clear for listeners by means of speeches . . .’. 29 I have taken over some turns of phrase from D. Graham’s translation, in his [9], 775. 30 There is a question whether the τάξεις referred to here are miltitary ranks, so that Palamedes would be claiming to have invented how to organize an army, via something like a ‘command structure’, or are rather the positions in which e.g. an infantry line (τάξις) was to be arrayed in battle, in order to fight more effectively. For our purposes, nothing depends on which alternative one chooses.

3 Gorgias and the Weakness of Logos

65

We might also recall here some lines from Euripides’ Palamedes, generally assigned to Palamedes’ defense speech, in particular, the comments about the value of writing as an instrument of memory and as a device for forestalling, or resolving, disputes.31 What are we to make of the de facto ineffectuality of recounting all these erga? One possibility is that Gorgias is relying on his audience’s knowledge that Palamedes’ ‘actual’ defense (whatever it might have been) did not succeed, and is accordingly incorporating within the speech a kind of ‘foreknowledge’ of the defense’s failure. There is, however, a more interesting and more sinister possibility. To bring it into view, I need to backtrack slightly from my earlier suggestion that Gorgias’s speech does not allude to the ‘evidence’ Odysseus produced against him. Recall what that ‘evidence’ was, according to tradition—the forged letter or letters (in some versions, even written in Phrygian script, i.e. in a writing system that would be recognizably ‘Trojan’), and the specific quantity of gold (mentioned in the letter or letters) that was planted under Palamedes’ tent. Thus it is Odysseus who is able to use erga (rather, pragmata) rather than logoi: he does not need to speak persuasively; he can just point to the evidence and say, ‘actions speak louder than words’. Moreover, and this is the sinister part, he would be using a number of Palamedes’ clever inventions against him: the written letters are no ‘instrument of memory’ here, but rather tools for subverting the truth; measures and weights, and number more generally, are neither ‘intermediaries of commerce’ nor ‘guardians of possessions’; the very precision they afford—gold of exactly this quantity—would serve to undermine Palamedes’ claims of innocence: if the letter mentioned, say, 30 bars of gold, and exactly that amount of gold was found under his tent, how could that be a coincidence? (Of course, it is not a coincidence—but that is another matter!) If we imagine that Gorgias thinks of his audience as having such details in mind, then the speech would be a de facto demonstration of how the things Palamedes invented in order to help make human life porimos, well-furnished for greater achievements and more leisure, were turned back against him and rendered him aporos— helpless and without resources (recall his use of aporein in §4). We have come a long way indeed from the strong claims made on behalf of logos in the ‘Encomium of Helen’.

31 See Euripides’ Palamedes, fr. 578 (the text is preserved at Stobeaus 2.4.8): ‘I alone correctly furnished ‘medications’ against forgetting (τὰ τῆς γε λήθης φάρμακ’ ὀρθώσας μόνος), which are speechless and speaking (ἄφωνα καὶ φωνοῦντα), by establishing syllables; I invented the knowledge of writing for human beings, so that someone absent over the ocean’s expanse might know well (καλῶς) all the matters back in his household, and so that a dying man might write down the measure of his possessions (χρημάτων μέτρον) when leaving them to his sons, and the one to whom they were left might know it. And the evils that beset humans when they fall into a quarrel—a written tablet does away with these and makes it impossible to speak falsely’. (Laks and Most [15], 341, ad loc., tentatively offer the attractive paraphrase, ‘consonants and vowels’, for ἄφωνα καὶ φωνοῦντα.) Again, we need to recall that the relative dates of Euripides’ drama and Gorgias’ ‘Palamedes’ cannot be established securely, so no inference about the supposed direction of influence is reliable. — With Euripides, Palamedes, fr 578. compare Sophocles, Nauplius, fr. 432 R which presents a very similar list of Palamedes’ inventions.

66

W.-R. Mann

Acknowledgements The first ‘offficial’ version of this material was presented at a workshop at the University of Toronto, in 2015. At the last minute, it proved impossible for me to attend in person. Rachel Barney very kindly read the paper out in my absence; many thanks also to David Wolfsdorf for providing me (in writing) with the comments he gave on that occasion. I am in addition grateful to have had the opportunitiy to present subsequent versions of the paper as talks at the NYU Ancient Philosophy Work-in-Progress workshop, at Northwestern University, and at the 2016 Berlin conference, ‘Argumentation in Classical Antiquity: Dialectic, Rhetoric, and Other Domains’. Comments and questions from the audiences at those events helped me to rethink various points. But I owe even deeper and longer-standing debts of gratitude to Kate Meng Brassel, Jonathan Fine, Usha Nathan, and above all Elizabeth Scharffenberger for countless valuable conversations over the years on many of the issues surrounding the relationship between philosophy and rhetoric in the fifth and fourth centuries BCE.

References 1. Barney, R. (2006). The sophistic movement. In M. L. Gill & P. Pellegrin (Eds.), A companion to ancient philosophy (pp. 77–97). Malden, MA: Blackwell. 2. Barney, R. (2016). Gorgias’s ecomium of Helen. In E. Schliesser (Ed.), Ten neglected classics of philosophy (pp. 1–25). Oxford. 3. Buxton, R. G. A. (1982). Persuasion in Greek tragedy. Cambridge. 4. Christopoulos, M. (2014). Odysseus, Diomedes, Dolon and Palamedes: Crimes of mystery and imagination. In M. Christopoulos & M. Païzi-Apostolopoulou (Eds.), Crime and punishment in homeric and archaic epic (pp. 153–166). Ithaka, Greece. 5. Cole, T. (1991). The origins of rhetoric in ancient Greece. Baltimore, MD. 6. Cooper, J. M. (Ed). (1997). Plato: Complete works. Indianapolis, IN. 7. Coulter, J. (1964). The relation of the Apology of Socrates to Gorgias’ Defense of Palamedes and Plato’s critique of Gorgianic rhetoric. Harvard Studies in Classical Philology, 68, 269–303. 8. Diggle, J. (1981). Review of Ruth Scodel, The Trojan Trilogy of Euripides. Classical Review, 31, 106–107. 9. Graham, D. (2010). The texts of early Greek philosophy. Cambridge. 10. Hudson-Williams, H. Ll. (1950). Conventional forms of debate and the Melian dialogue. American Journal of Philology, 71, 155–69. 11. Karp, A. J. (1977). The Homeric origins of ancient rhetoric. Arethusa, 10, 237–58. 12. Kennedy, G. (1957). The ancient dispute over rhetoric in Homer. American Journal of Philology, 78, 23–35. 13. Koniaris, G. (1973). Alexander, Palamedes, Troades, and Sisyphus—A connected tetralogy? A connected trilogy? Harvard Studies in Classical Philology, 77, 85–124. 14. Laks, A., & G. W. Most. (Eds.). (2016). Early Greek philosophy, Vol. 8: Sophists, Part 1. (Loeb Classical Library, 531).Cambridge, MA and London. 15. Laks, A., & G. W. Most. (Eds.). (2016). Early Greek philosophy, Vol. 9: Sophists, Part 2. (Loeb Classical Library, 532). Cambridge, MA and London. 16. Lloyd, G. E. R. (1979). Magic, reason and experience. Cambridge. 17. MacDowell, D. M. (1961). Gorgias, Alkidamas, and the Cripps and Palatine manuscripts. Classical Quarterly, 11, 113–24. 18. MacDowell, D. M. (1982). Gorgias: Encomium of Helen. Bristol. 19. Mann, W.-R. (1992). Rechtfertigung I. Griechische Antike; Logik und Dialektik. Historisches Wörterbuch der Philosophie, 8(Basel), 251–256. 20. Mann, W.-R. (2006). Was kann man von Euthydemos und seinem Bruder lernen? In C. Rapp & T. Wagner (Eds.), Wissen und Bildung in der antiken Philosophie (pp. 103–126). Stuttgart 21. Müri, W. (1944). Das Wort Dialektik bei Platon. Museum Helveticum, 1, 152–68.

3 Gorgias and the Weakness of Logos

67

22. Nehamas, A. (1990). Eristic, Antilogic, Sophistic, Dialectic: Plato’s demarcation of philosophy from sophistry. History of Philosophy Quarterly, 7, 3–16. 23. O’Sullivan, N. (1993). Plato and ἡ καλουμένη ῥητορική. Mnemosyne, 46, 87–89. 24. Pendrick, G. (1998). Plato and RHTORIKH. Rheinisches Museum, 141, 10–23. 25. Schiappa, E. (1990). Did Plato coin Rh¯etorik¯e? American Journal of Philology, 111, 457–470. 26. Schiappa, E. (1999). The beginnings of rhetorical theory in classical Greece. New Haven, CT. 27. Schiappa, E. (1991). Protagoras and Logos (2nd ed.). Columbia, SC. 28. Scodel, R. (1979). Dissertation summary: The Trojan Trilogy of Euripides. Harvard Studies in Classical Philology, 83, 407–409. 29. Scodel, R. (1980). The Trojan Trilogy of Euripides. Göttingen. 30. Sienkewicz, T. (1984). Review of Ruth Scodel, The Trojan Trilogy of Euripides. American Journal of Philology, 105, 482–484. 31. Solmsen, F. (1954). Review of L. Rademacher, Artium Scriptores (Reste der voraristotelischen Rhetorik). Gnomon, 26, 213–19. 32. Spatharas, D. G. (2001). Patterns of argumentation in Gorgias. Mnemosyne, 54, 393–408. 33. Toohey, P. (1994). Epic and rhetoric. In I. Worthington (Ed.), Persuasion: Greek rhetoric in action (pp. 153–167). London and New York, NY. 34. Usher, S. (1999). Greek oratory: Tradition and originality. Oxford. 35. Verdenius, W. J. (1981). ‘Gorgias’ Doctrine of Deception’. In G. B. Kerferd (Ed.), The Sophists and their legacy (pp. 116–128). Wiesbaden. 36. West, M. L. (2013). The epic cycle. Oxford. 37. Worman, N. (2002). The cast of character. Princeton, NJ.

Chapter 4

Plato’s Dialogues: Dialectic, Orality and Character Mathieu Marion

Abstract It is first argued that dialectic was a form of regimented debate, which grew out of public debates in Ancient Greece. A set of rules for dialectical bouts is then given and their meaning explained. The transition from oral to written arguments is briefly examined, leading to the formulation of a delimitation problem in Plato’s dialogues, as he inserted dialectical arguments within ordinary dialogue contexts, turning them into discussions where one of the participants reasons hypothetically to make the other realize that they are not entitled to their view. Doing so, Plato adjusted dialectic to a variety of dialogue purposes and in order to explore this variety, a study of the early tradition of classifying Plato’s dialogues in terms of their ‘character’ is suggested, the results of which are then compared with types of dialogues in contemporary Argumentation Theory.

4.1 Public Debates and Dialectic as Regimented Debate

Ut nihil affirmet ipse, refellat alios

G. E. R. Lloyd argued in Magic, Reason and Experience that beliefs about nature were subjected within Ancient Greek culture to the same “radical examination” as political views: they were openly challenged in public debates, where every assumption was liable to be scrutinized.1 In the opening section of On the Nature of Man, the author describes public debates between contending speakers concerning such beliefs: He who is accustomed to hear speakers discuss the nature of man beyond its relations to medicine will not find the present account of any interest. For I do not say at all that a man is air, or fire, or water, or earth, or anything else that is not an obvious constituent of man; such accounts I leave to those that care to give them. Those, however, who give them have 1 [42,

p. 248].

M. Marion (B) Département de philosophie, Université du Québec à Montréal, Montréal, QC H3C 3P8, Canada e-mail: [email protected] © Springer Nature Switzerland AG 2021 J. A. Bjelde et al. (eds.), Essays on Argumentation in Antiquity, Argumentation Library 39, https://doi.org/10.1007/978-3-030-70817-7_4

69

70

M. Marion not in my opinion correct knowledge. For while adopting the same idea they do not give the same account. Though they add the same appendix to their ideas – saying that ‘what is’ is a unity, and that this is both unity and the all – yet they are not agreed as to its name. One of them asserts that this one and the all is air, another calls it fire, another, water, and another, earth; while each appends to his own account evidence and proofs that amount to nothing. The fact that, while adopting the same idea, they do not give the same account, shows that their knowledge too is at fault. The best way to understand this is to be present when they contradict (αÙτoισ‹ν ¢ντιλšγoυσιν). Indeed, when the same speakers contradict one another (πρ`oς γαρ ` ¢λληλoυς ´ ¢ντιλšγoντες αÙτo„) in front of the same audience, the same man never prevails in the debate three times in succession, but now one is victor, now another, now he who happens to have the most glib tongue in the face of the crowd. Yet it is right that a man who claims correct knowledge about the facts should maintain his own argument victorious always, if his knowledge be knowledge of reality and if he set it forth correctly. But in my opinion such men by their lack of understanding overthrow themselves in the words of their very discussions, and establish the theory of Melissus.2

The debates witnessed by the author presumably took place on the coast of Asia Minor in the late fifth or early fourth century BCE, and the hypotheses discussed, coupled with the mention of Melissus, indicate that they involved philosophers. This oft-quoted passage—Lloyd himself refers to it frequently—3 certainly illustrates the importance of public debates within Ancient Greek culture. There are of course other references; for example, they are mentioned elsewhere outside philosophy by Thucydides4 as well as Isocrates5 and they are satirized in Aristophanes’ Clouds.6 Lloyd has investigated extensively the rise of dialectic and rhetoric over the background of these public debates, and their impact on the emergence of ‘demonstration’, leading further to the emergence of Greek mathematics and science.7 More recently, Reviel Netz argued in The Shaping of Deduction in Greek Mathematics that the central role of debate in Greek culture explains the importance of persuasion, which created in turn the conditions for the emergence of the deductive method in Greek mathematics.8 Independently from these, Benoît Castelnérac and I put forward more than a decade ago a proposal for the study of ancient dialectic as a regimented form of public debates,9 hence my speaking here of ‘dialectical bouts’, to mean what are 2 Nature

of Man I, quoting from [31, pp. 3–5], translation modified and my italics. example [41, p. 61] and [42, pp. 92–93]. 4 In the Mytilenean debate, Thuc. III, 38. 5 In Paneg. 45 and Antid. 295. 6 Nub. 937–947 and 1330–1345. 7 For example, see [42, Chapter 2], [43, Chapter 3] and [44, Chapter 1]. 8 See [51, Chapter 7]. 9 See [10]. There were related proposals put forth at roughly the same time by Laurent Keiff and Shahid Rahman [33] and Catarina Dutilh Novaes [22]—this last circulated for a few years before being published. The principal feature that distinguishes our proposal from these is the ‘Socratic Rule’ (see R5 below). The set of rules in [10] was modified in later versions: [11], [47], and [48]. In [11] Castelnérac and I discussed features specific to Eleatic dialectic (which we called ‘antilogic’ in the etymological sense the expression), and [47] reinterprets Zeno’s arguments on motion in this setting. In [48] and [15], this original project is extended to the study of the dialectical roots of Aristotle’s syllogistic. 3 For

4 Plato’s Dialogues: Dialectic, Orality and Character

71

also often referred to as ‘eristic moots’.10 We also see dialectic as a method for producing persuasive arguments. Following Aristotle, participants are named here after their role in the discussion: Questioner and Answerer. Questioner’s task is to elicit assent from Answerer, via a process of questions and answers, to premises from which a contradiction is derived in Answerer’s own beliefs. Given that anyone could take the role of Answerer, even if they do not share their beliefs, anyone could recognize the validity of the reasoning. Historians of philosophy and logic often ignore the fact that ideas make sense against the backdrop of their historical context, and there has been a tendency to personalize the study of dialectic. In history of logic, the traditional view is that Aristotle, who wrote Topics (and Sophistical Refutations) as a manual for participants in dialectical bouts, dropped the whole matter after discovering syllogistic—this view obscures the true relationship between dialectic and syllogistic by assuming they are independent from one another.11 In history of philosophy, some see dialectic as having originated with Socrates,12 others see it as having ended when Plato supposedly discovered its limitations.13 Others assume that Plato had a trademark, so that any new twist in his understanding of the usefulness of dialectic meant a change in the nature of dialectic itself.14 Richard Robinson is also known for the slightly different but related claim, that ‘dialectic’ was for Plato merely an honorific title which he applied to what seemed to him at the moment to be the best method, as if he applied 10 For

example, by Ryle in [59]. W. D. Ross’ claim that the discovery of syllogistic had rendered Topics “out of date”, as a reason for ignoring that book [57, p. 59]. On the dialectical origins of syllogistic, see the opposite view in [48] and [15]. 12 For example, Sidgwick’s early claim in [60]. Claims about the inventor of the method are certainly not new, to wit Aristotle’s claim that it was Zeno, reported in Diogenes Laertius, Vitae VIII, 57 and Sextus Empiricus, Ad. Math. VII, 7, or Diogenes Laertius’ claim in Vitae IX, 52–53 that it was Protagoras. At all events, there are reasons to believe that the practice is older than Socrates’ generation, harking back at least as far back as Eleatic philosophy. It is not possible to attribute with any degree of confidence the invention of dialectic to one person, and a better and more modest claim would be that Eleatic philosophers were the first ones fully to use features of adversarial debates for the purpose of philosophical argumentation, thus providing a more self-conscious regimentation. 13 I refer here to Vlastos’ doubtful developmental claims [67, p. 37], according to which Plato realized the uselessness of dialectic while writing Gorgias, this realisation leading to its supposed “euthanasia” in the next dialogues (Lysis, Euthydemus, Hippias Major and Meno). Henceforth, dialectic would have become inoperative in his philosophy. Later occurrences of dialectical exchanges are simply discounted, for example, Rep. I is assumed to be an earlier dialogue, which was recycled in order to provide a ‘pastiche’ [66, p. 250]. Among the many reasons why these developmental theses fail to convince is that they (ideally) rely on a precise sequencing of the dialogues that cannot be supported by the somewhat meagre results of stylometric studies. See [50, Chapter 5] for a further discussion of this point. Needless to say, rejecting developmental readings does not commit one to any of the alternative ‘unitarian’ readings. 14 See, for example [18]. Plato certainly sought better to understand the ‘form’ of dialectic and what one could achieve with it, and thus came up with ideas such as the ‘method of division’ (Phdr. 265d–266c, Soph. 253b–e & 218b–231e, Plt. 264b–267c) as a method for producing definitions that might in turn be tested in dialectical bouts. It would be incorrect to claim that he thus brought about a change in the very nature and practice of dialectic, especially since there are reasons to believe that this practice went on, virtually unchanged, for centuries afterwards. 11 See

72

M. Marion

it to different methods at different times.15 One might as well claim the opposite, namely that what appeared to Robinson as Plato giving the same label to supposedly different ‘methods’ was, rather, Plato trying to show how the very same ‘method’ can be put to different uses in different dialogue contexts. Of course, one cannot fully argue this point within a single paper, so I hope merely to examine the following suggestion. Against this tendency to personalize dialectic, Castelnérac and I assumed, as Lloyd and Netz did, that dialectic emerged from the practice of public debates, when they were regimented by a set of rules, so that what came variously to be called not only ´ but also ‘antilogic’ (¢ντιλoγ´ια or ¢ντιλoγικη´ τšκνη) in the ‘dialectic’ (διαλεκτικη) sense of ‘arguing against’ or ‘eristic’ (™ριστικη´ or ™ριστικη´ σoφ´ια) was a common verbal practice, which took place independently from any Ancient philosopher.16 One could even add ‘brachylogy’ (βραχυλoγ´ια) to this list, that is the practice of asking short questions with yes/no answers in order to elicit assent to premises leading to a contradiction in the adversary’s position—usually referred to as the ‘elenchus’ (λεγχoς). It is opposed to the practice called ‘macrology’ (μαχρoλoγ´ια) of providing long speeches—also called epideixis (™πιδε´ιχεις)—that argue one after the other on opposite sides of a question, as in, for instance, Dissoi Logoi.17 Although the passage from On the Nature of Man can be read both ways, there are reasons to believe that it describes debates that are instance of βραχυλoγ´ια.18 A proper reconstruction of the rules for dialectical bouts should thus be seen as forming part of a reconstruction of the context of origin of philosophical texts, aiming at a better understanding of the arguments they contain.19 A summary of our reconstruction will be presented in Sect. 4.2. 15 [56,

p. 70]. for example, the definition of the sophist at Soph. 226a. On this variety of names see [52] and the concomitant nascent distinction between ‘dialectician’ and ‘philosopher’ in Aristotle, see [49]. 17 Plato drew the distinction between βραχυλoγ´ια and μαχρoλoγ´ια, while rejecting the latter, in Alc. 1 106b, Hp. min. 373a, Prt. 329a–b or Grg. 449b–c. (One should not confuse the meaning of these expressions with that of their modern English counterparts, ‘brachylogy’ and ‘macrology’.) Gorgias had a related distinction in Hel. 13 between what he calls ‘contests of arguments’ (λ´oγων ¢γîνας), involving speeches, and ‘contests of philosophical arguments’ (ϕιλoσ´oϕων λ´oγων ¢μ´ιλλας), in which “it is revealed that rapidity of thought too makes the conviction of an opinion easily changeable” (quoting from [37, p. 179]). The description of these ‘contests of philosophical arguments’ reminds one of both On the Nature of Man, quoted above, and ‘eristic’ as discussed in Sect. 4.4 below. 18 Following, inter alios [59, p. 117]. Among reasons to believe that it is instances of βραχυλoγ´ια that are described is the fact that dialectical bouts were common among philosophers at the time of Melissus, and the idea that there are repeated bouts makes little sense if they were each made up of two long speeches. 19 To use the analogy with chess, to understand the point of a given move within a given game, it is necessary to reconstruct as context the series of moves in which it is embedded, not only in accordance with the rules of chess but also in strategical terms. Likewise, understanding the rules of dialectic would help clarifying the meaning of texts as moves in a game. Our approach to the history of ideas is thus largely in line with that of [61], which is inspired by R. G. Collingwood. 16 See,

4 Plato’s Dialogues: Dialectic, Orality and Character

73

Reviel Netz has also cleverly emphasized the role of orality and the importance of a proper study of the way in which arguments that had been developed in an oral setting would eventually be written down in a form that does not necessarily reflect their origin.20 In order better to understand the emergence and development of the deductive method in Greek mathematics, one has thus to reconstruct the oral context from which it emerged—its origins in public debates—as well as also to analyse carefully the implications of the passage from this oral context to a written one. The following is dedicated to a similar inquiry concerning the written transmission of arguments that were first formulated orally during dialectical bouts, looking specifically at Plato’s dialogues. We have of course no direct transcriptions of dialectical bouts, but it can be assumed that arguments that were first formulated within these found their way, albeit in a modified form, into texts such as Zeno’s ‘treatise’, Gorgias’ On Not-Being or Plato’s dialogues. There has been for decades now a trend within Plato scholarship studying the literary dimensions of his dialogues and bringing to the fore his use of a variety of literary devices: imagery, narrative suspense, etc.21 Plato is never the narrator of his own dialogues and since it cannot be granted that the narrator is Socrates, it is not even clear who speaks for Plato.22 In light of Netz’s point about orality, how are we to find traces of real dialectical bouts in Plato’s dialogues, if they are to be such complex works of literature? Auguste Diès once hypothesized that Plato arranged within the confines of his school dialectical bouts (‘joutes dialectiques’) with distinguished visitors—from places such as Syracuse, Tarentum or Elea—and used this as the material for his dialogues,23 while Ryle made the stronger claim that Plato’s “dialectical dialogues should be read as case-books of recent Moots, dramatized partly to help students to remember and digest the argument-sequences that finally crystallized out of these Moots”.24 Either way, granted that Plato’s dialogues are set against the backdrop of these public dialectical bouts, Plato was obviously not merely transcribing real bouts and the question becomes more complex: how was he inserting dialectical arguments within the rich tapestry of his dialogues? These issues will be discussed further in Sect. 4.3. My aim is not to provide a fully argued answer to these questions, but merely consists in looking at what help one could get from an ancient tradition of classifying ´ dialogues in terms of their ‘character’ (χαρακτηρ). The claim would thus be that arguments originating in dialectical bouts are adapted to a variety of aims within Plato’s dialogues (inasmuch as this is feasible), and that this tradition was based on the kindred idea that dialogues serve a variety of purposes, pointing precisely 20 [51,

pp. 292 and 297–298]. trend started with [26] and [34] and the literature is by now too vast to be accounted for in a footnote but, for other early representative samples, see [4], [32] [53], [58]. 22 See [55]. 23 [16, pp. 315–316]. 24 [59, p. 18]. Ryle had other hypotheses about Plato producing his dialogues for different audiences, to be performed (by him) at ‘cyclical gatherings’ [59, p. 222], etc. Recognizing that his dialogues are set against the backdrop of dialectical bouts does not commit one to these further claims. 21 This

74

M. Marion

to the diversity of uses of dialectic itself within Plato’s dialogues. Section 4.4 will offer an analysis of these classifications. To conclude, I briefly discuss in Sect. 4.5 the results of this analysis in relation to contemporary Argumentation Theory. Such a comparison is of interest not just because dialectic was an important source for Argumentation Theory, but also because classifying Plato’s dialogues in terms of their ‘characters’ can be seen as a form of Argumentation Theory avant la lettre.

4.2 Rules for Dialectical Bouts Aristotle believed that dialectic and rhetoric arose from the human tendency to defend one’s statements and attacking those of others: Both alike are concerned with such things as come, more or less, within the general ken of all men and belong to no definite science. Accordingly, all men make use, more or less, of both; for to a certain extent all men attempt to discuss statements and to maintain them, to defend themselves and to attack others.25

In contemporary jargon, one would talk here of the need to play the ‘game of giving and asking for reasons’.26 Aristotle also saw a need for regimentation, so that one moves from ordinary dialogue situations to dialectic as an art or τšχνη.27 Assuming ´ is but one of the above names for a particular form of that ‘dialectic’ (διαλεκτικη) regimented public debate, one must provide as precise as possible an account of the way in which it was regimented, the aim being to understand how it proceeds and how arguments are generated within these debates and what form they must have had before being written down. The passage from On the Nature of Man quoted above already indicates that bouts occurred in a public place,28 in front of an audience (of whatever size).29 Given that participants defend views associated with different philosophical schools, it is easy to imagine that training for these public encounters would take place away from the public, within the grounds of a given school.30 25 Rhet. 26 On

A, 1, 1354a1–7, quoting from [3, p. 2152]. the ‘games of giving and asking for reasons’ and the motivation for a ‘game semantics’, see

[46]. 27 Soph. el. 11, 172a34–35. 28 This is often left unspecified in Plato’s dialogues, but it could be the agora, a palaestra, a private house, etc. 29 References to the audience are frequent in Plato’s dialogues, for example when Socrates recounts in Prt. 339d–e, a round of applause from the audience, when Protagoras showed he had contradicted himself, that felt as if he had been hit by a boxer. See also references to the audience in Soph. el. 8, 169b31 & 15, 174a36. 30 See for example, the testimony in Cicero, Acad. pos. I. iv. 17. Plato speaks of a period of training of five years in Rep. 539e, that cannot be reasonably assumed to have occurred outside the Academy, while Aristotle refers in Soph. el. 34, 184a8–184b3 to Topics and Sophistical Refutations as the result of long arduous empirical research, which would not make much sense if it had been conducted in public debates as opposed to the privacy of the Lyceum.

4 Plato’s Dialogues: Dialectic, Orality and Character

75

Bouts would have involved two participants taking turns asking questions and giving answers. One must, however, avoid confusing the participants themselves and the roles they assume within a given bout. As Socrates tells in Grg. 462d, there are only two roles: “So now please do whichever of these you like: either ask questions or answer them”. (Socrates himself is portrayed by Plato as assuming both roles on different occasions.) As mentioned above, the names for these roles are those of Questioner and Answerer. A bout would start with Answerer committing to a thesis, say A, and Questioner’s role would be to elicit through a series of questions Answerer’s commitment to further theses that would then be shown to entail, when taken together with A, a contradiction (possibly ¬ A but not necessarily). Answerer’s task is simply to avoid being caught. One can thus establish the following structural rules R1–R7 for dialectical bouts, interspersed here with brief elucidatory comments: R1.

Bouts always involve two players, that are named Questioner and Answerer, in virtue of the role they assume at the beginning of the bout

As just pointed out, one must distinguish between discussants and the roles they assume. This explains why substitutions are permitted in mediis rebus, as in the opening of Philebus when Protarchus takes over the role of Answerer from Philebus, and why roles can be switched, as in Phlb. 13e, although Protarchus ends up not taking up the invitation to switch roles and act as Questioner. R2.

A play begins with Questioner eliciting from Answerer his commitment to an assertion or thesis A

In Top. A, 101b32–33, Aristotle calls the initial thesis ‘πρóβλημα’, as the Answerer is given the choice between the pair A or ¬ A. Plato usually has Socrates playing the role of Questioner and cunningly eliciting one of the alternatives from his opponent at the beginning of the bout. The theses that are thus forming the starting point are also described as ‘reputable opinions’ (νδoξα) by Aristotle Top. A, 100a23–25.31 Much has been written on the opening paragraphs of Topics and the distinction between ‘dialectical syllogisms’, that are said to be based on reputable opinions and ‘demonstrations’, that are syllogisms based on true premises (see Sect. 4.4 below). R3.

R4.

The play then proceeds through a series of alternate questions and answers. Questioner asks questions such that Answerer may give a ‘short answer’, ideally ‘yes’ or ‘no’ (Top. , 158a15). Proceeding thus, Questioner elicits further commitments from Answerer commitment to further assertions B1 , B2 , …, Bn . The set {A, B1 , B2 , …, Bn } is called Answerer’s ‘scoreboard’32 or ‘commitment store’.33

31 Aristotle

allows also paradoxes, if they have been held by some philosopher, Top. A, 104b19– 105a2. 32 See [40]. 33 For Hamblin’s notion of ‘commitment store’, see [27, p. 257]. This notion has been influential in Argumentation Theory, see for example [72, p. 8].

76

R5.

M. Marion

Questioner cannot introduce in Answerer’s scoreboard or use any premise (πρóτασις) that has not been explicitly granted by Answerer. (Socratic Rule)34

This last is called the ‘Socratic rule’ because it is related to Socrates’ repeated and well-known ‘avowals of ignorance’.35 Indeed, qua Questioner, Socrates must not introduce any premise in Answerer’s scoreboard. Otherwise, Answerer could simply refuse to concede the contradiction, arguing that it is based on a premise surreptitiously introduced in the scoreboard by Socrates. Thus, when questioning, Socrates must declare himself ignorant at all times simply in virtue of the demands of dialectic. The Socratic Rule is also related to another very important feature of dialectic, the need for Answerer to comply with the doxastic or ‘say what you believe’ constraint.36 The starting point is a thesis that Answerer is committed to and contradictions are to be derived from it along with further premises that are believed by Answerer. This is why so-called ‘refutations’ are said to be ad hominem.37 Furthermore, premises are neither known to be true prior to the bout nor proven to be true during it, and Questioner’s sole task is to reason hypothetically in order to infer a contradiction from Answerer’s own commitments. R6. R7.

Questioner can then ‘take together’ or ‘add up’ (συλλoγ…ζεσθαι) premises 38 from Answerer’s scoreboard to infer an impossibility (¢δνατoν). ´ If Questioner infers an impossibility, Questioner wins, otherwise Answerer wins. (Winning rule)

These last rules show that the Principle of Non-Contradiction plays an essential role in dialectic, so one finds it frequently stated.39 Dialectical bouts can thus be

34 The

name ‘Socratic Rule’ comes from [48, p. 210], where it is explained how it is analogous to the ‘Formal Rule’ in contemporary game semantics (in the school of Lorenzen). 35 For example, at Hp. min. 372b–e or Tht. 150c–d. 36 For example, Prt. 331c–d, Cri. 49c–e, or Chrm. 166d–e. 37 Although, of course, one can see that it is the beliefs in the scoreboard that are inconsistent. Otherwise the ‘refutation’ of a particular individual would be of little interest. 38 This is how Plato puts it in Soph. 230b: “they collect his opinions together during the discussion, put them side by side, and show that they conflict with each other at the same time on the same subjects in relation to the same things, and in the same respects” (Quoting from [12, pp. 250–251].). 39 The earliest statement is in Gorgias’ Defence of Palamedes (§ 25). Plato states it in Ap. 26e–27b, Grg. 495e–496a, Rep. IV 436e–437a & X 602e, and Soph. 230b, quoted in the previous footnote, while Aristotle states it in An. Pr. A 52a2–3 and Met.  3.1005b23. As a matter of fact, Aristotle states three variations of the principle of non contradiction in Met.  3, and famously defends it in Met.  4 with an argument involving dialectical bouts.

4 Plato’s Dialogues: Dialectic, Orality and Character

77

seen ‘consistency maintenance’ tests, to borrow an expression from Catarina DutilhNovaes.40 Assuming that inference is an act, the term ‘συλλoγισμ´oς’ is taken in R6 to refer to the act of inferring the conclusion.41 This inference to an impossibility (absurdity or contradiction) from Answerer’s scoreboard is what is usually referred to in the secondary literature as the ‘elenchus’ (λεγχoς). It is better to avoid translating that word with ‘refutation’, since this would be misleading: for sure, with the ‘elenchus’ the set of Answerer’s commitments is shown to be inconsistent, but the initial commitment to A has not been ‘refuted’ in the sense that it would have been shown to be false, for the simple reason that the argument is not a ‘demonstration’ on the basis of true premises, it is only the derivation of a contradiction on the basis of premises merely believed by Answerer. In his well-known discussion of the ‘Socratic elenchus’, Vlastos imputes this mistake to Socrates, who he portrays as supposedly believing that, having drawn a contradiction from A, he has shown that it is false and thus that ¬ A is true.42 One could represent the derivation of the contradiction thus: A, B1 , B2 , . . . , Bn  ⊥

(4.1)

Vlastos’ claim is that Socrates moves from (4.1) to: B1 , B2 , . . . , Bn  ¬ A

(4.2)

It is unlikely, however, that Plato had his character commit this blunder. His astuteness in these logical matters is rather striking, as one can see from Charmides, where Critias, qua Answerer, objects to Socrates, who has just derived the contradictory of his original claim: And if you [Socrates] think it necessary to draw this conclusion from what I admitted before, then I would rather withdraw some of my statements, and would not be ashamed to admit I made a mistake, in preference to conceding that a man ignorant of himself could be temperate.43

Here Answerer prefers to withdraw one or more of B1 , B2 , …, Bn rather than to concede ¬ A, for example moving from (4.1) to: A, B2 , . . . , Bn  ¬ B1

40 The

(4.3)

expression is taken from [21], where it is applied to the Medieval counterpart of dialectic, known as ‘obligationes’. 41 This point is argued in [14, p. 26]. 42 This is what Vlastos calls the ‘standard elenchus’ [67, p. 11]. 43 Chrm. 164c–d, quoting from [12, p. 651].

78

M. Marion

Perhaps one should think here of Diodorus Cronus’ Master Argument as an illustration.44 Indeed, if we are to follow Epictetus, Diodorus had shown the incompatibility of three premises taken together, and solutions vary according to which premise is rejected.45 One could extend this set of basic rules R1–R7 to include restrictions on the length of plays, etc.46 From the point of view of the history of logic, it is worth noting that the ´ at Top. , 157a34–37 & 160b1–6 is linked with to the rule for ‘induction’ (™παγωγη) dictum de omni in An. Pr. A, 24b28–29.47 This is not a structural rule as in R1–R7, but a particle rule stating that, if Questioner wishes to put forward a universal affirmative thesis, then he can only do so on the basis of having previously obtained from Answerer commitment to some cases.48 Once it has been so introduced, Answerer should either concede it or challenge it by putting forward a possible counterexample. This shows an elaborate understanding of quantification, akin to our contemporary view of universality in terms of the absence of a counterexample, and it is illustrated in numerous occasions in Plato’s dialogues.49 The foregoing rules R1–R7 provide a bare account of the structure according to which dialectical bouts appear to have proceeded, but I hope to have already indicated in this brief summary that they are rich in useful indications for reading Plato’s dialogues. So far, I have only described rules for the conduct of any particular bout. Aristotle even listed criteria for the evaluation of the resulting argument in Top. , 161b19–33. Repeated bouts on a given thesis A might yield different arguments (depending on what premises go into the scoreboard), likewise for ¬ A. Thus, dialecticians turn out to be useful on a controversial matter, as they draw out arguments on both sides, so that one would ideally be able to choose the better side after carefully examining arguments on both sides, assessing their strength. One should distinguish between the dialectical bouts themselves and strategical uses of dialectic itself to generate arguments. For example, the above suggests a particular use of dialectic illustrated by both Gorgias’ treatise and Plato’s Parmenides as they derive contradictions on both sides, A and ¬ A, thus potentially reaching an ‘aporia’ (¢πoρ´ια) or “equality between contrary reasonings”, to borrow Aristotle’s expression (Top. Z, 145b17). This much indicates that one could devise strategies to use dialectical bouts for more positive purposes such as exploring arguments ‘on both sides’, not just demolishing someone else’s view in public.50 44 I

owe this observation to Christopher Lemonnier. diss. II, 19, 1–5. See [68, Chapters 1–2] for an example of a reconstruction along similar lines. 46 See Top. , 161a10 and Soph. el. 34, 183a 25. 47 The link with the dictum de omni is argued for in [48] and the approach is extended to an explanation of ‘ecthesis’ in [15]. 48 Taken together with a limit on the time of the bout, the rule for ™παγωγη ´ would also motivate a further rule against delaying tactics, since a player could try and buy some time with repeated challenges, instead of conceding. For a discussion of Aristotle on this point, see [48, pp. 224–225]. 49 For an analysis of passages in Plato’s dialogues, see [48, pp. 219–224]. 50 Gorgias’ treatise and Plato’s Parmenides—with the third series of deductions starting at 155e— also exhibit what seems an ‘Eleatic’ trait, since one could argue that it harks back to Parmenides’ 45 Epict.

4 Plato’s Dialogues: Dialectic, Orality and Character

79

4.3 From Oral to Written Arguments: The Delimitation Problem What happens when the results of dialectical bouts are written down? To extract the argument, one can obviously write it down step by step, from the initial νδoξoν until the inference to the contradiction, simply stringing together Answerer’s commitments while editing out the exchange between Questioner and Answerer and any other matter deemed irrelevant, such as concessions irrelevant for the final argument. Zeno’s or Gorgias’ arguments were transmitted to us in this compressed form, where any trace of the live exchange and even of the second person has disappeared, but one could easily put them back in their original dialectical form. For example, Zeno’s ‘Millet seed’ is reported by Aristotle at Phys. H 5, 250a19–22 in compressed form, without trace of dialectic: Therefore Zeno’s argument is not true, that there is no part of a grain of millet that does not make a sound: for there is no reason why any such part should not in any length of time fail to move the air that the whole bushel moves in falling.51

However, when Simplicius, who probably had Zeno’s treatise in front of him when writing his Commentary, 1108.18, ad loc., comments on this passage, he deploys the argument dialectically: By this means he solves the conundrum which Zeno the Eleatic asked Protagoras the sophist. “Tell me Protagoras,” he said, “does a single grain of millet or the ten thousandth part of a grain make any sound when it falls?” And when Protagoras said it did not, “Then”, asked Zeno, “does a bushel of millet make any sound when it falls or not?” Protagoras answered that it did, whereupon Zeno replied, “But surely there is a ratio between a bushel of millet and a single grain or even the ten thousandth part of a grain’; and when this was admitted, “But then surely”, Zeno said, “the ratios of the corresponding sounds to each other will be the same: for as bodies which make the sounds are to one another, so the sounds will be to one another. And if this is so, and if the bushel of millet makes a sound, then the single grain of millet and the ten thousandth part of a grain will make a sound.” This was the way Zeno used to put his questions.52

Admittedly, the exchange between Protagoras and Zeno is imaginary, but this does not show that this dialectical version is contrived or that the origin of the argument is not dialectical: Simplicius’ claim that “This was the way Zeno used to put his questions” would not make sense if he did not believe this to be the case. An analogous point Poem and its three ‘paths’: contradictions are also derived from ‘A & ¬ A’. This appears at first sight in line with exclusive disjunction: (A ∨ B) & ¬ (A & B) and could be argued again to hark back to Parmenides’ Poem and its three ‘paths’. For steps towards reading Parmenides in this manner, see [9]. Searching for arguments, specific strategies can be adopted. In Zeno’s lost treatise, Gorgias’ On Not-Being, and in the second part of Plato’s Parmenides contradictions are generated along contradictory pairs of predicates, such as limited/unlimited, divisible/indivisible, etc. Parmenides’ Poem offers an earlier list of such contradictory pairs. A list of that kind might also be at the origin of the list of ‘categories’ in Top. A, 103b20–26 and, following [75], we can see De Interpretatione as a study of such contradictory pairs of predicates. 51 Quoted from [39, p. 109]. 52 Quoted from [39, p. 109].

80

M. Marion

can be made about the related Sorites paradox, attributed to Eubulides of Miletus by Diogenes Laertius as being specifically of a dialectical form.53 Again, Aristotle does not present it under that form Phys.  3, 253b14–22 or Soph. el. 24, 179a35, but, to mention only one example, Cicero’s exposition—again, later and imaginary—is in a dialogue form, which seems more natural.54 The disappearance of the second person from accounts of arguments originating in dialectical bouts has misled more than one competent reader, given that the arguments are then couched in the form of conditionals. Thus, Vlastos claimed that Zeno’s arguments differ fundamentally from Socrates’ because its “refutands […] are unasserted counterfactuals”,55 while Lloyd sees Eleatic arguments as instances of Modus Tollens.56 But we saw in the previous section that Questioner’s task is to reason hypothetically in order to infer a contradiction from Answerer’s own commitments. If we reconstruct Zeno’s arguments dialectically, we can then see that he did not argue ‘against movement’: as Questioner, he is not committed at all to what follows from the commitments of his adversaries, who hold that ‘what is’ is many and divisible (infinitely or not): he only derives this absurd consequence from this premise (and further concessions). Thus, Zeno argues against “‘what is’ is many” by showing that it entails the absence of movement, this being a difficulty for his adversaries, not a conclusion he is committed to or even wants to argue for.57 There is an analogous danger with Plato’s dialogues. The unadorned second part of Parmenides offers us the best-preserved illustration of a dialectical bout, although it is an imaginary one and it is probably a composite of arguments originating in more than one real bout. Many readers will certainly remember their first impression of these Stephanus pages, with the young Aristotle providing a total of 532 short answers to Parmenides’ questions, leading to what appears at first sight to be pointless contradiction after pointless contradiction. F. M. Cornford even incorporated in his commentary a translation expurgating the young Aristotle altogether, under the pretext that “nothing is gained by casting the arguments into the form of question and answer”, as “it only increases the difficulty of following the reasoning”.58 Cornford’s editing out of Aristotle shows how one could extract an argument from a dialectical exchange but, although Cornford’s mutilation may not appear unwarranted at first sight in the case of the second part of Parmenides, reading Plato’s dialogues as if dialectic is only a matter of superficial garb with no consequence for our understanding of the very point of the arguments is plainly incorrect. The explanation of dialectical bouts in the previous section help us understanding their aim: when a Questioner such as Zeno or Socrates derives a contradiction from 53 Vitae

II, 108. Pr. II. xxix. 92–95. See also Aspasius, In Eth. Nic., 56, 32–57. 55 [67, p. 2]. 56 [42, p. 77]. 57 See [47]. For spadework in this direction concerning Gorgias’ treatise, trying first to reconstruct the chain of inferences from the conflicting versions in Aristotle’s MXG and Sextus Empiricus, see [7] et [8]. 58 [13, p. 109]. 54 Acad.

4 Plato’s Dialogues: Dialectic, Orality and Character

81

premises to which Answerer assented, they are not committed to the contradiction, since they have not committed themselves to the premises. So, they cannot be said to have held the absurd conclusion thus derived, it is only meant as a sort of poke in the wheels for Answerer. Furthermore, the ‘elenchi’ in the second half of Parmenides can be parsed in (eight or nine) ‘deductions’, that form a structure quite clearly inspired by Eleatic dialectic. This structure implies that the ‘deductions’ have a presumptive equality in standing, which cannot be discounted as easily as it is in some commentaries, where one tries to make one set of deductions play a central role at the expense of the others.59 Plato’s use of the literary form of the dialogue has its own interesting issues. The second part of Parmenides also has the peculiarity that one can easily identify the bout: it begins when Aristotle volunteers at 137c for the role of Answerer, with Parmenides as Questioner, and it ends with the dialogue. There are other instances where identification is clear, for instance when Gorgias condescends to give ‘short answers’ at Grg. 449b–c or when Socrates elicits from Hippias the answer that starts the bout in Hp. min. 364c, etc. On the other hand, while a dialectical bout would stop once a contradiction is reached, this is not so in most of Plato’s dialogues, and this is more reflective of real-life ordinary discussions. For example, Hippias Minor ends on Socrates having derived that the better person is the one who voluntarily commit injustice, but in the course of the argument Socrates uses Hippias’ commitments to infer at 371e that ‘Ulysses is better than Achilles’, which is the contradictory of one of Hippias’ earlier statements. The exchange does not stop at this as a dialectical bout would, because an objection by Hippias restarts it immediately, leading it towards its final ‘elenchus’ at the end of the dialogue. This feature is typical of ordinary dialogue situations. There is another instance in the first stages of Philebus, where a double ‘elenchus’ occurs at 21d–e: although Socrates gets Protarchus to concede that the good in human life is not pleasure, which is the contradictory of the thesis Protarchus agreed to defend as Answerer, this is quickly followed by the concession that neither is it reason—a thesis Socrates was pretending to hold as Questioner. Socrates then suggests a third possibility, that it is a mixture of pleasure and reason, and gets Protarchus to agree (he does it with a double instance of the disjunctive syllogism at 22a–b), so that we go along towards the claim that in such a life, reason gets the second prize. This points to a delimitation problem, as Plato sharply distinguishes at times the beginning and end of a bout, while on other occasions he seamlessly blends in elements of dialectic as regimented dialogue within a dialogue situation that retains the flavour of real-life ordinary conversational back-and-forth, albeit with the feeling that the resulting exchange is perhaps a bit more contrived than a real dialogue. What makes Plato’s dialogues so peculiar is thus that he inserts dialectical exchanges within ordinary dialogue contexts, turning the latter into discussions where one of the participants reasons hypothetically to make the other realize that they are not entitled to their view. By the same token, he adjusts dialectic to a variety of dialogue 59 Neoplatonist readings come to mind, see [19]. For a critique, see [1, pp. 189–195] and for a brief overview of the symmetry of the deductions, see [24, pp. 54–59].

82

M. Marion

purposes. It is thus worth looking at the variety of aims of the ordinary dialogue situations in his dialogues, so to explore the variety of ways in which he was able to embed dialectical arguments. To achieve this, I suggest that we take a look at an early tradition—the earliest one, in fact—concerning Plato’s dialogues, that of classifying 60 ´ them in terms of their type or ‘character’ (χαρακτηρ).

4.4 Classifying Plato’s Dialogues Evidence of this early tradition is found in Diogenes Laertius,61 who attributes to Thrasyllus a classification of Plato’s dialogues.62 There are other classifications within Middle Platonism, by Albinus in his Prologue and Alcinous in his Handbook of Platonism, both second century CE.63 The point of these classifications was to establish a sequential reading of Plato’s dialogues on the basis of their character, and it has been replaced in the nineteenth century by ‘developmental’ approaches that explain the content of the dialogues in terms of the development of Plato’s thought (linearly, in one particular direction). Given that there are only a few textual indications of the order in which Plato wrote his dialogues, and given that developmentalists assume that each dialogue must have expressed Plato’s position at the time of writing so that they can plot the evolution of his thought one dialogue at a time,64 they have come to rely in typically nineteenth century positivist manner on the results of stylometric studies—the latest version being computer-assisted.65 By contrast with the early tradition, the aim here was to find the exact order in which Plato wrote his dialogues, so to be able to read them in this chronological sequence, plotting changes of mind as one goes along. The contrast could not be greater, but the exact sequence in which Plato wrote his dialogues is an ideal which is very clearly unattainable. This is very easily verified by comparing the chronology of two of the last attempts, by Brandwood and Ledger.66 Debra Nails has shown that the only broad agreement at the moment between stylometric and philological chronologies concerns a group of later dialogues (Timaeus, Critias, 60 The locus classicus

here is [2], especially p. 129f. There are useful discussions in [45, Chapter 2] and [6], and detailed historical treatments in [54, pp. 101–143], also providing an edition of Albinus’ text, pp. 30–34 and [64, pp. 46–57]. It is extremely difficult to draw any solid conclusions as to the origins of the practice of assigning ‘characters’ to Plato’s dialogues. Tarrant concludes that, although there are more ancient sources probably arising from the interpretation of Theaetetus, Thrasyllus should be seen as responsible for the full classification in Diogenes Laertius, Vitae III, 49–51. This issue is, however, only tangential to the concerns of this paper. 61 Vitae III, 48–66. 62 Vitae IX, 45. 63 Agreeing here with Whittaker in [76] in rejecting Freudenthal’s identification of the two. 64 For the contrary claim, see [65, p. 23]. Thesleff rejects both developmental and stylometric approaches, focusing on the evolution of Plato’s style. 65 See [38]. 66 Respectively [5] and [38].

4 Plato’s Dialogues: Dialectic, Orality and Character

83

Sophist, Statesman, Philebus and Laws).67 There is no other solid consensus for other dialogues or periods, and certainly even less concerning a particular sequence within a given period. It seems, therefore, that, for all its worth, an appeal to the results of stylometry provides no solid ground for keeping the early tradition buried. Of course, my point is not to revive the idea of a classification in terms of ‘characters’, while seeking perhaps to classify anew the dialogues, but simply to suggest that one might find more in the early tradition than odd reasons for editing Plato’s dialogues in trilogies or tetralogies. The claim would rather be that attributing characters to Plato’s dialogues was a practice somewhat analogous to our contemporary classification of the types of arguments within Argumentation Theory, and that a closer look at the dialogues within a perspective akin to the latter might help us dealing with the delimitation problem, and thus better to understand the relation between the real-life practice of dialectic and its embedding within Plato’s dialogues. Early editions of Plato’s dialogues68 contained further information, as they provided for each one a title and a subtitle, and Aristophanes of Byzantium’s edition grouped them in trilogies (diacritical signs were included as well), while Thrasyllus grouped them in tetralogies. These were devised to provide an order in which to read Plato’s dialogues—see below Albinus’ five stages of Platonic education. There is of course no lack of diverging views on that matter here too and it is better not to delve into these, and to stick to a discussion of the ‘characters’. These classifications also came with lists of dialogues appended to each character; another issue that I shall leave aside, as such pairings raise a raft of subsidiary questions of little value here. There is, however, an important objection involving what is known as ‘dialectical shifts’ in contemporary Argumentation Theory,69 that undermines the very idea of attributing a character to a dialogue as a whole: the aim of the discussion might change—even more than once—within a given dialogue, and this would involve a change in the type of argument used,70 and the attribution of a single character to a dialogue might simply hide those changes and bring about a superficial sense of unity. Proclus had already sensed it, given that, at the beginning of his commentary on the Republic,71 he classified that dialogue as a mixture between two types. At any rate, my point is not to validate any classification of the dialogues per se, but simply to explore the thinking that went on behind the attribution of characters, so that we could extract some ideas to look at the passage from the verbal arguments to their written versions within Platonic dialogues. It seems useful to preface the study of ‘characters’ with a look at classifications by Aristotle between types of ‘arguments’ (λ´oγoι) in conversation and types of 67 See

[50, Chapter 4].

68 The oldest edition is from the Academy, under the scholarch Xenocrates in the third century BCE,

followed by Aristophanes of Byzantium’s towards the end of the same century, and Thrasyllus’ in the first century CE. 69 See, for example [72, section 3.3]. 70 This point, implicit in Diogenes’ Vitae III, 65, is often noted in the secondary literature, for instance in [64, p. 46] or [45, p. 81]. 71 Proclus, In R. I.15, 20–22.

84

M. Marion

‘inferences’ (συλλoγισμo´ι), as opposed to classification between types of Platonic dialogues. In Soph. el. 2, 165a38–165b12, Aristotle distinguished four types of the former: Of arguments used in discussion there are four classes: didactic, dialectical, examinational, and contentious arguments. Didactic arguments are those that deduce from the principles appropriate to each subject and not from the opinions held by the answerer (for the learner must be convinced); dialectical arguments are those that deduce from reputable premises, to the contradictory of a given thesis; examinational arguments are those that deduce from premises which are accepted by the answerer and which anyone who claims to possess knowledge of the subject is bound to know […]; contentious arguments are those that deduce or appear to deduce to a conclusion from premises that appear to be reputable but are not so. The subject of demonstrative arguments has been discussed in the Analytics, while that of dialectical arguments and examinational arguments has been discussed [in Topics]: let us now proceed to speak of the arguments used in competitions and contests.72

Assuming that ‘demonstrations’ are but the inferences used in didactic arguments and noting that dialectical and examinational arguments are precisely those generated in bouts that follow rules R1–R7 above, we get: didactic − dialectical − examination − eristic Aristotle also distinguishes in Top. A 1, 100a27–101a4 between three types of inferences: demonstration − dialectical − eristic There is no counterpart to ‘examination arguments’ in this last classification, but Aristotle also tells us elsewhere they are but a type of ‘dialectical argument’.73 One should first note that demonstration—valid inference with true premises—appears suited for didactic purposes, while a dialectical syllogism seems suited to the examination of the νδoξα that forms its point of departure (rule R2).74 As for eristic or ‘contentious’ inferences, they are the ‘sophistical refutations’ defined as “what appears to be refutations but are really fallacies instead”.75 In other words, they are not really inferences as such—as Plato would put it at the very end of the Sophist (268c–d): they are ‘imitations’—used in eristic arguments, within, therefore, contexts where one seeks only to win at all costs. This may involve cheating: For just as unfairness in a contest is a definite type of fault, and is a kind of foul fighting, so the art of contentious reasoning is foul fighting in disputation; for in the former case those who are resolved to win at all costs snatch at everything, and so in the latter case do contentious reasoners.76 72 Quoting

from [3, p. 279]. el. 11, 171b3–6, or simply Top. , 5. 74 Top. A, 100a29–30. 75 Soph. el. 1, 164a20–22, quoting from [3, p. 278]. 76 Soph. el. 11, 171b21–25, quoting from [3, p. 291]. 73 Soph.

4 Plato’s Dialogues: Dialectic, Orality and Character

85

One can bring together these remarks and note that Aristotle devised a tripartite classification of arguments, each involving their own form of inference: didactic arguments proceed with demonstrations, dialectical or examination arguments with dialectical inferences, while eristic or contentious arguments make use of fallacies. Alcinous’ classification of arguments and Plato’s dialogues in his Handbook, 6.4 mirrors this conclusion: Plato uses demonstrative syllogisms in his expository dialogues, syllogisms based on widely held opinions when dealing with sophists and young people, and eristic ones when dealing with those properly called eristic, such as Euthydemus, for example, or Hippias.77

Of course, attributing to Plato the use of demonstrative syllogisms is controversial. Still, it is interesting to see Alcinous linking the use of particular types of arguments—in fact the same as Aristotle’s tripartite classification—with particular types of Platonic dialogues: ‘expository’, ‘dealing with sophists and young people’ and ‘dealing with eristic’. This classification appears very much, in turn, as one between: for exposition − for inquiry − eristic The introduction of ‘eristic’ as a type of arguments as well as of inferences raises some difficulties. Plato is always keen to distinguish between ‘dialectic’ and ‘eristic’, and it is commonly assumed that they are of a different nature. It is of course rather difficult to know where to draw the line: he insisted on distinguishing them in Rep. 454a, but could only recommend that one should practice dialectic at the age of 35, after five years of training, because youngsters would misuse it, turning it into “a kind of game of contradiction” (Rep. 539b).78 This already implies that there is no formal difference between dialectic and eristic, as the difference is merely an urge to win at all costs in younger practitioners. If, however, ‘dialectic’ is a game defined by the set of rules R1–R7 and if ‘eristic’ is meant to be the name of an alternative regimented form of debate, then it would have followed a different set of rules. But there is no shred of evidence for this, no mention whatsoever of an alternative rule in Ancient philosophy. Therefore, if ‘eristic’ is to proceed according to the rules of ‘dialectic’, then the difference we are seeking must lie elsewhere. Aristotle finds it in the use of fallacies: eristic inferences turn out to be “apparently dialectical” (ϕαιν´oμενoς διαλεκτικ`oς).79 Strictly speaking, there is no inference, so that we have in the end only really two types of inferences: from true premises (with valid inferences, thus sound arguments) or from reputable opinions. Aristotle also sees reasonings as ‘contentually’ fallacious, if they involve what are ‘reputable opinions’ only in appearance (see his criticisms of Zeno below). It is not easy without training to tell a sound from an unsound dialectical argument, and a specific rule banning fallacies seems not to have been formulated, hence Plato’s 77 Quoting

from [17, p. 11]. from [12, p. 1154]. Plato also describes them as ‘playing at contradiction for sports’ (Rep. 539c). 79 Soph. el. 9, 171b11. 78 Quoting

86

M. Marion

and Aristotle’s struggles with the issue. But this issue cuts deeper than a simple case of classification according to use or not of fallacies, because at bottom it points to a fundamental conceptual problem: dialectic and ‘eristic’ both proceed according to R1–R7, so how to draw a clear distinction? The case of Zeno might be interesting in this respect. As we saw in Sect. 4.4, his paradoxes of motion are in fact dialectical arguments about plurality, leading to the absurd conclusion that there is no movement. If one is to give credence to Parm. 128d–e, Zeno’s aim was purely contentious, namely deriving contradictions from the views of Parmenides’ adversaries. Plato even has Zeno distancing himself from his younger combative self, as he claims that the contentious character of his treatise is to be explained by his young age. Still, Plato also portrayed ambiguously Zeno as straddling the distinction between ‘dialectician’ and ‘eristic’ when he called him an ‘Eleatic Palamedes’ in Phdr. 261d—the figure of Palamedes implies that Zeno had bequeathed to humanity something important, presumably dialectic—and when he implied that he was both a philosopher and ‘god of refutation’ in Soph. 216a–b, while Theodorus calms Socrates saying the Eleatic visitor is “more moderate than the enthusiasts for debating are” (Soph. 216b). While Plato remained arguably ambiguous, Aristotle simply ranked Zeno’s arguments as fallacious arguments “that deduce or appear to deduce to a conclusion from premises that appear to be reputable but are not so”. His critique of Zeno’s arguments on motion relies precisely on this point, since he argued that Zeno assumed the false and merely appearing to be reputable premise that it is impossible “to come in contact with illimitable things in a limited time” (Phys. Z 233a21–31). Still, the distinction between ‘dialectic’ and ‘eristic’ cannot be said to rely only on the use or not of ‘apparently dialectical’ inferences or ‘apparently reputable’ opinions. This would mean that there could never be eristic uses of dialectic without any such cheating. But if it is the case that there are eristic uses of dialectic without cheating, how would these be distinguishable from ordinary ‘elenchi’ according to R1–R7? The difficulty is not solved by appealing to the idea that, dialectic being a common endeavour, somehow Questioner would be harsher towards Answerer. In Theaetetus Socrates makes it very plain that he is going to test Theaetetus without using tricks, but that does not mean that he is going to go easy on him, only that he is not going to trip him up (with fallacies). In other words, Socrates promises only to be fair.80 There is, however, a further reason for the need to be as uncompromising as possible without cheating: it is a necessary condition to generate the best arguments, after all this is what Questioner and Answerer ought to aim for. The analogy with chess might be useful here. Indeed, if a grand master trains a novice, it will not render the latter a great service if the grand master were intentionally to play badly to let the novice win. Likewise, if we want to look at endgames and figure out what possibilities are left for a given configuration, it will be of no use not to try one’s best ´ 81 that is Socrates’ to avoid checkmate. Plato’s imagery of ‘maieutic’ (μαιευτικη), ‘art of midwifery’ as described in the same dialogue at Tht. 150a–151d & 210b–d, 80 Tht. 167e–168a. As he puts it: “It is the height of unreasonableness that a person who professes to care for moral goodness should be consistently unjust in discussion” (quoting from [12, p. 186]). 81 Tht. 150a–151d.

4 Plato’s Dialogues: Dialectic, Orality and Character

87

conveys the same idea in its own way: as midwife Socrates is helping Theaetetus by leading him to see his contradictions but, as with giving birth, this is not going to be painless. There must therefore be sense in which Questioner must proceed without recourse to fallacy but also without mercy against Answerer, while dialectic is pursued with positive purposes, as with ‘maieutic’. But here the purpose of drawing out contradictions appears not to be essentially negative, as it would be in a public contest were the aim is to demolish your opponent’s views. So, while recognizing the essentially contentious character of dialectic, one must distinguish ‘positive’ uses of dialectic ‘for inquiry’,82 from the merely ‘negative’ task of demolishing the adversary’s standpoint, to which the word ‘eristic’ should be attached. These will be our two basic types. As already pointed out, none of this directly concerns the classification of Plato’s dialogues. But we cannot be that far, however, since classifying them in terms of character is to classify them in terms of their overall aim—answering questions such as ‘Was Plato aiming to expound his ideas or refute that of others?’. But the classifications we have involved types of arguments, as we shall see. With these preliminary remarks in mind, the classification of dialogues in terms of character in Diogenes Laertius’ Vitae III 49 can be schematized as follows: dialogue

for instruction theoretical

physical logical

practical

ethical political

for inquiry for training

maieutic

peirastic

aimed at victory

endeictic

anatreptic

The first tier is the all-important distinction between ‘hyphegetic’ and ‘zetetic ` Ùϕηγητικ´oς and χαρακτηρ ` ζητητικ´oς), thus between character’ (χαρακτηρ dialogues ‘for instruction’ and ‘for inquiry’. There is a similar classification in Albinus’ Prologue, Chapter 3, where Diogenes’ second tier has disappeared:

82 This

is the spirit that has been described by Christopher Gill as that of a “shared search” in [23, p. 284], in the sense that it is an ongoing collaborative activity. Gill is, however, of the opinion that, as ‘shared search’, dialectic is also directed at an aim, which he describes as “achieving knowledge of truth”, a view rejected above in Sect. 4.3. Dialectic can still be seen as primarily as ‘shared search’— what is called later in this section the ‘zetetic’ or ‘peirastic’ mode—without aiming automatically at truth as such.

88

M. Marion

dialogue

for instruction

for inquiry

physical logical political ethical

maieutic peirastic elenctic anatreptic

Albinus’ classification is slightly revised in his Chapter 6: dialogue

for instruction physical ethical political economic

for inquiry maieutic

peirastic logical anatreptic elenctic

I skip over the numerous differences between these classifications and recall their very point, which was to help devising a sequence for reading Plato’s dialogues. One could thus append to any of these classifications a scheme, such as Albinus’ five stages of Platonic education83 : i. ii.

iii. iv. v.

False opinions must be purged, with help of dialogues of ‘peirastic’ or ‘elenctic’ character. Notions that are deemed ‘recollectable’ in the sense of Meno or Phaedo are awakened and clarified, with help, precisely, of dialogues of ‘maieutic’ character. Thus prepared, the soul can receive proper instruction with help of ‘hyphegetic’ dialogues concerning moral and physical doctrines. The doctrines thus taught must be bound fast in the soul, with help of the ‘logical’ dialogues. Finally, one must learn to counter arguments dragging us away from this direction, with ‘endeictic’ or ‘anatreptic’ dialogues.

The point of classifying Plato’s dialogues in term of characters becomes obvious with schemes such as this, but it also becomes obvious why whole dialogues would be assigned a unique specific character in spite of the fact that these characters are at bottom about types of arguments (as I have hinted at, and more on this below) and despite the fact already noted that most Platonic dialogues, like ordinary dialogues, exhibit ‘dialectical shift’. It goes without saying that my aim here is a different one: to understand the variety of ways arguments devised in the oral practice of dialectic can be inserted in a Platonic dialogue. Thus, I suggest that one sets aside this problematic aspect and merely focus on the characters, ‘didactic’, ‘peirastic’, ‘maieutic’, ‘elenctic’, ‘endeictic’, ‘anatreptic’, ‘eristic’, etc. from a point of view

83 See

[64, pp. 40–41].

4 Plato’s Dialogues: Dialectic, Orality and Character

89

closer to Argumentation Theory, and with the aim of extracting from them what they tell us about the relation between orality and textuality. The classification reported by Diogenes proceeds in the manner of Platonic diaresis, and all three incorporate elements foreign to Plato about the divisions of philosophy. Dropping these and noticing the similarities, I would suggest further that we look at the following very simplified classification:

dialogue

for instruction

for inquiry dialectical

eristic

It is of little importance here how the content ‘for instruction’ is to be divided up. What seems crucial is that Plato’s dialogues are broadly divided between those that aim at exposition or instruction, and those that aim at inquiry. It should be uncontroversial to grant that Plato’s dialogues are not merely composed of instances of dialectic, even though it is pervasive, and that when there are instances of perfectly plain ordinary dialogues, that these passages may have expository or instructional purposes. That dialectic was a verbal practice means that even though some practitioners, such as Socrates, Arcesilaus or Carneades refused to write, those who chose to do so, even choosing to write dialogues as Plato did, were under no silly obligation merely to spend their time reporting dialectical bouts and never trying to write down their own positive views in, say, an expository fashion. Aristotle has passages where he contrasts dialecticians from philosophers, indicating a growing awareness of a division of labour between those who produce arguments and those who assessed them.84 One may wonder why there is no mention in these classifications of ‘protreptic’ (from ‘πρoτρεπτικ´oς’ or ‘hortatory’).85 According to Simon Slings, a text is said to be protreptic “if its design is to cause a change in behaviour of those for whom it is destined, or if within the text one character endeavours to cause such a change in another character or characters”.86 Exhortations to virtue and wisdom were common in Greek culture and the reason for the omission of this potential ‘character’ seem to be that, although Socrates occasionally engaged in explicit protreptic,87 exhortation was 84 Again,

see [49].

85 See, for example, mentions of ‘πρoτεπτικη ´ σoϕια’ and ‘πρoτρεπτικîν λ`oγoν’ at, resp., Euthy.

278c and 282d. 86 [62, p. 59]. Slings also distinguishes a stricter sense “applied to texts which are intended to impel readers to pursue a certain study” such as philosophy [62, p. 60], which is not relevant in the context of this paper. For the distinction between ‘explicit’ and ‘implicit’ protreptic, see [62, pp. 61–62]. 87 See for example Socrates’ protreptic speeches at Ap. 29d–e; 30a–b; 36c–d and Euthy. 278e–282d; 288d–292e or Cleit. 407b–408e of doubtful authenticity.

90

M. Marion

commonly associated with precisely the sort of ‘epideictic’ speeches of the Sophists at Hp. mai. 282b, that is the long speeches to which Plato explicitly opposed his uses of dialectic. He obviously favoured the dialogue form over hortatory speech or, as we saw, ‘brachylogy’ over ‘macrology’. This does not mean, however, that there is no ‘protreptic’ in Plato’s dialogues, outside the explicit exhortations just mentioned. There is after all another Ancient tradition, from Demetrius of Phalorum in De Elocutione to Cicero’s Academica Posteriora I. iv. 16, who is following closely Antiochus of Ascalon, that recognizes dialectic and the ‘elenchus’ in Plato’s dialogues as what one might call, with Slings, ‘implicit protreptic’.88 Although the Clitophon is probably not genuine, its main argument is an interesting condemnation of explicit protreptic in favour of the ‘elenchus’.89 Demetrius distinguishes in Eloc. 296–297 between three types of protreptic: ‘accusation’—and here one may think of Socrates’ exhortation at Ap. 29d–e—‘advice’ and the ‘Socratic type’, which is achieved through dialectic and the ´ proceeds dialectically without ‘elenchus’. Assuming that ‘maieutic’ (μαιευτικη) specifically saying that it does, we can see that the result of the ‘elenchus’ is a change of character in the person subjected to it. Socrates says indeed to Theaetetus that “you will be modest and not think you know what you don’t know” (210c),90 while in the Sophist the effect is described as follows: The people who are being examined see this, get angry with themselves, and become calmer toward others. They lose their inflated and rigid beliefs about themselves that way, and no loss is pleasanter to hear or has a more lasting effect on them. (230b–c)91

We can thus see μαιευτικη´ as use of dialectic to achieve implicit protreptic. This points to the centrality of ‘elenctic’ uses of dialectic in Plato. To recapitulate, we have to assume that we speak only of various uses of dialectic as it proceeds according to R1–R7 above. I have excluded possible ‘hyphegetic’ uses to concentrate on the ‘zetetic’ side, where we have our two basic types: ‘dialectic’ and ‘eristic’. These simply reflect the fact that inquiry could naturally be of two sorts: a ‘positive’ one, where there is a collaborative endeavour to explore arguments on opposite sides of an issue, in order to find out which possesses the stronger grounds, and a ‘negative’ sort, where one’s opinions are tested for consistency or one merely tries to demolish, in eristic fashion, the opposite view, without any other ulterior motive. The distinction between ‘dialectic’ and ‘eristic’ also overlaps that between training students within the school, and participating in adversarial bouts in public places, but does not reduce to it. The term ‘eristic’ thus covers purely adversarial uses of dialectic according to R1–R7, aiming merely to demolish the adversary’s position in “competitions and 88 The

evidence is marshalled in [62, pp. 83–89]. [62, p. 210]. Slings assumes that the dialogue is genuine, but this is very much a minority point of view. At all events, the point remains of interest even if the dialogue isn’t genuine. 90 Quoting from [12, p. 234]. 91 Quoting from [12, p. 251]. One recognizes here the theme of ‘catharsis’ (καθαρσις): ´ the cleansing of false pretence in Soph. 226b–231e, which is brought about via ‘elenctic’ uses of dialectic in Plato’s dialogues. 89 See

4 Plato’s Dialogues: Dialectic, Orality and Character

91

contests” as Aristotle puts it in Soph. el. 2, 165b12. Both Plato and Aristotle complain about the use of fallacies in eristic, but it is not a necessary feature: one could win in such a contest without cheating. In such contests, eristic serves essentially to demolish the Answerer’s positions—one should think here of Zeno’s arguments. Such uses of R1–R7 may go under the subtype ‘anatreptic’. Eristic also includes the subtype ‘endeictic’ to cover the more specific uses of uncovering Answerer as a fraud of some sort, and it is no surprise that, to take only one example, the character ‘endeictic’ is attached to Protagoras by Diogenes Laertius.92 Hippias, mentioned above, is also in for this sort of embarrassing treatment. There is a fundamental ambivalence, however, that arises from the fact that Greeks (and Athenians in particular) appear to have delighted in sheer display of quickness of mind in bouts—see Cleinais’ complaint in the Mytilenean debate (Thuc. III, 38) or Isocrates’ complaints about useless contests in Hel. 1–7, Antid. 269 and Panath. 26—but dialectic has more on offer than a mere demolition job in such contests: ‘consistency maintenance’ is an important test, and there is a positive lesson when one’s positions fail or not to pass it. This is the primary use of dialectic for philosophers, who argued ‘on both sides’ for that reason. We saw how Plato and Aristotle struggled with this ambivalence, and one should also recall that Plato portrayed Sophists as ‘eristic’—merely demolishing Answerer’s viewpoint, including with use of fallacies and that, while he was at pains to demarcate Socrates’ uses of R1–R7 from theirs, Socrates was also presented as a Sophist by Isocrates93 and Aeschines,94 not to mention Aristophanes’ satirical portrayal in Clouds 874sq.—this too reflects this fundamental ambivalence. What goes under the other basic type ‘dialectic’ is uses of R1–R7 with more positive goals in mind. Of its subtypes, ‘maieutic’ is closely related to, but distinct from ‘endeictic’. As opposed to ‘endeictic’ uses, which take place in a purely hostile context, Answerer is here willing to submit to the test—in the terminology of ‘midwifery’ in Theaetetus, Answerer is described as “pregnant and in labor” (151b– c)—and undergoes an improvement from the resulting ‘elenchi’. As we saw, arguing on both sides of a question might have an exploratory purpose, when participants have no idea which is the better option, either A or ¬ A, and no initial bias. Uses of R1–R7 in this spirit would correspond to the subtype ‘peirastic’. The second half of Parmenides is arguably a perfect example of such ‘peirastic’ uses of dialectic, laying out the arguments ‘on both sides’.

