From Puzzles to Principles?: Essays on Aristotle's Dialectic [converted] 0739100289, 9780739100288, 0739100297

Scholars of classical philosophy have long disputed whether Aristotle was a dialectical thinker. Most agree that Aristot

169 84 2MB

English Pages 288 [266] Year 1999

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

From Puzzles to Principles?: Essays on Aristotle's Dialectic [converted]
 0739100289, 9780739100288, 0739100297

Table of contents :
From Puzzles to Principles?
Contents
Introduction
Chapter One— For Dialectic Puts Questions about Matters which Philosophy Knows
Conclusion
Special Notes
Notes
Chapter Two— The Diodorean Modalities and the Master Argument
I
II
III
Notes
Chapter Three— Dialectic and Method in Aristotle
I— 'This in Itself Is Sufficient Proof'
II— Dialectical Argument and the Art of Dialectic
III— The Starting Points of Dialectic: 'Not Everyone's Opinions Count Equally'
IV— Dialectic and the Route to the Principles
V— Conclusions
Notes
Chapter Four— The Epistemological Basis of Aristotelian Dialectic
I— Conflicting Views of Dialectic
II— What Is Dialectic? A First Approach
III— Objections to the Sufficiency of Dialectic
IV— How to Do Dialectic: Some Initial Questions
V— Different Forms of Dialectic
VI— Give and Take in Dialectical Inquiry
VII— Reasoning from What Is Most Endoxon
VIII— Rank Ordering among Endoxa
IX— Dialectic and Peirastic in Scientific Inquiry
X— Some Main Features of Peirastic
XI— Dialectic in Science: A Standard View
XII— Revision of the Standard View
XIII— Dialectic in Science: The Traditional View
XIV— Dialectical Justification in Science
XV— The Merits of Peirastic Justification
Notes
Chapter Five— Choosing the Good in Aristotle's Topics
I
II
III
IV
V
Notes
Chapter Six— The Normalization of Perplexity in Aristotle
Notes
Chapter Seven— Dialectic, Contradiction, and Paraconsistency in Aristotle
I
II
III
IV
V
Notes
Chapter Eight— Perception and Dialectic in Aristotle's De Anima
I
II
III
Notes
Chapter Nine— Aristotle's Discovery of First Principles
Puzzles
An Irwin Excursus
An Aristotelian Solution
Conclusions
Notes
Chapter Ten— Dialectical Communities: From the One to the Many and Back
I— Dialectic, Definition, and Being
II— Unity of the Topics
III— Two Communities in Aristotle's Dialectic
IV— Rhetoric and Dialectic
Notes
Chapter Eleven— Poetry, History, and Dialectic
I
II
III
IV
Notes
Selected Bibliography
Primary Texts
Secondary Sources
Index
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
Q
R
S
T
U
V
W
Z
About the Contributors

Citation preview

Page iii

From Puzzles to Principles? Essays on Aristotle's Dialectic edited by May Sim

LEXINGTON BOOKS Lanham • Boulder • New York • Oxford

  

Page iv

Disclaimer: This book contains characters with diacritics. When the characters can be represented using the ISO 8859­1 character set (http://www.w3.org/TR/images/latin1.gif),  netLibrary will represent them as they appear in the original text, and most computers will be able to show the full characters correctly. In order to keep the text  searchable and readable on most computers, characters with diacritics that are not part of the ISO 8859­1 list will be represented without their diacritical marks. LEXINGTON BOOKS Published in the United States of America by Lexington Books 4720 Boston Way, Lanham, Maryland 20706 12 Hid's Copse Road Cumnor Hill, Oxford OX2 9JJ, England Copyright © 1999 by Lexington Books All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical,  photocopying, recording, or otherwise, without the prior permission of the publisher. British Library Cataloguing in Publication Information Available Library of Congress Cataloging­in­Publication Data From puzzles to principles?  :  essays on Aristotle's dialectic / edited by May Sim. p.   cm. Includes bibliographical references and index. ISBN 0­7391­0028­9 (cloth : alk. paper) ISBN 0­7391­0029­7 (pbk. : alk. paper) 1. Aristotle.  2. Dialectic.  I. Sim, May, 1962–  B491.D5 F76 1999 110—dc21                              99­32601                                                         CIP Printed in the United States of America T M The paper used in this publication meets the minimum requirements of American National Standard for Information Sciences—Permanence of Paper for Printed  Library Materials, ANSI/NISO Z39.48­1992.

  

Page v

For Charles Wesley DeMarco A man of principles who puzzles

  

Page vii

Contents

  

Introduction by May Sim

ix

Chapter One  For Dialectic Puts Questions about Matters which Philosophy Knows Martha Husain, Brock University

1

Chapter Two  The Diodorean Modalities and the Master Argument Lenn E. Goodman, Vanderbilt University

15

Chapter Three  Dialectic and Method in Aristotle Robin Smith, Texas A&M University

39

Chapter Four  The Epistemological Basis of Aristotelian Dialectic Robert Bolton, Rutgers University

57

Chapter Five  Choosing the Good in Aristotle's Topics Eugene Garver, St. John's University

107

Chapter Six  The Normalization of Perplexity in Aristotle Gareth B. Matthews, University of Massachusetts

125

Chapter Seven  Dialectic, Contradiction, and Paraconsistency in Aristotle J. D. G. Evans, The Queen's University of Belfast

137

Page viii

  

Chapter Eight  Perception and Dialectic in Aristotle's De Anima Michael Ferejohn, Duke University

151

Chapter Nine  Aristotle's Discovery of First Principles Allan Bäck, Kutztown University

163

Chapter Ten  Dialectical Communities: From the One to the Many and Back May Sim, Oklahoma State University

183

Chapter Eleven  Poetry, History, and Dialectic Edward Halper, University of Georgia

215

Selected Bibliography

229

Index

237

About the Contributors

249

Page ix

Introduction May Sim Dialectic is in dispute. Aristotelian commentators dispute the nature of Aristotelian dialectic along with whether Aristotle was a dialectical thinker in his (or anyone  else's) sense of 'dialectic'. Most agree that Aristotle contrasts dialectical reasoning with demonstrative reasoning, where the former reasons from generally accepted  opinions and the latter reasons from the true and primary. Starting with a grasp on truth, demonstration never relinquishes it. Starting with opinion, how could  dialectical reasoning ever reach truth, much less the truth about first principles? Is dialectic then an exercise that reiterates the prejudices of one's times, and at best  allows one to persuade others by appealing to these same prejudices, or is it the royal road to first principles and philosophical wisdom? Commentators argue both  these purposes. Both claims have textual support in the Topics and its counterpart, On Sophistical Refutations (De Sophisticis Elenchis, or SE). Both positions  mark the endpoints of a whole series of interpretative possibilities stretched between them. These possibilities are worth considering since each has philosophical  plausibility and textual support. This collection is dedicated to the exploration of these interpretative possibilities. In this introduction I shall outline the options and  arguments and link them to the most important texts. Given the pivotal significance of the question of first principles, thoughtful reflection on the nature and limits of  dialectic should play a crucial role in Aristotelian scholarship. Consider a series of options bounded by two extremes: at the one end are those who say dialectic attains to first principles, and at the other end those who say  dialectic has nothing to do with first principles whatsoever. Adherents to both extremes as well as to the possibilities in between are represented in this collection.  There are commentators who argue that dialectic never gets to first principles, need not invoke them in the course of its exercise, and even misdirects us from the  proper method of discovery—so that dialectic is the

  

Page x

complete antithesis of philosophy. Call this option one (1). At the other extreme, there are commentators who argue that Aristotelian dialectic is the only way to  achieve first principles since demonstration begins from them and induction assumes them. Other interpreters adopt an intermediate position, laying claim to a modest  function for dialectic which does not attain to first principles but helps in some way the search for first principles. There are a number of positions in this intermediate zone. Perhaps (2) dialectic is a rhetorical craft that helps us persuade others to truths. There are variations within  this construal of the function of dialectic as fundamentally persuasive. Perhaps dialectic is effective in convincing only those who have not made up their minds about  the issue at stake; or maybe dialectic can move even those who have made up their minds about something so that dialectic is an art of philosophical conversion. Another alternative (3) is to consider dialectic as peirastic; that is, not as a means of persuasion but of examination. Peirastic employs arguments which allow us to test  anyone who makes claims to knowledge and thus to establish his or her ignorance or expertise. These arguments do not persuade or convert those under examination;  they are calculated to provide the dialectician with knowledge of the state of understanding of some claimant. Peirastic as so far described provides the dialectician  with knowledge about claimants but not about any particular subject matter. Yet another reading of dialectic allows that dialectic provides some minimum of knowledge about some subject matter. Alternative (4) takes dialectic to be necessary  for philosophical inquiry in the sense that it produces the perplexity that is prerequisite to real inquiry, a state of mind that is essential to philosophy. Here dialectic is a  necessary propaedeutic to, but not a proper part of the discovery procedure; that is, dialectic is a stage in philosophical inquiry but not part of the achievement. It  concerns the questions, not the answers. Perhaps the function of dialectic is (5) recollection—providing reminders, or drawing out common sense into principled maxims. For instance, with respect to practical  subjects, dialectical topics such as 'ends are better than means', 'the proper cause of good is more choiceworthy than that which accidentally causes good', and 'that  which is more useful on every occasion is preferable to that which is only sometimes useful', act as reminders or maxims that are means to our decisions though they do  not make our decisions for us. Or again, dialectic (6) might unearth substantive new knowledge, helping to determine truths or facts, epistemic accomplishments necessary though not sufficient for  knowing first principles. Or finally, (7) perhaps there is a sort of dialectic that is the royal road to first principles. Suppose there is a road to first principles. The commentators who adopt the aforementioned positions would paint contrasting pictures of whether or how dialectic is  that roadway. (1) Interpreters at one extreme who seriously doubt dialectic's efficacy, who argue that dialectic is instead a false route, a hindrance to achieving first  principles, place dialectic off the road entirely. Continuing the

Page xi

same metaphor, (2) interpreters who doubt dialectic's ability to reach first principles but hold that dialectic can persuade others to adopt principles we know by other  means, depict it as useful for inviting others onto the road. (3) Defenders of dialectic as technique of examination present dialectic as allowing one to find out whether  another is or is not on the road to first principles. (4) Interpreters who take dialectic to produce puzzlement sketch it as the crossroads where one is lost and has to  stop to figure out the road signs before moving on. (5) Still other defenders of dialectic as a tool of self­persuasion depict it as providing reminders or signposts to help  the dialectician himself to stay on the road. (6) And those who think that dialectic provides us with some truths that lead toward first principles without attaining the  principles themselves are painting a picture where dialectic helps us to further our journey along the road to first principles. These stronger defenders of dialectic who  argue that dialectic gets us truths but not first principles are arguing that dialectic helps maintain us on the road or perhaps helps build the road but does not reach the  end of it. (7) Finally, the strongest defenders of dialectic portray it as an essential help in reaching the end of the road, in winning the prize of first principles. Martha Husain and Lenn Goodman provide versions of scene (1). Robin Smith's conviction that dialectic is persuasive makes him a detailer of (2), whilst Robert  Bolton's defense of dialectic's ability to find out if another is on the road to first principles finds him painting scene (3). Gareth Matthews's portrayal of dialectic's role in  provoking one to perplexity places him in picture (4). Eugene Garver's concern with dialectic as a tool of self­persuasion or reminder that helps the dialectician to stay  on the road is an instance of alternative (5). Pictures of dialectic which depict it as getting us truths but not the first principles (6) are presented by both J. D. G. Evans  and Michael Ferejohn. Finally, Allan Bäck, Ed Halper, and myself frame dialectic in the seventh way, portraying it as reaching and winning the prize of first principles.  The method of dialectic is, according to these interpreters, all pervasive in Aristotle's corpus, a large part of the method of his Metaphysics, both Nicomachean  Ethics and Eudemian Ethics, De Anima and Physics, just to name a few. The spectrum of possible pictures of the nature of dialectic and what it achieves is quite variegated, and the arguments harnessed to defend these views are no less  diverse and colorful. Starting from the contrast between dialectic and demonstration (where dialectic involves arguing for opposing sides to a claim and demonstration  argues to a valid conclusion from true premises), Martha Husain claims that dialectic is completely opposed to philosophy. Her essay "For Dialectic Puts Questions  about Matters which Philosophy Knows" thus marks one extreme of the spectrum. For her no sense of dialectic can get us to first principles at all; dialectic may even  be misleading in that search. Instead of trying to find a positive role for dialectic's arguing for opposite sides of a claim as some interpreters do, Husain exploits the  epistemological problems in such oppositions. By arguing for opposites, Husain maintains, dialectic disregards the priority of ousia and errs because it lays claim to  epistêmê in an unrestricted domain of being. By considering being in its full extension as involving

  

Page xii

opposites, dialectic does not capture ousia, for opposites do not belong in the category of ousia. Husain claims that opposites actually presuppose ousia and hence  any principles that appear as opposites cannot be first principles. Dialectic, however, tempts its practitioners to turn opposites into first principles. Metaphysical  epistêmê has the ability to unify opposites, unlike dialectic which leaves them irreconcilable. Dialectic then, on its own, blocks the way of inquiry, according to Husain,  by falsely supposing that it could argue for opposites and arrive at ousia as well. Husain ultimately accords a negative value to dialectic: by demonstrating that its own  questioning via opposites cannot attain to the contrariless ousia, the best thing dialectic can do is to demonstrate its own mistake and absurdity. Most commentators represented here take it to be beyond dispute that, if nothing else, Aristotle's dialectic is an art of persuasion, starting from common opinions and  culminating in something that is acceptable to the many or the wise. However, even such a moderate achievement for dialectic is challenged by Lenn Goodman in "The  Diodorean Modalities and the Master Argument." His concrete illustration shows how helpless dialectic is to change the mind of one committed to Diodorus's master  argument. Goodman sketches for us how for Diodorus, the modality of the possible is reduced to the actual. On this view, what is possible is what is or will be, whilst  the impossible is what is not and will never be. Goodman then shows that Aristotle's dialectic is powerless to persuade one committed to such modalities. One would  imagine that the reverse is also true. But Goodman's purpose is to defend Diodorus's argument by showing that it can refute Aristotle's notion of possibility from  Aristotle's own premises. Nonetheless, according to Goodman, Aristotle's dialectic is impotent to change the mind of one committed to Diodorus's modalities. It never  gets beyond the problem of begging the question when dealing with such a radical opponent. At best, Aristotle's dialectic can persuade only those who are not already  committed to a specific view incommensurable with Aristotle's own. Robin Smith's "Dialectic and Method in Aristotle" agrees with Husain that dialectic does not give us truth. Starting from the same contrast between demonstration and  dialectic, Smith argues however, that dialectic has important strengths in proceeding from common opinions. Thus, he also agrees with Goodman that dialectic's task is  to persuade others. More sanguine than Goodman about dialectic's ability to convince, Smith's view is that dialectic allows us to convert others and actually change  their minds. Smith takes Aristotle's distinction between dialectic and demonstration to be quite sharp and exclusive. Demonstration proceeds from true and primary  premises whilst dialectic depends on the answerer's responses to questions. Smith argues that the dialectical art needs to take into account the opinions of the  answerer as well as premises that will be acceptable to her. As such, dialectic, like rhetoric, studies the opinions of various types of people so that, knowing their  opinions, we will be able to argue successfully with them. Smith supports this rhetorical reading of Aristotle's dialectic by appealing to Aristotle's own comparisons of  dialectic to rhetoric. If dialectic is just a species of rhetoric, we must understand an

  

Page xiii

interlocutor's opinions in order to persuade her. If Smith is right, we can convert the otherwise persuaded to important truths. But on Smith's reading, dialectic offers  no help in reaching those truths in the first place. Robert Bolton offers another reading of dialectic that also starts from common opinions in ''The Epistemological Basis of Aristotelian Dialectic." This chapter is  especially helpful in setting up the debate on the use of dialectic in Aristotle's philosophical and scientific works. Charges against dialectic's usefulness in philosophical  and scientific works are fueled by its focus on opinions instead of empirical data. The case for dialectic's usefulness in such inquiries is propelled by Aristotle's claim  that theories should try to accommodate the endoxa and account for them as well as the phenomena. Bolton defends dialectic's usefulness in philosophical and  scientific inquiries by appealing to Aristotle's well­known view that inquiry moves from things that are more intelligible to us, to things that are in themselves intelligible.  Dialectic begins with endoxa, which are more intelligible to us. Dialectic is able to move toward what is intelligible in itself because of the 'common things' it uses in its  examinations. In Bolton's epistemic approach, the common things turn out to be premises that are commonly known because they are related to what we perceive.  Bolton argues that Aristotle's peirastic (as opposed to eristic) arguments can establish the truth or falsity of specific views rather than being merely sophistical. Peirastic  can decide whether a self­styled 'expert' is really ignorant. Dialectic for Bolton yields knowledge of common premises along with the knowledge whether the claimant  knows what he claims to know. Though far from being knowledge of first principles, this knowledge is related to first principles in following from them. Hence  dialectic's utility in peirastic examination. What dialectic does is collect the empirical data that the correct first principles will in turn explain. Discovery of first principles  is the scientist's or expert's province. Eugene Garver focuses on action rather than truth in his "Choosing the Good in Aristotle's Topics." But he takes his place here in the series since he accords dialectic a  more positive role, if minimally, arguing that topical dialectic is a help to decision making even though it yields no final answer. Other commentators who discuss the  topics in this collection take them to be items that help in the generation of the premises of arguments. Garver differs from these interpreters in claiming that dialectical  topics do not enter as premises or parts of arguments at all. Hence, though he uses the topics to understand the nature of deliberation and choice in Aristotle's ethics,  his view is that they stop short of determining choices for us. The topics are tools for reaching conclusions, not premises that justify conclusions. Hence, a topic such as  'ends are better than means', like all other topics that relate to action in Topics III, is at best a "reminder" or "maxim" for action. It reminds us that certain things are  more valuable than others, but in itself, does not make our choices for us. Ultimately, what choice one makes depends on the particulars of the context which delimit  what one can do, in spite of what one wants to do. The topics then indicate what is important when we choose. Dialectic is a tool of self­persuasion.

  

Page xiv

Bolton's willingness to grant that dialectic can establish the truth or falsity of specific views, along with Garver's view that dialectical topics are means of displaying what  is genuinely important for our choices, make positive though modest claims for the achievements of dialectic. Plainly, even these claims are too extreme for interpreters  who claim that dialectic has simply a rhetorical role, or those who claim that dialectic typically fails to persuade or actually misleads us. The question remains, however,  how the philosopher is to arrive at first principles. The remaining interpreters argue more positive roles for dialectic in this project. One standard feature of dialectic is its arrangement of rival opinions as contraries. Commentators in different ways find this feature significant or insignificant, a help or  a hindrance to inquiry. Gareth Matthews in "The Normalization of Perplexity in Aristotle" celebrates this contrariety because it ushers in the perplexity that starts  philosophical thinking. Matthews compares aporetic reasoning in Aristotle's works to Socratic perplexity as exemplified in the Meno. This method of addressing  contrary reasoning focuses on the significance of perplexity and wonder as states of mind appropriate to philosophical reflection. Matthews is particularly interested in  that strong kind of perplexity that is not easily—if ever—dispelled. He finds that Aristotle sometimes does and sometimes does not experience this sustained sort of  perplexity that continues wonder indefinitely. Two interpreters represented here try to dispel the claim that dialectic is nonlogical or illogical in character. Whilst Husain invoked the contrast between dialectic and  demonstration to undermine dialectic, these two commentators will try to elevate dialectic by showing how it could deal with opposing claims in logically  unobjectionable ways without collapsing into demonstration. Showing that dialectic has impeccable logical credentials lessens the antithesis between dialectic and  demonstration, so that if demonstration provides knowledge, dialectic need not lag far behind. Evans and Ferejohn show the role of dialectic in unraveling the different  senses of things and displaying the partial truths embellished in common claims. J. David G. Evans, a prominent voice in these debates since the appearance in 1977 of his Aristotle's Concept of Dialectic, tries to show in "Dialectic, Contradiction,  and Paraconsistency in Aristotle" that arguing for opposing sides of a claim is consistent with the principle of noncontradiction—the par excellence principle that  governs demonstration. Evans does so by exploring the relation between Aristotle's dialectic and more recent paraconsistent logics, i.e., logics that recognize true  contradictions. Evans tries to defend Aristotle against the charge that he does not take account of the paradoxes of truth in his handling of the principle of  noncontradiction. He shows that Aristotle accounts for the liar paradox in a way similar to the way he accounts for opposing views in dialectical reasoning. What is  more, Evans shows that Aristotle's dialectic does not compromise his principle of noncontradiction because he distinguishes between simple predication and qualified  predication. Dialectic schools its practitioners in qualified predication, an essential part of philosophical thinking.

  

Page xv

Evans ends by claiming that Aristotle is a dialectical logician who is the precursor of dialecticians like Hegel and Marx. (I hasten to add that Hegel and Marx insist on  real contradiction in things and view history dialectically, while Aristotle does not; they project into existence features of dialectic Aristotle reserves for opinion,  discourse, and knowledge.) Michael Ferejohn tries in "Perception and Dialectic in Aristotle's De Anima" to account for the presence of contradictory opinions in Aristotle's dialectic by analyzing  dialectic into four phases (which phases he inherits from G. E. L. Owen). These phases are: (1) aporetic survey of endoxa, (2) superimposition on investigation of  metaphysical concepts or distinctions, (3) disambiguation of question, and finally (4) resolution of problem. Ferejohn shows how one of these phases, the phase he  calls 'disambiguation', untangles different meanings of terms and can subsequently save the opposing opinions. This is done by showing how each opinion captures a  part of the phenomenon and so can share in the truth without agreeing entirely with its rivals. Ferejohn argues his case by using De Anima II.5 to illustrate how  Aristotle could resolve the oppositions between "like affects like" and "unlike affects unlike." Ferejohn, like Evans, relies on a discussion of how Aristotle uses  qualifications to explain that perception for instance, is neither alteration nor realization and yet shares in each of these. Most importantly, he tries to show that even  though the stage of 'superimposition' consists in superimposing metaphysical concepts or distinctions on the problem concerned, such a procedure, though a priori, is  not guilty of 'theoretical inflexibility'. This move has purchase against those inclined to believe that dialectical reasoning is limited to the prejudices from which it starts,  which would restrict faithfulness to the facts. Strategies of the sort deployed by Evans and Ferejohn aim at softening the threat of contradiction and hedging the oppositions of opinion on which dialectical  reasoning turns. They do so by showing how Aristotle could transcend opposing opinions without disregarding them, and hence reason from opposing opinions and  yet attain to the facts. These interpreters argue that dialectic can reach what is true without necessarily reaching what is primary. The final set of interpreters argues that  Aristotle's dialectic is the way to first principles, the way to what is metaphysically primary as well as true. Several commentators in this group argue versions of the thesis that dialectic provides the premises which demonstration (apodeixis) requires but cannot itself provide.  Aristotle asserts that induction (epagôgê) provides some premises for demonstration. This opens the issue of the relation between dialectic and induction. Allan Bäck  finesses the distinction between dialectic and induction, arguing that dialectic and induction are actually two stages of the same method, in "Aristotle's Discovery of  First Principles." By comparing Aristotle's dialectical reasoning and induction with the way modern science proceeds, Bäck shows how Aristotle could start with  ordinary beliefs and, through analyses, arrive at results that need not agree with ordinary beliefs. These results are the best­candidates­to­date for ultimate principles.  They subsequently become part of the inductive background for the further examination of principles that leads

  

Page xvi

to better candidates for even more general principles. Hence, through dialectic, past candidates for first principles are scrutinized and the best current candidates are  opened to future revision. In this way, Bäck merges dialectic and induction and hence brings dialectic and demonstration closer together. He argues the empiricism and  fallibilism of all knowledge for Aristotle, including knowledge of first principles. Readers convinced of the inerrancy to noûs might adapt Bäck's point, focusing on the  inability of us compound beings to know with certainty whether our noûs is acting. This would make Aristotle's view fallibilist, not by attributing fallibility to noûs, but  by ascribing to the compound whole a tendency to err concerning what is and what is not a deliverance of noûs. In "Dialectical Communities: From the One to the Many and Back," my strategy for dealing with dialectical and demonstrative reasoning in Aristotle is to argue for the  superiority of dialectic over demonstration in a decisive respect. By showing that dialectic captures the definition of being, I argue that dialectic is prior to  demonstration insofar as this definition appears in the premises of demonstrative metaphysical reasoning. To drive home the point that dialectic attains to true  definitions, I resort to an analysis of the Topics. I argue that a select group of topics plays, in Aristotle's dialectic, a dual role. First, they generate arguments that  comprise the materials of dialectical thinking. Second, they function in the examination or criticism that narrows these materials toward the truth. Following Aristotle, I  call the first function of dialectical topics a movement from the one to the many and the second a movement from the many to the one. In moving from one opinion to  many, Aristotle escapes the charge that his dialectic is limited to a pre­given set of opinions, since his topics generate an expanded set of alternatives. (For instance,  contraries are useful here; they help one see if all the options are covered by existing views.) In moving from a many of options to a single truth, Aristotle escapes the  charge that his dialectic is merely persuasive, or at best attains to a coherence. I argue that topics that are common in a strong sense can perform this truth­finding role  because they already embody principles of being. It belongs to philosophical dialectic—as opposed to dialectic for mental training or persuasion—to find the truths  and recognize the primary principles embedded in these topics. Ed Halper is another commentator who focuses on the use of the topics in Aristotle's dialectic. In "Poetry, History, and Dialectic," Halper defends the view that  dialectic for Aristotle proceeds from the setting out of facts or common opinions, which setting out Aristotle calls "history." These common opinions are then examined  by using the topics. This examination culminates in a definition of whatever is definable. Halper's case study compares the attempt to define tragedy and history in the  Poetics. History for Aristotle contrasts with poetry. It is not a collection of facts or opinions for any particular field, but rather a chronicle of particular human events.  Halper shows that Aristotle's dialectical method can define tragedy but not history because the former is general enough to be defined but the latter is too much a heap  of particulars to be defined or known. The topics allow us to see how the parts of tragedy are

  

Page xvii

united into a likeness of organic necessity so that tragedy can be defined and known. History lacks the unity and likeness of organic form—pace Hegel—and it is  dialectic that exhibits that history must fail of definition. This is why poetry, for Aristotle, is more philosophical than history. By stressing how dialectic can attain to first principles and definitions, Bäck, Halper, and I lie at the far end of the spectrum attributing the greatest philosophical import  to dialectic. Although the previous three commentators argue that dialectical reasoning can attain to facts or truths, they do not assert that it can arrive at first principles.  Their claim that dialectic can attain to truths at all is too strong for the first group of interpreters who attribute at best a rhetorical purpose to dialectic. Those who find  in dialectic a road to reach the principles find their claim too weak. Interpreters at each extreme find the interpreters at the other extreme not simply way too weak or  too strong but hard to fathom. Each author is a skilled and informed interpreter, and each supports his or her study by appeal to texts of Aristotle, under some construal. Naturally, each author  emphasizes certain texts over others. I want next to brief some of the most important of these texts. Pointing out the crucial passages in the Topics, Rhetoric, and On  Sophistical Refutations should clarify the challenges for each of these contributor's views and help us to judge their strengths and weaknesses. I begin with passages  that are most commonly cited, then present passages that challenge those who maintain that dialectic attains to first principles, followed by passages that challenge  those who maintain that dialectic does not attain to truth and is divorced from philosophy. Unsurprisingly, texts that suggest that there is a philosophical sort of dialectic  present the strongest challenge to those who argue that dialectic is no help or even a hindrance to philosophical inquiry. Conversely, texts that find Aristotle  discriminating sharply between dialectic and philosophy, or allying dialectic with rhetoric, present the strongest challenge for those who claim to a philosophical  dialectic that can reach first principles. The best interpretation will make the best use of the widest range of these texts. (The translations are mine unless otherwise  indicated. My purpose demands as literal a translation as possible, even if the style remains awkward.) We have already seen how the contrast between dialectical and demonstrative reasoning is key for most of our essayists. This contrast is found in Topics 100a26– b18: Reasoning, on the one hand, is demonstration when it is out of that which is true and primary, or out of such a kind that we have grasped the first principles (tên archên) of  knowledge of them through anything primary and true. On the other hand, reasoning is dialectical when it reasons out of endoxôn.

This passage is frequently interpreted as a challenge for those who claim that dialectic reaches truth or first principles or in any way escapes the orbit of the endoxôn  from which it begins. Strategies harnessed against this interpretation include elevating the status of dialectic so that it rivals the epistemic quality of demonstration while  remaining distinct from it, allying dialectic with induction

  

Page xviii

rather than demonstration, and arguing the superiority of dialectic over demonstration in respect of reasoning to first principles. The interpretation which limits dialectic to the orbit of opinion is buttressed by texts which suggest that dialectic not only starts from opinions but aims at the  preservation of opinion, such as Nicomachean Ethics 1145b2–7, here in Irwin's translation, Aristotle's First Principles, cf. EE 1235b13–18. We must, as in the other cases, set out the appearances, and first go through the puzzles. In that way we must prove, ideally, the truth of all the common beliefs about these  affections of the soul, or, if not of all, of most of them and the most authoritative (or 'important', kuriotata). For if the difficulties are resolved and the common beliefs are left  standing, that is an adequate proof. (30)

This text is, however, ambiguous. Those who cite this passage to support their view that dialectic starts out with appearances or common beliefs and argues to the  preservation of these common beliefs face the following difficulties. Namely, they face the difficulty of accounting for the puzzles that arise from the beliefs (which are  often inconsistent or conflicting), and the difficulty in the qualification that only those that are most important should be preserved. Interpreters who argue that dialectic  provides substantive philosophical outcomes fasten on this qualification, since the most important opinions are presumably the good or true ones. Other texts link the examination of opinion and puzzle solving with the finding of genuine truths. At Physics 211a7–11, Aristotle writes, We must try to conduct our investigation so that the 'what it is' of place will be given, so that the puzzles will be solved, and what seems to belong to place will belong to it, and  the cause of the discontent and the puzzles concerning it will be clear. For this is how each might be most beautifully shown (

).

The interpreter who uses Physics 211a7–11 to support a claim that dialectic aims at the essence or 'what it is' faces at least three difficulties. One challenge is the  reconciliation of the definition of place (or anything else for that matter) with the appearances about place. Is it always possible to reconcile the truth of something with  what seems? Another challenge, as Robin Smith points out, is that the beginning and end of the passage ("We must try ... this is how each might be most beautifully  shown,") seem to adumbrate an ideal that is to be aimed at rather than achieved by dialectic. Another difficulty is that the ideal of a beautiful showing which both yields  the truth and explains why the appearances and puzzles arise sounds hopelessly Platonic. Other challenges to interpretation come not from the finessing of distinctions or the many senses of key terms, but from Aristotle's syntax, or his use of connectives.  Consider the following text from Topics 101a35–37 (the context is a discussion about a dialectical problem's being an investigation that leads to choice and avoidance  or truth and knowledge, cf. 104b1–12).

  

Page xix With respect to the philosophic sciences [dialectic is useful] because if we can be raising difficulties on both sides, we shall easily make manifest the true and the false in each  thing. Still, in the first place, these [are useful] toward (pros) the first principles (tôn archôn) about each science.