4.5 Argumentation Theory I would like to conclude with some brief comments comparing the classification of Plato’s dialogues in terms of their ‘character’ with types of (ordinary) dialogues 92 Vitae

III, 51. Hel. 1, where both Socrates and Plato are implicitly grouped with Sophists. 94 Timarch. 173. 93 In

92

M. Marion

in contemporary Argumentation Theory. Krabbe and van Laar have drawn comparisons with Aristotle’s Topics and Sophistical Refutations,95 but no one has done so with ‘characters’. Walton and Krabbe recognized the following types of dialogues: ‘persuasion’, ‘negotiation’, ‘inquiry’, ‘deliberation’, ‘information-seeking’, ‘eristic’ and ‘mixed’.96 In devising a description for the type ‘eristic’, they were heavily ´ 97 and mixed influenced by Plato’s Euthydemus and Aristotle on ‘eristic’ (™ριστικη), dialogues are said to have subtypes, two of which are also defined with reference to dialectic: ‘debate’, which is described as a mixture of ‘eristic’ and ‘persuasion’,98 and ‘Socratic dialogue’, which involves ‘inquiry’ and ‘persuasion’, but with the ‘elenchus’ being seen as aiming at “cleaning the spirit” and “moral improvement”.99 This is directly related to the ‘protreptic’ nature of ‘catharsis’ and ‘maieutic’ just discussed. If there is an overlap concerning ‘eristic’ for the obvious reason that Ancient Greeks are the inspiration, it is not so clear where dialectic in its more positive uses should fit or even if Argumentation Theory can shed light on these uses. Dialectic does not seem to fit under the type ‘inquiry’ for a number of reasons. Surely, in both cases one could argue that the goal involves a reduction of the state of ignorance on a given topic. But in dialectic a single bout never gets rid of ignorance. In Argumentation Theory, the type ‘inquiry’ is also closely modelled on modern scientific inquiry and Walton considers its key property to be ‘cumulativeness’: once it is recognized as true, a particular proposition remains true at all later stages.100 This characteristic appears to be at odds with the open-ended nature of dialectic bouts, where assumptions are repeatedly challenged. Each new challenge will generate a new scoreboard, potentially revealing a hitherto unsuspected contradiction, so that defeasibility is more characteristic than cumulativeness. Worse, if dialectic can arguably be seen as helping one to gain knowledge of, say, the ‘good’, it can also be argued that it is not propositional.101 Furthermore, there seem to be no room for ‘peirastic’ exploration of arguments ‘on both sides’, which is a key characteristic of uses of dialectical bouts for inquiry. Finally, if the goal of ‘inquiry’ is to find a proof or destroy one,102 then the regimented dialogues of dialectic do not fit in, even though one may have in mind the case of the slave in Meno 82a–85e. It is interesting to note here that this last passage was construed by Jaakko Hintikka as an illustration of his own interrogative games of inquiry.103 As it turns out these are the inspiration behind 95 See

[36], as well as [35]. the sake of brevity, I limit myself to the classification at [72, p. 66]. See also the variant at [69, p. 31]. I leave aside altogether considerations of ‘argumentation schemes’ in [73]. 97 See [72, pp. 181–182] or [69, pp. 78–79]. 98 [72, pp. 83–84]. 99 [72, p. 85]. Walton and Krabbe rely essentially here on [56], as is Walton in [71, pp. 38–39]. 100 [69, p. 70]. 101 Arguments to that effect are marshalled in [25, Chapter 8]. 102 [72, p. 66]. 103 See [30, pp. 20–28]. This passage may be read as giving support to the idea that one of the goals of dialectic is obtaining knowledge. Hintikka realized late in his life that ‘knowledge’ is not the 96 For

4 Plato’s Dialogues: Dialectic, Orality and Character

93

the ‘information-seeking’ dialogues,104 as being those where one of the participants possesses information that the other seeks to obtain through a series of questions and answers.105 Now, one should add to our original list a new type of dialogue, ‘examination’, either as a new type to the list or as a mixed type.106 According to Walton, who develops the latter option, ‘examination’ dialogues have two related goals, with a shift from one to the other: extracting information and testing its reliability.107 The first goal would thus bring ‘examination’ under the type ‘information-seeking’, the other indicates that there is a mixture. Dialectic bouts as ‘consistency maintenance’ tests seem at first sight naturally to fit here: part of Questioner’s task is to gather information from Answerer, and part of it is to test the information thus obtained for consistency. Eristic is not far, given that testing the reliability of the information is also a test on the reliability of the informant, who can thus be exposed as a fraud. One would not want, however, to ignore altogether the ‘protreptic’ nature of dialectic. We saw the suggestion that ‘Socratic dialogue’ is a mixed form of dialogue that involves ‘inquiry’ and ‘persuasion’. For reasons just laid out, it seems to me wrong to involve the type ‘inquiry’ as defined in Argumentation Theory in the description of dialectic. On the other hand, to achieve its protreptic aim—to effect a change of character, to convince one to change their behaviour—dialectic must persuade, so ‘persuasion’ is at the heart of the matter. After all, Socrates says of himself that “I go around doing nothing but persuading” (Ap. 30a). Thus, it may be that the best characterization of positive uses of dialectic—‘maieutic’ in particular—is as a type of mixed dialogue involving ‘persuasion’ and ‘examination’ (thus involving both ‘information-seeking’ and ‘consistency maintenance’). Finally, there is an instance of ‘deliberation’ in Plato, when Socrates debates with Crito in the eponymous dialogue about reasons why, having been condemned to death, he should escape or why he should not. It can be argued that Socrates’ arguments are dialectical in the above sense, since his own arguments and those of the personified Laws of Athens (50a–54d) are both aimed ad hominem at Crito, and do not reflect any commitment on Socrates’ part, in compliance with the Socratic Rule (R5).108 This would therefore be a very good example of a Platonic dialogue where dialectic arguments are embedded in an ordinary dialogue, the aim of which remains ‘zetetic’: aiming at a decision on the side that seems after examination best supported—this being the wish expressed by Socrates in 46b. operative word in information-seeking games or dialogues: the key concept is ‘information’. See [29, pp. 24–28]. One could thus argue that this holds mutatis mutandis for dialectic too. 104 See Walton on ‘information-seeking dialogues’ [69, p. 127]. 105 Hintikka drew links between his own interrogative games of inquiry and dialectic, even up to his last book [29]. 106 Respectively [20] and [70]. 107 [70, p. 746]. See also [20, p. 1560]. 108 Reading here Crito more in line with [63], as Stokes sees continuity in the fact that the Laws of Athens argue against Crito as dialectically as Socrates does (thus without commitment to any thesis), as opposed to more orthodox readings in, say [28] and [74].

94

M. Marion

This paper began with endeavours by Lloyd and Netz to understand the rise of scientific thinking—demonstration in particular—over the background of open debates in Ancient Greek society. One interesting aspect of this inquiry is the passage from oral to written arguments. We saw that, as far as dialectic is concerned, abstracting the role of the second person led to essential distortion of the arguments, leading to potential misunderstanding of their point and meaning. In the case of Plato, the second person is preserved, but we have a different problem: that of delimiting the use of dialectic in his dialogues, understanding how he put it to a variety of uses, that the ‘characters’ of the early tradition were, as I claimed, meant to capture. Efforts in clarifying both the nature of dialectic—including with respect to Argumentation Theory—and Plato’s uses of it should eventually help improving our understanding of specific arguments, but at least for now one can see how to avoid detrimental mistakes such as those of Robinson and Vlastos pointed out above. One could object to dialectic as defined by rules R1–R7, that its artificiality makes it unsuitable for Plato’s endeavours. The simple fact that Plato’s dialogues or parts of them exhibit such a variety of uses of dialectic proves the contrary. It shows not that Plato kept re-inventing dialectic but that, anxious as he was to keep reflecting about it throughout his œuvre (and not just up to a certain point), he was able to find more than one way to incorporate arguments of dialectic origin in his dialogues, and show their relevance to a number of different aims in ordinary dialogue. Another objection one often hears about dialectic, especially in its Eleatic variant, is that its conclusions are purely negative, so the more mystical minded will seek to overcome the limitations of this form of ‘rationalism’, even claiming this overcoming for Plato himself, with an interpretation of well-known passages on knowledge of the ‘unhypothetical’ in Rep. 511b–c, backed up by further passages such as a well known one from the Seventh Letter, where Plato speaks of knowledge coming into the soul after a long practice of dialectic, as suddenly “as a kindled light” (341c). These passages certainly need careful interpretation, but Plato’s dialogues, taken collectively, show how far this attitude is misguided. When one realises how pervasively these dialogues were built upon the fruits of dialectical bouts, not in order to prove in a negative fashion the limits of dialogue and rationality—as if it could issue only in futile contradictions— but as they integrate arguments originating in these bouts in a variety of dialogue situations and aims, then the richness of Plato’s text comes to the fore, as opposed to the simplified picture suggested by this highly reductive approach.109

References 1. Allen, R. E. (1983). Plato’s Parmenides: Translation and analysis. University of Minnesota Press. 109 I would like to thank Benoît Castelnérac (Université de Sherbrooke), for help with Ancient Greek,

but above all for our numerous discussions on the topic of this paper and extensive comments on earlier versions of it, as it forms part of our collaborative work.

4 Plato’s Dialogues: Dialectic, Orality and Character

95

2. Alline, H. (1915). Histoire du texte de Platon. Honoré Champion. 3. Barnes, J. (Ed.). (1984). The complete works of Aristotle: The revised Oxford translation. Princeton University Press. 4. Blondell, R. (2002). The play of character in Plato’s dialogues. Cambridge University Press. 5. Brandwood, L. (1990). The chronology of Plato’s dialogues. Cambridge University Press. 6. Brisson, L. (1993). Les classifications des dialogues chez Diogène Laërce: enjeux interprétatifs. In A. Balansard & I. Koch (Eds.), Lire les dialogues, mais lesquels et dans quel ordre? (pp. 43– 57). Akademia Verlag. 7. Castelnérac, B. (2013). Gorgias et les joutes dialectiques: l’argument ontologique dans le traité Sur le non-être de Gorgias (première partie). Phoenix, 67(3–4), 263–283. 8. Castelnérac, B. (2014). Gorgias et les joutes dialectiques: l’argument ontologique dans le traité Sur le non-être de Gorgias (deuxième partie). Phoenix, 68(1–2), 78–94. 9. Castelnérac, B. (2015). Note exégétique sur le fragment 2 de Parménide (DK 28 B2). Revue des études grecques, 128, 291–308. 10. Castelnérac, B., & Marion, M. (2009). Arguing for inconsistency: Dialectical games in the academy. In G. Primiero & S. Rahman (Eds.), Acts of knowledge: History, philosophy and logic (pp. 37–76). College Publication. 11. Castelnérac, B., & Marion, M. (2013). Antilogic. Baltic International Yearbook of Cognition, Logic and Communication, 8. http://newprairiepress.org/biyclc/vol8/iss1/3/. 12. Cooper, J. M. (Ed.). (1997). Plato: Complete works. Hackett. 13. Cornford, F. M. (1939). Plato and Parmenides: Parmenides’ way of truth and Plato’s Parmenides. Routledge & Kegan Paul. 14. Crubellier, M. (2011). Du sullogismos au syllogisme. Revue philosophique, 1, 17–36. 15. Crubellier, M., Marion, M., McConaughey, Z., & Rahman, S. Dialectic, the Dictum de Omni and Ecthesis. History and Philosophy of Logic, 40(3), 207-233. 16. Diès, A. (1927). Autour de Platon. Essais de critique et d’histoire. Beauchesne. 17. Dillon, J. (Ed.). (1993). Alcinous: The handbook of Platonism. Clarendon Press. 18. Dixsaut, M. (2001). Métamorphoses de la dialectique dans les dialogues de Platon. Vrin. 19. Dodds, E. R. (1928). The Parmenides of Plato and the origin of the neoplatonic “one”. Classical Quarterly, 22, 129–142. 20. Dunne, P. E., Doutre, S., & Bench-Capon, T. J. M. (2005). Discovering inconsistency through examination dialogues. In Proceedings IJCAI-05 (pp. 1560–1561). Edinburgh. 21. Dutilh Novaes, C. (2005). Medieval obligationes as logical games of consistency maintenance. Synthese, 145, 371–395. 22. Dutilh Novaes, C. (2015). A dialogical, multi-agent account of the normativity of logic. Dialectica, 69, 587–609. 23. Gill, C. (1996). Afterword: Dialectic and the dialogue form in the late Plato. In C. Gill & M. M. McCabe (Eds.), Form and argument in late Plato (pp. 283–311). Clarendon Press. 24. Gill, M.-L., & Ryan, P. (Trans.). (1996). Plato: Parmenides. Hackett. 25. Gonzalez, F. J. (1998). Dialectic and dialogue: Plato’s practice of philosophical inquiry. Northwestern University Press. 26. Griswold, C. L. (Ed.). (1988). Platonic writings, Platonic readings. Routledge. 27. Hamblin, C. L. (1970). Fallacies. Methuen. 28. Harte, V. (1999). Conflicting values in Plato’s Crito. Archiv für Geschichte der Philosophie, 81, 117–147. 29. Hintikka, J. (2007). Socratic epistemology: Explorations of knowledge-seeking by questioning. Cambridge University Press. 30. Hintikka, J., & Bachman, J. (1991). What if…? Towards excellence in reasoning. Mayfield. 31. Jones, W. H. S. (Trans.). (1947). Hippocrates, volume IV. Harvard University Press. 32. Kahn, C. (1996). Plato and the Socratic dialogue: The philosophical use of a literary form. Cambridge University Press. 33. Keiff, L., & Rahman, S. (2010). La dialectique, entre logique et rhétorique. Revue de métaphysique et de morale, 115, 149–178.

96

M. Marion

34. Klagge, J., & N. D. Smith (Eds.). (1992). Methods of interpreting Plato and his dialogue. Oxford Studies in Ancient Philosophy, supplementary volume. 35. Krabbe, E. C. W. (2002). Meeting in the house of Callias: An historical perspective on rhetoric and dialectic. In F. H. van Eemeren & P. Houtlosser (Eds.), Dialectic and rhetoric (pp. 29–40). Kluwer. 36. Krabbe, E. C. W., & van Laar, J. A. (2007). About old and new dialectic: Dialogues, fallacies and strategies. Informal Logic, 27(1), 27–58. 37. Laks, A., & Most, G. W. (Ed. & Trans.). (2016). Early Greek philosophy: Sophists: Part 1. Harvard University Press. 38. Ledger, G. R. (1989). Re-counting Plato: A computer analysis of Plato’s style. Oxford University Press. 39. Lee, H. D. P. (Ed. & Trans.). (1936). Zeno of Elea: A text with translations and notes. Cambridge University Press. 40. Lewis, D. (1979). Scorekeeping in a language-game. Journal of Philosophical Logic, 8, 339– 359. 41. Lloyd, G. E. R. (1970). Early Greek science: Thales to Aristotle. Norton. 42. Lloyd, G. E. R. (1979). Magic, reason and experience. Cambridge University Press. 43. Lloyd, G. E. R. (1990). Demystifying mentalities. Cambridge University Press. 44. Lloyd, G. E. R. (2014). The ideals of inquiry: An ancient history. Oxford University Press. 45. Mansfeld, J. (1994). Prolegomena: Questions to be settled before the study of an author or a text. E. J. Brill. 46. Marion, M. (2009). Why play logical games? In O. Majer, A.-V. Pietarinen, & T. Tulenheimo (Eds.), Logic and games, foundational perspectives (pp. 3–26). Springer. 47. Marion, M. (2014). Les arguments de Zénon selon le Parménide de Platon. Dialogue, 53, 393–434. 48. Marion, M., & Rückert, H. (2016). Aristotle on universal quantification: A study from the perspective of game semantics. History and Philosophy of Logic, 37, 201–229. 49. Merry, D. (2016). The philosopher and the dialectician in Aristotle’s topics. History and Philosophy of Logic, 37, 78–100. 50. Nails, D. (1995). Agora, academy, and the conduct of philosophy. Kluwer. 51. Netz, R. (1999). The shaping of deduction in Greek Mathematics: A study in cognitive history. Cambridge University Press. 52. Nehamas, A. (1990). Eristic, antilogic, sophistic, dialectic: Plato’s demarcation of philosophy from sophistry. History of Philosophy Quarterly, 7, 3–16. 53. Nightingale, A. (1993). Genres in dialogue: Plato and the construction of philosophy. Cambridge University Press. 54. Nüsser, O. (1991). Albins Prolog und die Dialogtheorie des Platonismus. Teubner. 55. Press, G. (Ed.). (1999). Who speaks for Plato? Studies in Platonic anonymity. Rowman & Littlefield. 56. Robinson, R. (1953). Plato’s earlier dialectic. Clarendon Press. 57. Ross, W. D. (1949). Aristotle. Methuen. 58. Rutherford, R. B. (1995). The art of Plato: Ten essays in Platonic interpretation. Harvard University Press. 59. Ryle, G. (1966). Plato’s progress. Cambridge University Press. 60. Sidgwick, H. (1872/1874). The Sophists. Journal of Philology, 4, 288–307 & 6, 66–80. 61. Skinner, Q. (2001). The rise of, challenge to and prospects for a Collingwoodian approach to the history of political thought. In D. Castiglione & I. Hampshire-Monk (Eds.), The history of political thought in national context (pp. 175–188). Cambridge University Press. 62. Slings, S. R. (Ed. & Trans.). (1999). Plato: Clitophon. Cambridge University Press. 63. Stokes, M. C. (2005). Dialectic in action, an examination of Plato’s Crito. The Classical Press of Wales. 64. Tarrant, H. (1993). Thrasyllan Platonism. Cornell University Press. 65. Thesleff, H. (1989). Platonic chronology. Phronesis, 34, 1–26. 66. Vlastos, G. (1991). Socrates: Ironist and moral philosopher. Cornell University Press.

4 Plato’s Dialogues: Dialectic, Orality and Character

97

67. Vlastos, G. (1994). Socratic studies. Cambridge University Press. 68. Vuillemin, J. (1984). Nécessité ou contingence. L’aporie de Diodore et les systèmes philosophiques. Éditions de Minuit. 69. Walton, D. N. (1998). The new dialectic: Conversational contexts of argument. University of Toronto Press. 70. Walton, D. N. (2006). Examination dialogue: An argumentation framework for critically questioning an expert opinion. Journal of Pragmatics, 38, 745–777. 71. Walton, D. N. (2013). Methods of argumentation. Cambridge University Press. 72. Walton, D. N., & Krabbe, E. C. W. (1995). Commitment in dialogue: Basic concepts in interpersonal reasoning. SUNY Press. 73. Walton, D. N., Reed, C., & Macagno, F. (2008). Argumentation schemes. Cambridge University Press. 74. Weiss, R. (1998). Socrates dissatisfied: An analysis of Plato’s Crito. Oxford University Press. 75. Whitaker, C. W. A. (1996). Aristotle’s De Interpretatione: Contradiction and dialectic. Clarendon Press. 76. Whittaker, J. (Ed. & Trans.). (1990). Alcinoos, Enseignement des doctrines de Platon. BellesLettres.

Chapter 5

What is the Difference Between the Dialectical and Rhetorical Use of Arguments? A Paradigm-Based Approach to Plato’s Socrates Christopher Roser Abstract Plato sharply contrasts Socrates’ dialectical manner of argumentation to the rhetorical way of using arguments. The upshot of this contrast seems to be clear: the dialectical use of arguments is epistemically valuable, while the rhetorical use is epistemically deficient. What is less clear, however, is how Plato justifies this sharp contrast. In this paper, I consider seven different ways scholarship has tried to answer this question. I argue that none of them manage to explain Plato’s clear-cut distinction. This has two important consequences. First, it shows that the distinction between the rhetorical and dialectical use of arguments is considerably more delicate and multilayered than is commonly thought. Second, it also shows that the commonly accepted definitions of dialectic fail to sufficiently distinguish dialectic from rhetoric. That does not mean, however, that there is no essential difference between them. Rather, I suggest, Plato distinguishes rhetoric from dialectic by following a paradigm-based approach: he points to paradigmatic differences between them and, in so doing, he reveals what is essential about each of them.

5.1 Introduction What is the difference between the way language (logos) is used in Socrates’ dialectical conversations and the way rhetoricians use it in their rhetorical speech? At first sight, this might seem to be an easy and straightforward question to answer. There is, first, a clear similarity between the two approaches: Both are argumentative. It is an essential feature of each way of speaking that the speaker provides a set of premises that are supposed to stand in some inferential or justificatory relationship to a conclusion. In short, both proceed by providing reasons. Second, it is apparent that, according to Socrates, there is a clear difference between these two uses of arguments: the one is an epistemically adequate use and the other an epistemically inadequate one; or as he puts it in the Gorgias, the one is worth something with regard to the truth and the other is not. These two ways of using arguments are C. Roser (B) Department of Philosophy, Humboldt University Berlin, Berlin, Germany © Springer Nature Switzerland AG 2021 J. A. Bjelde et al. (eds.), Essays on Argumentation in Antiquity, Argumentation Library 39, https://doi.org/10.1007/978-3-030-70817-7_5

99

100

C. Roser

so different from each other that they are even linked to different forms of living: the philosophical life and the rhetorical life, respectively. Consequently, it seems reasonable to expect that any account of Socrates’ dialectical argumentation, or of what Socrates takes to be the epistemically adequate form of argumentation, should explain how the two forms of speech differ. Such an account will need to distinguish the dialectical from the rhetorical use of arguments, given that it needs to pick out Socratic argumentation from other uses of arguments. However, in this paper I argue that this question is significantly more complex than it seems at first sight. We will see that many accounts of dialectical argumentation do indeed point out differences between Socrates’ argumentation and that of the rhetoricians. The problem is that none of them individually, nor any combination of them, provide a general criterion that can be used to distinguish all the different cases of rhetorical argumentation from the cases of dialectical argumentation. This raises the question as to whether it is even possible to define dialectical argumentation, given that such a definition would need to include a condition that is sufficient to distinguish it from other kinds of argumentation, including the rhetorical. Ultimately, I propose that there may very well be no definition at all of dialectical or rhetorical argumentation. Yet even if this is the case, it does not mean that there are no significant differences between these two kinds of argumentation or between Socrates’ way of arguing and that of the rhetoricians. I argue that the main problem with the various criteria that have been proposed by scholars so far is that they fail to meet either one or both of the desiderata that we can reasonably expect from a criterion. On the one hand, they do not sharply distinguish rhetorical from dialectical argumentation. This is to say, they do not manage to distinguish all the possible cases of rhetorical argumentation from all the cases of dialectical argumentation. On the other hand, they do not provide an epistemically explanatory distinction. This is to say, the distinction they propose cannot explain why dialectical argumentation is supposed to be epistemically superior to rhetorical argumentation. The problem is, however, that we can reasonably expect a criterion of distinction to be sharp and epistemically explanatory in this way. For, both Plato and his Socrates engage in what we can call the rhetorical challenge. Both the writer and his character distinguish the Socratic and epistemically adequate kind of argumentation from the rhetorical kind that they take to be epistemically inadequate. Because of this, we may, first, expect there to be a sharp distinction between these two kinds of argumentation. It should not be the case that several cases of rhetorical argumentation are indistinguishable from dialectical argumentation and vice versa. If this were the case, either dialectical argumentation is not always epistemically adequate, or rhetorical argumentation is not always epistemically inadequate. It would also follow from such a partial indistinguishability that the two ways of living, the rhetorical and the philosophical, would not seem to be that different.1 Second, with this backround, 1 This, to be clear, does not imply that there cannot be a piece of argumentation that is both dialectical

and rhetorical. Rather, it implies that such a piece of argumentation would be rhetorical and would be dialectical on the basis of meeting different conditions or on the basis of different features that it bears. What is more, there might not be such a sharp distinction between dialectic and what

5 What is the Difference Between the Dialectical …

101

we should also expect the distinction between dialectical and rhetorical argumentation to have an epistemically explanatory value. Such a distinction should be able to play a significant role in an explanation of why it is dialectical argumentation that is of direct epistemological value, rather than rhetorical argumentation. Only if this is the case will Plato and his Socrates be able to argue in a non-ad hoc way that his dialectical argumentation is epistemically superior to rhetorical argumentation. In the following, in Sect. 5.2 I use these two desiderata as a test-case to evaluate seven common definitions or specifications of Socratic argumentation, all of which identify or imply prima facie plausible conditions for how it is different from rhetorical argumentation. I argue that while each of these interpretations can draw on strong evidential support, they do not satisfy our desiderata, either individually or in combination. What they do manage to show successfully is that there are important differences between the two kinds of argumentation. However, they do not provide a general criterion of distinction that applies to all the different cases of rhetorical and dialectical argumentation. This being the case, they also do not provide sufficient conditions for dialectical argumentation.2 I propose in Sect. 5.3 that we should use the features of dialectical argumentation identified by these interpretations as paradigmatic examples or features that point us towards what dialectic is, even if they do not define it. It is important to note that this has crucial consequences for the question of what Plato, or Socrates, considers to be an epistemically adequate use of arguments, i.e., a use of arguments that is conducive in the search for truth and knowledge. The absence in Plato of a general criterion that successfully defines dialectical argumentation suggests that Plato and Socrates also do not distinguish which arguments are epistemically adequate and which are not by reference to such a general criterion, that can be applied to all cases. Instead, they distinguish them by using a variety of paradigmatic examples and features of adequate and of inadequate argumentation. These paradigmatic examples and features are supposed to provide an understanding of what makes an argumentation adequate.

Socrates calls ‘true rhetoric’, as he describes it in Gorg. 517a, which might very well simply be an ability to argue dialectically. However, for now, we can set this special use to one side and, instead, concentrate on the ‘traditional’ rhetorical argumentation that Socrates criticizes as epistemically inadequate. 2 More precisely, they do not provide a condition or combination of conditions that apply to all cases of dialectical argumentation and to no cases of rhetorical argumentation, nor do they provide a condition or combination of conditions that apply to all cases of rhetorical argumentation, and to no cases of dialectical argumentation. Nevertheless, I do not deny that some features might be sufficient to identify an argumentation as being rhetorical and not dialectical (e.g., trying to persuade a large crowd in a short time). However, if there are such features, they are not necessary features of rhetorical argumentation.

102

C. Roser

5.2 The Problems of the Common Interpretations and the Manifold Differences Between Rhetoric and Dialectic To begin with, it will be worth restating the main similarity between rhetorical and dialectical language usage: Both use arguments. That is to say that both use a set of sentences which they present as inferentially linked. This raises the question: What is the difference in their use of arguments? This is the question which I will focus on here. To see this similarity, we can draw on a preliminary understanding of an argument as a set of claims (or other propositional or linguistic entities) that are inferentially structured, such that there are a set of premises and a conclusion. In arguments, there is one sub-set, the premises, which speaks for or against the other sub-set, the conclusion, or it is, at least, indicated or implied that this is supposed to be the case. Argumentation is, then, the exchange of such arguments. It is the linguistic activity in which one person (or a group of people) puts forward a set of claims (or other linguistic entities), directs them at other people, and suggests or implies that these claims speak for or against a different claim.3 This understanding of what an argument is enables us to identify the use and understanding of arguments in both rhetorical and dialectical uses of speech in ancient texts. While one cannot identify a single standardized word in these texts that corresponds to the English “argument”, there are several clear ways in which these authors mark their arguments and by which they express their ideas concerning what makes an argument valid, sound, and good. This is mainly done (i) by their use of inferential particles such as “for” ´ and “ε‡περ”), “then” or “thus” (e.g., “¥ρα”, “oâν”, and or “since” (e.g., “γαρ” ´ and “but”, in the sense of marking premises (especially the inferential use “δη”), of “δš”); (ii) by their use of a set of verbal expressions which can be translated as ‘something shows/demonstrates/makes clear that something is true or correct’ (e.g. “™πιδεικνναι”, ´ “¢πoδεικνναι”, ´ “δÁλoν” and “ϕανερ´oν πoε‹ν”); and (iii) by their use of certain nouns that, under some circumstances, correspond to “demonstration”, “proof”, or “argument” (e.g., “π´ιστις”, “¢π´oδεξις”, and “™π´ιδειξις”). These words are often used as meta-discursive markers, with which the speaker draws a distinction between those parts of a speech that are premises and that part which serves as the conclusion. In drawing this distinction, they suggest that the conclusion follows from the premises and that the premises speak in favor of, or provide reasons with varying degrees of strength for, the conclusion being true. In doing so, they address questions about what makes an argument valid, sound, or, more generally, good. What is crucial for us, is that this kind of language is used both in rhetorical writings and speeches and in Socratic conversations. Socrates and rhetoricians both use this kind of vocabulary to describe the inferential connection between sentences. Given these similarities, 3 Alternatively,

it can also be suggested that these claims, or other linguistic entities speak for and against certain behavior, states, or attitudes of people, such as asserting and believing a claim, or acting in accordance with it.

5 What is the Difference Between the Dialectical …

103

the question we must answer is what, precisely, distinguishes how such arguments are used or understood in rhetorical and in dialectical argumentation. In the following, I consider seven possible ways of distinguishing rhetorical from dialectical speech. The most common approaches can be grouped together into four camps. The approaches in the first camp concentrate on the aim or motivation of those who argue rhetorically or dialectically and suggest that it is in this that their difference lies. We can call this the aim approach (options 1 and 2). The approaches in the second camp, which we can call formal approaches, concentrate on the rules or formalizable features of these two kinds of argumentation and suggest that it is by reference to these that we can distinguish rhetorical and dialectical argumentation. I consider three possible distinctions (options 3–5) that fit into this camp: a formal distinction in the procedure of the argumentation, a formal distinction in the arguments used, and a distinction concerning the role of forms. The third camp concentrates on the content (option 6). The fourth, and last, camp suggests that there are differences in the context of the two kinds of argumentation. These are primarily differences concerning the settings in which the two kinds of argumentation typically happen, the political implications of the arguments, or the kinds of people who argue in one way but not the other (option 7). The arguments advanced by most scholars tend not to focus exclusively on just one of these options but a combination of them.

5.2.1 Option One: The Role of Persuasion Probably the most common understanding of rhetoric is that it is the form of speech that uses its arguments to persuade. Plato invokes such a characterization almost every time he refers to rhetoric. In the Gorgias, for instance, Gorgias describes rhetoric as the ‘craftsman of persuasion’. Similarly, in the Phaedrus, rhetoric is described as ‘some kind of soul-leading by logoi’,4 as ‘the art of persuasion’ (πιθανoà τšχνη),5 as that which ‘enables one to persuade artfully’ (πε´ιθειν τšχν),6 and as ‘what is 4 “ψυχαγωγ´ια τις δια ` λ´oγων” (Phaedr. 261a). The “τις” indicates that this not a full definition. A similar usage of the indefinite pronoun can be found at Gorg. 462c. There he says about rhetoric that it is “some empeiria” (’Eμπειρ´ιαν γωγš τινα). Socrates emphasizes that he uses “τινα” because rhetoric is not the only empeiria. Analogously, this suggests that rhetoric is not the only type of soul-leading through logoi. Another kind might be the philosophical soul-leading that takes place through dialectical arguments. This speaks against Jessica Moss’ [21] interpretation, which identifies dialectic as a kind of rhetoric. Moss shows persuasively that Socrates often engages in soul leading, as does Diotima in the Symposium (ibid., 3–6), and that dialectic is likely “a soul-leading in a higher sense than that applicable in large gatherings” (ibid., 17). However, our interpretation suggests that this does not imply that dialectic is a kind of rhetoric as it could be a soul-leading of a different kind. Nevertheless, Moss’ claim that the Phaedrus as a whole is concerned with soulleading is persuasive if we add a slight modification: It is concerned with soul-leading of different kinds—the soul-leading of rhetoric and the soul-leading of philosophy. 5 Phaedr. 269c–d. 6 Phaedr. 260d9.

104

C. Roser

concerned with that which is persuasive’ (τoà πιθανoà).7 These passages strongly suggest that the ability or art of rhetoric is primarily characterized by reference to persuasion and that persuasion is precisely the end for which rhetoric uses arguments. The problem, however, is that this characterization of rhetorical argumentation does not really distinguish it from dialectical argumentation,8 given that Socrates also aims at having a persuasive effect on others, as a number of scholars have convincingly shown in recent publications.9 Indeed, this effect seems to be an intrinsic aim of his kind of argumentation. That is to say, persuasion is not aimed at only by his non-dialectical utterances, such as the myths, nor is it merely his personal aim or an extra benefit of Socrates’ arguments that is external to the aims of dialectical argumentation. On the contrary, dialectical argumentation as such is concerned with persuading or with getting agreement from an interlocutor. One of the strongest pieces of evidence for this is Socrates’ description of his elenchus in the Gorgias. There he states that it is his aim when arguing to make the interlocutor his witnesses, i.e., to make her agree. He also makes clear that this is not a personal goal, but a standard that successful dialectical refutation needs to meet: If one does not make the interlocutor a witness, one has achieved ‘nothing worthy with regard to what the logos is about’.10 This is to say, to make the interlocutor agree is an intrinsic aim of dialectical argumentation. This feature of dialectical argumentation is prevalent throughout the actual dialectical conversations that we encounter in the dialogues. The language of persuasion and agreement is an integral part of these conversations and the agreement of the interlocutor is often described as a natural aim and end.11 Disagreement is regularly 7 Phaedr.

272d–e.

8 It has been suggested that this interest in persuasion can also distinguish rhetorical from dialectical

argumentation. While rhetorical argumentation aims at persuading, dialectical argumentation has other aims, such as the truth, epistemically adequate arguments, or the revealing of inconsistencies. This position, in this simplistic form at least, seems not to be present in the recent literature on the Gorgias or Phaedrus. However, it is still important to consider it. James Doyle [10] comes close to accepting it: “The distinction Socrates is really interested in is one that the distinction between speechmaking and discussion characteristically stands for; and this is a distinction between the basic stances of orator and discussant as such: put briefly, the orator seeks to persuade, while the discussant seeks the truth” (ibid., 92). Doyle also concedes that persuasion plays some role in dialectical argumentation, but he thinks that dialectical argumentation only ascribes to “persuasion of the interlocutor a value strictly conditional upon the truth of the propositions he thereby comes to believe (495 A, 505 E-506 A, 471 E472 C)” (ibid., 91). However, it is questionable whether this concession is sufficient to account for the role of persuasion as it is described in Gorg. 471e–472c. There, Socrates suggests that only the arguments that persuade the interlocutor achieve anything worthy, including especially anything worthwhile ‘with regard to the truth’. This suggests that the relation between truth and persuasion is much more complex: truth seems not simply to be a condition of the value of persuasion, but persuasiveness seems to be a condition for the truthconduciveness of arguments. 9 Cf. Elizabeth Asmis [22]; Hugh H. Benson [23]; Rolf Geiger [24]; Marina Mccoy [14]; Jessica Moss [21] [25]; Tushar Irani [1]; Sandra Peterson [26‚ Ch. 6]; Alejandro Santana [27]; Jacob Stump [28]. 10 Gorg. 472c. 11 Cf. Prot. 314c7, Theaet. 169e.

5 What is the Difference Between the Dialectical …

105

taken as the inciting incident and reason for a conversation, giving it its purpose, which is precisely the resolution of this disagreement.12 Socrates is also frequently explicit in his use of words for persuasion when describing his argumentative aims, e.g., when he says that his arguments are meant to persuade the interlocutor of the conclusion.13 That persuasion and agreement play such a role seems to be built into the typical setting for dialectical argumentation. Socrates does not argue primarily for himself alone. Rather, he directs his arguments towards others, and, as such, the arguments that he uses in this way are supposed to have some effect on those towards whom they are directed. Socrates describes this desired effect explicitly as persuasion in a particularly telling passage of the Gorgias: [Soc:] These things [i.e., the things that Socrates has just said] are probably a bit odd, yet they make clear that about which, by showing it to you [Callicles], I want to persuade you to change (πε‹σαι μεταθšσθαι) your mind, if I am somehow able. (Gorg. 493c)14

Socrates wants to persuade Callicles by showing something to him, i.e., by using his arguments. These arguments are supposed to show something to the interlocutor and, thereby, affect them in such a way that they come to accept the desired conclusion.15 This aspect of Socrates’ approach is not just a feature of the intersubjective or external dialectical conversations. Rather, according to the description of thought in the Theaetetus, it is also an intrinsic feature of intrasubjective thinking and judging. In the Theaetetus, Socrates describes thinking and judging as aiming at self-persuasion: Soc: And do you think that anyone else, in her right mind or not, ever dared seriously to say to herself to persuade herself (¢ναπε´ιθoντα αØτ`oν) that ‘A cow is necessarily a horse’ or ‘Two is necessarily one’? (Theaet. 190c)

This passage describes the formation of a judgment (doxa). Socrates describes this process as an internal dialectical conversation, and he suggests that even this internal dialogue is aimed at persuasion. As such, even the judgment formation that one does by oneself is just a case of persuasion, namely of self-persuasion.16 With these considerations in mind, we can extend the list of similarities between dialectical and rhetorical argumentation. It is not just that both use arguments. Both also typically use their arguments in an interpersonal setting in which they are directed at other people. As such, in both kinds of argumentation, arguments are 12 Cf. Gorg. 465a; Phil. 50 d–e; Charm. 165b; the discussions with Polus and Callicles in the Gorgias

also both start with dissent. for example, Gorg. 493c–d, 494a, 505d, 527d; Crat. 391a; Phaedo 63a, d, 77a, 89a, 91b; Theaet. 176a2–3, 195c; cf. Soph. 182b, 259a. 14 Cf. Gorg. 493d, 494a. 15 Jessica Moss [25, p. 229] argues extensively for a similar conclusion. 16 In this regard, Socrates’ model of thought does not differ from that of Isocrates. In both cases, thinking and judging is conceived as an attempt at self-persuasion. Isocrates develops his model of thinking in the hymn to logos (Antidosis 255–257): “It is those very same arguments with which we persuade when we speak to others that we use when we deliberate (with ourselves) and we call, on the one hand, ‘able in speaking’ (·ητoρικoς) ` those who are able to speak among the many and, on the other hand, we deem ‘able in deliberating’ (εÙβoλoυς) ´ those who speak in the best way themselves to themselves on these matters” (15.256). 13 See,

106

C. Roser

typically meant to affect these other people—those at whom they are directed. They are supposed to bring others to say, believe, or do something. In short, the arguments are used to persuade others. The difference between the two kinds of argumentation, thus, is not that the one kind tries to persuade with their arguments and the other not. Instead, the difference rather lies in how they try to persuade or in what kind of persuasion they aim at. The question we have to ask is what distinguishes these two ways or kinds of persuasion from each other.

5.2.2 Option Two: The Good vs. The Bad Persuasion Another set of interpretations accepts that both dialectic and rhetoric use arguments for the sake of persuasion, but still holds onto the claim that there is a difference in motivation or aim. While both Socrates and the rhetoricians use their arguments to persuade, they do this with different ultimate motivations or ends, and, as such, they persuade in a different way. Most interpretations that take such an approach say that the crucial difference in motivation concerns the role of epistemic ends in both kinds of argumentation. Rhetoricians are often described as being concerned with persuasion for the sake of winning, overpowering the opposition, personal gain, or public esteem, and as not being concerned with truth and knowledge in their persuasion. By contrast, dialectical argumentation and philosophy are supposed to be primarily concerned with knowledge and truth, and their interest in persuasion is supposed to be, as it were, subordinated to this concern. Some interpretations have dialectical persuasion aim at additional objectives, such as the improvement of the interlocutor, or the protreptic aim of turning them towards philosophy. Recently, Tushar Irani [1] has provided an extensive and coherent renewal of this line of interpretation. He argues that rhetoricians try to persuade others in order to support their own “personal interests” (33), to “overpower others” (36), and to “outdo or dominate others” (44). By contrast, Socrates aims at “mutual understanding” (33) and has a “commitment to others” (36) to improve them.17 17 Similarly James Doyle [10, p. 91]: “Rhetoric is characterised (i) as concerned only with persuading the audience, indifferent to the truth of the 5 propositions of which they are persuaded (454 E, 459 A–C); (ii) as manipulating the appearance of good (centrally, pleasure) with no concern as to where the real good lies (464 B–465 D); and (iii) as having tyranny—the usurpation of political power—as its natural end (452 E). Socrates consistently depicts philosophy (or dialectic), on the other hand, as (i) ascribing to persuasion of the interlocutor a value strictly conditional upon the truth of the propositions he thereby comes to believe (495 A, 505 E–506 A, 471 E–472 C); (ii) concerned above all with the real good and indifferent to the apparent good (482 A, 494 E–495 A; and (iii) assigning no value to the accumulation of (so-called) political power for its own sake (468 C).” Marina Mccoy [14] also defends such a view when it comes to Gorgias. Gorgias’ rhetoric is concerned with a persuasion which does not require “knowledge, whether of technical or moral matters” (90). She argues that “[i]f there is any point of difference, it is that Socrates seeks a kind of knowledge to which Gorgias is indifferent” (92). This is the political knowledge about the just and unjust. G. B. Kerferd [29‚ pp. 62–63 and 78–80] follows such an interpretation primarily with regard to the difference between dialectic and eristic, but he also suggests that this might be the difference

5 What is the Difference Between the Dialectical …

107

This reading surely finds strong support in the way Plato portrays some of the rhetorical figures in his dialogues, especially Gorgias, Polus, and Callicles, all of whom suggest that rhetoric is a means to gain great power over others. It also finds support in the way that Socrates describes the rhetorical arguers and their rhetorical argumentation as “flattery”, an “irrational thing”,18 as not “worth anything with regard to the truth”,19 and as wanting to please the people (demos).20 However, in spite of this evidence, this interpretation does not provide a criterion that sharply distinguishes rhetorical from dialectical argumentation. The main problem is that there is a wide variety of possible motivations and ends that may motivate different rhetorical arguers. While some rhetoricians might indeed ultimately be motivated by such non-rational ends, others are not. Instead, some, such as Gorgias in the Helen, take themselves as benefitting the audience, and others, such as Isocrates, take rhetorical argumentation to be epistemically valuable and a means for acquiring the truth. Plato and his Socrates are aware of this fact. We can see their awareness in the Gorgias. For, if we consider the actual argumentative behavior of the rhetoricians who appear in the dialogues, they seem to be driven to argue by similar motivations to those of Socrates: Callicles tries to persuade Socrates in his long speech with the declared aim of benefitting him by making him stop doing philosophy, while Polus tries to prove that Socrates’ claims are false by refuting Socrates.21 However, that they share this motivation does not mean that they argue dialectically. Callicles’ long speech is full of rhetorical features and Polus’ attempted refutations are explicitly described by Socrates as “rhetorical”.22 The second problem is that, when seen in the light of the motivations of at least some rhetoricians, many of Socrates’ comments about the epistemically problematic or harmful features of rhetorical argumentation are not to be understood as definitional, reference-fixing, or even neutral descriptions. Rather, they should be interpreted as criticism and as having an antagonistic character. This is especially true when he claims that rhetoric is alogos, or that rhetorical refutation isn’t ‘worth anything with regard to the truth’. These comments do not describe how those who between rhetoric and philosophy. Cf. Erik C.W. Krabbe [8, p. 211]; J. Anthony Blair [9, p. 159f.]; George Duke [30]. 18 Gorg. 464b–465a. 19 Gorg. 471e–472a. 20 Cf. Gorg. 481c–e. By contrast, Socrates emphasizes that he aims at the truth and knowledge, claiming that his argumentation is a communal activity and beneficial for all the participants, including the refuted. Gorg. 458a–b, 472c–d, 487e–488a, 492d, 515b. 21 The only option that remains for this interpretation is to claim that the rhetoricians are ignorant about themselves; they think they are concerned with the truth, but they are not. Yet the behavior of the rhetorical figures in the dialogues suggests that they are concerned with the truth. They make truth-assertions, they attempt to justify their assertions, and they attempt to refute Socrates. This suggests that they have at least some concern with the truth. In fact, Socrates does not seem to suggest that they have no concern with the truth. He emphasizes that conversation concerned with truth and knowledge is supposed to motivate the interlocutor to enter into a discussion with him, a strategy which presupposes that the interlocutors have at least some concern with the truth. 22 Gorg. 471e.

108

C. Roser

argue rhetorically want to argue. Rather, they criticize the fact that rhetoricians argue in such an epistemically valueless way, whether they intend to do so or not. To be clear, the problem is not primarily that Socrates’ comments might be unfair or false. Instead, the problem is that they are a criticism, and as such, they do not provide the kind of impartial or neutral description of rhetorical argumentation that we are looking for. Before we can understand this criticism, we need to understand exactly what kind of language usage is being criticized, i.e., what kind of argumentation is described as rhetorical in the first place. Only with such a description can we adequately understand why this language use is supposed to be epistemically inadequate.23 What is more, Socrates’ own criticism actually indicates that some rhetoricians do want to argue in an epistemically adequate way. Socrates directs his criticism directly at the rhetorical characters themselves, and, as such, the criticism serves an argumentative function. By arguing that the rhetorical way of arguing is epistemically inadequate, he tries to show to his rhetorical interlocutors that they should change how they argue. This argument presupposes that they accept that one should argue in an epistemically adequate manner. For it is only with this presupposition in place that Socrates is able to show the rhetoricians in a dialectically legitimate way that they should argue differently by demonstrating that their way of arguing is epistemically valueless. Only then will his criticism be able to motivate them to change the way in which they argue.24 That some rhetoricians accept epistemic adequacy as a norm of argumentation is also supported by the fact that epistemic criticism cuts both ways. It is not just Socrates who criticizes his rhetorical interlocutors for the epistemic inadequacy of their arguments, but they also criticize him for the same kind of failing. Most notably, Callicles criticizes Socrates for not being interested in the truth, but only in winning: Cal: Oh Socrates, you seem to be youthful in these logoi, as a true demagogue. Indeed, now you say this demagogically (δημηγoρε‹ς). (Gorg. 482c) Cal: This man does not stop talking nonsense. Tell me, Socrates, are you not ashamed of hunting for words while you are such an age, and of considering it a godsend if someone slips up in their words? (Gorg. 489c) Cal: You are a lover of victory, Socrates. (Gorg. 515b)25 23 Such a neutral understanding of rhetoric seems especially necessary for grasping the conceptual debate between rhetoricians and Plato about the nature of epistemically valuable argumentation. These assertions about the epistemic value of rhetoric are part of this debate. They express a criticism of the rhetoricians and reject their claim to epistemic superiority. However, in criticizing their endeavor and their argumentation as epistemically valueless, Socrates already exploits a preconception of rhetoric and rhetorical argumentation. If Socrates nominally defined rhetorical argumentation as argumentation that is epistemically valueless (alogos), a substantial debate about whether rhetorical argumentation is epistemically valuable would not be possible. 24 This does not mean that they necessarily have an interest in arguing in an epistemically adequate manner as an end in itself. It could well be that their interest in this is derived from their interest in not acting in a way that seems shameful to others, combined with the idea that people would take it to be shameful if they were able to argue in an epistemically adequate way. 25 Polus also suggests that Socrates merely speaks for the sake of winning and neither speaks honestly nor argues fairly (Gorg. 461b–c, 468e, 471a–d, 471e).

5 What is the Difference Between the Dialectical …

109

In these critical comments, Callicles evaluates Socrates’ arguments in epistemic terms. This suggests that Callicles also thinks one should argue in an epistemically adequate manner. What distinguishes him from Socrates is not that the one evaluates arguments in epistemic terms and the other does not, but that they disagree about which arguments are epistemically adequate, or about how one needs to argue if one is to argue in an epistemically adequate way—in the way Socrates argues or in the manner of Callicles. In fact, Socrates explicitly suggests that accusations such as those directed towards him by Callicles are rather typical of those who do not argue in a dialectically adequate way. He claims that if someone who does not argue in a dialectically adequate manner disagrees with someone else, then he quickly becomes angry and believes that the other does not agree out of envy (κατα` ϕθ´oνoν) and love for victory.26 These assertions are either means to discredit the interlocutor or, in the best case, means to motivate him to give up his allegedly absurd claims. As such, assertions of this type are tools used to gain the upper hand in the discussion and to attempt to refute one’s opponent.27 However, while Socrates suggests that the use of such tools in the pursuit of refuting one’s opponent are illegitimate, this does not imply that those who use them are solely concerned with winning the argument. On the contrary, their use of such tools suggests that they do take epistemic adequacy to be a norm for arguments. For such assertions can only help to refute or discredit an interlocutor if it is assumed that the interlocutor should not speak out of envy and love of victory but, rather, out of an interest in truth and epistemic adequacy. We can see then that, while the motivations of the dialectical and rhetorical arguers often differ, this does not by itself provide a sharp distinction between rhetorical and dialectical argumentation, since the motivation behind rhetorical arguments can vary. Some who argue rhetorically appear to have motivations that are fundamentally the same as those that drive Socrates when he argues dialectically. The difference between them appears not to come from the motivation or self-conception of the arguer, but rather from the actual argumentation they engage in. That is to say, the difference appears not to be in how they conceive of their argumentation, but in how they actually argue. What we require is a sufficiently neutral or impartial description of the difference between the two ways of arguing so that we can make sense of Socrates’ criticism that the rhetorical argumentation is not, in fact, epistemically adequate.28

26 See

Gorg. 457d. is suggested by the comparison between this description and Callicles’ usage of ‘lover of honor’. Furthermore, ‘envy’ also seems to invoke an idea present in rhetorical texts. In rhetorical texts, one can find the idea that judgments or assertions are to be discredited if they judge out of ‘envy’ (ϕθ´oνoν). Cf. Gorgias, Palamedes 3; Isocrates 12.21, 81, 172, 251; 15.31, 130, 142, 163, 259. 28 A satisfactory interpretation should render intelligible how the rhetoricians could think that their rhetorical way of arguing is epistemically valuable. As such, the Socratic definition of rhetoric as a kind of flattery which is irrational does not provide us with such an interpretation of rhetorical argumentation. This might also be true of Aristotle’s description of rhetoric. For Aristotle’s reception of Isocrates, see the dissertation by Laura Viidebaum [31]. 27 This

110

C. Roser

5.2.3 Option Three: The Procedure and Form of the Argumentation Another set of interpretations try to specify the distinguishing features of dialectical argumentation by identifying a set of formal features or rules that it adheres to and promotes. Peter Stemmer [2, pp. 96–100] and Benoît Castelnérac, Mathieu Marion [3] [4] have developed this kind of idea most fully by using game semantics to describe dialectic as a rules-based game.29 This approach can also be used to draw out the differences between rhetoric and dialectical argumentation. The basic idea is that these two kinds of argumentation follow different rules and, as such, are different ‘games’. It follows from this that they are different in kind and persuade in two different manners. According to this approach, the difference between the two kinds of argumentation does not lie in the motivation or the aim for the sake of which the argumentation is laid out. Rather, the difference is found in how the argumentation proceeds. Most interpretations of this sort propose differences along the lines of the distinction between makrologia and mikrologia, or between ‘speech’ and ‘conversation’. Dialectical argumentation is typically characterized as taking place in a conversation between two people or within a small group of interlocutors. Questions and answers are both short (answers ideally being restricted to yes and no) and clarificatory questions are allowed. In such a conversation, the two sides can alternate the roles of asking or answering the questions and the audience for the conversation, if there is one, is small. By contrast, rhetorical argumentation is typically characterized as consisting of long speeches that take place in a public setting and are directed at a larger audience. Here, clarificatory questions are typically strictly limited, and the audience and speaker role does not usually alternate. Charles H. Kahn [5, pp. 303–304], for example, emphasizes this difference by claiming that A further interpretation of the difference has been advanced by Alexander Nehamas [19, pp. 10– 11] who suggests that there is a non-partisan and a difference “more in purpose than in method”. He claims that “we can distinguish between Socrates and the sophists by means of his refusal to present himself as a teacher of others” (ibid.). Nehamas formulates this claim with reference to sophistry, but given that he includes Gorgias among the sophists, he suggests that this might also distinguish Socrates from those who Plato describes as rhetoricians (in the Gorgias, Gorgias is not described as sophist, but as a rhetorician, and rhetoric is distinguished there from sophistry). In fact, this might very well distinguish Socrates from Gorgias as a person. However, it does not seem to distinguish their ways of using arguments. For, while Gorgias claims to be able to teach rhetoric, he does not claim to teach each time he argues. In fact, he explicitly says that rhetoric does not succeed in teaching the audience (cf. Gorg. 454e–455a). What is more, this also does not seem to distinguish Socrates’ way of arguing from the approaches of Polus or Callicles. In the Gorgias, all of them argue for positive claims, as does Socrates (namely claims such as that suffering injustice is better than doing it). Neither of them claims that they will fully teach the interlocutor with their arguments. In particular, Polus does not claim to teach. Rather, he tries to refute Socrates, and Socrates describes this refutation as rhetorical (Gorg. 471e). 29 Cf. David Timmerman, Edward Schiappa [32, p. 25]: “The term [dialegesthai] describes a practice that (1) is rule governed, (2) is a definable event in space and time, (3) involves question and answer, and (4) aims at reaching a decision”.

5 What is the Difference Between the Dialectical …

111

“[Plato’s] focus on dialegesthai as the method of question-and-answer is designed to make clear that oratory and the lecture format are not the appropriate vehicles for serious philosophical inquiry”. Likewise, Hugh H. Benson [6, p. 36] claims that “Socrates’ distinctive method consists in reasonably short questions and answers”, and that it is this that distinguishes it from the long speeches of rhetoric. Daniel Werner [7, pp. 44–45] combines this approach with a distinction between private and public contexts. He argues that “while the sophists offer long speeches in public, the philosopher offers short speeches (arguments) in private”.30 This idea has also been picked up by contemporary argumentation theorists such as Erik C.W. Krabbe [8] and J. Anthony Blair [9], who identify rhetoric with speeches and dialectic with conversations.31 As with the other views that we have considered, there is a range of sound evidence that can be offered in support. Socrates emphasizes again and again that the argumentation is supposed to follow these kinds of rules, and this does, indeed, seem to be linked to what he means by “διαλšγεσθαι”, namely to converse. What is more, in the Gorgias he clearly contrasts his preference for this kind of conversations with his rhetorical interlocutors’ preference for long speeches. Soc: Would you be willing to continue, Gorgias, just as we converse (διαλεγ´oμεθα) now, the one asking, the other answering, and to put aside for another time this sort of length of logoi which indeed Polus started? But, don’t cheat with what you have promised, instead, be willing to answer what is asked briefly. (Gorg. 449b)32 30 To be clear, Werner argues that, in the Phaedrus, “rhetoric” is reformed in such a way that dialectic

is a kind of rhetoric. As such, he does not claim that every example of rhetorical argumentation happens in public. However, he speaks here of the more classical understanding of rhetoric, which he takes to be the rhetoric of the Sophists. 31 The influential pragma-dialectical approach in contemporary argumentation theory is also based on understanding dialectic as an interaction between two interlocutors. It is worth noting that this approach is introduced explicitly with reference to Socrates, see Frans H. van Eemeren, Rob Grootendorst [33‚ pp. 15–18] [34, pp. 279–282]. 32 It is unclear whether “διαλεγ´ oμεθα” is used in a technical or an ordinary sense here. Socrates clearly has a special kind of conversation in mind, but it is not clear whether it is the word “διαλεγ´oμεθα” that specifies this special kind of conversation and that distinguishes it from other kind of conversations. Instead, the term could very well mean conversing in an ordinary sense. The ‘just as now’ would, then, specify the particular kind of conversing that Socrates has in mind. This would, for example, allow that makrologia should also be taken as a form of conversing, yet not the type done ‘just as now’. However, Socrates does use “διαλšγεσθαι” in a technical manner shortly before this point. There he juxtaposes διαλšγεσθαι with rhetoric as being linked to different ways of speaking (cf. Gorg. 448d). This suggests that the word “διαλεγ´oμεθα” might also be used technically as referring to the specifically Socratic kind of dialectical conversation. It is worth noting that in the Protagoras, mikrologia is also linked to διαλšγεσθαι, and here it seems to be used in a technical sense, referring especially to the specific kind of conversation in which Socrates engages: “[Alcibiades:] For Socrates agrees here that he does not have a claim to makrologia and concedes to Protagoras, but it would astonish me if he yields to any human when it comes to the ability to converse dialectically and the knowledge to give and receive a logos.” (Prot. 336b-c: ωκρατης ´ μν γαρ ` Óδε Ðμoλoγε‹ μη` μετε‹να´ι oƒ μακρoλoγ´ιας κα`ι παραχωρε‹ ρωταγ´oρ, τoà δ διαλšγεσθαι oŒ´oς τ’ εναι κα`ι ™π´ιστασθαι λ´oγoν τε δoàναι κα`ι δšξασθαι θαυμαζoιμ’ ´ ν ε‡ τ ¢νθρωπων ´ παραχωρε‹.) Both of these passages link mikrologia to a non-ordinary use of “διαλšγεσθαι”.) Here, we clearly find a special usage of the term, given that only a special kind of conversation,

112

C. Roser

This theme of the preferred format of discussion runs throughout the dialogue. Socrates repeatedly insists that the conversation should follow this short-questionand-answer format; his rhetorical interlocutors, by contrast, again and again fail to follow this rule or even protest against it.33 However, despite this strong evidence, these formal rules do not amount to a general and clear-cut distinction between the two kinds of argumentation. As James Doyle [10‚ pp. 92–93] and Charles L. Griswold [11] among others, have pointed out, Socrates also occasionally gives longer speeches of his own, as he himself concedes, without, we may suppose, arguing rhetorically.34 Another problem is that it is possible to do mikrologia in a rhetorical way. In the Phaedrus, mikrologia (βραχυλoγ´ιας) is mentioned as one among the artful devices and forms of logos (ε‡δη λ´oγων) the rhetorician may use, alongside, for example, figurative speaking (ε„κoνoλoγ´ια) or piteous appeal (™λεινoλoγ´ια).35 The ability to do mikrologia is also part of the skillset of a rhetorician,36 as we can see in Gorgias’ self-description in the Gorgias. Gorgias, the paradigmatic rhetorician, claims that he is just as able to succeed in mikrologia as he is in makrologia. He presents this as one of the skills he claims to have, stating that, “no one can say the same things in fewer words than I” (Gorg. 499c: μηδšνα ν ™ν βραχυτšρoις ™μoà τα` αÙτα` ε„πε‹ν). This suggests that Gorgias is able to do mikrologia and makrologia due to the same ability: rhetoric.37

i.e., the Socratic type, is described with this word. Alcibiades juxtaposes this διαλšγεσθαι with a makrologia, suggesting that διαλšγεσθαι happens necessarily in the mikrological format. 33 Cf. Gorg, 448d–e, 449b–c, 461d–462; 505d–506a. This difference can also be seen if one compares the Platonic and Isocratic models of thinking. In the Platonic model, thinking is described as a dialogue consisting of short questions and short answers. In the Isocratic model, it is modelled on Isocratic speeches, i.e., those in which one person speaks and the other listens. For Plato’s model, see Theaet. 189e–190a, Soph. 263e–264a; for the Isocratic model, see Antidosis 255–257). 34 For Socrates’ acceptance that he gives long speeches, see Gorg. 465e–466a. James Doyle [10, pp. 92–93] formulates this objection more fully as follows: “The very casualness with which this distinction, between speechmaking and discussion, is introduced may lead us to suspect that it is too superficial, too much a matter of mere style of discourse, to bear the tremendous ethical and philosophical weight Socrates comes to place upon it. We may also object that it is in the end a distinction of degree rather than of kind: doesn’t an answer to a question posed in discussion become a piece of speechmaking once it exceeds a certain length? In a sense, these objections should be conceded. The distinction in question can’t literally be a matter of mere form. A ‘discussion’ with a sufficiently compliant or mesmerized interlocutor is surely not preferable, by any criteria that could interest Socrates, to a speech, as many passages in Plato’s ‘non-Socratic’ dialogues illustrate all too well; the paradigmatic discussions recounted in the ‘Socratic’ dialogues, for their part, can easily be imagined read aloud by a single speaker, thereby becoming, technically, ‘speeches’! As for the distinction being one of degree, Socrates himself concedes this in acknowledging that certain discussions make necessary long answers that amount to speeches (465E1–466A2)”. 35 Phaedr. 267b, 269a, 272a. 36 Phaedr. 272a. 37 This is confirmed by Phaedrus 267a–b, where it is said that Tisias and Gorgias invented both conciseness (συντoμ´ια) and the endless length of logoi. Both are described as something that is present in the books written about the technê concerning logoi. This clearly suggest that both are rhetorical techniques.

5 What is the Difference Between the Dialectical …

113

Furthermore, in both the Gorgias and the Phaedrus, Socrates explicitly extends the term “rhetoric” or “rhetorical” beyond the typical contexts and forms of the speeches given by rhetors of the time, who primarily gave their speeches in legal contexts, in front of the assembly, or in some other public setting. In the Gorgias, he argues that Polus’ refutation is rhetorical even though Polus gives short speeches in a private setting, just as Socrates does. According to Socrates, Polus argues rhetorically in spite of all this because he refutes “just like those who think to refute in law courts”.38 Similarly, in the Phaedrus, he argues that rhetoric does not just concern law courts, but also antilogical arguments, such as those of Zeno. The reason for this is that Zeno is also concerned with making his subject matter appear in contradictory ways. In both of these cases, Socrates explores a similarity between the typical speeches of rhetors of the time and speeches that have different formal features. He uses this similarity to argue that these other speeches are also rhetorical, in spite of their formal differences from the speeches typically considered to be rhetorical. This argumentation makes clear that the formal features of the typical speeches are not necessary or essential for rhetorical argumentation. They are not what makes an instance of argumentation rhetorical. Instead, what makes a speech rhetorical is something that the typical speeches of rhetors have in common with Zeno’s speeches and Polus’ argumentation. These considerations not only show that the length of a speech cannot be what distinguishes rhetorical argumentation from dialectical argumentation, but they also raise problems for other possible attempts to use formal procedural features to make the distinction. They thereby raise problems for any attempt to specify what it is that makes the use of arguments dialectical by reference to the forms that this usage follows. For, when Socrates describes the argumentation of Polus and Zeno as rhetorical, he does not really care about the form of their argumentation. Rather, he says that the argumentation they deploy bears similarities to the typical speeches of rhetors in some other way. This suggests that it is not the form that makes the use of arguments rhetorical, and, as such, it is also not the form that makes their use dialectical either. Instead, it seems possible that one can follow the formal rules of dialectical argumentation and still argue in a way that is rhetorical. Socrates’ description of how Gorgias undertakes mikrologia points towards the same kind of problem: Soc: Indeed, this is needed, Gorgias: Also make a display to me of exactly this, mikrologia (βραχυλoγ´ιας), but of makrologia (μακρoλoγ´ιας) later. (Gorg. 449c)

Socrates suggests that Gorgias undertakes mikrologia in just the same way as he undertakes makrologia, namely by giving an epideixis, a display of his ability in logos, i.e., a display of rhetoric. As such, Gorgias still seems to argue rhetorically, even when he follows the rules that Socrates proposes. The defining feature seems to be how he follows these rules: He follows the rules of mikrologia in the very same way as he follows the rules of makrologia, by exercising his rhetorical ability. This issue is again not limited to the formal rules governing mikrologia, but also raises 38 Cf.

Gorg. 471a.

114

C. Roser

a problem for other potential rule-based definitions of dialectical argumentation. If one tries to define dialectal argumentation by reference to a set of rules it follows, it seems possible that a rhetorician could acquire an ability to follow these rules by simply learning them. If a rhetorician would then give a display of this ability, she would follow these rules and still argue in a rhetorical way rather than dialectically. These considerations suggest that the difference between rhetorical and dialectical argumentation depends not on which rules the arguers follow, but rather how they follow them. Gorgias 449c points out a further problem for the formal approach: the rules of mikrologia do not, themselves, provide means for explaining why argumentation that follows these rules should be thought of as epistemically valuable. When Gorgias follows the rules of mikrologia, he gives a display (epideixis) of his ability to follow these rules. His success in following the rules without being refuted is simply supposed to show that he has an ability to do so. It does not show anything about his understanding of the subject matter or the truth of his claims. The possibility of succeeding in mikrologia in this way shows that it cannot be the rules of mikrologia that make dialectical argumentation ‘worth something with regard to the truth’, and nor can it be these rules that make dialectical refutation a viable test for whether one knows. Instead, one might succeed simply because one has a general rhetorical ability, or one might fail to succeed because one lacks such an ability.39 This provides a problem for rules-based approaches more generally: For any given set of procedural rules, it seems possible that someone can learn to follow these rules and, merely because of this, avoid being refuted in a conversation that demands one

39 Gorgias, in fact, gets refuted. This suggests that he does not actually have a general ability to follow the rules of mikrologia without contradicting himself. However, this does not show that such an ability is impossible. After all, it is not the short-question format that forced Gorgias to give the answers he gave. Instead, it would have been possible for him to adhere to this format without accepting the premises that eventually led to his refutation. In fact, this seems to be the tactic used by Euthydemus and Dionysodorus to avoid refutation. As such, their ability to argue comes close to a general ability to follow mikrologia such that one can refute others and yet not be refuted in turn—independently of whether one says the truth or knows anything about the subject matter of the conversation (cf. Euthyd. 294b–295a, 269e–270a).

5 What is the Difference Between the Dialectical …

115

adhere to such rules, even if one has no understanding of the subject matter of the conversation.40,41

5.2.4 Option Four: Requirements for the Arguments Used In addition to specifying dialectical argumentation by reference to the procedural rules that govern it, many scholars have also sought to specify it by reference to the rules that determine which kinds of arguments can legitimately be used within it. In particular, this approach is concerned with the question of which kinds of premises and inferences are allowed to be used in dialectical argumentation. The most common rule of this sort is the say-what-you-believe requirement. According to this requirement, dialectical arguments need to use premises that the interlocutor actually believes.42 Besides this, there are also rules that pertain to the inferences leading to a conclusion, such as that the interlocutor needs to use valid arguments and is not allowed to use fallacies or other epistemically illegitimate moves, such as delaying

40 To be more precise, any attempt to characterize the difference formally will run into the following dilemma. (i) Either one formulates a limited set of rules in such terms that one can apply them somewhat easily. Then it is likely that the rhetorician could acquire knowledge of these rules and use this knowledge to succeed in this discussion—without having any understanding of the content. (ii) Or one formulates these rules with more demanding, even normative, terms: e.g. “If it is rational/adequate/contributing to knowledge to say x, say x.” However, rules of this kind beg the question, since both rhetorical and dialectical argumentation are understood in rational terms. Such rules would, thus, fail to capture the essence of the rational conflict. This is even a problem for the complex rules laid out by Benoît Castelnérac, Mathieu Marion [3] as even these rules seem not to distinguish Gorgias’ way of doing mikrologia from the dialectical way. For most of these rules, it seems possible that Gorgias would simply learn them and give a rhetorical display of their use, without having any understanding of what he speaks about. We may add that, even if one were able to find formal rules that cannot be followed merely on the basis of knowing the rules, there would still be a question about why dialectic and rhetoric have the rules they do, and why the dialectical rules are supposed to be the ones that carry epistemic significance. 41 Alexander Nehamas [19] and Rachel Barney [35] have argued that that there are no formal differences between Socratic argumentation and eristic argumentation, especially of the sort used by Euthydemus and Dionysodorus. This is not quite the same claim as mine, given that I am concerned with rhetorical argumentation. When it comes to the way Euthydemus and Dionysodorus argue, there seem to be some candidates that could count as formal differences. For instance, Euthydemus and Dionysodorus do not allow Socrates to add anything to clarify his answer, nor do they allow clarificatory questions. By contrast, Socrates’ allowance that one may ask if one has not understood seems to be a specifically Socratic rule (cf. Euthyph. 295b–c). However, even if these count as formal differences, such differences seem merely to be evidence that there is a distinction between Socrates’ form of argumentation and that of Euthydemus and Dionysodorus, rather than explaining in what that distinction consists. 42 According to Vlastos, this requirement, together with mikrologia, represents one of the two constraints that must be observed “for success in this enterprise [i.e., the Socratic conversation]” (Gregory Vlastos [13‚ p. 7]). This requirement is also used by Benoît Castelnérac, Mathieu Marion [4, p. 11].