Interpreters who use this passage to crown philosophical dialectic with the prize of first principles have to struggle with the slippery preposition 'toward' (pros). Does  this mean that dialectic is simply a road 'toward' the first principles, a road that requires other vehicles such as demonstration to deliver us to the end of our journey?  Or is it the case that dialectic itself is the road 'to' first principles, a road that delivers us to the door of the house of wisdom? This passage, like the others, makes it  clear that no single text can settle the issue of Aristotle's view of dialectic. Let us turn now to those passages that seem to support a modest role for dialectic, often highlighted by interpreters who are skeptical about the ability of dialectic to  give us knowledge and first principles. Consider Topics 105b30–34: On the one hand, with respect to philosophy, we must carry on according to truth, on the other hand, from the point of view of dialectic herself, according to opinions (doxan). All  propositions must be taken for the most part generally, and one should be made into many.

This passage with its radical distinction between philosophy and dialectic, bears some semblance to the distinction between demonstrative and dialectical reasoning in  Topics 100a26–b18, noted above. This passage says that philosophy proceeds according to truth while dialectic proceeds according to opinions. The contrast  between the two seems to condemn dialectic to the realm of opinions permanently since surely if it proceeds from opinions, it will remain with opinions. The separation  of science from dialectic is further reinforced by these passages from the Rhetoric. For just as we said at first, [rhetoric] is a part of dialectic and its likeness, for neither of them is a science about anything definite, but [about] how faculties are for procuring  arguments. (Rhet. I.2 1356a7) But as much as anyone tries to make dialectic or this [rhetoric] not just as they are, i.e., faculties, but science, one will prepare, without knowing it, a disappearance (aphanisas)  [making away with] of their nature, but one restores by changing into science subject matters that are certain things not only arguments (logôn). (Rhet. I.4 1359b6)

These passages liken rhetoric to dialectic, making both deal with arguments and not things. They also stress the nature of dialectic and rhetoric as faculties, i.e., mere  potentialities which contrast sharply with science that by definition actually is knowledge. The gulf between dialectic and science seems further widened when Aristotle  tells us that if we do not understand dialectic and rhetoric properly, i.e., as faculties, but as sciences, we would threaten the destruction of science itself.

  

Page xx

The case against those who argue for philosophical dialectic's ability to attain to first principles is strongest in Rhet. I.2 1358a21–22: And on the one hand, the former [i.e., common topics] will not make one prudent concerning any one genus for they are not a subject matter [hupokeimenon] either; on the other  hand, these latter [i.e., specific topics], as much as one is better in his choice of propositions, he will make another science and [he] ceases from dialectic and rhetoric. For if he  happens upon the first principles, it will no longer be dialectic or rhetoric but that being whose first principles he has.

Notice how Aristotle explicitly contrasts the attainment of first principles with dialectic and rhetoric. It is special sciences that deal with specific topics that achieve the  first principles or knowledge of anything definite. On the one hand, it seems that dialectical topics are too general to be able to tell us anything about any one subject  matter. On the other hand, Aristotle at other times suggests that dialectical topics are not general enough to provide the principles that encompass all things. That point is made  at SE 172a1–20: But now the dialectical is not about a definite genus, nor does it demonstrate anything, nor is it of the nature of the universal; for there is not any one genus that all is in, nor if  there were, it is possible for these things (

).

Furthermore, not only does dialectic proceed by questioning, but it relies on the answerer's responses to the questions raised. This seems to prevent it from  demonstrating anything, for there is nothing to prevent the answerer from responding incorrectly. Aristotle claims that if dialectic were to show anything, it would not  question the primary principles. On the other hand, since no special science questions the "primary and proper principles," perhaps this role of dialectic plays some  positive role. The difficulty is that the sciences depend on these principles and if dialectic knows them, and if the deliverance of dialectic is insecure, then the principles  of the sciences would be insecure. Our last group of commentators reasons that if dialectic investigates first principles, then this investigation must, somehow, deliver the goods rather than the sciences.  These commentators lean on passages that present challenges to interpreters who are skeptical about the ability of dialectic to attain definition, truth and knowledge of  first principles. Some of these passages also stress the commonality of the topics, and hence their usefulness across disciplines for establishing truth. Topics VIII.3  158b1–4 talks about the importance of first principles ( definitions.

  

) and how they must be defined. Aristotle asserts that often interlocutors refuse to provide 

Page xxi On the one hand, this kind of thing happens for the most part with respect to first principles (

). (Topics VIII.3 158b1–4)

Books I–VIII of the Topics focus on how we arrive at definition. Aristotle shows how dialectic gives us definitions of first principles, amongst definitions of other  things. Interpreters skeptical with respect to philosophical dialectic and its ability to deliver first principles must give an account of both the detailed discussions of  definition throughout the Topics and his claims that dialectic provides the definition of first principles. That Aristotle regards dialectic as at least some assistance in the  philosophic quest, at least in the Topics, seems beyond dispute. Aristotle often extols the benefit of arguing for both sides of a claim so that the results of both are clear  enough for a choice to be made. Such an accomplishment, Aristotle claims, is crucial to the attainment of knowledge and philosophic wisdom. To be able to take and to have taken in at a glance the results of each of two hypothesis is no small work with reference to knowledge and philosophic wisdom ( ) one of them. (Topics 163b9–13)

While texts such as On Sophistical Refutations 172a12–20 (above) seem to restrict the generality of the dialectician's common topics, other passages recommend  dialectical topics for another sort of generality, which we might call 'commonality'. Topics are the common things (ta koina), and Aristotle says we search for first  principles through the common things. Both assertions challenge the opponent of philosophical dialectic. For what I mean by dialectical and rhetorical syllogisms are those which concern what we call "topics" ( and many different in kind, such as the topic of the more and the less. (Rhet. 1.2 1358a21)

), which are those in common to law, physics, politics, 

It is clear then that we need not grasp the topics (tous topous) from all refutations but dialectic, for these are common (koinoi) to every art (technê) and faculty (dunamin). On the  one hand, it is according to science to look at the refutation of each science and see if it is apparent and if it is not real, against what is real, and why. On the other hand, [it is] of  the dialectician [to look at the refutation] out of the common (tôn koinôn) and [which is] not under one art. For if we have the source (ex hôn) of the generally accepted reasonings  (hoi endoxoi sullogismoi) about whatever [subject], we have the source (ex hôn) of the refutations; for the refutation is a reasoning of a contradictory, so that one or two  reasoning of the contradictory is a refutation. We have then as many [refutations] as all such, and if we have these, we have the solutions, for the solutions are the oppositions of  these [refutations]. (SE 170a35–b5)

  

Page xxii

In these passages, the dialectical topics are said to be common to all disciplines. In contrast to the limited task of a particular science that deals only with what is  specific to it, dialectic deals with what is common to all arts and faculties. Far from being a limitation, dialectic's concern with what is common is here portrayed as an  access to the source of all refutations, and ultimately, an access to what is common to all. The advocate of philosophical dialectic argues that this access is denied to  every special science but open to dialectic. The sources of which Aristotle speaks in these texts, according to the advocate, are the first principles. Furthermore, not only are the topics of dialectic common to all arts and sciences, but they are also commonly known and used by everyone. This comes across in  Aristotle's discussion of peirastikê, which is a kind of dialectic. Thus, it is clear that peirastikê is not knowledge of anything definite, it follows that it is about every subject, for every art uses certain common (koinois) principles. Accordingly,  everyone including the unscientific (hoi idiôtai) uses in some manner dialectic and peirastic. (SE 172a27–31)

This is a passage that finds Aristotle putting scientists and nonscientists, including dialecticians, on equal footing. That they are on equal footing is not surprising given  Aristotle's claim that peirastikê is a kind of dialectic that argues to truth and falsehood instead of being merely combative or persuasive. For instance, Aristotle says, Dialectic is the same as peirastikê. (SE 172a22) Further, since such kinds of arguments (tôn logôn) are for the sake of exercise (gumnasias) and examination (peiras) but not instruction (didaskalias), it is clear that they were  arguing not only to truth but also falsehood. (Topics 161a24–27)

Topics 161a24–27 not only asserts that peirastic arguments aim at truth, but also reinforces the view that there are different uses of dialectic in Aristotle. A more  detailed examination of these various uses of dialectic is visible in this next passage: Still of arguments (logôn) in discussion (dialegesthai) there are four kinds, didactic, dialectical, examinational and contentious (eristikoi). Didactic on the one hand are out of the  proper first principles of each learning and not out of the opinions of the answerer [for it is necessary that the learner is to trust the teacher]. Dialectical [arguments] are those out  of endoxôn, reasoning to a contradiction. Peirastikoi on the other hand, are not of the opinions held by the answerer and must be known to one who claims to have knowledge....  Eristikoi are those that reason or seem to reason out of what appears to be endoxon but are not. (SE 165a37­b9)

Since Aristotle sometimes uses 'dialectic' as inclusive of peirastic, eristic, etc., this passage shows that Aristotle sometimes uses 'dialectic' in a narrow sense and  sometimes in a broad or more inclusive sense. This passage fixes a narrower sense. Which sense of 'dialectic' Aristotle is using or talking about, and whether

  

Page xxiii

it argues to truth, or to teach, or to win or persuade, depends on the context. Difficulties in interpretation arise when one tries to limit Aristotle's use of dialectic to any  one of these (i.e., didactic, or peirastic, or eristic) uses to the exclusion of the others. A similar point pertains to Aristotle's uses and mentions of 'dialectician'. A good rule of thumb in understanding Aristotle's dialectician appears in Topics 164b3–6: ... to put it simply, the maker of propositions and objections is a dialectician. To put forward a proposition is on the one hand to make a many ( put forth an objection is to make a one into many.

)), on the other hand, to 

More elaborately, whatever else he is, a dialectician is one who can move from a one to a many and from a many back to a one. The advocate of Aristotelian dialectic  interprets this to mean that, throughout his making propositions and objections, the dialectician has his eye toward the thing itself, which allows him to argue ultimately  toward the end which is a single whole or essence. Thus, these commentators aim to remind us that Aristotle never tires of saying that dialectical arguments (in the  philosophical sense of dialectic, i.e., those that aim at truth), argue about things and not names. As he puts it, On the one hand an elenchos is a contradiction of one and the same—not name but thing (pragmatos). (SE 167a23–24) On the one hand therefore, the dialectician looks at the common according to the things (ta koina kata to pragma), on the other hand the sophist does this apparently. (SE  171b6–7)

Interpreters who argue for a positive role for dialectic believe that the philosophical dialectician's use of common topics is ''according to the things" and this shows both  that the things themselves are the measure of truth and that the common topics help us get these truths. The most important preliminary point to glean from this survey of texts and interpretation is that Aristotle uses 'dialectic' in various ways. Problems invariably arise for  interpretation when this variety is neglected. As Aristotle tells us at Topics 101a26–27, there are three uses of dialectic: ... for mental training (pros gumnasian), for conversations (pros tas enteuseis), and for philosophic sciences (pros tas philosophian epistêmas).

The most satisfying interpretations of dialectic in Aristotle will take account of the fullest range of texts, and will draw on Aristotle's own distinctions in regard to his  manifold uses of 'dialectic' to sort these texts out. The question then remains what to make of Aristotle's philosophical dialectic—dialectic "with reference to  philosophic sciences." Is it the job of this philosophical dialectic to discover or defend first principles? If so, is dialectic successful at this job? If Aristotle believed this  at one time, is it a belief he abandoned? Does

  

Page xxiv

Aristotle employ dialectical method in his mature works such as the Physics, Metaphysics, or Ethics? We might also use these three functions of dialectic enumerated by Aristotle to classify the seven options defended by our authors. So far as they acknowledge the  three (or four) basic uses of dialectic according to Aristotle, our first group of interpreters emphasizes the pedagogical function and limits philosophical use of dialectic  to a negative role. Or members of this first group might claim that dialectic in the Topics is an early, failed stab at first philosophy, supplanted later by metaphysics  proper. The second and third groups of commentators emphasize a rhetorical mode. For them dialectic refers to conversational subject matter, the topics a dialectician  aims to make persuasive to others. Beyond this, dialectic is philosophically fruitful in that dialectic examines those who claim to know a subject and finds out if they  really do. With the fourth group, we turn to the philosophical use of dialectic, at least, in a preliminary way, so far as it is an art of raising and framing puzzles vital to  proper inquiry. The fifth commentators fasten on the conversational function of dialectic in that discourse with oneself by which one reminds oneself what is important  and so seeks to persuade oneself. The sixth group of interpreters further emphasizes the philosophic function of dialectic. For them dialectic procures truth and  knowledge thus paving the way to first principles. The seventh and final group of commentators accords maximum weight to Aristotle's talk of a specifically  philosophical sort of dialectic. For these interpreters, dialectic in its philosophical use attains to first principles and definitions and hence secures philosophical wisdom  about the sources of all things. A number of the chapters in this collection had their origin at the Fourteenth Annual Joint Meeting of the Society for Ancient Greek Philosophy and the Society for the  Study of Islamic Philosophy and Science in Binghamton, New York, 1995. I am indebted to both Tony Preus and Parviz Morewedge for bringing me to the Program  Committee of these meetings. Parviz's warmth, generosity, and encouragement have been the source of this as well as an earlier collection. For his friendship and  guidance, I am pleased to express my gratitude. I am thankful to the participants for sessions stimulating enough to suggest a collection on Aristotle's dialectic. I am  grateful to all the contributors to this volume, for their insightful essays, and for their patience over the years this project has taken to reach completion. My thanks go  to Éditions CNRS for permission to reprint Robert Bolton's "The Epistemological Basis of Aristotelian Dialectic" (In Biologie, Logique et Métaphysique chez  Aristote, edited by D. Devereux. and P. Pellegrin. Paris: Éditions de CNRS, 1990, 185–235), and I owe thanks to both Rob Bolton and Pierre Pellegrin for their  help in obtaining permission for the reprint. I am indebted to Murry Hodgson for her assistance in formatting the essays of this collection, and her patience in showing  me the necessary skills to manage the rest. It pleases me to acknowledge my sons, Aris Ezra and Ambrose Yosha, now six and two, for the joys they bring to my life.  Most of all, I wish to thank my husband, C. Wesley

  

Page xxv

DeMarco, for the encouragement to work on yet another collection, and for pressing the importance of Aristotle's dialectic. Wes has been an indispensable source of  editorial suggestions, and help with the numerous drafts the introduction has undergone. Most importantly, it is the excellent way in which Wes has borne the burden of  caring for our young sons that made possible my peace of mind and much needed time for this work and other projects. For his intellectual and emotional support, I  am most grateful.

  

Page 1

Chapter One—  For Dialectic Puts Questions about Matters which Philosophy Knows Martha Husain Aristotle is generous in according the status of epistêmê to a wide variety of human cognitive endeavours. An epistêmê is defined as having a distinctive subject  matter, a distinctive method, and the ability to attain truth or knowledge. 1  The subject matters of all epistêmai are contained within the extension of being because that  extension is all­inclusive. A given epistêmê can therefore have a distinctive subject matter only by investigating a distinctive part of that extension. The methods of all  epistêmai are contained within the intension of being because that intension, too, is all inclusive. A given epistêmê can therefore have a distinctive method only by  investigating its subject matter qua a distinctive part of that intension. Its ability to attain truth or knowledge accordingly rests on its investigating a distinctive part of the  extension of being qua a distinctive part of its intension. But here it must be added that the intensional qua has to be appropriate to the extensional demarcation of the  subject matter. For example, investigating number qua poson (the category of quantity) is an epistêmê while investigating it qua poion (the category of quality) is not.  Given that the status of epistêmê rests on these conditions, Aristotle denies it to any human cognitive endeavour that fails to meet them. In the Metaphysics he denies  it to dialectic. Aristotle's recognition of a plurality of epistêmai is made possible by his extensional and intensional ontological pluralism. This is expressed in his metaphysical  leitmotif: "Being is spoken in many senses," which his epistemological realism allows us to expand as follows: "Being is truly spoken in

  

Page 2

many senses because it is in many ways." Being is extensionally in many ways because there are many individual substantial entities in many species and genera,  which have many properties in different secondary categories. It is intensionally in many ways because each of the individual substantial entities has its own primary  physis or intensional nature, which as governing archê imprints itself on its generic nature (its intelligible matter or potentiality), on its physical matter (its perceptible  matter or potentiality), and on all its properties in the different secondary categories by way of formal­final causality. Aristotle's ontological pluralism is radical because  he pluralizes being not only extensionally but also intensionally. Yet it is cosmic because the total being of any one individual substantial entity is a microcosm whose  internal complexity stands in ordered intensional bonds of unity, pros hen focused on the entity's own primary physis. And the aggregate of all these microcosms forms  a macrocosm because they stand in ordered intensional bonds of specific and generic kinship, they are connected relationally, and they interact in active and passive  causal ways. All these are governed by the distinctive primary physis of each entity. Being is common to all beings not kath hen but pros hen because it is always the  being of a substantial entitative being and therefore governed by the primary physis of that being. This means that all categories, even pros ti and poiein and paschein, are always pros hen focused on the primary physis of an individual substantial entity. Just as  there are no unattached nonrelational properties, so there are no unattached relational ones. For an individual substantial entity relates to and interacts with others only  as what it is in its own substantial nature, which imprints itself on all its categorial modes of being: "the substance of each thing and the essence of each; for this is the  limit of knowledge, and if of knowledge, of the object also" (Met. V 1022a 8–10). "For there is nothing either great or small, many or few, or, in general, relative to  something else, which without having a nature of its own is many or few, great or small, or relative to something else" (Met. XIV 1088a27–9; cf. III 996b14–8). This radical yet cosmic ontological pluralism allows Aristotle to accord the status of epistêmê both to philosophy, whose investigation of being is completely general  extensionally and intensionally, and to several special sciences whose investigation of being is restricted in scope extensionally and intensionally. The former is  metaphysical epistêmê, which investigates the full extension of being (panta ta onta) qua its full intension (qua on). Each of the latter investigates a part of the  extension of being (a genus or a property) qua a part of its intension (qua a generic nature or a category). Dialectic, however, does neither and so is denied the status  of epistêmê. For while it is like philosophy in that it investigates the full extension of being (panta ta onta), it is unlike any epistêmê in that it does not investigate its  subject matter qua either the full or a partial intension of being: Dialecticians and sophists assume the same guise as the philosopher, for sophistic is Wisdom which exists only in semblance, and dialecticians embrace all things in their  dialectic, and being is common to all things; but evidently their

  

Page 3 dialectic embraces those subjects because they are proper to philosophy. (Met. IV 1004b 17–22)

Instead, dialectic investigates panta ta onta qua formal­abstract general concepts, the most important of which are opposites (antikeimena). These, however, are  not the intension of being, either in full or in part, for the intensional nature of being is always a physis. And antikeimena neither are nor constitute a physis. Rather,  they presuppose it, for they are derived from one and many, which are pros hen focused on each individual substantial entity's own primary physis. 2 Aristotle's language in the Metaphysics is vividly extensional when he characterizes the distinctive subject matter of each epistêmê. Philosophy investigates panta ta  onta, neither generically nor categorially restricted. Each special science cuts off from panta ta onta either a generically or a categorially restricted part (meros) and  hence is a partial epistêmê (en merei): "[n]ow this is not the same as any of the so­called special sciences; for none of these others treats universally of being as being.  They cut off a part of being and investigate the attribute of this part; this is what the mathematical sciences for instance do" (Met. IV 1003a22–6; cf. IV 1004a2–6).  And his language is just as vividly intensional when he characterizes the distinctive qua of each epistêmê. Philosophy investigates qua being, i.e., qua the properties  that belong to being in its own physis. Each partial epistêmê investigates qua either a generically or categorially restricted part of that physis. So zoology investigates  the denizens of the animal kingdom qua their shared generic nature, and mathematics investigates number qua poson (the category of quantity). The hê, the feminine  dative of the relative pronoun that we translate with qua or its cognates, is a precise singling­out expression. If being were not plural in its intension as well as in its  extension, the hê would be vacuous because it would have no work to do.3 It is clear from this that Aristotle's synonymous use of qua on and qua ousia means that qua ousia is not categorially restricted. For if it were, philosophy would be  but another epistêmê en merei. Qua ousia is qua on because ousia is not exclusive but rather inclusive as pros hen focus. It does not absorb all being into itself but  only governs it as formal­final causal archê. To investigate being qua ousia is to investigate it qua on. Conversely, to investigate it qua anything else is not to  investigate it qua being: "all these sciences mark off some particular being—some genus, and inquire into this, but not into being simply nor qua being" (Met. VI  1025b7–10; cf. XI 1061b19–33). From this his denial of the status of epistêmê to dialectic follows readily enough. He denies it that status, not because dialectic does not have a distinctive  methodological qua of its own but because it has the wrong one. It does not have the ability to attain truth or knowledge because it does not investigate being as it is,  i.e., qua on or qua ousia, as governed by the primary physis of each individual substantial entity. It investigates panta ta onta kath hen, not pros hen. But being

  

Page 4

always is pros hen. For the being of an individual substantial entity always is kath hauto, i.e., on its own terms, in terms of its own primary physis. Therefore it can be  understood only kath hauto, on its own terms, in terms of its own physis. To use a double spatial metaphor, the being of a being must always be understood from the  inside out and from the top down, that inside and that top being its own primary physis. 4 It is a good argument, but like many such it proves too much. For if only the investigation of being qua on or qua ousia is an epistêmê because only it understands  being as it is, then only philosophy can have that status. Only it investigates being qua on or qua ousia, i.e., pros hen. All other human cognitive endeavours  investigate it qua something else and so kath hen rather than pros hen. By parity of reasoning, the denial of the status of epistêmê to dialectic implies a like denial to  all the special sciences. Aristotle shows some awareness of this implication in the text of the Metaphysics, for it creates a rather embarrassing problem for him. On the one hand he is  concerned to argue for the possibility of philosophy as a general metaphysical epistêmê, and the only way he can secure this pressing desideratum is by linking qua on  with qua ousia and with the pros hen unity of being: As, then, there is one science which deals with all healthy things, the same applies in the other cases also. For not only in the case of things which have one common notion does  the investigation belong to one science, but also in the case of things which are related to one common nature; for even these in a sense have one common notion. It is clear then  that it is the work of one science to study the things that are, qua being. (Met. IV 1003 b 11–6)

On the other hand, to declare that such recognized and established disciplines as zoology and mathematics are not epistêmai because they do not understand being as  it is, is to court ridicule and to deprive himself of the rich variety of epistemic investigations which constitute the bulk of his work. So we find him more than a little  defensive on this point, arguing especially in the case of mathematics that though it investigates number qua poson and not qua ousia, kath hen and not pros hen, it is  nonetheless a genuine epistêmê. One cannot help but ask why he is not equally defensive on behalf of dialectic. Dialectic errs in not recognizing the priority of ousia to  the opposites, but mathematics does not err in not recognizing the priority of ousia to poson: "[a]nd those who study these properties err not by leaving the sphere of  philosophy, but by forgetting that substance, of which they have no correct idea, is prior to these other things" (Met. IV 1004b8–10). Again, Aristotle says, "but the  mathematical sciences, i.e., arithmetic and geometry, treat of no substance" (Met. XII 1073b6–8). He adds: Therefore if we suppose attributes separated from their fellow­attributes and make any inquiry concerning them as such, we shall not for this reason be in error.... Each question  will be best investigated in this way—by setting up by an act of separation what is not separate—as the arithmetician and the geometer do. (Met. XIII 1078a17–23)

  

Page 5

There are undoubtedly some historical reasons for this differential treatment. After all, dialectic was much the older and more formidable rival for the status of general  metaphysical epistêmê. Aristotle notes with obvious disapproval that all thinkers made the opposites prior in their investigation of being. He refers to friends closer to  home who exalted dialectic as the near­divine method of philosophy (Met. XIV 1087a29–b4; XII 1069a26–8; XIII 1084b23–5). He defends the superiority of his  own metaphysical epistêmê by a neat reversal of perspective: dialectic makes the opposites prior to ousia, metaphysical epistêmê makes ousia prior to the opposites:  "for there was as yet none of the dialectical power which enables people even without knowledge of the essence to speculate about contraries and inquire whether the  same science deals with contraries" (Met. XIII 1078625–7; cf. III 995b 20–5). For the attributes of this in so far as it is being, and the contrarieties in it qua being, it is the business of no other science than philosophy to investigate ... while dialectic and  sophistic deal with the attributes of things that are, but not of things qua being, and not with being itself in so far as it is being. Therefore it remains that it is the philosopher who  studies the things we have named, in so far as they are being. (Met. XI 1061b4–11)

By contrast, he can dismiss the attempt of some Platonists to promote mathematics to the status of general metaphysical epistêmê by caging it in the secondary  category of poson. There mathematics can stay and enjoy itself harmlessly, for secondary being cannot be prior to primary being (Met. VII 1038623–9;  XIV1088b2–4). Being is pros hen focused on ousia, not on poson. But while such historical reasons may make the differential treatment understandable, they do not  justify it. The position that the neglect of the priority of ousia is an error in the case of dialectic but not in the case of mathematics must be defended by systematic  arguments if it is to be more than special pleading. Aristotle does offer some systematic arguments, and some others can be inferred from positions he takes. For the question of error must be assessed in relation to the  epistemic claim being made. Mathematics (barring some over­enthusiastic Platonists) lays claim only to the status of a partial categorially restricted epistêmê, while  dialectic lays claim to the status of an unrestrictedly general metaphysical epistêmê. Aristotle holds that disregarding the priority of ousia within a categorially or  generically restricted domain of being is not an error, while disregarding it within the unrestricted domain of being is an error. Why does he hold this? He does so  because the former has an adequate basis in the being of things while the latter does not. For the former presupposes the categories of being while the latter denies  them. A categorially or generically restricted domain of being always lies within a category, but dialectic sees its opposites as univocally operative across all categories.  Thereby the former does not deny the transcategorial pros hen unity of being, while the latter does. Mathematics investigates being kath hen only within the category  of quantity, but dialectic investigates it

  

Page 6

kath hen across all categories. Therefore the mathematical kath hen is compatible with the pros hen priority of ousia, while the dialectical kath hen is not. 5 Still, this argument will not carry him all the way. It shows the incompatibility of the dialectical kath hen with the pros hen priority of ousia, but it does not yet show  the compatibility of the mathematical kath hen with the pros hen priority of ousia. For the latter has to be proved within the category of poson itself. In that category,  as in every other, being is pros hen focused on ousia. It has to be shown that investigating number qua poson is compatible with its being qua ousia, for qua ousia is  the same as qua on. And therefore qua poson is not qua on and so does not investigate being as it is. While investigating number qua poson implies accepting the  categorial restriction and therefore the categories, mathematics still abstracts from the priority of ousia. Why does this abstraction not amount to the kind of disregard  of the priority of ousia that constitutes an error? The answer lies in showing that number is both qua poson and qua ousia. For poson is a distinct even if secondary categorial mode of being. As distinct it is qua  poson and as secondary it is qua ousia. What needs to be shown is that its distinctness is compatible with its secondary status. This, I believe, Aristotle can do, and  so he can vindicate the claim of mathematics to the status of epistêmê. As a distinct categorial mode of being, poson must have some intrinsic characteristics, some characteristics qua poson. For otherwise its distinctness would be lost, it  would be indistinguishable from the other categorial modes of being and so could not stand in a distinctive pros hen relationship with ousia. Such characteristics must  attach intrinsically to this mode of being. They are set out in Chapter 6 of the Categories: discreteness or continuity, having no contrary, not admitting of a more and a  less, being equal and unequal rather than similar and dissimilar (Cf. Met. IV 1004b10–5; XIII 1078a24–8; XII 1077b22–33). They must be preserved even in the  pros hen priority of ousia, for this distinct categorial mode of being must be pros hen focused on ousia as what it is, qua poson, not qua poion or qua any other  categorial mode. Ousia must govern all the distinct secondary categorial modes of its being, but it must not destroy them. Qua poson must be compatible with qua  ousia. Therefore the abstraction of qua poson from qua ousia does not constitute an error, and mathematics can be accepted as an epistêmê. But the truth or knowledge it  attains is only partial. The full truth or knowledge of being can be attained only qua ousia, which incorporates qua poson both in its categorial distinctness and in its  pros hen focus on ousia, i.e., both as a distinct and as a secondary categorial mode of being. Aristotle is prepared to accept partial truth or knowledge as genuine  epistêmê. But he calls it exactly what it is: partial epistêmê, epistêmê en merei. Partial truth is obtained by abstraction, and mathematics abstracts the distinctness of  the category of poson from its secondary status. It still is truth because poson really is a distinctive categorial mode of being. Partial truth can only be abstracted from  complete truth, hence qua poson only from qua ousia = qua on. But each secondary categorial mode of being has its own (even if

  

Page 7

secondary) ti esti in terms of which it can be known and so yield epistêmê (Cf. Met. XI.3, 4; VII 1028a36­b2). The differential treatment of the epistemic claims of dialectic and of mathematics is no mere case of special pleading. Mathematics is a partial epistêmê because its  disregard of the priority of ousia has an adequate basis in the being of things, while dialectic is not an epistêmê because its disregard of the priority of ousia has no  such basis. For being in unrestricted generality is not qua opposites at all. Opposites are excluded from the categories of ousia and poson, and they function most  properly as the akra or extremes of alloiosis, qualitative change. But even there, they are not prior but presuppose a hypokeimenon, so that they cannot be archai.  While disregarding the priority of ousia is foundational for mathematics, it is fatal for dialectic. Still, Aristotle does not deny dialectic any cognitional value. It cannot be gnostikê, but it is peirastikê. It cannot be an epistêmê because it does not know, but it can  be a technê because it puts questions. Moreover, it puts questions about the same matters which philosophy knows, i.e., about being in its full extension, about panta  ta onta. Nor is its questioning idle or unprofitable. In fact, dialectical questioning is an indispensable preliminary to metaphysical epistêmê. Aristotle accords it this  recognition not only theoretically but also practically by making use of dialectical questioning in the Metaphysics. For holding that the human quest for wisdom arises  out of wonder, which takes the form of asking "why?" he realizes that the ability of metaphysical epistêmê to find answers depends on the correct framing of the  questions. An aporia cannot be resolved if it has not been set out properly (Met. III 995a24–8; 995b2–4). But here a problem arises. For the correct framing of a question to which metaphysical epistêmê can find an answer, the correct setting out of an aporia, which  metaphysical epistêmê can resolve, would seem to presuppose continuity between question and answer. It would seem to presuppose that both look at their common  subject matter from the vantage point of the same methodological qua. A question asked qua opposites cannot be answered qua ousia. Dialectic disregards the  priority of ousia even in its questions, not only in its attempted answers. Its questioning is itself in error. But how could questioning that is in error be helpful or even  indispensable to finding answers that are not in error? Clearly, only by leading to the recognition of that error. For only that can lead to the correct reformulation of the question. Metaphysical epistêmê is gnostikê only  because it is correctly peirastikê. Dialectical technê is not gnostikê because it is not correctly peirastikê. The most significant achievement of philosophy does not lie  in the answers it finds but in its reformulation of the metaphysical question after the being of all beings. The most famous and comprehensive of these Aristotelian  reformulations involves the priority of ousia, the recognition that the question "what is being?" is primarily the question "what is substance?" (Met. VII. 1).