116

C. Roser

or giving too much redundant information.43 These rules are more substantial than the merely formal procedural rules of mikrologia,44 and, as such, they do not seem to run into the same problems. Being able to refute a claim by following these more substantial rules might very well show something about the epistemic state of the interlocutor and about the falsehood of their claim. Showing that the claims can withstand this kind of refutation might also provide good reasons for believing that they are not false and that an interlocutor holds them in an epistemically adequate way. However, the problem is that these kinds of rules do not provide what we are looking for: While they might distinguish the dialectical form of argumentation from the eristic or sophistic forms, used, for example, by Euthydemus and Dionysodorus, they do not distinguish it from the rhetorical form. We can begin by considering the say-what-you-believe requirement. Gregory Vlastos [12, p. 712] [13, pp. 8f. and 135f.] suggests that it is this requirement that distinguishes the Socratic elenchus from eristic refutations.45 Hugh H. Benson [6, pp. 52–53] applies the same requirement to the difference between Polus and Socrates, arguing that this requirement also distinguishes the Socratic elenchus from the rhetorical refutation as it is described in the Gorgias.46 One widely discussed problem for this reading is that Socrates is sometimes willing to let the requirement 43 The requirement that one uses valid inference is a feature of the standard account of the elenchus, as Gregory Vlastos [12‚ p. 712] and others have described it. The absence of fallacies is used by Benoît Castelnérac, Mathieu Marion [3, 4] in their rule-based formulation of dialectic. A similar interpretation has been proposed by Peter Stemmer [2, pp. 106–107] who suggests that, in eristic, the arguer merely produces seeming contradictions or consistency: “Denn die Täuschung, die einen eristischen Scheinelenchos ausmacht, besteht nicht darin, daß einer der Diskutanten etwas sagt, was er nicht für wahr hält. Sie besteht vielmehr darin, daß der Fragende versucht, trotz konsistenter Antworten den Schein eines Widerspruchs zu erwecken, oder der Antwortende versucht, trotz widersprechender Antworten den Schein der Konsistenz zu erwecken.” Some have suggested further rules concerning the inferential structure. For example, in the standard model of the elenchus, Vlastos suggests that not just the premises, but also the inferential steps, need to be accepted by the interlocutor, in accordance with what they believe. Hugh H. Benson [6] adds to this the requirement that the arguments do not just need to be valid or sound, but also that the interlocutor needs to be persuaded of the conclusion because the arguments are sound. Here, we can set this complication to one side. 44 These rules do seem to be constitutive rules for dialectical argumentation, for it seems possible that someone who argues dialectically provides fallacious arguments without knowing that they do so, or uses premises that the interlocutor actually does not believe. In this case, we may suppose that they still argue dialectically and so we can take the rules to be rules of success. They imply that one can only ‘win’ a dialectical refutation if one uses arguments that meet these rules, or that only arguments that follow these rules are considered to be good or successful. 45 Cf. Alexander Nehamas [19‚ p. 11]: “Socrates’ insistence on receiving answers his interlocutors truly accept reflects a crucial difference between him and the sophists.” See also Roslyn Weiss [15, p. 73]. 46 Benson uses Gorg. 471e–472c as evidence. He writes: “Socrates’ point here is clear. What distinguishes the Socratic elenchos from the elenchos as Polus practices it or as it is practiced in the jury courts is that the former but not the latter requires that its premises are believed by the interlocutor” (Hugh H. Benson [6, p. 53]). Unfortunately, he does not say how this could be Socrates’ point in the passage. Socrates does not use a Greek equivalent for “premise” and nor does he suggest that the premises are the important issue here. Socrates speaks about the importance of the agreement of the interlocutor, stating: “if I can’t produce you, all alone by yourself, as a witness agreeing on

5 What is the Difference Between the Dialectical …

117

drop. It has been persuasively argued that this requirement cannot, thus, be a necessary or essential feature of dialectical argumentation.47 However, for our purpose, there is a further problem with this reading: A requirement of this kind is typically adhered to in rhetorical argumentation. Let us distinguish two possible versions of this requirement. First, Vlastos has suggested that this requirement is meant to rule out hypothetical arguments. His interpretation finds its main support in the Protagoras, where Socrates explicitly says something along these lines: I won’t have this. For it isn’t this ‘if you wish’ and ‘if you think so’ that I want to be refuted, but you and me. I say ‘you and me’, for I think that the thesis is best refuted if you take the ‘if’ out of it. (Prot. 331c, trans. Gregory Vlastos [13, p. 8])

According to this passage, the arguments are not supposed to start from mere hypothesized premises, but from the premises actually believed by the speaker.48 However, while this requirement might be able to distinguish dialectical from eristic argumentation, for which ‘the prime object is to win’, according to Vlastos, it does not promise a sharp distinction between dialectical and rhetorical argumentation. For, after all, rhetorical arguments are meant to persuade and change the mind of the audience, and this is typically done by using premises or claims the interlocutor believes.49 Arguments with hypothetical premises play a subordinated role, if any at all. Alternatively, the requirement could be taken to rule out dishonesty. The interlocutor should not simply say something because, in this way, she avoids being refuted, but she should be honest about what seems to be right to her. This seems to be the main point of Socrates’ use of the say-what-you-believe requirement when he criticizes Callicles: By the god of friendship, Callicles, don’t think that you can play games with me and answer whatever comes to your head, contrary to your real opinion… (Gorg. 500b, trans. Gregory Vlastos [13, p. 8])

the things I’m talking about, I think I have achieved nothing of any account” (Gorg. 472b, Benson [6, p. 52]). Yet this passage refers to the conclusion, not the premises. This is clear from the context in 472d–472a, where Socrates says that “I shall try to make you too, comrade, say the same things to me.” Here “the same things” refer to the paradoxes that the doer of injustice cannot be happy and that he is less wretched if he is punished. These paradoxes are the conclusion of the refutation, not the premises (cf. Gorg. 475e, 479d). 47 For example, Peter Stemmer [2, pp. 101–113] has argued that this rule cannot be taken to be a necessary rule of dialectic. 48 This is what Vlastos stresses about the say-what-you-believe requirement (Vlastos [13, p. 8]). 49 Hugh H. Benson [6, p. 53] suggests that the difference is that rhetoric “aims at persuading the audience to the discussion rather than the interlocutor himself” (ibid., n. 71). However, Benson’s rather weak formulation already suggests that it is not a criterion that distinguishes them sharply. If a rhetorician speaks to a judge in the presence of the accused, he would surely aim to bring about the agreement of the single interlocutor, rather than the audience. This would then, according to Benson himself, be like the Socratic elenchus in which “the jury and/or judge… is the interlocutor himself” (ibid.). This criterion thus leaves certain cases of rhetorical argumentation as indistinguishable from Socratic argumentation.

118

C. Roser

However, such a requirement is again quite typical of rhetorical argumentation. For example, the interrogation of a witness in a law court also follows this principle: The witness should only say what she believes. What is more, as Marina Mccoy [14, pp. 93–97] has persuasively shown, Polus adheres to such a requirement as he tries to refute Socrates. Again and again, Polus asks what Socrates believes as he attempts to refute him. By doing this, he tries to find a belief Socrates holds that contradicts Socrates’ claims about rhetoric and justice. Eventually, he tries to discredit Socrates’ answers by asserting that Socrates does not say what he believes.50 In this attempt to discredit Socrates, Polus invokes the norm that the interlocutor should say what he believes. This finds further support in the fact that Polus criticizes Socrates’ apparent victory. Gorgias was only led into asserting a premise, Polus claims, because he would be ashamed to deny it. As such, Socrates has not really succeeded in refuting Gorgias because Gorgias did not truly believe the premise he accepted.51 Here, Polus invokes again the principle that a refutation is only valid if the necessary premises which the interlocutor asserted are actually believed by the interlocutor. Next, let us turn to the requirement that dialectical arguments need to be valid and are not allowed to use fallacies. While this requirement is also primarily used to distinguish dialectical from eristic argumentation, there are reasons to think that it might also distinguish dialectical argumentation from rhetorical argumentation. Socrates often points out the fallacious character of rhetorical argumentation when criticizing it. Most notably, he criticizes Polus for his use of several illegitimate tactics, such as his use of the verdict of the many, his ridiculing of Socrates’ assertions, or his description of their terrifying consequences.52 Socrates’ criticism seems to be that Polus uses fallacies, namely the fallacies of argumentum ad populum, which says that something is true because many people believe it to be true, of the appeal to ridicule, which suggests that something is false because it is ridiculous, and of the argumentum ad consequentiam, which tries to show that something is false because its practical consequences are undesirable. The difference between Polus’ rhetorical arguments and the eristic arguments of Euthydemus and Dionysodorus seems to be that while the latter exploit what are today called formal fallacies, Polus rather uses informal fallacies, which are also sometimes called rhetorical fallacies.53,54 This reading is problematic for a number of reasons. One problem is that, as Roslyn Weiss [15] has argued, there are a number of passages in which Socrates may very well also be thought to exploit fallacies. Indeed, even if Weiss is mistaken about these passages, there seems to be no good reason why a dialectal arguer could never 50 Gorg.

471d. Gorg. 461b–c. 52 Cf. Gorg. 471e–472b, 473d-3. 53 Hans Hansen [36] specifies this difference as follows: “Formal fallacies are invalid inferences which “bear a superficial resemblance” to valid forms of inference, so these we may think of as deductive fallacies… Informal fallacies are not characterized as resembling formally valid arguments; they gain their allure some other way”. 54 The similarity to eristic argumentation finds also further support in the Phaedrus. There, Socrates suggests that antilogic and the ability to argue for both sides is part of rhetoric. As such, we may suppose that a good rhetorician is able to argue fallaciously. 51 Cf.

5 What is the Difference Between the Dialectical …

119

use a fallacy either accidentally, not knowing that it is fallacious, or deliberately, for an epistemically adequate purpose, such as when the speaker wants to test whether the interlocutor can identify the fallacy or wants to provide a puzzle that motivates the interlocutor to find a solution. Another even more straightforward problem for this reading is that not all rhetoricians use fallacies all the time, but they still do not argue dialectically just because they sometimes use valid arguments. What is more, at least some rhetoricians who argue non-fallaciously do not do so merely by accident. Rather, they actually accept the rule that one should not argue fallaciously and think that their own rhetorical arguments fulfill this norm, believing that they are valid and sound rather than fallacious. Polus appears to be one such rhetorician and Isocrates another. Both seem to think that their arguments as adequate from an epistemic perspective. Socrates’ assertion that Polus argues fallaciously does not deny this. Rather, it criticizes Polus for using fallacies even if he does not mean to use them. However, Polus’ use of fallacious arguments without being aware of them does not distinguish him from someone who argues dialectically and uses a fallacy by accident.55 This problem already provides a sufficient reason for abandoning this approach to distinguishing the dialectical form of argumentation from the rhetorical form. Nevertheless, it will be worth saying a little more about the difficulties inherent in this approach. One further problem is that the question of which arguments are fallacious and which are not is, itself, a topic of debate between Socrates and the rhetoricians. For instance, Polus does not agree that using many witnesses is a fallacious argument. Rather, he seems to treat it as an adequate way of showing that Socrates’ claims are false and should be rejected. The disagreement between Socrates and Polus on this issue provides an especially promising point of departure for the broader question of what Socrates takes to be a fallacious argument and why he does so. Analyzing the manner in which Socrates engages in this debate, how he distinguishes the adequate way of arguing from the rhetorical way, and how he tries to show that the rhetorical way is fallacious, will help us gain a better understanding of Socrates’ views concerning what it is that makes an argument epistemically adequate and of his reasons for these views. However, the assertion that rhetorical argumentation is fallacious does not by itself provide the kind of description that enables us to achieve this end, even if that assertion is true. Instead, to understand what kind of arguments Socrates takes to be fallacious, we need a more informative description.56 What we need is a description of rhetorical argumentation that is both fair and informative, one 55 Someone who wants to take fallacies to be the key difference could try to avoid this problem by identifying all the fallacious argumentative patterns that rhetoricians in fact use, and then specifying rhetorical argumentation as the argumentation that uses one of these patterns. While this would clearly provide an interesting result, it would probably not provide a satisfying general criterion. Such a definition would likely be insufficiently dynamic, given that it would not account for the fact that rhetoricians might find and then use new fallacies. 56 While there are surely answers to the question of why certain language games are epistemically more valuable than others, the kind of formal characterizations we find in Plato, such as short speeches and limited number of interlocutors, seem to be either insufficiently rich to be the subject of such debates, or are question-begging, given that the rhetoricians also claim that their kind of argumentation fulfils these rules. To illustrate this, let us consider how dialectic relates to two

120

C. Roser

that does not simply presuppose that it is fallacious but instead provides the means to explain why this is supposed to be so.

5.2.5 Option Five: The Importance of Methodology Before we move on to non-formal or not rule-based approaches, we should briefly consider one contemporary version of a formal approach to the general distinction between rhetoric and dialectic that does not, on the face of it, seem to run into the same problems identified for Options Three and Four above. Joseph W. Wenzel [16] and Jürgen Habermas [17, pp. 47ff.] have developed closely related accounts according to which rhetoric and dialectic are concerned with two different perspectives on argumentation.57 Rhetoric is concerned with argumentation as a “natural result-oriented process of persuasive communication” (Wenzel, 1990, 9); dialectic is concerned with the procedural or methodological perspective on argumentation. As such, rhetoric concentrates on how to achieve the natural end of communication, which is, according to Wenzel, persuasion. By the means of persuasion, rhetoric is then able to “help people in a social group make wise decisions” (ibid., 14, my emphasis). By contrast, dialectic is concerned with “making critical decisions” (ibid., 9) and with ‘[producing] good decisions’ (ibid., 14, my emphasis). “The focus of this perspective… is on rules, standards, attitudes and behavior that promote critical decision-making”. As such, dialectic “embraces all methodological, procedural approaches to organizing argumentative discussions” (ibid., 16). Thus, according to this approach, the difference between rhetoric and dialectic is not to be found in the rules that each follows but, rather, in their primary concerns: persuasion as the end of argumentation, in the case of rhetoric, and the rules of argumentation, in the case of dialectic. recent game-theoretical accounts. First, one might be tempted to specify dialectic as the game that Robert Brandom [37] calls the “game of giving and asking for reasons”, which, according to him, includes the various players keeping score of each other’s deontic status, i.e., what they are committed to, entitled to, ought to do, and so on, given what they have said and how they have behaved previously (cf. ibid., xiv, 141ff.). While this kind of scorekeeping does indeed seem to be a crucial part of dialectical argumentation, it does not distinguish it adequately from rhetoric. For, set against this background, the correct way to keep score is itself a point of significant controversy in the debate between Plato and the rhetoricians. The difference between Socrates’ argumentation and that of some rhetoricians is not that the one includes scorekeeping and the other does not. Rather, the difference seems to lie in how they keep score. Second, another option is presented by contemporary accounts of logical games. Logical games adhere to the kind of rules that make every allowable move necessarily truth-preserving (see Jaakko Hintikka [38]). The problem with applying this idea to Plato is that he seems not to formulate rules of this kind, and it is questionable whether he had available to him any sort of notion of necessary truth-preservation in the logical sense. What is more, it is not the case that all moves in Socratic argumentation are necessarily truth-preserving. 57 According to J. Anthony Blair [9‚ pp. 148f.], this is the most widespread view these days (cf. Frans H. van Eemeren, Rob Grootendorst [33, p. 17]). However, there are difference in the ways this view is formulated. Habermas and Wenzel, for example, seem to understand ‘process’ rather differently.

5 What is the Difference Between the Dialectical …

121

This contemporary theory draws attention to an interesting feature of the difference between dialectical and rhetorical argumentation in Plato’s writings. Socrates is significantly more concerned with following certain rules. He repeatedly engages in methodological conversations in which he demands that the conversation follows certain rules: one should argue mikrologically; one should start from the premises believed; and one should not use fallacies. By contrast, rhetoricians seem to be much less concerned with these kinds of issues. Instead, their interest lies primarily in achieving the persuasive end. This distinction can be used to explain both why rhetoricians are so flexible concerning the forms or rules they follow in their arguments, e.g., whether they proceed makrologically or mikrologically, and why they tend to use fallacies, even if they do not mean to. The rhetoricians are just not as focused as Socrates on the correctness of arguments, being primarily concerned instead with their success. Despite its value, this approach does not map perfectly onto Plato’s or Socrates’ understanding of the difference between rhetorical and dialectical argumentation. The problem is that, according to the perspectival view, neither rhetoric nor dialectic offers a full account of argumentation, but merely one perspective on the phenomenon. Nor do they stand in tension with each other; they are just two parts that should be unified in coming to a complete view of argumentation. This is not so in Plato, and especially not in the Gorgias. In this dialogue, Plato clearly has Socrates present rhetorical and dialectical argumentation as two distinct and rival ways of arguing, rejecting the one in favor of the other.58 While the difference between the two may be explained in part by the fact that rhetoric and dialectic are each concerned with different perspectives on, or aspects of, argumentation, it is not the case that Plato links the former exclusively with the persuasive aim and the latter exclusively with the method. On the contrary, rhetoric and dialectic each constitute a view of correct argumentation as a whole, and, as such, both include ideas about the method and about the persuasive end. Just as Plato’s rhetoricians make methodological remarks about correct argumentation, so too does Socrates try to persuade.59 In Plato’s presentation, the difference between a dialectician and a rhetorician lies partly in their different ideas about the way in which one should argue, and partly in their differing ideas about what kind of persuasion is the proper aim of argumentation. So, it is not that they simply approach argumentation from different but compatible perspectives, but, rather, that they disagree on the very question of what constitutes a good argument.

58 He

criticizes Polus for refuting in a rhetorical manner and calls upon him to refute dialectically instead (see Gorg. 471e–472c). 59 This is most apparent in the original texts of the rhetoricians of the time. Isocrates, in particular, makes extensive methodological remarks. 101–113. Rhetorical characters in the Gorgias also make similar methodological remarks (see Gorg. 461b–d, 473e, 482c–483a).

122

C. Roser

5.2.6 Option Six: A Difference in Content and the Role of the Ti-Estin Question If we accept that the difference between rhetoric and dialectic lies neither just in the aim nor in the form of the argumentation, the most plausible remaining candidate is that it lies in the content of the argumentation, i.e., in what each of the two kinds of argumentation argues about. Socrates and the rhetoricians tend to argue about different things and to argue about them in a different order. While both are concerned with normative issues, Socrates has a tendency to speak about them in a more general or abstract way. The role of the ti-estin question stands out here. Socrates often claims that the ti-estin question needs to have a prior place in the conversation. Before one can speak about how a thing is, one first needs to determine and speak about what it is.60 This norm is not usually adhered to in rhetorical argumentation. Socrates explicitly criticizes both Polus in the Gorgias and Lysias’ speech in the Phaedrus for failing to satisfy this order of priority.61 Indeed, it is typical to find Plato’s rhetoricians speaking directly about how to evaluate certain subject-matters without first discussing what those subject matters are. For example, they turn their attention straight to evaluative questions such as whether rhetoric is fine, whether teachers of rhetoric deserve to be punished for the injustice their students commit, to what extent it is good to do philosophy, or whether a young man should yield to a lover. This difference in content seems to be closely related to a difference on the pragmatic side of rhetorical and dialectical argumentation. The speeches of the rhetorical characters are the typical forms of normative rhetorical speech acts: They are praise-, defense-, accusation-, and advice- speeches. These speech acts aim to directly affect the behavior of the intended audience. By contrast, Socrates’ argumentation concerning the ti-estin question is not one of these kinds of normative speech acts, since it does not aim to change the behavior of his interlocutors in such a direct way. Instead, their primary aim is to acquire and transmit an understanding of the subject matter or to test someone’s understanding. While this understanding is also supposed to help in making decisions, such as those concerning how to educate students, this link is less direct than in the case of rhetorical speech acts. However, while the ti-estin question allows us to identify an important difference, it does not provide a generally applicable criterion of distinction. One reason for this is that, while Socrates generally adheres to the priority of the ti-estin question, he also waives this priority on some occasions. Most notably, in the later parts of the Meno, Socrates considers whether virtue is teachable without having first defined what virtue is. A similar pattern can also be detected in other dialogues.62 In the Gorgias, for instance, Socrates argues that doing injustice is worse than suffering it and he puts forward this claim without first defining what doing justice is.63 Socrates 60 Cf.

Euthyph. 5d, Laches 190c, Meno 71b; for a discussion of priority of the ti-estin question phenomena see Peter T. Geach [39], William J. Prior [40], David Wolfsdorf [41]. 61 Cf. Gorg. 448d–e, 462c–463e, Phaedr. 263d–e. 62 Cf. Meno 86d–e: oικεν oâν σκεπτšoν εναι πo‹´ oν τ´ι ™στιν Ö μηπω ´ ‡σμεν Óτι ™στ´ιν. 63 Cf. Gorg. 469aff.

5 What is the Difference Between the Dialectical …

123

still argues dialectically in these cases, even if he argues in a slightly different way. It follows that any full account of dialectical argumentation will also need to account for argumentation that does not follow the normal ti-estin priority of question. A further problem with this proposal is that even if rhetorical argumentation does not necessarily speak first about definitions, it is still possible for a rhetorician to define what they speak about, and at least some rhetoricians indeed seem to do so. For example, Isocrates defines philosophy in the Antidosis. Similarly, large parts of Callicles’ speech are concerned with the definition of justice.64 However, it is not the case that, just because Callicles and Isocrates discuss definitions, they also argue dialectically. On the contrary, they still give long speeches replete with typical rhetorical forms and using a range of typical rhetorical devices. This shows that the difference in content between rhetorical and dialectical argumentation cannot solve our problems. What Socrates and the rhetoricians argue about does not sharply distinguish their kinds of argumentation, since they can argue about the same things, but do so in different ways. The difference must then lie in how they argue about things, rather than in what they argue about. And here we find ourselves back again at the problem identified above: What is it that distinguishes how dialecticians and rhetoricians argue?

5.2.7 Option Seven: Context or General Worldview Having considered differences in content, in aim or motivation, and in formal features or rules, we can now turn to consider two last sets of related features that have been said to distinguish dialectical from rhetorical argumentation. First, one significant difference between dialectical and rhetorical argumentation is that they often seem to occur in different contexts. Rhetorical speeches typically occur in public settings, are directed at many people, propose to show a truth the speaker already knows, and follow rather conventional beliefs. Dialectical argumentation, by contrast, typically happens in private, is directed at one person or a small group of people, is a means of discovery that does not require that the arguer knows the truth, and has a tendency to result in paradoxical claims. Another difference between the two kinds of argumentation is that they are carried out by different people with different value-systems and world-views. It has been suggested that rhetorical arguers value the conventional idea of power, prefer to refute over being refuted, present themselves as political insiders, have a rather conventional set of beliefs, and value justice to an indifferent degree.65 By contrast, Socrates prefers being refuted over refuting, presents himself as a political outsider, has an unusual set of beliefs, and values justice highly.

64 Gorg.

482e–484c. James Haden [42]; Andrea Wilson Nightingale [43‚ Ch. 1]; David Roochnik [44]; George Duke [30]. Charles L. Griswold [11] writes that “‘Rhetoric’ is taken here to constitute an entire world view”. 65 Cf.

124

C. Roser

However, as with the other options discussed above, despite these features showing some differences between rhetoricians and Socrates as the paradigmatic dialectician, they again fail to provide a general distinction between the two ways of arguing. Rhetorical argumentation might have a certain typical form, but it seems to be useable in a wide variety of contexts and by a diverse range of people with different values. As for the context, even though rhetorical argumentation typically happens in public settings and is addressed to groups, it is also possible to argue rhetorically in private and in a way that is addressed to a single person, as when Polus tries to refute Socrates. One further line of evidence for this contextual flexibility is that even poetry and antilogical argumentation are described as parts of rhetoric.66 As for value-systems, even though the rhetoricians we find in Plato’s works typically have the values described above, rhetorical argumentation is clearly not restricted to those who do. Socrates himself provides a case in point when he says in the Menexenus that he might be able to argue rhetorically and was educated in rhetoric. This shows that one and the same person with a single set of beliefs and values can argue both dialectically and rhetorically.67

5.2.8 Summary Our consideration of this range of possible distinctions yields a negative conclusion: None of the options discussed here allow us to pinpoint a defining feature that sharply distinguishes dialectical argumentation from rhetorical argumentation. The primary recurrent problem is that dialectical and rhetorical kinds of argumentation are too varied and flexible to be tied down with a single distinction. Rhetorical argumentation, in particular, can be pursued with a variety of motivations or aims, in different forms, in diverse settings, and about a range of topics. Dialectical argumentation also proves to be extremely flexible concerning motivation, form, and content, even if not to the same extent as rhetorical argumentation. The overlaps that follow from this high degree of flexibility have meant that none of the possibilities considered above successfully yield a sharp distinction between the two kinds of argumentation: They are not necessarily different in their aim or motivation, or in the form or rules they follow, or in their content, or in the context in which they take place, and nor can they be distinguished by reference to the type of person who argues in that way. We may add that an attempt to identify a unique ‘signature’ for each type of argumentation by combining elements of difference is a not a promising approach either. A conjunctive combination would not provide such a distinction, since not all cases of dialectical argumentation exhibit all the typical features. A disjunctive combination would be no more successful, since, as we have seen, several cases of rhetorical argumentation also exhibit typically dialectical features.

66 Cf.

Gorg. 502c–d, Phaedr. 261a–e. 235e–236a.

67 Menex.

5 What is the Difference Between the Dialectical …

125

In addition to this main problem, a second issue has also recurred prominently in our consideration of the possible options: Several of the proposed differences were not sufficiently informative, impartial, or explanatory. They simply posit that dialectical argumentation is epistemically adequate, and do not provide the means to specify what makes it so, nor what it is that makes rhetorical argumentation inadequate. As such, they also fail to provide a means to explain why the one is epistemically adequate and the other not, and to say what it is about them that is epistemically adequate. However, Plato, or his Socrates, seems to specify what an epistemically adequate argument consists in and why it is epistemically adequate by describing what dialectical argumentation is and by demarcating it from rhetorical argumentation. Rather than simply positing that it is epistemically adequate, an interpretation should, thus, be able to provide us with some means to explain why dialectical argumentation is supposed to be epistemically adequate and to identify that in which its epistemic adequacy is supposed to reside. This result suggests that the difference between the dialectical and epistemically adequate kind of argumentation, on the one hand, and the epistemically inadequate rhetorical kind of argumentation, on the other, is much more delicate than is normally thought. Plato and Socrates do not seem to distinguish them by reference either to a single general condition or to a combination of general conditions. As such, they also do not appear to give a definition of dialectical argumentation, since a definition would need to combine general conditions that are both necessary and sufficient.

5.3 The Consequences What are the consequences of our failure to find a general criterion for distinguishing between the two kinds of argumentation? We might now be tempted by this failure to give up on the expectation with which we started that there is a sharp distinction between dialectical and rhetorical argumentation in Plato at all. Either we were mistaken in thinking that Plato wanted to draw such a sharp distinction, or, even though he wanted to, he simply did not manage the task successfully. As Charles L. Griswold [11] puts it: It is striking that while Socrates wants to contrast “rhetorical” speech-making with his own approach of philosophical dialogue, in practice the differences blur.68

Tempting as such a conclusion may be, I believe it should be rejected. All the features of the two types of argumentation that we have considered in this paper show that there are indeed many crucial differences between rhetorical and dialectical argumentation. We should not think that these differences are vague or blurred just because none of them, or no conjunctive combination of them, can distinguish every case of rhetorical and dialectical argumentation. That there is no such general applicable difference 68 Marina Mccoy [14, p. 110] comes to a similar conclusion: “There is no final or complete distinction

between philosophy and rhetoric in the Gorgias because a key part of Platonic philosophical activity is to continue to test philosophy itself.”

126

C. Roser

does not imply there is not a feature of each case of dialectical argumentation that distinguishes it clearly from each case of rhetorical argumentation. Instead, I propose that what we should do is to give up our expectation that there should be a single, or combined, generally applicable condition. Rather, we should use a different conception to account for Socrates’ distinction between the two types: A conception drawing on Wittgenstein’s notion of the family resemblance. With this conception at hand, we can see that while Plato and his Socrates do not distinguish rhetorical and dialectical argumentation by providing a definition of each, they do distinguish them by identifying their many paradigmatic features. These paradigmatic features are supposed to show that the different instances of dialectical argumentation bear a certain resemblance with each other—in Wittgenstein’s terms, a family resemblance—and so too do the different instances of rhetorical argumentation. Whilst our consideration of the different options provided a negative result in terms of the search for a defining criterion, it also yielded an important positive result. None of the options discussed are sufficient to distinguish all cases of the two kinds of argumentation, but they do each reveal features that are characteristic of the two kinds. Dialectical argumentation aims at truth and knowledge; it typically uses the form of mikrologia; it typically uses arguments that start from premises the interlocutor believes; dialectical speakers tend not to use fallacies; dialectical argumentation normally adheres to the order of priority that comes out of Socrates’ treatment of the ti-estin question; it typically takes place in a private setting; and dialectical argumentation is normally pursued by a speaker who values truth more than power (in the conventional sense). All of these are important and informative features of dialectical argumentation, even if not all cases of dialectical argumentation exhibit all these features. The different options also reveal several important and informative features of rhetorical argumentation: it is typically directed at many people with the aim of persuasion; it can easily be used as a means of exercising power; it typically takes the form of a long speech; it often uses a set of arguments that can be described as informal fallacies; it typically does not define the things it speaks about; it normally takes place in a public setting, and especially in law courts. Comparing these two sets reveals important differences between the two kinds of argumentation. The two sets show both that what is typical, central, important, or preferential for each of the kinds of argumentation is different, and that what counts as an outlier also differs for each kind. We should not simply push these differences aside just because some cases of rhetorical argumentation exhibit features that are typical of dialectical argumentation, and vice versa. On the contrary, these differences in what is typical, central, important, preferential, and what are outlier cases, shows quite clearly that these two kinds of argumentation are different. In each of our analysis of the seven options, we found such a typical or paradigmatic difference between dialectical argumentation (D.A.) and rhetorical argumentation (R.A.). We can summarize the differences that we found as follows:

5 What is the Difference Between the Dialectical …

1.

2.

3.

4.

5.

6.

7.

127

PERSUASION: One of the most central or essential features of R.A. is that it aims at persuasion. R.A. seems to be directed at persuasion of any kind, but typically persuades many people at once and often persuades in a way that is admittedly inferior from an epistemic perspective. //D.A. aims at persuasion of a special kind: It mainly tries to persuade the interlocutor to give his agreement, and it tries to do this in an epistemically adequate way (a way that is “worth something with regard to the truth”). AIM/MOTIVATION: R.A. can be used a means to pursue personal gain for the arguer, and as a means to overpower others, without concern for the epistemic adequacy of the conclusion or for what is beneficial for the persuaded. This is one of the typical ways in which R.A. is used. //D.A. is mainly, or even essentially, used for epistemically valuable ends (to discover the truth, to test or obtain knowledge, to reveal ignorance). Its attempt to persuade is normally subordinate to this overall end; the persuasion is also normally supposed to benefit the persuaded. FORM: R.A. is not strongly bound to any form: it is typically carried out in a makrological form (long speech, many listeners, without interrupting questions, etc.), but it can also easily be done in other formats. //In D.A., there is a strong preference to follow the mikrological form (two interlocutors, short questions and answers etc.). ARGUMENTS: R.A. typically starts from premises that are believed by many people, preferably including the listener. It typically uses a set of strategies that Socrates takes to be fallacious, such as the appeal to the many, to fear, or to ridicule. Some rhetoricians also seem to use fallacies deliberately as a means of deception. (i.e., to persuade the interlocutor of the conclusion of the fallacy). //In D.A., it is preferable to use arguments that start from premises that are believed by the interlocutor and the arguer typically avoids the use fallacies (at least insofar as they are a means of deception). META COMMENTS: R.A. is very much concerned with success in the goal of persuasion, and less with the adequacy of the procedure or with the path that leads to the persuasion. //D.A. is highly concerned that the argumentative procedure should be adequate, and that the persuasion comes about in an epistemically adequate manner. As such, it typically includes a significant amount of meta-conversation about how one should argue. CONTENT: In D.A., it is strongly preferred to first specify what something is before arguing how it is. As such, the first topic discussed, and often the most extensively discussed topic, of D.A. is a ti-estin question. //R.A. typically does not have such a preference. It typically speaks directly about comparatively concrete normative questions and most often does so without first defining the thing it speaks about. POLITICAL VS. PRIVATE: R.A. is typically done in public, by, in an ordinary sense, political figures, and with ends such as honor, power, and personal political success (all taken in an ordinary sense). //D.A. is typically pursued in a private setting and is linked to a disregard for honor or power in an ordinary sense.

128

C. Roser

All these differences are pointed out by Plato, either through the dialogical setting or via his main character Socrates. This raises the question of why they point out this diverse set of differences, and what unifies these differences. I propose that we should take these to be paradigmatic features of each kind of argumentation.69 We have seen that none of these features, nor any combination of them, is sufficient to define the two forms of argument, and most of them are also not necessary. This suggests that neither Plato nor his character Socrates mean to define dialectical and rhetorical argumentation with general necessary and sufficient conditions that distinguish the one from the other. Instead, I propose, he specifies them by point out a variety of paradigmatic features and examples, such as those in the list above. This way of specifying concepts can be further analyzed with the help of some ideas that Wittgenstein deploys in connection with the idea of family resemblance in his Philosophical Investigations. Wittgenstein shows, there, how one can explain concepts to others without defining them. He illustrates how this is possible most extensively with the concept of ‘game’, suggesting that there might not be one feature that all different games have in common and that only games have. This is to say, there is no necessary and sufficient condition. Instead, the different examples of games bear a family resemblance to each other. As a consequence, one cannot explain what a game is by defining it, so other means must be deployed instead. Wittgenstein suggests that examples can be used as just such a means: How would we explain to someone what a game is? I believe we would describe games to him, and we could add to the description: “this and similar things are called ‘game’.”70 Imagine that I stood with someone in a city square and said that [“stay roughly there”]. In doing this, I do not draw any boundaries, but I make a pointing gesture with my hand—as if I showed him a specific point. And in this way, one explains, for example, what a game is. One gives examples and wants them to be understood in a certain way.71

This kind of explanation uses examples or descriptions of examples as means by which to specify or convey what something is, and it does so without giving general necessary and sufficient conditions. The listener can understand which other things also count as games by looking for the similarities to the examples provided. Here, the examples work as paradigms; other things are understood by the listener to be games 69 This interpretation develops a suggestion briefly made by James Doyle [10, p. 93]. Doyle suggests that the “distinction Socrates is really interested in is one that the distinction between speechmaking and discussion characteristically stands for” (ibid.). However, he differs from my argument here in suggesting that the distinction consists in “the basic stance of the orator and discussant as such: put briefly, the orator seeks to persuade, while the discussant seeks the truth.” The problem with this reading is that, as we have seen, such a difference in aim does not provide a sharply defined difference that allows us to distinguish all cases of rhetoric from all cases of dialectic. In contrast to Doyle, I propose that this difference in aim is just one paradigmatic feature among others, and not what all the paradigmatic features stand for. Instead, what all the paradigmatic features stand for is a difference between two kinds of persuasion-directed argumentation, or two ways of persuading with arguments, and not between a truth- and a persuasion-directed argumentation. 70 PI 69. 71 PI 71.

5 What is the Difference Between the Dialectical …

129

because they are similar to the example. To be clear, not everything that is somehow similar to the paradigmatic example are games, but only those things that are similar to it in the relevant way. Thus, in addition to getting to know what the paradigmatic examples are, the listener has to grasp what is in the relevant way similar to them. She has to grasp ‘in which way the explainer wants the examples to be understood’, i.e., what it is about the examples that is supposed to be paradigmatic for games.72 This kind of explanation, I propose, can help us better understand how Plato specifies what dialectical and rhetorical argumentation are. This is done by providing several examples that can serve as paradigms and by pointing out which of their features are paradigmatic. In the case of dialectical argumentation, the most extensive paradigmatic example is the argumentative behavior of Socrates himself in the dialogues; in the case of rhetorical argumentation, it is the rhetorical speeches and the argumentation of the rhetorical characters. What is more, when explicitly commenting on argumentation, such as when Socrates and his interlocutors describe how one should argue or make claims about what is rhetorical or dialectical, the features they identify can be understood as that which is paradigmatically rhetorical or paradigmatically dialectical. The juxtaposition between dialectical and rhetorical argumentation plays a special role here. I suggest that Plato does not just give the examples of dialectical argumentation that specify what dialectical argumentation is and the examples of rhetorical argumentation that specify what rhetorical argumentation is. Rather, he intends for the juxtaposition between them to help specify what each is in itself. That is to say, the examples Socrates gives of rhetorical argumentation can help specify what dialectical argumentation is, and vice versa. This is possible because one can also use paradigmatically examples of things that do not instantiate a certain concept to explain this concept. This is also pointed out by Wittgenstein: Could one point for the sake of explaining the word ‘red’ to something that is not red? This would be as if one is supposed to explain the word “bescheiden” (“modest”) to someone who can’t speak German, and, as an explanation, one pointed to an arrogant human and said “Dieser is nicht bescheiden” (“This one is not modest”).73

Both Plato and his character Socrates use the juxtaposition between the two kinds of argumentation in precisely this way. By considering the paradigmatic examples and features of rhetorical argumentation, they can point out what dialectical argumentation is. They simply need to add, as they often do, that this is not the way in which dialectical argumentation proceeds. The reverse is also true. Plato goes some way towards specifying what rhetorical argumentation is by distinguishing it from the way in which Socrates argues. I suggest that we can understand the process of demarcation in Plato identified by Alexander Nehamas [19] in precisely this way. It is a specification of one thing by describing its opposite, and vice versa. The specifications of dialectical and of rhetorical argumentation are codependent. 72 This

does not necessarily mean that she will be able to say what is paradigmatic about the examples. In PI 71, Wittgenstein suggests that it is sufficient that she is able to use the example in the right way. 73 PI, p.14, note.

130

C. Roser

We may note yet another feature of this manner of specification. Neither Plato nor his Socrates simply presuppose that there is a difference between rhetorical and dialectical argumentation and then exploit this presupposition to explain what each of the two kinds is. Rather, he also tries to show that there is such a difference in the first place. This seems to be a further function of the paradigmatic examples and features summarized in our list above. While none of these differences can distinguish all possible instances of rhetorical and dialectical argumentation, each of them presents a paradigmatic example of what the differences between the two kinds of argumentation consist in. In this way, Plato seems to point towards a distinction about which the rhetoricians themselves were probably at least partly unaware.74 This also helps to explain how Plato introduces, or significantly shapes, the terms “rhetoric” (“·ητoρικ´oς”) and “dialectic” (“διαλšγεσθαι”). Neither of these terms is used in the way in which Plato uses them in any source earlier than his dialogues. This strongly suggests that Plato uses the terms in a novel manner that departs from any use to which they were put before him. Our considerations can also explain how Plato introduces these terms: He conveys their meaning by using paradigmatic examples ´ and “·ητoρικ´oς” are also almost completely absent from and features. (“·ητoρικη” the extant literature before Plato,75 while Plato uses this language 138 times.76 While cognates of “διαλšγεσθαι” are often used before Plato, there is no good independent evidence that any of them are used in Plato’s technical sense, referring to a specific

74 To be clear, this is not to deny that Plato associates ‘rhetoric’ and ‘dialectic’ with forms of language usage that were already familiar. Rhetorical language usage is clearly associated with the activity of the rhetor and with long speeches. Dialectical language usage is associated with short-speech competitions. However, when Plato uses and shapes these words in their non-ordinary sense, he does not simply identify them with these well-known usages. Instead, he uses the familiar usages merely as examples or paradigms, with the aim of pointing towards an important and epistemically significant distinction. 75 There are two extant usages which may have preceded Plato. Alcidamas uses “¹ ·ητoρικη” ´ twice at the beginning of On Sophists (On Sophists 1, 2). There, he suggests it is an ability to speak, which is closely connected to ϕιλoσoϕ´ια. Isocrates uses ·ητoρικ´oς in the so-called Hymn to Logos. There, he suggests that someone who is able to speak in front of a crowd is rhetorical (Isocrates 3.8, 15.256). Isocrates also uses ·ητoρε´ια three times, apparently referring to a linguistic practice or expertise (Isocrates 5.26, 12.2, 13.21). The term ·ητωρ ´ precedes the usage of ·ητoρικη´ (cf. Mogens Herman Hansen [54], Jeffrey Arthurs [46]). 76 If Alcibiades II is taken as spurious, the number of instances is reduced to 137. Besides the appearances already mentioned, there are also 71 or 70 more in Aristotle, depending on whether Aristotle gave the Rhetoric its title. It is also used twice in Aeschines 1.71, 3.164 (cf. Stephen Halliwell [47]; Harvey Yunis [48, p. 235]). This is a result of a TLG lemma search using the terms “·ητoρικη”, ´ “·ητoρικ´oς”, and “·ητoρικîς” on 20.10.2018.

5 What is the Difference Between the Dialectical …

131

kind of conversation with specific standards.77 Plato introduces these terms in a very limited number of places.78 ) To be clear, here, I have not shown that this paradigm-based approach does, in fact, map onto the specification of dialectical argumentation found in Plato. However, our consideration of the paradigm-based approach already shows that the absence of a sharply defining distinction between dialectical and rhetorical argumentation does

77 The only two extant usage of “διαλεκτικ´ oς” that might be older than Plato can be found in Xenophon’s Memorabilia 4.5.12 and 4.6.1. There are 22 uses in Plato’s dialogues. The most detailed description of usage can be found in Müri Walter [49]. He claims: “Die Ableitungen διαλoγoς ´ und διαλεκτικ´oς finden sich erst bei Platon.” Ernst Heitsch [50‚ p. 132] [51‚ p. 149] has used this suggestion to make the even stronger claim that the word itself is formed by Plato. Similarly, Richard Robinson [52, p. 92] writes: “The notion of dialectic which we find in Plato’s dialogues was invented by Plato.” Both seem to have overlooked Xenophon’s usage, which makes the question of the oldest extant use turn on the relative dating of the Memorabilia and certain Platonic works. Furthermore, Xenophon’s usage might be evidence that the word was genuinely used by Socrates, rather than being a Platonic invention. As to διαλšγεσθαι, there is some evidence that it might already have referred to a special kind of conversation involving short questions and answers, as David Timmerman, Edward Schiappa [32, Ch. 2]. However, the evidence is indecisive, given that in many passages it is unclear whether διαλšγεσθαι means simply ‘to converse’ or refers to a special way of conversing. There is some evidence that may be independent of Plato. First, in Dissoi Logoi 8.1 and 8.12, διαλšγεσθαι is connected with short speech: τî δ’ αÙτî ¢νδρ`oς κα`ι τας ˜ αÙτας ˜ τšχνας νoμ´ιζω κατα` βραχ´ τε δνασθαι ´ διαλšγεσθαι and Öς δ κατα` βραχ` διαλšγεσθαι δναται, ´ δε‹ νιν ™ρωτωμενoν ´ ¢πoκρ´ινασθαι περ`ι παντων. ´ Second, the Antitheses expresses the idea that the wise are clever in conversing dialectically (oƒ σoϕo`ι δεινo´ι ε„σι διαλšγεσθαι, Fr. 51, 6). Third, In Aristophanes’ Clouds, the character Socrates says that the clouds “supplied us with mind, dialectic and nous” (“γνωμην ´ κα`ι διαλεξιν ´ κα`ι νoàν ¹μ‹ν παρšχoυσιν”, 425). This is reminiscent of Philebus 16e–17a, where dialectic is described as a gift from the gods, and of the Politeia’s connection of nous to dialectic. Later in the Clouds, the character Strepsiades uses an apparently invented word, διαλεπτoλoγoàμαι (as he speaks of “chopping logic with the wood of your house”, 1496). This might be a comic reference to Socratic dialectic. Finally, Xenophon speaks in the Memorabilia of being most able in conversing dialectically (διαλšγεσθαι), and he connects this to separating something into kinds: “But, he said “dialectic” is named after wanting to separate the things in accordance to kinds while conversing in common” (διαλšγεσθαι δυνατωτατoυς: ´ ϕη δ κα`ι τ`o διαλšγεσθαι ÑνoμασθÁναι ™κ τoà συνι´oντας κoινÍ βoυλεεσθαι ´ διαλšγoντας κατα` γšνη τα` πραγματα. ´ Mem. 4.5.12). Unfortunately, the temporal relationship between these texts and Plato’s first reference to dialectic is uncertain. The date of the Dissoi Logoi, in particular, is uncertain (cf. Thomas M. Conley [53]; Sebastiano Molinelli [54]). Yet the usage in Antisthenes, Xenophon, and the Clouds suggests that the historical Socrates may well have developed a special usage of διαλšγεσθαι. 78 More than 90% occur in the Gorgias and the Phaedrus (21 uses in the Phaedrus and 107 in the Gorgias). While the non-ordinary usages of τ`o διαλšγεσθαι are less easy to count, due to the lack

132

C. Roser

not imply that the difference between the two is particularly vague or blurred.79 Most people are unable to provide a general criterion that distinguishes between blue and red things. However, in spite of this, their use of “blue” and “red” is not especially vague. They can identify the colors with a high rate of success, and they can also teach others, such as children or non-English speakers, the same concepts by simply pointing to red and blue things. The same may very well be true for the way in which Plato and his Socrates distinguish dialectical from rhetorical argumentation. It is certainly possible that, even though he does not provide a general informative criterion, he can reliably tell them apart and can successfully explain or convey what distinguishes them, with the help of the paradigmatic examples and features. To see how this approach allows us to avoid the conclusion that the distinction between the two forms of argumentation is vague or blurred, we can consider an example of such a criticism. Charles L. Griswold [11] suggests that the distinction is blurred because Socrates sometimes also gives long speeches as part of his dialectical argumentation.80 However, this kind of argument does not hold up if giving long speeches is neither a sufficient nor necessary feature of rhetorical argumentation. If long speeches of rhetoricians are merely paradigmatic examples of rhetorical argumentation, then the use of such a speech as part of Socrates’ dialectical argumentation is unproblematic. It is not the case that every long argumentative speech is rhetorical, but only those that are similar in the relevant sense to the long speeches of the rhetoricians. Similarly, if Socrates’ short speeches are ‘merely’ paradigmatic examples of dialectical argumentation, then it does not follow that any short speech is dialectical, but only those that are similar in the relevant way to Socrates’ short speeches.81 If this is correct, we should not conclude that Socrates’ giving long speeches blurs the distinction. Rather, we have to ask what is in the relevant sense similar to and different from the paradigmatic examples provided by Socrates’ short speeches and the long speeches of the rhetoricians.82 of clarity in certain cases, they also seem to appear only in a limited number of places. This kind of evidence has to be taken with a pinch of salt, given that many earlier texts have been lost, and there may well have been some common usage which has either not survived or was not written down in the first place. 79 Cf. Wittgenstein PI §71. 80 That “Socrates too starts to speak at length” is one of the main arguments use by Charles L. Griswold [11] to suggest that the difference is blurred. The others are that Socrates “sounds rhetorical at times and ends the discussion with a myth.” However, this argument is not quite as relevant for our question, since it might well be that the myths are an exercise in rhetorical argumentation. At least, many of them seem not to be conceived of as epistemically adequate arguments, i.e., as showing that the conclusion is true. 81 To give an analogy, let us suppose someone tries to explain the difference between shapes and colors. He uses two pieces of papers. On the one piece of paper, there are several differently colored squares. On the other piece, there are a number of different blue colored shapes, including one square. Here, all of the paradigms for color have a shape, and all of the paradigms for shapes have a color. There might even be the very same blue square on both sheets. However, this does not imply that the difference between colors and shapes is blurred or vague. 82 As part of this, we should ask what similarities all these paradigmatic features have to each other. However, this does not mean that we should look for the one identifiable feature that is common to

5 What is the Difference Between the Dialectical …

133

5.4 Conclusion In sum, in this paper I have argued that the common specifications of dialectic do not adequately distinguish it from rhetoric. I have proposed that we can explain this by accepting that Plato does not distinguish these two by providing a fixed definition. Instead, he follows a paradigm-based approach: he uses paradigmatic examples and features to point out their differences. This account has crucial implications for Plato’s understanding of what it is for an use of argument to be epistemically adequate or valuable. For, as Alexander Nehamas [20] has rightly pointed out, by distinguishing the dialectical from the rhetorical use of argument, Plato does not merely distinguish two ways to use arguments, but he also specifies what dialectic is. By doing so, he specifies the use of arguments that he considers to be of direct epistemic value. The dialectical use of arguments is the use of arguments that is ‘worth something with regard to the truth’, while the rhetorical use is not. This, to be clear, does not mean that any use of rhetorical arguments is false, or that it cannot contribute to knowledge in some indirect way. It rather means that the dialectical use of arguments is in a direct way of value for figuring out what is true and for acquiring knowledge, in a way in which rhetoric is not. Somewhere else I argued that we can take this to mean that dialectical argumentation is a giving and receiving of reason. It is the use of arguments in which a person needs to be able to justify and defend her claims to count as a knower.83 If this is right, my account implies that Plato does not specify the epistemically valuable use arguments, the giving and receiving of reasons, by providing a general definition of it. Instead, he specifies it by providing a set of paradigmatic examples that distinguish it from uses of arguments that are not in such a way of epistemic value. What is more, the conclusion of this paper even suggests that the epistemically valuable use of arguments cannot even be specified by a general definition. The problem appears to be that any criterion could also be adopted by a rhetorician and used in an epistemically inadequate manner. It can be used to undermine, impede, or hinder the rational considerations of an interlocutor and to disqualify her point of view. This suggests that, in considering which argumentation is epistemically adequate and valuable, we cannot rely on a fixed criterion that we can learn at some point and then use to answer this issue definitely and safely. Instead, we have to be constantly ready to test and check each particular case in its own right, engaging in the messy business of considering to what extent it is similar or dissimilar to the paradigmatic cases of epistemically adequate argumentation. Often, we can settle this question by pointing out one or the other feature of a line of argumentation that all of them and with the help of which we can then define dialectical and rhetorical argumentation. Instead, it may well be that the similarity between the cases of dialectical argumentation is what Wittgenstein has called family resemblance. There is a “complex net of similarities that overlap and crisscross” (PI 67). 83 See my Dissertation “The Epistemic Value of Logos: Gorgias, Isocrates, and Plato on the Possibility of Giving Reasons to Others” submitted in 2019 at the Department of Philosophy, Humboldt University Berlin.

134

C. Roser

shows that it is epistemically inadequate and problematic. But, at times, when we encounter someone with a different view of how we should argue, as Socrates did in the rhetorical figures of his time, we have to be ready to argue with them about this very question. Then, we have to try to find common ground that can be used to settle this question argumentatively.

Bibliography 1. Irani, T. (2017). Plato on the value of philosophy. The art of argument in the Gorgias and Phaedrus. Cambridge University Press 2. Stemmer, P. (1992). Platons Dialektik: die frühen und mittleren Dialoge. De Guyter. 3. Castelnérac, B., & Marion, M. (2009). Arguing for inconsistency: Dialectical games in the academy. In G. Primiero & S. Rahman (Eds.), Acts of knowledge: History, philosophy and logic: Essays dedicated to Göran Sundholm (pp. 43–82). College Publications. 4. Castelnérac, B., & Marion, M. (2013). Antilogic. Baltic International Yearbook of Cognition, Logic and Communication, 8, 1–31. 5. Kahn, C. H. (1996). Plato and the Socratic dialogue: The philosophical use of a literary form. Cambridge University Press. 6. Benson, H. H. (2000). Socratic wisdom: The model of knowledge in Plato’s early dialogues. Oxford University Press. 7. Werner, D. (2012). Myth and philosophy in Plato’s Phaedrus. Cambridge University Press. 8. Krabbe, E. C. W. (2000). Meeting in the house of Callias: Rhetoric and dialectic. Argumentation, 14, 205–217. 9. Blair, J. A. (2012). Rhetoric, dialectic, and logic as related to argument. Philosophy & Rhetoric, 45(2), 148–164. 10. Doyle, J. (2006). The fundamental conflict in Plato’s Gorgias. Oxford Studies in Ancient Philosophy, 30. 11. Griswold, C. L. (2016). Plato on rhetoric and poetry. In E. N. Zalta (Ed.), The Stanford encyclopedia of philosophy (Fall 2016 Edition). https://plato.stanford.edu/archives/fall2016/entries/ plato-rhetoric/. 12. Vlastos, G. (1982). The Socratic Elenchus. The Journal of Philosophy, 79(11), 711–714. 13. Vlastos, G. (1994). Socratic studies. Cambridge University Press. 14. Mccoy, M. (2008). Plato on the rhetoric of philosophers and sophists. Cambridge University Press. 15. Weiss, R. (2000). When winning is everything: Socratic Elenchus and Euthydemian Eristic. In L. Brisson & T. M. Robinson (Eds.), Plato Euthydemus, Lysis, Chamides, Proceedings of the V. Symposium Platonicum (pp. 67–75). Academia Verlag. 16. Wenzel‚ J. W. (1990). Three perspectives on argument. Rhetoric, dialectic, logic. In R. Trapp & J. Schuetz (Eds.), Perspectives on argumentation: Essays in honor of Wayne Brockriede. Waveland Press. 17. Habermas, J. (1981). Theorie des kommunikativen Handelns (Vol. 1). Suhrkamp. 18. Wittgenstein, L. (1953). Philosophical investigations. Macmillan. 19. Nehamas, A. (1990). Eristic, antilogic, sophistic, dialectic: Plato’s demarcation of philosophy from sophistry. History of Philosophy Quarterly, 7(1), 3–16. 20. Nehamas, A. (1998). The art of living: Socratic reflections from Plato to Foucault. University of California Press. 21. Moss, J. (2012). Soul-Leading: The unity of the Phaedrus, again. Oxford Studies in Ancient Philosophy, 43, 1–23. 22. Asmis, E. (1986). Psychagogia in Plato’s Phaedrus. Ill. Classical Studies, 11(1/2), 153–172. 23. Benson, H. H. (1987). The problem of the Elenchus reconsidered. Ancient Philosophy, 7, 67–85.

5 What is the Difference Between the Dialectical …

135

24. Geiger, R. (2006). Dialektische Tugenden. Untersuchungen zur Gesprächsform in den Platonischen Dialogen. Mentis. 25. Moss, J. (2007). The doctor and the pastry chef: Pleasure and persuasion in Plato’s Gorgias. Ancient Philosophy, 27(2), 229–249. 26. Peterson, S. (2011). Socrates and philosophy in the dialogues of Plato. Cambridge University Press. 27. Santana, A. (2011). Reasons and the problem of the Socratic Elenchos. In G. Anagnostopoulos (Ed.), Socratic, Platonic and Aristotelian studies: Essays in honor of Gerasimos Santas (pp. 35– 52). Springer. 28. Stump, J. (forthcoming). Protreptic and Socrates’ erotic art. 29. Kerferd, G. B. (1981). The sophistic movement. Cambridge University Press. 30. Duke, G. (2018). Plato’s Gorgias and the power of ´oγoς. Archiv für Geschichte der Philos., 100(1), 1–18. 31. Viidebaum, L. (forthcoming). Creating the rhetorical tradition. 32. Timmerman, D., & Schiappa, E. (2010). Classical Greek rhetorical theory and the disciplining of discourse. Cambridge University Press. 33. van Eemeren, F. H., & Grootendorst, R. (1984). Speech acts in argumentative discussions. De Gruyter. 34. van Eemeren, F. H., & Grootendorst, R. (1988). Rationale for a pragma-dialectical perspective. Argumentation, 2, 271–291. 35. Barney, R. (2006). The sophistic movement. In M. L. Gill & P. Pellegrin (Eds.), A companion to ancient philosophy (pp. 77–97). Wiley Blackwell. 36. Hansen, H. (2019). Fallacies. In E. N. Zalta (Ed.), Stanford encyclopedia of philosophy (Fall 2019 Edition). https://plato.stanford.edu/archives/fall2019/entries/fallacies/. 37. Brandom, R. (1994). Making it explicit. Harvard University Press. 38. Hintikka, J. (2002). Hyperclassical logic (A.K.A. IF Logic) and its implications for logical theory. Bull Symbolic Logic, 8(3), 404–423. 39. Geach, P. T. (1966). Plato’s Euthyphro. An analysis and commentary. Monist, 50(3), 368–382. 40. Prior, W. J. (1998). Plato and the ‘Socratic Fallacy’. Phronesis, 43(2), 97–113. 41. Wolfsdorf, D. (2004). The Socratic fallacy and the epistemological priority of definitional knowledge. Apeiron, 37(1), 35–68. 42. Haden, J. (1992). Two types of power in Plato’s Gorgias. Classical Journal, 87(4), 313–326. 43. Nightingale, A. W. (1995). Plato, Isocrates, and the property of philosophy. In A. W. Nightingale (Ed.), Genres in dialogue Plato and the construct of philosophy (pp. 13–59). Cambridge University Press. 44. Roochnik, D. (1995). Socrates’ rhetorical attack on rhetoric. In F. J. Gonzales (Ed.), The third way: New directions in Platonic studies (pp. 81–94). Rowman & Littlefield. 45. Hansen, M. H. (1983). Rhetores and Strategoi in fourth-century Athens. Greek, Roman and Byzantine Studies, 24, 151–180. 46. Arthurs, J. (1994). The term rhetor in fifth- and fourth-century B.C.E. Greek texts. Rhetoric Society Quarterly, 23(3/4), 1–10. 47. Halliwell, S. (1994). Popular morality, philosophical ethics and the rhetoric. In D. Furley & A. Nehamas (Eds.), Aristotle’s rhetoric: Philosophical essays (pp. 211–230). Princeton University Press. 48. Yunis, H. (2008). Dialectic and the purpose of rhetoric in Plato’s Phaedrus. The Boston Area Colloquium in Ancient Philosophy, 24, 229–248. 49. Walter, M. (1944). Das Wort Dialektik bei Platon. Mus. Helv., 1(3), 152–168. 50. Heitsch, E. (1997). Dialektik und Philosophie in Platons Phaidros. Hermes, 125(2), 131–152. 51. Heitsch, E. (1997). Platon. Phaidros. Vandenhoeck & Ruprecht. 52. Robinson, R. (1953). Plato’s earlier dialectic. Clarendon Press. 53. Conley, T. M. (1985). Dating the so-called Dissoi Logoi: A cautionary note. Ancient Philosophy, 5(1), 59–65. 54. Molinelli, S. (2018). Dissoi Logoi: A new commented edition. Doctoral thesis, Durham University.

Chapter 6

Platonic βραχυλoγ…α and Aristotle on Say-What-You-Believe David Crane

A say-what-you-believe requirement (SWYB) is often considered distinctive of how Plato’s Socrates engages with interlocutors, distinguishing his method(s) and/or therapeutic aims from those of sophistic and eristic questioners. The treatment of SWYB in Aristotle’s Topics and Sophistical Refutations (SE), however, challenges the notion that SWYB would have distinguished Socrates in the eyes of Plato’s earliest readers. In what follows, I motivate applying an Aristotelian lens to the Platonic Socrates by suggesting that Platonic βραχυλoγ…α, or “brief-speaking,” may be identified with conventions of question-and-answer argumentation similar to those which Aristotle describes. After highlighting the special status of Gorgias for Aristotle’s discussion, I then explore Aristotle’s treatment of SWYB as an argumentative trope whose functions vary with respect to particular modes of question-and-answer argument (didactic, dialectical, peirastic, and eristic or sophistic). Finally, I argue that Socrates’ use of SWYB to refute Calliclean hedonism exemplifies Aristotle’s account of the trope’s “sophistic” application, thus in a (weak) sense substantiating Callicles’ accusation to that effect.

Versions of this paper have benefitted greatly from sympathetic ears and sharp tongues. In particular I would like to thank Elias Avinger, Agnes Callard, John Ferrari, and Charles Ham; audiences at Classics and Philosophy working groups at UC—Berkeley, GVSU, and MSU; and—last but not least—Joseph Bjelde and the organizers of the 2016 Argumentation and Classical Antiquity Conference at Humboldt-University, Berlin. D. Crane (B) Grand Valley State University, Allendale, MI, USA e-mail: [email protected] © Springer Nature Switzerland AG 2021 J. A. Bjelde et al. (eds.), Essays on Argumentation in Antiquity, Argumentation Library 39, https://doi.org/10.1007/978-3-030-70817-7_6

137

138

D. Crane

6.1 A Manner of Speaking When Plato’s Socrates questions Gorgias in typical Socratic fashion, he codes their exchange as a kind of βραχυλoγ…α.1 This label raises an interpretive crux: is Gorgias already skilled in the argumentative practice to which Socrates refers, or is Socrates introducing something new? In Phaedrus, Plato assigns the origin of βραχυλoγ…α, as a rhetorical device, to Gorgias (or perhaps to his pupil, Polus).2 In Gorgias, Gorgias claims for himself expertise in “brief-speaking”3 :

soc:

gor:

soc: gor: soc:

soc: gor:

Then would you be willing, Gorgias, to continue this conversation in the manner we are speaking now, with one person asking questions and the other answering, and to set aside for another time the long style of speeches (τ`o δ`ε μη˜κoς τω˜ν λ´oγων) that Polus began? Don’t renege on your promise, but be willing to answer briefly (κατα` βραχÚ) what is asked. There are some answers, Socrates, that require lengthy speech-making (δια` ` λ´oγoυς πoιει˜σθαι). Nevertheless I shall try to respond as μακρîν τoυς briefly as possible (æς δια` βραχυτ£των). For indeed this is also one of the things I claim, that no one could speak on the same topics more briefly (™ν βραχυτšρoις) than I. That is exactly what’s needed, Gorgias. Give me a display (™π…δειξιν) of this very thing, brief-speaking (βραχυλoγ…ας), and give a display of lengthyspeaking (μακρoλoγ…ας) some other time. So I shall, and you will claim to have heard no one briefer of speech (βραχυλoγωτšρoυ). Let’s have it, then. You claim that you are knowledgeable (™πιστ»μων) about the rhetorical art and that you could make another person a rhetor, too. What, of the things that are, is rhetoric about? Just as weaving is about the production of clothes – isn’t that right? gor. Yes. soc. So too, then, music is about the making of melodies? gor. Yes. By Hera, Gorgias, I marvel at your answers, that you give them as briefly as ´ possible (δια` βραχυτατων). Yes, Socrates, I believe I’m doing this rather well. (Grg. 449b4–449d7)

Yet, despite Gorgias’ claims to expertise at speaking briefly, in this passage Gorgias must be cajoled and flattered into doing so, and likely complies only under pressure of his earlier offer to answer any question.4 Gorgias engages, finally, after an invitation to treat βραχυλoγ…α as ™π…δειξις—a genre of rhetorical display for which

1 Grg.

449b4–449d7; cf. Prt. 329b1–5, 334e4–335c3. 267a6–267b1, 267b10–267c3, 269a5–269a8. 3 Translations are my own, except where noted. 4 Grg. 447c6–8. Cf. Dodds [22, p. 104]. 2 Phdr.

6 Platonic βραχυλoγ…α and Aristotle on Say-What-You-Believe

139

Gorgias was famous,5 though one that seems at odds with the style of brief, questionand-answer exchange traditionally associated with Socrates. Indeed, Gorgias and Polus appear to be entirely unfamiliar with the Socratic style of argument, if we are to judge from the coaching that they receive here and in the ensuing exchanges. Can this representation be squared with the skill in βραχυλoγ´ια that is attributed to Gorgias in Phaedrus, and that he claims for himself in Gorgias? It can, but only if Gorgianic βραχυλoγ…α differs from Platonic βραχυλoγ…α—that is, βραχυλoγ´ια as it is practiced by Plato’s Socrates.6 To identify what is distinctive about Platonic βραχυλoγ…α, it is helpful to read Gorgias against the background of Aristotle’s Topics 1 and 8, along with his Sophistical Refutations (SE).7 These works are composite texts,8 and these portions likely reflect conventions of argument current in Plato’s Academy and in the cultural milieu from which the Academic conventions themselves emerged, notwithstanding Aristotle’s claims to originality.9 In them, we get a rich picture of how controversial or substantive theses would have been debated in adversarial fashion, through questionand-answer exchanges, in order to test their philosophical merit—i.e. as “arguments without winners or losers.”10 We also get a picture of what such debates would have 5 At

Rhet. 1.3.3 Aristotle defines epideictic as a rhetorical genre that assigns praise or blame, and he later cites Gorgias’ Olympiacus as an example (3.13.1414b30–32; cf. DK 82 B6–8a). Cf. Grg. 447a5–447c4; H. Ma. 282b7–282c1. 6 For discussion of βραχυλoγ´ια vis-à-vis the sophists and Plato’s Socrates, see Sidgwick [73, pp. 323–371]; Hudson-William [41]; [22, p. 195]; Kerferd [49, pp. 32–34]; Trapp [79]; Murray [60]; Nehamas [62]; Kurke [52, pp. 138–142, 301–308]. Suv´ak’s [77] discussion of Antisthenes raises the issue of possible differences in usage between Socratic writers, hence my use of “Platonic” as a label for the species of βραχυλoγ´ια that interests me here. 7 This approach has gained traction in Plato studies since a pioneering article by Frede [35]. Cf. Bolton [11]; Ostenfeld [63]; Baltussen [4]; Castélnarac and Marion [19]; Ferrari [31, pp. 25–27]; Fink [33]; Kakkuri-Knuuttila [48]. For older scholarship that flirts with the idea, see Thionville [78, pp. 77–85]; Robinson [68, pp. 438–440] and [69, pp. 18–19, 72, 88]; Ryle [71, pp. 102–108, 119–123]. 8 On the composite nature and dating of Topics and SE, see Huby [40]; Brunschwig [15, pp. I.lxxxiii– xcvi]; Dorion [24, pp. 24–32]; and Slomkowski [74, pp. 6–11, 46–47]. For the interpretation of Plato we may omit Aristotle’s theory of predicables, around which the treatment of τÒπoι in the central books of Topics is organized. But we should include what Aristotle refers to as the “most opportune” or “most common” topoi, which target relations between types of opposites; relations involving coordinate concepts; and different inflections of the same term. Cf. Top. 1.10.104a20– 33; 2.8.113b27–114a6, 2.9.114b6–15; 4.6.119a36–38; 7.3.153a26–29; 7.4.154a12–15 and [74, pp. 142–145]. 9 The claim at SE 34.183b34–184a9 is curious, since Aristotle criticizes the prior teaching of the subject as “quick and dirty (ταχε‹α μν ¥τεχνoς δ’)” and the mere learning by rote of “rhetorical works (·ητoρικoÚς)” and “works of question-and-answer (™ρωτητικoÚς).” Presumably Aristotle’s claim to originality regards the formal presentation of dialectical tropes, etc. Elsewhere Aristotle appears to claim originality vis-à-vis the theorizing of the answerer’s role (Top. 8.5.159a125–127; SE 34.183a37–b6), but the evidence of Plato’s dialogues belies this claim. Cf. Lloyd [53, p. 62]; [11, pp. 121, 141–143]. 10 This felicitous phrase is Brunschwig’s [16], who gives a general picture of the practice described by Aristotle. Cf. [71, pp. 102–108]; [15, pp. I.xxii–xlv]; Moraux [59]; [53, pp. 62–65]; Bolton [10, pp. 196–205] and [12]; Primavesi [66, pp. 34–48] and [67]; [74, pp. 9–42]; [19, pp. 53–62, 67–77].

140

D. Crane

looked like when they deviated from this ideal, as we may imagine happened often in forensic and eristic contexts. Aristotle makes it quite clear that, while the master of these tactics must know both their uses and their abuses, abuses are alien to their exercise in his paradigmatic, “dialectical” situation.11 In these respects Topics and SE form natural companions to Platonic dialogue, the dramatic genre uniquely suited to illustrating the vagaries of question-and-answer debate. There is a cost, however, to this approach: when seen through Aristotelian eyes, Plato’s Socrates seemingly becomes an Academic Socrates in his method(s). This appearance creates difficulties for standard histories of ancient dialectic, which find their telos in Aristotle’s strict attention to argumentation.12 Argument for argument’s sake, abstracted from the relation between an individual’s life and the ethical stances that he is willing to entertain, would seem to be orthogonal to the Platonic Socrates’ avowed aims in examining his interlocutors: either to learn from their examples, or to expose their ignorance and thereby (hopefully) improve their psychic well-being.13 Insofar as the aims of Socratic questioning are taken to establish parameters for his method(s), there is potentially a gap between Academic method(s) and the example set by Socrates. To this objection it may be conceded that the argumentative tactics described by Aristotle likely had a status at the Platonic Academy that was independent of Plato’s Socrates. For, while many of these tactics appear to have a Platonic provenance,14 often they are not quite appropriate to the relevant Platonic contexts, nor does Aristotle explicitly attribute them to his erstwhile mentor. Rather than viewing this omission as a systematic suppression of his principal source, it is preferable to view the tactics themselves as generic. In other words, in cases where resemblances between Platonic dialogue and Aristotle’s theory of dialectic extend beyond trivialities, and in the absence of direct attribution to Plato, it is a plausible interpretive principle to assume both Platonic and Aristotelian texts are reflexes of a shared tradition. But reading Plato’s Socrates through the lens of this shared tradition does not deny Socrates a moral purpose absent from Aristotle’s account; nor, in my view, does Socrates’ concern for the psychic well-being of his interlocutors preclude the pro forma use of established argumentative conventions, on the part of Plato the

11 SE 11.172b5–8; Top. 1.18.108a26–37. Aristotle conceives of different contexts in which questionand-answer exchanges take place, as well as different modes—didactic, dialectical, peirastic, and eristic—of conducting such exchanges. I take Aristotle’s dialectical mode to be paradigmatic for the theory outlined in Topics; see discussion below as well as treatments in the sources listed in the previous footnote. 12 [53, pp. 59–125] gives a broad-ranging historical account. 13 Dorion [25] emphasizes how Aristotelian dialectic is “depersonalized” in comparison with Plato’s Socratic dialectic. This account grounds Dorion’s cautions against applying an Aristotelian lens to Plato’s Socrates, which he outlines in [26] and [27], framed largely in response to Bolton [11]. On Socrates’ avowed aims, see e.g. Ap. 22a8–b5, 23b4–7, 29d4–30b4, 36e5–d1, 39c4–8, 41e1–42a1. 14 Düring [30] provides a remarkable catalogue that indexes Aristotelian tropes with Platonic passages.