  

Page 8

Toward this achievement dialectic gives it invaluable help. It does so by demonstrating the wrongness of its own methodological qua by reducing it to absurdity  (atopon, adynaton). Investigating panta ta onta qua opposites leads to such (to Aristotle) obvious absurdities as denying that opposites cannot act directly on one  another but presuppose a hypokeimenon: All philosophers make the first principles contraries: as in natural things, so also in the case of unchangeable substances. But since there cannot be anything prior to the first  principle of all things, the principle cannot be the principle and yet be an attribute of something else. To suggest this is like saying that the white is a first principle not qua  anything else but qua white, but yet that it is predicable of a subject; i.e., that its being white presupposes its being something else; this is absurd, for then the subject will be  prior. But all things which are generated from their contraries involve an underlying subject; a subject, then, must be present in the case of contraries, if anywhere. All contraries,  then, are always predicable of a subject, and none can be apart, but just as appearances suggest that there is nothing contrary to substance, argument confirms this. No contrary,  then, is the first principle of all things in the full sense; the first principle is something different. (Met. XIV 1087a29­b4; cf. XI 1059a22–3; XII 1075b21–2)

Generally, it leads to unbridgeable oppositions since opposites cannot both be true at the same time in full actuality. If being is both one and many in full actuality, if it is  both individual and general in full actuality, if being is actually individual while knowledge is actually universal, etc., epistêmê cannot be attained. The proper  metaphysical reformulation depends on recognizing the priority of ousia and on understanding it as the priority of actuality to potentiality: ''For the same thing can be  potentially at the same time two contraries, but it cannot actually" (Met. IV 1009a34–6). Dialectic is a valuable preliminary peirastikê technê because it debunks its own metaphysical pretensions when it is allowed to frame its questions qua opposites.  There is a good deal of interesting scholarly debate as to the extent to which the extant text of the Metaphysics is aporetic. I offer the thought that perhaps what we  call aporetic has the dual function of debunking the wrong framing of the metaphysical question and of leading to the correct framing. 6 If this thought has merit, then Aristotle's engagement with dialectic can be seen to be subtle indeed. Defeating one's main rival by allowing him to destroy himself is a  Machiavellian tactic in metaphysics no less than in politics. Silencing dialectic would have been much less effective than allowing it to speak. Suppressing any voice in a  scholarly debate or in the conversation of mankind is not only always wrong, it is also always foolish. For if Socrates is right that the unexamined life is not worth living  for a human being, then the examination offered by rivals is an inestimable service. Aristotle avails himself of that service fully. His metaphysical leitmotif: "Being is truly  spoken in many senses because it is in many ways" is validated by the shipwreck of its rival. Panta ta onta must be understood pros hen because they cannot be  understood kath hen.

  

Page 9

However, such validation is only preliminary and negative. Even if it is shown that the investigation of panta ta onta kath hen cannot be metaphysical epistêmê, it  does not follow that Aristotle's own pros hen is. The true validation of philosophy's claim to the status of metaphysical epistêmê must be positive. It must lie in  illumining the being of panta ta onta qua ousia. Aristotle, I believe, can do this. For he can show how one and many are reconciled in the being of each individual  substantial entity because its pros hen unity is delicate, intensional, and so not destructive of its inner multiplicity. Instead, it takes its many­hued richness from that  multiplicity and makes it cosmic under the imprint of its own primary physis. 7  His arguments for the status of mathematics as a partial epistêmê are important because  they are an instance of this reconciliation of one and many, namely, of qua ousia and qua poson. He denies the status of epistêmê to dialectic above all because it  takes one and many to be opposites that are both actual and so forever irreconcilable. But according to Aristotle, the being of each being is at peace within itself, not  at war. For one is actuality while many is potentiality, and the actual and the potential are correlatives, not opposites. They not only can but must both be true at the  same time in the being of each complex individual substantial entity. The priority of ousia, which the philosopher understands while the dialectician does not, lies in this correlation of actuality and potentiality as modes of being. An  individual substantial entity's actuality is the pervasive unifying constitutive formal­final causality of its primary physis as determinant, for which all other aspects of its  being are potentiality as determinables. Their status as determinables is linguistically indicated by paronymy, i.e., by adjectival forms that are derived from nouns (Cat.  I; Met. IX.7). The only exception is the generic nature which, though a determinable as intelligible matter, is stated as a noun because it is part of a thing's essence and  definition. But both the physical matter and the properties in the secondary categories proclaim their metaphysical status as determinables by paronymy. It is this  parallel between the physical matter and the properties in the secondary categories which holds the key to Aristotle's reconciliation of one and many, of qua ousia and  qua poson. Aristotle draws explicit attention to this parallel, even coining a new technical term, ekeininon, and he explains that both an individual substantial entity's physical  matter and its properties in the secondary categories are potential or determinable because they are indeterminate (aoriston, apeiron). As he puts it, "[a]nd it is only  right that 'thaten' should be said with reference both to the matter and to the accidents, for both are indeterminates" (Met. IX 1049a36­b2). Again, he says, "for it is  that which is potentially and not in complete reality that is indeterminate" (Met. IV 1007b28–9). As relatively, not absolutely, indeterminate they are determinable to  each individual substantial entity's own being because they are actualized by the entity's primary physis as determinant. Only as so determined can they actually be, for  only determinate

  

Page 10

being is actual and "it is not possible for anything indeterminate to be" (Met. II 994b26–7; my translation). They are receptive of the imprint of the entity's primary  physis. Aristotle does not explicitly call the accidents dektika, but the above parallel implies it. Metaphysically, any property in a secondary category can actually be  only as actualized and determined to the being of the individual substantial entity whose property it is. That is what it means for a category to be a secondary mode of  the being of a substance. Its pros hen reference to ousia is not detachable in reality nor in metaphysical understanding, for qua ousia equals qua on entelecheia. 8 But it is detached for the category of poson in mathematical abstraction, in fact, this detachment is what is meant by abstraction. By virtue of it, quantitative properties  acquire actuality qua poson, not in the order of being but in the order of knowledge. And along with it, they acquire mathematical akribeia. Both philosophy and  mathematics are epistêmai because both the metaphysical and the cognitional actualization of poson have an adequate basis in being, namely, in the doubly potential  and determinable being of quantitative properties. If potentiality were not a mode of being, Aristotle could not reconcile qua ousia and qua poson nor give both  epistemic status: "[t]hus, then, geometers speak correctly; they talk about things that are, and their subjects are; for being has two forms—it is not only in complete  reality but also in the manner of matter" (Met. XIII 1078a28–31, my translation. Cf. XIII 1084b9–13).9 Conclusion What does it mean concretely to understand individual substantial entities both qua ousia and qua poson? More pointedly, what does it mean to understand the same  quantitative property both qua ousia and qua poson? Take the 100­pound weight of two different individual substantial entities, say, a pumpkin and a pony and  assume each to be a normal specimen of its kind. Each weighs 100 pounds. Qua poson the two weights are precisely equal, ison. For qua poson they are understood  without regard to the substances whose weights they are and so are comparable with mathematical akribeia. But qua ousia they are not equal at all for they are  understood as the weight of the pumpkin and of the pony, respectively. A 100­pound pumpkin is a champion pumpkin, unusually large and healthy and sure to win a  citation at the county fair. A 100­pound pony is an emaciated near­skeleton, feeble and ill and near death, which is apt to win a citation from the police or the SPCA.  The 100­pound weights are comparable with metaphysical determinacy qua their significance for and their contribution to the substantial being of the pumpkin and of  the pony. It is good for the pumpkin to weigh 100 pounds, but it is not at all good for the pony to do so. Mathematical abstraction abstracts from this intrinsic metaphysical significance. The being of an individual substantial entity is intrinsically

  

Page 11

significant in all its aspects in all categories, for all make their own unique contribution to its being. That is why its formal cause coincides with its final cause and why its  primary physis imprints itself upon all aspects of its being by way of formal­final causality. In not considering the good, mathematics also does not consider the entity's  formal cause. It purchases akribeia at the cost of significance. For 100 pounds of weight, abstractly taken qua poson, have no metaphysical significance. The 100  pounds are not aoriston qua poson, for qua poson they are perfectly determinate. They are paronymous in their being and so aoriston and determinable only qua  ousia, qua the role they play in the individual substantial entity's total being. The metaphysical understanding of opposites, like the metaphysical understanding of poson, recognizes the priority of ousia and so does not commit the error of  dialectic. I suggest that it sees opposites as a further parallel to the status of matter and accidents as determinables, though the text does not explicitly say so. Qua  ousia, opposites too are determinable to the being of each individual substantial entity. For example, the akra of alloiosis are opposite poles of qualitative change  which have a determinate range characteristic of and determined by the ousia whose properties they are. Only as so determined do they have metaphysical  significance. A vivid example of such significance are the akra or opposite poles of the range of temperature variation characteristic of and determined by the human  organism. Changes within a very narrow range mean that the organism is warmer or colder, while changes beyond that range mean that it is dead. 10 Opposites  understood qua ousia are part of metaphysical epistêmê, but they pertain only to some categories, not to all, and they play a different role in different categories. They  cannot be abstracted from the individual substantial entities in general or across all categories, for they have no being in general or across all categories. Therefore  dialectic, which so abstracts them, is not an epistêmê. Special Notes Textual references are exclusively to the Metaphysics in order to develop a picture of how dialectic is understood in that work. For a fuller picture other texts will of  course need to be considered. The Oxford translation has been used unless otherwise indicated, but 'exist' has been changed to 'be' throughout. The thesis of this chapter is based on taking pros hen equivocity as the intensionally unifying feature of a syntheton's entire being in all categories. The author is  indebted to Joseph Owens for having recognized the central importance of pros hen equivocity in Aristotle's Metaphysics. However, her interpretation of pros hen  differs fundamentally from his. His thesis in The Doctrine of Being in the Aristotelian Metaphysics (Toronto, 3rd rev. ed. 1978) is based on extending pros hen to  the relationship between syntheta and the Prime Mover, regarding the

  

Page 12

former as secondary and the latter as the primary instance of being. The primary physis of a syntheton is then not the cause of its being, while the Prime Mover is:  "Correspondingly all sensible things have their own natures, but they are beings through reference to one nature that is primary.... Each of the secondary instances has  its own nature, which is not the nature of being ... the nature of being ... is found only in the primary instance" (37). The present chapter's thesis, by contrast, sees a  syntheton's primary physis as the cause of its being. Notes 1. epistêmê attains truth through causal understanding: "It is right also that philosophy should be called knowledge of the truth. For the end of theoretical knowledge is  truth.... Now we do not know a truth without its cause ... as each thing is in respect of being, so it is in respect of truth" (Met. II 993b19–31). 2. "Further, with regard to the same and other and like and unlike and contrariety, and with regard to prior and posterior and all other such terms about which the  dialecticians try to inquire, starting their investigations from probable premises only" (Met. III 995b22–4). "To the one belong, as we indicated graphically in our  distinction of the contraries, the same and the like and the equal, and to plurality belong the other and the unlike and the unequal" (Met. X 1054a29–32). "Now most  things are called one because they either do or have or suffer or are related to something else that is one, but the things that are primarily called one are those whose  substance is one" (Met. V 1016b6–9).       J. D. G. Evans, Aristotle's Concept of Dialectic (Cambridge: Cambridge University Press, 1977), 5, agrees that dialectic studies panta ta onta, but not that it  studies them qua opposites or indeed qua anything. Instead he says, "[D]ialectic must be distinguished from the sciences in that it does not work with any set view of  reality ... should not embody any view of reality—neither a correct one ... nor an incorrect one." It does not seem historically likely that Aristotle or any ancient Greek  thinker could have conceived of such an "ontologically neutral" technê.       Robert Bolton's "Aristotle's Conception of Metaphysics as a Science," in Unity, Identity and Explanation in Aristotle's Metaphysics, T. Scaltsas, D. Charles,  M. L. Gill, eds. (Oxford: Clarendon Press, 1994), does not interpret the distinction peirastikê­gnostikê (Met. IV 1004b25–6) as denying dialectic the status of  metaphysical epistêmê. Interpreting Met. IV in the light of the Sophistical Refutations, he holds that peirastic dialectic demonstratively proceeds from true premises  to true conclusions. It knows the same things as philosophy, but by a different method.       This notion is difficult to square with Aristotle's negative comments about dialectic in the Metaphysics, and it is difficult to accept that Aristotle could have  understood metaphysics to have two different but equally truth­yielding methods. 3. There is no agreement about the qua­locution in Aristotle. I understand the first on in the locution on qua on extensionally, as referring collectively to all the beings  that there are. I understand the second on intensionally, as referring to each being's own constitutive physis. The qua then grounds an entitative on (beings) in a  nonentitative on

  

Page 13

      (causally constitutive archai) (Cf. Met. IV.1; 1004b15–7; 1005a13–8). Each special science correspondingly investigates a part of the first on qua a part of the  second on: "Hence to investigate all the species of being qua being is the work of a science which is generically one, and to investigate the several species is the work  of the specific parts of the science" (Met. IV 1003b21–2). "Many properties attach to things in their own nature as possessed of each such character, e.g. there are  attributes peculiar to the animal qua female or qua male (yet there is no 'female' nor 'male' separate from animals); so that there are also attributes which belong to  things merely as lengths or as planes" (Met. XII 1078a5–9; cf. IV 1004b10–7; XI 1065b23–8).       J. Cleary, Aristotle on the Many Senses of Priority (Carbondale: Southern Illinois University Press, 1988), 74–5, denies that an Aristotelian epistêmê has a  distinctive method: "... Aristotle's tendency to differentiate the sciences according to their characteristic objects rather than by their respective methods." 4. I use the phrase "from the inside out" to express an individual substantial entity's immanently constitutive self­governance and self­closure. I borrow the phrase "from  the top down" from F. Sparshott, "Aristotle's World and Mine," in Aristotle Today, M. Matthen, ed. (Edmonton: Academic Printing and Publishing, 1986). Cf. F. A.  Lewis, "Aristotle on the Relation between a Thing and its Matter,'' in Unity, Identity, and Explanation in Aristotle's Metaphysics. (Cf. Met. IV.2; V. 18; VII. 1,  2). 5. Dialectic's kath hen use of general metaphysical terms such as 'one' and 'many' and the opposites derived from them is contrasted with philosophy's pros hen use  in Met. IV.2. Mathematics uses its terms kath hen since it is a demonstrative epistêmê, so that any equivocity in its terms (even pros hen equivocity) would invalidate  its arguments. But its kath hen use is restricted to the category of poson: "But number qua number differs in quantity" (Met. XIII 1083a4). 6. Alan Code, "The Aporematic Approach to Primary Being in Metaphysics Z," CJP Supp. 10 (1984): 41–65, has offered a sustained argument for regarding book  Z as aporematic and dialectic as the method of philosophy. John Cleary. "Working Through Puzzles with Aristotle," The Journal of Neoplatonic Studies 1, no. 2  (1993): 83–179, holds that Aristotle assumes his own doctrinal standpoint in the Metaphysics, and Terence Irwin, Aristotle's First Principles (Oxford: Clarendon  Press, 1988), 274, also does not consider it "purely or primarily aporetic."       I disagree with both David Evans (see n. 2 above) and John Cleary that the dialectical framing of metaphysical questions is, or can be, from no doctrinal  standpoint or "metaphysically neutral." Even endoxa must be formulated from some standpoint, from some determinate qua, and so must questions. The proper  reformulation of the metaphysical question after the being of all beings necessarily involves Aristotle's own standpoint, so that he could not achieve such reformulation  (e.g., VII.1) as a "neutral referee."       Terence Irwin's interpretation of Aristotle's metaphysical method as 'strong dialectic' is admittedly Kantian and so, while interesting, almost certainly too  anachronistic to fit the text: "Aristotle asks the Kantian question, 'How is knowledge possible?', by asking how things must be if there is scientific knowledge of them ...  the parallels are suggestive, and the differences may be less deep than they seem" (170). 7. This intensional unification of an individual substantial entity's total being, its hen to pan, depends on its qualitative primary physis: "But essence depends on quality,

  

Page 14

      and this is of determinate nature" (Met. XI 1063a27–8). This physis causally determines all determinable aspects so that the entity is one and whole and "so  contains the things it contains that they form a determinate unity" (Met. V 1023b27–28, my translation). This is a condition of the entity's substantial being.       T. Scaltsas, "Substantial Holism," in Unity, Identity and Explanation in Aristotle's Metaphysics, expresses this intensional pros hen determination of  determinable aspects as "reidentification." In the same volume, M. Ferejohn, "The Definition of Generated Composites in Aristotle's Metaphysics," gives an excellent  account of the paronymous status of the material cause. 8. Charlotte Witt, Substance and Essence in Aristotle (Ithaca: Cornell University Press, 1989), chs. 2 and 5, gives a fine discussion of this parallel in the  paronymous being of matter and accidents.       M. L. Gill, Aristotle on Substance (Princeton: Princeton University Press, 1989), ch. 5, sees the paronymous being of both matter and accidents as necessary for  a substance's unity: "Aristotle ... compares the matter itself to nonsubstantial properties. Thus the matter is indefinite ... because it is itself a sort of property that, like  nonsubstantial properties, depends for its definiteness on some actual subject to which it belongs" (154). She contrasts this successful metaphysical account of  substantial unity in book IX with the unsuccessful one of book VII. 9. Potentiality as the basis of mathematical objects in the being of things is recognized by John Cleary, "Working Through Puzzles with Aristotle," 127–8: ... any genuine science must have a real or existent object. Since Aristotle shares that assumption, he would probably find it unthinkable that the objects of mathematics should  not exist in some way because that would leave the paradigmatic Greek sciences without foundation ... the dispute will not be about their being (peri tou ontos) but rather about  their mode of being (peri tou tropou).

      Like me, he understands abstraction as the conceptual detaching of accidents from the substance whose properties they are: "Conversely, 'abstraction' would be  the process of taking away the subject and defining white separately ... priority in substance is being denied to the so­called 'result of abstraction'" (152. Cf. 164–72).       A clear account and summary of recent literature on Aristotle's mathematical ontology is given by Edward Halper, Form and Reason (Albany: SUNY Press,  1993), ch. 7. "In philosophy of mathematics Aristotle provides a common sense mixture of moderate realism and constructivism" (153). 10. M. Husain, "The Multiplicity in Unity of Being qua Being in Aristotle's pros hen Equivocity," New Scholasticism, vol. LV, 2 (1981): 208–218.

  

Page 15

Chapter Two—  The Diodorean Modalities and the Master Argument Lenn E. Goodman Since its inception as a science, logic has been called upon continually to serve as a handmaiden to metaphysics. Nor is this surprising. The mind's search for purchase  on the infinitude of the categorical will naturally be arrested by the striking capabilities for universality afforded by systematic abstraction—as though here surely lay the  hidden power of words to govern reality.* One might almost believe that some lingering faith in the magic of words lay at the base of the enduring preoccupation with  language (as an externalization of thought) that characterizes so much of Western philosophy. One might even suspect that certain lightnesses in the harvest might  represent a sort of natural retribution for the superstitious distraction of the mind by externalities of form from its true object, reality itself. For positive as well as  negative sentence regarding metaphysics has been sought (and avowedly found) in the abstract vastnesses of logic. The ease with which incoherent notions come  acrupper when tried for consistency has lent credence to the metaphysician's dream that somewhere (beyond the philosophical blue heaven and red earth) lies a truth  that complements in its rationality and *

 I fondly remember A. N. Prior's generosity in discussing the Diodorean modalities with me some thirty years ago and dedicate this chapter in his memory. Warm thanks to my  colleagues Jeffrey Tlumak and Henry Teloh and to Allan Bäck of Kutztown University and May Sim, the editor of the present volume, for their helpful suggestions.

  

Page 16

coherence the irrationality and incoherence of the inconsistent, without descending into triviality. The Parmenidean esti is clearly the eldest offspring of that philosophical dream, but not much younger are the splendid truths promised by Plato's anamnesis. Despite  the glittering promise of such rewards, neglect of logical neutrality can lead to but one result: abrupt and often premature entry into metaphysics. The assumption that  logic affords sufficient guidance to metaphysics is predicated on the metaphysically problematic presumption that there is just one logically possible world. Diodorus Cronus (ca. 350 B. C.–ca. 284 B. C.) was a contemporary of Ptolemy Soter, at whose court he lived in Alexandria. He was so celebrated in logic that the  sobriquet 'logician' became a kind of surname for him, which was passed on, along with his profession, to his five daughters. His celebrated efforts to reduce the  modality of the possible to the terms made available by the modality of the actual, was, I will argue, an attempt to bring the authority of logic to the support of the a  problematic metaphysical notion, the Megarian rejection of potentiality. 1 The belief that only those events which do happen can happen may seem to be of the essence in any serious determinism. Yet determinism alone does not call for the  elimination of all notions of potentiality. At least in a curtailed, monovalent form, potentiality may survive even in a radically deterministic universe.2  Indeed, it must. For  any determinism trades on the credibility of the idea of necessity, and that in turn requires some notion of possibility, either as its dialectical backstop or as matter out  of which the stronger stuff of necessity is distilled. But 'real' potentiality, a modality that allows the possibility of what will never be, is incompatible with the kind of  global determinism that many exponents of determinism have sought in arguing that what does not happen cannot happen, and what does occur must occur, cannot be  otherwise, and cannot ever have been such that it could have been otherwise.3 Even the most comprehensive determinism need not insist that the ideas of real possibility and contingency be deemed logically incoherent. Determinism may be  deemed simply "the way of the world," without acquiring or demanding the authority of logic. But determinism itself can become a stake in a larger game. For the  Parmenidean esti, the first attempt by a philosopher to draw an account of being from an examination of language, did not merely colonize logic but founded it, in the  service of monism. To the Megarian heirs of Parmenides and his bold attempt to deduce monism from the seeming self­inconsistency of 'is not',4  real potentiality  clearly seemed to threaten not just determinism but monism as well. For time as well as space affords a dimension to multiplicity—since change bespeaks diversity, and no stronger evidence of difference can be found than the  procession of variance that manifests itself in time. It was the phenomenal irreducibility of temporality, perhaps, that kept Parmenides from reconciling the Way of  Truth with the Way of Seeming. For even the illusion of change remains a real change. To the Megarian heirs of the Eleatics, for this reason, potentiality

  

Page 17

became a critical target. For potentiality was the linchpin of the analysis of change that Plato had suggested 5  and Aristotle had developed in response to the monists'  critique of temporality. To demonstrate the incoherence of Aristotelian potentiality would undermine Aristotle's defense of time and vindicate the Parmenidean claim  that change and time itself are mired in contradiction. The Megarian attack on potentiality is grounded intuitively in the authentically Parmenidean insight that only the actual is real. The Aristotelian attempt to distinguish  potentiality from actuality seems to the Megarian no more than ontological trifling—an effort to ignore the core law of being (and logic), that a thing either is or is not.  What neither is nor is not but 'can be' seems a suspect category to the ontological seriousness of the Megarians.6  Had not Zeno shown that it is impossible for the  arrow to move from bow to bull's­eye, when at every instant it has a definite position, and so is always, obviously, at rest? The paradox here was a graphic realization  of the Parmenidean demand for determinacy. For (at every location) either the arrow was here or it was not here. How then, without equivocation, could it move? The  Megarian denial of potentiality was a parallel return to the demand for determinacy,7  and the critique of potentiality was not just a dialectical ploy but a committed  recognition that here too, as in any account of motion or change, difference means equivocation, the affirmation of that which is not. Aristotle's appeals to experience, his claims that processes of learning, or manufacture, in which 'dispositions' are acquired that are incomprehensible without the notion  of potentiality (Metaphysics Theta 3, 1046b32ff.), will of course appear to Megarians to be arguing in a circle. To rest potentiality on the idea of process, when  process itself is the target of the attack is egregiously to beg the question.8 Aristotle's attempt (1047a4ff.) to reduce Megarians to "the doctrine of Protagoras," i.e., to subjectivism, is telling. Without dispositional properties, Aristotle urges,  "nothing would be cold or hot, sweet or perceptible at all, unless people were perceiving it." But Megarians are far from being subjectivists. Their aim is a critique of  appearances, from which, presumably, they may turn to the investigation of reality and the discovery of an ontological plane or focus that is more responsive than  phenomena to the epistemological tests of monist, or realist/logicist rigor. In shoring up his own ontology, the Megarian might reason, Aristotle is free by all means to appeal to experience. But how can he appeal (without begging the  question) to ontological phantoms and bridge passages like dispositions and abiding capacities that are experientially unknown: the visibility of objects unseen, the  breakability of objects that will never be broken. One need not admit that a man was deaf when he was not hearing in order to claim that it is otiose to assign him a  hearing ability that he does not ever (at such times) use. Aristotle may justly claim that unless we posit such an ability, the deaf man and the unimpaired man who is

  

Page 18

not hearing at the moment are indistinguishable. But the Megarian cheerfully admits this. Aristotle might appeal to the need to support a language and conceptual scheme that make possible our practice of science and our reference to and reliance upon the  continuities of nature. Similarly one might say that the warnings we issue, e.g., that certain objects are fragile, would be senseless on the Megarian account, in those  cases where an object is never to be broken. But such cautions as we use in ordinary discourse, the Megarian would answer, serve only as admissions of our  ignorance. And all appeals to science, nature, or pragmatics assume a prior commitment to the phenomena of change, which, I think, is the problem at issue.  Regardless how far­fetched Megarianism may seem against the backdrop of ordinary language and experience—let alone science, or the writing of insurance  policies—it rests not in such language, practice, and experience but in a metaphysical commitment that purports to offer a vantage point superior to that afforded by  Aristotle's naturalism and empiricism. Aristotle knows that what he is rejecting are the implications of firmly held premises radically different from his own. 9  He can plead for premises by demonstrating  their effectiveness in marshaling certain classes of experience; and he can, or demonstrate the unwelcome consequences of his opponents' view, but he cannot prevent  his adversaries from accepting the consequences of their position, as corollaries of their fundamental premise. Thus his arguments "against the Megarians" are not really  directed at them at all, but at the undecided or indecisive inquirer who may not have grasped the full cost in implications of the Megarian view. They are critical  examples of the practice of dialectic, critical in that they seek for common ground with hearers not­yet committed or converted to Megarianism. In the celebrated discussion of the "Sea Battle (De Interpretatione 9)," then, when Aristotle presents a familiar type of situation, in hopes that a wholesome intuitive  response will point up our common recognition of real possibility, he takes for granted precisely what Megarians would find problematic—Just as he finds problematic  what they take for granted. He sees that future contingency requires a nonce suspension of the demand for a determinate truth value for all sentences. Where reality is  not yet determinate, neither are the truth values of the corresponding judgments.10 But to the Megarians Aristotle seemed to be fudging on the law of the excluded middle.11 All judgments, they argued, must be either true or false—why exempt those  about the future? Aristotle's answer, of course, was that failure to suspend the determinacy of the truth values of sentences about future events leads to a variety of  determinism that would, in effect, obviate the role of natural or voluntary causes.12 But this outcome would hardly faze the Megarians. By drafting his own Megarian definitions of the crucial modal terms, 'possible', 'impossible', 'necessary', and 'not necessary', Diodorus Cronus was able to excise from  the account of modality the suspicious "gray area" that Aristotle's doctrine

  

Page 19

of 'interpretation' seemed to leave between being and nonbeing. And, to the confusion of Aristotle's followers, Diodorus was able to bolster his doctrine of possibility,  through what seemed to be a powerful appeal to logic itself. My purpose in this chapter is to examine the modalities of Diodorus and the logic of his argument in their  behalf, the so­called Master Argument. 13 I Diodorus defines the possible as what either is or will be, the necessary as what is and will never not be, the impossible as what is not and will never be, and the not  necessary as what either is not or will not be. The modalities of Diodorus thus use tense and quantification14 to reduce the modality of the possible to the terms of the  actual. What is possible is what is at some time actual. Possibility, in other words, is here defined extensionally: only what is or will be real is deemed possible. No  possibility is allowed that is not at some time realized. Thus possibility of the robust sort familiar in ordinary speech and speculation about events that seem to be  contingent is collapsed into actuality—and ultimately, to necessity in the familiar sense of the term, since everything that Diodorus will deem possible must at some time  occur. The impossible is coextensive with the nonexistent. No distinction is recognized between what Aristotelians might call a purely accidental or factitious necessity  and the necessities of universal causal law, or of logic. The strength of Diodorus's claim to have established his revisionist definitions was vested in his celebrated Master Argument. No one in antiquity seems to have found  a means of refuting it. So it offered welcome shelter to thinkers of Megarian sympathies. Epictetus is our main source for its content. He writes: The Master Argument appears to be based on some such considerations as these:The following are mutually inconsistent— 1. In all cases what is past is true necessarily. 2. The impossible cannot follow from the possible. 3. There is some possibility that neither is nor will be true.

Perceiving this contradiction, Diodorus used the credibility of 1 and 2 to establish 'Nothing is possible which neither is nor will be true'.15 We lack the nexus through which the inconsistency was "perceived," but Diodorus apparently had little trouble in convincing his interlocutors that the third was the  suspect item. Moderns have as much difficulty finding the force of Diodorus's argument as the ancients did in discovering its weakness. This being so, I want to suggest  a reconstruction of the argument as an attempt to exploit concessions derived from Aristotle himself.16 Paraphrasing, I would put the argument as follows: Of the following three claims, one must be false:

  

Page 20 1. All truths about the past are necessary. 2. The possible does not imply the impossible. 3. What neither is nor will be may yet be possible.

The dialectic, as I conceive it, would run as follows: The claim that (1) every truth about the past is necessary is Aristotle's own: Nothing that is past is an object of choice, thus no one chooses to have sacked Troy; for no one deliberates about the past, but about what is future and contingent, while what is  past is not capable of not having taken place; hence Agathon is right in saying: For this alone is lacking even to God, To make undone things that have once been done. (Nicomachaean Ethics VI 1139a32–1139b11)

And similarly: No capacity relates to being in the past, but always to being in the present or future. (De Caelo I 283b12)

And again: There is no contingency in what has now already happened. (Rhetoric III. 17 1418a3–5)

By the same token, on Aristotle's account every false past statement is impossible. But there is a bond between times present, past, and future. 17 For every present truth implies some past one. In the midst of an elaborate examination of causality  Aristotle moots an argument for the view that, "Everything that is to be will be of necessity." The argument rests on causal considerations, as Aristotle explains: "For  example, it is necessary that he who lives shall one day die; for already something has happened that would make this so—e.g., the presence of contraries in the same  body." Aristotle heads off the fatalistic inference that some might wish to base on this reflection by adding: "—but whether he dies by disease or by violence, is not yet  determined, but depends on the occurrence of something else." Causes, for Aristotle, are emergent, and only the general pattern, he reasons, can be presumed to be  fixed eternally.18 Thus the generality of causal determinations does not entail the universality of logical determinism. Causes must act in their own due time, and until  they do the future is not determined, and the truth about it, not determinate. But in passages like the one in which Aristotle speaks of the prior existence of causes that will produce a future effect (the contraries already at war within my body),  the Megarians may sense that they have found just the concession they need—tensions or contradictions within Aristotle's own thinking that render possible, or indeed  inevitable, their own dialectical victory. For if Aristotle holds that capacities need not be exercised constantly19 or that potentiality allows futures

  