6 Platonic βραχυλoγ…α and Aristotle on Say-What-You-Believe

141

writer.15 And when we do read Plato’s Socrates through this lens, we acquire a nearcontemporaneous, technical apparatus as a means of capturing the idiosyncrasies of Platonic βραχυλoγ…α, which we have already seen must differ from at least one other understanding of βραχυλoγ…α that was available to Plato’s audience. This apparatus has two further features to recommend it: it provides insight into how Plato’s earliest readers may have evaluated the arguments of the dialogues,16 and it flags aspects of modern approaches to Socratic argumentation that merit reconsidering. One such aspect concerns what, for many, is a distinctive and general feature of Socratic argumentation: some form of “say-what-you-believe” requirement (SWYB). Plato’s Socrates frequently enjoins an interlocutor not to answer contrary to his opinions (παρα` τα` δoκoàντα), not to qualify his affirmations, and to say what he thinks (λšγειν ¤περ διανoÍ).17 Following Richard Robinson and Gregory Vlastos,18 many commentators view SWYB as crucial to Socratic questioning because, generally speaking, Socrates examines a particular interlocutor’s belief-set for consistency,19 in the context of discussing how one should lead one’s life. An interlocutor would not be made to face his own ignorance, so the thought goes, nor would any practical action be likely to be derived from the candidate belief-set, if the beliefs in question were not being advocated sincerely. In cases where Socrates faces a recalcitrant interlocutor, furthermore, SWYB may contribute to the literary effect of producing sincerity in Plato’s readers, who seek to fill the void left by the unsatisfactory interlocutor.20 Other commentators remain skeptical of the notion that SWYB is a general requirement of Socratic questioning, and instead favor ad hoc explanations of individual occurrences, or restricting the application of SWYB to a particular mode of Socratic questioning.21 As this brief survey demonstrates, commentators 15 On the interpretive freedom that comes from differentiating Plato the writer from his character Socrates, while generally taking Socrates at his word, see Ferrari [32]. 16 As Sedley [72] argues it did for the anonymous commentator on Theaetetus. 17 E.g. Cr. 49d1–2; Grg. 495a9, 500b7; Prt. 331c4–d1; Rsp. 1.346a2, 349a5–6, 350e5. Cf. Ethd 275d7–e2, where Socrates encourages Clinias to be sincere in his exchange with Euthydemus; and note that similar expressions occur in extra-dialectical exchanges, as well (e.g. Lch. 178b1–3; Prt. 312a7, 319a9–10). 18 [69, pp. 15–17]; [81, pp. 35–38] = [82, pp. 8–11]. For a range of perspectives influenced by Robinson and Vlastos, cf. Benson [5, p. 596]; [6, pp. 36–53]; [7]; [8, pp. 287–193]; Nehamas [62, pp. 10–11]; Irwin [45]; Brickhouse and Smith [13, pp. 13–14]; Murray [60, pp. 122–123]; Dorion [25, p. 600]; [26, pp. 566–568]; [27, pp. 265–267]; Cain [17, pp. 18–23, 39–43]; Doyle [29, pp. 72–76]; White [85, pp. 24–27]; McCabe [56, pp. 17–31]; [57, pp. 174–178], and esp. [55, pp. 29–30, 54–59], which provides a nuanced account of different ways in which sincerity may be significant in different Platonic contexts. For literature on SWYB that is critical of SWYB as a generic requirement, see references in n21. 19 For a stimulating account of Socratic examination as not (primarily) directed toward exposing the inconsistencies in an interlocutor’s belief-set, see May [54]. 20 Bailly [1, pp. 72–75] draws attention to this under-appreciated literary effect of SWYB, adducing evidence from ancient rhetorical theories. 21 Cf. Kahn [47, pp. 251–256]; Nails [61, pp. 286–288]; Beversluis [9, pp. 37–58]. Benson (see n19) would restrict his “doxastic constraint,” which entails SWYB and which he believes to be uniquely Socratic, to a particular elenctic mode of Socratic argument. Brickhouse and Smith [14,

142

D. Crane

differ on a range of issues related to SWYB. An Aristotelian lens, however, raises a more basic question: was SWYB a distinctively Socratic requirement? Surprisingly, Aristotle treats SWYB as just one among many argumentative tropes, utilized in a variety of applications, including sophistic ones; and nowhere does Aristotle indicate that SWYB is distinctively associated with Socrates. From this wealth of unattributed applications, we may infer that SWYB was not a distinguishing feature of Socratic argumentation for early readers of Plato—at least not without significant qualification. But Aristotle presents a further wrinkle where SWYB is concerned: Plato’s Socrates, when refuting Callicles’ hedonistic thesis in Gorgias, applies SWYB in a manner that Aristotle would label “sophistic,” raising the perennial specter of a Socrates who “cheats” in his arguments.22 The remainder of my discussion examines the different applications of SWYB described by Aristotle, in order better to understand how the label “sophistic” fits the application of SWYB in the refutation of Calliclean hedonism. We shall see that this label does not, in fact, compel the reader to regard Socrates as the kind of sophistic or eristic questioner who “cheats.” Thus, another virtue of approaching Platonic βραχυλoγ…α through an Aristotelian lens is that it makes possible a charitable interpretation of Socrates, while doing justice to the fact that Socrates comes across as a sophist—even, or perhaps especially, to an audience familiar with Academic debate.

6.2 Plato’s Gorgias and Aristotle It may seem odd, in a discussion of Socrates vis-à-vis the sophists, to focus on Gorgias, rather than e.g. Euthydemus, Protagoras, or a Hippias. On the surface, Gorgias does not even target sophistic argumentation. Instead it pits (traditional) rhetoric against philosophy. More precisely, it pits the status of rhetoric as a craft, and its value to its practitioners and audiences, against similar claims that might be made for philosophy. But it is an inescapable subtext of this conflict that the loser will be collapsed with sophistry in the imagination and lexicon of its audience.23 As is often observed, initially Gorgias is not represented by Plato as a sophist, but rather identifies himself as a rhetorician.24 Moreover, Socrates distinguishes between

pp. 147–149, 152–154], in their criticism of Benson, explore problems with adopting a hard-line stance. Rowe [70, pp. 204–206] suggests SWYB may just be a caution against “making false steps.” Bolton [11, pp. 126–128, 132–135] views SWYB through an Aristotelian lens, as I do below, though he does not draw attention to its sophistical applications. 22 Vlastos’s [83] “Does Socrates Cheat?” echoes Callicles’ complaint at Grg. 483a2–3, but Vlastos focuses his discussion on Socrates’ refutation of Polus. 23 Recent studies of Gorgias have rightly emphasized that Gorgias leaves open the possibility of a true art of rhetoric (hence my use of “traditional” in parentheses above). Cf. Grg. 501d7–504e4 and Carone [18]; Stauffer [76, pp. 123–167]; Trivigno [80, pp. 119–128]; Irani [42, pp. 62–66]. 24 Grg. 449a5–7.

6 Platonic βραχυλoγ…α and Aristotle on Say-What-You-Believe

143

sophistry and rhetoric in his taxonomy of knacks aimed at pleasure.25 Yet Socrates collapses this distinction in his next breath, citing general confusion on the matter—a likely nod to the ad hoc nature of the taxonomy—and repeats this conflation at the end of his exchange with Callicles, with the caveat that, to the extent that rhetoric and sophistry are different, sophistry is “finer (κ£λλιoν).”26 Socrates’ own method of examination is made suspect, in turn, by Callicles’ accusations following the refutation of Polus.27 And when Callicles’ hedonistic thesis is refuted, Callicles calls out Socrates once again for employing sophistry: “I don’t know what kind of sophistic games you are playing (oÙκ oδ’ ¤ττα σoϕ…ζ), Socrates.”28 The sophistic subtext of the conflict in Gorgias helps to explain why Gorgias is the single Platonic dialogue to which Aristotle alludes explicitly in Topics and SE. This mention occurs in SE 12, where Aristotle catalogues sophistic tactics aiming to show that an answerer has employed a fallacy, or to lead an answerer to a paradoxical claim. Aristotle cites specifically Callicles’ allegation that Socrates employs against Polus a well-known sophistic trope that pits the standard of nature (ϕÚσις) against that of law and custom (νÒμoς).29 Other tropes catalogued in SE 12 also resemble argumentative tactics for which Socrates is criticized in Gorgias. So, for example, Aristotle describes questioners who claim that they are seeking to learn something from an answerer; questioners who lead the answerer to subjects on which the questioner is well supplied with arguments; and questioners who present dilemmas wherein either thesis is “implausible (¥δoξoν),” such as “whether…it is more choice-worthy to suffer injustice or to do harm (πÒτερoν…¢δικε‹σθαι αƒρετèτερoν ½ βλ£πτειν).”30 It would appear as though Aristotle, at least, read Gorgias through the lens of his own theory. While the significance of Gorgias to SE 12 highlights some of the tactical similarities between Socratic argumentation and the question-and-answer exchanges described by Aristotle, there are similarities in the language and structures relating to argument, as well. I limit myself here to what is necessary to appreciate the structure of Socrates’ refutation of Calliclean hedonism and his application of SWYB 25 Grg. 465c1–9. Commentators disagree on how to interpret the distinction drawn between sophistry and rhetoric, along with its implications for Plato’s view of Gorgias: cf. [22, pp. 6–10, 216–217]; Harrison [39]; Irwin [43, pp. 125–126]; Cooper [21, p. 39n12]. 26 Grg. 520a6–520b3. 27 Grg. 482c4–483a7. 28 Grg. 497a6. 29 SE 12.173a7–18. 30 SE 12.173a20–22. For instances in Gorgias of the tropes mentioned above, see 453a8–453b3, 457c4–458b1, 461b3–461c4, 482e2–4, 489d2–3, 490c8–490d1, 490d10, 490e9, 491a1–3, 494e7– 8, 506a3–4, 515b6–8. One further trope described by Aristotle is especially relevant to Gorgias, though I do not have space to discuss it in full: that of arguing both from the beliefs people espouse and from their hidden wishes (SE 12.172b36–173a6). At Rhet. 2.23.1399a30–34 Aristotle singles it out as the most effective at generating paradoxes, and it corresponds to the most general criticism directed at Socrates’ argument in Gorgias, i.e. that he shames Gorgias and Polus into expressing beliefs that do not accord with their hidden wishes. Cf. Grg. 461b3–4, 482c4–483a2, 487a7–487b2, 494d2–3, 508c1–3 and Kahn [46, p. 79]; Irwin [44, p. 71n40]; Cooper [21, p. 49n27]; Beversluis [9, p. 309]. On “frankness” in Gorgias, see McCoy [58, pp. 85–110].

144

D. Crane

therein. The first, linguistic similarity concerns vocabulary related to “encounters.” In Topics 1.2 Aristotle mentions three contexts for which his exposition of dialectical theory will be useful: “training” (γυμνασ…α); “encounters” (™ντεÚξεις); and the testing of axioms (τα` πρîτα) of theoretical disciplines (™πιστÁμαι).31 Plato, in Gorgias and elsewhere, often uses vocabulary related to ντευξις in order to characterize the interactions between Socrates and his interlocutors, both on-stage and off-stage, as it were.32 It is a plausible hypothesis that Plato’s fictive “encounters” model interactions similar to what Aristotle envisions as “encounters.” Consider, in this light, the mention of SWYB in Aristotle’s elaboration of ντευξις: It is useful for encounters (™ντεÚξεις) because, after we have taken stock of the opinions of the many (τας ` τîν πoλλîν…δÒξας), we shall converse with them not on the basis of the opinions of others, but on the basis of their own opinions (oÙκ ™κ τîν ¢λλoτρ…ων ¢λλ’ ™κ τîν o„κε…ων δoγμ£των), getting them to change whatever they don’t appear to us to say well (καλîς).

Though unlikely, if “whatever they don’t appear to us to say well (καλîς)” expresses a judgment on the beliefs of the interlocutor, then this passage suggests that there is an ethical dimension to its SWYB requirement,33 readily understandable on certain views of the Platonic Socrates’ therapeutic aims. This interpretation of καλîς, however, contradicts the impression that Aristotle goes to great lengths in Topics 8 to dissociate the answerer from the position that he defends.34 Indeed, the entire thrust of Aristotle’s treatment of dialectic is arguably an attempt to clarify the distinction between a thesis or argument and its proponent, so as to gain a better grasp of the correct and incorrect application of argumentative tactics. This impression is strengthened if, instead, καλîς expresses judgment on the argumentation and not the interlocutor’s beliefs.35 But even on this understanding of the adverb, Aristotle’s focus on argumentation should not be taken to mean that Plato’s Socrates does not engage in the kind of “encounters” to which Aristotle alludes in Topics 1.2. Again, Plato the author can both represent Academic conventions associated with “encounters,” such as enforcing SWYB, and grant his character Socrates a therapeutic intention in doing so. Another important linguistic and structural resemblance lies in the way both ´ Plato and Aristotle formulate “problems” or πρoβληματα for question-and-answer debate. Aristotle defines a πρÒβλημα as a question concerning the choice-worthiness or truth of a proposition. In order to be worthy of debate, two opposed sides of the issue must have a respectable partisanship, either among the general population or among experts; or else the nature of the subject must be such that neither of the opposed 31 Top.

1.2.101a25–101b4.

32 On-stage: Ethph. 2a4; Prm. 126a2; Chr. 154a4; Lys. 203a3; Prt. 309c2, 334d4; Grg. 486e2–487a4;

Ethd. 272e1. Off-stage: Ap. 29d6, 30a3; Tht. 143e5–6, 144a1; Sph. 239c5, 251c2; Prm. 126c1; Prt. 361e2; Grg. 487a4–5, 509a6; M. 71c4–5, 71d8; H. Min. 364a9; Rsp. 7.531e3, 10.598c8. 33 [26, pp. 568–569]. 34 Cf. Galston [37, pp. 81–84] and Baltussen [3, pp. 34–39]. 35 [15, I.116].

6 Platonic βραχυλoγ…α and Aristotle on Say-What-You-Believe

145

sides has a committed following. The opposed claims that constitute a πρÒβλημα are expressed by a disjunctive question: “πÒτερoν…À oυ;” This formula distinguishes the question up for debate from the yes-or-no questions asked by the questioner in the course of the debate. While many yes-or-no questions might easily be reformulated ´ as disjunctive πρoβληματα that are worthy of argument, the structural difference between these types of questions is important for focusing the debate, as subsidiary πρoβλ»ματα could be introduced.36 Plato’s use of the term πρÒβλημα indicates familiarity with the Aristotelian meaning, though at times Plato uses the term more vaguely. In Sophist, for example, Theaetetus despairs of the difficulty of ever being able to capture the kind to which the sophist belongs, since the sophist is “well supplied with problems (πρoβλ»ματα)” and is able to “toss one out (πρoβ£λ)” before being apprehended.37 In Theaetetus we get a more specific idea of how Plato’s Socrates, at least, understands the term. When Theodorus criticizes the present-day Heracliteans for not conducting Socratic examinations and advocates “taking the doctrine out of their hands and considering it for ourselves, as we should a problem (πρÒβλημα) in geometry,”38 Socrates adopts this language and then introduces a thesis contrary to Heraclitean flux—Eleatic monism. We are thus confronted with a πρÒβλημα in the Aristotelian sense. More importantly, the Aristotelian formula for stating πρoβλ»ματα appears often in Plato’s dialogues39 ; so often, in fact, that one recent description of Socratic argument makes these statements its primary focus. Vasilis Politis locates the distinctiveness of Socratic argument in the fact that (1) it is inquiry into a whether-or-not dilemma; (2) it reaches a “radical aporia”40 regarding the dilemma; and (3) the resolution of aporia and dilemma requires the definition of a key concept. Politis, too, connects the dilemma and its related aporia to Aristotelian πρoβλ»ματα.41 Consider, in this light, the following exchange between Socrates and Polus in Gorgias:

soc:

…Consider, then, whether you will be willing in turn to supply a refutation by answering the questions asked. For I think both you and I and the rest of humankind believe that committing injustice is worse than suffering injustice, and that not paying a penalty is worse than paying a penalty.

1.11.104b1–17. The occurrence of subsidiary πρoβλ»ματα will be important for understanding the structure of the refutation of Calliclean hedonism. 37 Sph. 261a5–9. Cf. DK 82 A1a, where Gorgias is said to do the same. 38 Tht. 180c5–6. 39 Phd. 70b5–70c5; Lch. 180a1–3, 181c7–9, 185c2–4; Prt. 324d7–324e1, 329c6–329d1, 349a6– 349b6, 351e5–7; Grg. 472d1–7, 476a3–6; H. Min. 365c3–365d5, 373c6–9; Rsp. 348b2–10. This list is by no means exhaustive. 40 Politis [65, p. 216] defines a radical aporia as “an aporia, articulated by the question whether or not  is , that renders questionable whether particular things that are commonly recognised as exemplars of things that are Φ (or Ψ ) genuinely are exemplars of things that are Φ (or Ψ )”. 41 [65, pp. 134–135]. I differ from Politis in that I interpret some instances of whether-or-not statements in Plato (e.g. “Is Charmides temperate, or not?”) as serving to naturalize the formulaic statement of a πρÒβλημα, in advance of when such a statement introduces a dialectical exchange. 36 Top.

146

pol: soc: pol: soc: pol: soc:

D. Crane

And I think neither myself nor a single other person believes that. After all, would you yourself prefer suffering injustice over committing injustice? Indeed you would, too, and everyone else. That’s way off the mark! Neither I, nor you, nor anyone else would. Won’t you answer, then? Absolutely, for in fact I want to know what you will say. In order that you may know, speak to me as if I were asking you from the beginning (™ξ ¢ρχÁς): Which (π´oτερoν) seems to be worse to you, Polus, committing injustice or (½) suffering injustice? (Grg. 474b1–474c6)

This passage dramatizes the formulation of a whether-or-not dilemma—in this case the very one that Aristotle mentions in SE 12—in such a way as to impress the force of both sides of a πρÒβλημα and, arguably, to make its formulaic nature seem as though it could appear naturally in the course of a friendly intellectual “encounter.” With these tools at our disposal, we are now in a position to describe Socrates’ complex exchange with Callicles and, with it, the application of SWYB that Aristotle would label “sophistic.” Upon Socrates’ conclusion of his exchanges with Polus, Socrates voices a number of paradoxes that follow from his dual position that it is better to suffer injustice than to do injustice, and that it is worst of all to do injustice and not to be punished for it. Callicles interrupts, expressing disbelief. In a lengthy response to Callicles, Socrates reintroduces the fundamental πρÒβλημα from his exchange with Polus, adding dramatic emphasis with one of the most paradoxical statements in the dialogue: So, refute her [philosophy], regarding what I was just now saying, by proving that it is not the worst of all evils to do injustice and not pay the penalty for it. Or if you will allow this to remain unrefuted, by the dog, the god of the Egyptians, Callicles will not agree with you, Callicles, but he will be out of tune for your entire life. (Grg. 482b2–6)

This remark occasions another outburst from Callicles, in which he enumerates several of Socrates’ allegedly sophistic tactics (including the ones mentioned in SE 12), before criticizing the life spent in pursuit of philosophy and elaborating his position that, by nature, the strong should rule. Socrates’ response to this diatribe is to describe his “encounter” with Callicles as a stroke of luck (“I believe that, having encountered (™ντετυχηκèς) you, I have encountered (™ντετυχηκšναι) a gift from Hermes”); and he picks back up the dialectical thread with an allusion to the verse of Pindar that Callicles quoted in order to give his position some rhetorical panache.42 The drama of Callicles’ entrance into the discussion establishes the ensuing question-and-answer exchange as, on the one hand, a debate concerning a formal πρÒβλημα and, on the other hand, an ντευξις where we may expect to see Socrates enforce SWYB. And so he does, many times.43 Calliclean hedonism is itself introduced as a subsidiary πρÒβλημα to the overarching πρÒβλημα of whether it is worse to suffer injustice or to do injustice. At this point in their exchange, Socrates has thrown a wrench in Callicles’ position that the “superior” (κρε…ττoνες) should 42 Grg. 43 Grg.

484b1–c3, 488b2–3. 487d7–487e7, 495a7–9, 500b5–7, 501d8–9.

6 Platonic βραχυλoγ…α and Aristotle on Say-What-You-Believe

147

rule and have more than the ruled, by asking Callicles whether the rulers rule over themselves, by which he means their own pleasures, in addition to others. Callicles bites the bullet and says that a person cannot become happy if he is a slave to anything whatsoever, including constraints on his own appetites. After two cautionary tales fail to persuade Callicles, in which Socrates likens the soul of an intemperate person to leaky jars that perpetually need filling, Socrates introduces the example of the catamite, who wants only to satisfy his own passive, homosexual desires. It is a trivial consequence of Callicles’ extreme thesis that the life of a catamite must be a good one. In Classical Athens a reputation for this lifestyle was shameful, however, to the point of being a contributory factor in a court case that famously resulted in a citizen’s political disfranchisement.44 This outcome is not one Callicles would find desirable. Socrates’ outrageous suggestion prompts the formal statement of a subsidiary πρÒβλημα, and the application of SWYB that Aristotle would label “sophistic”:

cal: soc:

cal: soc:

cal: soc:

cal: soc: cal: soc: cal: soc:

Aren’t you ashamed to lead the arguments (τoς ` λÒγoυς) to these kinds of things, Socrates? Do I lead them there, my good man, or does the one who claims in such an absolute way that those who take pleasure, however they take pleasure, are happy, and does not define what sort of pleasures are good and what sort are bad? But state whether (πÒτερoν) still even now you claim that the same thing is pleasurable and good, or (½) is there some pleasure that is not good. So as not to be inconsistent in my argument, if I say something different, I claim they [pleasure and good] are the same. You are destroying your first words [i.e. the commitment to speak frankly], Callicles, and would no longer sufficiently be examining with me the real issues (τα` Ôντα), if you speak contrary to your own beliefs (παρα` τα` δoκoàντα σαυτù). Well you do it, too, Socrates. Then I don’t do so correctly, either, if in fact I do this, nor you. But, blessed man, consider that this cannot be the good, i.e. to take pleasure in every way. For those many shameful things hinted at just now apparently follow, if this is the case, and many others. So you think, Socrates. Do you truly persist in maintaining these things, Callicles? I do. Should we attempt this argument, then, on the understanding that you are serious? Absolutely. Come on, then, since this is the case, distinguish the following… (Grg. 494e7– 495c3)

44 Aeschin.

1. Cf. Dover [28, pp. 19–31]; Kahn [46, pp. 105–107]; Gagarin [36, pp. 183–188].

148

D. Crane

Modern critics have taken stabs at characterizing the sophistical air of this passage by arguing that Callicles is saddled with a straw man. Could not Callicles have argued a less extreme version of hedonism? On one such account, Plato felt it necessary to refute hedonism in order to refute the thesis that rhetoric is powerful, but felt that refuting anything other than an extreme form of hedonism would have been too complex for the scope of this dialogue.45 On another such account, the extreme hedonist thesis is unnecessary; the refutation of it is sophistical through and through; and both Socrates and Callicles know it.46 Callicles was manipulated into defending this extreme position by his desire to appear shameless, and Socrates proceeds with a formal, if sophistical, refutation because of the type of interlocutor Callicles has proved to be. In what follows I do not address the question of whether the refutation of extreme hedonism is integral to Socrates’ argument,47 though from a literary perspective the allegories of the leaky jars would seem to signal its importance. Neither do I evaluate the legitimacy of all of Socrates’ dialectical moves in the course of the refutation. Instead I focus on just one tactic, the invocation of SWYB, and I suppose, for the sake of argument, that the rest of Socrates’ dialectical moves are legitimate. Even then, on an Aristotelian reading Socrates deploys SWYB in a “sophistic” fashion, adding a greater pointedness to Callicles’ accusation that Socrates argues sophistically.

6.3 Say-What-You-Believe Requirements Aristotle refers to SWYB several times throughout Topics and SE; so far, we have considered its mention in connection with the context of question-and-answer argument that Aristotle dubs “encounters.” The deployment of SWYB, however, also varies with the mode of question-and-answer argument in which interlocutors engage. Aristotle states his basic typology of question-and-answer argument in SE 248 : ,

_

There are four modes of question-and-answer argument (γšνη τîν ε ν τî διαλ´εγεσθαι λÒγων): didactic, dialectical, peirastic, and eristic. Didactic arguments are , , , ones reasoning from the first principles proper to each discipline (ε, κ τω˜ν oι κε´ ˜ν ε\ καστoυ ´ , ιων α ρχω μαθηματoς) ´ and not from the beliefs of the answerer (ε κ τω˜ν τoυ˜ α πoκρινoμ´ενoυ δoξω˜ν)

45 Klosko

[50, pp. 137–138]. [38, pp. 30–42]. 47 For contrast with Klosko and Gentzler, see [46, pp. 102–107]; Irwin [44, p. 69]; [42, pp. 88–96]. 48 My treatment of this passage should be compared with the sources cited in n10. 46 Gentzler

6 Platonic βραχυλoγ…α and Aristotle on Say-What-You-Believe

149

_

_

49 ]); dialec(for the one learning must be convinced [δει˜ γαρ ` πιστευειν ´, τ`oν μανθανoντα ´ tical arguments are deductions from reputable opinions (ε νδ´oξων), leading to a contradic, tion; peirastic arguments are [deductions] from premises believed by the answerer (ε κ τω˜ν , δoκoυντων ´ τî α πoκρινoμ´ενω ) and necessarily known by anyone claiming to have knowledge (in what manner has been specified in other places); eristic arguments are deductions from premises that appear reputable, but actually are not, or are deductions that only appear to employ valid reasoning (ϕαιν´oμενoι συλλoγιστικo´ι).

(SE 2.165a38–165b8)

Here I translate ™ν τù διαλšγεσθαι as “question-and-answer argument,” owing to the explicit mention of an answerer in Aristotle’s description of didactic and peirastic modes (I take Aristotle not to exclude question-and-answer altogether from his didactic mode50 ). Aristotle differentiates his dialectical mode, which I take to be paradigmatic for much of his discussion in Topics and SE, from peirastic by the relative expertise of the participants and by the presence of SWYB. In peirastic, the questioner is ignorant of the subject matter and, despite his ignorance, is able to expose the pretension to knowledge of a reputed expert.51 The definition of eristic is derivative of dialectic. As a hermeneutic for interpreting arguments in the Platonic dialogues, this typology permits us to describe apparent inconsistencies between the Platonic Socrates’ rhetoric and behavior as oscillations between modes of question-andanswer. For we must imagine some slippage between modes taking place within any given encounter, as a natural consequence of the various characters participating in the exchange. As Aristotle himself points out, dialectic is always directed toward another person (πρ`oς ›τερoν).52 But what is the role of belief in each of these modes? Aristotle states that the beliefs of the answerer do not generate the premises of didactic question-and-answer, though insofar as didactic is a mode of question-andanswer, the questioner surely seeks assent to the premises from his interlocutor. This implicit differentiation between “assent” and “belief” raises an issue for interpreting SWYB in Platonic dialogue, for the Greek meaning of an expression such as δoκε‹ μoι can be ambiguous between the two. Several passages in Gorgias require Plato’s Socrates to have understood this distinction,53 which partially accounts for the range 49 The translation of δε‹ γαρ ` πιστεÚειν τ`oν μανθ£νoντα given here follows that of PickardCambridge in Barnes [2]. Contrast [24]: “celui qui apprend doit en effet accorder sa confiance” and Forster [34]: “for he who is learning must take things on trust[.]” I take it that the learner trusts the teacher to put forward true premises that are relevant to the subject, but must nevertheless be convinced of the premises and the reasoning that proceeds from them in order to assent to them, if the didactic mode is one of question-and-answer. Wolf [84, pp. 27–28] imagines a pedagogical scenario in which a student initially takes on trust the certain premises, the reasoning from which is then developed through question-and-answer. A thorough investigation of Aristotle’s meaning here must reconcile this comment with Top. 8.3.159a4–14, 8.5.159a25–37 and SE 10.171a38–11.171b4. 50 SE 11.172a15–17 encourages this interpretation, but contrast Top. 8.5.159a25–37 and [59, p. 288n2]; [24, p. 213]. 51 SE 8.169b25–26; 11.171b4–6, 172a21–32; 34.183b7–8. Cf. [10, pp. 212–219]; [11, pp. 122–123]; [12, pp. 272–276, 278–286]. 52 Top 8.1.155b10. 53 Grg. 480e1–4, 495a7, 513c4–6. Cf. [1, p. 70].

150

D. Crane

of force SWYB has been given, from Vlastos’ idea that “one has given one’s opinion the weight of one’s life” to the view that SWYB need be no more than a caution against “making false steps.”54 This issue, as it relates to Aristotle’s didactic mode, is moot, however, since the refutation of Calliclean hedonism does not count as didactic. Formally, Socrates argues not from first principles, but from νδoξα or “reputable opinions,”55 which classifies the exchange as one of Aristotle’s other three, endoxic modes. Aristotle defines νδoξα as popular beliefs, the beliefs of experts in their area of expertise, or the beliefs of individuals who have a reputation for being wise.56 When Socrates examines Calliclean hedonism, the propositions that courage and knowledge are different, or that pleasure and knowledge are different, qualify as νδoξα.57 Certainly they could be debated in their own right, but they are also plausible positions to which an interlocutor may be willing to commit without further scrutiny. Of Aristotle’s three endoxic modes, the dialectical mode does not, in fact, have a say-what-you-believe requirement.58 The absence of SWYB in Aristotle’s dialectical mode makes good sense for, as has already been remarked, Aristotle systematically divorces the respondent from any ownership over the thesis he is defending. Aristotle writes, for example, that “it belongs to the answerer to give the appearance of the impossibility or implausibility resulting not because of him, but because of the thesis.”59 He thereby suggests that the answerer’s role in debate is to run with an idea as far as reason permits, and furthermore that the answerer’s success is measured not by vindicating his thesis, but rather by not permitting any errors in reasoning along the way. In this way the answerer would be absolved of any responsibility, if the questioner succeeds in refuting the answerer’s thesis. Thus, in Topics 8.5–6, Aristotle treats cases where an answerer must adopt theses and propositions that are not p. 36] = [82, p. 9]; [70, pp. 204–205]. See n18 and n21 for further references. pp. 40–44] = [82, pp. 13–17] and Dorion [26, pp. 566–567] dispute the claim that Plato’s Socrates relies on νδoξα, arguing that to do so is in conflict with the stress Plato’s Socrates places on SWYB. Pace Dorion, I find compelling the arguments made against this point in [51]; Polansky [64, pp. 249–253], and [11, pp. 122–127, 132–135]. It is characteristic of Socrates to draw upon νδoξα that his interlocutor will agree to, and both of these criteria appear integral to Socratic method and Aristotelian peirastic. 56 Top. 1.1.100b21–23. 57 Grg. 495c3–8. 58 At Top. 8.9.160b17–22, squarely in the middle of his discussion of the answerer’s role in dialectical argument, Aristotle cautions against adopting a thesis that is ¥δoξoς, by which he appears to mean “disreputable” rather than the more usual Aristotelian sense of “implausible.” Aristotle states that a thesis may be unacceptable if it is either absurd or indicative of a weak character. His examples of the latter are that pleasure is the good or that it is better to commit injustice than to suffer injustice. This passage runs against the grain of the rest of Book 8; perhaps it may be construed as evidence of a dialectical practice in transition. But it is not evidence of a therapeutic dimension to dialectic such as commentators associate with Plato’s Socrates. Significantly, the reason Aristotle recommends caution is that “they will hate him, not as one upholding a thesis for the sake of argument, but as one saying what he believes (τα` δoκoàντα λšγoντα)”—Aristotle focuses on the response of the audience in making this recommendation, and not the moral correction of the interlocutor. I owe this last point to Elias Avinger. Cf. Smith [75, p. 136]. 59 Top. 8.4.159a20–22. 54 [81, 55 [81,

6 Platonic βραχυλoγ…α and Aristotle on Say-What-You-Believe

151

plausible, introduce the opinions of others, defend the opinions of others, and take over from others the role of answerer. All of these cases have parallels in Platonic dialogue. It is Aristotle’s peirastic mode, however, that bears the greatest resemblance to the manner of Plato’s Socrates, when Socrates is refuting an interlocutor.60 Arguing from the answerer’s beliefs belongs to the definition of peirastic; in peirastic the questioner may be ignorant of the subject; and the primary purpose of peirastic is to expose the ignorance of an interlocutor pretending to knowledge. One immediately thinks of the Platonic Socrates’ frequent profession of ignorance and of his tendency to ascribe expertise to his interlocutors, to which Callicles presents no exception.61 It makes sense, too, that this mode might arise in the course of an “encounter,” where the questioner cannot be sure in advance of the answerer’s disposition. So, it is tempting to describe Plato’s Socratic question-and-answer exchanges as oscillating between an ideal of dialectic, as a joint pursuit of truth, and peirastic, when an answerer claims expertise that he does not possess.62 There still remains, however, Aristotle’s eristic or sophistic mode, for which he elaborates two different applications of SWYB: Asking many questions, even if the subject of question-and-answer argument is defined, and requiring [the answerer] to say what he believes (τα` δoκoυ˜ντα), provides an easy means of leading [the answerer] to an implausible statement or falsehood; and if the one being questioned assents to or denies any of these things, [the questioner] proceeds to subjects for which he is well-stocked with means of attack. But they are now less able to cheat (κακoυργε‹ν) through these means than before, for [those answering] ask what this has to do with the original thesis. (SE 12.172b16–21) There is also the sophistic tactic (σoϕιστικÒν) of demanding, after a thesis contrary to the generally accepted view (παραδ´oξoυ) has been laid down, that [the answerer] answer as it appears [to him] (τ`o ϕαιν´oμενoν), when , , the generally accepted view was put forward initially (πρoκειμ´ενoυ τoυ˜ δoκoυ˜ντoς ε ξ α ρχη˜ς), and to ask such things as, “Is it your belief… (π´oτερ´oν σoι δoκει˜)?” For, necessarily, if the question is drawn from the premises of ´ the deduction, there will be either refutation (™λεγχoν) or paradox (παραδoξoν): ´ refutation, on the one hand, if he grants [the premise]; but if he does not grant [the premise], nor even ´ allow that it is generally accepted (δoκει˜ν), then implausibility (¢δoξoν); and if he does not grant [the premise] but admits that it is generally accepted (δoκει ˜ν), then something like , refutation (ε λεγχoειδ´ες). (SE 15.174b12–18)

The first example, which once again comes from SE 12, identifies SWYB as a tactic designed (a) to secure answers to so many questions that it is easy to catch the respondent in an inconsistency, and (b) to steer the respondent toward topics on which the questioner has more secure footing than the answerer, and is therefore likely to prove quicker in argument. This tactic, along with many others in SE 12, is not fallacious in itself.63 Presumably it is only its potential to be used in the service of 60 [11]

makes the case for comparison. Contra: [25], [26], [27]. 486e5–487d7. See [8, p. 183n15] for a list of parallels with other interlocutors. 62 There are perhaps a few specimens of didactic, as well; one thinks particularly of Parmenides. Cf. [35, pp. 213–214]. 63 Dorion [23] discusses the peculiar status of the tactics enumerated in SE 12, which are not open to generic critique as sophisms. 61 Grg.

152

D. Crane

making an irrelevant argument appear relevant that qualifies this tactic as eristic. Note that Aristotle describes the eristic use of this tactic as “cheating” (κακoυργε‹ν)—the same word Callicles uses of Socrates, and the only time this word appears in Topics or SE. It is the second eristic application of SWYB, however, that more closely resembles Socrates’ tactic in the refutation of Calliclean hedonism. This tactic requires the respondent explicitly to state his belief in the implausible thesis that he defends, in order (a) to refute him if he grants the more plausible propositions that establish the contrary thesis, or (b) to make his position appear implausible or downright paradoxical. Again, when Aristotle calls this tactic “sophistic,” he cannot mean that there could be nothing true or even philosophically interesting about a thesis that is generally considered implausible or unacceptable. Presumably this strategy becomes sophistic when the appearance of refutation is all that matters to the questioner. This was likely a good strategy in forensic argumentation64 ; I myself have seen it used to excellent effect in an American courtroom.65 Socrates’ refutation of Callicles does, in fact, have a forensic air, which is highlighted by the invocation of each man’s deme when their respective positions on the relations obtaining between pleasant, good, knowledge, and courage are set down.66 But the refutation is also represented as a formal dialectical exchange. As we saw earlier, when, following the leaky jar images, question-and-answer begins, a subsidiary πρÒβλημα is established. This πρÒβλημα consists in Socrates’ thesis that an orderly life is better than a life of unrestrained appetite-fulfillment, and in Callicles’ opposed thesis that one should cultivate every appetite and devote oneself to fulfilling each appetite as the need arises. Callicles’ extreme position that the pleasant is identical with the good corresponds to the “thesis contrary to the accepted view (παραδ´oξoυ),” in Aristotle’s description. Socrates then emphatically reminds Callicles to say what he believes, and proceeds to refute Callicles by constructing a syllogism requiring Callicles’ assent to several generally accepted premises about the relations between knowledge and courage and the experience of pain and pleasure. The conclusion of the refutation is stated at 497a3–5, at which point Callicles accuses Socrates of playing “sophistic games.” This instance is clearly one of Socrates invoking SWYB in preparation to refute an interlocutor who has adopted an implausible thesis—what Aristotle labels a “sophistic tactic”—and being called out for it. 64 [38,

pp. 21–22] makes this point with reference to the potential use of Euthydemus’ and Dionysodorus’ eristic tactics in an Athenian courtroom. 65 In the January 2010 Proposition 8 trial, in a federal district courtroom in San Francisco, a political science professor from a reputable institution of higher learning took the stand to defend the claim that gays and lesbians had more political power in the U.S.A. than African-Americans. At the time, the President of the U.S.A. was himself African-American. There may have been a serious point to this claim: in order to settle the issue, one would have to consider carefully the different kinds of systemic injustice faced by these minority groups and the means available to each group for addressing injustice. But the “optics” of this position, under scrutiny in a San Francisco courtroom, were decidedly not good. 66 Grg. 495d1–7. Cf. [22, p. 308], along with the parallels that Dodds cites.

6 Platonic βραχυλoγ…α and Aristotle on Say-What-You-Believe

153

Nevertheless, in this instance we may still read Socrates charitably, even through an Aristotelian lens. In as much as both of these sophistic applications of SWYB could have valid dialectical applications, Socrates may be arguing sophistically only in the weak sense that his tactics are recognizable from their frequent misuse in forensic and eristic contexts. In SE 12 Aristotle states, furthermore, that a certain trope, singled out here and elsewhere as “the sophistic tactic (τÒπoς Ð σoϕιστικÒς),” has both “fair (καλîς)” and “unfair (μη` καλîς)” applications.67 Aristotle thereby indicates that, with reference to the conventions he is describing, “sophistic” can have a neutral sense, i.e. perhaps something along the lines of “sophists characteristically employ this device.”68 Interpreting Platonic βραχυλoγ…α as code for a style of argumentation along the lines of what Aristotle describes in Topics and SE has thus given us access to a technical apparatus that would describe Plato’s Socrates as occasionally sophisticseeming, but not necessarily a sophist. Naturally, we cannot expect that the generic reader of Gorgias would have been attuned to these nuances, unless perhaps they were an admirer of Prodicus. Given what is at stake in Gorgias, where either the philosopher or the rhetorician stands to be identified with the sophist, Plato poses a terminological problem requiring further study of dialectic to resolve. In any case, Aristotle’s discussion makes it clear that “say what you believe” is not a purely Socratic requirement.

References 1. Bailly, J. A. (1999). What you say, what you believe, and what you mean [Special Issue]. Ancient Philosophy, 19, 65–76. https://doi.org/10.5840/ancientphil199919Special50. 2. Barnes, J. (Ed.). (1984). Complete works of Aristotle: The revised Oxford translation (2 Vols.). Bollingen Series LXXI-2. Princeton University Press. 3. Baltussen, H. (2000). Theophrastus against the Presocratics and Plato: Peripatetic dialectic in the De Sensibus. Brill. https://doi.org/10.1163/9789004321113. 4. Baltussen, H. (2007). Dialectic in dialogue: The message of Plato’s Protagoras and Aristotle’s Topics. In E. A. Mackay (Ed.), Orality, literacy, memory in the ancient Greek and Roman world (pp. 203–225). Brill. 5. Benson, H. H. (1989). A note on eristic and the Socratic elenchus. Journal of the History of Philosophy, 27, 591–599. https://doi.org/10.1353/hph.1989.0083. 6. Benson, H. H. (2000). Socratic wisdom: The model of knowledge in Plato’s early dialogues. Oxford University Press. 7. Benson, H. H. (2002). Problems with Socratic method. In G. A. Scott (Ed.), Does Socrates have a method? (pp. 101–113). The Pennsylvania State University Press. 8. Benson, H. H. (2011). Socratic method. In D. Morrison (Ed.), Cambridge companion to Socrates (pp. 179–200). Cambridge University Press. https://doi.org/10.1017/CCOL97805218 33424.008. 67 SE

12.172b25–28; Top. 2.5.111b35–112a2. Cf. [23, pp. 47–49, 55–60] and [24, pp. 63–69, 304]. understanding tracks with the distinction that [37, p. 92] observes between “sophistic” and “sophist.” Classen [20] offers a nuanced analysis of Aristotle’s use of these terms, according to which Aristotle recognizes the benefits of analyzing sophistic argumentation, while taking a dim view of the referents of these terms when speaking generally. 68 This

154

D. Crane

9. Beversluis, J. (2000). Cross-examining Socrates: A defense of the interlocutors in Plato’s early dialogues. Cambridge University Press. 10. Bolton, R. (1990). The epistemological basis of Aristotelian dialectic. In D. Devereux & P. Pellegrin (Eds.), Biologie, Logique, Et Métaphysique Chez Aristote (pp. 57–105). Editions du Centre National de la Recherche Scientifique. 11. Bolton, R. (1993). Aristotle’s Socratic elenchus. Oxford Studies in Ancient Philosophy, 11, 121–152. 12. Bolton, R. (2012). The Aristotelian elenchus. In J. L. Fink (Ed.), The development of dialectic from Plato to Aristotle (pp. 270–295). Cambridge University Press. 13. Brickhouse, T. C., & Smith, N. D. (1994). Plato’s Socrates. Oxford University Press. 14. Brickhouse, T. C., & Smith, N. D. (2002). The Socratic elenchos? In G. A. Scott (Ed.), Does Socrates have a method? (pp. 145–157). The Pennsylvania State University Press. 15. Brunschwig, J. (1967–2007). Aristote: Topiques (2 Vols.). Éditions Les Belles Lettres. 16. Brunschwig, J. (1986). Aristotle on arguments without winners or losers. Wissenschaftskolleg, 1984(1985), 31–40. 17. Cain, R. B. (2007). The Socratic method: Plato’s use of philosophical drama. Continuum International Publishing Group. 18. Carone, G. R. (2005). Socratic rhetoric in the Gorgias. Canadian Journal of Philosophy, 35(2), 221–241. Retrieved from https://www.jstor.org/stable/40232246. 19. Castélnarac, B., & Marion, M. (2009). Arguing for inconsistency: Dialectical games in the academy. In G. P. Primiero & S. Rahman (Eds.), Acts of knowledge: History, philosophy, logic: Essays dedicated to Göran Sundholm (pp. 45–84). College Publications. 20. Classen, C. J. (1981). Aristotle’s picture of the sophists. In G. B. Kerferd (Ed.), The sophists and their legacy (pp. 7–24). Franz Steiner Verlag GMBH. 21. Cooper, J. M. (1999). Reason and emotion: Essays on ancient moral psychology and ethical theory. Princeton University Press. 22. Dodds, E. R. (1959). Plato: Gorgias. Oxford University Press. 23. Dorion, L.-A. (1990). Dialectique et éristique dans les Réfutations sophistiques, 12 et 15. Revue de philosophie ancienne, 8(1), 41–74. Retrieved from https://www.jstor.org/stable/24353889. 24. Dorion, L.-A. (1995). Les refutations sophistiques. Presses de l’Université Laval. 25. Dorion, L.-A. (1997). La ‘depersonalization’ de la dialectique chez Aristote. Archives de Philosophie, 60(4), 597–613. Retrieved from https://www.jstor.org/stable/43037598. 26. Dorion, L.-A. (2011). Aristote et l’elenchos socratique. Les etudes philosophiques, 4, 563–582. Retrieved from https://www.jstor.org/stable/41418749. 27. Dorion, L.-A. (2012). Aristotle’s definition of elenchus in the light of Plato’s Sophist. In J. L. Fink (Ed.), The development of dialectic from Plato to Aristotle (pp. 251–269). Cambridge University Press. 28. Dover, K. J. (1978). Greek homosexuality. Bloomsbury. 29. Doyle, J. (2010). The Socratic elenchus: No problem. In J. Lear & A. Oliver (Eds.), The force of the argument: Essays in honor of Timothy Smiley (pp. 82–95). Routledge. 30. Düring, I. (1968). Aristotle’s use of examples in the Topics. In G. E. L. Owen (Ed.), Aristotle on dialectic: The Topics (pp. 202–229). Oxford University Press. 31. Ferrari, G. R. F. (2010). Socrates in the Republic. In M. McPherran (Ed.), Plato’s Republic: A critical guide (pp. 11–31). Cambridge University Press. https://doi.org/10.1017/CBO978051 1763090.002. 32. Ferrari, G. R. F. (2015). Plato the writer. Epoché, 19(2), 1–13. https://doi.org/10.5840/epoche 2014121728. 33. Fink, J. L. (2012). Introduction. In J. L. Fink (Ed.), The development of dialectic from Plato to Aristotle (pp. 1–23). Cambridge University Press. 34. Forster, E. S. (Trans.). (1955). Aristotle: On Sophistical Refutations. Loeb Classical Library 400. Harvard University Press. 35. Frede, M. (1992). Plato’s arguments and the dialogue form. In J. C. Klagge & N. D. Smith (Eds.), Methods of interpreting Plato and his dialogues (pp. 201–219). Oxford University Press. 36. Gagarin, M. (Ed.). (2011). Speeches from Athenian law. University of Texas Press.

6 Platonic βραχυλoγ…α and Aristotle on Say-What-You-Believe

155

37. Galston, M. (1982). Aristotle’s dialectic, refutation, and inquiry. Dialogue, 21(1), 79–93. https:// doi.org/10.1017/S0012217300017388. 38. Gentzler, J. (1995). The sophistic cross-examination of Callicles in the Gorgias. Ancient Philosophy, 15(1), 17–43. https://doi.org/10.5840/ancientphil199515134. 39. Harrison, E. L. (1964). Was Gorgias a Sophist? Phoenix, 18(3), 183–192. https://doi.org/10. 2307/1086795. 40. Huby, P. (1962). The date of Aristotle’s T opics and the treatment of the theory of ideas. The Classical Quarterly, 12(1), 72–80. Retrieved from https://www.jstor.org/stable/638030. 41. Hudson-William, H. L. (1950). Conventional forms of debate and the Melian dialogue. American Journal of Philology, 71(2), 156–169. https://doi.org/10.2307/292194. 42. Irani, T. (2017). Plato on the value of philosophy: The art of argument in the Gorgias and Phaedrus. Cambridge University Press. https://doi.org/10.1017/9781316855621. 43. Irwin, T. (1979). Plato: Gorgias. Oxford University Press. 44. Irwin, T. (1986). Coercion and objectivity in Plato’s dialectic. Revue Internationale de Philosophie, 40(156/157, 1/2), 49–74. Retrieved from https://www.jstor.org/stable/i23946600. 45. Irwin, T. (1993). Say what you believe. Apeiron, 26(3/4), 1–16. Retrieved from https://www. jstor.org/stable/40913723. 46. Kahn, C. (1983). Drama and dialectic in Plato’s Gorgias. Oxford Studies in Ancient Philosophy, 1, 75–121. 47. Kahn, C. (1992). Vlastos’s Socrates. Phronesis, 37(2), 233–258. Retrieved from https://www. jstor.org/stable/4182412. 48. Kakkuri-Knuuttila, M. (2012). The role of the respondent in Plato and Aristotle. In J. L. Fink (Ed.), The development of dialectic from Plato to Aristotle (pp. 62–90). Cambridge University Press. 49. Kerferd, G. B. (1981). The sophistic movement. Cambridge University Press. 50. Klosko, G. (1984). The refutation of Callicles in Plato’s Gorgias. Greece & Rome, 31(2), 126–139. 51. Kraut, R. (1983). Comments on Gregory Vlastos’ ‘The Socratic elenchus.’ Oxford Studies in Ancient Philosophy, 1, 59–70. 52. Kurke, L. (2011). Aesopic conversations. Princeton University Press. 53. Lloyd, G. E. R. (1979). Magic reason and experience: Studies in the origins and development of Greek science. Hackett Publishing. 54. May, H. E. (1997). Socratic ignorance and the therapeutic aim of the elenchus. Apeiron 30(4), 37–50. Retrieved from https://www.jstor.org/stable/40913823. 55. McCabe, M. M. (2000). Plato and his predecessors: The dramatisation of reason. Cambridge University Press. https://doi.org/10.1017/CBO9780511552601. 56. McCabe, M. M. (2007). Irony in the soul: Should Plato’s Socrates be sincere? In M. Trapp (Ed.), Socrates from antiquity to the enlightenment (pp. 45–60). Routledge. 57. McCabe, M. M. (2011). ‘It goes deep with me’: Plato’s Charmides on knowledge, selfknowledge, and integrity. In C. Cordner (Ed.), Philosophy, ethics, and a common humanity: Essays in honour of Raimond Gaita (pp. 161–180). Routledge. 58. McCoy, M. (2008). Plato on the rhetoric of philosophers and sophists. Cambridge University Press. 59. Moraux, P. (1968). La joute dialectique d’après le huitième livre des Topiques. In G. E. L. Owen (Ed.), Aristotle on dialectic: The Topics (pp. 277–311). Oxford University Press. 60. Murray, J. S. (1994). Interpreting Plato on sophistic claims and the provenance of the ‘Socratic method.’ Phoenix, 48(2), 115–134. Retrieved from https://www.jstor.org/stable/1088311. 61. Nails, D. (1993). Problems with Vlastos’s Platonic developmentalism. Ancient Philosophy, 13, 273–291. 62. Nehamas, A. (1990). Eristic, antilogic, sophistic, dialectic: Plato’s demarcation of sophistry from philosophy. History of Philosophy Quarterly, 7(1), 3–16. Retrieved from https://www. jstor.org/stable/27743916. 63. Ostenfeld, E. (1996). Socratic argumentation strategies and Aristotle’s Topics and Sophistical Refutations. Méthexis, 9, 43–57. Retrieved from https://www.jstor.org/stable/43738886.

156

D. Crane

64. Polansky, R. (1985). Professor Vlastos’s analysis of Socratic elenchus. Oxford Studies in Ancient Philosophy, 3, 247–260. 65. Politis, V. (2015). The structure of inquiry in Plato’s early dialogues. Cambridge University Press. 66. Primavesi, O. (1996). Die Aristotelische Topik: Ein Interpretationmodell und seine Erprobung am Beispiel von Topik B. C. H. Beck. 67. Primavesi, O. (2010). Dialektik und Gespräch bei Aristoteles. In K. Hempfer & A. Traninger (Eds.), Der Dialog im Diskursfeld seiner Zeit: Von der Antike bis zur Aufklärung (pp. 47–74). Steiner. 68. Robinson, R. (1931). The historical background of Aristotle’s Topics VIII. In G. Ryle (Ed.), Proceedings of the 7th International Congress of Philosophy (pp. 437–442). Oxford University Press. https://doi.org/10.5840/wcp7193176. 69. Robinson, R. (1953). Plato’s earlier dialectic (2nd ed.). Oxford University Press. 70. Rowe, C. (2011). Self-examination. In D. Morrison (Ed.), Cambridge companion to Socrates (pp. 201–214). Cambridge University Press. https://doi.org/10.1017/CCOL9780521833424. 71. Ryle, G. (1966). Plato’s progress. Cambridge University Press. 72. Sedley, D., & Brown, L. (1993). Plato, ‘Theaetetus’ 145–147. Proceedings of the Aristotelian Society, Supplementary Volumes, 67, 125–151. Retrieved from https://www.jstor.org/stable/ 4106986. 73. Sidgwick, H. ([1871–1872]/1905). Lectures on the philosophy of Kant and other philosophical lectures. Macmillan. 74. Slomkowski, P. (1997). Aristotle’s Topics. Brill. https://doi.org/10.1163/9789004320994. 75. Smith, R. (1997). Aristotle: Topics books I and VIII. Clarendon Press. 76. Stauffer, D. (2006). The unity of Plato’s Gorgias: Rhetoric, justice, and the philosophic life. Cambridge University Press. https://doi.org/10.1017/CBO9780511499005. 77. Suv´ak, V. (2018). On the dialectical character of Antisthenes’ speeches Ajax and Odysseus. In A. Stavru & C. Moore (Eds.), Socrates and the Socratic dialogue (pp. 141–160). Brill. 78. Thionville, E. (1855). De la théorie des lieux communis et des principales modifications. A. Durand. Retrieved from https://babel.hathitrust.org. 79. Trapp, M. (1987). Protagoras and the great tradition. In M. Whitby, P. R. Hardie, & M. Whitby (Eds.), Homo Viator: Classical essays for John Bramble. Bristol University Press. 80. Trivigno, F. (2011). Is good tragedy possible? The argument of Plato’s Gorgias 502b–503b. Oxford Studies in Ancient Philosophy, 41, 115–138. 81. Vlastos, G. (1983). The Socratic elenchus. Oxford Studies in Ancient Philosophy, 1, 28–58. 82. Vlastos, G. (1991). Socratic studies. Cambridge University Press. 83. Vlastos, G. (1995). Socrates: Ironist and moral philosopher. Cornell University Press. 84. Wolf, S. (2010). A system of argumentation forms in Aristotle. Argumentation, 24, 19–40. https://doi.org/10.1007/s10503-009-9127-1. 85. White, N. (2011). Definition and elenchus. In G. H. Anagnostopoulos (Ed.), Socratic, platonic, and Aristotelian studies: Essays in honor of Gerasimos Santas (pp. 19–34). Springer. https:// doi.org/10.1007/978-94-007-1730-5.

Chapter 7

Rhetoric, Dialectic and Shame in Plato’s Gorgias Benoît Castelnérac

Abstract This paper deals with the philosophical purpose of the Gorgias. I argue that this dialogue, both in its form and content, yields a dramatic demonstration that the success of the Socratic inquiry depends on the character of his interlocutors and their sense of what is shameful or not. To read the Gorgias is to inquire whether Socrates’ refutations have demonstrated anything. Although there is no definition of justice, happiness or the art of rhetoric, the dialogue nevertheless shows that justice and trustworthiness concern both the practitioners of rhetoric and those of dialectic.

7.1 Introduction The Gorgias is the most substantial work in which Plato lashed out at rhetoric. At multiple times in the dialogue Socrates puts into question its status as an expertise—a technê, a technical knowledge—in persuasion.1 He would readily agree with Callicles that rhetoric is a technê, on the condition of finding one example of a real ‘knower’ in this field. Callicles proposes the names of the most famous orators in Athens (Themistocles, Cimon, Miltiades and Pericles), but the hope is vain that rhetoric will ever reach Socrates’ standard of what a technê is, if this means the rhetorician would be a master at steering the audience towards a better life (503c–504a). The Gorgias does not complete the inquiry into who a real technician of rhetoric is, and the possibility of rhetoric being an art remains, in the end, a mere possibility. This paper is an attempt to spell out the purpose of the Gorgias as a piece of philosophical writing. Other than Socrates’ attack on rhetoric and his desperate efforts to convince Polus and Callicles that rhetoric is bad, what does Socrates have to say 1 See

Gorgias’ speech at 454c–455a and Socrates’ response at 462b–463c and 464b–466a (unless specified otherwise, all references are to the Gorgias, Lamb trans.). This runs against a common idea that rhetoric is an art, along with medicine. B. Castelnérac (B) Département de philosophie et d’éthique appliquée, Université de Sherbrooke, Sherbrooke, Canada e-mail: [email protected] © Springer Nature Switzerland AG 2021 J. A. Bjelde et al. (eds.), Essays on Argumentation in Antiquity, Argumentation Library 39, https://doi.org/10.1007/978-3-030-70817-7_7

157

158

B. Castelnérac

on justice and the happy life?2 As it turns out, the play shows to what extent the success of the Socratic elenchus (refutation) depends on the moral character of the interlocutors and, more to the point, on their sense of shame. The public practise of the dialectic inquiry, just like rhetoric, can impact the feelings of the interlocutors and the audience in many ways. The Gorgias delivers an acute and dramatic illustration that the elenchus raises a series of questions about public opinion, trustworthiness and justice. That Plato’s Gorgias brings to light examples of the ‘defects’ of dialectic is a source of concern among historians of ancient philosophy. In a famous passage towards the end of the dialogue, Callicles stops playing the game of questions and answers, and lets Socrates go on with the argument by himself (from 501c to 505d). Thus, Socrates is left alone musing on justice after death, imagining how souls will be judged after the body dies. This sudden flood of metaphysics offered as a conclusion to the dialogue leaves the reader puzzled, one central question being: can dialectic operate with a partner saying the contrary of what he believes? Is telling a myth acceptable as a final argument, when the opponent has stopped debating in earnest? On that matter, the epilogue of the Gorgias stands in stark contrast with most of the other Socratic dialogues, where long speeches (and therefore myth-telling) are prohibited, and people need to willingly take part in the discussion.3 In his critique of rhetoric, Socrates argues forcefully that being in front of an audience compels the orator to please his crowd. This servile relation to the audience—thus the analogy with cooking (465b–d)—is why he compares the orator to a chef and rhetoric to a knack. The exact opposite is none other than Socrates, whose dialectic is dull, scorching and hard to swallow. As Socrates keeps repeating, he does not care about pleasing public opinion, which is why he could never be an orator. What he wants is the truth, and he thinks the tiresome test of the elenchus is the way of finding it. This test requires only short questions and answers, and starts with setting a topic for discussion with one’s interlocutor, e.g., ‘Is justice good or bad?’ The roles are then distributed, one ‘player’ being the defendant of the proposition that justice is either good or bad and the other, the opponent, trying to ‘refute’ the defendant through a series of questions.4 The defendant tries to answer all the questions without making

2 “Why

should there be so long a dialogue with such recalcitrant opponents? Socrates silences all three, but he seems to have convinced none of them” ([1], p. 5). In a nutshell, this is the problem any sceptical reading of the dialogue must confront. 3 The Crito finishes in likewise manner: Socrates ends his discussion with Crito in a hypothetical debate with “the Laws” personified, who maintain that Socrates must drink the hemlock if he does not want to betray them (Crito 50a–54d). Another mute character is Philebus, in the eponymous dialogue. These passages should also be connected with what Socrates says to a speechless Meletos in the Apology (27c); “I take it that you agree, since you do not answer”. 4 For a more detailed presentation of dialectical games in the late IVth c. B. C. see Castelnérac and Marion in [4].

7 Rhetoric, Dialectic and Shame in Plato’s Gorgias

159

any mistakes in his argumentation.5 The opponent tries to fault the defendant’s position by showing there is something unsatisfactory in it; the opponent’s aim is to ‘gather up’, from the answers of the defendant, a demonstration a contrario that there is something unconvincing in the defendant’s position.6 Plato offered many different examples of this procedure. The Gorgias focuses on one particular problem with the practice of the elenchus: it sometimes brings shame on the interlocutors. This points to one of its most serious ‘defects’. As historians have noted, a respondent might feel ashamed at failing the test of the elenchus.7 Socrates’ inquiry focused on questions about virtue that he put to specialists claiming to know, as specialists, what is beauty, courage, piety, etc.8 It would be hard to understand why Socrates was put to death for the crime of refuting everyone all day long, if his ‘way of living’—continuous refutations—had not infuriated and brought shame to some of the notorious people who failed the test in front of their ‘lovers’. In Plato’s narrative, Socrates’ trial and condemnation were consequences of the public pressure the elenchus put on people in front of him. As a writer of Socratic dialogues, Plato could not ignore the topic of dialectic and shame. As I will highlight further down, this thread is tightly knitted in the dense fabric of the Gorgias. Public opinion weighs a lot in the dramatic setting of the Gorgias.9 Even if Socrates spells out as a rule of conduct that, in the procedure of the refutation, one should not use any other witnesses than himself and his interlocutor,10 shame, or the lack of it, persistently creeps back in the dialogue either as a topic of discussion or to qualify a character’s behaviour. The main reason for this is that argumentation, either rhetoric or dialectic, is in some way morally driven (as this analysis of the Gorgias will show).11 The Gorgias spells out a major point of contention because the pressure coming from the interlocutors and the bystanders is equally palpable in rhetoric and dialectic, and the question of trustworthiness is a major issue for both. In terms of dialectic, to play the game one needs to know what constitutes a valid move, but there 5 See

Symposium where Diotima refutes Socrates (201d–202e). Under specific conditions like the just and almost perfect city of the Republic, passing through the elenchus and not making any mistake counts as a proof of knowledge (VII 534b–535a). 6 In the Gorgias Plato uses syllogizomai to describe the “syllogismos” achieving the elenchus; it means “gathering up” the premises that lead to a contradiction (i.e. a deduction that something is inconsistent, incomplete, ambiguous, etc. in the premises agreed during the exchange). Among Plato’s dialogue of his so-called early career, the Gorgias stands out as a place where technical terms that are used in Aristotle’s Prior Analytica abound. Crubellier (in [5]) first noted the presence of syllogizomai (“to syllogize”: to gather together) in the Gorgias. The parallel with Aristotle’s vocabulary extends to other words used in the Gorgias (homologomenon, sumbainein, perainein, antiphasis, etc.; see Castelnérac, in [3], p. 316). 7 See Lesher (in [6], pp. 21–23). 8 See Katja Maria Vogt (in [15], pp. 27–50). 9 See Callicles, at 486d where doxa means “the opinion which others have of one, estimation, repute” (LSJ, scil. δ´oξα [III], p. 444). 10 471e–472c; 474a; 475e–476a. 11 The themes of rhetoric and morality are not two but one ([1], p. 6). For Benardete, it was a way of showing that rhetoric should be excluded from justice (in [1], p. 7). I disagree on this point, the Gorgias proposes a reflexion on justice and speaking in public in general, including dialectic.

160

B. Castelnérac

is also the question of what constitutes a fair player. So, the Gorgias raises questions of this type: is it fair to accuse Socrates of foul play at Gorgias’ expense, as Polus does? Why spend time arguing with someone else about justice if it is preferable to win at all costs? And if someone confesses not feeling any shame while saying the opposite of what he thinks, how is this person ever going to be trusted or accept defeat? I will start with a summary of the dialogue, paying special attention to what it says about rhetoric, dialectic and shame, and then propose a reading of the dialogue with these points in mind.12

7.2 Gorgias, Polus and Callicles The play starts with a very hospitable tone. Everyone is thrilled to hear a discussion between Socrates and Gorgias (458c–d), and this is so, even though Socrates is critical of rhetoric from the start. Socrates cannot understand what sort of knowledge it is; it remains a mystery for him (456a). Still, Gorgias wins the first match (up to 457c), and makes the point that rhetoric has a power, a dunamis of its own.13 It prevails over assemblies and against the speech of any other specialist untrained in the art of rhetoric. As Gorgias explains, if two physicians were to be pitted against one another in front of an assembly, the one trained in rhetoric would win the argument (456b–c). Thus, rhetoric has some kind of ‘power’. Socrates and Gorgias agree to have another go at defining rhetoric. During the second match Socrates finds, at last, a contradiction in Gorgias’ answers. His interlocutor had agreed at the end of the first exchange that someone trained in rhetoric could be unjust. This makes sense: as is the case with any other dunamis, rhetoric can serve various purposes, both good and bad. Yet in the next exchange Gorgias claims to be able to teach justice along with rhetoric.14 Gorgias thus appears to be inconsistent: he first admitted that rhetoric was an art that could be used against justice, and he seems now committed to teaching justice with it. The exchange is abruptly cut short. Polus is outraged that Socrates finds fault with his master, and becomes Socrates’ opponent. Gorgias and Socrates will never have the chance to get another go at defining the power of rhetoric. Let us pause here to note that shame is already present in the opening section of the dialogue. First, Socrates asks Gorgias what is his technê, invoking as a pretext that some of his pupils might be ashamed to ask (455c). Secondly, two pages later, just before starting to refute Gorgias’ claim to know what justice is, Socrates tells 12 Although

I offer a general interpretation of the entire dialogue, I do not have the space to discuss every other interpretation of the dialogue. I rather focus on asking what is the philosophical relevance of the Gorgias: what does it demonstrate, if anything? This explains why I mostly restrict myself to works discussing the ‘Socratic elenchus’ (see Sects. 7.3 and 7.4). 13 456c: “this, Socrates, is the potency [dunamis] of rhetoric, and its art [technê]”. 14 See 460c–461b, along with 459c–460 and 456a–457c.

7 Rhetoric, Dialectic and Shame in Plato’s Gorgias

161

everyone that his request for a definition of rhetoric is not driven by any ‘love of victory’, because this philonikia often causes shameful behaviour on the part of the players, and brings about “abusive expressions on each side”, which even bystanders are angry to spend time listening to (457d–e). Thirdly, the last mention of shame comes from Gorgias, who says he should be ashamed of himself if he refused to answer Socrates’ questions, especially because he boasts being a master at brachylogia (‘short speech’, i.e., an exchange of questions and answers as short as possible).15 About the first and third mention of shame, there is nothing intrinsically wrong in being ashamed to ask a question, or in Gorgias’ desire to avoid shame, which is consistent with his stature as a professional orator. Philonikia, the second cause of shame—i.e., the will to win at all costs—foreshadows what will unfold next with Polus, in the second part of the dialogue (461b–481). The second part is at once lighter and darker than the first. The argumentation is getting worse, mostly for two reasons. Firstly, Polus makes it a matter of pride. Pursuing philonikia, he wants to beat Socrates at his own play, and exact revenge on behalf of his master. Secondly, he tries to champion Gorgias with a fatal flaw in his attacks on Socrates. He makes time and again the error of thinking that it is the audience he must convince, not Socrates. To exonerate his teacher, Polus tells Socrates that Gorgias “was ashamed to say he could not teach justice” (461b), and scorns Socrates saying “it is a sign of very bad taste to lead the discussion into these arguments”.16 Socrates agrees that telling the truth about rhetoric might be “too rude”,17 since people might think he is making fun of Gorgias. But Gorgias steps in and urges him to speak freely, with no shame in front of him.18 Socrates suggests next that one is better off suffering from injustice than imposing it on someone else. Polus thinks he can defeat Socrates easily because this goes flatly against common sense. Polos bets on the locus communis that tyrants, although unjust, lead a happy life. Lots of unjust men are happy (470e), and he praises the tyrant Archelaos, a classical example of a happy unjust man. He then tries to ridicule Socrates’ argument by pushing it to its ultimate consequences: If a man be caught criminally plotting to make himself a despot, and he be straightway put on the rack and castrated and have his eyes burnt out, and after suffering himself, and seeing inflicted on his wife and children, a number of grievous torments of every kind, he be finally crucified or burnt in a coat of pitch, will he be happier than if he escape and make himself despot, and pass his life as the ruler in his city, doing whatever he likes, and envied and congratulated by the citizens and the foreigners besides? […] And this, according to you, is impossible to refute at all? (473c–d)

Socrates is careful in saying that none of the two horns of the dilemma—suffer injustice or get away with it—is enviable (473d–e), but he holds that the latter is 15 See

458d, following up on Gorgias’ contention that he can answer any possible question, 447d– 448a. 16 461c3–4; see Socrates’ answer at 462e–463a. 17 agroikoteron, 462e; see also 509a. On this particular word in the Gorgias see Michelini (in [8]). 18 mêden eme aiskhunthêis, 463a.

162

B. Castelnérac

worse (469b–c, 472d–473a). Polus insists on saying that it is really strange that an argument nobody would hold could be considered irrefutable (473e). This makes Socrates the laughing stock of Polus, as the old man points out immediately: “Oh Polos, I am not a public man, and only last year when my tribe was serving as prytanes, and it became my duty as the president to take the votes, there was a laugh at me, because I did not know the procedure” (473e–474a). Yet, there is one thing Polus cannot object to Socrates: one should feel ashamed to commit an injustice, not to suffer from one (474c–475d; 477c–d). Polus is not good at dialectic, but not stupid either: he comes to realize he is stuck with the weakest position in the eyes of the audience. His advice on justice is frankly disturbing if the tyrant Archelaos is his model of the happy life. In the end, although he thinks Socrates’ position is absurd, he admits he has to agree with him (480e).19 Here, we can see plainly that, although Polus concedes defeat, the outcome of the game is in sharp contrast with his set of beliefs. In the last part of the play the tension reaches its peak with an overly self-confident Callicles mad at Socrates taking up the challenge (481b–482e). He pretends to show that it is better to be unjust than being unjustly accused no matter what. Callicles prides himself in speaking with no shame (486c). Since, for him, true justice is dictated by nature, not by convention, trying to satisfy one’s appetite at the expense of others is not being unjust (491e–492c). After a set of two long speeches replying to one another (482c–486d, 492e–493d) and a series of short questions and answers (489b–492c, 495a–500a), Callicles is caught up in a contradiction, and ends up complacently answering Socrates’ questions, as a favour to Gorgias to “complete the argument”.20 This part—the climax of the play—is rife with mentions of shame and the lack thereof. Callicles first accuses Socrates of having used shame against his first two opponents (482c–d), an accusation that Socrates takes into consideration twice (487b, 494d). As a consequence, Socrates asks Callicles to speak with no shame (487d–e, 494d). As for Callicles, he thinks it is shameful to indulge in such inquiries. He ridicules philosophers for behaving in an unmanly way (485b–c) and tells Socrates he should be ashamed of his refutations (489b, 491a). He dismisses this practice altogether as a pure waste of time and, although in silence, he sticks to his initial position until the end. The issue of fairness and justice is closely tied with the dramatic setting of the dialogue. All this talk about shame is not accidental. Gorgias’ fame and status, Socrates’ lack of manners, Polus’ poor strategy and Callicles’ bad temper all revolve around the same philosophical issue. If, according to Plato, humans do rely on doxa, i.e., on ‘public opinion’, judging others and being judged by them, he must then accept as a consequence that, in some situations—for instance, a debate on rhetoric with rhetoricians—our sense of what is shameful or not can interfere in our way of handling the debate. Whether they do not want to lose the game by any means or they suddenly realize they have lost the exchange, dialectical players are liable 19 His 20 See

position does not merely appear to be paradoxical ([1], p. 58), it is blatantly so. 501c and 505d.

7 Rhetoric, Dialectic and Shame in Plato’s Gorgias

163

to feel shame in front of the audience. As the following sections hope to show, in some situations like the one depicted in the Gorgias, dialectic and the practice of the elenchus is not an infallible method to reach the truth, since it is not immune to outside tampering.

7.3 Socrates’ Theses Before articulating how I think the Gorgias tackles this problem, I must deal with the issue of exactly what Socrates is attempting to show in the course of the dialogue. If the Gorgias had any direct demonstration to propose to its readers— either showing that rhetoric is bad or providing a definition of justice—this could downplay the philosophical relevance of all the drama going on around it.21 On the contrary, if the dialogue does not end with a clear definition of either rhetoric, justice or the good life, it could be that its message lies somewhere else. The dialogue starts by asking what rhetoric is. Gorgias says it is efficient for persuading large amounts of people, while Socrates criticizes it as a knack. Soon enough, the discussion veers towards the examination of Socrates’ soul and the search for the best life (486d–488b). Socrates is not persuaded that rhetoric can teach what the just life is. This is a standard turn of events when Socrates is making an inquiry. The Gorgias resembles the other ‘Socratic dialogues’, because, most of the time, Socrates goes on and on in his quest for knowledge and ends up unsatisfied by the answers he finds. He always starts with benign questions of expertise, then it turns out that his interlocutor and his life are the real subject of inquiry.22 This is how in the course of the Gorgias Socrates’ position is criticized (461b–c; 481b–c) and the roles are inverted. Now, it is Socrates who has to defend himself against the attacks of Polus and Callicles. In this respect, this dialogue is akin to the (‘demonstrative’) Republic, where Socrates starts in the role of the opponent, and ends up playing the defendant for the rest of the dialogue.23 In the Gorgias, Socrates defends the stunning and bluntly counterintuitive thesis that any unjust and unpunished person is worse off than the one suffering from injustice. Even though Polus (473b–474b) and Callicles disagree with him (508d–509c)—and still disagree after the exchange is over—Socrates makes it plain that he can defend his thesis on the basis of his previous experience. The problem with reading the Gorgias as a treatise against rhetoric is that nowhere in the dialogue is it secured that rhetoric inevitably leads to the unjust life. Surely, Socrates believes rhetoric is a knack comparable to cooking (464b–465d). He is convinced it is bad (463d), but he is equally aware that he has come short of 21 The

best known candidate for a dogmatic reading of the Gorgias is Vlastos: Socrates’ elenchi “articulate intuitions which prove practically viable in his own experience; they tell him who is happy and who isn’t; he does what they tell him and he is happy” [14, p. 26]. This optimistic (and partially esoteric) reading is hard to endorse as I will show in this section and the next. 22 See Laches, 187d–188a. 23 His position is challenged in the opening of books II and VI of the Republic.

164

B. Castelnérac

convincing Gorgias and Polus he is right. It is not because rhetoric often lets people get away with injustice that it is irremediably bad. The possibility of finding a good rhetorician or a good purpose to rhetoric is entertained in the Gorgias (503c–504a). In other dialogues, Plato expanded more on the idea of what rhetoric would be if it were to be good. According to the Phaedrus, it should be based on knowledge of the soul,24 in the Laws it comes under the form of preambles to the jurisdiction of the land.25

7.4 Challenging the Dogmatic Reading If Socrates does not show that rhetoric is bound to be bad, the other option for a dogmatic reading is a definitive demonstration that suffering from injustice is better than being unjust. As this section will show, although Socrates is confident that he can win his debate against Callicles, on his view this does not mean he knows anything. Let us read Gorgias 508d–509a, where Socrates sums up what conclusion he reached after discussing with Callicles. I hold that to strike or cut me or mine wrongfully is yet more of a disgrace and an evil, and likewise stealing and kidnapping and housebreaking, and in short any wrong whatsoever done to me or mine, are both worse and more shameful (aiskhiston) to the wrongdoer than to me the wronged. All this, which has been made evident in the form I have stated some way back in our foregoing discussion, is held firm and fastened — if I may put it rather bluntly — with reasons of steel and adamant (so it would seem, at least, on the face of it) which you or somebody more vigorous26 than yourself must undo, otherwise it is not possible to speak nicely, saying something else than I just said. For my story is ever the same, that I cannot tell how the matter stands, and yet of all whom I have encountered, before as now, no one has been able to state it otherwise without making himself ridiculous. (Lamb transl. mod.)

In my view, this passage is pivotal mainly for two reasons. First, because it is a striking representation of what Plato’s Socrates thinks he can best achieve with a refutation. Secondly, because it is the place where Socrates recapitulates what he has achieved in the last stretch of the dialogue.27 24 Phaedrus

261a–b. IV, 772e–773e; IX, 854a–b. 26 neaniskoteros (younger, having more vigour). It echoes what Callicles’ said about Socrates: “roistering recklessly” (neanieusthai, 482c; Lamb transl.). Polus is young but not that neaniskos (“powerful”), otherwise he would have won his exchange with Socrates, who starts his exchange with Polus on a sarcastic tone: “Oh, beautiful Polus, it is indeed very handy that we have friends and sons, that when we get old and stumble, you younger ones may be around to reform our lives in words as well as deeds” (461c5–8; see Polus’ remark at 470c). Again, Polus does not match his own expectations. 27 These lines and the place of the Gorgias in Plato’s works are a turning point in Vlastos’ reconstruction of Plato’s intellectual life (for his last take on the elenchus, see Vlastos, [14], pp. 1–37). His understanding of the elenchus, especially in the Gorgias, has been met with scepticism, most notably by Benson ([2], Chaps. 3 and 4). For a summary of this debate, see Scott ([12], p. 4). For an alternate reading to Vlastos and Benson, see Penner (in [10]). 25 Laws

7 Rhetoric, Dialectic and Shame in Plato’s Gorgias

165

Socrates seems to be contradicting himself when he says he has “reasons of steel and adamant” to say he is right, but “cannot tell how the matter stands”.28 The contradiction is aggravated by his own reference to what he said before. If this passage goes back—as Vlastos and numerous others have pointed out—to two other passages where Socrates says that “Truth cannot be refuted”29 or “Really, you and me agreeing will be the achievement of truth”,30 then he implies that what he will say is true with no other qualifications needed: all things considered, it is better to suffer than to commit injustice. Thus, his refutations of Polus and Callicles are “reasons of steel and adamant”, i.e., they should count as “strongly bound” and “unyielding”31 arguments that he is right. Yet, in the same breath he maintains that he does not know where the matter stands. How to make sense of Socrates’ contradictory talk here?32 This passage yields decisive information about Socrates’ understanding of truth and proof. For Plato’s Socrates, the procedure of the elenchus is the best way of inquiring into something. In the context of our passage, that amounts to saying that something could be considered true as long as there is no refutation available for it. This is summed up in two lines quoted above: “somebody more vigorous than yourself [scil. Callicles] must undo [my arguments], otherwise it is not possible to speak nicely, saying something else than I just said”. This is how the game goes: refute your opponent or remain silent. This should strike one as implementing a very low standard of truth, but it seems that Plato did think of truth and falsehood in this way: proving something true or false is based on a test, but the test in itself is no guarantee that we have grasped something intrinsically true. Even if Socrates says that “truth has never been refuted” (473b), Plato did not state the converse, that “something not refuted is true” in any universal and definitive sense of the term. All of Plato’s early ‘Socratic’ dialogues reflect Socrates’ ignorance and offer vivid examples that, while the elenchus does not offer a straight road to truth, it is still the most potent test available. A quick review of what the ‘mature’ Plato said on the elenchus assures us that he still thought of it as the best test for truth. While Plato points out in the Parmenides that Zeno does not think he has proved Parmenides’ theses true against his critics—he has merely shown that their theories were more ridiculous than the ones of his friend and master33 ; this mirrors Socrates’ position in 28 The contradiction just underlined must be kept in mind when considering Vlastos’ hesitation about the Socratic elenchus. Vlastos explains how he changed views, veering from a sceptical towards a “dogmatic approach”. The late Vlastos maintained that Socrates is “ironical” when he says he does not know anything, since he has “true beliefs”. See for instance the introduction to Vlastos (in [13]). 29 470c; 473b, with Polus. 30 487e with Callicles. 31 The strength of the proof is of “iron” (sidêron), and “adamant” (diamond?), the latter being a transcription of (the Homeric), a-damas, literally “that cannot be bent; not giving away”. 32 Vlastos calls “true” Socrates’ undefeated proposition. It is his way of solving the contradiction, i.e. “a systematically dual use of his words for knowing […] where Socrates’ claim to know p is simply the claim that p is elenctically viable, i.e., if he were to pit it against its contradictory in elenctic argument it would prevail” [14, p. 67]. This is mostly right, although Socrates never says he knows p, or that p is true all other things considered. 33 Parmenides 128b–e.

166

B. Castelnérac

front of Polus and Callicles—it is in this dialogue that Plato’s own ‘Theory of Ideas’ is fully tested through Parmenides’ elenchi.34 The Republic eschews a program of scientific knowledge ending up in a dialectical exchange to determine if the philosopher knows what is the ‘Good’.35 And in one of Plato’s latest works, the Timaeus, scientific arguments (logoi) must be as irrefutable as possible.36 Accordingly, when Timaeus spells out his theory about the shape of the basic triangles, he reminds his audience that if someone else would speak better and refute his theory, he would welcome this person as a friend, not as a foe.37 Another important thing to remember about the discussions with Polus and Callicles on the ‘unjust happy life’ is that they are based on a first dilemma one could wish to reject. Socrates stresses that the unjust person and the one suffering from injustice are both in a bad position. When confronting two adverse and questionable positions, the elenchus can only evaluate which position is least defendable (or the more problematic), not which position is ‘true’ or ‘false’, and ‘just’ or ‘unjust’ simpliciter. Thus, the discussion is not about who is living a happier life, but who is the more wretched (473d). Socrates makes plain that none of these situations is desirable, i.e., neither is in the smallest way happy. It is only if he had no other choice that he would rather suffer from injustice than be unjust (469c). To go against his point of view, someone like Callicles needed to vindicate injustice.

7.5 Socrates’ Inconclusive Elenchus This being said, why would Socrates be so confident that he is right, that the wronged person is better off than the one committing injustice? The best reason he can give is the observation that no one, placed in front of this dilemma, has been able to argue the contrary of his position “without making himself ridiculous”. Had Socrates discussed this problem one time only, he would probably not have had the same confidence that he could win. This points to the fact that, as a process of inquiry, the elenchus needs to be reiterated. Socrates, who spent all his time inquiring in this manner, was the living example of the idea that it does not take a single elenchus to set the matter straight about the topics he was inquiring into. See for example the exchange with Gorgias, at the end of which a victorious Socrates tells Gorgias “it would require no short sitting to distinguish properly if it is possible or not for the rhetorician to use his rhetoric unjustly or consent to do wrong” (461a–b). Although he caught Gorgias in a contradiction, this does not mean he knows whether the rhetorician can be unjust or not. To Socrates, Gorgias’ answers are a sign that there is something further to

34 Parmenides

128e–135c. VII 534b–535a. See above note 5, p. 002. 36 Timaeus 29b. 37 Timaeus 54a–b. 35 Republic

7 Rhetoric, Dialectic and Shame in Plato’s Gorgias

167

inquire about. On his part, the orator is eager to know what the philosopher has to say on rhetoric and political power (463e). The elenchus is an open-end process that should be driven by the desire to know. In his whereabouts Socrates defends his position that the unjust person is worse off than the victim of an injustice. He keeps score, probably over many years of practice. Yet, he is aware that no game is decisive, and refrains from making the final induction that he is right simpliciter when he beats his opponent (especially when discussions derail as in the Gorgias). Although he is confident he can defeat his opponents on a particular question, Socrates does not think he has reached the definitive truth with regards to the issue discussed in the Gorgias. This explains why, even if he boasts so much in front of Polus and Callicles about winning the argument, he still reminds everyone that he cannot tell “how the matter stands”; to wit, if rhetoric is bad, or what is a happy life in terms of injustice. If the Gorgias exposes merely good—not decisive—reasons to think that rhetoric is bad and suffering from an injustice is better than being unjust, then it could be argued that this dialogue is a pamphlet against rhetoric and injustice disguised as a fake triumph of dialectic over rhetoric.38 The conclusion of the dialogue is even more outrageous if one believes that Socrates did win against his three opponents by shaming them into silence. If dialectic relied on the audience’s sense of what is shameful or not to silence one’s opponent, then this practice could be criticized for using public shame as much as rhetoric is criticized for relying on public gratification. So, if the Socratic exchanges are to derail into ad hominem arguments—as is the case in the final part, with Callicles—39 what sort of guarantee can Socrates have that the elenchus will lead to any positive result? This is, I think, the real philosophical issue of the dialogue. It is closely related to two elements of the practice of the elenchus. First, the elenchus requires free speech, which entails ‘saying what one believes’. Secondly, it requires some respect for justice as an intrinsic value of moral agents: is there any genuine sense of shame in moral agents, or are they just refraining from doing injustice for fear of the possible consequences of being found guilty?

38 This is a worry both Vlastos and Benardete have to face, although for opposite reasons. For the former, Socrates calls “true” his undefeated theses, with the possibility of “true” having two different meanings (see above, note 32 p. 165): Socrates does not have decisive reasons to press his case, yet the middle premises granted to Socrates will always end up rejecting the adverse thesis. Benardete is more pessimistic: “The triumphs of Socratic reasoning only go to show the impotence of reason. We thus seem forced to choose between a dialectic that alters no one’s convictions and a rhetoric that is effective but knows neither how it is effective nor what it effects. Rationality is empty; rhetoric is blind” ([1], p. 13). 39 Socrates calls Callicles a “lover of the mob” (481d–e). For Callicles, practitioners of dialectic behave “in an unmanly way” (492b) and he calls Socrates a “demagogue” (482c), a “sophist” (483a) and an abuser of dialectic (487b).

168

B. Castelnérac

7.6 Elenchus and Free Speech The Gorgias makes clear that, especially in Athens, free speech was not a problem.40 In terms of what the players needed to keep in mind, ‘free speech’ meant ‘saying what they believed’.41 But Polus and Callicles criticize Socrates, because (so they say) Socrates won the first two games by pressing his opponents to avoid a shameful position. Since Gorgias and Polus were ashamed to say what they believed, they did not speak freely, and no one should rely on the outcome of the discussion—Socrates did not truly win the exchanges. The refutation is non-demonstrative because Polus and Gorgias did not speak their mind for fear of having a bad reputation. This is the origin of Socrates’ bad name: free speech and speaking one’s mind have direct consequences on someone’s reputation. Dialecticians be warned: on some issues, the elenchus implies crowd management and can have dire consequences for its practitioners. Thus, the following question arises: to what degree does shame play a decisive role in the elenchus? According to Callicles it matters a lot since, “if a man is ashamed and dares not say what he thinks, he is forced to contradict himself” (482e–483a). This is a trivial point. It happens all the time when someone is inconsistent because they lied. It is also a reminder that Socrates and Callicles agree at least on one point: speaking freely and putting shame aside is an advantage for the inquiry. Contrary to Polus and Gorgias, Callicles will thus speak with no shame at all, and Socrates is uniquely optimistic that their exchange will yield some result. But, as pointed out earlier, after Socrates caught him in a contradiction (497a), Callicles lost interest in the exchange. “I agree with you,” he says, “in order that your argument may reach a conclusion, and to be pleasing to Gorgias here”.42 At this point of the play, Callicles does not seem to be ashamed of saying what he does not believe, thus breaking his own ‘honesty indictment’ of saying what he believes. It is a hard blow for Callicles.

7.7 Elenchus and Fairness It is important first to underline that dikaios—‘just’—is the adjective used in the Gorgias to describe a valid move in the debate.43 The plot of the Gorgias gives a clear indication that Plato reflected on the topic of what constitutes a fair player. Recall that rhetoric and the elenchus use two different types of ‘proof’, respectively the 40 Recall the invitation of Gorgias to speak freely in front of him. Polus does not understand why he could not say what he wants in Athens, the city of free speech (461d–e). 41 See Vlastos (in [14], pp. 13–17), Benson (in [2], p. 70), and Castelnérac and Marion (in [4], pp. 55–56) among others. 42 501c, Lamb trans. mod. 43 In 465e–466a, Socrates’ long speech is a ‘fair’ move, because Polos, apparently untrained in dialectic (461e–462a and 471e–472b), would not understand short speech. See also 453c; 454b kata dikaion logon, 504e; 522b–c. The same wording occurs in the Apology (18a, 20d).

7 Rhetoric, Dialectic and Shame in Plato’s Gorgias

169

approval of the crowd and the mutual agreement of the players. The free and honest participation of all is what Socrates wishes the elenchus to be, but the Gorgias shows that it is a condition which is hard to achieve. How can someone hope to inquire in any matter by means of the elenchus after realizing that its simple setting—two players, and a crowd of listeners—implies conditions sometimes going against a successful inquiry? What the Gorgias says explicitly is that, if the elenchus is confused with a shaming contest, then someone might either try to win by whatever means, and even use free speech to hurl insults—genuine ad hominem replies—at his opponents.44 On this point, the Gorgias spells out some rule of personal conduct. For the sake of learning something, not of winning: restraint and judgement are inherent to the behaviour of the fair dialectician and free speech must be used in a benevolent manner. This is why, for Socrates, the greatest reward is when he is reminded how ignorant he is. On this point, his position is egoistic: he prefers being refuted than refuting someone else (458a). Because he is so keen in his search for the truth, his disposition to learn undercuts the problem of wining at all costs and any sense of shame. The love of victory or any abuse of shame disappears since Socrates has already agreed he is ignorant. This should not mean he is not aware of how much public opinion can hinder the practice of refutations, since it often brings shame onto his opponent or some of the bystanders.45 Inevitably, the next question is this: Did Socrates make any use of shame to silence his opponents in the Gorgias? As recapitulated earlier, contrary to what Polus says, shame was invoked in Socrates’ dialogue with Gorgias rather as a way of getting started and setting the discussion on a good course (to avoid philonikia), certainly not to pressure him.46 Against Polus and Callicles, his exchanges mostly hinged on the question whether it is better to get away with injustice or be unjustly punished, with Socrates arguing that getting away with injustice is shameful. How far can he be confident that it is shameful to get away with injustice? To the extent that—and this is a vital point, both for rhetoric and dialectic—if someone was indeed feeling no shame while getting away with injustice, no one would trust what this person is saying. Polus seems to have understood this when he tacitly agreed that injustice brings shame. In his exchange with Callicles, Socrates makes plain that some ‘infinite pleasures’ can be bad and ugly (494c–495a),47 while his opponent goes against the injunction of ‘saying what he believes’. If justice is not simply an empty convention, but a set of rules one can only infringe at the cost of severe consequences, then you would not be likely to take advantage of your audience merely to get away with victory. If, on the contrary (like, say, Antiphon the sophist) you think that human justice (vs natural justice) is only a matter of convention within a community; that human justice has no real consequences except when one gets caught, then you are very likely to see 44 See

above, note 39, p. 167. See Plato’s Apology 22e–23a among others passages. 46 See above, Sect. 7.2. 47 See also the exchange 499b–d, where Socrates shows that some pleasures are better than others (Moravczik gives a clear summary of the argument, in [9], pp. 125–126). 45 457d.

170

B. Castelnérac

no problem in using demagogic arguments or other tricks of the trade, either in a philosophical dispute or in a trial. This reveals that Callicles might not be the best example of a ‘fair’ player, because of his opinion that justice is but a convention.

7.8 A Myth About Real Justice This is where the drama and content of the Gorgias bind together. The Gorgias deals with injustice and public speech; the characters and their arguments illustrate how far justice and fair argumentation should concern the orator and the Socratic enquirer. It delivers a tangible observation on speech and public opinion: a sound management of shame and reciprocal fairness are key elements to benefit from, when speaking in public. Although they are at odds on so many other issues, rhetoric and dialectic operate on a common base of justice and trustworthiness. This point leads to Socrates’ defence of the wronged man. It is based on the sense of guilt one ought to feel after doing something unjust. He reminds the audience until the very end, that talking against his position will inevitably bring shame on ‘the defendant’.48 Callicles challenges this point in words and deeds: this dialogue makes abundantly clear that without a desire for justice and an aversion for injustice, the elenchus would have no chance to be the means of inquiry Socrates is so confident in using; even more so if the debate is ‘widely broadcasted’. The problêma ‘Is justice real or not?’ will trigger the plot of the Republic—do we, humans, have a real conception of justice that brings about shame when we act against the idea of what is just or not? Is shame only a consequence of the weight social conventions put on us, especially when we appear in front of others? What is the purpose of the final myth considered from this point of view? There Socrates imagines that the souls of humans will be judged after death; the gods pay special attention that no crime remain unpunished (523b–524a). Here are two different reasons why Plato might have thought purposeful to end the dialogue with a myth. The first is Callicles’ final attitude. After discussion through short questions and answers (brachylogia) ended up in an impasse, long speech (makrologia) is the last option for Socrates to express his thoughts. That is to say, rhetoric starts where dialectic ends. When no one is willing to argue with Socrates, long speech is the only means of exposition that remains in order to complete the dialogue. It does not show that dialectic in itself is to be abandoned, but when some specific conditions for dialectical inquiry are lacking, rhetoric will indeed ‘put an end’ to dialectic. The eschatology of the Gorgias also expands on the results of previous discussions with Polus and Callicles. If there is a strong reason why it is better to suffer injustice than to escape justice while being unjust, it could be because justice is not a matter of convention among humans, but stems from a genuine sense of shame. From there, Socrates hypothesizes that justice might be real in a strong, realistic, sense of the term 48 In

508d, see above, Sects. 7.4 and 7.5.

7 Rhetoric, Dialectic and Shame in Plato’s Gorgias

171

(522e). His concluding myth implies that no one escapes justice, even after death. Hence this story of the ultimate trial of the soul, stripped of all disguise after the body has died (523c–e).49 It might seem ironical coming from Socrates, the champion of rational thinking and ‘elenctic’ uses of dialectic, but I hope to have made clear that this final myth is a sufficient reason to think that the Gorgias is not simply an attack on rhetoric. It is also the manifestation that the practice of the elenchus is dependent on at least two things, public opinion, and how far one believes that justice is real and worthy of being pursued.50

References 1. Benardete, S. (2009). The rhetoric of morality and philosophy: Plato’s Gorgias and Phaedrus. University of Chicago Press. 2. Benson, H. H. (2000). Socratic wisdom: The model of knowledge in Plato’s early dialogues. Oxford University Press. 3. Castelnérac, B. (2015). Impossibility in the prior analytics and Plato’s dialectic. History and Philosophy of Logic, 36(4), 303–320. 4. Castelnérac, B., & Marion, M. (2009). Arguing for inconsistency, dialectical games in the academy. In G. Primiero & S. Rahman (Eds.), Acts of knowledge: History, philosophy and logic, Essays Dedicated to Göran Sundholm (pp. 35–71). College Publication. 5. Crubellier, M. (2011). Du sullogismos au syllogisme. Revue philosophique de la France et de l’étranger, 136(1), 17–36. 6. Lesher, J. H. (2002). Parmenidean elenchos. In G. A. Scott (Eds.), Does Socrates have a method? Rethinking the elenchus in Plato’s dialogues and beyond (pp. 19–35). Pennsylvania State University Press. 7. Liddell, H. G., Scott, R., & Jones, H. S. (1996). A Greek-English Lexicon. Clarendon Press. 8. Michelini, A. N. (1998). Pollê agroikia. Cl. Ph., 93, 50–59. 9. Moravczik, J. (2007). Goodness Trumps pleasure-loving in the Gorgias. In M. Erler & L. Brisson (Eds.), Gorgias-Meno: Selected papers from the Seventh Symposium Platonicum (pp. 122–127). Academia Verlag. 10. Penner, T. (2007). The death of the so-called ‘Socratic elenchus’. In M. Erler & L. Brisson (Eds.), Gorgias-Meno: Selected papers from the Seventh Symposium Platonicum (pp. 3–19). Academia Verlag. 11. Plato. (1967). Plato in twelve volumes (Vol. 3, W. R. M. Lamb, Trans.). Harvard University Press; William Heinemann Ltd. 12. Scott, G. A. (2002). (Ed.). Does socrates have a method? Rethinking the elenchus in Plato’s dialogues and beyond. University Park [PA]; Pennsylvania State University Press. 13. Vlastos, G. (1991). Socrates, ironist and moral philosopher. Cambridge University Press. 14. Vlastos, G. (1994). Socratic studies. Cambridge University Press. 15. Vogt, K. M. (2012). Belief and truth: A skeptic reading of Plato. Oxford University Press.

49 The myth of Gyges’ ring is another place where Plato expressed the worry that people would chose injustice if there was any guarantee it would remained unnoticed (Republic II 360b–d). 50 The first stages of this paper were presented in Atlanta (Platonic Realism, Mid-Term Meeting of the International Plato Society, Emory University, March 13–15, 2015) and Berlin (Argumentation in Classical Antiquity: Dialectic, Rhetoric, and other Domains, Humboldt-Universität zu Berlin, June 23–25, 2016). I thank the participants in both venues for engaging in discussion.

Chapter 8

The Prospects for Rhetoric in the Late Plato Christopher W. Tindale

Abstract Plato’s engagement with rhetoric continues past the early and middle dialogues, like the Gorgias and the Phaedrus, contrary to the views of commentators. And that engagement recognizes a positive value to rhetoric as a necessary tool for leading people to justice. The paper explores rhetoric’s relation to Platonic (rather than Socratic) dialectic through an examination of its role in late dialogues where the method of dialectic is most pronounced.

8.1 Introduction With rare exceptions, the judgement of the literature is that Plato’s reaction to rhetoric is generally negative, and that the dialogues in which rhetoric is at issue are the Gorgias and Phaedrus, with the latter allowing a more positive conception.1 Tom Conley, for example, in his well-received study of rhetoric in the European tradition (in [3]) attests that “Plato had nothing good to say about rhetoric or its practitioners,” a sentiment that strongly echoes that of Brian Vicker’s (in [20]) indictment of Plato. Conley notes with regard to the Phaedrus account that rhetoric does not corrupt and is guaranteed legitimacy by the role of dialectic in it. But he still hints that the dialectic involved is Socratic dialectic, and notes that this rhetoric is intolerant “of doxa” and is grounded in truth ([3]: 12). James Herrick, commenting on the Phaedrus in his history of rhetoric (in [6]), finds the issue to be rhetoric’s relationship with truth. “Scholars have often noted that in Phaedrus Plato hints at a true art of rhetoric. Clearly, this would not have been the same art as that practiced by the Sophists and 1 One exception would be Wardy (in [22]), who reads the Phaedrus account as consistent with that of the Gorgias. The sense of “rhetoric” confirmed in the Phaedrus is so far from what is “recognizably practiced” as to represent something that only “perfect philosophers” could acquire (55). Another exception of note is Edwin Black (in [2]), who devotes the last few pages of his paper to the later Plato.

C. W. Tindale (B) Centre for Research in Reasoning, Argumentation and Rhetoric, University of Windsor, Windsor, ON, Canada e-mail: [email protected] © Springer Nature Switzerland AG 2021 J. A. Bjelde et al. (eds.), Essays on Argumentation in Antiquity, Argumentation Library 39, https://doi.org/10.1007/978-3-030-70817-7_8

173

174

C. W. Tindale

criticized in Gorgias. In fact, this Platonic art of rhetoric may not have been practiced by anyone in Athens, except Plato himself!” ([6]: 63). George Kennedy (in [9]) allows that in the Phaedrus “the art of rhetoric is conceivable for use in advancing truth, which must, however, be known by the orator first” (78). Indeed, in the Phaedrus Plato proposes a true art of rhetoric in which the rhetorician is transformed into a philosopher. This atopic philosopher (because he is out of place when taken beyond the activities of the polis) has no audience other than himself. But returning to the city after his time in the sun the philosopher-rhetorician of the Phaedrus must find the discourse and strategies with which to address others.2 The task is not leading people to philosophy per se, since so many of the lessons learned “in the sun” are not applicable to the wider populace. In this respect Susan Jarratt (in [7]), contra Kennedy above, is correct in pointing out that rhetoric’s role is not the discovery of truth; that is always the role of dialectic. She sees rhetoric’s role to lie in advocacy. But, then, how, if at all, does this fit into Plato’s later philosophy, where, as Rosenstock ([16]), observes “Plato seems to have shifted from a view of philosophy as practice to a view of philosophy as theory” (84). James Kastely (in [8] and [9]) comes closest to what interests me here. While he does not explore the later dialogues, he does raise questions about the relationship between refutation and persuasion ([9]: 51–52). Refutation is at the heart of Socratic dialectic, involving the examination of an interlocutor’s statements in a way that results in the contradiction of his initial thesis. But the dialogues also raise questions about what persuasion involves and how it should function in the everyday life of the polis. The contrast suggested by Kastely is important. It is one that appears to be overlooked by those who claim that Socrates is actually the true politician because the one-on-one discussion of his dialectic (the refutation) is the only authentically political sort of talk (Wardy in [21]: 82–83). This ignores the failure of Socratic dialectic to meet the general needs of the polis (Tindale in [18]) where a positive discourse is required (rather than Socrates’ negative nay-saying). It is this, I maintain, that is on Plato’s mind towards the end of his life. This is long after Plato has abandoned the method of Socratic dialectic with its failure to move beyond aporia and replaced it with his more distinctive brand of dialectic. At this later stage, rhetoric is rehabilitated and called upon to play an important role. The philosopher discovers truth through dialectic, but dialectic cannot itself suffice for communicating ideas informed by that truth.

2 On

a parallel course, we could track Plato’s own employment of rhetorical styles and devices throughout the dialogues, since that also bears on his attitude to rhetoric. But that would be a different paper. Here, I am interested in the theoretical role that rhetoric has in Plato’s thought, especially in the late dialogues.

8 The Prospects for Rhetoric in the Late Plato

175

8.2 Vickers’ View of Plato In titling his chapter ‘Plato’s Attack on Rhetoric’, Vickers makes no attempt to hide his judgment. And given the importance that this work has had on rhetorical studies and Plato’s reputation as a rhetorician, it is quite deserving of a careful analysis. The treatment begins with the Gorgias, a dialogue that, claims Vickers, treats rhetoric “as subservient to politics, being indeed the main tool of politics in Athenian democracy” ([20]: 84). Indeed, it is the main tool, and would not be otherwise. As we will see, Plato does not stray from this judgment in later dialogues. But as a tool of politics, it is not necessarily subservient to it. We might equally observe that without rhetoric, politics could not succeed. But Vickers accuses Plato of stacking the deck against rhetoric by, for example, enlisting incompetent people to defend it, a strategy which he puts down to Plato’s strong desire to discredit rhetoric (93). He finds the Phaedrus to be “Plato’s longest exposition of a reformed, philosophical rhetoric” (133), without fully acknowledging what has been accomplished there. And there, in the Phaedrus, Plato’s engagement with rhetoric effectively ends. Vickers enlists the judgment of E.L. Hunt to suggest that Plato’s “‘later utopia’ allowed ‘no freedom of utterance’, so denying rhetoric any chance to develop.” (138). Unfortunately, he fails to consider the chronology of the dialogues to explore any such development. His acknowledgements of dialogues after the Phaedrus are cursory and confused. He thinks the Socrates of the Statesman is Socrates when he was younger, which does not help matters. And he observes of that dialogue that while oratory is distinct from statesmanship, it “is openly given the task of state propaganda” (142). I will correct that judgment below. On the last page of the chapter (147) there is mention of the Philebus. But it is only a passing reference. He makes note of Gorgias’s presence and his citing of rhetoric’s greatness, but without exploring the context. After all the details of the treatments of earlier dialogues, it is disappointing to see such a minimal discussion of the later Plato, and with no commensurate appreciation of the development of Platonic dialectic. In fact, one of the confusions concerns the nature of the Platonic dialectic, which, like other commentators before him, he reduces to the Socratic method of dialectic. He talks of Plato’s privileging of the Socratic elenchus (125), and never allows that Plato’s own method is distinctively different. This puts into question much of his interpretation of how rhetoric stands to philosophy in Plato’s work, since the latter cannot be understood without a full appreciation of the dialectical method. It is in discussing the role of rhetoric in the Laws that Vickers comes close to the point. If he were not predisposed to conclude otherwise, he would recognize the value of what he observes. With reference to a passage at 719e–720a he says “The appointed lawmaker must not just ‘tell us curtly what we are to do or not to do, add the threat of a penalty, and then turn to the next enactment, without one word of exhortation or advice to the recipients’; rather, he must add a ‘prefatory statement in front of his code’,” and then: “So, ironically enough, he now needs rhetoric again!” (143) In fact, he never stopped needing it and a deeper analysis might have recognized this.

176

C. W. Tindale

8.3 Dialectic and Rhetoric It follows from ideas in the previous section, or is at least strongly suggested by them, that the prospects of rhetoric (in Plato) are closely tied to those of dialectic (and, dare we suggest, vice versa). How should we understand these central terms of dialectic and rhetoric? It has been claimed that dialektikê and rhêtorikê were terms invented by Plato (Kahn 2013: 132; Schiappa 1999: 14). Of the latter, Schiappa writes: “The term rhêtorikê simply cannot be found in any text that has been dated prior to Plato’s Gorgias, usually dated to the 380s B.C.E.” (14).3 Of course, while the terms may originate with Plato, only the practice of one of them does. Rhetoric itself has a lively pre-Platonic life, captured by various terms, usually logos. This makes it all the more interesting that Plato would apparently adopt a term to describe a practice that he does not initially embrace. One plausible reason for this is a desire to clarify what was at stake in the contrast between Socrates’ practice and that other mode of discourse. Initially, dialectic describes the art of philosophical conversation, and its technical sense is lent to the method of refutation employed by Socrates. In the course of the Republic a more developed sense still of dialectic will emerge, and this will be the sense fully adopted by Plato himself and discussed in the late dialogues. Part of the problem with the view about dialectic expressed by rhetoric scholars like Conley and Vickers is that they overlook the specific nature of Plato’s dialectic, confusing it with the earlier Socratic version.4 For example, Conley writes: “Dialectic has a game-like quality to it. Two people are needed to engage in dialectic” (10). This is characteristic of the early conversations Socrates has with interlocutors, as he seeks to establish whether they know what they claim to know. This activity is one of refutation, where the statement put forward by the interlocutor is examined dialectically (through the asking of questions) to the point where the examination arrives at the opposite of the statement and it is refuted. Essentially, the inquiries here were definitional: What is piety? What is courage? And the results were aporetic, failing to achieve any satisfactory definition. The Platonic dialectic that emerges in the Republic and Phaedrus and that is illustrated in later dialogues like the Sophist, Statesman or Philebus refines the interest in definitions by looking to place things under the right classes. Such inquiries into classification are best understood in terms of one of the descriptions given to them—the method of division and collection. Here, participants in Plato’s later dialogues take concepts and divide them in order to find where something—the sophist, the statesman—naturally belongs and thus reveal the truth about it. Hackforth observes in his commentary on the Philebus that it tends to be more a method of division than of collection (Hackforth in [5]: 26). But insofar as the participants engage in finding the correct classes they are effectively composing those classes (or uncovering what is collected in them). Hence, we have a method of 3 Laurent Pernot’s (in [11]) review of the evidence is more cautious, allowing only that the word is present in sources available from about 390. 4 Admittedly, there are some important similarities—they both involve a “shared search,” for example (See Gill in [4]: 296)—but the difference in goals and what can be achieved far outweigh any commonalities.

8 The Prospects for Rhetoric in the Late Plato

177

division and collection. The Stranger observes in the Statesman that those listening to him should not make the mistake of thinking he believes that the class and part are separate from one another. Rather: “when there is a class of anything, it must necessarily be a part of the thing of which it is said to be a class; but there is no necessity that a part be also a class. Please always give this, rather than the other, as my doctrine” (263b).5

8.4 The Position of the Later Dialogues (Sophist, Statesman, Philebus, and Laws) Contrary to any belief that Plato moves away from practice and toward theory,6 several late dialogues both illustrate the practice of dialectic and appeal to the use of rhetoric. Nor should we make the mistake of thinking that Plato’s attitude toward rhetoric has not developed (Vickers in [20]). Dated around 360 B.C.E., both the Sophist and Statesman depend for their success on the central role of dialectic as a method. The Sophist, for instance, contains four examples of division, giving four different lines of descent. But the Statesman is of more interest here because of the way it introduces rhetoric. It also has a focused interest on the state, and so questions of how one operates within the State are natural here. Indeed, it is about midway through the dialogue, after several divisions that establish the statesman within the class of those who care for herds, and after a subsequent analogy with the art of weaving, that the investigation turns to the arts that have to do with the state itself (Statesman, 287b).7 The analyses continue to explore all the arts that contribute in any way to the state. These are described as contingent causes, without which statecraft could not exist (287d). Painters and even priests are in turn considered and left behind, until the Stranger and Young Socrates arrive at three arts that are “precious and akin to statecraft” (303–304a), having separated what is different from it, alien to it and incompatible with it. These three are “the arts of the general and of the judge and that kind of oratory which partakes of the kingly art because it persuades men to justice and thereby helps to steer the ship of state” (304a). But for all of these there stands above them one premier science which decides when and whether they should be used. This is statecraft. Thus, “the science which decides whether to persuade or not should control that which can persuade” (304c). The power of persuading through speech (rather than teaching) is given to rhetoric (rhetoriké), thus separating it from statecraft and making it subservient to it. But, it is closely akin to it. Likewise, generalship (304e–305a) and the capacity of judges who judge correctly (305b–c) are akin to the kingly art, but also subservient to it. Plato is concerned here to distinguish 5 The

source for quotes from the Statesman is Harold N. Fowler’s Loeb edition ([14]). than in his own practice, as noted in fn2. 7 For a more detailed discussion of the rhetorical nature of the Statesman, including Plato’s choice of imagery see Tindale (in [17]). 6 Other

178

C. W. Tindale

these three arts from statecraft (or politikén) lest they be confused with it. But for our purposes, the point to note is the importance given to rhetoric, correctly directed, as a means of persuading people to justice.8 It is also noteworthy that while two of the close arts can be assigned to a particular figure—the general or the judge—that of the third has no such association, suggesting that it may be more generally acquired and applied. Later (309d), we learn that it is part of statecraft’s role to implant true opinions about honour, justice, goodness and their opposites into those who have the correct education. Rhetoric’s function as an important tool for achieving this and communicating to the multitudes should not be lost sight of here. The attitude toward rhetoric is thus far from the dismissive sentiments of the Gorgias, and helps to explain how the informed rhetoric of the Phaedrus should operate in the correct governing of the state. Turning to the Philebus, we find a similar discussion and illustration of dialectic before some poignant remarks about rhetoric are introduced. The Philebus is another late work of Plato’s, connected through a number of themes to the Timeaus (Hackforth in [5]: 3), reflecting the same illustration of dialectical method that we find in the Sophist and Statesman. The focus is on the nature of the better life: whether it is that of pleasure (Philebus’ position) or intelligence. In examining both the nature of pleasure and of intelligence, Socrates and Protarchus perform the role of dialecticians and apply the method of Division (and Collection) to each. The importance of the method is emphasized at the outset. Socrates discusses the problem of the one and the many, focusing on what he calls the real problem (15a– b).9 He then explains how the answer is to apply dialectic (15b–18e) and to show how each (collected) unity has within itself a number of “kinds” that mediate between it and particulars. The philosopher (or dialectician) must know these intermediaries and their number. That this method is not the Socratic method of earlier dialogues, where an interlocutor is led into contradiction (refuted) in order to learn what they do not know, is illustrated by the presence of a brief case of contradiction to illustrate the contrast (21a–d). Socrates asks Protarchus whether he would care to live his whole life enjoying the greatest pleasures. On receiving an affirmative response, he proceeds to examine Protarchus’ commitment to that statement until Protarchus experiences what many interlocutors had experienced before him: “Your argument, Socrates, has reduced me for the moment to complete speechlessness” (21d).10

8 Vickers

overlooks both the development of ideas through the dialogues and the nature of the problem facing Plato when he suggests that rhetoric here is “openly given the task of state propaganda” ([20]: 142). This is the power of persuading through speech that Vickers had looked for (and failed to find) in the Gorgias, and it is needed for those who are not philosophers to be brought to justice through correct opinion. 9 That problem, it will be recalled, is how the One is distributed across the many without being divided from itself. This had been the opening concern of the Parmenides. In the Philebus, Socrates chides people for jumping too hastily from the One to the many without noticing the intermediaries that Collection and Division now bring to light. 10 The source for quotes from the Philebus is the Hackforth translation ([5]).

8 The Prospects for Rhetoric in the Late Plato

179

The method of dialectic is easy to indicate but hard to employ (16c). It is introduced as a gift from the gods, comparable to the gift of fire given to Prometheus. Plato in this sense is a Promethean figure. According to the gods’ account: we ought, they said, whatever it be that we are dealing with, to assume a single form and search for it, for we shall find it there contained; then, if we have laid hold of that, we must go on from one form to look for two, if the case admits of there being two, otherwise for three or some other number of forms: and we must do the same again with each of the ‘ones’ thus reached, until we come to see not merely that the one that we started with is a one and an unlimited many, but just how many it is. But we are not to apply the character of unlimitedness to our plurality until we have discerned the total number of forms the thing in question has intermediate between its One and its unlimited number: it is only then, when we have done that, that we may let each one of these intermediate forms pass away into the unlimited and cease bothering about them (16d–e).

There is a lot that could be analyzed here. But the main point is that the method, however we understand it, is not the dialectic of old that was illustrated in that brief exchange between Socrates and Protarchus. It is a new dialectic, enquiring into the nature of things. And while it still requires the combined efforts of several interlocutors working together to reach understanding (7th Letter 344b), it is no longer concerned with what those interlocutors believe or their commitments. This dialectical method is illustrated in various ways throughout the dialogue, exploring, for example, all that now exists in the universe (23c), and the pair Becoming and Being (54a). But the division that most interests us is that of Knowledge (55c–59c). Here, the original division (55d1–3) is between technical knowledge (production) and that concerned with education and culture. The first involves measuring (like in the art of carpentry), and the second involves guesswork (like in the art of music). Again, there are two types of arithmetic, one ordinary and the other philosophical (56e). With respect to knowledge, then, there are two types, one purer than the other. This may also be the case for dialectic, since when Socrates says that no other art can be preferred to dialectic, Protarchus responds “plainly everyone will recognize he whom we now speak of” (58a). As Hackforth notes (119n2), the phrasing may imply two kinds, one with rigor and the other just conversation. The method of dividing knowledge has arrived at an understanding of dialectic itself, the purest kind of knowledge, having the exactness that is appropriate to philosophers. What is plain to everyone is the “cognition of that which is, that which exists in reality, ever unchanged, is held, I cannot doubt, by all people who have the smallest endowment of reason to be far truer than any other.” (58a). In this discussion of knowledge, when Protarchus hears Socrates talk of what everyone believes to be this pure kind of knowledge he responds that he thinks this is the art of rhetoric of which Gorgias spoke: “On many occasions when I used to listen to Gorgias, he regularly said, Socrates, that the art of persuasion was greatly superior to all others, for it subjugated all things not by violence but by willing submission, and was far and away the best of all arts” (58a–b). Socrates’s response suggests that there is, indeed, an important use for rhetoric. As he explains things: “What I wanted to discover…was not which art or which form of knowledge is superior to all others in respect of being the greatest or the best or

180

C. W. Tindale

the most serviceable, but which devotes its attention to precision, exactness, and the fullest truth” (58b–c). This allows that Gorgias’s art may indeed be “the greatest or the best or the most serviceable.” And Socrates would seem to confirm this when he goes on to explain: “you won’t give offense to Gorgias, if you allow his art the property of doing paramount service to mankind, while assigning to the procedure to which I have just referred just that property of possessing paramount truth” (58c). Seth Benardete (in [1]) does not see Plato endorsing rhetoric here. He sees Protarchus advocating the rhetoric of self-knowledge, the art that allowed him to do whatever he wanted. But this rhetoric, suggests Benardete, “seems to be comedy, which Socrates had argued was always grounded in self-ignorance” (224). And, of course, one could insist that all Socrates is doing is inviting Protarchus to allow that Gorgias’ art is of paramount service to humanity (which, indeed, he had already done). But this fails to appreciate the statement in the context of the division of knowledge that has been performed. Nor does it explain why Plato should phrase things as he does and not reinforce the earlier negative view of the Gorgias. In fact, the text recognizes what the late Plato has been drawn to see, that grasping the fullest truth with precision and exactness, while having great personal value, has no social currency unless the dialectician can then communicate the insights gained in a way that informs society at large and thereby improves the nature of that society. The subsequent advocacy in the Philebus of the mixed life would seem to acknowledge as much (61b). Hackforth’s gloss on this is instructive: “We now proceed to select the ingredients of the mixture. First as to knowledge, we must of course have the ‘truest part’ of this, namely the knowledge of true, immutable Being; but it is agreed that an inferior kind must be included as well; in fact we shall allow any and every sort of intellectual ability…a place. This decision is taken out of regard for the needs of the practical life” (127). For example, as he explains things, knowledge of pure arithmetic is not sufficient to build houses. Indeed, we can make a parallel case for the discourses required to govern. The pure is not sufficient. The only strict provision in taking all the other sorts of knowledge is that the first sort must be in place before the others (62d). Finally, it is appropriate to offer some remarks on the Laws, the unfinished investigation by Clinias, Megillus and the Athenian into the best legislation for the state. The discussion in Laws XII makes clear that the Guardians (the ruling nocturnal council), comprising older and younger members, must practice dialectic as we have seen it described elsewhere. In any field, the skilled individual must be able to see not just many individual instances of a thing “but also win through to a knowledge of the single central concept, and when he’s understood that, put the various details in their proper place in the overall picture.” (XII: 965b).11 This holds for things like virtue, goodness and beauty, with respect to which the guardians should know not only that these terms are a plurality but also understand the sense in which they are unities (966a).

11 The

[12]).

source for quotes from the Laws is Trevor J. Saunders’ translation in the Hackett edition (in

8 The Prospects for Rhetoric in the Late Plato

181

Then the Athenian asks: “what if they understood the point, but couldn’t find the words to demonstrate it?” To which Clinias replies that this is absurd. So, the Athenian proceeds: “Well then, isn’t our doctrine going to be the same about all serious questions? If our guardians are going to be genuine guardians of the laws they must have genuine knowledge of their real nature; they must be articulate enough to explain the real difference between good actions and bad, and capable of sticking to the distinction in practice.” (966b). How they come to master the methods of articulation is not explained. Earlier in Book VII their education had been discussed, and rhetoric was not included. But earlier still in the discussion of Book IV, as dawn has become noon (722c),12 the Athenian recognizes that legislators cannot just impose laws, the “dictatorial prescription.” There must be a preamble such as “the spoken word, and in general all compositions that involve using the voice” employ. This is the persuasive element “analogous to that of a preamble in a speech” (722d–723a). Vickers finds this recourse to the tradition of rhetoric ironic ([20]:143) and seems surprised by it. But on my reading, it is a natural acknowledgement of the utility of what has never been abandoned; it just needed to be adapted to the right ends. At the very end of the Laws, at the close of Book XII, they return to the question of education and decide they cannot legislate in advance the full nature of the education until the council’s activities have been established. “Its curriculum must be decided by those who have already mastered the necessary branches of knowledge” (968c), and “it will be a waste of time to produce written regulations about the order in which the various subjects should be tackled” (968e). This discussion, then, certainly points towards the need for the Guardians to be able to use language to articulate what they see. And in the context of other late dialogues, this need points towards the value of rhetoric that has been identified in those dialogues. That not just any discourse in the social arena will do had been made clear earlier in Book XI, where the Athenian observes: There is a certain kind of immoral practice, grandly masquerading as a ‘skill’, which proceeds on the assumption that a technique exists—itself, in fact—of conducting one’s own suits and pleading those of others, which can win the day regardless of the rights and wrongs of the individual case; and that this skill itself and the speeches composed with its help are available free—free, that is, to anyone offering a consideration in return. Now it is absolutely vital that this skill, if it is a skill and not a knack born of casual trial and error—should not be allowed to grow up in our state if we can prevent it. (937e–938a)

The reference to a “knack” echoes the same clause used to dismiss the negative rhetoric of the Gorgias (463b). Here, as there, what is at issue is an unskilled rhetoric associated with discourse of the Sophist—Plato’s perennial foe—and the desire to use any means to win the argument, regardless of the truth of the matter (Tindale 12 An

echo, perhaps, of the heat of the noonday sun that prompted the important action of the Phaedrus.

182

C. W. Tindale

in [19]: 30–43). For this, Plato has never lost his disdain. But this contrasts clearly with the positive rhetoric Plato promotes, one that is informed by knowledge of the truth as it speaks justice.

8.5 Conclusion: The Reluctant Rhetorician—Plato’s Recognition of the Power of Rhetoric Unlike Socrates, Plato was not an egalitarian thinker. The views of the Seventh Letter, authentic or not, express well what we find elsewhere: that not everyone can access the truth and that not everyone can write, and no one can write the truth. Books are not to be trusted. Recall the derogatory remark of the Theaetetus (162a) that has Protagoras speaking from the “impenetrable sanctuary of the book.”13 The sense of inaccessibility that attaches to this metaphor is the sense that Plato wishes to promote. Thus, it falls to speech to be the primary means of communicating ideas, and it falls to the few to convey what is important to the many. After all, on a certain level correct opinion is indistinguishable from knowledge (Meno 97b). Thus, Plato’s engagement with rhetoric continues past the middle dialogues. He requires it to forge the social environment in which justice can flourish. The late dialogues reflect this positive attitude in the comments made, duly understood within the contexts developed there. As we see in Statesman, rhetoric has a primary role, one of three gathered close to statecraft and necessary for it to function well. It implants true opinion about matters like honour and justice and their opposites to those who have the appropriate preparation. This explains the usefulness of rhetoric acknowledged in the Philebus and illustrated in the Laws, where details of the tradition of rhetoric (discussed in contrast to “Platonic” rhetoric in the Phaedrus) are brought back as aids to the last project that Plato attempted in his life. The traditional rhetoric discussed in the Phaedrus has not been supplanted by the new; it has been supplemented by that to which it is naturally inferior. Dialectic is a method of shared discovery, not mass communication. But the philosopher does not just speak to other philosophers or those being led to philosophy. There is the speech of persuasion, which checks the desire to compel and speaks of justice. This is how rhetoric is employed in the late Plato.14

13 From

the revised translation by Myles Burnyeat ([15]) in the Hackett edition. versions of this paper were read at the annual meeting of the Classical Association of Canada, Toronto, May 2015, and to the fellows of the Centre for Research in Reasoning, Argumentation and Rhetoric at the University of Windsor, September 2015. In each case, I am grateful to my audience for constructive discussion. 14 Earlier

8 The Prospects for Rhetoric in the Late Plato

183

References 1. Benardete, S. (1991). The tragedy and comedy of life: Plato’s Philebus. University of Chicago Press. 2. Black, E. (1958). Plato’s view of rhetoric. The Quarterly Journal of Speech, XLIV (4), 361–374. 3. Conley, T. M. (1990). Rhetoric in the European tradition. University of Chicago Press. 4. Gill, C. (1996). Afterword: Dialectic and the dialogue form in late Plato. In C. Gill & M. M. McCabe (Eds.), Form and argument in late Plato (pp. 283–311). Clarendon Press. 5. Hackforth, R. (1972). Plato’s Philebus. Cambridge University Press. 6. Herrick, J. A. (2001). The history and theory of rhetoric: An introduction. Allyn and Bacon. 7. Jarratt, S. (1991). Rereading the sophists: Classical rhetoric refigured. Southern Illinois University Press. 8. Kastely, J. L. (2015). Rhetoric and Plato’s Republic: Democracy and the philosophical problem of persuasion. University of Chicago Press. 9. Kastely, J. L. (1997). Rethinking the rhetorical tradition: From Plato to postmodernism. Yale University Press. 10. Kennedy, G. (1963). The art of persuasion in Greece. Princeton University Press. 11. Pernot, L. (2005). Rhetoric in antiquity (W. E. Higgins, Trans.). The Catholic University of America Press. 12. Plato (1997). Laws (T. Sanders, Trans.). In Plato: Complete works. J. M. Cooper (Ed.). Hackett Publishing Company. 13. Plato (1925). Philebus (H. N. Fowler, Trans.). Loeb. Harvard University Press. 14. Plato (1925). Statesman (H. N. Fowler, Trans.). Loeb. Harvard University Press. 15. Plato (1997). Theaetetus (M. F. Burnyeat, Trans.). In J. M. Cooper (Ed.), Plato: Complete works. Hackett Publishing Company. 16. Rosenstock, B. (1997). From counter-rhetoric to askesis: How the Phaedo rewrites the Gorgias. In B. Schildgen (Ed.), The rhetoric canon (pp. 83–105). Wayne State University Press. 17. Tindale, C. W. (2018). Persuasion and statecraft: Plato and protreptic argument. Paradigmi: Rivista Di Critica Filosofica. XXXVI (Settembre-Dicembre), 443–460. 18. Tindale, C. W. (2018). Alternative lives: The power of rhetoric in Plato’s polis. Eranos: Acta Philologica Suecana, CIX(2016–2018), 93–103. 19. Tindale, C. W. (2010). Reason’s dark champions: Constructive strategies of sophistic argument. University of South Carolina Press. 20. Vickers, B. (1988). In defence of rhetoric. Clarendon Press. 21. Wardy, R. (1996). The birth of rhetoric: Gorgias, Plato and their successors. Routledge. 22. Wardy, R. (2009). The philosophy of rhetoric, and the rhetoric of philosophy. In E. Gunderson (Ed.), The Cambridge companion to ancient rhetoric. Cambridge University Press.

Chapter 9

Does Plato Have a Theory of Induction? Epag¯og¯e and the Method of Collection “Purified” of the Senses Holly Moore

Abstract Although Socrates’ use of induction and epagogic argumentation in Plato’s dialogues is well studied, scholarship on Platonic methodology lacks a clear account of Plato’s own view of epag¯og¯e. In this paper, I refute Richard Robinson’s claim that Plato had no awareness of epag¯og¯e, arguing that the “method of collection” serves as Plato’s theory of dialectical induction. Using the evidence of both the Statesman and the Sophist, I maintain that the abstraction characteristic of collection may be ‘purified’ of its empirical origins in its dialectical application to the inquiry into forms. In addition to providing a unifying thread from Socratic to Platonic and, finally, Aristotelian epag¯og¯e, this view also offers evidence for the consistency of dialectical methodology within the dialogues.

9.1 Introduction In response to Aristotle’s claim in Metaphysics M that Socrates engaged in epagogic arguments (epaktikoi logoi), several modern scholars have attempted to identify the structure and function of “Socratic” induction1 as well as its relationship to both elenchus and the search for definitions [9, 13, 20, 21]. Despite the fact that the majority of the evidence for Socratic induction is drawn from Plato’s Socratic dialogues, very little scholarly attention has been given to assessing Plato’s view of 1 Despite the fact that, as Julia Annas [2], Gregory Vlastos [23] and others have noted, Aristotle’s use of epaktikos does not correspond to modern logic’s definition of induction, this terminology remains the most recognizable translation of epag¯og¯e. Though I agree with these authors, my argument does not hinge on a technical definition of the form, but instead upon establishing that Plato had an interest in setting up collection as a fruitful ground for dialectical reasoning.

An abridged, earlier draft of this paper was published in Spring 2019 as: Platonic epag¯og¯e and the “purification” of the method of collection. Epoché: A Journal for the History of Philosophy 23 (2), 353–364. H. Moore (B) Associate Professor of Philosophy, Luther College, Decorah, IA, USA e-mail: [email protected] © Springer Nature Switzerland AG 2021 J. A. Bjelde et al. (eds.), Essays on Argumentation in Antiquity, Argumentation Library 39, https://doi.org/10.1007/978-3-030-70817-7_9

185

186

H. Moore

the role of epag¯og¯e in dialectic. Given Aristotle’s claim that all dialectical argument is either syllogism or epag¯og¯e [3, Topics I.12], it is therefore surprising to find Richard Robinson claim that, while the character Socrates regularly practices epagogic induction in the Platonic dialogues, “Plato shows virtually no consciousness of epagoge as a method,” and again that “Apparently we must conclude that his depiction of it in the early dialogues made no impression on his own theory of method” [20, pp. 45, 46]. Such a definitive statement in what remains the authoritative, systematic account of Platonic dialectic appears to have halted scholarly discussion of the topic entirely. However, an examination of Plato’s ‘later’ methodological discussions of dialectic, particularly the role of the method of collection, suggests that the subject should be revisited. In what follows, I maintain that Robinson’s argument against Plato having an account of epag¯og¯e fails because Plato’s account of collection and its ability to operate ‘purified’ of the senses provides an account of epagogic reasoning as well as offering a better understanding for Aristotle’s two forms of epag¯og¯e. Despite arguing that Socratic epag¯og¯e is not founded on sense perceptions and that “if [Plato] realized the logical nature of epagoge, he did not connect it with sensation,” Robinson rejects the possibility that ‘the upward path’ (Republic) of dialectic might include epag¯og¯e because “If the upward path were generalization, it would surely have to be empirical” [20, pp. 46, 163]. Robinson’s argument begins from the notion that since collection is said to be based in sensation (as stated in the Phaedrus’ depiction of dialectic), it cannot function as part of the ‘upward path’ of dialectic described in the Republic. Nevertheless, I will argue that what appears to be an inconsistency among the dialogues’ methodological claims about dialectic is in fact a misunderstanding of the way that collection functions as both the starting point and the driver of further dialectical investigation. I argue that collection is Plato’s term for the process of generalization, which Robinson rightly associates with epag¯og¯e as the movement from more particular to more universal, but I provide here an account of collection as liberated from the senses. In addition to the Phaedrus’ discussion of collection as beginning with the senses, using both the Sophist and the Statesman, I show that collection may be ‘purified’ through the dialectical method, such that its fundamental movement from ‘many to one’ may be applied not only in the movement from a sensible many to an intelligible one, but also from the intelligible many produced through the method of division to an intelligible one, the higher cause of each form, as suggested by the depiction of dialectic in both the Republic and the Philebus.

9.2 The Meaning of Epag¯og¯e Much of the difficulty in assessing the role of epag¯og¯e in Platonic dialectic resides in the ambiguity of the term itself. As Robinson notes, the substantive form is not used extensively in Plato’s works, and when it appears, it usually carries the sense of ‘incantation’ or ‘inducement’ (viz., Republic 364c). However, the verb epagein is a relatively common one, showing up in 13 distinct dialogues, with 27 total uses

9 Does Plato Have a Theory of Induction …

187

[4, p. 747]. The Statesman invokes epagein in a way more akin to the logical sense attributed to the term by Aristotle, where the teacher is said to ‘lead along’ the student in the use of paradeigmata (examples). As it happens, epag¯og¯e and paradeigma are again related in Aristotle’s accounts, to which we turn now. Aristotle’s accounts of epag¯og¯e appear in several of his logical works. In Topics I.12, he asserts that epag¯og¯e, along with syllogism, are the two species of dialectical logoi, saying “…epag¯og¯e is a passage from particulars to universals, e.g. the argument that supposing the skilled pilot is the most effective, and likewise the skilled charioteer, then in general the skilled man is the best at his particular task” [3, 105a12–15]. Here, Aristotle describes the process of developing a general principle based on the similarity among cases, which resembles his discussion of paradeigma in Posterior Analytics II.242 and seems closely connected the discussion of paradeigma in the Statesman, where epagein is invoked, as noted above. In addition to these depictions of epag¯og¯e as concerned with argument from general cases to a rule suggested by the cases, Aristotle also emphasizes in several works that epag¯og¯e is more persuasive because it is “more readily learnt by the use of the senses” [3, 105a16–17]. Here is some ambiguity in Aristotle’s account of epag¯og¯e. By using the same term, Aristotle appears to conflate argumentative epag¯og¯e with the inductive apprehension of universals through the perception of particulars, represented in the infamous passage of Posterior Analytics II.19 as well as I.1 and I.18 of the same work. Book VI of the Nicomachean Ethics also alludes to what Robinson terms “intuitively certain epagoge” [20, p. 36] as the basis for knowledge of first principles (1139b26–31). Though we will have to leave this curiosity aside for now, in my conclusion, I return to re-assess this apparent conflation, given my account of epag¯og¯e in Plato. Now, despite confusion regarding the precise meaning of epag¯og¯e for Aristotle (whose account of dialectic as composed of epag¯og¯e and syllogism drives Robinson’s examination of inductive argument in Socratic method), Robinson articulates three types of epagogic argument as the basis for his assessment of Socratic induction: “inferences from a case or cases to another case” or “from the cases to the universal” or “from cases to a universal followed by an inference from the universal to a new case” [20, p. 34]. It is significant that Robinson eschews Aristotle’s account of epag¯og¯e as induction from sensation, for his claim about collection’s ‘empirical’ basis provides the purportedly contradictory evidence of the Republic’s account of dialectic and the later dialogues’ accounts of collection [20, p. 163]. In what follows, I will demonstrate that the dialogues’ depiction of collection not only reflects; this is a matter of avoiding anachronism. Aristotle’s two-sided account of epag¯og¯e but also that, contra Robinson, the logical form of collection’s generalizing method is a viable candidate to serve as Plato’s explicit methodological understanding of the 2 There

is a great deal of interpretive ambiguity in these accounts, since Aristotle insists in Prior Analytics II.24 that paradeigma differs from epag¯og¯e in that the latter begins from “all the particular cases,” rather than from some exemplary ones [3‚ 69a17]. The form of epagogic argument, which Robinson designates “enumeratively certain epagoge” [20, p. 36] is elaborated in more detail in Prior Analytics II.23.

188

H. Moore

Republic’s account of the ‘upward path’ of dialectic, which demands a method purified of sensation. I defend the view that the dialectical method of collection stands as Plato’s account of epag¯og¯e, serving as the fundamental pattern of generalizing— both from sensible particulars to universals as well as from intelligible forms to their higher order causes.

9.3 Robinson’s Rejection of Platonic Epag¯og¯e Robinson’s claim that Plato does not have an understanding of epag¯og¯e is founded on two pieces of evidence: (i) the absence of the sense of the term indicated (or a terminological equivalent), and (ii) the lack of an account of anything like it in the methodological passages [20‚ p. 46]. Robinson explains his comparatively high burden of proof for assessing Plato’s views: “I have tried not to attribute to Plato any inference that he does not make in so many words, or any abstraction that he does not have a name for…I have assumed that to possess a single name for an idea is a later stage than to be able to express it only in a sentence, and that, if the author neither names nor states the idea, it requires a very special evidence to say he had it” [20, p. 5]. This approach accounts for Robinson’s insistence that Plato’s limited use of the term epag¯og¯e provides evidence that the notion is itself missing from Plato’s thought. I do not plan to directly address this aspect of Robinson’s argument, since I believe that my challenge to his second premise regarding the methodological works functions to provide precisely the “very special evidence” required to assert that Plato did indeed have a notion of epag¯og¯e, while not using this term or any other systematically. Additionally, it should become clear in what I argue that, effectively, Plato’s account of epag¯og¯e simply is his account of “collection,” though this notion is itself not terminologically stable. Plato’s thinking is rarely terminologically systematic, and this suggests Robinson’s expectation is somewhat inappropriate for studying Plato. Since I will direct the majority of my argument against Robinson’s second premise, I will first elaborate in some detail the basis of Robinson’s claim that none of Plato’s methodological discussions includes reference to epag¯og¯e. Robinson begins his argument by referencing the Republic’s discussion of dialectic, in particular, its depiction of the ‘upward path’. At the climax of the divided line passage of Republic VI, Socrates describes the manner by which the soul interacts with the forms through dialectic: …Go on to understand that by the other segment of the intelligible I mean that which logos itself grasps with the power of dialectic, making the hypotheses not beginnings but really hypotheses—that is, steppingstones and springboards—in order to reach what is free from hypothesis at the beginning of the whole. When it has grasped this, logos now depends on that which depends on this beginning and in such fashion goes back down again to an end; making no use of anything sensed in any way, but using forms themselves, going through forms to forms, it ends in forms too. [19, 511b4–c21]

This account of dialectic is famously difficult to interpret and appears to be a quite different depiction of the philosopher’s art than those we find in other works, where

9 Does Plato Have a Theory of Induction …

189

dialectic is associated with the methods of collection and division, which receive no mention here. Robinson takes pains to reject interpreters’ (Zeller, Maier and Rodier) suggestions that the ‘upward way’ of dialectic could correspond to the methods of collection and division, central to the dialectical methodologies of the Phaedrus, Sophist, Statesman, and Philebus, a view he refers to as the “synthesis-theory.” Key to Robinson’s argument is the fact that the Phaedrus quite explicitly depicts collection as grounded in sensation, whereas the ‘upward way’ explicitly denies that dialectic has any connection to sensation [20, p. 164].3 Having shown that the upward path cannot include sensation, Robinson closes his case against epag¯og¯e with this: “If the upward path were generalization, it would surely have to be empirical” [20, p. 163]. My challenge to Robinson’s view, then, depends upon establishing the case that the later methodology includes an account of collection untethered from sensation. Beginning from Robinson’s assertion that Socratic epag¯og¯e consists of generalization that does not depend upon sensation, I argue that collection ‘purified’ of sensation is again a suitable candidate for the method of the ‘upward path’ assigned to dialectic in the Republic.

9.4 A Refutation of Robinson In so far as Robinson aims to account only for Plato’s earlier dialectic, it can be expected that he would not necessarily address in any detail the more elaborated methodology of dialectical inquiry discussed in the later dialogues. Nonetheless, Robinson does argue at length that the explicit methodological discussions of the later dialogues do not offer evidence of Plato’s ‘consciousness’ of epag¯og¯e. This view is surprising, given Robinson’s avowal that Socrates, in keeping with Aristotle’s attestation in Topics, did indeed practice epagogic argumentation in many of the early dialogues. It is prima facie improbable that the man who depicted these very arguments would not himself have any cognizance of the nature of their form as epagogic, but Robinson holds fast to his burden of proof, namely that the word (or its surrogate) be used in the sense meant, or that there be a very strong reason to suppose it is conceived as such, which in this case, he finds to be unlikely. In addition to the paradoxical claim that the Socrates of Plato’s dialogues often practices epag¯og¯e but that Plato has no ‘conception’ of it, Robinson admits that Socratic epag¯og¯e is in no way tied to sensation [20‚ pp. 42, 46], and that “it is 3 In

addition, Robinson argues that each method treats quite different objects: in the case of the ‘upward path’, Robinson characterizes it as concerning ideas, such as ‘the good’, whereas he claims collection and division treat genus and species [20‚ p. 163]. There has been great debate in the scholarly literature over the nature of what is treated by collection and division. I add my voice to those who reject the notion that they treat genus and species, advocating for a view similar to Cristina Ionescu, who argues persuasively that these methods may treat a variety of objects: “Depending on the level at which we collect and divide, we have as our objects intelligible forms as they are in themselves or as inherent in something else” [10, p. 48].

190

H. Moore

probable that, if [Plato] realized the logical nature of epagoge, he did not connect it with sensation” [20, p. 46]. Furthermore, as noted above, Robinson grants that “The connexion which Aristotle laid down between epagoge and sensation (Anal. III 18 and IV 19) is not necessary to epagoge in the broad sense given to it in this chapter for the purposes of describing what actually happens in the early dialogues” (ibid.). Accordingly, if I can show that the later dialogues’ discussions of collection include a depiction of generalization from instances to kind, and that this same operation is compatible with the ‘upward path’ of the Republic, I will have provided evidence for the claim that Plato does indeed have a consciousness of epag¯og¯e and that this may be identified with the method of collection, the partner to division, in dialectical method. The primary obstacle to my interpretation remains the dialogues’ depiction of collection as employing the senses, since this is precisely what is excluded from the description of the ‘upward path’ in the Republic. However, I shall argue in what follows that the formal nature of collection, gathering a multiplicity into a unity according to similarities, corresponds precisely with the method of epag¯og¯e identified by Robinson as a feature of Socratic elenchus, and may be applied to a non-sensible many, as is described in the Philebus. Further, this interpretation of the method of collection will also help to explain Aristotle’s ambiguous depictions of epag¯og¯e as both cognitive and argumentative in its application.

9.5 Collection as Generalization from Sensibles In the Phaedrus [16], Socrates offers a description of two ‘forms’ (eidoin) present in his speech on eros, which he later identifies with the divine art of dialecticians at 266c1: That of perceiving (sunor¯onta) and bringing together in one idea the scattered particulars, that one may make clear by definition the particular thing which he wishes to explain…[and the other] that of dividing things again by classes where the natural joints are, and not trying to break any part, after the manner of a bad carver. [ibid., 265d5–9, e1–3]

Socrates goes on from this to declare himself “a lover of these divisions and collections (t¯on diairese¯on kai sunag¯og¯on)” [ibid., 266b5–6, my translation]. Again, in the Sophist [18], the Eleatic Stranger describes the “free person’s science” [18, 253c9], referring to these same two methods: Shall we not say that the division of things by classes and the avoidance of the belief that the same class is another, or another the same, belongs to the science of dialectic?...Then he who is able to do this has a clear perception (diaisthanetai) of one form or idea extending entirely through many individuals each of which lies apart and of many forms differing from one another but included in one greater form, and again of one form evolved by the union of many wholes, and of many forms entirely apart and separate. [ibid., 253d1–e1]

Finally, in the Statesman, the Eleatic Stranger again describes the proper manner of distinguishing classes, invoking both methods:

9 Does Plato Have a Theory of Induction …

191

…when a person at first sees (aisth¯etai) only the unity or common quality of many things, he must not give up until he sees (id¯e(i)) all the differences in them, so far as they exist in classes; and conversely, when all sorts of dissimilarities are seen in a large number of objects he must find it impossible to be discouraged or to stop until he has gathered into one circle of similarity all the things which are related to each other and has included them in some sort of class on the basis of their essential nature. [17, 285a9–b6]

Through all these depictions (and many others),4 a few salient features become clear about collection. It is first of all characterized by what Kenneth Sayre calls “movement from plurality to unity” [22, p. 41]. The other major feature of collection is its attention to the ‘koinon’ or common—that feature of the many that unites them all under a single class or form.5 Division, on the other hand, is described as the process of dividing a given form into (preferably two) sub-classes (see Statesman 287c5–6). This process is a matter of locating the ‘natural joints’ between kinds, and thereby rests upon the attention to relevant differences, depending on the target of inquiry. Already, we may notice that collection and division are contraries: while collection focuses on similarity, division focuses on difference. Another key feature of collection here is its association with perception and sensation—aesth¯esis and its cognates are common in these passages, and collection is often explicitly defined with respect to its movement from the sensible many toward a common, intelligible form. This is, indeed, Robinson’s primary basis for rejecting the view that collection could function as Plato’s account of the ‘upward way’, and, thus, his strongest case that Plato has “no consciousness of epag¯og¯e as a method.” Recall that Robinson grants that Socratic epag¯og¯e, as it appears in the earlier dialogues, demonstrates a method that proceeds without ‘sense data’ and that he also says if Plato did have a theory of induction, it would be one that does not begin from sensation [20, p. 46]. In what follows, I will show the way that collection is transformed and ‘purified’ of its encounter with the objects of sensation, thereby becoming capable of collecting forms themselves, an activity fundamental to dialectic and the ‘upward way’. However, before turning to my account of collection as epag¯og¯e, I will here outline the trajectory and basis for my argument against Robinson. I believe that Robinson’s error lies not in overlooking the evidence of the later dialogues but in unnecessarily and artificially distinguishing perception and understanding, the passage through which is precisely the accomplishment of inductive thinking. If we consider epag¯og¯e as the generalizing characteristic of a ‘movement of thought’, as Aristotle puts it, rather than as a procedure defined by its objects, then collection, which may be applied to either a sensible or intelligible many, becomes an excellent candidate for representing Plato’s own view of epag¯og¯e. This view also has the virtue of allowing for a better understanding of the variations in Aristotle’s own accounts of epag¯og¯e. Insofar as it names the movement of thought 4 Collection is described or demonstrated at: Phaedrus 249b–c, 265d, 266b5–6; Sophist

219b11–c7, 234 b3, 253d7; Philebus 23e5, 25a3. Division is described at: Phaedrus 265e1–3, 273e2; Sophist 218d5, 220a8–9, 221b3, 253d1, 262b, 264d10–e2, 282a1–4; Statesman 282a1–4, 287c. 5 This process of abstraction does indeed appear to coincide with the depiction of recollection as the ‘recognition of similarities’ in the Phaedo, but also with the common hypothesis of the forms, referenced at Republic 507b.