Page 21

that will not be realized, 20 he seems, the Megarian can argue, already to have severed the connection of the present to the past and violated his own affirmation that  true judgments about the past are necessary. The Stoics typically argued that a present scar 'indicates' a past wound,21 where 'indication' is a Stoic surrogate for implication.22 Similarly, if the names of all and only  the fighters in the 442nd are listed in the register of that combat team, but I am not and have never been a member, my name is not listed. The false statement that my  name is listed in that register implies the false claim that I once served in the 442nd. Since (2) the impossible cannot follow from the possible, the present false  statement, entailing, as it does, a false statement about the past, must be impossible as well. But it would seem that all false statements about the present or the future entail some false statement about the past. This, I think, Aristotle is presumed to grant, in  view of his causal assumptions about the spatio­temporal and physical linkage of the present to the past—a linkage which becomes all the more prominent in Stoic  cosmology. In view of this linkage, all false statements—past, present, or future—are deemed (in the familiar sense) impossible by the Megarian. Those about the past  Aristotle admits to be impossible, since the past cannot be altered. Likewise with the present, since it consists, one is expected to suppose, entirely of present facts.  But false statements about the future are drawn into the same circle of impossibility by the implications that link them to false statements about the past.23 Similarly,  every true statement becomes (in the familiar sense) necessary. There remain no statements that merely happen to be false, and so there can be none that merely  happen to be true. Nor can this conclusion be confined to the realm of judgments. For the Aristotelian correspondence theory of truth can be used to motivate the demand that the modal  character of reality match that of the truths about it.24 So reality itself—past, present, and future—becomes necessary. Global determinism is established a priori, and  the Diodorean modalities come with it: (3) there is no possibility for what is not and never will be. Diodorus has eliminated the "middle class," the possible but at no  time actual. Q. E. D. G. E. M. Anscombe noted years ago that the Master Argument contains "a concealed assumption," namely, that "If it always will be impossible that something has  happened, then it always is impossible that it will happen."25 Anscombe calls this principle "highly specious." In harmony with Prior's analysis,26 I would prefer to  rephrase the hidden and indeed specious assumption somewhat differently, and initially in terms of a special case. What Diodorus seems to me to be saying is that if  there is a state of affairs that will never occur, then (for example) it was always true that it would never occur (and always true to say so, as in the version of the  argument that Aristotle considers). So the event cannot occur without falsifying that truth or claim, thus, without making over the past, which Aristotle concedes is  impossible. Diodorus seems to me to generalize this approach, by assuming that there is always some truth about the past that any falsehood about the future will

  

Page 22

contradict. The inference is that all falsehoods about the future are impossible and all truths about the future necessary. We need not speculate that this line of argument was in fact taken by Diodorus, for we have the testimony of Cicero's De Fato to confirm our reconstruction. The  core of the argument that Cicero links to the name of Diodorus is this: If p ('Fabius was born at the rising of the Dogstar') entails q ('Fabius will not die at sea'), then  the entailment of q by p (a necessary proposition, since it deals with the past) renders q necessary as well, and it becomes impossible for Fabius to die at sea. 27 The strength of the Diodorean position lay in its capacity to refute the conception of possibility put forward by Aristotle, and that by arguing from his own premises.28  If real possibility is indeed an incoherent notion and there are no possibilities that are not at some time to be realized, then all truths become necessary and the time  dimension itself is called into question. For necessity is timeless. Even though Diodorean necessity was achieved through quantification over time, it has the effect of  bracketing or erasing time. Carthago ruit becomes "There is a time such that the fall of Carthage occurs in it." Peripatetic philosophers were wrong, then, it appears,  if the Master Argument is accepted, in claiming that no law of logic so abstract as the law of the excluded middle can determine what will actually take place.29 What  will take place must take place, because it is what will take place. What will not take place cannot take place because it is what will not take place. This outcome,  clearly, will have been of moment to the Stoics, for it displaces natural necessity from its Aristotelian lodgement in the essences of things and lays it in the larger facticity  of the cosmos as a whole, and in the overmastering sentence of fate: Anything can happen,30 but whatever does happen must be. With the elimination of real possibility, modality as an independent category is effectively suppressed. A system of modality which does not allow for possibilities that  may never be realized cannot distinguish possibility from actuality except in point of temporal placement. Modality as such becomes vacuous; and the distinction of,  say, natural from logical necessity, otiose. The Aristotelian, then, and indeed anyone who would retain modal concepts in describing change, causality, or nature has an  interest in refuting the Master Argument. II It is possible to defend real possibility on many grounds, but not all of these will be equally effective against the Megarians. Thus the ironing of possibility into actuality  may be attacked on intuitive grounds as an arbitrary reduction of categories: there ought not to be fewer things dreamed of in our philosophies than are in heaven and  earth. But this smacks of a petitio principii. For the question at issue is precisely which categories are real and which otiose. Appeals to experience and its  categories—change, causality, contingency, and the like—can also be

  

Page 23

dismissed as question begging. For the Megarians are engaged in a radical critique of experience. Perhaps we do habitually suppose that the future is open and  dependent on events that have not yet occurred. But the question at issue is whether such assumptions are ever justified or whether they can be shown to be illusory.  The Megarian reply to any ''ordinary language" defense of real possibility would be the same. Our ability to evolve or devise linguistic categories and distinctions that  faithfully reflect our experience is no defense of the veracity of that experience or the legitimacy of the categories it appears to favor. Indeed, it may be the case that  language does not faithfully reflect our experience, 31 just as experience may not faithfully reflect the ultimate character of reality. Since Megarians find a base for their  critique of experience in an appeal to logic, it is only through a critique of the Megarian logic that Megarian metaphysics can be addressed. Clearly Diodorus's modalities are both more interesting and more challenging than many familiar modal definitions. They are more challenging because of their  metaphysical claims and more stimulating because they seem to reduce modal to nonmodal terms. Ordinary definitions might define the necessary as what is so and  cannot be otherwise, the possible as what may be so but might be otherwise, and the impossible as what is not so and cannot be. Unlike Diodorus's definitions, these  commonplace formulations do not imply a particular metaphysic, although they do allow for one, by permitting real possibility, which the Diodorean modalities  explicitly exclude. I say that the humdrum modal definitions do not legislate because they allow for the vexed case of the possible but never actual, yet do not require it. Conventional  modalities might, of course, be taken to require real possibilities. But if we are to clean up the reductionisin of the Megarian alternatives, our best argument is their  metaphysically tendentious character. We do not remedy the Megarian projection of determinism onto logic by injecting a rival metaphysical bias. So I take the less  muscular modalities only to permit real possibilities, not to require them. The Megarian modalities were less tolerant. But any information that Megarians may have had  that would exclude real possibility should have been presented on its merits, not stowed away in logical conventions. Al­Farabi is at his best on this point, rejecting the  aid of logicians who want to define modal terms so as to canonize real possibility. His reason: that such an approach would compromise the topic neutrality of logic. In  his words, it "would put logic in the position of examining the natures of things and the manner of their existence."32 Ordinary modal talk allows (in principle) that some  metaphysical foundation might be found for determinism. What remains to be rejected in Megarianism is its logicism. The metaphysic that the commonplace definitions seem to favor, by preserving access to genuinely modal categories, is one in which both 'might' and 'must' figure,  neither being reduced to the other. That is hardly as exciting as the reduction of all modalities to actuality or necessity. But it is more faithful to the givens of human  opinion and experience.33

  

Page 24

The commonplace modalities do preserve the modality of the definiendum in the definiens. It is doubtless for this reason that they looked circular to Quine 34—they  preserve something beyond facticity. And that is what the Megarians found most problematic. This sort of 'circularity' is inevitable in any attempt to define the primitive  or elemental. We face here precisely the same problem with modality that Moore (and Hume before him) confronted with morality35 or that Socrates faced when he  worried that Thrasymachus would ask him, in effect, to define '12' but then demand, "Don't you be telling me that twelve is twice six or three times four." (Republic I  337b). Barring reductionism, the modal can be defined only in modal terms; no others will do justice to its unique intension. Thus modal 'definitions' will necessarily be  disappointing, since they only tell us what we already know, even if we do not wish to admit we know it. Strictly speaking they are not definitions at all in the formal,  analytical sense, in which things break down into what they themselves are not, but only paraphrases. But this fact, the fact that they are not definitions, is what allows  the commonplace modalities their purchase on truth. The commonplace modalities are not merely accurate reflections of linguistic usage and our categories of thought  and expectation. Nor even can we claim, without some proof, that they are reflections of reality. But they are part of a scheme into which all metaphysical pictures may  be fitted, including the metaphysics of Megarianism and determinism and those of contingency­potentiality­causality and the open future. They allow for the world,  whatever its metaphysical character may be. Or (to put the matter conventionally) they are true in (or of) all possible worlds. Diodorus wants to prove that this is not so. Here lies both the elegance and the flaw of his position: The elegance, in that he offers an extensional account of possibility  without reference to the notion of possible worlds—thus avoiding the nonreductive (or ultimately intensional) character of such accounts. As Nicholas Denyer puts it,  "Diodorus's definitions are quite innocent of circularity: not one of the modal terms to be defined recurs in any of the definitions given"36—a statement that cannot be  made of Philo's definitions—or of those who define modalities by resort to the idea of possible worlds. The flaw, that Diodorus has not allowed for what is crucial in  the idea of possibility, the notion of alternative futures. The Megarians believed that being is such that all possibilities are realized. This is a metaphysical belief, since it governs all the things that are and speaks of them  inasmuch as they are. To be is now to be incapable of being otherwise. In following the Parmenidean monolithic interpretation of 'being' the Megarians follow  Parmenides into his confusion of metaphysics with logic—the science of being with the science of formal possibility. Thus their presumption that their metaphysical  doctrine can be demonstrated on logical grounds. Megarians could hardly claim to justify their modalities on any but a priori grounds. For experience is frustrating enough to give the lie to the naive belief that all  possibilities are realized; and from the same teacher we soon learn to dismiss as sour grapes the ex post facto rationale, 'If I did not do it, it cannot be done'.

  

Page 25

But how can dialectic gain a purchase on the slippery Megarian eel? Besides process and disposition, causality, and time itself, space too, we might argue, would be  severely curtailed, if each particle and being as a whole were restricted in its latitude of movement to the locations it actually will occupy. Are we mistaken to presume  "unimpeded" space to be of freer access? If so, by what mechanism would the new class of invisible barriers be supplied? Would we not be warranted in distinguishing  these by some idiom from the operationally more substantial barriers that impede what we more conventionally call possible, as opposed to actual motion? If there  were no possibility of walking down "the passage we did not take," what can we mean in saying that any passage was there at all? But, of course, such lines of  argument are entirely wrongheaded. For if the phenomenal categories are the objects of the Megarian attack, then no appeal to causality (as when we demand a  mechanism for the interposition of invisible "barriers") or to the linguistic recognition of potentiality (as when we ask 'What can we mean when we give directions'?) is  possible. Space too, we must admit, along with time, is just another of the items on the Megarian "little list." A metaphysical argument may fare slightly better against the Megarian preconception. Thus, beyond mere space, it might be said, some doubt remains which  metaphysic is more economical—if ontological economy were an authentic goal in the elimination of potentiality and thus of real dispositions—a system in which each  entity, and being as a whole, bears its own ontological straitjacket (to restrain it from everything but what it will in fact do) or one in which each thing can respond to  causes in accordance with their character and its own? The latter would involve dispositions, yet the former would equally imply "indispositions." Megarians, of course, might not admit ontological standing to such incapacities, any more than they would allow ontological status to positive capacities. Their aim is to  take 'reality' more seriously than that. Thus they remain faithful not only to the object but also to the method of Parmenidean philosophy. III What logical feat can make it impossible for all the dogs in New York to bark at once merely because they never have and never will? If this were permissible, would  it not be equally possible for a logician to cast a truth­functional spell condemning the vast majority of honey bees to helplessness merely because they will never use  their stings? The Peripatetics are rightly suspicious of the Diodorean definitions, since these efforts at redefinition seek to smuggle dubious metaphysical baggage into  the first class compartment reserved for logic. But the question in antiquity was hardly confined to so simple a matter as the proper role of logic. Instinct, intuition,  experience itself seemed to rebel against the Megarian elenchus, yet this did not move the followers of Diodorus. In the face of attacks by Diodorus's student Philo of  Megara and by the great Stoic Chrysippus, they stoutly

  

Page 26

maintained that a block of wood that will remain forever at the bottom of the sea is not combustible; a sunken shell that will never be seen is ipso facto invisible. They  shrugged off the implications of their view for the coherency of nature and experience, the possibility of science, the relevance of precautions, and the appositeness of  admonitions, resting their case on the Master Argument. Where, then, does the argument go wrong? Perhaps some of the confusion is due to a failure to distinguish modal contexts. A woman is due to have a baby.  Necessarily, we say, it will be either a boy or a girl. But that necessity is on quite different footings before and after the baby is conceived. Prior to conception it is the  necessity of a truism, dependent upon the assumption of certain definitions that fairly sharply bifurcate the offspring of our species. But after conception, the  disjunction, originally invoked to forestall presumptions as to the (then indeterminate) sex of the (then nonexistent) babe, can and must be suspended. One of the two  alternatives has descended into reality. The sex that in three months will appear anatomically is now genetically determined. It makes as much nonsense of natural  science to draw any inferences from the truth that might have predicted the outcome as it does to claim that the outcome was always (e.g., prior to the existence of its  causes) determinate. For, in either case, what is merely a natural necessity has been transformed into a necessity of metaphysics or of logic. If this were allowed, there  would be no significance in discovering the functions of X and Y chromosomes. They would have no functions, properly, except as "signs" of what timelessly was to  be; the outcome was determinate before such particles as chromosomes existed. Thus, despite the appeals to causal experience that typify various brands of  determinism, Megarian determinism is inherently anti­causal. Its flavor is captured timelessly in the fatalism of Cicero's Stoic: "These portents were sent by the immortal  gods to Caesar that he might foresee his death, not that he might prevent it" (De Divinatione I lii, cf. II vii). Megarianism dreams of elevating all facts to the status of  necessity, but by so doing it only succeeds in debasing the meaning of the term. The force of modal terms, I would suggest, is context dependent; for every modal term or concept works at a variety of different levels of intension. Thus the moral  necessity that the president be an ethical person is of a different order from the presumed political necessity that the president be male. The constitutional necessity that  he be over 35 years of age differs from the biological necessity that he belong to the human rather than any other terrestrial species. The metaphysical necessity that he  be a concrete individual differs from the logical necessity that he not be a spherical cube. In each case, the parameters of necessity are dictated by a particular system  of assumptions regarding possibility, assumptions based on our knowledge or belief or the rules and principles that are germane (or are thought to be) within a certain  sphere. Both in speaking and in thinking we constantly invoke such frames of reference, which provide postulates to guide expectations in accordance with our modal  intensions.