192

H. Moore

from particulars to universals, epag¯og¯e is indeed the cognitive starting point of the development of concepts from sense data. However, in so far as we use dialectic and begin always already with generalized classes, such as ‘the expertise of the charioteer’, the movement of epag¯og¯e is simply the relative move from the more particular (which is to say, the many examples or instances) to the more universal (the single, overarching similar trait they all share). This ambiguity is reflected in Plato’s own account of collection, as I shall demonstrate in what follows.

9.6 Purification in the Sophist Scholars have debated considerably the role of collection and its relative significance in Plato’s dialectical method.6 Elsewhere, I have argued that the logical priority of collection extends beyond the initial positing of a class, and continues with each successive identification of sub-classes [15, pp. 185–188].7 Extending this account with a reading of the Sophist, I here emphasize the ultimately purifying influence division has upon collection in order to show that collection has the ability to treat not only sensible particulars but to collect intelligible objects into a unified whole. This reading of collection provides a basis for reconciling the Republic’s depiction of the ‘upward way’ with the later dialogues’ discussion of collection and division, thus offering a glimpse of Plato’s view of epag¯og¯e, even in the absence of his use of this term. Because the search for the definition of the sophist is expected to be difficult, the Eleatic Stranger, who leads the discourse of the Sophist, suggests that they begin with the easier quarry of the angler, to get practice at the method of division, which they’ll be using. This is, of course, no innocent example, since the choice of a type of hunter goes on to influence (perhaps not unduly) the first five definitions of the sophist. With its initial foray alone, the Sophist has already implicitly introduced both the role of hypothesis as the starting point for dialectic and the function of a model for providing a concrete example of the abstract dialectical method that will be employed throughout both the Sophist and the Statesman. 6 Although

Ambuel asserts that “there is no collection in the Sophist and there can be none” [1‚ p. 38] and, more generally, Deborah De Chiara-Quenzer claims that demonstration of collection is missing from the later dialogues because it is an easier method to acquire [7, p. 98], many scholars have argued that collection does not simply provide the starting point for dialectic but remains an integral part of dialectical reasoning, appearing in new guises, viz., Cornford regarding the Statesman [6‚ p. 171]. Kenneth Sayre claims that collection becomes paradigmatism [22] and, similar to Cornford’s view of the Sophist, Mitchell Miller [14] claims that the first six failed divisions of the Statesman constitute an extended collection. Kenneth Dorter makes the intriguing suggestion that the method of division is the focus of the later dialogues only because its philosophical use requires that it not remain value-neutral, as it is in sophistry, and that division must not be allowed to remain dianoetic but convert the thinking of the one who engages in it to no¯esis [8, pp. 113–117]. What I argue here might be best understood as extending this same claim to collection. 7 Cristina Ionescu makes a similar claim in her work on the Sophist [10].

9 Does Plato Have a Theory of Induction …

193

Given the discussions of both the Gorgias and Phaedrus, few readers of the Sophist are surprised that the Eleatic Stranger might begin from the form of techn¯e (nor, for that matter, that the Statesman’s definitions begin from epist¯em¯e). However, this starting point is not exactly neutral. It is hypothetical, in the sense that it is a reasonable assumption, based on prior, discursive inquiries. Christina Ionescu [11] has argued that these starting points (as well as each subsequent definition) are generated on the basis of the tests produced through elenchus and that the final definition of the sophist ‘collects’ what is learned from all the prior attempts.8 Sayre [22] and Klein [12] both argue that collection occurs at the end of each series of division, in the summary of each definition. I turn now to the sixth definition of the sophist, the so-called ‘noble sophist’, where the Stranger introduces an oblique discussion of dialectic, and where I believe the function of collection as it comes into contact with division is given deeper, if subtle, treatment. While it too begins from the positing of techn¯e as the generic class within which to find the sophist, the sixth definition diverges from all the prior attempts in that it does not initially divide the technai into those that are productive and those that are acquisitive (hunting) but instead begins by describing a range of technical practices, calling them all “discriminatory (diakritik¯en)” [18, 226c8]. Typically, when the Stranger makes a division, he announces the two sub-classes, sometimes explaining the meaning of each at some length. In this case, however, after having collected together under one name all the arts that share a diairetic nature [ibid., 226c2], the Stranger immediately begins to divide that class, making no mention of what other arts have been left behind. Commentators tend to fill in this blank on their diagrams of the division, assuming that the other sub-class consists of the combinatory arts, but it is worth noticing that precisely where we might expect to see something akin to collection, this goes unremarked by the Stranger. Having identified the discriminatory arts, the Stranger divides this class of arts into those that separate worse from better and those that separate like from like. Although the Stranger proceeds to identify the discrimination of worse from better as purification (katharmos) and pursues the ‘noble’ sophist along this path, something important is again left unremarked [18, 226d10]. The discrimination of like from like appears to be precisely what characterizes the method of division, as it is described by the Stranger later when they discover the free person’s science: “Shall we not say that the division of things into classes and the avoidance of the belief that the same class is another, or another the same, belongs to the science of dialectic?” (cited above). Additional evidence that the method of division does not necessarily belong to the art of purification comes almost immediately. While coaching Theaetetus on the next division to be made,9 the Stranger remarks that “the method of argument [division] is 8 See also Hugh Benson’s Clitophon’s challenge: Dialectic in Plato’s “Meno”, “Phaedo”, and “Republic” [5] on the relationship between hypothesis and dialectic. 9 Note that while the Stranger suggests living creatures (z¯ oo¯ n) and soulless thing (apsych¯on) as the basis for the following division of purifications, a division he’s invoked earlier in the division of hunting, Theaetetus innovates, crafting a far more bifurcatory and comprehensive division of soul from body. Clearly Theaetetus does not need to be taught the lesson that Young Socrates receives in the Statesman regarding ‘cutting too quickly’ to a target class (Statesman 262b).

194

H. Moore

neither more nor less concerned with the art of medicine than with that of sponging, but is indifferent if the one benefits us little, the other greatly by its purifying. It endeavors to understand what is related and what is not related (to xuggenes kai to m¯e xuggenes) in all the arts, for the purpose of acquiring intelligence; and therefore it honors them all equally…” [ibid., 227a8–b3]. That is, the method of division makes no distinction between better and worse (e.g., medicine and sponging); it attends instead only to their relative similarity in the given context. Again, this seems to place division more in line with the art that separates like from like rather than the art that purifies forms of what is unlike. Given all this, we should perhaps consider whether purification, while defined as an art of separation, might nonetheless include the method of collection and where the method of division properly belongs in the separation of the various discriminatory arts. Having noted these curiosities about the divisions relating to purification in the Sophist, let us turn to the discussion of purification in the Statesman for a different vantage on the subject, as well as a solution to the meaningfulness of these difficulties.

9.7 The Statesman on Purification As it happens, the discussion of purification in the Statesman follows immediately on the heels of an explicit reference to the subject of the Sophist’s discussion. In the process of trying to separate true statecraft from its illegitimate competitors, the Eleatic Stranger—who again guides this search—summarizes their discussion of various earthly regimes: “Then those who participate in all those governments— with the exception of the scientific one—are to be eliminated as not being statesmen, but partisans; and since they preside over the greatest counterfeits, they are themselves counterfeits, and since they are the greatest of imitators and cheats, they are the greatest of all sophists” [17, 1925, 303b7–c4]. The Stranger continues, noting that the remaining divisions are more difficult than this first one, since they require discrimination among those things most akin (genei) to the ‘kingly art’, which include generalship, judgeship, and rhetoric (1925, 303d5). He describes their situation with an analogy: “I think we are in somewhat the same position as refiners (kathairousi) of gold” [ibid., 303d6]. The Stranger explains that, while the first refinement removes the dross of stone and earth that clings to the ore, subsequent purification is required in order to draw off other precious (though less valuable) substances such as copper and silver. The result of rigorous “smelting and tests” is pure unalloyed gold, “auton monon eph’eautou” [ibid., 303e5–6]. The depiction of these two processes of refinement appear to map onto the first division of the discriminatory arts in the Sophist. The first refinement separates those that are generically different: earth, stone, and metal. A far further refinement, however, is necessary in order to separate the better and worse among the metals. Thus, the second refinement is an ultimate purification, the result of which is not only a unified, self-same form but also the rendering of other forms as different. This

9 Does Plato Have a Theory of Induction …

195

second refinement, then, appears to accomplish both the separation of like from like as well as the purification of the worse from the better.

9.8 The Purification of Collection What does all of this mean? The fact that these discussions of purification invoke processes attributable to both collection and division is simply a reflection of the fact that the use of the method of division in dialectical inquiry requires ongoing practices of collection. Let us outline, then, the way that collection and division are interwoven and operate in tandem at each point in the dialectical analysis that generates a definition. First (i), a class, recognized as generically inclusive of the definiendum is posited (techn¯e in the case of the search for the sophist and epist¯em¯e in the case of the statesman). The identity of this class presupposes a prior collection of many sensible instances (according to their perceived similarity) into a general form, identified by a feature common to all its instances. Next (ii), this class is divided according to additional classes that, while they share the common feature of the higher-order form, are themselves distinct forms. This process of distinguishing like from like, requires attending to difference, but it also presupposes, once more, the collection into self-same classes such that they may be subsequently compared.10 Third (iii), each of the classes so divided must be inspected according to whether the common feature of that form includes an indispensable attribute of the target class— this ‘selection process’ requires both attention to similarity between the sub-class and the target class as well as attention to their differences; in short, this is precisely the purificatory discrimination described in both the Sophist and Statesman, which, we have seen, blends the work of collection and division. This process (i–iii) is repeated until divisions no longer contribute any essential defining characteristics for the target form. The definition is complete, however, only when each of the selected classes that extended the division are gathered together into a final unification of all the selected forms, all of which are jointly implicit to the target class. This final collection is notably different from the initial one, however, in that it unifies not a sensible but an intelligible many. To this extent, it appears that collection is not only presupposed by every division, it also finalizes each definition. In addition to its continued employment throughout the dialectical search for the target, collection is transformed by its encounter with division in dialectic. While collection attends to similarities and gathers into unities, division directs the gaze of the dialectician back toward the inherent multiplicity within each form, seeking out seams11 implicit within the class—that is, division 10 This

process has its analogue in the Stranger’s explanation of the “example of example” at Statesman 278 b–c. 11 This image is employed by the Stranger in the Sophist, as he reviews their definition: “Then let us examine the opinion-imitator as if he were a piece of iron, and see whether he is sound or there is still some seam (diplo¯en) in him” [ibid., 267e7–9].

196

H. Moore

directs collection not back toward its original gathering of sensible instances but toward the multiplicity of the classes that compose each form. Division, then, is responsible for instigating the work of collection by presenting multiple ‘likes’ among those sub-classes, which must be further discriminated according to those that are ‘better and worse’, in this case better and worse fits for the target class.12 In terms of levels of analysis (which are not to be simply conflated with levels of abstraction13 ), division drives collection ‘downward’, toward multiplicity, though not as far as the utter multiplicity of sensible instances. Thus, although collection is necessary for each subsequent separation, division renews collection’s work, turning the dialectical inquiry back toward multiplicity. The kind of multiplicity division presents for collection, however, is an intelligible one, such that collection may draw out the similarities present among the multiple forms presented to it. Influenced by division in this way, collection proceeds “from forms to forms” rather than drawing out similarities among a multiplicity of sensible instances (cited above). At this point it should be fairly clear how division and collection together form the basis of the dialectical method, which proceeds by “using forms themselves, going through forms to forms, it ends in forms too” (cited above). In addition to providing initial hypotheses—themselves complex forms such as techn¯e and epist¯em¯e, which serve as the starting point for division, collection also makes these forms “not beginnings but really hypotheses—that is, steppingstones and springboards—in order to reach what is free from hypothesis at the beginning of the whole (t¯en tou pantos arch¯en)” (cited above). That is, by attending to the differences embedded within an hypothesized complex form, division produces an intelligible multiplicity (the multiple forms revealed through the diairetic search for the target), rather than a sensible one, upon which collection may work in order to synthesize a definition that involves all the forms necessary to account for the nature of the target. The Republic suggests that such definitions are in effect the logoi of causes, ontological recipes, as it were, that serve as the arch¯e for the being of each thing. This ontological beginning point allows logos a communion with the intelligible that is purified of the sensible: “When it has grasped this, argument now depends on that which depends on this beginning and in such fashion goes back down again to an end; making no use of anything sensed in any way” (cited above). As a result, the Republic does not rule out the possibility that, while collection does indeed serve the role of epagogic generalization from sensible instances to intelligible kinds, collection can also offer these intelligible kinds up as hypotheses and starting points for dialectical division. Furthermore, through its purification by encounter with the intelligible many provided by the method of division and through its generalizing procedure collection may wield its purified method within the intelligible, revealing an account for the order of intelligible causes in the target class’s nature. 12 This

maintains consistency with the Stranger’s claim, noted earlier, that division’s ‘purificatory’ work is not based on absolute values, but instead upon pragmatic ones depending on the circumstances of the inquiry. 13 Sayre [21, p. 61] makes a compelling case against the Aristotelianizing view that division is a matter of distinguishing genus and species.

9 Does Plato Have a Theory of Induction …

197

9.9 The God-Given Path As a way of testing the plausibility of my reading of the ‘purified’ method of collection as a fitting elaboration of the ‘upward way’, I complete my argument with a reflection on Socrates’ discussion of the ‘god-given’ dialectical method discussed in the Philebus. There, Socrates describes the doctrine of the one and the many as a foundation for what sounds like the dialectical elaboration of the nature and relationship of forms by means of division and collection: …the ancients, who were better than we and lived nearer the gods, handed down the tradition that all the things which are ever said to exist are sprung from the one and many and have inherent in them the finite and the infinite. This being the way in which these things are arranged, we must always assume that there is in every case one idea of everything and must look (z¯etein) for it—for we shall find that it is there—and if we get a grasp of this, we must look (skopein) next for two, if there be two, and if not, for three or some other number…The gods, then, as I said, handed down to us this mode of investigating, learning, and teaching one another; but the wise men of the present day make the one and the many too quickly or too slowly, in haphazard fashion, and they put infinity immediately after unity; they disregard all that lies between them, and this it is which distinguishes between the dialectic and the disputatious methods of discussion. [17, 16c8–d4, 16e3–17a5]

This proposal recalls the procedure employed after the discussion of purification in both the Sophist and Statesman, which, I have argued, results in the transformation of collection into a means for reevaluating the constrained multiplicity of sub-classes. This ‘modest’ multiplicity is, in the Philebus, characteristic of the dialectical form of discussion that deals with the intelligible alone. An orientation toward a finite multiplicity is here used to distinguish the dialectician from the sophist, who, rather than tarry with the intelligible, instead seeks to drive the inquiry into forms back to the infinite multiplicity of the sensible by highlighting the paradoxes that arise from considering forms alongside their instances. This is the kind of immoderate use of division that the Stranger bemoans in the Sophist as “utterly uncultivated and unphilosophical” [18, 259d9–e2]. To do this is to engage the many as if it were the same as the infinite, rather than to recognize that the intelligible many is itself finite. The view that the many is infinite takes as its principle the many of the sensible world, where, indeed, each sensible thing may be imitated infinitely through images, whether visible or spoken (viz., Republic X, 596d–e). And the infinite many of imitation is indeed the many of the sophist, which provides the key insight by which the final definition of the Sophist begins. Collection, then, stands out as the dialectician’s path of purification, whereby the sensible many may be gathered according to forms, which may themselves be again collected according to the finite (and thereby intelligible) many that figure into the logos of each form. In this way, not only is the intelligible distinguished from the sensible, but so too is dialectic distinguished from sophistry; again, the difference is one of purification. And this is perhaps the greatest lesson of the Eleatic Stranger’s inquiry into the definitions of the sophist, statesman and philosopher: that they may only be distinguished so long as we attend to both the many and the one, moderately and in turn.

198

H. Moore

9.10 Reflections on Epag¯og¯e in Aristotle and Plato By looking to the Eleatic Stranger’s division of the ‘discriminatory’ arts in the Sophist, I have shown that collection is purified through its ongoing interaction with division through the dialectical process, which invokes each method in turn. Returning to the Republic’s account of dialectic reveals that the ‘purification’ of collection consists of its application to noetic objects (“from forms to forms”), rather than sensible ones. Finally, what is otherwise a rather cryptic methodological digression in the Philebus in turn clarifies that this ‘purified’ method of collection operates by gathering into one the plurality of forms produced by division. In this paper, I have tried to show that, with the later dialogues’ discussions of the method of collection, Plato indeed provides an account of epag¯og¯e, albeit one less conspicuous than Robinson may have wished. To conclude, let us now consider the similarities and differences between what I have argued is Plato’s account of epag¯og¯e as collection and Aristotle’s various depictions of epag¯og¯e, for surely, we might not expect them to be precisely the same. First, one of the main advantages of my account of collection as Platonic epag¯og¯e lies in the fact that it aligns with a curious feature of Aristotle’s account of epag¯og¯e, which Robinson does not address. In order to focus on epagogic argumentation, Robinson neglects an explanation of the relationship between Aristotle’s cognitive (see Post. An. I.1, 18, II.19, Nic. Ethics VI.3) and argumentative accounts of epag¯og¯e (see Pr. An. II.23–24, Topics I.12). Aristotle himself links these two aspects of epag¯og¯e: “All teaching and learning come about from already existing knowledge…both deductive and inductive arguments proceed in this way; for both produce their teaching through what we are already aware of, the former getting their premises as from men who grasp them, the latter proving the universal through the particular’s being clear” [3, Post. An. I.1, 71a1–8]. That is, dialectical argument depends upon an individual’s existing epagogic recognition of universals; indeed, all predication is thus dependent upon epag¯og¯e. Moreover, this same connection is drawn in the Philebus, cited above: “The gods, then, as I said, handed down to us this mode of investigating, learning, and teaching one another.” That is, in both Aristotle and Plato’s accounts of epag¯og¯e, the cognitive process by which one comes to know universals (which in Aristotle is the epag¯og¯e depicted in Post. An. II.19 and in Plato is the Phaedrus’ account of collection from sensibles) is itself the basis upon which one might acquire any knowledge through argument. Where Aristotle and Plato of course differ is in the nature of ‘particulars’. The distinction between perception of particulars and knowledge of universals is essential for Aristotle, and the depiction of their connection is by no means fully evident in Post. An. II.19. Plato, on the other hand, despite the rift between the mimetic nature of the sensible world and its origins in the noetic, seems to maintain flexibility in his epistemology because the mechanism by which one comes to know the nature of sensibles (collection) is the same as that by which one recognizes (learning) and points out (teaching) general features common to multiple forms. That is, whereas Plato’s metaphysics is saddled with dualism, while his epistemology is not, in order

9 Does Plato Have a Theory of Induction …

199

to avoid dualistic metaphysics, Aristotle’s epistemology ends up bearing the burden of accounting for the gulf between particulars and universals. In conclusion, it now seems ironic that what Robinson viewed as damning counterevidence against collection being considered as Plato’s elaboration of the ‘upward way’—namely, that collection is depicted as concerned with sensibles—is in fact a feature of its superiority as an account of epag¯og¯e. For, by viewing the method of collection as a generalizing operation, we discover that its logical form maps onto Aristotle’s ambiguous account of epag¯og¯e as the name for both the pattern by which we come to know universals on the basis of the perception of sensibles and for the form of an argument that moves from the adducement of cases toward identification of a shared principle among those cases.

References 1. Ambuel, D. (2007). Image and paradigm in Plato’s “Sophist.” Parmenides Publishing. 2. Aristotle. (1976). Aristotle’s “Metaphysics”: Books mu and nu (J. Annas, Trans.). Clarendon Press. 3. Aristotle. (1984). The complete works of Aristotle: The Revised Oxford Translation (J. Barnes, Ed.). Princeton University Press. 4. Ast, F. (1835). Lexicon Platonicum sive vocum Platonicarum index. In Libraria Weidmanniana. 5. Benson, H. H. (2015). Clitophon’s challenge: Dialectic in Plato’s “Meno,” “Phaedo,” and “Republic.” Oxford University Press. 6. Cornford, F. M. (1935). Plato’s theory of knowledge. Routledge & Keegan Paul. 7. De Chiara-Quenzer, D. (1998). The purpose of the philosophical method in Plato’s “Statesman.” Apeiron: A Journal for Ancient Philosophy and Science, 31(2), 91–126. 8. Dorter, K. (1987). Justice and method in the “Statesman.” In Justice, law and method in Plato and Aristotle. Academic. 9. Gulley, N. (1968). The philosophy of Socrates. Retrieved from https://books.google.com/ books?id=qyhXAAAAMAAJ. 10. Ionescu, C. (2013). Dialectic in Plato’s “Sophist”: Division and the communion of kinds. Arethusa, 46(1), 41–64. https://doi.org/10.1353/are.2013.0003. 11. Ionescu, C. (2014). Elenchus and division in Plato’s ‘Sophist’. In Society for Ancient Greek Philosophy Annual Conference. Fordham University, New York, NY. 12. Klein, J. (1977). Plato’s trilogy: “Theaetetus”, the “Sophist”, and the “Statesman.” University of Chicago Press. 13. McPherran, M. L. (2009). Santas, Socrates, and induction. Philosophical Inquiry: International Quarterly, 31(1–2), 61–85. 14. Miller, M. (1990). The god-given way. Proceedings of the Boston Area Colloquium of Ancient Philosophy, 6(1), 323–359. 15. Moore, H. (2015). Animal sacrifice in Plato’s later methodology. In J. Bell & M. Naas (Eds.), Plato’s animals: Gadflies, horses, swans, and other philosophical beasts (pp. 179–192). Indiana University Press. 16. Plato. (1914). Phaedrus (H. N. Fowler, Trans.). Harvard University Press. 17. Plato. (1925). The Statesman, Philebus, and Ion (H. N. Fowler & W. R. M. Lamb, Trans.). Harvard University Press. 18. Plato. (1961). Theaetetus and Sophist (H. N. Fowler, Trans.). Harvard University Press. 19. Plato. (1991). The “Republic” of Plato (2nd ed.; A. Bloom, Trans.). Basic Books. 20. Robinson, R. (1962). Plato’s earlier dialectic (2nd ed.). Clarendon Press. 21. Santas, G. (1979). Socrates: Philosophy in Plato’s early dialogues. Routledge & Keegan Paul.

200

H. Moore

22. Sayre, K. M. (2006). Metaphysics and method in Plato’s “Statesman.” Cambridge University Press. 23. Vlastos, G. (1991). Socrates: Ironist and moral philosopher. Cornell University Press.

Chapter 10

Endoxa and Epistemology in Aristotle’s Topics Joseph Bjelde

Abstract What role, if any, does dialectic play in Aristotle’s epistemology in the Topics? In this paper I argue that it does play a role, but a role that is independent of endoxa. In the first section, I sketch the case for thinking that dialectic plays a distinctively epistemological role—not just a methodological role, or a merely instrumental role in getting episteme. In the second section, I consider three ways it could play that role, on two of which endoxa play at least a co-starring role. But in the third section, I argue that endoxa don’t play that co-starring role, by arguing for a deflationary account of what endoxa are. For that reason, the third view emerges as the most plausible; but the third view is, I’ll suggest, Platonic.

10.1 The Epistemological Role Envisioned for Dialectic in the Topics Immediately after the introductory sketch of the method for acquiring dialectical skill in Topics I.1, Aristotle turns to explain what the point of acquiring it will be. Two uses in particular have some connection with episteme: What follows on what’s been said so far would be to say for how many things, and for which, this treatise is useful. It is useful for three things: for exercise, for encounters with the public, and for the philosophical epistemai…. [1st use for episteme] for the philosophical epistemai, because being able to puzzle through both sides of an issue, we will more easily discern the true and the false in each case. [2nd use for episteme] And besides these (τι δ), [it is useful] for (1) the first of those [principles] concerning each episteme (πρ`oς τα` πρîτα τîν περ`ι καστην ´ ™πιστημην). ´ For it is impossible to say anything about the [first principles] appropriate to the relevant science, [by] starting from them, since the starting-points of every [science] are first. Rather, it necessary to go through the endoxa concerning each of the principles. But this is unique or most appropriate to the dialectical [techne]; for, since it is examinative, (2) it has a way

J. Bjelde (B) Humboldt Universität zu Berlin, Berlin, Germany e-mail: [email protected] © Springer Nature Switzerland AG 2021 J. A. Bjelde et al. (eds.), Essays on Argumentation in Antiquity, Argumentation Library 39, https://doi.org/10.1007/978-3-030-70817-7_10

201

202

J. Bjelde

to the starting-points of all investigations (πρ`oς τας ` a`πασîν τîν μεθ´oδων ¢ρχας ` Ðδ`oν χει).1

The first use is clear enough, for present purposes: dialectical skill improves one’s ability to resolve the sorts of aporiai which are important for specifically philosophical episteme. So I’ll call this the diaporetic use of dialectic. The diaporetic use makes a kind of instrumental contribution to the acquisition of episteme, and for that reason doesn’t count as playing an epistemological role, as I understand what that would mean. To see why, consider the role that prescription lenses play in coming to know things by sight. They clearly enable you to see better, and make you more likely to get visual knowledge. But they do so by changing the evidence you gather: they are instrumental for gathering higher-quality visual evidence. But it is the evidence itself which plays the epistemological role of making your belief count as knowledge. That is why, if you already know that some sign says “STOP,” your belief does not become more well-supported when you don the glasses. In the same way, the diaporetic use of dialetical skill may help you acquire philosophical episteme you would otherwise lack. But it doesn’t do so by making your cognitive state amount to episteme; instead, it is instrumental for carrying out an activity, i.e. puzzling through both sides of an issue, and that activity either itself plays a properly epistemic role, or is instrumental for something else which does. So the diaporetic use of dialectical skill need not play any more of an epistemological role than leisure. Leisure is also necessary for coming to have philosophical episteme, on Aristotle’s account. But no-one would describe it as playing an epistemological role. Now, one might also think that Aristotle employs a kind of dialectical method in philosophy, so that the diaporetic use isn’t instrumental in the same way as leisure and acquiring glasses are. For acquiring dialectical skill, one might think, will enable one to follow the right philosophical method (cf. [14]). And this would be to assign an epistemological role to dialectical skill, if in addition Aristotle thought that following the right method would lead to episteme. But Aristotle is not that kind of proceduralist about episteme. There is no reason to think that, on Aristotle’s account, following a particular method is either necessary or sufficient for episteme. In brief: methodology is not epistemology, and methodological significance does not transmute into epistemological significance. Moreover, suppose that acquiring dialectical skill does make one a better resolver of aporias, or a better arguer. Still, the same might be said of studying math. But studying math will be merely instrumental for acquiring non-mathematical episteme. In the same way, studying dialectic will be of merely instrumental use to acquiring episteme, if the skills acquired are content-neutral.2 So the diaporetic use of dialectical skill seems to me not to assign it any specifically epistemological role.

1 Topics

I.2, 101a25-b4. Otherwise unattributed page references in this paper are to the Topics. while I don’t dispute that dialectical skill following the rules set out in Topics VIII could lead to better argumentative skills, and so to “good rational argument” ([8, p. 89]), this consequence of dialectical skill does not count as a specifically epistemological role for dialectic, in my sense. Nor do other effects of dialectical skill and exercise considered in [8] and [9].

2 Thus,

10 Endoxa and Epistemology in Aristotle’s Topics

203

Many things are helpful for episteme without being part of the story of why some people have episteme and others do not. By contrast, the second use for episteme is more promising. But it is also considerably less clear. Indeed, even to translate the passage as I have is controversial. In particular, Robin Smith takes Aristotle to be talking at (1) about common principles (e.g. PNC) rather than definitions in general, and at (2) takes him to be saying that dialectical techne is unique in that, since it is examinative of the starting points, it has a way (to proceed).3 These readings are mutually supporting: what’s unique to dialectic with respect to the common principles is that it is not domain-specific, as for instance physics is. For instance, physicists are not necessarily in a position to discuss the use of PNC in mathematics, but dialecticians will be. But Smith also gives independent reasons for preferring his understandings of the text at points (1) and (2). At (2), Smith points out quite rightly that Greek idiom for “way to” takes mostly “™π´ι” or “ε„ς;” not, as we have here, “πρ´oς.” But Smith’s preferred reading, on which ´ is no better Greek, for “™ξεταστικη” ´ the “πρ´oς” is controlled by “™ξεταστικη,” would standardly take the genitive. So Greek idiom doesn’t favor one reading over the other. The same goes for Smith’s philological reasons to read (1) as referring to common principles rather than definitions in general. Perhaps it is true, as Smith says, that Aristotle’s idiom for “the principles of each science” would generally ´ ´ omit the “τîν” in the phrase “τα` πρîτα τîν περ`ι καστην ™πιστημην.” But it ´ ´ would also be odd to include the rest of the phrase, “περ`ι καστην ™πιστημην”, if Aristotle meant “τα` πρîτα τîν ¢ρχîν,” as Smith would have it. So philological considerations by themselves don’t push us in Smith’s direction. Moreover, the broader context pushes against Smith’s reading of (1). Much of Aristotle’s Topics is focused on definitions. Books VI and VII are concerned entirely with them. And while books II-V do consider the other three predicables distinguished in book I, Aristotle justifies his attention to two of these (genus and property) on the grounds that they are elements of definitions, remarking that dialecticians rarely concern themselves with genus and property for their own sakes.4 By contrast, nothing in the rest of the Topics seems aimed at discussing common principles. So it seems positively perverse to sever the present passage on the benefits of dialectic with the overwhelming interest in definition we find elsewhere in the Topics. Smith’s reading would sever that connection. The traditional reading of (1) preserves it. A weightier case for Smith’s reading of (2) could be made if the alternative readings would land Aristotle in inconsistency. So Smith writes that “it would be amazing for Aristotle to say that dialectic can actually establish the indemonstrable first principles… given his frequent insistence that dialectic has no power to prove anything” ([16, p. 353]). But why does Smith think that, on the alternative translation, Aristotle would be committed to dialectic establishing or proving anything? 3 See

especially [16, pp. 352–354] for these readings and the justifications for them which I gloss below. 4 At the start of Topics IV. Accidents presumably get left out not because they are of no interest in their own right but because they aren’t elements of definitions; cf. 120b12–15.

204

J. Bjelde

The reason comes in two parts. The first part is about what it would mean for dialectic to have a way to the first principles: namely, that it would help one achieve the right kind of grasp of the first principles. That seems right. For what we are in need of with respect to many first principles is not awareness of their truth, but the right kind of grasp on their truth. We are already aware that triangles are closed three sided plane figures, for instance. What we need, in order to have geometrical episteme, is the right kind of grasp on that truth. At least part of what is entailed by one’s having the right grasp on a truth, for Aristotle, is that that truth be more familiar to one than the consequences we will deduce from them.5 So, for instance, for dialectic to be helpful by having a way to the first principles might mean that it has a way to make those principles more familiar to us than what follows from them. But why would this commit Aristotle to dialectic proving or establishing anything? Perhaps the thought is that anything which is better known than something proved must itself be proved. Thus, whenever some consequences are proved from premises which are better known than them, the premises must themselves also be proved. At any rate, Smith must be assuming something like that, if he thinks that (2) would require Aristotle to be saying that dialectic proves first principles. And Smith would be right, if that assumption were right. For dialecticians do not prove first principles, even when they give valid syllogisms from endoxa to first principles. But the assumption is wholly unmotivated. There is simply no reason to think that Aristotle requires better known premises to be proved, and hence there is no reason to think proofs of first principles are required in order for them to become more familiar to us. So the significance of the fact that dialectic isn’t in the business of proving first principles is not that dialectical plays no epistemological role. Instead, it is that dialectic can play its epistemological role without proving or establishing anything.6

5 Posterior Analytics I.2. 71b20–24; 72a25ff makes clear that this will include being more familiar to the cognizer. 6 I’m intending to distinguish in this paragraph between proving (which I take to be a modern notion), establishing (which I take to be a quasi-technical notion in the Topics, of getting the respondent to assent to the premises of a syllogism), and demonstration in the sense of the Posterior Analytics. To be charitable to Smith, the tradition to which he is responding (e.g. [13], [7]) does often speak as if, on their view, Aristotle thinks dialectic is going to "establish" the definitions - without clarifying what that would mean, or distinguishing it from demonstration or proving. Also against the background of that tradition, Wlodarcyzk defends the thesis that Aristotle in the Topics has a unified view of how definitions can be established, but she does not take this to amount to proving them. Rather, on her view, establishing a definition helps one achieve a "more articulated level of understanding" ([18, p. 191]). While that is one way to understand how dialectic plays its epistemological role, I’m inclined to doubt that there is any one argument which Aristotle takes to achieve the goal, as I explain in the next footnote.

10 Endoxa and Epistemology in Aristotle’s Topics

205

10.2 How Could Dialectical Skill Help There are roughly three ways dialectical skill might make first principles more familiar to us. The simplest of these would say almost what Smith is afraid it would say, namely that dialecticians will reason from endoxa to first principles; those who follow this reasoning will thereby acquire the right grasp on first principles. It would differ from Smith’s bogeyman, however, by denying to this reasoning the status of a proof. But then what about this reasoning would be unique to dialectic? Unless syllogisms employed in dialectical reasoning are unique, it must be the premises which are unique to dialectical reasoning which account for its unique epistemological role: that is, the endoxa. Some special status of endoxa—as acceptable, reputable, or what have you—is simply transmitted to the first principles by the reasoning. This transmission view, however, is for good reasons without defenders.7 Indeed, on the views of Smith’s intended targets, this would be a radical mischaracterization of the epistemological role of dialectic. For, on coherence views like Owen’s and Irwin’s, the role of dialectic is essentially to induce coherence in a set of beliefs.8 It does not do this by reasoning from endoxa to the first principles. Rather, dialectic induces coherence on some set of beliefs, and the first principles get their status by virtue of the place they hold in that belief set. But both Owen’s and Irwin’s views of how dialectic produces coherence are importantly limited. For, on both their views, dialectic produces coherence by producing consistency, when one finds one’s beliefs inconsistent with the conclusion of some deductive argument, and so adjusts one’s belief set to preserve consistency. But not all consistent sets of belief will count as coherent, on any plausible coherence theory of justification (e.g. [4]). It seems pretty plausible, for instance, that coherence among the beliefs in a set requires that the members of set are relevant to each other, and bear rich confirmation relations to each other—confirmation relations that go significantly beyond lack of inconsistency. So, granting to coherence views that dialectical reasoning removes inconsistencies, why think that it will promote not just consistency but also coherence? One reason to think so would be if the belief sets on which dialectical reasoning operated were already restricted, for instance to sets where the primary bar to higher coherence was inconsistency. Take for the sake of a simple example a set of three beliefs: that virtue is knowledge, that knowledge is teachable, and that virtue is not teachable. Making this set consistent by rejecting one belief and retaining the other two plausibly promotes its coherence. 7 Among

those reasons: Aristotle nowhere gives any reason to think we gain knowledge of first principles by giving a single argument. When Aristotle does talk about a dialectician establishing a definition, in Topics VII.3, the sample argument uses premises of a form which do not occur in "conventional wisdom" or "reputable views," and anyway to "establish" a definition in this sense is simply to win one dialectical encounter - not to prove them in any way that is not ad hominem against a very special opponent who happens to admit the premises. 8 [7, p. 49]: "If a dialectical inquiry succeeds, we will have achieved coherence among our beliefs;" cf. [13, p. 242].

206

J. Bjelde

Irwin makes just this move.9 “Strong dialectic… selects only some of the premisses that pure dialectic allows. Strong dialectic requires us to stick to premisses that we have some good reason for accepting.”10 Irwin himself is quite clear that, on his view, endoxa do not constitute the right restricted set, since we will not have good reason to accept all endoxa. But one might make Irwin’s basic move in a more liberal way, and think that dialectic is able to work its magic when it produces consistency in the endoxa—a body of beliefs which, though we will not have good reason to accept them all, nonetheless rise to some other epistemic status like acceptability. Indeed, this is the epistemic cousin of a widespread view about Aristotle’s methodology.11 The basic thought would be that Aristotelian science involves some tidying up of conventional wisdom, and the result has its epistemic status in part because it is tidied up, but also in large part because of what was tidied up, viz. conventional wisdom. One obvious question about this move is whether it would really suffice to produce coherence. For just as not all consistent sets of beliefs are coherent ones, not all restrictions on the sets of beliefs to be made consistent will result in the production of coherent sets. Why think that either the restriction Irwin favors or the more liberal restricton to endoxa will help produce coherence? But I won’t press this question. For at this point one might want to respond by retaining the basic picture, but dropping “coherence” for “consistency.” On that view, the epistemic role of dialectic is to make a body of endoxa consistent. If the endoxa were already privileged—as the coherence theorist has to say anyway—perhaps being rendered consistent is enough to make them more familiar. What I want to argue next, though, is that this would rely on a seriously flawed picture of endoxa—as the transmission view and coherence view also do. For endoxa are not conventional wisdom, nor are they a restricted set of beliefs like the “acceptable” or “reputable” or simply those believed by everyone or most people or the wise. Rather, on my account, the endoxa are simply whatever a dialectical opponent would assent to if asked. If that account is right, then dialectic must play a different epistemological role than transmitting the privileged status of the endoxa, or inducing coherence among them (or some of them). But there’s a contemporary model for the epistemological role of dialectic on which it doesn’t depend on the premises having any special status: a Platonic model. On that Platonic model, dialectical testing makes one’s grasp on first principles stable. It is in that way, I will go on to suggest in the last section, that dialectic helps make the first principles more familiar.

9 Albeit

for other reasons: Irwin takes for granted that Aristotle’s scientific realism is incompatible with a coherence theory of justification. A reader of [4], published over a decade earlier, would not have taken that view for granted. 10 [7, p. 476]. 11 [12] inter alia.

10 Endoxa and Epistemology in Aristotle’s Topics

207

10.3 What Are Endoxa? The late 20th century consensus about endoxa has recently come in for rough treatment at the hands of very able commentators, whose critical work I do not aim to repeat or summarize here.12 Rather, I want to present a view of what endoxa are, and some textual reasons for that view. On that extreme deflationary view, to be endoxon is simply to be something a dialectical opponent would assent to if asked. Thus, on this view, what it is to be endoxon is always at least implicitly relative to a cognizer. The central insight here is Smith’s, on whose own deflationary view “Aristotle’s endoxa are then just various opinions: ‘things that people think’”—and so will accept in dialectical debates.13 That is, as I understand it: what it is to be one of the endoxa is just to be something someone thinks.14 The extreme deflationary view sticks even closer to the verbal assent which is of prime importance for the dialectician, in that it is neutral on the cause of their assent, and so seems to me to have an edge on Smith’s account. Whereas on Smith’s view, to be endoxon to X is to be believed by X, on the extreme deflationary view, X need not believe what they assent to—what counts is that they assent to it, whether because they believe it, or would be ashamed to deny it, or for whatever other reason. These two deflationary views face three objections together, and either would solve three thorny interpretive puzzles. So, for the rest of this section, I will consider the two deflationary views together. One objection is that Aristotle twice gives lists of endoxa which are much more narrowly restricted than on the deflationary views. That would be bad news for the deflationary views if Jonathan Barnes were right that the list gives “Aristotle’s Óρoς or definition of τα` νδoξα… (It does not explain what “τα` νδoξα” means: it defines τα` νδoξα inasmuch as it determines their scope; it lays down the criteria of reputability.)”15 But there is no reason to think the lists determine what does and doesn’t count as endoxon. Instead, I suggest that the lists are intended to help a student of dialectic realize the goal of the Topics: roughly, to find a method to win dialectical debates in both questioner and respondent roles.16 Since humans can’t know what their opponents are likely to believe or assent to, they can’t learn all the things that are endoxon. But they can learn some common things that different groups of people are likely to assent to. And they can prepare for the toughest debates they are likely to have—i.e. with the wise—by learning in particular what the wise think. 12 Recent

entries into the debate: Dorothea Frede’s attack in [6] on [10]; see also [11] & [15]. p. 347]. 14 [17, p. 42]: “As I interpret [endoxos] it is in fact a relative term: a proposition is endoxos with respect to some definite group of persons…” It is not clear that Smith takes all things people think to count as endoxa, since he regards the lists as "specifying the different types of endoxa" [emphasis mine], p. 78. But it is also unclear why he would not do so. 15 [1, p. 500]. 16 One quibble with Smith on this point: the goal is to maximize the student’s dialectical ability, not their performance on each occasion. So there is no reason to think that dialecticians will find "the best argument available for a given conclusion and a given respondent" ([17, p. 55]). Even experts have bad days, and the best all around basketball player might not be the best free-throw shooter. 13 [16,

208

J. Bjelde

So the lists contain only those endoxa which will be most important for the student of dialectic to learn. The reply to this objection helps solve the first puzzle by explaining a difference between the two lists. The puzzle is that two inconsistent answers to the following question seem to be found in the texts—and are certainly to be found among interpreters. Do the paradoxon views of the wise count as endoxa? All the views of the wise seem to count an endoxa in the list in I.1: What seems so to all or to most or to the wise, and either to all of them or most or the most familiar and endoxon - [these] are endoxa.17

So, if we take this list to determine the extension of “endoxon,” as Barnes does in [1], the paradoxon views of the wise are endoxa. But according to Reeve18 they are not—for they seem to be specifically excluded from the otherwise closely parallel list in I.10: And a dialectical premise is the asking of something endoxon either to all or to most or to the wise, and either to all or most or the most gnorimon - provided it is not paradoxon.19

On the deflationary views, of course these paradoxon views are endoxon—to those wise people who hold them. But they are not endoxon to anyone else. And this explains why they are left out of the second list: for the paradoxon views of the wise will be less useful for students of dialectic.20 Since the first list is just intended to give the broadest overview of the subject (which is expanded in other ways in I.10), this exception is a detail that needn’t be mentioned in it – just as the dialectical premises drawn from technai did not need to be mentioned in the initial overview. A second puzzle which the deflationary views resolve concerns Aristotle’s advice about what people should assent to when they are defending the views of others. Aristole makes clear at the end of VIII.5 that they are not supposed to grant what that person would deny. Aristotle’s example is that someone defending Heraclitus’s views should not grant that “contraries cannot be present simultaneously in the same thing.” But if the extension of “endoxon” is fixed by these lists, then that surely is endoxon! So why shouldn’t the respondent admit it? Other explanations are possible.21 But the deflationary views offer a particularly simple and elegant explanation. The Heraclitean should not grant this, because it is not endoxon to Heraclitus. 17 100b21-23: νδoξα δ τα ` δoκoàντα πασιν ˜ À τo‹ς πλε´ιστoις À τo‹ς σoϕo‹ς, κα`ι τoτoις ´ À πασιν ˜ À τo‹ς πλε´ιστoις À τo‹ς μαλιστα ´ γνωρ´ιμoις κα`ι ™νδ´oξoις. 18 [14, p. 158]. 19 104a8-11: στι δ πρ´ oτασις διαλεκτικη` ™ρωτησις ´ νδoξoς À πασιν ˜ À τo‹ς πλε´ιστoις À τo‹ς σoϕo‹ς, κα`ι τoτoις ´ À πασιν ˜ À τo‹ς πλε´ιστoις À τo‹ς μαλιστα ´ γνωρ´ιμoις, μη` παραδoξoς. ´ 20 One might well wonder what the exception comes to: exactly which of the views of the wise will count as paradoxon. Para whose doxon? The present point does not require answering this question; but there are at least some clear cases: if the view is paradoxon to literally everyone else. 21 In particular, interpreters have often thought that VIII.4/5 use "endoxon" differently than the rest of the Topics. This seems to me a mistake, but the issues are too tangled to work out here.

10 Endoxa and Epistemology in Aristotle’s Topics

209

The third puzzle is over whether the set of endoxa is consistent. Dorothea Frede thinks they need not be,22 for there are inconsistent claims on the list of endoxa in EN VII.1, i.e. that the phronimos cannot be akratic, and that the phronimos is akratic (1145b17-19). On the other hand, Reeve assumes that the endoxa must be consistent,23 and would be justified in doing so (though he does not cite this as a reason) by Topics VIII.5, where it is assumed that the contradictory of a thesis is endoxon if the thesis itself is adoxon, and vice versa.24 That would not be the case if the contradictory of a thesis and the thesis itself could both be endoxon. Again, the deflationary views explain why sets of endoxa might both seem to be necessarily consistent but at the same time seem to admit inconsistencies. Since a single individual generally won’t assent to contradictory things in the same counterfactual circumstances, the set of what is endoxon to someone in particular is generally consistent. But things that are endoxon to different people obviously need not be consistent, if either of the deflationary views are correct. Solving these two puzzles, though, would not be enough to prefer the deflationary views, if they could not make sense of two other aspects of Aristotle’s usage. First, they have to make sense of merely apparent endoxa:25 for, as Aristotle says in Topics I.1, “not everything which appears to be endoxon actually is so: for none of the endoxa mentioned has this appearance purely on its surface…”26 There seem to be two ways for something to be merely an apparent endoxon on the deflationary views. On the one hand, one might make a mistake about whether something is endoxon to someone else. For instance, if one is engaged in defending Heraclitus’ views, as envisioned at the end of Topics VIII.5, one is supposed to answer in terms that are endoxon to Heraclitus. It is clearly possible—indeed, quite easy—to make a mistake about what Heraclitus would have assented to. And, on the other hand, it is not always clear to us what we believe or will assent to.27 For instance, it’s not clear to me whether I would assent to the claim that knowledge is teachable, just after assenting to the claim that virtue is not teachable—if I were attracted to the view that virtue is knowledge. Or, for a different kind of case, it’s not clear to me whether I will assent to things whose meaning I don’t fully grasp. For instance, I’m with the Socrates of Plato’s Euthydemus in not knowing how to answer the question “Do you know by means of that by which you have knowledge?”28 So deflationary views have no special trouble making sense of merely apparent endoxa. Second, the deflationary views have to make sense of what Aristotle calls unqualified endoxa, for instance in Topics VIII.5:

22 [6,

p. 189]. p. 159]. 24 159b4-5: ¢δ´ oξoυ μν oâν oÜσης τÁς θšσεως νδoξoν ¢ναγκη ´ τ`o συμπšρασμα γ´ινεσθαι, ™νδ´oξoυ δ’ ¥δoξoν. 25 I am grateful to Colin King for pressing me on this issue. 26 100b26–28. 27 cf. [17, pp. 48–49]. 28 Euthydemus 295b ff. 23 [14,

210

J. Bjelde

Now, the respondent must necessarily take up the argument by agreeing to a thesis that is either endoxon or adoxon or neither; and one that is endoxon or adoxon either without qualification or in a definite way (as to a specific person, to himself, or to someone else).29

Here and in the rest of that chapter, Aristotle appears to be drawing a contrast between being endoxon to someone, and being endoxon not to anyone in particular. Doesn’t that contrast require a non-relative notion of endoxon? In short: no. Being endoxon in an unqualified way might just mean being endoxon to someone with episteme. Compare Aristotle on being familiar in Topics VI.4: And perhaps what is familiar without qualification is not what is familiar to all, but to those in a state of dianoia, just as also what is healthy without qualification is what is healthy to those in a good bodily condition.30

This suggestion is necessarily speculative, but it has two significant things in its favor. First, it makes good sense of how dialectic might be used for teaching and learning, as mentioned at the start of Topics VIII.5. A learner should evaluate what is endoxon not to them, but to the teacher. Since the teacher will have episteme, this means that the learner should accept claims when they are endoxon without qualification. In Aristotle’s opaque and unparalleled formulation: they should accept things which “seem so without qualification.”31 On my view, this thought about teaching is why the talk of being endoxon without qualification arises in VIII.5.32 Second, this suggestion simplifies the relationship between being endoxon and being familiar. For, on this suggestion, they come together, in the sense that what is endoxon without qualification will also generally be what is familiar without qualification: those in good intellectual condition will be the standard for both. That is welcome both because the two terms seem to be roughly equivalent in ordinary pre-Aristotelian Greek,33 and because Aristotle himself supposes in Topics VIII.5 that what is more familiar is also more endoxon. However, Robert Bolton objects to this otherwise plausible connection between being endoxon and familiar, on the grounds that the things that are most intelligible simply [i.e. most familiar without qualification] include as paradigms the first principles of the sciences as such… But these cannot be as such among the more authoritative premises for dialectic. For these need not even be endoxa… endoxa must be things that are accepted.34

29 159a38-b1: ’Aναγκη ´ δη` τ`oν ¢πoκριν´oμενoν Øπšχειν λ´oγoν θšμενoν ½τoι νδoξoν À ¥δoξoν θšσιν À μηδšτερoν, κα`ι ½τoι a`πλîς νδoξoν À ¥δoξoν À æρισμšνως, oŒoν τδ´ι τινι, À αÙτù À ¥λλ. 30 142a9-11: ‡σως δ κα`ι τ` o a`πλîς γνωριμoν ´ oÙ τ`o πασι ˜ γνωριμ´ ´ oν ™στιν ¢λλα` τ`o τo‹ς εâ διακειμšνoις την ` διανoιαν, ´ καθαπερ ´ κα`ι τ`o a`πλîς Øγιειν`oν τ`o τo‹ς εâ χoυσι τ`o σîμα. 31 159b24-25: τα ` δoκoàντα a`πλîς. 32 This is not to say that inquiry will have a different goal; indeed, as I understand VIII.5, it will not. The point is instead that those with episteme—for instance, a teacher—set the standard for what is endoxon without qualification. 33 As Karbowski appreciates in [11]. 34 [3, p. 75].

10 Endoxa and Epistemology in Aristotle’s Topics

211

Bolton’s worry is that the things that are familiar without qualification—the first principles—may not be endoxa at all, and so a fortiori not endoxon without qualification. That worry does not, however, apply to the deflationary accounts. For on those accounts, being endoxon without qualification need not depend on being actually accepted—if no-one has episteme. And if someone does have episteme, then on the present suggestion anything endoxon without qualification will be endoxon—to them. In addition, Bolton assumes that what is endoxon without qualification will also be “most endoxon.” This is unwarranted. In my view, being more/most endoxon and more/most familiar are distinctions between things that are endoxon to some particular person. The rough idea is that what is more endoxon or more familiar is closer to the center of a Quinean “web of belief.” That is, things that are more endoxon or familiar are things that someone would be less likely to give up. Roughly: if someone becomes convinced that p and q are inconsistent, and they give up q rather than p, then p was more endoxon to them than q. Something like this way of making out what it is to be more familiar is needed to make sense of one of Aristotle’s reasons for thinking that first principles must be more familiar to those with unqualified episteme: that otherwise those with unqualified episteme might be persuadable.35 Regardless, it is clear that on my account something can be most endoxon to an individual, without being endoxon without qualification. That will be true whenever the individual lacks episteme. So the apparent problems for the deflationary views are easily and fruitfully solved, and they enjoy a considerable interpretive advantage besides, in virtue of solving the puzzles above. But, on either of the deflationary views, endoxa have no special epistemic status.36 For that reason, the transmission and coherence views considered above must be wrong, since they rely on endoxa having some special status. So in the next section I consider the obvious alternative: that dialectic plays an epistemological role in Aristotle which is fundamentally Platonic.

10.4 The Platonic Account On Aristotle’s view, we generally start seeking for episteme with the first principles less familiar to us than the things that follow from them. Coming to have episteme is partly a matter of reversing that situation, so that the first principles come to be more familiar to us than the things that follow from them. Exercising dialectical skill plausibly accomplishes this transition. So it plays more than an instrumental role in coming to have episteme, since it ensures that a necessary condition on episteme is satisfied. 35 Posterior

Analytics 72a25-b4, esp. b3-4. course, on this account, things that are endoxon without qualification may have some special status - but it will be possible to believe them without your belief having a correspondingly privileged status. Moreover, not all endoxa will be endoxon without qualification.

36 Of

212

J. Bjelde

But it doesn’t accomplish this by giving arguments which have the first principles as their conclusions, as the transmission view would have it. Nor does it do so by puzzling through difficulties and crowning the endoxa which survive the process, as the coherence view would have it. Rather, I suggest that it does so when it is used for evaluation of first principles. Active witnesses to this dialectical evaluation, who properly appreciate the force of the arguments, may change what they assent to. In particular, they will often be more likely to assent to the winning definitions. Repeated applications of this procedure, if the participants were knowledgeable enough and dialectically skillful enough, will change what the participants and active witnesses would assent to, and by doing so will change both what is endoxon to them and what is familiar to them. In particular, I suggest that this procedure will change what is most endoxon and most familiar to those inquirers into the things that are most familiar and most endoxon by nature. The same procedure would plausibly make those inquirers unpersuadable. And that is why, I suggest, Aristotle avows that those with episteme are unpersuadable, although Jonathan Barnes famously could not “think of any satisfactory argument” for this “academic commonplace.”37 What Barnes misses here is that this is not just an academic commonplace, but a direct consequence of how a particular kind of dialectical meeting changes peoples’ minds. In short, Aristotle thinks that having episteme is partly a matter of having the right things at the center of your web of belief. Dialectical skill enables you to put them there. This view of how dialectic makes things more familiar to inquirers is Platonic in this way, that it is properly speaking the defensive part of dialectical skill which matters for epistemology. Just as in Republic VII those with episteme must be able to defend their accounts in dialectical combat, so too on this view knowers must have their defenses in order, in the sense that they would assent to only the right things in defense of their definitions. That could conceivably be the case quite independently of whether anyone were around to attack their definitions, or whether they were able to attack others’ definitions. Clearly, much more would need to be said to argue that this Platonic view is actually Plato’s. Equally clearly, there is ample scope for Aristotle’s epistemology to differ from Plato’s all sorts of ways. My point here is just that, while Aristotle’s view of the epistemological role played by dialectical is fundamentally different from the transmission and coherence views, it is in this way continuous with a plausibly Platonic view of the epistemological role of dialectic. This way of understanding Aristotle’s epistemology in the Topics underscores the importance of separating methodology and epistemology. For Aristotle’s philosophical method does not extend to the full dialectical evaluation which, on this account, would be necessary for episteme. To count as having episteme, on this view, Aristotle would have to be able to defend the claim that there is change against a Parmenidean opponent, in terms as endoxon to that Parmenidean opponent as possible. He needn’t do that in the Physics, of course, and I suppose one might think Aristotle could have defended it, but for one reason or another we don’t have his defense of it. It is however less plausible that Aristotle’s defense of PNC in Metaphysics  is incomplete, and 37 [2,

p. 103].

10 Endoxa and Epistemology in Aristotle’s Topics

213

that defense certainly does not seem to proceed from premises which are as endoxon as possible to Heraclitus or Gorgias. So it seems to me that Aristotle’s philosophical method does not require the kind of dialectical testing that his epistemology in the Topics does. Perhaps Aristotle lowers his philosophical standards, or changes his epistemological views; but it seems to me more likely that Aristotle regards his treatment of philosophical questions as falling short of generating episteme in his students. The Platonic view of the epistemological role of dialectic in the Topics is not a complete epistemological story: it is a story about satisfying one necessary condition on episteme. That is clearly compatible with there being other necessary conditions. For instance, in the case of enmattered things, there might also be perceptual preconditions. Where dialectic helps cement our grasp on the right choice of definition from among a given set of options, perception might determine those options. After all, perception clearly has to play some role in why we might entertain the definientia “rational animal” and “featherless biped,” but would not consider “circular sponge,” when we try our hand at defining humans. So it would be natural for another treatise to complete the story in the Topics. Whether the Analytics in fact completes, or instead replaces this story, is quite another question, which I will take up here only to make the briefest of suggestions by way of closing. Bronstein’s reevaluation of the first question Aristotle deals with in Posterior Analytics II.19 opens up a tantalizing possibility. As Bronstein points out, that question is not “how do we come to have nous of the principles.”38 In fact, Aristotle’s question is “how do the principles come to be familiar.”39 An answer to that question does not replace an account of how first principles come to be more familiar, like the one I’ve suggested we find in the Topics. Instead, it positively requires such an account if it is to be complete.40

38 [5],

Chapter 13. On Bronstein’s positive account, the question "in effect asks ‘how do they begin to become known’" (p. 231). That reading seems to me quite strained in comparison with the natural reading I present here, though of course Bronstein’s reading is partly motivated by considerations of fit—i.e. how II.19 fits with the rest of Posterior Analytics. Whether Bronstein’s view leaves room for an epistemological story like the one here is a difficult question; it explicitly leaves room for dialectic only to do the work, prior to real scientific inquiry, of discovering that genuses exist (p. 188). 39 Posterior Analytics II.19, 99b17–8: περ`ι δ τîν ¢ρχîν, πîς τε γ´ινoνται γνωριμoι. ´ 40 Thanks for helpful feedback on this paper are due to audiences in Berlin and Gothenburg and especially to Jakob Fink, Ronja Hildebrandt, Colin King, Giouli Korobili, Ana Maria Mora Marquez, David Merry, Stephen Menn, Robert Roreitner, Christopher Roser, Johanna Schmitt, Gisela Striker, and Benjamin Wilck.

214

J. Bjelde

References 1. Barnes, J. (1980). Aristotle and the methods of ethics. Revue Internationale de Philosophie, 34(133/134), 490–511. 2. Barnes, J. (1994). Aristotle: Posterior analytics. Clarendon Press. 3. Bolton, R. (1990). The epistemological basis of aristotelian dialectic. In D. Devereux & P. Pellegrin (Eds.), Biologie, Logique Et Metaphysique Chez Aristote (pp. 57–105). Editions du Centre National de la Recherche Scientifique 4. Bonjour, L. (1976). The coherence theory of empirical knowledge. Philosophical Studies, 30(5), 281–312. 5. Bronstein, D. (2016). Aristotle on knowledge and learning: the Posterior analytics. Oxford University Press. 6. Frede, D. (2012). The endoxon mystique: What endoxa are and what they are not. Oxford Studies in Ancient Philosophy, 43, 185–215. 7. Irwin, T. (1988). Aristotle’s first principles. Clarendon Press. 8. Kakkuri-Knuuttila, M. (2012). The role of the respondent in Plato and Aristotle. In J. Fink (Ed.), The development of dialectic from Plato to Aristotle. Cambridge University Press. 9. Kakkuri-Knuuttila, M. (2005). The relevance of dialectical skills to philosophical inquiry in Aristotle. Rhizai, 1, 31–74. 10. Kraut, R. (2006). How to justify ethical propositions: Aristotle’s method. In R. Kraut (Ed.), The Blackwell guide to Aristotle’s Nicomachean ethics (pp. 76–95). Blackwell. 11. Karbowski, J. (2015). Complexity and progression in Aristotle’s treatment of endoxa in the topics. Ancient Philosophy, 35, 75–96. 12. Nussbaum, M. (1982). Saving Aristotle’s appearances. In M. Schofield & M. Nussbaum (Eds.), Language and logos (pp. 267–294). Cambridge University Press. 13. Owen, G. E. L. (1961). Tithenai ta phainomena. In S. Mansion (ed.), Aristote et les problemes de methode (pp. 239–251). Symposium Aristotelicum. 14. Reeve, C. D. C. (2012). Aristotle’s philosophical method. In C. Shields (Ed.), The Oxford handbook of Aristotle (pp. 150–168). Oxford University Press. 15. Reinhardt, T. (2015). On endoxa in Aristotle’s topics. Rheinisches Museum für Philologie, 158, 225–246. 16. Smith, R. (1993). Aristotle on the uses of dialectic. Synthese, 96, 335–358. 17. Smith, R. (1997). Aristotle: Topics books I and VIII. Clarendon Press. 18. Wlodarczyk, M. (2000). Aristotelian dialectic and the discovery of truth. Oxford Studies in Ancient Philosophy, 18, 153–210.

Chapter 11

Was Aristotle a Virtue Argumentation Theorist? Andrew Aberdein

Virtue theories of argumentation (VTA) emphasize the roles arguers play in the conduct and evaluation of arguments, and lay particular stress on arguers’ acquired dispositions of character, that is, virtues and vices. The inspiration for VTA lies in virtue epistemology and virtue ethics, the latter being a modern revival of Aristotle’s ethics. Aristotle is also, of course, the father of Western logic and argumentation. This paper asks to what degree Aristotle may thereby be claimed as a forefather by VTA. A tempting way of addressing this question might be to say that since Aristotle was a virtue theorist and an argumentation theorist, surely he must have been a virtue argumentation theorist! But not only is this an invalid argument, it is not even clear that the premisses are true. There is no word corresponding exactly to ‘argumentation’ in Aristotle: λ´oγoς is too broad; συλλoγισμ`oς is too narrow (Rapp & Wagner, in [36], p. 15). So, although Aristotle wrote about many aspects of what we now describe as argumentation, we cannot infer that he conceptualized these aspects as comprising a single field. What’s more, even the familiar perception of Aristotle as paradigm virtue ethicist is open to challenge (Santas, in [39]). So, taken as a purely historical question, the answer to my title must be no. However, my project is an anachronistic one: I hope to demonstrate that a careful reading of some of Aristotle’s work on argumentation is of enduring relevance to VTA. The roots of VTA lie in much grander projects, which I survey briefly in §11.1. §11.2 turns to what Aristotle has to say about the relationship of character to argument, and how that may be of use to VTA. §11.3 addresses one of those things that holds particular promise for VTA: the status of the phronimos as ideal arguer.

A. Aberdein (B) School of Arts and Communication, Florida Institute of Technology, Melbourne, USA e-mail: [email protected] © Springer Nature Switzerland AG 2021 J. A. Bjelde et al. (eds.), Essays on Argumentation in Antiquity, Argumentation Library 39, https://doi.org/10.1007/978-3-030-70817-7_11

215

216

A. Aberdein

11.1 What Are Virtue Theories of Argumentation? I have made a point of referring to virtue theories of argumentation in the plural, since, despite its youth and comparatively small scale, the programme has been undertaken in several distinct ways. That heterogeneity owes something to the range of inspirations behind VTA, three of which I survey below.

11.1.1 Virtue Epistemology One big question confronting any virtue theorist is “What counts as a virtue?”. In virtue epistemology, the standard way of splitting up the possible answers is between reliabilist and responsibilist virtues (Axtell, in [14]). The former are most often associated with Ernest Sosa, who includes as virtues many things which don’t much look like virtues, such as the faculty of sight (Sosa, in [42]). Reliabilist virtues are involuntary, but reliable, knowledge-producing processes. They stand in contrast to responsibilist virtues, notably proposed by Linda Zagzebski (in [45]). These virtues, such as open-mindedness, look more like ethical virtues. They are voluntarily exercised aspects of character that require conscious effort of will. Many virtue epistemologists incorporate both types of virtue into a mixed approach (for example, Battaly, in [16]). When we turn to argumentation, the roll call of virtues may look much the same. Maybe there are argumentation-specific virtues that need to be added. Conversely, some virtues lack obvious relevance to argumentation, so may be downplayed. Or perhaps we will need to specifically fine-tune the virtues for argumentation: openmindedness in arguing and open-mindedness in knowing are not necessarily the same thing, for example (Cohen, in [26], p. 57). But VTA need not depend on novel virtues for novelty. And the disputes over precisely which virtues to include seem to closely track their counterparts in virtue epistemology. The other main dimension distinguishing virtue theories concerns the choice of project. For epistemology, this is a distinction between ‘classical’ projects addressing core questions such as “What is knowledge?”; and projects addressing more neglected questions, such as “What is the value of knowledge?”. Some virtue epistemologists just focus on the former, by trying to find a virtue response to the standard questions; some just look at the latter, trying to see what virtue epistemology can say about the neglected questions; some people apply the virtue apparatus to both sorts of project. The classical project of argumentation does not ask “What is the nature of knowledge?”, it asks “What distinguishes good argument from bad argument?”. But there are neglected questions in argumentation theory too: questions about the value of argumentation, and what makes an argument truly satisfying; about how arguments contribute to human flourishing, and whether we need a more eudaimonistic conception of argument. Once again, one can pursue both projects, as I have tried to do in my own contributions to VTA. Many other people who have some allegiance with VTA focus exclusively on the neglected questions. That is an understandable choice,

11 Was Aristotle a Virtue Argumentation Theorist?

217

since it is where the most obvious dividends lie, but I maintain that a more ambitious project is defensible (for a taxonomy of such projects, see Paglieri, in [34], p. 77). Returning to the question of which virtues to follow, let us review some core virtues of virtue epistemology to see if we can appropriate them for argumentation. Heather Battaly lists as “paradigm” reliabilist virtues “sense perception, induction, deduction, and memory” (Battaly, in [16], p. 644). Induction and deduction have an obvious relevance to argumentation. Indeed, they lend themselves to what one might call ‘easy road’ VTA: accept these as virtues, and any account of argument becomes a virtue theory, but in the trivial sense that if deductive and inductive reasoning are virtues then every theory of reasoning is a virtue theory. This may be an unilluminating approach to VTA (but for further discussion, see my [4]). It seems that any sufficiently interesting VTA will not be strictly reliabilist, but will also recognize some responsibilist virtues. Indeed, Zagzebski’s Virtues of the Mind, although a work in epistemology, not argumentation theory, proposes several virtues that look like virtues of argument. She actually includes “fairness in evaluating the arguments of others”, while virtues such as “intellectual humility” are certainly important to argumentation; likewise “thinking of coherent explanations” and “being able to recognize reliable authority” (Zagzebski, in [45], p. 114). Such virtues could contribute to an understanding of the value of argument. Perhaps they can also recapture the idea of cogency, of what makes arguments good ones. One way such a project might get off the ground would be to define a cogent argument as an argument that a virtuous arguer would make in relevant circumstances and then cash out what a virtuous arguer looks like in terms of the virtues they would need.

11.1.2 Critical Thinking Dispositions If one root of VTA lies in contemporary epistemology, another lies in education. Some of the earliest treatments of critical thinking in educational theory framed it solely in terms of skills. More sophisticated accounts argued that skills should be supplemented with dispositions: tendencies or inclinations to actually use these skills in a way that fosters successful critical thinking. In the hands of such pioneers as Robert Ennis and Peter Facione, this work led to detailed inventories of critical thinking dispositions and instruments whereby they could be assessed (Ennis, in [29]; Facione, in [30]). The debate whether or not mastery of critical thinking involves the inculcation of dispositions was won by the advocates of dispositions: “Virtually all the major theorists of critical thinking … make dispositions central to their accounts” (Bailin & Siegel, in [15], p. 193). This early research was conducted significantly in advance of the emergence of VTA, and seldom drew a direct connection to the use of virtues in other nearby fields, such as epistemology. Nonetheless, the dispositions in question look a lot like virtues. After all, we follow Aristotle in defining a virtue as a disposition (Aristotle, in [11], 1106b35), and these specific dispositions have quite virtue-like qualities:

218

A. Aberdein

The primary disposition consists in valuing good reasoning and being disposed to seek reasons, to assess them, and to govern beliefs and actions on the basis of such assessment. In addition, most theorists outline a subset of dispositions that are also necessary for critical thinking, including open-mindedness, fair-mindedness, independent-mindedness, an inquiring attitude, and respect for others in group inquiry and deliberation. (Bailin & Siegel, in [15], p. 183)

Here’s a more extensive list from Robert Ennis, the doyen of critical thinking literature: 1.

2.

Care that their beliefs be true, and that their decisions be justified; that is, care to “get it right” to the extent possible, or at least care to do the best they can. This includes the interrelated dispositions to do the following: (a)

Seek alternatives (hypotheses, explanations, conclusions, plans, sources), and be open to them;

(b)

Endorse a position to the extent that, but only to the extent that, it is justified by the information that is available;

(c)

Be well-informed; and

(d)

Seriously consider points of view other than their own.

Represent a position honestly and clearly (theirs as well as others’). This includes the dispositions to do the following: (a)

3.

Be clear about the intended meaning of what is said, written, or otherwise communicated, seeking as much precision as the situation requires;

(b)

Determine, and maintain focus on, the conclusion or question;

(c)

Seek and offer reasons;

(d)

Take into account the total situation; and

(e)

Be reflectively aware of their own basic beliefs.

Care about the dignity and worth of every person. This includes the dispositions to: (a)

Discover and listen to others’ view and reasons;

(b)

Take into account others’ feelings and level of understanding, avoiding intimidating or confusing others with their critical thinking prowess; and

(c)

Be concerned about others’ welfare (Ennis, in [30], p. 171).

Much of this looks as much like ethics as epistemology. But so it should: argumentation straddles the two. Argumentation obviously has connections to epistemology, in so far as it is a source of justification for our beliefs; but it also, since it involves interacting with other people, raises questions about how to do that in the best way. So the interpersonal nature of argumentation requires ethical as well as epistemic virtues.

11.1.3 Virtues of Argument VTA emerged in the 2000s, with its own sets of virtues. Daniel Cohen, probably the best-known advocate of VTA, proposes four basic virtues: willingness to engage in serious argumentation; willingness to listen to others; willingness to modify one’s

11 Was Aristotle a Virtue Argumentation Theorist?

219

own position; willingness to question the obvious (Cohen, in [25], p. 64). Cohen’s account is Aristotelian at least in the sense that his virtues are all situated as means between vices of deficiency or excess. This yields a rogue’s gallery of bad arguers, some of them quite recognizable types: for example, the Concessionaire concedes things too readily; the Deaf Dogmatist never listens to anyone; the Eager Believer believes whatever they’re told; and the Unassuring Assurer assures you of things in a way that makes you doubt what you otherwise wouldn’t (for the full picture, see my [3], p. 416). My own list of virtues was directly inspired by Cohen’s: 1.

willingness to engage in argumentation (a)

being communicative

(b)

faith in reason

(c)

intellectual courage i.

2.

(a)

intellectual empathy i.

(b)

(c) (d)

ii.

insight into problems insight into theories

fairmindedness i.

justice

ii.

fairness in evaluating the arguments of others

iii.

open-mindedness in collecting and appraising evidence

recognition of reliable authority recognition of salient facts sensitivity to detail

willingness to modify one’s own position (a)

common sense

(b)

intellectual candour

(c)

intellectual humility

(d)

intellectual integrity i.

4.

insight into persons

iii.

i. 3.

sense of duty

willingness to listen to others

honour

ii.

responsibility

iii.

sincerity

willingness to question the obvious (a)

appropriate respect for public opinion

(b)

autonomy

(c)

intellectual perseverance i.

diligence

ii.

care

iii.

thoroughness (Aberdein, in [1], p. 175).

220

A. Aberdein

(1), (2), (3), and (4) are Cohen’s virtues, but their subdivisions are from other sources (in particular, Zagzebski, in [45], p. 114). Elsewhere, I have also presented these virtues as means between vices (in [3], p. 416). Despite the hierarchical structure of this list, I do not mean to imply that any such taxonomy of virtues could ever be definitive. But, although many of the details of this list are open to dispute, it at least indicates how closely intertwined VTA can be with more traditional accounts of virtue.

11.2 What Does Aristotle Say About Character and Argument? In exploring Aristotle’s potential as a precursor of VTA, two broad strategies present themselves. One option would be to analyse Aristotle’s own argumentational practice for traces of an implicit virtue theory. I shall not take that tack in this paper (but, for an example of how this might work, applied to Plato, see Aikin & Anderson, in [8]). The other prospect is to examine what Aristotle explicitly asserts about argumentation. This significantly narrows the focus of the enquiry: Aristotle has a lot to say about virtue and character, but little of it is in the context of logic. Indeed, I am aware of no direct virtue references in any of his works on formal logic (the Categories, On Interpretation, or the Prior Analytics). However, the Topics and the Rhetoric, where he addresses informal logic, are richer sources.

11.2.1 Topics VIII The Topics is an early work and focussed on a somewhat artificial task: success in competitive debate exercises intended for students. This may raise the concern that any insights gleaned from it will be of narrow application. However, as Sara Rubinelli remarks, “as a result of designing this method, which basically comprises the rules of a game, [Aristotle] created the first treatise on argumentation theory of the Western world!” (Rubinelli, in [38], p. 7). Catarina Dutilh Novaes goes further, arguing that rethinking the normativity of logic in explicitly dialogical terms, thereby essentially returning to the perspective of the Topics, provides a more attractive account of the relationship of logic to reasoning than the more familiar monological picture (Dutilh Novaes, in [28], p. 596). There is not a great deal of discussion of character or virtue in the Topics, but there are some significant passages. I shall focus on four. With regard to the giving of answers, we must first define what is the business of a good answerer, as of a good questioner. The business of the questioner is so to develop the argument as to make the answerer utter the most extravagant paradoxes that necessarily follow because of his position: while that of the answerer is to make it appear that it is not he who is responsible for the absurdity or paradox, but only his position: for one may, perhaps,

11 Was Aristotle a Virtue Argumentation Theorist?

221

distinguish between the mistake of taking up a wrong position to start with, and that of not maintaining it properly, when once taken up. (Aristotle, in [12], 159a16–24)

Here Aristotle addresses the contrasting strengths of the two parties to the sort of semi-formal debate he is discussing: the questioner and answerer, or proponent and respondent. Although he does not invoke virtue directly here, he does discuss what is required to play each role well. He also draws a distinction between the arguer and the position defended by the arguer; something characteristically overlooked in much argumentation theory, but of greater salience in VTA. This distinction is also key to the following passage: He [the answerer] should beware of maintaining an implausible hypothesis … which a bad character would choose, and which [is] opposed to men’s wishes (e.g. that pleasure is the good, and that to do injustice is better than to suffer it). For people then hate him, supposing him to maintain them not for the sake of argument but because he really thinks them. (Aristotle, in [12], 160b17–23)

On the one hand, this passage stresses the importance to the arguer of not being thought vicious, or in other words, of (being seen as) arguing as a virtuous arguer would argue, a core principle of VTA, as we have seen. On the other hand, the emphasis is firmly on the vicious aspects of the hypothesis, not the argument by which that hypothesis is defended, so this is not a virtuistic analysis of argument as such. Moreover, it is presented in instrumental terms: implausible hypotheses are to be avoided for fear of reputational damage, not because they are inherently wrong. However, the concern with reputation is consistent with its playing a role in argument evaluation. More specifically, it raises a question that has recently attracted interest from a VTA perspective: can a virtuous arguer be a devil’s advocate? (Stevens & Cohen, in [43]; see also Callard, in [23]). There is more in Topics VIII about bad argumentational practice, some of it perhaps inveterate, and thereby rising to the level of vice: Accordingly it sometimes becomes necessary to attack the speaker and not his thesis, when the answerer lies in wait for the points that are contrary to the questioner and becomes abusive as well: when people lose their tempers in this way, their argument becomes contentious, not dialectical. (Aristotle, in [12], 161a21–23)

Focussing the attack on the speaker indicates that Aristotle sees ad hominem tactics as at least sometimes justified. This is a prerequisite for VTA, since appraising arguments in terms of the arguers’ characters cannot get off the ground if all such appraisal is deemed fallacious (see my [2], p. 89, in which I argue that ad hominem arguments are legitimate if and only if they address specifically argumentational character flaws). Other passages suggest more strongly that some arguers are more prone to bad practice than others, again reinforcing an analysis of argument in terms of vices: Do not argue with every one, nor practise upon the man in the street; for there are some people with whom any argument is bound to degenerate. For against any one who is ready to try all means in order to seem not to be beaten, it is indeed fair to try all means of bringing about one’s conclusion; but it is not good form [εÜσχημoν]. Therefore the best rule is, not lightly to engage with the man in the street, or bad argument is sure to result. For you see how in practising together people cannot refrain from contentious argument. (Aristotle, in [12], 164b7–15)

222

A. Aberdein

As Robin Smith paraphrases, “The problem with engaging in this kind of argument is that it encourages bad habits of argumentation” (Smith, in [41], p. 163). This concern to avoid bad argument, even at the cost of avoiding arguing altogether, foreshadows an important theme in VTA of attending closely to the value of arguments, and more specifically contradicts the Socratic commonplace that one should never give up on arguing (see Campolo, in [24], p. 721; cf. my [5]).

11.2.2 Rhetoric I The Rhetoric is a richer source for insights relevant to VTA than the Topics. I shall concentrate on two passages, one from I.2, one from II.1. I should reiterate that I am quarrying Aristotle, not proposing a definitive reading. Indeed, some scholars suggest that the Rhetoric exhibits traces of more than one stage of Aristotle’s thought, and therefore cannot sustain such a reading (let alone a synoptic reading with Aristotle’s other virtue theoretic texts) (Fortenbaugh, in [31], 232 ff.). Of those proofs that are furnished through the speech there are three kinds. Some reside in the character of the speaker, some in a certain disposition of the audience and some in the speech itself, through its demonstrating or seeming to demonstrate. Proofs from character are produced, whenever the speech is given in such a way as to render the speaker worthy of credence—we more readily and sooner believe reasonable men on all matters in general and absolutely on questions where precision is impossible and two views can be maintained. But this effect too must come about in the course of the speech, not from the speaker’s being believed in advance to be of a certain character. Unlike some experts, we do not exclude the speaker’s reasonable image from the art as contributing nothing to persuasiveness. On the contrary, character contains almost the strongest proof of all, so to speak. (Aristotle, in [13], 1356a)

This passage says several things encouraging for VTA. Firstly, the “character of the speaker” and the “disposition of the audience” have been made equal partners with what is actually said. Secondly, the particular relevance of character to “questions where precision is impossible and two views can be maintained” indicates we are concerned with informal reasoning, and more narrowly, pro and con arguments, or “conductive argumentation” (Blair and Johnson, in [19]). VTA are likewise aimed at informal argumentation. And, at the very least, this passage shows Aristotle taking ethotic argument very seriously: “almost the strongest proof of all”. As Gregory Beabout comments on this passage, “In other words, the pursuit of mastery in the art of rhetoric involves practicing the virtues of character and intellect. In order to be persuasive, a speaker must know how, in the speech, to present oneself as a trustworthy person with good judgment” (Beabout, in [18], p. 161). This passage also raises an intriguing issue: proofs from character “must come about in the course of the speech, not from the speaker’s being believed in advance to be of a certain character”. There’s quite a lot to be said about this. Firstly, there seems to be some tension with the concern with the speaker’s reputation that we saw exhibited in Topics VIII. Perhaps this should not be surprising: as Janja Žmavc points out,

11 Was Aristotle a Virtue Argumentation Theorist?

223

Aristotle is also at odds with an earlier conception of ethos, sometimes referred to as ™πιε´ικεια, as revealing an already established character (Žmavc, in [46], p. 184). This suggests that Aristotle departed in the Rhetoric from a traditional view he had implicitly endorsed in earlier work (More tentatively, an analogy might be drawn between the temporally united conception of ethos in the Rhetoric and the unities of time and place proposed in the Poetics [Aristotle, in {22}, 1451a32]). Some modern authors explicitly deploy both conceptions of ethos. For example, Ruth Amossy: “the pragmatists’ ethos, descended from Aristotle, is constructed within verbal interaction and is purely internal to discourse; the sociologists’ ethos, on the other hand, is inscribed in a symbolic exchange governed by social mechanisms and external institutional positions … however, these two approaches can be complementary rather than conflictual” (Amossy, in [10], p. 5). Aristotle’s exclusion of prior ethos seems a departure from some (tacit) assumptions in VTA, but maybe a valuable one. One immediate argument in favour of a no track-record approach is that track record requires keeping track! That is, it would simplify the process of character-based argument evaluation if only immediately available character aspects need be taken into account. Moreover, as we saw above, VTA must reject the twentieth-century textbook tradition that any reference to an arguer’s character is automatically an ad hominem fallacy. But that’s not to say that ad hominem arguments are never fallacious; rather, we now need to draw a line between the fallacious and non-fallacious cases. Aristotle’s no track-record constraint may inspire a compromise proposal for the location of that line: reference your interlocutor’s character all you please, but only as it is presented on this occasion—no harping back to previous encounters. However, this may be too restrictive: perhaps some legitimate ad hominem arguments would be ruled fallacious. Consider this example from Larry Powers: “I would strongly urge that you not listen to him. You know in advance that his argument is going to sound extremely convincing and compelling. His arguments always sound extremely convincing and compelling, quite independently of the truth or falsity of what he is arguing for” (Powers, in [35], emphases added). Powers defends the extreme position that no ad hominem arguments are fallacious, but one need not endorse that position to see that track record might be relevant to the determination of whether an ad hominem argument is legitimate. Furthermore, a no track-record constraint may raise difficulties for the definition of argument. Communication theorists speak of “serial arguments”: arguments that basically go on forever, like a couple who have been arguing about whose turn it is to do the washing up for forty years (Trapp & Hoff, in [44]). There are many examples of such interminable arguments: the same participants engaging each other on the same issues over a long period, often with the same manoeuvres. If aspects of a speaker’s character revealed in a serial argument are fair game throughout that argument, then track record seems to be back. Indeed, one serial argument might outlast several regular arguments; why should track record be admissible in the former but not the (more temporally united) latter? Indeed, the VTA-adjacent argumentation theorist Michael Gilbert has explicitly endorsed a track-record account of ethos: “An individual’s ethotic rating [ER] comes first and most assuredly from previous

224

A. Aberdein

interactions. Even when encountering someone for the first time the associations they carry, the context they bear, and the situation in which that encounter ensues all form a basis for at least a preliminary ER. Who introduced you, the purpose of the meeting, it’s importance to you, the initial power standings of those involved, all serve to create an initial tentative ER” (Gilbert, in [32], p. 468). On balance, this seems the best way forward for a treatment of ethos in VTA.

11.2.3 Rhetoric II Everything in the above passage from I.2 is consistent with an account of character as veridical. That is, it could be construed as assuming that the audience’s perception of the speaker’s character is always an accurate one. Less so the next passage, from II.1, which displays a greater emphasis on audience perception of character as something potentially distinct from the speaker’s actual character: But since the objective of rhetoric is judgement (for men give judgement on political issues and a court case is a judgement), we must have regard not only to the speech being demonstrative and persuasive, but also to establishing the speaker himself as of a certain type and bringing the giver of judgement into a certain condition. For this makes a great difference as regards proof, especially in deliberative oratory, but also in court cases—this appearance of the speaker as to be of a certain kind and his making the audience suppose that he is disposed in a certain way towards them, and in addition the condition that they are disposed in a certain way to him. Now the appearance of the speaker to have a certain character is more useful for political oratory, and the given disposition of the audience for the courts. For things do not seem the same to those who love and those who hate, nor to those who are angry and those who are calm, but either altogether different or different in magnitude. For to the friend the man about whom he is giving judgement seems either to have committed no offence or a minor one, while for the enemy it is the opposite. And to the man who is enthusiastic and optimistic, if what is to come is to be pleasant, it seems to be both likely to come about and likely to be good, while to the indifferent or depressed man it seems the opposite. There are three causes of the speakers themselves being persuasive; for that is the number of the sources of proof other than demonstration. They are common sense [ϕρ´oνησις], virtue [¢ρετη] ´ and goodwill [εÜνoια]. For men lie about what they are urging or claiming through either all or some of the following: they either have the wrong opinions through stupidity, or, while having the correct opinions through perversity they fail to say what they think; or they have common sense and integrity but are not well-disposed, whence they might not give the best advice, though they know it; and there are no other causes besides. So it must be that the man who is thought to have all of our first list is persuasive to the audience. Now the means of appearing to have common sense and integrity can be drawn from the distinctions we have given in connection with the virtues. For from the same points one might make both oneself and another seem to be of this kind. But about goodwill and friendship we must speak in the discussion of the emotions. (Aristotle, in [13], 1377b–1378a)

As Antoine Braet observes, the Rhetoric is “a nearly totally descriptive treatise of the means of persuasion” (Braet, in [20], p. 316). As such, it need not concern itself with whether an appearance of virtue may be misleading. But this turn to the audience perception of character raises a problem for VTA when understood as

11 Was Aristotle a Virtue Argumentation Theorist?

225

making evaluative claims. Even if there was no more to VTA than an exhortation to argue more virtuously, then recognizing the vices of our interlocutors would still seem a prerequisite. Aristotle’s solution to this conundrum would be that the phronimos would not be fooled by a mere appearance of virtue: “it is generally accepted that the good man’s view is the true one” (Aristotle, in [11], 1176a). (For a challenge to this aspect of the phronimos, see Baumtrog in [17]). The main attraction for VTA in this passage is that Aristotle introduces a threefold analysis of the role of character in argument: “common sense [ϕρ´oνησις], virtue ´ and goodwill [εÜνoια]”. He then gives us the flip side [in the moral sense, ¢ρετη] of this, the failure modes of these three causes: stupidity, a failure of ϕρ´oνησις; ´ and ill-will, a failure of εÜνoια. For William Fortenperversity, a failure of ¢ρετη; baugh, Aristotle’s use of this threefold account acknowledges established Greek practice: the division is “a tradition that goes back as far as Homer” (Fortenbaugh, in [31], p. 216). As such, he suggests that ϕρ´oνησις in particular should be read in its everyday sense here, “wisdom” generally construed, rather than as a term of art in Aristotle’s ethical theory (Fortenbaugh, in [31], p. 220). If nothing else, this explains how anyone could have ϕρ´oνησις and ¢ρετη´ but lack εÜνoια. Nor need we assume that Aristotle’s threefold analysis is exhaustive: “if e¯ thotic argument is sometimes justified in nondeliberative contexts, then some other aspects of Ãθoς, above and beyond the three mentioned by Aristotle, some more intellectual ones for example, will be of importance” (Brinton, in [21], p. 255). Aristotle also acknowledges in this passage the distorting potential of friendship, a correlate of εÜνoια; later passages which discuss character suggest the speaker may be most persuasive by reflecting back the audience’s character, for better or for worse, “since all men accept speeches directed at their own age and their kind” (Aristotle, in [13], 1390a). One remedy which VTA might adopt to this issue would be to constrain ¢ρετη´ and εÜνoια to their argumentational aspects. VTA can productively reflect on all three of the components of character Aristotle identifies. I will pay particular attention to ϕρ´oνησις in the next section; but ¢ρετη´ and εÜνoια raise issues it does not. In principle, all three could be exemplified by the arguer or directly argued for, although in each case “one can better suggest ethos than prove it because attempting to prove ethos would produce doubt about the ethos!” (Braet, in [20], p. 312). Even so, the arguer may successfully persuade an audience by exemplifying aspects of ¢ρετη´ and εÜνoια that are strictly irrelevant to the argument. Goodwill is a neglected feature of argument in both ancient and modern sources; one ambition for VTA should be to remedy this oversight. As Alan Brinton observes, “What ¢ρετη´ is in a moral advisor, εÜνoια is in a prudential advisor” (Brinton, in [21], 254 f.). In his apparently independent account of ethos, mentioned above, Gilbert also proposes a threefold analysis, in terms of knowledgeability, trustworthiness and likability (Gilbert, in [33], p. 275). Each of these components is at least loosely analogous to one of the three Aristotelian components: likability is closed allied to goodwill, trustworthiness is a (moral) virtue, and knowledgeability is at least a component of common sense.

226

A. Aberdein

11.3 Is the Phronimos an Ideal Arguer? Whether or not the phronesis of the Rhetoric is identical to that of the Nicomachean Ethics, it is to phronesis in the latter, stronger sense that I turn in this section. In addition to his well-known roll-call of ethical virtues and their corresponding vices, Aristotle also identifies intellectual virtues: art or technical skill (τšχνη); scientific ´ knowledge (™πιστημη); prudence (ϕρ´oνησις); wisdom (σoϕ´ια), and intelligence (νoàς) (Aristotle, in [11], 1139b16). Of these virtues, ϕρ´oνησις plays the most central role in this project, since the role which Aristotle assigns to the phronimos, the person who exemplifies phronesis, most directly connects to the twenty-first century project of VTA. I think you can make a plausible case that the phronimos is the exemplar of ideal argumentation. In fact, a plausible case has already been made, independently of VTA, by Lois Self and Amélie Rorty: “There are important theoretical and practical relationships between rhetoric and phronesis and it is the man of practical wisdom who has both the capacity and incentive to be an ideal practitioner of the Aristotelian art of rhetoric” (Self, in [40], p. 143); “In short, a brilliant, successful Persuader need not be a phronimos, but a phronimos must— among other things—rightly as well as successfully exercise the skills of a talented Persuader” (Rorty, in [37], p. 733). They make essentially the same point: that all true phronimoi must be ideal arguers, but not vice versa: “the Persuader is not, qua rhetorician, a virtuous phronimos” but “it seems that the phronimos must have the abilities and skills of the best of Persuaders” (Rorty, in [37], p. 721). You could master the skills of the Rhetoric, learn the tricks, without exemplifying the virtue of phronesis. But that wouldn’t make you an ideal arguer in the full sense; you might have great technical fluency and sophistication, but you would fall short of being an ideal arguer because you would be, potentially at least, an unprincipled rogue! “A Persuader can successfully craft an astute and even insightful legal defense for an unjust cause, but he does not qualify as a person of practical wisdom unless his desires and ends are genuinely good” (Rorty, in [37], p. 732). So the model to which we ought really to aspire is someone who knows when to use these skills appropriately. It is that broader, more well-rounded figure of the ideal arguer that we are trying to advocate for in VTA. It is an important observation for VTA that there can be more people involved in an argument than just proponent and respondent; there can also be an audience, an arbiter, and other interested parties (Cohen, in [27], p. 480). If we accept that the ideal arguer should be a phronimos, where does that leave the other parties? This question has given rise to a conflict in the interpretation of Aristotle’s work. For Susan AllardNelson, “An orator, if confronted by a phronimos, could only ‘persuade’ her to do precisely what she would have done, under the same conditions, anyway. The art of rhetoric, in order to be successful and to have a meaningful product, is dependent upon an audience composed mainly of stereotypes (i.e., the average person) and not of the ideal (i.e., the phronimos)” (Allard-Nelson, in [9], p. 258); whereas for Arash Abizadeh, “ethical rhetoric appears to require that phronêsis obtain not just in the rhetorician but additionally in the audience” (Abizadeh, in [7], p. 278). On

11 Was Aristotle a Virtue Argumentation Theorist?

227

Abizadeh’s account, Aristotle requires the audience to be phronimoi; on AllardNelson’s he requires that they not be. Although Abizadeh does not directly address Allard-Nelson’s work, a resolution to the apparent inconsistency may be inferred from his account. He proposes that “the structural constitution of the art of rhetoric produces an internally generated propensity to induce judgments consistent with the outcomes of phronetic deliberation” (ibid.). Rhetoric thereby “enable[s] political institutions to reach correct outcomes despite the ethical shortcomings of the polity’s members” (ibid.). In effect, this envisages an epiphenomenal phronimos, distributed over the multiple participants in a well-conducted dialogue. Hence the paradox is resolved: the audience has phronesis collectively, but lacks it distributively. For Aristotle, phronesis is primarily concerned with deliberation, deciding upon the right course of action. VTA have a broader concern of argumentation in general, so any special role for phronesis in VTA would require a wider construal. Such a move is at least consistent with similar steps taken by some virtue epistemologists: “we ought to consider the virtue of phronesis, or practical wisdom, as a higher-order virtue that governs the entire range of moral and intellectual virtues” (Zagzebski, in [45], p. 229). There is an even broader implication: this privileged role for phronesis lends a privileged role to VTA too, making VTA necessary for a wide range of other projects. If to properly exemplify phronesis you need to be an ideal arguer, then VTA will be central to the virtue theory of anything. In order to properly exemplify the ethical virtues, or the epistemic virtues, or the virtues specific to any other practice, you would first need to be an ideal arguer. So there is potential here for an ambitious role for VTA as the capstone of virtue theories, a common point that they all need to share; that what the true virtue theory needs to have is an account of arguing in a way that is conducive to human flourishing (for further discussion, see my [6]).

References 1. Aberdein, A. (2010). Virtue in argument. Argumentation, 24(2), 165–179. 2. Aberdein, A. (2014). In defence of virtue: The legitimacy of agent-based argument appraisal. Informal Logic, 34(1), 77–93. 3. Aberdein, A. (2016). The vices of argument. Topoi, 35(2), 413–422. 4. Aberdein, A. (2018). Inference and virtue. In S. Oswald & D. Maillat (Eds.), Argumentation and Inference: Proceedings of the 2nd European Conference on Argumentation, Fribourg 2017 (Vol. 2, pp. 1–9). College Publications. 5. Aberdein, A. (2019). Courageous arguments and deep disagreements. Topoi Forthcoming. 6. Aberdein, A. (2020). Eudaimonistic argumentation. In F. H. van Eemeren & B. Garssen (Eds.), From argument schemes to argumentative relations in the wild: A variety of contributions to argumentation theory (pp. 97–106). Springer. 7. Abizadeh, A. (2002). The passions of the wise: Phronêsis, rhetoric, and Aristotle’s passionate practical deliberation. The Review of Metaphysics, 56(2), 267–296. 8. Aikin, S. F., & Anderson, M. (2006). Argumentative norms in Republic I. Philosophy in the Contemporary World, 13(2), 18–23. 9. Allard-Nelson, S. K. (2001). Virtue in Aristotle’s Rhetoric: A metaphysical and ethical capacity. Philosophy and Rhetoric, 34(3), 245–259.

228

A. Aberdein

10. Amossy, R. (2001). Ethos at the crossroads of disciplines: Rhetoric, pragmatics, sociology. Poetics Today, 22(1), 1–23. 11. Aristotle. (1976). Ethics (J. A. K. Thomson, Trans.). Penguin. 12. Aristotle. (1984). Topics (W. A. Pickard-Cambridge, Trans.). In J. Barnes (Eds.), The complete works of Aristotle (Vol. I, pp. 167–277). Princeton University Press. 13. Aristotle. (1991). The art of rhetoric (H. Lawson-Tancred, Trans.). Penguin. 14. Axtell, G. (1997). Recent work in virtue epistemology. American Philosophical Quarterly, 34(1), 1–27. 15. Bailin, S., & Siegel, H. (2003). Critical thinking. In N. Blake, P. Smeyers, R. Smith, & P. Standish (Eds.), The Blackwell guide to the philosophy of education (pp. 181–193). Blackwell. 16. Battaly, H. (2008). Virtue epistemology. Philosophy Compass, 3(4), 639–663. 17. Baumtrog, M. D. (2016). The willingness to be rationally persuaded. In P. Bondy & L. Benacquista (Eds.), Argumentation, objectivity and bias: Proceedings of the 11th International Conference of the Ontario Society for the Study of Argumentation (OSSA), May 18–21, 2016. OSSA. 18. Beabout, G. R. (2013). What contemporary virtue ethics might learn from Aristotle’s Rhetoric. Proceedings of the American Catholic Philosophical Association, 87, 155–166. 19. Blair, J. A., & Johnson, R. H. (2011). Conductive argument: An overlooked type of defeasible reasoning. College Publications. 20. Braet, A. C. (1992). Ethos, pathos and logos in Aristotle’s Rhetoric: A re-examination. Argumentation, 6(3), 307–320. ¯ 21. Brinton, A. (1986). Ethotic argument. History of Philosophy Quarterly, 3(3), 245–258. 22. Butcher, S. H. (1951). Aristotle’s theory of poetry and fine art, with a critical text and translation of the Poetics. Dover. 23. Callard, A. (2019). The devil’s advocate’s advocate. The Point Magazine, online at https://the pointmag.com/2019/examined-life/the-devils-advocates-advocate-agnes-callard. 24. Campolo, C. (2019). On staying in character: Virtue and the possibility of deep disagreement. Topoi, 38(4), 719–723. 25. Cohen, D. H. (2005). Arguments that backfire. In D. Hitchcock & D. Farr (Eds.), The uses of argument (pp. 58–65). OSSA. 26. Cohen, D. H. (2009). Keeping an open mind and having a sense of proportion as virtues in argumentation. Cogency, 1(2), 49–64. 27. Cohen, D. H. (2013). Virtue, in context. Informal Logic, 33(4), 471–485. 28. Dutilh Novaes, C. (2015). A dialogical, multi-agent account of the normativity of logic. Dialectica, 69(4), 587–609. 29. Ennis, R. H. (1996). Critical thinking dispositions: Their nature and assessability. Informal Logic, 18(2–3), 165–182. 30. Facione, P. A. (2000). The disposition toward critical thinking: Its character, measurement, and relationship to critical thinking skill. Informal Logic, 20(1), 61–84. 31. Fortenbaugh, W. W. (1992). Aristotle on persuasion through character. Rhetorica, 10(3), 207– 244. 32. Gilbert, M. A. (2015). Rules is rules: Ethos and situational normativity. In B. J. Garssen, D. Godden, G. Mitchell, & A. F. S. Henkemans (Eds.), Proceedings of ISSA 2014: Eighth Conference of the International Society for the Study of Argumentation (pp. 467–474). Sic Sat. 33. Gilbert, M. A. (2016). Ethos, familiars and micro-cultures. In F. Paglieri, L. Bonelli, & S. Felletti (Eds.), The psychology of argument: Cognitive approaches to argumentation and persuasion (pp. 275–285). College Publications. 34. Paglieri, F. (2015). Bogency and goodacies: On argument quality in virtue argumentation theory. Informal Logic, 35(1), 65–87. 35. Powers, L. H. (1998). Ad hominem arguments. In H. V. Hansen, C. W. Tindale, & A. V. Colman (Eds.), Argumentation and rhetoric. Vale. 36. Rapp, C., & Wagner, T. (2013). On some Aristotelian sources of modern argumentation theory. Argumentation, 27, 7–30.

11 Was Aristotle a Virtue Argumentation Theorist?

229

37. Rorty, A. (2011). Aristotle on the virtues of rhetoric. The Review of Metaphysics, 64(4), 715– 733. 38. Rubinelli, S. (2009). Ars topica: The classical technique of constructing arguments from Aristotle to Cicero. Springer. 39. Santas, G. X. (1997). Does Aristotle have a virtue ethics? In D. Statman (Eds.), Virtue ethics: A critical reader (pp. 260–285). Georgetown University Press. 40. Self, L. S. (1979). Rhetoric and phronesis: The Aristotelian ideal. Philosophy and Rhetoric, 12(2), 130–145. 41. Smith, R. (1997). Aristotle: Topics: Books I and VIII. Oxford University Press. 42. Sosa, E. (1991). Knowledge in perspective: Selected essays in epistemology. Cambridge University Press. 43. Stevens, K., & Cohen, D. H. (2020). Devil’s advocates are the angels of argumentation. In C. Dutilh Novaes, H. Jansen, J. A. Van Laar, & B. Verheij (Eds.), Reason to dissent: Proceedings of the 3rd European Conference on Argumentation, Groningen 2019 (Vol. 2, pp. 161–174). College Publications. 44. Trapp, R., & Hoff, N. (1985). A model of serial argument in interpersonal relationships. Journal of the American Forensic Association, 22, 1–11. 45. Zagzebski, L. (1996). Virtues of the mind. Cambridge University Press. 46. Žmavc, J. (2012). The ethos of classical rhetoric: From epieikeia to auctoritas. In F. H. van Eemeren & B. Garssen (Eds.), Topical themes in argumentation theory (pp. 181–191). Springer.

Chapter 12

Seneca’s Argumentation and Moral Intuitionism David Merry

Walter Sinnott-Armstrong argues that moral disagreement and widespread moral bias pose a serious problem for moral intuitionism. Seneca’s view that we just recognise the good could be criticised using a similar argument. His approach to argumentation offers a way out, one that may serve as a model for a revisionary intuitionism.

Moral intuitionism refers to a family of views according to which belief in at least some moral propositions is justifiable without inference.1 Walter Sinnott-Armstrong has argued against moral intuitionism. He argues that widespread disagreement and extensive bias documented by empirical psychology strongly suggest that the vast majority of our moral beliefs are false. But if the vast majority of our moral beliefs are false, we cannot be justified in believing them without justification, as we need arguments to pick out the small number of true beliefs.2 On this definition, Seneca could retroactively be granted membership in the club of moral intuitionists, for he held that belief in some moral propositions was justified non-inferentially. But Seneca also believed that most of most people’s moral beliefs were badly mistaken, especially about what was good and bad, in the sense of what would or would not benefit them. For he denied that wealth, beauty and pleasure were good, and that pain and death were bad. When Michael Huemer argued that intuitionists tended, unnecessarily, to be on the side of everyday morality, he was not thinking of Seneca [5]. When we look at how argumentation works for Seneca, we can see how he would respond to Sinnott-Armstrong, and his response is deeply challenging, unsettling,

1 This is a particular use of the word ‘intuitions’ which appears to be popular in meta-ethics. It brackets the question of whether intuitions are fast, phenomenologically distinctive, particularly compelling, and so on. 2 Sinnott-Armstrong has developed this argument in a number of different articles, most importantly: [7–9].

D. Merry (B) Institut für Philosophie, Humboldt-Universität zu Berlin, Berlin, Germany © Springer Nature Switzerland AG 2021 J. A. Bjelde et al. (eds.), Essays on Argumentation in Antiquity, Argumentation Library 39, https://doi.org/10.1007/978-3-030-70817-7_12

231

232

D. Merry

and promising.3 For he appears to think that, to deal with disagreement, and to overcome bias, arguments are needed. But he also appears to think that the appropriate arguments are non-inferential. The premises do not provide the support for the conclusion. Rather, the arguments, if used correctly, provide a better and more complete picture of the moral proposition being considered, which enables a more accurate non-inferential judgment of its truth.4 Most intriguingly, in several of Seneca’s Letters to Lucilius (henceforth EM or Letters), namely, EM 23, 82, 83, and 87, Seneca offers a kind of proto-argumentation theory, in which argumentative tropes can be assessed and accepted or rejected, depending on how well they support the non-inferential recognition of the moral truth. This suggests a way in which Senecean intuitionists can systematically study ways of improving moral judgment, and ground a claim to philosophical rigour. Senecean intuitionism would also lead to changes in the practice of ethics which many would welcome.

12.1 Moral Intuitionism and Sinnott-Armstrong’s Objection For the purposes of this paper, I will follow Sinnott-Armstrong in defining moral intuitionism as the view that there are some moral beliefs that can be adequately justified without any inference [7]. Given this definition, moral intuitionism covers a family of views, including the view that some moral facts are knowable a priori, that some moral facts are known perceptually (so that we can hear that a statement is unjust in the same way that can hear that it is ungrammatical) and moral sense theories. This general definition suits my purposes in this paper, because SinnottArmstrong intended his argument against moral intuitionism to work at this level of generality, and because I think Seneca offers a helpful and interesting approach for moral intuitionists in general. Moral intuitionism is often understood as a solution to the problem of regress in morality. Some of our moral beliefs are justified inferentially. But we are only justified in believing them if we are justified in believing the premises of the inferences which support them. If these must be supported inferentially, then we are left with a regress. If some moral beliefs are justified non-inferentially, then we can stop the regress.

3 Other

lines of response have been developed to Sinnott-Armstrong’s argument. Jonathan Smith criticises the details of Sinnott-Armstrong’s case that our moral beliefs are unreliable in [10]. Julia Hermann has questioned Sinnott-Armstrong’s assumption that moral reasoners assign probabilities to their moral beliefs in [4]. As we will see, what is appealing about Seneca’s view is that it, like Huemer’s response, shows how philosophical practice could be altered to take into account the very real concerns that underlie Sinnott-Armstrong’s argument. 4 These ‘arguments’ don’t correspond to many definitions of argument. But their role in Seneca’s philosophical writing—namely, that of rational persuasion—and his systematic study of them alongside things that clearly count as arguments, speak in favour of applying the term.

12 Seneca’s Argumentation and Moral Intuitionism

233

Sinnott-Armstrong distinguishes between a strong and a weak version of noninstrumental justification. According to a weak version of non-instrumental justification, a belief can be non-inferentially justified if we believe it without drawing any inferences, but hold appropriate beliefs from which inferences might be drawn in support of it. I might believe that I will almost certainly be dead by 2080 without drawing the necessary inferences from the beliefs I have—that I would be very old, that medical technology’s ability to increase life expectancy appears limited, that my grandparents were not particularly long-lived people, and that my lifestyle is only moderately healthy. If my justification for moral beliefs rests on being able to draw inferences, even when I have not actually drawn the inferences, then this is not enough to stop the regress problem. Moral intuitionists, Sinnott-Armstrong argues, must then be committed to the view that I can have justified moral beliefs for which my justification rests neither on inferences I have actually drawn, nor indeed on inferences I could draw, given the beliefs I have. Sinnott-Armstrong’s argument against non-inferentially justified moral beliefs draws on two sources of doubt about the truth of our moral beliefs in general. First, he points to widespread moral disagreement. When two people disagree with each other, they cannot both be right. Since people disagree with each other a lot about moral matters, we can conclude that a lot of the moral beliefs people actually do hold are false. Secondly, Sinnott-Armstrong points to a range of experiments indicating that our moral beliefs are subject to a large number of biases. But if we tend to be biased in our moral beliefs, then many of them are likely to be false. From these arguments, we can conclude that a significant number of our moral beliefs are false, which suggests they stand in need of inferential justification. Nathan Ballantyne and Joshua Thurow have questioned the assumption that the unreliability of our moral beliefs places them in need of inferential justification [2]. They argue that the dialectic between Sinnott-Armstrong and the intuitionist should be reconstructed in terms of defeaters and counter-defeaters as follows. The intuitionist claims a moral proposition is true, because it gives a strong appearance of being so, a defeasible justification. Sinnott-Armstrong responds that such appearances are often misleading, defeating the justification. Now, the moral intuitionist will in some cases respond to this defeater, by arguing, for example, that this appearance is particularly widely shared, or is, as Huemer has suggested, against the direction of common biases. Here, they do not replace the original, non-inferential, justification, but rather support its adequacy. They illustrate this point with an example. In their example, McCoy visits the local widget factory and sees a widget that appears red. A stranger approaches him and tells him there are red lights shining on the widgets, undermining McCoy’s evidence that the widgets are red. Another stranger then approaches and informs McCoy that this stranger likes to pull pranks on people visiting the factory, and that there are no red lights at all. McCoy’s belief that the widgets are red is now justified. The important point about this is that McCoy’s justification for believing the widgets are red is still that he can see that they are. Certainly, he needs to establish that this is a situation in which seeing can count as evidence; but that doesn’t show that his perceptually grounded knowledge of their redness is actually supported by an inference.

234

D. Merry

Ballantyne and Thurow’s response to Armstrong is promising, but depends on moral intuitionists being able to defeat the defeater of the unreliability of moral beliefs. One way to do this is to offer some kind of theory about why people’s moral intuitions tend to go wrong, and suggest techniques for overcoming this. Various philosophers with intuitionist leanings have attempted this route. Hume, for example, suggests that we have a tendency to be self-serving, and provides instructions for taking up an impersonal point of view. Moore, similarly, suggests that it is difficult to work out how good something is, because of the phenomenon of organic unities, and so we must consider whether a universe would be good if it had only the thing in question in it. If we have followed these exercises, and can assume those who disagree with us have not, then this may give us grounds for trusting our own views over those who disagree with us. One problem one might have with solutions like that of Moore and Hume is that the argument needed to resolve moral disagreement is generally far more robust and dynamic than simply getting one’s opponent to take a less personal view, or to think about something existing in an empty universe. These seem like opening gambits in a discussion, but if they are all we have, then we don’t seem to get very far. More recently, Huemer has suggested that moral intuitionists can respond by seeking intuitions that respond to five criteria: (1) they are part of a substantial body of intuitions that cohere together well, (2) they are widely shared and do not preferentially favour the practices of one’s own culture, (3) they do not promote evolutionary fitness, (4) they are not self-serving and (5) they do not line up with one’s emotions about the thing in question [5]. Each of these five criteria is desirable, but none of them is a necessary condition for believing an intuition. Huemer’s view is, however, that intuitions that meet these criteria are not undercut by Sinnott-Armstrong’s arguments, as they are neither subject to widespread disagreement nor likely to be the result of bias. Huemer winds up suggesting that intuitionists should focus on highly abstract beliefs. Huemer’s view certainly offers better hopes of having something to say to people who disagree with us than does Moore’s view, or Hume’s. Huemer’s view, however, seems to risk introducing its own biases through over-correction. Sometimes, a belief that lines up with my biases will be quite correct. The virtue of Hume and Moore’s approach is that they offer a corrective to bias which allows for some beliefs in the direction of the bias to be vindicated. For example, I may realise that I am being unfairly self-serving when I think in a dispassionate way about a business transaction I am involved in occurring between two distant strangers, taking into account both perspectives impartially. But my view that I am in the right may also, rightly, be reinforced by such an exercise. Huemer may hope that by focusing on abstract principles, he will be able to show that sometimes our views are right in concrete cases even when they line up with our biases, but the worry here is methodological: Huemer’s approach doesn’t offer the appropriate confidence of resisting over-correction. Seneca’s approach to argumentation offers a prototype for a framework that would allow for rich and robust discussion of moral beliefs, of the sort adequate for responding to widespread disagreement and bias, while allowing for an ongoing discussion and improvement of the approach. In this way, it offers a robust picture of

12 Seneca’s Argumentation and Moral Intuitionism

235

how moral argumentation may be applied in an intelligent and sophisticated way to overcome the shortcomings of an intuitionist framework, without the use of inference.

12.2 How Seneca Argues By Sinnott-Armstrong’s definition, we can count Seneca as an intuitionist. But he offered a radically counter-intuitive moral theory. Although he thought we have noninferential moral knowledge, he also thought most people are the victims of extensive moral illusions, and badly mistaken about the very things which he claims we know non-inferentially.5 If Sinnott-Armstrong’s argument applies to anyone, it applies to Seneca. In this section I unpack the resources in Seneca’s moral philosophy that would allow him to respond to Sinnott-Armstrong’s argument.

12.2.1 The Source of Moral Illusions One of Seneca’s best resources for responding to an argument like SinnottArmstrong’s is a theory about where people’s moral intuitions go wrong, and why. Seneca thought that moral truths were difficult to apprehend, for reasons that connect with his theory of the indifferents. The Stoics divided the world into good, bad, and indifferent things—things that were neither good nor bad. Among the indifferents, the Stoics classed most things people usually imagine are good or bad. Although there was some disagreement even within the Stoa, many placed pleasure and pain, wealth, health, and poverty among the indifferents. Stoics also distinguished indifferents into preferred and dispreferred indifferents. This division is relatively well known. What is less well known is that different Stoics understood the division differently. For Cicero, the distinction is a normative one. Preferred indifferents are to be selected, and dispreferred indifferents are to be rejected.6 But for Seneca, the difference is psychological. It is difficult to reject a preferred indifferent, or to tolerate its absence. It is difficult to accept dispreferred indifferent, or to tolerate its presence. A fully virtuous person has the strength to do so, but most people do not.7 Unlike the Stoics represented by Cicero’s Cato in De Finibus, Seneca thought progress toward virtue made a substantive difference. People who have made some progress towards virtue will be able to hold virtue-like confidence at times, but only inconsistently. Contact with a heavily dispreferred indifferent, such as death, 5 Seneca

does not use the word ‘illusions’, but he does say that we are ‘deceived by the things’, for example at EM 45.5–6. 6 De Finibus III.50–54. 7 EM 71.24–36.

236

D. Merry

may cause even people who have made considerable progress towards courage to despair—a state of mind associated with the vice of cowardice, and actually bad on the Stoic theory.8 But, because the virtue required to die bravely is very uncommon, we tend to see death as intrinsically tied up with vice. We have trouble seeing that death is not bad, because of its association with things that actually are bad. Similarly for preferred indifferents. According to Seneca, wealth is a preferred indifferent. It is easier therefore to keep one’s mind in balance when accepting it than when turning it down. In turning it down, a person who has made considerable progress towards virtue may succumb to bitterness, vanity, or depression. Because we see few examples of truly humble and cheerful rejection of wealth, and more examples of gracious and diligent selection of it, we are likely to come to see wealth as good, as something we ought to select. For Seneca, there is something importantly right to the thought that one doesn’t want to be one of those pious do-gooders: even if we should give our wealth away, we should not do so if we will do it bitterly or smugly. In this way, Seneca would explain why people are so widely deceived by indifferents. Explaining the illusion is not enough, however. One must also show how the illusion is to be undone. In the next section, I argue that Seneca would do this through argument, and in the following section I argue that this argument should be seen as enabling non-inferential moral knowledge, rather than as providing inferential support for moral claims.

12.2.2 Semi-Formal Strategies for Undoing Illusions In this section, I argue that we find in Seneca the beginnings of a theory of argument designed to enable recognition that preferred and dispreferred indifferents were neither good nor bad. These strategies involve drawing the interlocutor’s attention to preferred indifferents in connection with vice, and dispreferred indifferents in connection with virtue, and so correcting the error described in the previous section. In EM 24, Seneca describes how we are to deal with things we are afraid of. Lucilius is facing a lawsuit. Seneca instructs him to list all the bad things that could happen to him, if he were to lose the trial, and then to mention names of people who scorned them.9 This exercise, involving considering these dispreferred indifferents directly, will remove the mask from them, and see that the only fearful thing in them is the fear itself.10 The point of the exercise, then, is to imagine undergoing the bad things, but to imagine undergoing them fearlessly. This allows us to separate the indifferent (death, pain, exile, etc.) from the bad (fear, despondency, etc.), and to get a proper recognition of its value.

8 EM

71.30–34.

9 24.4. 10 24.11–13.

12 Seneca’s Argumentation and Moral Intuitionism

237

My purpose in this paper is not to argue that Seneca’s technique works for eliciting the right moral intuitions, but to argue that it may be the right kind of resource for moral intuitionists. If use of the technique Seneca describes makes dispreferred indifferents appear indifferent, then this helps explain why dispreferred indifferents are commonly seen as bad: we see them so often associated with something actually bad, namely, a cowardly emotional state, that we think they are bad. Our ability to identify what is good and bad is not fine-grained enough to identify the badness in a cowardly death. Seneca can therefore be seen as proposing a technique that allows us to overcome a limitation in our direct grasp of goodness and badness. By describing the technique in general, he allows for its epistemic merits and disadvantages to be discussed. What is particularly interesting about Seneca’s approach is that he systematically develops a set of such techniques. In EM 87, he presents a series of ‘syllogisms’ designed to show that a preferred indifferent is not good. The syllogisms are presented as formal arguments, whose premises are supposed to support their conclusions. However, when Seneca defends them against challenges to the plausibility of the premises, he does not defend the truth of the premises, but uses the premises to vividly outline a situation in which the preferred indifferent is combined with a vice. The five syllogisms Seneca have as premises the following principles: 1. 2. 3. 4. 5.

That which does not make us good is not good. That which can fall to the lot of a terrible person is not good. Good does not result from evil. That which, while we are desiring to attain it, involves us in many evils, is not a good. Things which bestow upon the soul no greatness or confidence or freedom from care are not goods.

In EM 87, Seneca uses these principles to create syllogisms showing that wealth is not a good. He means them to apply more widely, however, and to be generally usable for showing that an apparent good is merely a preferred indifferent. Importantly, as premises in arguments that pleasure, or wealth, are not good, these principles are particularly poorly chosen: Seneca’s opponents are very unlikely to accept them, and indeed, his discussion of each of the syllogisms involves responding to arguments against these premises. There is not the space here to provide a full reading of EM 87, but I will illustrate the point with the example of the fourth syllogism. When we are pursuing riches, the Stoics point out, we become involved in evils. Their opponents present them with a dilemma. Either, the Stoics must mean that riches subject us to misfortune, but in that case the same can be said about virtue: we may become shipwrecked when travelling to undertake studies. Or, the Stoics must mean that riches cause us to commit evil deeds, in which case they ought to be an evil, and not an indifferent. Seneca’s preferred response to this is to follow Posidonius in differentiating between the efficient and the antecedent cause of evil. Bad things are efficient causes of evil. Indifferents may cause us to do evil by tempting us, but they do so only by appealing to weaknesses of our character. This response is unlikely to be very

238

D. Merry

convincing to people not yet convinced by the Stoics’ position, however, who will doubtless maintain that we may very well be tempted to do evil by good things. Indeed, it is exactly because money is good, we might think, that we may be tempted to do evil to obtain it. Seneca uses the form of the syllogism to generate an argument for these opponents: For wealth has this as the efficient cause: it inflates minds, brings about haughtiness, invites envy, and constantly makes a person crazy, because the fame of being wealthy pleases us, although it is harmful. It belongs to good things to be free from all blame; they are pure, they do not corrupt minds, they do not seduce; indeed they raise up and expand minds, but without swelling. Good things create dependability, wealth creates audacity; good things give a mind greatness, wealth gives it arrogance. For arrogance is just a false appearance of greatness.11

Here, Seneca invites his reader to compare the effects on a person of pursuing wealth and pursuing virtue. A person pursuing wealth will become haughty, and will take delight in the—ultimately harmful—reputation for being wealthy. A person pursuing virtue, on the other hand, will become dependable, and will gain a greatness of spirit without becoming arrogant.12 The hope here is that when the reader considers this contrast between how people become when they pursue wealth and how they become when they try to become good people, wealth will start to lose some of its appearance of goodness. Seneca is not here asking for people to draw an inference; he is vividly outlining the difference between wealth and virtue, so that they can see it for themselves. This fits in with the source of moral illusion described in the previous section. If we wish to consider whether wealth is good, we may think of examples of it being used or acquired by people who are already good enough not to become corrupted by it. Certainly many people are capable of making and spending money without becoming arrogant. We may think of people using money to provide security for family members, or to support friends in need, and acting generously. In these cases we will find the generosity of the action difficult to separate from the money involved, and we may wrongly conclude that money is good. It is by seeing how money interacts with bad parts of a person’s character, and comparing this to how virtue might interact with it, that we can see it for what it is. When we think this is what it takes to see what’s wrong with money, we might well think that although we saw directly what was wrong with it, we understand why many people have not—it takes some fairly careful and precise reflection to see the matter clearly. If we are convinced by Seneca on wealth, we are likely to think that many of our beliefs about what is good are mistaken, which may lead to an initial distrust of 11 My

translation of EM 87. Hanc praecedentem causam divitiae habent: inflant animos, superbiam pariunt, invidiam contrahunt, et usque eo mentem alienant ut fama pecuniae nos etiam nocitura delectet [32]. Bona autem omnia carere culpa decet; pura sunt, non corrumpunt animos, non sollicitant; extollunt quidem et dilatant, sed sine tumore. Quae bona sunt fiduciam faciunt, divitiae audaciam; quae bona sunt magnitudinem animi dant, divitiae insolentiam. Nihil autem aliud est insolentia quam species magnitudinis falsa. 12 Presumably we have to be careful here to differentiate a person trying to become virtuous from a person trying to gain approval.

12 Seneca’s Argumentation and Moral Intuitionism

239

them. Seneca’s 87th letter, however, gives us a formula for working through them. We subject them to an exercise based on each of the syllogisms in the letter. If Seneca is right, the appearance that health, fame, and pleasure are good will fade as a result of the exercise. On the other hand, the appearance that virtue is good will be reinforced. As this is a form of the response to Sinnott-Armstrong that intuitions that have been adequately subject to reflection are reliable, it is worth considering his response. He argues that if reflection is adequate, then it will involve a lot of information that can fit into an inferential structure, and so the beliefs in question will turn out only to be non-inferentially justified in the weak sense [9]. Here, however, it is more plausible to see this story about how we come to know that wealth is a preferred indifferent, and not a good, as something we grasp through considering what wealth is like, and what effects it has on people who desire it, rather than as seeing it as something we somehow infer from these effects. For the reflection works to overcome a shortcoming in our abilities to recognise what is good and what is not, rather than supplanting these abilities as a source of knowledge.

12.2.3 The Argumentative Context Seneca used non-inferential argument strategies to help people acquire justified non-inferential moral beliefs. An interesting aspect of Seneca’s approach to these arguments is his rejection of dialectical argumentative contexts aimed at refutation. Such a context, argued Seneca, was not conducive to the acquisition of justified non-inferential moral beliefs. In EM 82 and EM 83, Seneca criticises Zeno of Citium’s dialectical approach to arguing that death is not bad (in 82) and that good people don’t get drunk (in 83). Since the reason good people do not get drunk is that the pleasure of drunkenness is not a good thing, and is perhaps even a bad one, these letters are both relevant to our discussion of non-inferential knowledge of the good and bad in Seneca. Interestingly, Zeno of Citium tried to support the claim that death is not bad, and that the good person does not get drunk, inferentially. Zeno’s syllogism about death is that it is (sometimes) glorious, and what is (sometimes) glorious is not bad, so death is not bad. His syllogism about drunkenness is that you wouldn’t trust a secret to a drunk person, but you would trust a secret to a wise person, and so the wise person does not get drunk. Each of these syllogisms starts from premises that an interlocutor is likely to accept, and these premises are supposed to justify belief in the conclusion. Neither syllogism does much to help the interlocutor form a clear opinion of death, or drunkenness. Seneca has a raft of criticisms of these syllogisms: the one on drunkenness is dishonest; the one on death is too indirect; they are unpersuasive; they are trivial. For our purposes, however, the most interesting criticism he raises against them is that they lead the interlocutor to think about the wrong things. Somebody, or some people, produced parallel arguments to these syllogisms, designed to show that the

240

D. Merry

conclusion of the argument doesn’t follow from the premises.13 Against the drunkenness argument, the parallel is that you would not trust a secret to someone who was asleep, but you would trust a secret to a wise person, so the wise person does not sleep.14 Here, the idea must be that the argument form is invalid, for if the Stoics accept the argument form and the first premise, then they would be stuck in absurdity. The parallel syllogism against the death argument is that death is (sometimes) glorious, and what is (sometimes) glorious is not indifferent, therefore death is not indifferent.15 Here again, the idea is that the parallel argument has the same structure as the original Stoic argument, and premises that the Stoic would accept, leading to a conclusion they would reject, and so the argument must be invalid. Later Stoics came up with responses to these parallels, often by drawing distinctions. Seneca complains about the Stoics perpetuating this practice. He argues that Zeno’s arguments are designed to make an interlocutor feel trapped and compelled into accepting a conclusion before they have really come to believe it.16 The effect of this is to turn the interlocutor’s attention away from the matter at hand, and towards how to escape refutation.17 Seneca urges the Stoics to move away from these arguments, towards other methods: in the case of death, we have seen that this involves thinking about courageous deaths. In the case of drunkenness, this involves thinking clearly about the effects drunkenness has on a person’s deportment and mental states.18 In both cases, the idea is that this will enable non-inferential moral justification. Refuting an opponent, however, will not help such justification to occur. The contemporary moral intuitionist can perhaps take something over from Seneca’s idea. Creating an argumentative context in which the goal is, through discussion, to create as clear, comprehensive and unbiased a view of a candidate for noninferential moral knowledge may be an important discipline to create. Arguments would, in this context, be seen as good if they added clarity, comprehensiveness, and balance to a person’s picture of the item in question, and bad if they did not. An inconclusive argument in this context may still be a very valuable step towards establishing the requisite position to know.

12.3 Updated Senecean Intuitionism Whilst Seneca’s account of why we are badly mistaken in our attributions of goodness and badness are challenging and worth considering, moral intuitionists may be more interested in the overall approach than Seneca’s implementation of it. In this section, 13 Malcolm Schofield has argued, through an appeal to Occham’s razor, that Alexinus was the author of all the Zeno, in [6]. 14 83.9. 15 82.10. 16 82.19. 17 82.22. 18 83.27–36.

12 Seneca’s Argumentation and Moral Intuitionism

241

I outline how this approach may be brought up to date, and how the updated approach would allow moral intuitionists to respond to Sinnott-Armstrong’s objection. On Seneca’s theory, there is a clear explanation of why moral intuitionism is compatible with widespread moral disagreement. Moral beliefs can be justified noninferentially, but this justification is exigent, and moral error is common. Most people simply have not done the work that would allow them to have a reliable set of moral intuitions. Contemporary intuitionists already make a move in this direction; Audi, for example, is careful to make the point that being self-evident is not the same as being obvious, and that moral truths may very well be simultaneously self-evident and non-obvious [1]. But for Seneca, this goes further: not only are moral facts nonobvious , they are extremely difficult to recognise, and it’s easy to wrongly believe that one has been recognised. Disagreement is just what we would expect under such conditions. With this recognised, it falls to the intuitionist to explain how to get good moral intuitions. Seneca’s model of studying arguments aimed at correcting moral bias, in an argument-theoretical way, offers a solution. Contemporary intuitionists should work both to identify techniques that actually do reduce bias, as well as non-inferential fallacies, arguments that appear to reduce bias, but do not actually succeed in doing so. To some extent, such an exercise will likely vindicate that the way we currently do ethics. Good thought experiments might be seen as ways of reducing bias. For example, Judith Jarvis Thomson’s violinist case may be seen as a way of reducing the effects of partisanship on how people think about the abortion debate [11]. Nevertheless, if we follow Seneca in giving prominence to establishing a very clear view of the issue we are considering, then this will make clear some real shortcomings of the way we do ethics at the moment. For ethicists seem all too willing to discuss a wide variety of issues, while retaining only a distant, vague, and abstract understanding of the matter at hand. To take an issue close to Seneca’s example of drunkenness, students in an applied ethics class may be asked to consider arguments for and against drug regulation without first coming to terms with the current state of scientific knowledge about how the drugs function and what their effects are, the experiences of a wide range of users of the drugs and those around them, or the experience and efficacy of rehabilitation programmes. There is no good reason, in such a context, to expect much reliability in students’ non-inferential beliefs about whether or not to regulate drugs or enforce drug control. Another difference with current practice is that ethicists will need to pay more attention to developing psychological evidence about the ways in which we tend to be biased. Here, experimental ethics may play a role, as techniques designed to counteract particular biases may be tested psychologically, although in practice such testing will be difficult. By showing that they hold a set of techniques that are informed by a good understanding of biases, moral intuitionists can better defend their view that they are in a good position to know. Just as anyone who follows Seneca’s exercises in building up a clearer and more comprehensive picture of wealth and death can reasonably argue that their epistemic position relative to these matters has improved relative to how it was before, so could people who have systematically worked to eliminate bias.

242

D. Merry

I will close by considering two objections to this approach. The first is that it seems to hand ethical expertise over to people who are particularly experienced in an area. For example, the opinions about doctors in medical matters will count for more than those of ethicists, as they have a far clearer picture of the issues they are considering. Yet experience in an area is no protection against moral confusion. The first thing to note here is that, although ignorance of a matter may be one source of erroneous intuitions about it, it is far from the only source. A philosopher familiar with a wide range of non-inferential moral arguments might still have something to offer even a far more experienced person, in terms of correcting for bias. Further, some moral knowledge will be inferential, and philosophers will be of considerable assistance in detecting faulty inferences. On the other hand, Senecean intuitionism gives particular respect to the perspectives of people with first hand experience of the situations that ethicists may discuss, and this is a virtue of the approach. The second objection is that the valuing of non-inferential argumentation will undermine one of philosophy’s strongest values, that of critical, sceptical inquiry. The fear is that if philosophers do not insist on demonstrating moral claims, they will merely be moralists [3]. This is why reflection on the quality of non-inferential argument is so important to this project. Senecean intuitionists may use non-inferential arguments in a more collaborative way to enable clear intuitive judgment of moral situations. But they will be careful to limit their use to arguments that actually do assist in getting a clearer, more balanced view of the case at hand. Some non-inferential arguments introduce bias or obscure a matter further; the critical skill of the Senecean intuitionist is being able to tell the difference.

References 1. Audi, R. (2008). Intuition, inference, and rational disagreement in ethics. Ethical Theory and Moral Practice, 11(5), 475–492. 2. Ballantyne, N., & Thurow, J. (2013). Moral intuitionism defeated? American Philosophical Quarterly, 50(4), 411–422. 3. Cooper, J. (2004). Moral theory and moral improvement: Seneca. In Knowledge, Nature, and the Good (pp. 309–335). Princeton University Press. 4. Hermann, J. (2017). Sinnott-Armstrong’s empirical challenge to moral intuitionism: A novel critique. Ethical Theory and Moral Practice, 20, 829–842. https://doi.org/10.1007/s10677017-9822-1. 5. Huemer, M. (2008). Revisionary intuitionism. Social Philosophy and Policy, 25(1), 368–392. https://doi.org/10.1017/S026505250808014X. 6. Schofield, M. (1983). The syllogisms of Zeno of Citium. Phronesis, 28(1), 31–58. https://doi. org/10.1163/156852883X00031. 7. Sinnott-Armstrong, W. (2002). Moral relativity and intuitionism. Philosophical Issues, 12(1), 305–328. 8. Sinnott-Armstrong, W. (2006). Moral intuitionism meets empirical psychology. In T. Hogan & T. M. Timmons (Eds.), Metaethics after Moore (pp. 339–366). Oxford University Press. 9. Sinnott-Armstrong, W. (2011). An empirical challenge to intuitionism. In J. Hernandez (Ed.), The new intuitionism (pp. 11–28). Bloomsbury.

12 Seneca’s Argumentation and Moral Intuitionism

243

10. Smith, J. (2010). On Sinnott-Armstrong’s case against Moral Intuitionism. Ethical theory and moral practice, 13(1), 75–88. 11. Thompson, J. J. (1971). A defense of abortion. Philosophy and Public Affairs, 1(1), 47–66.

Chapter 13

Aristotelian Dialectic, Argumentation Theory and Artificial Intelligence Douglas Walton

Abstract It is shown that Aristotelian dialectic can be analyzed as having two parts: a core formal model that has a formal dialogue structure and a set of ten definable supplementary characteristics that lie outside the core structure. Some current argumentation tools used in artificial intelligence and multi-agent systems are applied to the task of extending the core formal model to include the supplementary characteristics. Using these tools it is explained how the structure of a dialogue can be mapped into an argument diagram that can be analyzed and evaluated using standard argumentation techniques such as argumentation schemes, types of dialogue, critical questions and a dialectical concept of burden of proof. Disputed issues on how the elenchus fits the standard dialogue typology are discussed, and it is concluded that it fits best into a type of dialogue called examination dialogue that is closely related to persuasion dialogue.

13.1 Introduction It is generally accepted that contemporary argumentation theory has its roots in ancient Greek philosophy, and especially in Aristotelian dialectic (Walton and Krabbe, in [48]; Krabbe, in [19]). In this chapter it is shown how the standard dialogue typology used in argumentation theory, artificial intelligence and multiagent systems fits even better than one might have thought with Aristotelian dialectic. Starting from the finding that Aristotelian dialectic can be minimally modeled as a core formal dialogue structure (Krabbe, in [19]), the chapter goes on to show that it can be even more fully captured by a set of ten features surrounding the core. These findings bring out some new aspects of how Aristotelian dialectic fits with and is a predecessor to current formal dialogue systems for argumentation. Argumentation theory is a distinctive philosophical approach that has achieved notable successes by building and applying practical methods to help a user identify, analyze and evaluate arguments in everyday conversational discourse and in more D. Walton (Deceased) (B) University of Windsor, Windsor, ON, Canada © Springer Nature Switzerland AG 2021 J. A. Bjelde et al. (eds.), Essays on Argumentation in Antiquity, Argumentation Library 39, https://doi.org/10.1007/978-3-030-70817-7_13

245

246

D. Walton

structured areas such as debate, law and scientific fields. This undertaking is inherently pragmatic in that it studies not just the semantic properties of arguments but also their pragmatic properties concerning how argument is used in a multiagent conversational framework where groups of agents are putting arguments forward and responding to each other’s arguments. In the simplest case there is a proponent who puts forward an argument or some other type of speech act such as asking a question, and a respondent who replies following procedural rules. The method of argumentation evaluates arguments by weighing the pro arguments against the con arguments to judge, after a sequence of such turn-taking moves, which participant had the stronger argument (Gordon, in [12]). This method is essentially what defines argumentation as a distinct field of study. It is widely accepted that the leading historical antecedent of this approach is the Aristotelian method of dialectic described in Aristotle’s Topics and other logical works. Although there were many attempts during the Middle Ages to try to do something with Aristotelian dialectic (Hamblin, in [16]), after the Enlightenment the subject became unfashionable and was overshadowed by the advent of modern mathematical logic. Now there is a renewed interest in the subject with the rise of argumentation as a field of study. Research on argumentation and Artificial Intelligence (AI) began to consolidate and build mass in the early eighties. The initial efforts were spurred on by finding that adopting the argumentation approach offers a natural way of conceptualizing commonsense reasoning by bringing out its defeasible nature. Different dialogue models have been proposed and applied in argumentation and AI that raise new questions about the roots of dialectical argumentation in ancient Greek philosophy. Computing has advanced argumentation theory as a useful approach for research in many fields of computing such as AI and multi-agent systems (Rahwan and McBurney, in [28]). Recent successes include argumentation-based models of evidential relations and legal processes of examination and evaluation of evidence, argumentation dialogues in multiagent systems, argumentation-based models of evidential reasoning in law, design and the use of argument-based structures for autonomous reasoning in AI, and computer-supported collaborative argumentation, including the implementation of software tools for enabling online argument in domains such as education and electronic government, as well as collaborative learning tools that use argumentation models of deliberation (Simari and Rahwan, in [32]; Prakken, in [27]). These advances are shown in this chapter to make the ancient Greek idea of dialectic better understandable to us, as well as giving it traction for those working in computational argumentation by clarifying historical foundations that can help them to continue to build applicable formal models of computational argumentation. In Sect. 13.2 the reader is given a brief introduction to the seven types of dialogue currently recognized as dialectical structures providing frameworks for different uses of arguments in varying contexts. Later, in Sect. 13.10 of the chapter these will be systematically compared to the different species of Aristotelian dialectic described in the earlier parts of the chapter. In Sect. 13.3 a brief introductory account is given of the different types of arguments and reasoning recognized by Aristotle

13 Aristotelian Dialectic, Argumentation Theory …

247

in his Topics, Posterior Analytics and On Sophistical Refutations. In Sect. 13.4, it is explained how the central core of Aristotle’s procedure of dialectical argumentation can be modeled by the formal dialectical system ACADEMIC1 of Krabbe (in [19]). This model is shown to capture the central core of Aristotelian dialectic through its formal specification of the different kinds of rules and other formal properties of ACADEMIC1 . Section 13.5 specifies ten rules for the practice of dialectic described by Aristotle in the Topics that are not captured by the formalization in ACADEMIC1, even though it remains possible that some or even all of them could be captured by extending Krabbe’s formal dialectical model. The different systems of dialogue specified in these first five sections of the paper can be classified as species of persuasion dialogue, one of the types of dialogue recognized in the current sevenfold typology. Sections 13.6 and 13.7 outline two other types of dialogue: deliberation dialogue and examination dialogue. These types of dialogue are also described by Aristotle in a way that makes them easily recognizable to readers familiar with current argumentation studies. They are fundamentally important in argumentation theory. Deliberation dialogue is about a rational agent making a decision on what to do in a given set of circumstances. Curiously enough, deliberation has recently been formalized in artificial intelligence in a way that is highly evocative of Aristotle’s remarks on deliberation in the Nicomachean Ethics. Examination dialogue is best known to us from its use in law where a witness is cross-examined often with the aim of what is called impeachment in law, trapping the answerer in an inconsistency (Walton, in [41]). But readers of this chapter will identify it with the Socratric questioning technique called the elenchus, and possibly also with Aristotelian dialectic itself. Hamblin (in [17], p. 59) cited three passages in On Sophistical Refutations (169b25, 171b4, 172a23) to suggest that Aristotle is not sure whether dialectical and examination arguments are the same or different. However, now the conjecture is put forward in Sect. 13.7 that examination dialogue fits best as a species of exetastic examination dialogue (Guthrie, in [15], p. 155) where the proponent has the goal of attacking the commitments of the respondent, and most characteristically, specifically carrying out this goal by finding an inconsistency in the commitments of the respondent. Section 13.8 shows how argumentation schemes currently important in formal computational systems of the kind used in artificial intelligence are connected to dialogue structures. Using a simple example of a natural language dialogue on the issue of whether videogames lead to violence, Sect. 13.8 explains the connection between argumentation schemes, of the kind associated with the Aristotelian topics, and formal dialogue structures such as ACADEMIC1 . Using this example, it is explained how the structure of the dialogue can be mapped into an argument diagram (a graph structure) that can be analyzed and evaluated using standard argumentation techniques now widely used in artificial intelligence, such as argument diagrams, argumentation schemes, critical questions and dialectical burden of proof. Displaying this capability of current computational systems of the kind used in artificial intelligence helps us get a much better grasp of the unity of Aristotle’s dialectic

248

D. Walton

by linking it more closely, through connections recently being developed in computational argumentation studies, to his writings on what are essentially argumentation schemes in the Topics. Section 13.9 offers some conclusions. One of these is a classification system displayed as a graph indicating how the current standard dialogue typology relates to the different kinds of dialectic found in Aristotle’s writings considered in the chapter. A second one is a list of ten defining features of Aristotelian dialectic lying outside the core structure of Krabbe’s formal model of the core of dialectic ACADEMIC1 . In Sect. 13.10 it is also suggested how some issues only discussed in a tentative way in the chapter could be further investigated.

13.2 Standard Types of Dialogue in Argumentation Theory The standard dialogue typology (Walton and Krabbe, in [47]; Walton, in [47, p. 9]) for identifying the goals and other characteristics of each type of dialogue is displayed in Table 13.1. This is merely a brief introductory explanation of these types of dialogue. Later on throughout the chapter more details will be given. A persuasion dialogue is defined as a type of dialogue where the initial situation, codified at the opening stage, takes the form of what the Greeks called a stasis. There is an ultimate proposition to be proved by the party designated as the proponent and this proposition is set in place as representing the proponent’s claim at the opening stage. The proponent’s goal is to find an argument that has her claim as the conclusion and that has a set of premises that the other party either accept or can be persuaded to accept. The framework of a persuasion dialogue, just as is the case with all the other types of dialogue, is that Table 13.1 The standard dialogue typology

Type of dialogue

Initial situation

Participant’s goal

Goal of dialogue

Persuasion

Conflict of opinions

Persuade other Resolve issue party

Inquiry

Need to have proof

Verify evidence

Prove hypothesis

Discovery

Need an explanation

Find a hypothesis

Support hypothesis

Negotiation

Conflict of interests

Get what you want

Settle issue

Information

Need information

Acquire information

Exchange information

Deliberation

Practical choice

Fit goals and actions

Decide what to do

Eristic

Personal conflict

Hit out at opponent

Reveal deep conflict

13 Aristotelian Dialectic, Argumentation Theory …

249

the proponent and also the other party, called the respondent, has what is called a commitment set or commitment store (Hamblin, in [16] and [17]). A commitment store is a database containing all of the propositions accepted by either party at the opening stage, and propositions in the database of each participant are continually added to or deleted from as the party makes a move in the dialogue (Prakken, in [27]). They take turns making moves, and each move contains a speech act, such as the speech act of making an assertion or the speech act of putting forward an argument (Prakken, in [26]). The rule for the speech act of making an assertion is that when a participant makes an assertion, the proposition asserted automatically goes into his or her commitment set. The participants actually do not need to be human beings. They can also be automated entities such as agents in a multiagent dialogue conducted on the Internet. I take the critical discussion type of dialogue (van Eemeren and Grootendorst, in [34]), to be a species of persuasion dialogue (Walton, in [47]). One of the most important of the procedural rules for this type of dialogue is the one codifying the requirement of burden of proof. If one party makes an assertion and the other party questions the assertion by expressing doubts about it, the first party must, at its next move, either retract the proposition put forward as its assertion from its commitment set, or provide an argument that supports its acceptance by the other party. This requirement is called the burden of proof (BoP) rule. The kinds of successful arguments used in a persuasion dialogue would not generally be considered to be proofs, in any strong sense of this term appropriate say for uses in a scientific context. In order to be successful, they only need to be based on premises accepted by the agent to be persuaded. In contrast, the goal of an inquiry dialogue is to provide a proof of the proposition designated as the claim or hypothesis at the opening stage. Essentially, the goal of the inquiry dialogue is to present a sequence of argumentation that is so carefully supported by evidence and so carefully shown to be immune to refutation by counterarguments, that there will be no need for it to be retracted in the future, as new evidence comes in Black and Hunter (in [7]). A standard example of the inquiry type of dialogue would be a judicial inquiry on air disaster that has the goal of determining the cause of the crash. Such an inquiry can take years of assembling the evidence and carefully evaluating it to build arguments pro and con hypotheses concerning the probable cause of the crash. If it is necessary to have such a high standard of proof when insufficient evidence is available to conclusively determine the cause in this case, then the outcome of the inquiry will be that the cause cannot be attributed to one single factor. We could contrast the persuasion dialogue with the inquiry dialogue by saying that there are different standards of proof for each of these types of dialogue ([14]). To win a persuasion dialogue, generally you only need to have an argument that is slightly better than that of your opponent. This type of dialogue is highly adversarial in nature even though it is governed by rules of politeness and orderly procedure. But to prove the conclusion based on an inquiry, you need to collect all the relevant evidence that you can. However, very often an inquiry dialogue is preceded by a discovery dialogue, where there is a search for hypotheses to explain some event.

250

D. Walton

The formation of these hypotheses is to provide a basis for continued testing and collecting of further evidence. Negotiation is a type of dialogue that we are all familiar with, and we do not need to say much about it here, because there appears to be no reason to think that this kind of argumentation would have been considered dialectical by the ancient Greek philosophers such as Aristotle. It needs to be said, however, that negotiation dialogue is interest-based, because each participant is trying to maximize its own interests. It might not be appreciated at first that deliberation dialogue is different from persuasion dialogue, so some care will be taken below to look at its special characteristics in formal models of deliberation dialogue of the kind currently being worked on in AI. Below some consideration of inquiry dialogue will also be given. Eristic dialogue is one of the types of dialogue in the typology above, but Aristotle excluded eristic arguments from dialectic. Both Plato and Aristotle were careful to distinguish between dialectical and eristic arguments, associating the latter with fallacies and deceptions. According to Plato (Sophist 231e), eristic arguments represent the arts of the Sophist, the Sophist is portrayed negatively by Plato as an arguer who has no regard for the truth of a matter being discussed, and who is motivated by financial gain, suggesting a bias of self-interest that might be typical of negotiation methods. According to Aristotle (On Sophistical Refutations 171b23–171b29), sophistic is an eristic type of dialogue where the goal is to defeat an opponent by any means. Plato even more openly used ‘eristic’ as a term of abuse, and famously used it to discredit the Sophists (of course, not classifying Socrates as a Sophist). This condemnation of eristic dialogue is not so severe in some contemporary argumentation approaches. Walton (in [41], Chapter 7) allows that quarreling can have some value, even though fallacies can arise from subtle shifts from other types of dialogue, such as persuasion dialogue, to eristic quarreling. The most obvious kind of example would be the use of negative campaign tactics in political debates through the use of personal attack (ad hominem arguments). Sometimes personal attack arguments on an opponent’s ethos can be reasonable, even though obviously in many cases they are used inappropriately as sophistical tactics.

13.3 Aristotle on Classifying Types of Arguments and Reasoning There are many passages in Aristotle’s writings where he makes remarks about dialectic that seem to relate to the contemporary typology of different types of dialogue outlined in Table 13.1. However, it is not at all easy to see how these remarks should be interpreted and to see in any precise way how they relate to the contemporary typology. Aristotle classified three types of reasoning in the Topics (100a27– 101a4) which he called demonstration, dialectical reasoning and eristic (contentious) reasoning. The latter category fits the category of quarrelsome argumentation in the

13 Aristotelian Dialectic, Argumentation Theory …

251

contemporary typology. Demonstration might also fit into the information-seeking or the inquiry type of dialogue. There is a difference of opinion in the literature about which type of dialogue demonstration is used in or fits into. Walton (in [41]) thinks it fits into inquiry dialogue. According to Walton (in [41], p. 76), the inquiry typically takes a form similar to what Aristotle called demonstration (apodeixos) in the Posterior Analytics. An Aristotelian demonstration is a chain of reasoning in which the initial premises are accepted as primary, meaning that they are assumed to hold for the purpose of an inquiry, but do not require demonstration in order to be classified as known. As noted by Walton (in [41], p. 77), Aristotle stated (Posterior Analytics 71b29) that in a demonstration these initial premises function as the grounds or reasons supporting the conclusion by a chain of reasoning that leads to it. As he stated in the Posterior Analytics (71b18–71b22), a demonstration is a syllogism that produces scientific knowledge as the conclusion of a sequence of valid reasoning from premises that are true, primary, immediate, and better known than the conclusion to be proved from them. Dialectical reasoning not based on syllogistic reasoning, in his view, would not be a demonstration, because the result will not be knowledge (71b25). To put it simply, the goal of a demonstration is to prove a proposition, based on a chain of syllogistic reasoning originating from premises known to be true. According to Krabbe and van Laar (in [20], p. 37) however, demonstration fits the information-seeking category, because demonstration is always associated with a didactic context (teaching and learning), rather than one of research. This view of the matter suggests that the proposition to be proved in a demonstration has already been at least provisionally accepted by the arguer constructing the demonstration to be something that is known to be true. The purpose of the demonstration is for the arguer building it to prove the claim that this proposition is known to be true by deducing it from a set of premises all known to be true that can be used by the arguer to show to the students (learners) that the conclusion also belongs to knowledge. The difference between these two methods of classification could be a good subject for further research, but it needs to be outside the scope of this chapter. One reason is that attempting to define demonstration, and see whether it can be represented as a type of dialogue, is a difficult subject that needs to be treated in its own right at some length. The other reason is that the category of didactic dialogue, although it is obviously important for the study of argumentation in the field of education, is also a complex subject with a growing literature in which many different views have been expressed. So for the purpose of this chapter we have put any attempt to classify and analyze didactic dialogue aside. Dialectical reasoning appears to be a more general category than persuasion dialogue, as will be borne out by the remarks on Aristotelian dialectic below. As will be shown below, dialectical arguments are ones that use reason to establish a contradiction starting from generally accepted opinions (endoxa) and answers of the respondent (Topics 165b4–5). Using some of the same classifications, but approaching them from a different angle, Aristotle classified four types of arguments in On Sophistical Refutations

252

D. Walton

(165a40–165b12): didactic arguments, dialectical arguments, examination arguments, and contentious (eristic) arguments. Didactic arguments are ones used for the purpose of teaching. This category of arguments has special characteristics of its own, and although it does not appear to have been discussed widely, it is a fundamentally important category for the field of education. Aristotle tells us that examination arguments (perastikoi logoi) are based on opinions held by the answerer and known to one who claims knowledge of the subject involved. This type of dialogue appears to fit the type of cross-examination of an expert witness of the kind commonly found in common-law trials. Such peirastic arguments are supposedly used by a questioner to test and possibly even criticize the answers of the respondent by subjecting them to careful scrutiny. In the same place in On Sophistical Refutations (165b8–10), contentious arguments (eristikoi) are defined as ones that reason, or seem to reason, from opinions that only appear to be generally accepted, but are not really generally accepted. Aristotle also counts in arguments that actually start from opinions that are generally accepted but only seem to reason correctly (Topics I.1, 100b23–25). This suggests that Aristotle is not simply defining eristic arguments as quarrelsome ones, but also ones that bear relation to endoxical arguments, arguments based on generally accepted premises. It is interesting that here, eristic arguments are identified with sophistry as well as quarrelling. In these passages it appears that Aristotle’s distinctions between different kinds of reasoning in different kinds of arguments overlap considerably with contemporary classifications of types of dialogue in current argumentation. But in the context of present-day argumentation studies, it is quite hard to see what Aristotle is driving at, and how it represents a systematic approach that could be useful. Sometimes he seems to be classifying different kinds of reasoning, while in another passages he is classifying different kinds of arguments. He doesn’t appear to be classifying different dialogue frameworks in which arguments are used, or in which reasoning is used in arguments. And trying to relate his distinctions to the contemporary typology is complicated by the obstacles making it very hard for those of us working in the current area of argumentation studies to get a precise grasp of what Aristotle is writing about when he uses the terms ‘dialectic’ and ‘dialectical argument’.

13.4 The Core Structure of Aristotelian Dialectic Aristotle in the Topics (101a26–30) gives three purposes to which the argumentation methodology presented in the Topics is supposed to be applicable, training in disputation, casual conversations and philosophical sciences. What are the kinds of disputations that Aristotle has in mind? They are formulated by Aristotle and other authors of the time as dialectical debates that have a well-defined formal structure. What is meant by the word ‘formal’ is that they have what would now be called a dialogue structure (Hamblin, in [16] and [17]; Krabbe in [19, p. 72]). There are two parties involved called the questioner and the answerer, and each of them has to take

13 Aristotelian Dialectic, Argumentation Theory …

253

turns making moves in a game like sequence according to rules, or what are now called protocols, defining the kinds of speech acts that are allowed at each move in the procedure. The idea that such dialogues have a formal structure constraining argumentation is an exciting one because it is the basis of the current resurgence of work on argumentation and formal dialogue systems. Once we get an example that can be identified as fitting a dialogue type, we can use the profiles technique and the other tools described in Sect. 13.2 to analyze and evaluate it. But where could Aristotle have gotten this brilliant idea? Apparently he was not the inventor of such dialectical debates or even the first author to write about them (Slomkowsi, in [33], p. 12), for his way of writing about them suggests that the terms he uses to describe them were widely used. In this kind of stylized debate, the answerer defends a designated proposition called a thesis and the questioner attacks this thesis by putting questions to the answerer. During the procedure, an audience can interact with the two primary participants to try to make them follow the dialogue protocols. These dialectical debates seem to be an important part of Greek culture. One source Aristotle may have gotten them from, and where we can see a procedure that strongly resembles Aristotle’s model is the use of the elenchus technique by Socrates in the Platonic dialogues. Castelnerac and Marion (in [9]) have applied dialogical logic of the kind developed by Lorenzen and Lorenz (in [22]) to examples of the Socratic elenchus found in Plato’s dialogues and to Plato’s comments on the dialectical structure of the procedure of elenchus. Castelnerac and Marion have shown that a set of fourteen dialectical rules can be used to logically model the question-andanswer procedure of the elenchus. Before getting to the discussion of such formal dialectical rules, we need to grasp the procedure of dialectical question-and-answer as outlined in Aristotle’s model explaining how it works. According to the widely used formula, the questioner begins the procedure by posing an issue or so-called problem (problema) in the form of a yes-no question. At his first move, the answerer has to select one option or the other as the proposition he will defend. This proposition is often called the answerer’s thesis by Aristotle ([33], p. 16). The goal of the questioner is, by means of asking questions, to get the answerer to become committed to the opposite of the proposition he originally selected as the proposition he will defend. This opposite proposition is called the questioner’s thesis. The method of the questioner is to derive conclusions from the previous answers of the answerer by deductive or inductive inferences. The answerer’s goal is to avoid being refuted. The answerer is refuted, and hence the dialogue ends at that point, if the opposite of his thesis is so derived by the questioner. This central core of the Aristotelian game of dialectical argumentation has been shown by Krabbe (in [19]) to have a structure that can be formally modeled. He formalizes this central core in a formal dialectical system called ACADEMIC1 . This formal model does not attempt to capture the Aristotelian use of inductive arguments or argument by analogy and is only directed to modeling deductive arguments. But it does allow the questioner to argue by reductio ad absurdum, the method of refuting the answerer by generating an inconsistency from the answerer’s initial commitments and the additional commitments he incurs along the way. Krabbe (in [19], p. 76) models the core structure of the Aristotelian dialectical game using Hamblin-style

254

D. Walton

rules (Hamblin, in [16] and [17]) of the kind used in [48] to model dialogues in which argumentation between two parties takes place. The formal language is composed of a set {C, H, P, S, T, …}, a fixed decidable set VA of valid elementary arguments (P1 , …, Pn /C), and a fixed decidable set IP of sentences that express impossible propositions. There are four kinds of rules (Walton and Krabbe, in [48, p. 76]). First, there are locution rules, that define the speech acts that can be used by each of the two parties at each move. Q stands for the Questioner and A for the Answerer. There are four locutions for Q: (1) Problem T, ¬T (T not being itself a negation); (2) Question P, ¬P; (3) Argument P1 , …, Pn /C; (4) Reductio P1 , …, Pn, H/C. There are only two locutions for A: (1) Thesis T; (2) Concession P. Second, there are nine structural rules that define the order of turn taking and what speech acts are acceptable as responses: (1) Each speaker contributes one locution in each turn; (2) Q moves first, putting forward “Problem T, ¬T ”. (3) A move of type “Problem T, ¬T ” can be executed only once; (4) “Problem T, ¬T ” must be followed by “Thesis T ” or by “Thesis ¬T ”; (5) “Question P, ¬P” must be followed by “Concession P” or by “Concession ¬P”. (6) The move “Argument P1 , …, Pn /C” can be used only if P1 , …, Pn /C ∈ VA and P1 , …, Pn are commitments of A; (7) “Argument P1 , …, Pn /C” must be followed by “Concession C”; (8) The move “Reductio P1 , …, Pn , H/C” can be used only if (i) P1 , …, Pn , H/∈ VA, (ii) P1 , …, Pn are commitments of A, and (iii) C ∈ IP; (9) “Reductio P1 , …, Pn , H/C” must be followed by “Concession ¬H”. Third, there are two commitment rules that define which propositions go into each arguer’s commitment set at each move made: (1) “Thesis T ” places T in A’s commitment store; (2) After a move “Concession P”, P is in A’s commitment store. Fourth, there are three rules for winning and losing, defining what sequences of moves by each party count as winning or losing. (1) As soon as ¬T appears in A’s commitment store, the game is over, and Q has won while A has lost; (2) As soon as T appears in A’s commitment store, the game is over, and Q has won and A has lost; (3) If the number of moves equals a previously fixed limit and neither W1 nor W2 applies, the game is over, and A has won and Q has lost. An informal outline of how ACADEMIC1 works follows. There are two participants, the questioner and the answerer. In the first move of the dialogue, the questioner poses a problem by asking the answerer to accept one of two contradictory statements. The statement accepted is now identified and set into place henceforth as the answerer’s thesis. The questioner’s thesis is now set in place as the remaining proposition in the problem that is the opposite of the answerer’s thesis. The questioner asks the answerer to accept propositions and the answerer has to reply whether she accepts the given proposition or not. In order to win the game, the questioner must deduce his own thesis using propositions accepted by the answerer as premises in the arguments used for this deduction sequence. For use later, it is worth emphasizing that ACADEMIC1 permits the type of argumentation called reductio ad absurdum in logic. According to this form of argumentation, if you can derive a contradiction by an acceptable sequence of inferences from an opponent’s position (the set of propositions he has previously accepted),

13 Aristotelian Dialectic, Argumentation Theory …

255

then you have shown that his position is untenable. The questioner’s purpose in the core dialectical structure is to get the answer to commit itself to a proposition that is the opposite of its initial thesis to be defended, and when this happens, the answerer’s position is defeated. There are some obvious resemblances between this ancient dialectical structure and contemporary argumentation theory, because the sequence of argumentation does take the form of a dialogue in which there are three stages, an opening stage, an argumentation stage and a closing stage. Slomkowsi (in [37], p. 15) noted that the formulation of the problem, the so-called problema, comes at the beginning of the discussion. The sequence of moves in the argumentation stage is clearly defined by the turn taking of questioning and answering. What determines the closing stage is whether the answerer succeeds in achieving his goal of deriving the opposite of the questioner’s thesis designated at the opening stage or not. If the questioner fails to achieve this goal after a predetermined number of moves, the answerer wins. We can see from this criterion of winning and losing that the dialogue is of an adversarial kind. It can be rightly classified as a kind of competition or game with rules defining what moves can be made and that determine which sequence of moves count as winning or losing. One might compare this set of rules modeling the procedure of the Socratic elenchus given by Castelnerac and Marion in [9]. These rules are quite closely comparable, point by point, to the rules in Krabbe’s formal model ACADEMIC1 and other rules proposed below are extensions of this model that help us to grasp the nature of Aristotelian dialectic more fully. The correspondences between these sets of rules suggest the hypothesis that the Socratic elenchus, as it can be found in so many of the Platonic dialogues, is an instance of Aristotle’s dialectic. This hypothesis is controversial however, and will be discussed further at the end of this chapter. Krabbe (in [19]) has shown how the central part of the critical discussion type of dialogue, corresponding to the argumentation stage and a small part of the opening stage, can be formally modeled in such a way that it fits with the central core of Aristotle formal dialectic modeled by ACADEMIC1 . He calls this formal model of this central part of the critical discussion CD1 (for critical discussion). CD1 is a Hamblin style formal model that has locution rules, structural rules and rules of winning or losing structured after the format of Hamblin’s main formal dialogue [17]. Krabbe ([19], pp. 81–83]) has also shown that another Hamblin-type system called CB, which is parallel to and highly comparable to ACADEMIC1 , has special features of strategic argumentation. CB ([45], pp. 133–135) contains locution rules (that determine permissible locutions, such as questions and challenges and retractions), commitment rules (indicating when a proposition can be added to or deleted from a participant’s commitment store) dialogue rules (that state pre and post conditions for a given type of move), and strategies (that present kinds of sequences of moves that can help a participant towards his goal). One strategy rule of CB that is also important in Aristotelian dialectic is called a dividing strategy. According to this strategy, the questioner should present premises separately for the answerer’s acceptance so that the answerer cannot easily foresee where the line of questioning is going. Another is called a spreading strategy that works by inserting many intermediate steps in the

256

D. Walton

derivation. Again the rationale is that the answerer should not easily foresee that the line of argumentation is moving toward proving the opposite of his ultimate conclusion.

13.5 Dialectical Features Surrounding the Core Structure It is also interesting to find that outside of this formal core structure there were ten other characteristics of Aristotelian dialectic of the kind described in the Topics. The ten features describe rules for the practice of dialectic that were applied in practice to reflect how Greek academic debates actually took place. The first feature is that, according to the account of this procedure given by Aristotle, there were rules of “fair play”, presumably enforced by the audience (Topics 8 160b3, Soph. El. 8 169b31, Slomkowsi, in [33], p. 15). The second feature is that if the answerer does not accept the conclusion that was deduced by syllogism, he is supposed to give some sort of explanation called a “solution” ([33], p. 15). It is interesting to note here that explanations as well as arguments are involved. The third feature is explained by Krabbe (in [19], p. 75) as follows: when the answerer is asked to choose which one of two opposite propositions he will accept, he may alternatively choose to ask for clarification or object to the question by pointing out an ambiguity (Topics 8 160a17–34). Although there could be two kinds of responses here, they are connected, because pointing out a linguistic ambiguity in the formulation of the question would typically be a request for clarification of the usage of language of the question. It is uncertain therefore whether we have one feature here or two. In general, if the answerer does not understand the question, he is always permitted to reply “I do not understand” (Topics 8 160a18–21). The fourth feature is that the answerer is allowed to contest the validity of an argument, but this option is not formalized in ACADEMIC1 ([19], p. 75). Krabbe specifically states (p. 75) that these two possible responses are not formalized in the formal structure of ACADEMIC1 . The fifth feature is the allowing of inductive reasoning in addition to deductive reasoning (Topics 8 157-18-20). Inductive generalizations of the kinds that are subject to exceptions could also be used. This feature brings out an important characteristic of Aristotelian dialectic: that universal generalizations were not the only inference rules allowed as means to infer conclusions. The sixth feature is the rule that if the answerer does not accept a conclusion derived by inductive reasoning, he is supposed to give a counterexample. Neither of these kinds of moves is allowed in ACADEMIC1 . Both rules are very interesting from a point of view of recent developments in argumentation theory, because they make it evident that more than just deductive logic (syllogistic reasoning) was involved in the way these games were actually played. One can see from these additional rules that there appears to be room for defeasibility by allowing for counter-examples to generalizations. The seventh feature described by Aristotle is that a problem of the kind to be argued about dialectically cannot just be any pair of propositions where one is the

13 Aristotelian Dialectic, Argumentation Theory …

257

opposite of the other. In the Topics, Aristotle tells us that a problem has to have aporetic content, meaning that the disagreement needs to be one that is subject to puzzlement. Its subject is something about which either men have no opinion either way, or most people hold an opinion contrary to that of the wise, or the wise contrary to that of most people, or about which members of each of these classes disagree among themselves ([2], 104b1–5).

The seventh feature, the aporetic feature of puzzlement, brings us to the eighth feature. Puzzlement is based on the notion of an endoxical proposition, referring to a proposition generally accepted by the wise and/or by most people generally. The eighth feature is the use of endoxical propositions. In On Sophistical Refutations (165b4), Aristotle defined dialectical arguments as “those which, starting from generally accepted opinions (endoxa), reason to establish a contradiction”. However, as made evident from the quotation from the Topics above, endoxa are not just popular opinions of any sort, for in the Metaphysics, Aristotle wrote that we do not have to “consider the views of the multitude” for they “talk without consideration about almost everything” ([1], 1214b29–1215a15). An endoxical proposition is also one that has some standing, representing a view that should be taken seriously. Hence Barnes (in [5], p. 500) translated the word endoxon as a “reputable opinion”, suggesting that an endoxon should be seen in the Aristotelian sense as referring to an opinion that is more than just a commonly held belief. In book one of the Topics, Aristotle defines endoxical propositions as those which “appear to be correct to everyone or the majority or the wise—that is to say, all of the wise or to the majority or to the most famous and distinguished of them” ([2], 100b21–23). He goes on to discuss these notions in the Topics (104b18–28), drawing a number of subtle distinctions about endoxical propositions. For example, he suggests that a thesis that is contrary to general opinion but propounded by a famous philosopher would be a suitable dialectical problem. What has been brought into the mix here is the feature called common knowledge in current work on argumentation and AI (Walton and Macagno, in [44]). The idea is that dialectical disputation is based on the idea that some propositions can be taken for granted as acceptable to an audience because they are so widely accepted as true by the general population or by the experts that they can be set aside as beyond dispute, for the purposes of delimiting a dialectical discussion. This notion may not coincide exactly with the Aristotelian idea of an endoxical proposition, but it is closely enough related to that it should be interesting, both to argumentation theorists and Aristotelian scholars. It raises some other issues of interest for studying Aristotelian dialectic. Bolton ([8], p. 69) convincingly argues for the thesis that some endoxical propositions can be more endoxical than others for Aristotle. This implies that endoxical propositions can form an ordering. Bolton ([8], pp. 69–71) also has the view that for Aristotle, the different kinds of endoxical premises are used differently in different uses of dialectic, such as using dialectic for mental training or for gymnastic dialectical discussion. This view of the matter is interesting because it suggests that different kinds of dialectic can be distinguished, based on how dialectic is being used in a

258

D. Walton

given case, which in turn depends on how endoxical propositions are used in the type of dialectic. These distinctions appear to turn on the feasibility of constructing an ordering of endoxical propositions with the category as endoxical as possible at the top. After surveying the different ways that Aristotelian dialectic has been interpreted in the literature, Wlodarczyk ([51], p. 156) states that her interpretation is different from all of them in that it emphasizes what she calls the overarching principle. This principle requires that in order for an argument put forward in a dialectical dialogue to be considered to be persuasive, it has to proceed from more knowable premises to a less knowable conclusion. She adds that this principle has to be followed by both the questioner and the answerer. For example, she cites Aristotle’s statement (Topics 8 160a11–14) that the questioner will achieve his argument provided that all its premises are more endoxical than the conclusion he aims to prove from them. A little below that (8 160a14–16), Aristotle states that those who attempt to reason from premises that are more adoxical than the conclusion do not reason well. This principle could be seen as a defining characteristic of Aristotelian dialectic because it states a necessary requirement for the usefulness or success of any argument put forward in a dialectical framework of dialogue. In general it could be called the principle of premise priority: the premises of a persuasive argument in a dialectical exchange need to be epistemically prior to the conclusion in order for the argument to be properly taken as successful in the respect that the premises support the acceptability of the conclusion. If this principle applies to the category of persuasion dialogue generally, it might not be an especially distinctive characteristic of Aristotelian dialectic. On the other hand, it does seem to be especially important in Aristotelian dialectic because it offers either party in the exchange a way of responding to and successfully challenging an argument put forward by the other party. Hence it does represent a kind of argumentative move that should be allowed as a speech act or type of locution in any formal model of Aristotelian dialectic. There are two remaining features of Aristotelian dialectic that need to be taken into account. The ninth characteristic of Aristotelian dialectic is its use of the topoi or topics listed by Aristotle in the Topics. But we will have more to say below about the role of the topics. The tenth characteristic of Aristotelian dialectic is that one can identify certain strategies that give the participants advice on how to construct a persuasive argument. For example, the demonstration of the thesis to be proved to the other party should not be too obvious to the other party nor should it be too remote (Topics 8 156a12f). If it is thought to be too near to destroying the answerer’s thesis, he has the right to object to conceding it, even if it seems true (Topics 8 160a3–6; [51], p. 171). On the other hand, if it were to be too obvious that problema would not really be aporetic, but if it were to be too remote, this could involve difficulties for dialectical training (Slomkowski, in [33], p. 17). Aristotle writes in book 8 of the Topics (6 160a1–6) that if the proposition put to him to be answered as true or false is too near his original thesis, he should say that if it is conceded his thesis will be destroyed. As Slomkowski ([33], p. 28) puts this point, the questioner needs to conceal his conclusion because if the answerer realizes that a concession he is asked to make leads to the opposite

13 Aristotelian Dialectic, Argumentation Theory …

259

of this thesis, he will refuse to grant it. The strategy that appears to be suggested by these remarks is that the chain of inferences should be sufficiently long or complex so that the answerer does not easily see that he will be refuted. See the discussion of this point by Krabbe ([19], pp. 82–83). To sum up this section, we have identified ten defining features of Aristotelian dialectic that lie outside the core structure of Krabbe’s formal model of the core of dialectic ACADEMIC1 : (1) rules of fair play, (2) offering a “solution” that explains the problem, (3a) asking for clarification of the question, and/or (3b) pointing out an ambiguity in it, (4) contesting the validity of an argument, (5) allowing inductive reasoning, (6) offering a counter-example as a response to an inductive argument (7) the aporetic aspect of the initial problem to be discussed, (8) the acceptance and use of endoxical propositions, for defining the problem, for use in judging the acceptability of premises in dialectical arguments and for distinguishing different uses of dialectic enabling us to postulate different categories of dialectical argumentation (9) use of topics, and (10) use of argumentation strategies. These ten characteristics bring out the important point that Aristotelian dialectic cannot be completely defined by the formal characteristics of ACADEMIC1 as a type of dialogue. The ten additional informal characteristics of it need to be taken into account. Taking these ten additional characteristics into account enables us to distinguish different uses of Aristotelian dialectic. This, in turn, implies that Aristotelian dialectic is not simply a formally structured and abstract kind of dialectical question-andanswer game, a view that historically often appears to have been adopted. Indeed, seeing it in this way as a simple formally regulated game is probably the reason why its importance in relation to real arguments was not appreciated for such a long time. But now we can see how the core formal structure of it can be so modeled, as precisely shown by Krabbe (in [19]), we can also identify the features outside this core structure. These developments when taken together with the advent of argumentation theory provide a much fuller, richer and more sophisticated view of Aristotle’s dialectic.

13.6 Deliberation Dialogue Aristotle does not see deliberation as a dialectical model of argumentation that would fit his account of dialectic, but he does give a very clear account of the characteristics of deliberation in the Nichomachean Ethics 1 112a20–1 113b25 that fits very well with contemporary models of deliberation in AI. One of the important characteristics of deliberation and Aristotle’s account is that we deliberate about things that are inherently subject to change. Deliberation is employed in matters which, though subject to rules that generally hold good, are uncertain in their issue: or where the issue is indeterminate, and where, when the matter is important, we take others into our deliberations, distrusting our own capacity to decide. [3], 1112b10)

260

D. Walton

This account of deliberation fits in very well with the recent modeling of it in AI, where the defeasible nature of argumentation is stressed, and where deliberation is seen as a type of argumentation that needs to be sensitive to the agent’s knowledge of the changing circumstances of the case. This quotation is especially significant because it suggests another characteristic of deliberation that has been stressed as important in recent work in AI: that we can take others into account in our deliberations, suggesting that deliberation could be a group activity, even though Aristotle appears to see it primarily as an individualistic kind of decision-making. As noted above, persuasion dialogue is highly adversarial because the individual goal of each participant is to win out over the other side by finding stronger arguments that defeat its contention or cast it into doubt. On the contemporary view of deliberation as a type of dialogue as studied in AI, in contrast, the participants try to devise a collective plan of action that can help them to move forward to agree on a policy that can take their individual interests and positions into account, but is not dictated by these factors. The first in a series of formal models of deliberation dialogue put forward in AI was the computational argumentation system of McBurney, Hitchcock and Parsons (in [25]), the MHP model. Deliberation on this model was formalized as a resourcebounded procedure that starts from an initial situation where a choice formulating a so-called governing question needs to be made. The sequence of deliberation proceeds from the opening stage to go through seven other stages as it moves towards a closing stage where the decision is arrived at on the basis of the pro and con arguments that were considered in the preceding stages. The procedure ends in a closing stage. The governing question stays in place then applies over the whole following seven stages of the dialogue procedure, including the closing stage. The issue to be deliberated on stays the same, in a way comparable to the stasis proposition, the ultimate conclusion to be proved, according to the ancient theories of argumentation. Initially, the type of examples of such issues to be decided were posed as yes-no questions such as “Should we invade Iraq?” or “Should we go for pizza?”, but later examples also considered problem-solving questions such as “Where should we go for dinner?” or “How do we stop the Iraqi’s war?” (Kok et al., in [18]). More recent models use an open knowledge base to allow sequences of practical reasoning that can more fully take changes of the circumstances into account (Walton, Toniolo and Norman, in [50]). There are certain argumentation schemes, comparable to the forms of topical reasoning studied by Aristotle in the Topics, that are highly characteristic of deliberation dialogue. These include the scheme for practical reasoning, often called teleological or goal-directed reasoning, and the schemes for argument from positive and negative consequences. Arguments from positive consequences are species of argument from positive value, and arguments from negative consequences are species of argument from negative value. Argument from values is also closely related to the scheme for practical reasoning. The simplest form of practical reasoning, called the heuristic form of the scheme, is called practical inference ([49], p. 323). In this version an agent, represented by the first-person singular pronoun ‘I’, is deliberating on what to do.

13 Aristotelian Dialectic, Argumentation Theory …

Major Premise

I have a goal G

Minor Premise

Carrying out this action A is a means to realize G

Conclusion

Therefore, I ought (practically speaking) to carry out this action A

261

There are three identification conditions that can be used to recognize an instance of this kind of argumentation in a given text of discourse in a natural language. (1) There is a decision about an action, or a proposal for action, or a practical problem that needs to be solved, (2) pro and con arguments need to be weighed against each other as all the arguments are summed up, and (3) an agent or group of agents arrive at a conclusion on how to proceed based on a goal or set of goals. The simplest kind of practical reasoning, such as the kind used to solve the problem of fixing a broken machine, can be purely instrumental. But in a more complex variant, called value-based practical reasoning, the agent’s goals are based on its values (Atkinson and Bench-Capon, in [4]). Premise 1

I have a goal G

Premise 2

G is supported by my set of values, V

Premise 3

Bringing about A is necessary (or sufficient) for me to bring about G

Conclusion

Therefore, I should (practically ought to) bring about A

In this form of reasoning an agent is advocating going ahead to carry out a particular action, or policy for action, based on a goal, and the goal is held to be based on values held by the agent shared by the collective group of agents trying to arrive at decision on what is the best thing for them to do when they are confronted with a problem (Black and Atkinson (in [6]). One subject currently being discussed in AI (Walton, Toniolo and Norman, in [23]) is whether the issue of the deliberation is fixed at the opening stage, and remains applicable though to the closing stage, or whether the issue can be changed once new information about the circumstances has come to be known by the agents. The issue (stasis) remains fixed in persuasion dialogue, but deliberation may be different in this respect.

13.7 Examination Dialogue A central problem for further studies posed by the investigations in this chapter is that of classifying Aristotle’s dialectical system, defined in this chapter as made up of the core structure defined by ACADEMIC1 along with the ten supplementary features. These characteristics taken together all appear to suggest that Aristotelian dialectic is supposed to be a subtype of persuasion dialogue. The reason supporting this interpretation is that the questioner and the answerer each has an ultimate proposition to be proved, and the one proposition of the pair is the opposite of the other.

262

D. Walton

Moreover, it looks like the questioner is trying to prove her thesis by getting the answerer to commit himself to an inconsistency, and that achieving this result defeats the answerer’s attempt to prove his thesis. The outcome of this procedure, it would thus appear, is that the questioner has proved that only her thesis has survived the procedure, and therefore she has proved her thesis to the answerer and the audience. Looking at Aristotelian dialectic this way, it seems to be a species of persuasion dialogue, even though it is a special type of persuasion dialogue in which the options of the answerer are limited. As the questioning and answering procedure unfolds, neither side can put forward arguments in the standard way of a persuasion dialogue. The questioner can only ask yes-no questions, and the answerer has the highly limited option of only being able to answer yes or no. Hamblin ([16], p. 59) cites three passages in On Sophistical Refutations (169b25, 171b4, 172a23) indicating that Aristotle is not sure whether dialectical and examination arguments are the same or different. From the point of view of argumentation theory, it makes sense to see examination dialogue as being a distinctive type of dialogue in its own right, separate from persuasion dialogue, even though the two types of dialogue have common elements and appear to be very closely related [41]. Very generally speaking, according to the way we commonly use the term ‘examination’, an examination would appear to be careful and critical scrutiny of the evidence or known facts in a case by testing them, for example by saying whether they are consistent or not. This kind of examination is carried out by the asking of questions, or in scientific pursuits, the putting forward of hypotheses that can be confirmed or refuted by the evidence. Examples would be a physician examining a patient, or a lawyer examining a witness in a trial setting. In this view of the matter, a normative model should conceive an examination dialogue as having two goals, the extraction of information and the testing of it for reliability. Having the first goal suggests that examination should be classified as a species of information-seeking dialogue in the typology given above. This first goal is achieved by the examiner putting questions to the respondent in order to obtain information about something. The second goal is achieved by using critical argumentation used to judge whether the information presented is reliable, or even by trying to show that it is unreliable. To achieve this goal, the information can be tested against the respondent’s other answers or by propositions accepted as factual evidence in order to find an inconsistency (Dunne et al., in [10]). But for a different view that classifies examination dialogue as a mixed type of dialogue with two subtypes subsumable under inquiry or persuasion dialogue, see van Laar and Krabbe (in [35]). Their conclusion (Van Laar and Krabbe, in [35], p. 34) is that it seems not really necessary to model examination dialogue as a separate main type of dialogue, but that even so, the case could be made for classifying it as an inquiry dialogue or a persuasion dialogue. They conclude (p. 35) that examination should be classified as a type having instances of both persuasion dialogue and inquiry, and often being a mixed type of dialogue that contains embedding of other dialogue types. They also stress the importance of modeling examination dialogues in many instances as species of inquiry dialogue. They also discuss several distinctive subtypes of examination dialogue, including

13 Aristotelian Dialectic, Argumentation Theory …

263

the expert examination dialogue, a type of dialogue that is very important in law, most notably in cross-examination of expert witnesses in a trial setting. However examination dialogue is to be modeled, it is closely related to Aristotelian dialectic. What peirastic arguments are, in particular, seems to be a disputed question. Hamblin (in [17]) pointed to ambiguities suggesting that Aristotle may have sometimes held that the view that dialectical and examination arguments are in the same category. According to this view, examination arguments could be subspecies of dialectical arguments, and so there would only be three basic categories. Hamblin ([17], p. 59) cited the passage in Topics 159a25, as evidence for this view. In this passage, Aristotle contrasted examination arguments with contentious arguments, but there does not seem to be a clear distinction made between dialectical and examination arguments. Guthrie ([15], p. 155) also saw peirastic as being part of dialectic, or tied in with it, but appeared to see it as somewhat distinct as well. The issue for argumentation theory is how peirastic dialogue should be analyzed and classified. Is it a species of information-seeking dialogue? It does seem to be, at least partly. For when you are examining somebody, presumably you are trying to get information out of them. But there is also a testing element involved. Bolton ([8], p. 80) describes the term peirastic (peirastike) as used by Aristotle to refer to “the art of testing”, an art used to test claims to knowledge. Examining seems to be kind of probing in which the information, or what is claimed to be knowledge is tested by a question-answer procedure. The Greek words used for examination by Aristotle strongly suggest that this type of dialogue is a kind of testing of a claim that is a contestive procedure where the questioner attacks the answerer’s claim or supports the opposite claim (Fink, in [11], p. 5). This argumentative activity of the questioner is called ‘testing or refuting’ (™λšγχειν), ‘examining’ (™ξετ£ζειν), ‘attacking’ (™πιχειρε‹ν), ‘destroying’ (¢νασκευ£ζειν) the thesis of the respondent, or ‘constructing’ (κατασκευ£ζειν) a claim in case the respondent defends a negative thesis.

But this kind of testing is not just attacking or destroying a claim. The attacking has a purpose and a value of leading the two parties closer to the truth by weeding out falsehoods. The suggestion is that if a claim is tested, for example a scientific hypothesis, the testing procedure provides evidence supporting the rational acceptability of the hypothesis, or undermining the assumption that the hypothesis is rationally acceptable. Fink ([11], p. 5) suggests that this notion of examination as testing corresponds to the Socratic notion of taking up an argument for examination, corresponding to the Greek expression λÒγoν λαβε‹ν used in Plato’s Meno (75d1–2). This proposal fits with the approach of Castelnerac and Marion (in [9]) that sees Aristotelian dialectic as representing the kind of Socratic examination dialogues found in Plato’s dialogues. Perhaps then examination dialogue is best seen as a blending of information-seeking dialogue with some other kinds of dialogue, such as persuasion dialogue or inquiry. In many instances, Socrates is trying to extract knowledge by examining an expert using the procedure of finding an inconsistency between the expert’s previous commitments and then seeing how the expert reacts to this finding.

264

D. Walton

Guthrie ([15], p. 155) drew a distinction between two subtypes of examination, peirastic and exetastic arguments. On his way of drawing the contrast, peirastic arguments use testing or probing questions to extract information from a respondent, while exetastic arguments probe even more deeply into the respondent’s answers by examining them critically and even trying to refute them using counterarguments ([15], p. 155). The interpretation of exetastic argument as having a critical aspect or use of counterargument can be found in the Rhetorica Ad Alexandrum (1427b12– 1428a17). Exetasis is described in this manual as using circumstantial ad hominem arguments that attack the questioner for not practicing what he preaches by arguing that his words and deeds are not consistent with each other (1427b13–1427b14). The peirastic type of examination could represent the familiar kind of testing procedure used in education. The teacher presumably already has the information asked for in her question, but she needs to determine whether the student can answer the question. The exetastic type of examination dialogue could be typified by the cross-examination of a witness in court. In cross-examination, the attorney may typically try to show that answers of the witness are inconsistent. How peirastic and exetastic dialogues should be modeled is an unsolved problem, one of special importance for those applying argumentation techniques in the field of education. Examination dialogue generally seems to be basically a type of information-seeking dialogue combined with persuasion dialogue. Examination may also have some role to play in scientific inquiry of the kind described by Aristotle in the Analytics. On the other hand, Aristotelian dialectic as a species of formal dialogue structure defined in this chapter the dialogue seems more like what has been classified in this chapter as an examination dialogue, at least on the interpretations of this type of dialogue in current argumentation studies (Walton, in [41]). In particular, it seems to fit best as a species of exetastic examination dialogue, where the proponent has the goal of attacking the commitments of the respondent, and most characteristically and specifically carrying out this goal by finding an inconsistency in the commitments of the respondent. The complication is that the exetastic type of dialogue, as defined in this chapter, is a species of dialogue that combines persuasion dialogue with examination dialogue. To make progress with solving this problem on how to classify Aristotelian dialectic within the current classification system of types of dialogue, it has to be determined what the purpose of the type of dialogue characteristic of Aristotelian dialectic was supposed to be. Was it supposed to be a training exercise? Was it supposed to represent the type of argument exemplified in the Socratic elenchus illustrated in the Platonic dialogues? The kind of argumentation that takes place in Socratic elenchus has a testing function at its center. By examining a series of claims made as answers to leading questions, and probing into them by critical argumentation, the questioner can guide the answerer to a contradiction that forces the answerer to improve and refine his initial claim. The purpose of the dialogue, so conceived, is to subject the answerer’s position to testing. Philosophical views typically cannot be refuted by empirical evidence, as a scientific hypothesis can be. They can only be refuted by getting the theorist to accept other commitments and then drawing inferences from the expanding set of commitments which lead to an unacceptable

13 Aristotelian Dialectic, Argumentation Theory …

265

result, such as an inconsistency in the set. Socrates’ strategy representing his goal in the dialogue is to show that the answerer’s replies do not make sense because they collectively lead to contradictions (Vlastos, in [39]). But the examination dialogue also has a wider goal which is to provoke the answerer and the audience to learn from it by seeing how refinements need to be made in the arguer’s initial claim. The examination is a procedure of testing a claim to see if it stands up to critical scrutiny. In this sense of the term ‘examination’, the Socratic elenchus can rightly be classified as a distinctive type of examination dialogue.

13.8 Schemes and Dialogues in Artificial Intelligence Earlier commentators on Aristotle such as Ross ([30], p. 59) had accepted the view that the earlier method of dialectical argumentation described in the Topics was later superseded by a better method, namely the way of laying out demonstrations described in the Analytics. In this view the method of scientific inquiry described in the Analytics supersedes and replaces the method of dialectic ([8], p. 58). On this view, scientific inquiry, in Aristotle’s view, is not a form of dialectical argumentation of the kind defined in the Topics. But more recently, the generally accepted opinion among Aristotelian scholars has turned the other way ([8], p. 59). Some Aristotelian scholars now accept the view that the method of dialectic described in the Topics is applicable not only to philosophical discussions but also to scientific inquiry of the kind described in the Analytics. This difference of opinions poses an issue that is here left as a problem for further study. Another problem for further study is the relationship between the current argumentation schemes and Aristotle’s topics. I think I can speak for many in the argumentation community when I say that we have found it very hard to make much use of Aristotle’s list of topics and the descriptions he presented of them in the Topics ( [21]). They don’t appear to match, or even look very much like, the argumentation schemes listed in [49]. Of course there are certain exceptions, such as the argument from analogy, where the form of argument as described in Aristotle is easily recognizable as at least somewhat comparable to the current scheme for argument from analogy. But in general, the topics have tended to appear elusive and mysterious. Nobody seemed to know quite how to interpret them or what to do with them, and opinions differed widely in this regard. It is undoubtedly for this reason that Aristotle’s Topics has been a much neglected work in the past. It has been hard for Aristotle commentators to see how it fits into the rest of his works. One case of special interest is the relationship between endoxical premises of the kind used in arguments in Aristotelian dialectic and the argumentation scheme for argument from general acceptance, also called argument from popular opinion (Walton, Reed and Macagno, in [49], p. 311).

266

D. Walton

General Acceptance Premise A is generally accepted as true Presumption Premise

If A is generally accepted as true, that gives a reason in favor of A

Conclusion

There is a reason in favor of A

There are two critical questions matching this scheme. CQ1 What evidence, like a poll, or an appeal to common knowledge, supports the claim that A is generally accepted as true? CQ2 Even if A is generally accepted as true, are there any good reasons for doubting it is true?

This form of argument was traditionally called the argument from popular opinion in the logic textbooks, where it was typically regarded as an informal fallacy. More recently it is regarded as a defeasible form of argument, a heuristic or short-cut argument that can support a conclusion weakly as a presumption, but that can be made stronger by testing it against critical questions and counterarguments. Some other argumentation schemes of the kind that have now become familiar in computational argumentation, and argumentation studies, are listed in Table 13.2. All these forms of argument are defeasible (Walton, Reed and Macagno, in [49]), meaning they are subject to critical questioning, as illustrated in the example below. Some of these schemes are closely related to Aristotelian topics (Macagno, Reed and Walton, in [24]). Dialectical argumentation also takes into account the context of use of an argument by looking at the argument in a procedural way as a sequence of questions and answers (and other sorts of moves) by two or more parties in a dialogue. This dialectical approach reaches a decision on whether to accept a claim or not based on the arguments both for and against the claim, and therefore on a dialectical point of Table 13.2 Some standard argumentation schemes Argument from witness testimony

Argument from a verbal classification

Argument from rule

Argument from expert opinion Argument from appearances (perception)

Argument from threat

Argument from analogy

Argument from positive consequences

Argument from generally accepted opinion

Argument from precedent

Argument from negative consequences

Direct Ad Hominem argument (personal attack)

Practical goal-based reasoning to action

Circumstantial Ad Hominem argument

Argument from correlation to cause

Argument from Evidence to a Hypothesis

Abductive reasoning (IBE)

Argument from commitment

Argument from Ignorance (negative evidence)

Argument from sunk costs

Slippery slope argument

13 Aristotelian Dialectic, Argumentation Theory …

267

Table 13.3 The videogames example shown in the dialogue format T

Proponent

Respondent

T1

Videogames do not lead to violence.

Why do you think so?

T2

Dr. Smith says so, and she is an expert.

Is Dr. Smith trustworthy?

T3

What evidence do you have for questioning her trustworthiness?

She is biased

T4

What evidence do you have for claiming she Her research was funded by the videogame is biased? industry

T5

What evidence do you have for claiming that?

It was shown by a 2001 investigation of the Parents’ Defense League

view seeing an argument as always having two sides, the pro and con sides in the cases of a persuasion dialogue such as the videogame example ([47], p. 8) set out in in Table 13.3. At turn T1 the proponent makes a claim, and the respondent puts forward a challenge by asking for evidential support for the claim. At T2, the proponent advances an argument from expert opinion and the respondent asks a critical question. At T3 the proponent then asks for evidence to support the critical question, and the respondent offers the claim that Dr. Smith is biased. At T4, the proponent asks for evidence to support the allegation of bias and the respondent offers some backup evidence, by making the further statement that her research was funded by the videogame industry. When this claim is also a challenge by the proponent at T5, the respondent cites an investigation. Notice that this small dialogue definitely contains argumentation, but modern logic is ill-equipped to evaluate argumentation that takes place in a dialogue format. What we need is some way to extract the individual arguments put forward in this dialogue so that some sort of logical procedure could be used to link them together showing how they are connected and to analyze and evaluate the sequence of argumentation. There is a method for doing this called the profile of dialogues tool. The profiles tool works by mapping a dialogue sequence containing argumentation into an argument diagram into a pair of argument diagrams (Walton, in [42]). One dialogue provides a diagram of how the argument actually went while the other dialogue provides an idealization of how the argumentation should have gone if it were to properly conform to the normative model of rational dialogue. Dialogue rules, called protocols in AI, define what types of moves are allowed at each stage (Reed, in [29]). What is important for our purposes here is that the profiles technique is based on the capability to visually represent a given sequence of argumentation using a standard argument diagram, a graph structure called an argument map in computing. An argument map is essentially a graph, in the mathematical sense of the term, meaning a set of points and a set of pairs of points called lines, or sometimes arcs or arrows. The points represent propositions, the premises and conclusions joined together by the lines which represent transitions from a set of premises to a conclusion. Many software systems and tools from artificial intelligence use argument diagrams along with

268

D. Walton

argumentation schemes to identify, analyze and evaluate arguments. In Sect. 13.10 a simple example is used to show how the profiles technique works by mapping a dialogue structure into an argument structure of the kind used in artificial intelligence.

13.9 Mapping Dialogues into Argument Diagrams To help to explain to the reader how formal dialogue systems are related to argumentation schemes, in this section it is briefly shown in a simple manner how the sequence of argumentation illustrated in Table 13.3 can be identified, analyzed and evaluated by mapping it into an argument diagram. For example, the argumentation in the videogames dialogue could be evaluated using argument graphs of a kind associated with the Carneades Argumentation System (Walton and Gordon, in [43]; Gordon, Friedrich and Walton, in [12]) or the ArguMed argumentation system of Verheij (in [37, 38]). Consider the argument from expert opinion in Table 13.3 at T2 where proponent claimed that videogames do not lead to violence based on the two premises that Dr. Smith is an expert on video violence and that she says that videogames do not lead to violence. Underlying this argument is a conditional premise expressing a generalization to the effect that experts are generally reliable. Let’s say however that the respondent questions the trustworthiness of Dr. Smith. There is also an underlying unexpressed premise once the respondent’s attack is framed as a counterargument. This is the conditional premise that if Dr. Smith’s research is funded by the videogame industry she is not trustworthy. Once this implicit premise has been filled in, both the pro argument and the con argument attacking it can be framed as an instance of the argumentation scheme for defeasible modus ponens. Verheij (in [36, p. 232]) showed how defeasible argumentation schemes such as the scheme for argument from expert opinion fit a more general scheme he called modus non excipiens, which can be expressed as taking the following form. As a rule, if P then Q P It is not the case that there is an exception to the rule that if P then Q Therefore Q This scheme was called defeasible modus ponens (DMP) by Walton (in [40]). It tells us that a defeasible argument such as an argument from expert opinion (a form of argument that is not infallible, because experts themselves are not infallible) can be defeated if an exception is found to the general rule taking the form of a conditional premise in the DMP scheme. But using the device of entanglement, as illustrated in Figs. 13.1, 13.2 and 13.3, it can be shown more simply on an argument diagram how a defeasible argument can be undercut using entanglement. A brief sketch of how it is possible in principle to carry out such a task using such tools can be given, starting with the argument diagram shown in Fig. 13.1. In this diagram, the propositions making up the premises and conclusion of the argument are

13 Aristotelian Dialectic, Argumentation Theory …

269

Fig. 13.1 An argument diagram of the first step in the videogames example

Fig. 13.2 An undercutter in the videogames example

Fig. 13.3 An argument diagram of the second step in the videogames example

shown in the rectangular nodes. The argument, and argument from expert opinion, is shown in the circular node labeled +a1. The plus sign indicates a pro argument, an argument supporting a claim. In its simplest form, the argument from expert opinion states that so-and-so is an expert and so-and-so states that a particular proposition is true, therefore this particular proposition is true. Such an argument is taken in argumentation theory to be a linked argument, meaning that the two premises go together to support the conclusion. This linked structure is displayed in Fig. 13.1 using the conventions of Carneades. The two premises are shown as being linked to each other by the two lines that both go into the round argument node representing the argument a1. Arguments can often be identified as linked because of an argumentation scheme (Walton, Reed and Macagno, in [49]) that joins the premises to the conclusion, as in the argument from expert opinion used at turn T2 by the proponent. Here is the scheme for argument from expert opinion as configured by Walton, Reed and Macagno (in [49, p. 310]). Major Premise

Source E is an expert in subject domain S containing proposition A

Minor Premise

E asserts that proposition A is true (false) (continued)

270

D. Walton

(continued) Conclusion

A is true (false)

Let’s say that both these premises are accepted by the audience, and are not subject to doubt at this particular point in the sequence of argumentation displayed in the graph. To denote this acceptance, each of these two propositions is shown in a green background. In a linked argument, both premises need to be accepted in order to prove the conclusion. But also, the argument is defeasibly valid, meaning that if both premises are true the conclusion should be tentatively accepted unless some reason has been given so far to think that it should not be accepted. On the basis of this argument, the conclusion, the proposition that video games do not lead to violence, is automatically drawn by the system as having a green background. Hence this proposition, the ultimate conclusion of the whole argument is shown as “proved” (acceptable to the audience) on the basis of the argument so far. Generally in Carneades, if a proposition is rejected, it is shown in a rectangle with a red background. If a proposition is neither accepted nor rejected, it is shown with a white background. Hence this model can be seen as a three-valued logic system. Once the user inputs the initial values of premises, Carneades calculates the values of the conclusions drawn from them as outputs. However, it has to be kept in mind that the argument from expert opinion is taken in Carneades and other comparable computational argumentation systems, such as ArguMed, to be defeasible, meaning that although it may hold tentatively in a given case, it may be attacked and overturned as new evidence comes in. In general, in computational systems of the kind currently being used in artificial intelligence and multiagent systems, there are three ways of attacking an argument: you can attack one of its premises, you can attack its conclusion, or you can attack the argument connecting the premises to the conclusion. In the example dialogue, note that the respondent at T2 has raised the critical question of whether Dr. Smith could be biased. In such a case, it needs to be noticed that there are two kinds of critical questions. With the first kind of critical question, merely asking the question immediately defeats the argument that was put forward by its proponent. But with the second kind of critical question, the argument will only be defeated if the respondent (the other party) offers some evidence. This could be explained as a matter of burden of proof. Some critical questions have a burden of proof attached while others do not. The argumentation scheme for argument from expert opinion is defeasible, since it is subject to the asking of critical questions matching the scheme. The scheme for argument from expert opinion has six such critical questions ([49], p. 310). CQ1

Expertise Question. How credible is E as an expert source?

CQ2

Field Question. Is E an expert in the field that A is in?

CQ3

Opinion Question. What did E assert that implies A?

CQ4

Trustworthiness Question. Is E personally reliable as a source? (continued)

13 Aristotelian Dialectic, Argumentation Theory …

271

(continued) CQ5

Consistency Question. Is A consistent with what other experts assert?

CQ6

Backup Evidence Question. Is E’s assertion based on evidence?

In general, in formal and computational argumentation systems, there are three ways of attacking an argument. You can attack one or more of the premises, you can attack the conclusion, or you can attack the inferential link between the premises and the conclusion, as shown in Fig. 13.3. The putting forward of the bias critical question at turn T2 by the respondent is of the third sort. It could defeat the argument from expert opinion, but only if the respondent backs it up with some evidence. The burden of proof is on the respondent to carry this out or else this critical question will not be sufficient by itself to defeat the argument. Let’s say that the audience knows that there was an investigation into Dr. Smith’s research on the videogames industry in 2001 and that it revealed her funding by the industry. On this basis, let’s say, the user marks the text box about the 2001 investigation by showing it with a green background meaning that premise is accepted as evidence. This situation is shown in Fig. 13.2, as it was in Fig. 13.1, but then there are differences to be explained. We still have to consider the remaining moves in the dialogue shown in Table 13.2. The respondent has a counterargument, in the form of the allegation that Dr. Smith is biased, and it has been backed up by some evidence, as shown in Fig. 13.2. What impact will this have on the argument diagram? To represent this sequence of argumentation in a way that is easy to understand, we will show how an undercutter is visually represented in ArguMed by the device of entanglement which displays an instance where one argument attacks another one by displaying an arrow going from one round argument node to another. This does not occur in typical argument diagrams of the kind that are widely used, but it is possible to represent it in an argument using this device of entanglement. How the videogames example can be visualized using an argument diagram with this modified version of Carneades using entanglement from ArguMed is shown in Fig. 13.2. Pro argument a1 is undercut by a con argument a2, the argument having the premise that Dr. Smith is not trustworthy. Entanglement is a dialectical notion in this instance because we can see that the effect of the con argument a2 is the same as the effect that would be achieved by asking the critical question of whether Dr. Smith is trustworthy. But once we translate the effect of this critical question into an argument diagramming format, it is shown as an entanglement in which one argument attacks another. But there is more to the chain of argumentation than that. We still have to consider the remaining moves in the dialogue shown in Table 13.2. Suppose that the proposition at the far right, the statement that there has been an investigation of the Parents’ Defense League, is accepted by the audience as evidential, it will be shown in a green background. Once the evidential proposition at the right citing the investigation of the Parents’ Defense League is shown as colored with a green background, Carneades will automatically color the next three propositions to the left, the conclusion of a5, the

272

D. Walton

conclusion of a4, and the conclusion of a3, with a green background. In a singlepremised argument showing that the one premise is accepted is enough to prove the conclusion is acceptable (provided that the form of the argument is valid). Assuming all these three arguments are defeasibly valid, once the first proposition is accepted as evidential by the audience, there is a propagation of acceptance that moves to the left in a sequence indicating acceptance of the conclusion that Dr. Smith is not trustworthy. The system now automatically colors the proposition citing funding by the videogame industry with a white background, changing it from green to white. This shows that the con argument a2 has successfully attacked and has defeated the ultimate conclusion that video games do not lead to violence. Because the argument labelled a1 has been undercut, the ultimate conclusion, the proposition that videogames do not lead to violence is shown in a white background, indicating that it is not accepted. This does not mean that it is rejected, but only that its former acceptance status has been removed. The reason for this is essentially that the untrustworthiness allegation has been successfully supported by evidence through a chain of intervening reasons. Burden of proof is an important tool for evaluating argumentation by AI dialectical methods such as the Carneades Argumentation System (Gordon, Prakken and Walton, in [13]). By this means the argument diagram is used as a tool to help analyze and evaluate the argumentation moves. In the profile of dialogue, burden of proof (BoP) shifts back and forth as each move is made. At the first turn in the dialogue shown in Table 13.2, the proponent makes the assertion that videogames do not lead to violence. At the same turn, the respondent requests some evidence to back up this claim by asking a why-question, a request for support of a claim. Since this dialogue is a persuasion dialogue, the requirement of BoP applies. If the proponent fails to give an answer which meets the requirements for the speech act of making an assertion, she will immediately lose the dialogue, according to the BoP protocol. However, the proponent responds by putting forward an appropriate argument, an argument from expert opinion. Because this argument fits the argumentation scheme for argument from expert opinion, provided that the premises are accepted by the respondent, it will shift the burden of proof back to the respondent’s side. Natural language examples of dialogue such as the videogames example generally have a lot of complications because they are expressed in natural language with all its attendant ambiguities possibilities of different interpretations. Even so, this relatively simple analysis of the videogames example gives the reader some idea of how a dialogue can be mapped into a sequence of argument diagrams that can be used to identify, analyze and evaluate key moves in a sequence of argumentation. This argument diagram and the method of evaluating it, is displayed in a simplified form representing the way of evaluating argumentation used in Carneades. Carneades is merely one of several systems currently being used to analyze and evaluate argumentation in artificial intelligence by using argument diagrams (also known as argument graphs). The illustrative point being made here is that of showing the reader in general

13 Aristotelian Dialectic, Argumentation Theory …

273

how a dialogue can be mapped into an argument diagram to help the user of a computational argumentation system to evaluate the sequence of argumentation extracted from the dialogue. Once we get a better grasp of the characteristics of Aristotelian dialectic, the idea of fitting the topics (or argumentation schemes) into the framework of Aristotelian dialectic really helps to show how the new dialectic is related to the old dialectic. Particularly important here is Slomkowsi’s theory that the underlying logical structure of the Aristotelian topics is the argument form called hypothetical syllogism. It appears evident from reading Slomkowsi’s chapter on the topics that the form of argument he calls hypothetical syllogism is the same as the one now called defeasible modus ponens (DMP) in argumentation studies ([9], p. 366). For these reasons, I would suggest that an important topic for future research is to consider whether the Aristotelian topics fit DMP.

13.10 Conclusions The arguments put forward in this chapter support the view that all the types of argumentative dialogue identified in the standard dialogue typology in Table 13.3 were studied by Aristotle under the category of dialectic (with the exception of negotiation and possible exception of deliberation). This conclusion is highly significant because it shows that the current approaches to argumentation theory are much more closely associated with Aristotelian dialectic than has been previously thought. Of course it is generally accepted that contemporary argumentation theory has its roots in ancient Greek philosophy, and especially in Aristotelian dialectic. But now there is new evidence showing that the standard dialogue typology used in argumentation fits even better than one might have thought with Aristotle’s structure of dialectical argument. Van Laar and Krabbe ([35], p. 43) showed through a series of examples that although examination dialogue should be seen as comprising two subtypes, persuasion dialogue and inquiry, it also needs to be seen as a mixed dialogue type because realistic examination dialogues often embed dialogical testing methods. More work needs to be done to get a better picture of how the model of Aristotelian dialectic provided by ACADEMIC1 , supplemented by the ten additional features, fits with real examples of the types of dialogue studied in the recent literature in argumentation and AI. To help get an overview of the terrain, Fig. 13.4 is a graph that represents, in rough outline, how these dialectical structures are related to each other. Even though examination dialogue can in some instances include negotiation, how significant this embedding is has yet to be determined. So no connection between negotiation and examination has been put in Fig. 13.4. Aristotle did not appear to include deliberation under dialectical argumentation either, but his discussion of it in places in his writings on ethics fits very nicely with the current work on modeling deliberation dialogue in argumentation in AI. So deliberation is also included in Fig. 13.4, but no known connection so far between it and examination dialogue has

274

D. Walton

Fig. 13.4 Relating Aristotelian dialectic to types of dialogue

been established. Similarly, it has not yet been determined what the relationship is between eristic dialogue and negotiation, so no connection between them is represented in Fig. 13.4. As noted in Sect. 13.3, this chapter has not tried to analyze or classify didactic (pedagogical) dialogue, an important subject for the study of examination dialogue. The system of connections shown in Fig. 13.4 is only a hypothesis that is meant to be open to improvement by subsequent research. Examination is classified as a mixed type of dialogue that combines informationseeking, inquiry and persuasion dialogue. The Socratic elenchus is tentatively classified as a species of examination dialogue, but this hypothesis is controversial for specialists in Greek philosophy, as shown by Scott ([31], p. 4) in his introduction to a collection of papers entitled Does Socrates Have a Method? He writes that there is no general agreement among Greek scholars on the meaning of the elenchus, but if it refers to a distinctive kind of process, it could mean “to cross-examine” or “to put to the test”. If this hypothesis is on the right track, it could be at least one interpretation of the elenchus that could reconstruct it as a species of examination dialogue of the kind modeled in [31]. Moreover, if there turns out to be sufficient evidence to support this approach, Krabbe’s formal model of Aristotelian dialectic could provide the first step to also providing a model of the central core of the elenchus as a species of cross-examination dialogue seen as a philosophical method. This approach to examining the structure of Aristotelian dialectic as a framework of argumentation has been made possible by the precise articulation of its central core by means of the formal argumentation model of Krabbe in [19]. What is especially notable here is that Krabbe formally modeled the central part of the structure of the pragma-dialectical critical discussion type of dialogue (CD1 ) and a Hamblinstyle formal model of persuasion dialogue (CB) that has strategies for preventing the answerer from seeing the end result of a line of questioning. These results showed that there is a very strong provable connection between Aristotelian dialectic and current theories of argumentation. However, the precise details of how these different models

13 Aristotelian Dialectic, Argumentation Theory …

275

of dialogue fit with each other needs to be left as a project for future research. This project is now made possible by the identification of the defining characteristics of Aristotelian dialectic. Ten defining features of Aristotelian dialectic that lie outside the core structure of Krabbe’s formal model of the core of dialectic ACADEMIC1 were identified in Sect. 13.5. (1) (2) (3a) (3b) (4) (5) (6) (7) (8) (9) (10)

Rules of fair play enforced by the audience. The possibility of offering a “solution” that explains the problem. The possibility of asking for clarification of the initial question posing the problem. The possibility of pointing out an ambiguity in the question. The possibility of contesting the validity of an argument. The possibility of allowing inductive reasoning. The possibility of offering a counter-example as a response to an inductive argument. The necessity for the aporetic aspect of the initial problem to be discussed. The acceptance and use of endoxical propositions. The use of topics for representing kinds of arguments. The use of argumentation strategies.

A research program suggested by this finding is to extend the core formal structure to build other formal models that include more and more of these ten features, using the profiles of dialogue tool, and other tools such as argument mapping and argumentation schemes. This program should start by sequentially building stronger and stronger formal systems of dialectic that gradually model more and more of these features. The ultimate goal should be to build a full formal system of Aristotelian dialectic that models all ten features, generating new dialogue models that can be applied to examples of ancient arguments known to be interesting. Following Sect. 13.5 some other issues were discussed in a tentative way that posed some problems needing further investigation by collaborative work between those working in ancient philosophy and those working in contemporary argumentation theory. One of these concerns the structure of deliberation dialogue, a current topic being investigated by the use of formal argumentation models in AI (McBurney et al., in [25]; Walton, Toniolo and Norman, in [50]). Another topic was the place of examination dialogue in Aristotelian dialectic and its applicability to current concerns in argumentation and AI (Dunne et al., in [10]). A final area which received only the briefest commentary, despite its importance, is the role of the topics in Aristotelian dialectic and their relationship to argumentation schemes. Acknowledgements The author would like to thank the Social Sciences and Humanities Research Council of Canada for Insight Grant 435-2012-0104: The Carneades Argumentation System. The author would also like to thank Benoit Castelnerac, Jakob L. Fink and Christopher Roser for discussions and help with translations.

276

D. Walton

References 1. Aristotle. (1933). Metaphysics. Loeb Classical Library. Harvard University Press. http://www. hup.harvard.edu/catalog.php?isbn=9780674992993. 2. Aristotle. (1939). Topics (E. S. Forster, Trans.). Harvard University Press. 3. Aristotle. (1984). Nicomachean ethics (W. D. Ross, Trans., J. O. Urmson, Revised). In J. Barnes (Ed.), The complete works of Aristotle. Princeton University Press. 4. Atkinson, K., & Bench-Capon, T. J. M. (2007). Practical reasoning as presumptive argumentation using action based alternating transition systems. Artificial Intelligence, 171, 855–874. https://doi.org/10.1016/j.artint.2007.04.009. 5. Barnes, J. (1980). Aristotle and the methods of ethics. Revue Internationale de Philosophie, 34(3), 490–511. 6. Black, E., & Atkinson, K. (2009). Dialogues that account for different perspectives in collaborative argumentation. In Proceedings of the Eighth International Conference on Autonomous Agents and Multiagent Systems (pp. 867–874). 7. Black, E., & Hunter A. (2009). An inquiry dialogue system. Journal of Autonomous Agents and Multi-Agent Systems, 19(2): 173–209. 8. Bolton, R. (1999). The epistemological basis of aristotelian dialectic. In M. Sim (Ed.), From Puzzles to Principles: Essays on Aristotle’s Dialectic. Lexington Books. 9. Castelnerac, B., & Marion, M. (2009). Arguing for inconsistency: Dialectical games in the academy. In G. Primiero & S. Rahman (Eds.), Acts of knowledge: History, philosophy and logic (pp. 43–82). College Publications. 10. Dunne, P. E., Doutre, S., & Bench-Capon, T. J. M. (2005). Discovering inconsistency through examination dialogues. In Proceedings IJCAI-05 (International Joint Conferences on Artificial Intelligence), Edinburgh (pp. 1560–1561). Available at: http://ijcai.org/search.php. 11. Fink, J. L. (2012). Introduction. In J. L. Fink (Ed.), The development of dialectic from Plato to Aristotle (pp. 1–23). Cambridge University Press. 12. Gordon, T. F., Friedrich, H., & Walton, D. (2018). Representing argumentation schemes with constraint handling rules. Argument & Computation, 9(2), 91–119. 13. Gordon, T. F., Prakken, H., & Walton, D. (2007). The carneades model of argument and burden of proof. Artificial Intelligence, 171(10–15), 875–896. 14. Gordon, T. F. (2010). The carneades argumentation support system. In C. Reed & C. W. Tindale (Eds.), Dialectics, dialogue and argumentation (pp. 145–156). College Publications. 15. Guthrie, W. K. C. (1981). A history of Greek philosophy. Cambridge University Press. 16. Hamblin, C. L. (1970). Fallacies. Methuen. 17. Hamblin, C. L. (1971). Mathematical models of dialogue. Theoria, 37(2), 130–155. 18. Kok, E. M., Meyer, J. J. C., Prakken, H., & Vreeswijk, G. A. W. (2011). A formal argumentation framework for deliberation dialogues. In P. McBurney, I. Rahwan, & S. Parsons (Eds.), Argumentation in multi-agent systems. Lecture Notes in Computer Science (Vol. 6614, pp. 31–48). Springer. 19. Krabbe, E. C. W. (2013). Topical roots of formal dialectic. Argumentation, 27(1), 71–87. 20. Krabbe, E. C. W., & van Laar, J. A. (2007). About old and new dialectic: Dialogues, fallacies and strategies. Informal Logic, 27(1), 27–58. 21. Leff, M. (2006). Up from theory: Or I fought the Topoi and the Topoi won. Rhetoric Society Quarterly, 36(2), 203–211. 22. Lorenzen, P., & Lorenz, K. (1978). Dialogische Logik. Wissenschaftliche Buchgegesellschaft. 23. Macagno, F., Walton, D., & Tindale, C. (2014). Analogical reasoning and semantic rules of inference. Revue Internationale de Philosophie, 207(4), 419–432. 24. Macagno, F., Reed, C., & Walton D. (2017). Argumentation schemes, history, classifications and computational applications. IFColog Journal of Logics and Their Applications, 4(8), 2493– 2556. 25. McBurney, P., Hitchcock, D., & Parsons, S. (2007). The eightfold way of deliberation dialogue. International Journal of Intelligent Systems, 22, 95–132. https://doi.org/10.1002/int.20191.

13 Aristotelian Dialectic, Argumentation Theory …

277

26. Prakken, H. (2006). Formal systems for persuasion dialogue. The Knowledge Engineering Review, 21(2), 163–188. 27. Prakken, H. (2011). An overview of formal models of argumentation and their application in philosophy. Studies in Logic, 4(1), 65–86. 28. Rahwan, I., & McBurney, P. (2007). Argumentation technology: Introduction to the special issue. IEEE Intelligent Systems, 22(6): 21–23. 29. Reed, C. (2006). Representing dialogic argumentation. Knowledge-Based Systems, 19(1), 22– 31. 30. Ross, W. D. (1949). Aristotle (5th ed.). Methuen. 31. Scott, G. A. (2002). Introduction. In G. A. Scott (Ed.), Does socrates have a method? (pp. 1–16). Penn State University Press. 32. Simari, G., & Rahwan, I. (2009). Argumentation in artificial intelligence. Springer. 33. Slomkowski, P. (1997). Aristotle’s topics. Brill. 34. Van Eemeren, F. H., & Grootendorst, R. (2004). A systematic theory of argumentation: The pragma-dialectical approach. Cambridge University Press. 35. Van Laar, J. A., & Krabbe, E. C. W. (2010). Examining the examination dialogue. In C. Reed & C. W. Tindale (Eds.), Dialectics, dialogue and argumentation (pp. 31–44). College Books. 36. Verheij, B. (2001). Legal decision making as dialectical theory construction with argumentation schemes. In ICAIL 2001: The Eighth International Conference on Artificial Intelligence and Law. ACM. The full paper is available at http://www.ai.rug.nl/~verheij/publications/pdf/arg sch.pdf. 37. Verheij, B. (2003). DefLog: On the logical interpretation of prima facie justified assumptions. Journal of Logic and Computation, 13(3), 319–346. 38. Verheij, B. (2005). Virtual arguments: On the design of argument assistants for lawyers and other arguers. TMC Asser Press. 39. Vlastos, G. (1983). The socratic elenchus. In J. Annas (Ed.), Oxford studies in ancient philosophy (pp. 27–58). Oxford University Press. 40. Walton, D. (2002). Are some Modus Ponens arguments deductively invalid? Informal Logic, 22(1), 19–46. 41. Walton, D. (2006). Examination dialogue: An argumentation framework for critically questioning an expert opinion. Journal of Pragmatics, 38(2006), 745–777. 42. Walton, D. (2015). Profiles of dialogue: A method of argument fault diagnosis and repair. Argumentation and Advocacy, 52(2), 91–108. 43. Walton, D., & Gordon, T. F. (2015). Formalizing informal logic. Informal Logic, 35(4), 508– 538. 44. Walton, D., & Macagno, F. (2006). Common knowledge in argumentation. Studies in Communication Sciences, 6(1), 3–26. 45. Walton, D. (1984). Logical dialogue-games and fallacies. University Press of America. http:// www.dougwalton.ca/books/LDG84bk.pdf. 46. Walton, D. (1998). The new dialectic: Conversational contexts of argument. University of Toronto Press. 47. Walton, D. (2013). Methods of argumentation. Cambridge University Press. 48. Walton, D., & Krabbe, E. C. W. (1995). Commitment in dialogue. State University of New York Press. 49. Walton, D., Reed, C., & Macagno, F. (2008). Argumentation schemes. Cambridge University Press. 50. Walton, D., Toniolo, A., & Norman, T. J. (2014). Missing phases of deliberation dialogue for real applications. In Proceedings of the Eleventh International Workshop on Argumentation in Multi-Agent Systems. http://www.dougwalton.ca/papers%20in%20pdf/14ArgMAS.pdf. 51. Wlodarczyk, M. (2000). Aristotelian dialectic and the discovery of truth. Oxford Studies in Ancient Philosophy, 18, 153–210.