  

Page 27

Thus, when speaking of nature, we often intend the assumption of certain causal theories; other frameworks of discourse may assume laws or customs, linguistic  conventions, common "understandings," or the special premises of an art, a science, a subculture, a conversation, or perhaps a work of fiction. 37 It is characteristic of  natural as opposed to artificial languages that they allow and even encourage constant shifts in the assumptions that frame our modal expectations. This fact has  perhaps encouraged some philosophers to suppose that there are no objective modal facts. But, of course, the flexibility of our language tells us nothing about the  modal character of reality, except perhaps about the usefulness of varying our views of it. All the resources of communication—tone, gesture, facial expression, quotation marks, italics, to name only a few of the least complex means of communicating what  is not always expressed—are put to use in setting a stage, outside of which the naked and unsupported propositional content is as devoid of sense as if it were  semantically vacant or opaque.38. Modal intensions and assumptions or understandings are an essential part of such a setting. For modal notions can operate only  when they have found or defined a specific framework—whether of hope or prejudice, law or nature, morals, metaphysics, logic, or the nuances of irony and  humor.39 To those who accordion these levels, the necessity of logic can mean no more than that of chemistry or convention. Now the question may be asked, and it is asked by my friend May Sim, how do we (or should we) distinguish modal contexts and the modal usages appropriate to  each, and what light is shed on the relations between logic and metaphysics by the notions that such contexts should be kept distinct and that there are right and wrong  ways of operating in or with each such context? My answer is that this is a place to avoid a priorism. We should heed (but not blindly or mechanically) the cues of the  discourse we are engaged in and the mode(s) of experience it addresses. The modal terms themselves will differ little, if at all, in their syntax or relations from one  context to another.40 Where they will differ is in their force, that is, in the farrago of associated assumptions they drag along with them. These will not readily be  captured in some pre­established taxonomy but will vary with experience and with the growth and change of our modes of conversation and interaction as the arts and  sciences shift their gaze and as we continually renegotiate the shared presumptions of our discourse from one moment to the next. My only plea is that we not mistake  the varieties of hypothetical necessity for the firmer stuff of logical necessity, the abstracted or background modality that supplies their form or format.41 The Master Argument harbors more than one fallacy, but perhaps the most glaring is its equivocation on 'necessary'. Does 'necessary' in premise 1 mean simply out of  reach, or does it attract some of the aura of logical necessitation? The easy slip of forgetting that the "unalterable" past was not always unalterable is sufficient to  obliterate future contingency. But there is no need to allow a purported reform of logic to beg a question of metaphysics—and that in a direction that

  

Page 28

subverts all the categories of familiar experience. Granted that the truth of a judgment entails the reality of a corresponding state of affairs, we are given truth in just this  sense only after the fact. There is an asymmetry in things that does not allow deduction from a truth to the necessity of just any fact. That asymmetry is time. But for the Megarian, we must always recall, it is time that is most problematic; and, for that reason, it is precisely the distinction between logic and metaphysics that  the Megarian is not prepared to concede. In refusing to make such a distinction, the Megarian, we can see, is doing metaphysics, not logic. 42 But it is hard to drive  home the point without a petitio. Megarian thinking is not a heresy of Aristotelianism that can be brought back to the fold by dialectical application of Peripatetic  principles. It is or was an independent philosophy with a (decidedly anti­empirical) base and bias of its own. We may prefer a system that affords at least an  opportunity for recognition of the reality of time and change, but we cannot invoke those categories against Diodorus except to point out how much the Megarian is  missing. Aristotle Saw clearly that change could not long survive without potency. He wrote: Since that which has no potency is incapable, it would follow that what is not actually occurring is incapable of occurring; and of course it must be erroneous—by the very  meaning of 'incapable'—to say that what is incapable of occurring either is or will be. Whence it is clear that the argument in question does away with motion and even with all  becoming: whatever is standing will always stand and what is sitting will always sit. I.e., if it is sitting it will never stand up, for it is obviously impossible that anything should  stand up that is incapable of standing up. Since the conclusion is untenable it becomes clear that there is a difference between potency and actuality, and those theories that  identify them are trying to abolish a distinction of no little importance.43

The reductio assumes, as Aristotle sees, that what has no potency is impossible. But the absurdity to which Aristotle seeks to reduce his adversary is a welcome  outcome to a Megarian. AS Aristotle seems here to recognize, the argument he finds conclusive would not seem so to a committed monist. Diodorus Cronus in fact denied the reality of motion,44 and this, as far as Aristotle was concerned, implied a denial of all change, and of time itself, since all changes  were species of motion and time was motion's measure. Thus the consistent Megarian is faithful to his Eleatic heritage. His critique of potentiality is just one step in his  attempt to unravel the fabric of appearance, by tugging hard at the Aristotelian analysis of change and the Aristotelian defense of time implicit in it. Aristotle's reductio  led to an "absurdity" that was not the endgame but the point of departure for the Megarian (anti­empirical) metaphysics. In fact, there will not be an endgame here. For, should an Aristotelian claim conclusiveness in these arguments, that would be to claim that a formal rather than

  

Page 29

a material error lies at the base of the Megarian philosophy. To make such a claim would be to undermine one of the great achievements of Aristotelianism, the  distinguishing of logic from metaphysics. The Megarian, who fails adequately to make such a distinction, can denounce Aristotle's modal scheme as logically  incoherent, but the Aristotelian cannot reply in kind. He can only exhibit to the fullest extent possible the cost in implications of adopting the Megarian view. Here we  see one of the critical limitations of the dialectical method in philosophy. It can call an interlocutor to his senses by asking him to acknowledge the premises that his  own thoughts and actions seem to presuppose. But it cannot guarantee that he will acknowledge them or construe them in the same way that others do, or that we feel  others should. We underrate the vigor of the Megarian position if we suppose that Diodorus's denial of potentiality was independent in his mind from his denial of  motion and, by implication, of time. A rejection of time, and indeed of all multiplicity can, in fact be developed from Diodorus's modalities themselves. The process might work somewhat as follows. First  Diodorus employs the Master Argument to establish his modalities. This gives him determinism, but it also gives him a new, extensional surrogate of the modal  categories, defined exclusively in terms of truth­functions, quantification, and time. Next, Diodorus proceeds to define implication in terms of this modality­surrogate: A  conditional sentence, according to Diodorus 45 is true if and only if its antecedent cannot (in Diodorus's own sense) be true while its consequent is false. That is, an  inference is valid if and only if it is not and never will be the case that its antecedent is true while its consequent is false.46 This position on the heated issue of the truth conditions of hypotheticals is, of course, implicit in the Diodorean rejection of modality as an independent category. For it  amounts to an extensional definition of implication. If constant coincidence is the stuff of necessity, then only constant coincidence can anchor inferences. Essences,  dispositions, even causes play no role. Coincidence, fact, and necessity (of all types) are placed on a par; and any combination of two sentences one of which is not  and never will be true while the other is false, is an implication. Implication, then, like modality, is defined by Diodorus as a truth­function quantified over time.47 But can time itself survive such usage? Suppose all judgments, to be  true, require universal time quantification. Time in that case would drop out of the equation. It would lose its use in differentiating sentences now true and now false  from those that are always true or always false. Moreover, all truths, being true always, would imply one another; all would become simple truth functions, dependent  upon one another and derivable from one another—an outcome suggestive indeed for the Stoic notion of sympatheia48 and clearly redolent of Parmenidean monism  as well, since truths deemed interchangeable might well be deemed identical: If necessity is no more than constant coincidence, why should identity be more than  interchangeability? If all truths are identical, so are all beings. A Parmenidean unity may be derived, we suggest, along with the collapsing of time, by the addition to the  Diodorean notion of implication of the assumption that

  

Page 30

all truths must be true at all times. As Jaakko Hintikka put it, "Take any sentence of the type 3 [= 'p at time t0']. If this sentence is true once, it is true always. If  necessity equals omnitemporal truth, this means that (3) will be necessarily true if true at all; and by the same token it will be impossible if false. Hence all statements  about events that are individual in the sense of being tied to a particular moment of time will be either necessarily true or necessarily false. Everything that happens thus  apparently happens necessarily." 49 But we know that Diodorus did demand that true propositions must be true at all times, for that was the basis of yet another of his disagreements with his student  Philo. What Diodorus said, in effect, was that only temporally quantified propositions can be deemed well­formed; and these, of course, are true at all times. The sort  of sentences that Philo would countenance, e.g., 'Socrates is swimming', Diodorus seems to reason, are true because what they state is the case at some time.50  Michael White argues that it is "counter­intuitive with a vengeance" to read Diodorus as holding that past truths are necessary. For this would mean that, for a given  proposition p(t), say, 'Socrates is suntanned' (really a propositional function containing one free variable, 'Socrates is suntanned at t', if there is a time in the past at  which it was true, then it is presently necessary.) Interpreting that claim in terms of Diodorus's own modalities (see Boethius, in De Int., ed. Meiser, 234), White  derives the conclusion (66) that "if Socrates was once suntanned, then he presently is suntanned and ever hereafter will be suntanned." But I think the conclusion  Diodorus intended was that if Socrates was once suntanned, or is, or ever will be suntanned, it is forever true (and thus necessary, now in the Diodorus's sense) that  he be suntanned at that time. For Diodorus, there is a place, if you will, beyond all temporal appearances, where the suntanned Socrates (or whatever beings, however  qualified, are durable in such a place) abides eternally. The Diodorean assimilation of the possible to the actual and the actual to the necessary suggests that such an  assumption was not alien to the Megarian way of thinking. For why, if a proposition is true, should it not be true always, sub specie aeternitatis? Aristotle's object in the Sea Battle discussion was to show that necessity is not distributive across a disjunction. The necessity of 'yes or no' does not imply the  necessity of 'yes' or the necessity of 'no'.51 This, in effect, is what the Megarians deny. By pressing the law of the excluded middle as a rigid demand for an immediate  determinate truth value for all claims, they are in effect demanding universal time quantification for all propositions now. Diodorus may appear to be more liberal in  speaking of 'what is or will be', but in insisting that what we experience as future is actual, and not in any robust sense merely possible, he is not granting the futurity of  truth values. Truth values on the Megarian reading, are always definite, whether known to us or not. Thus, for a Megarian, a proposition not true at all times is not true  at all. But if a true proposition is to be true at all times, the reality to which it corresponds must be equally real at all times. The conclusion may appear strange,

  

Page 31

but it was precisely what the Megarians were seeking. Time itself is unreal. In the words of Cicero's Stoic, "the passage of time is like the unwinding of a rope: it  creates nothing new, but only unfolds each event in its order" (De Divinatione I lvi 127). Time here is made subjective, like the movie whose action is revealed as the  pre­existing celluloid unwinds; the order that is conceded is only the order of appearances. Thus determinism is grounded in ontology, and becoming is treated as  creation (time "creates nothing new"), and so brought within the withering fire of Parmenides's charge that nonbeing can never come to be without a violation of the  most fundamental rule of reason, the transformation of is not into is. The Stoic affirms his Eleatic allegiance, by treating all becoming on a par with absolute becoming.  52

It is only in view of the passage of time that Aristotle feels justified in asking no more for the exclusion principle than that ultimately all propositions will have just one  of the two truth values. It is the monism of the Megarians that demands the sorting be done immediately. Megarians do not mortgage truth values, for the simple reason  that for them the time dimension does not exist. Our finite minds, it may be true, are capable only of linear and thus temporal experience. But, to the Megarian, the unreality of future contingency suffices to prove the  unreality of indeterminacy, or rather—to draw the properly Eleatic conclusion—the reality of determinacy now and always. Just as time and the related categories of  potency and change are illusions, too transparent to withstand candling by the absolute standard of monistic ontology, the pure light of the One, so too they will not  withstand the test of monistic epistemology. In reality, and in the Absolute mind that knows all things as they really are and sees all things as they actually exist, time  and all kindred phenomena are nonexistent. All truths imply each other and are thus equivalent, reducing to the one great fact of being, esti. Reality, which truth  mirrors, is likewise focused to a point. Or rather, since the ideal or real mind which transcends time and plurality would find no diversity within itself, the differentiation  of subject and object are overcome, the Parmenidean dream of uniting thought and fact under a single sense of being is achieved, and the monism so characteristic of  one variety of mystic experience becomes evidence of the ultimacy of that unity. So far more is at stake in the Master Argument than a mere exercise in logic or even a vindication of logicist determinism. Diodorus had within the train of his subtle  reasoning the potential—if use of that word is not unkind—for undermining the Aristotelian defense of future contingency and change by refuting the idea of potentiality  on which both contingency and change depend. But he had as well the potentiality of vindicating his own denial of motion and the rejection of time that it implied. He  even had the opportunity of reducing all truths to one and restoring the Eleatic Unity to the solitary reign that reason had seemed to promise it. How far Diodorus went  in this direction, beyond his denial of motion, will perhaps never be known. There is reason to believe that the potentials opened up by his arguments against  potentiality and motion did not remain unexploited by his successors in the

  

Page 32

Stoa and the Kalam. But that question goes beyond the scope of the present investigation. Notes 1. D. N. Sedley questions Diodorus's Megarianism, placing him instead in the "Dialectical School," but there is ample testimony to Diodorus's commitment to the  Megarian school founded by Socrates's disciple Euclides. Klaus Döring, Phronesis 34 (1989): 293–310, has argued convincingly that the "dialectical school" is in fact  an imaginary construct, indistinguishable from the Megarian. 2. Thus the 'capacity' of the Ash'arites. See al­Ash'ari, Kitab al­Luma' in The Theology of al­Ash'ari, R. J. McCarthy, ed. and tr. (Beirut: Catholic Press, 1953), ch.  5. Monovalent capacities like those of the Ash'arites cut off the Aristotelian objection (Metaphysics Theta 3, 1047a11) that if there were no potential for an event it  would not occur. But they do not leave room for naturalistic counterfactuals or unrealized possibilities. Whatever happens had the potential to happen, but there is no  shadow potentiality for what did not happen, what "would have happened, had it not been for ..." All is determinate. The Ash'arite theologian al­Ghazali writes in his  Ihya? 'Ulum al­Din that phrases like 'Had it not been for ...' are demonic temptations that open the door to shirk, idolatry; see Ihya', book XXXV, Tawhid, Bayan  2 (Cairo: Aleppo Foundation and Publishing and Distribution Companies, 1312 A. H.), vol. 4, 190. 3. Cf. Gilbert Ryle's brilliant analysis in "It Was to Be," in Dilemmas (Cambridge: Cambridge University Press, 1954; reprinted 1962), 15–35. 4. See Parmenides, Fragment 6, ap. Simplicius, in Phys. 86, 27–28; 117, 4–13. 5. See Hippias Minor 367bc, Sophist 247a. 6. Cf. Aristotle, De Interpretatione 9, and al­Farabi, Sharh fi­l­Ibara (Commentary on De Interpretatione), W. Kutsch, S. Marrow, eds. (Beirut: Catholic Press,  1960), 100; F. W. Zimmermann, tr., in Al­Farabi's Commentary and Short Treatise on Aristotle's De Interpretatione (London: Oxford University Press, 1981),  95. 7. See Diogenes Laertius, ix 72. That Diodorus saw the connection is known from the testimony of Sextus; see Against the Physicists II 85–87; cf. Michael J.  White, "The Spatial Arrow Paradox," Pacific Philosophical Quarterly 68 (1987): 71–77. 8. Cf. the response to Hintikka in Michael J. White, "Facets of Megarian Fatalism: Aristotelian Criticisms and the Stoic Doctrine of Eternal Recurrence," Canadian  Journal of Philosophy 10 (1980): 189–206. The argument that the Diodorean account of the modalities "takes away all possibility of change and of coming into  being" (193) "must be regarded as an argumentum ab extra" (201). 9. Thus his treatment of Melissus, at Topics I. 11 104b21–22. 10. See De Interpretatione 9, 18a38; al­Farabi's, Sharh fi­l­Ibara, Kutsch and Marrow, eds. 82, l. 23; Zimmermann, tr., 76. 11. See al­Farabi, Sharh 83 ll. 8 ff. Whether future contingency requires a "many valued" logic has remained a subject of philosophical speculation; cf. A. N. Prior  "Diodoran Modalities," Philosophical Quarterly V (1955): 205–13. Influenced by Ryle's Dilemmas and his own discovery of several ways of mapping future  contingency upon a two valued logic, Prior argues that there is no necessary conflict between an open future and the law of the excluded middle; see his John Locke  Lectures of 1955–56, Time and Modality (Oxford: Oxford University Press, 1957), 94ff., but cf. 86.

  

Page 33

12. See De Interpretatione 9, 18b5ff. A propos Aristotle's charge that Megarian determinism does not allow each event to occur ''as it might," al­Farabi comments,  "When he says "as it might," he means without some determinate essential cause arising from its own nature." At 18b26, al­Farabi explains the endpoint of Aristotle's  reductio: "Events would not result from our sound determination that an action will occur; nor would they be prevented by our sound determination that it will not."  Instancing the rival efforts of two persons to bring about or prevent a given event, he adds, ad 18b33: "the actions of neither of them would make any causal difference  in bringing about or preventing that event. The outcome would simply occur of itself." 13. See Boethius ad De Interpretatione 9; cf. William Kneale and Martha Kneale, The Development of Logic (Oxford: Oxford University Press, 1962), 117 ff.;  Benson Mates, Stoic Logic (Berkeley: University of California, 1961), 36–41. 14. Quantification is required because the Diodorean reduction treats necessity in terms of what is true at all times, possibility at some, impossibility at none. 15. Epictetus, Dissertations, ed. Schenkl II 19.1. 16. For the reconstructions by Zeller, Copleston, and Rescher, see Peter Øhlstrom, "A New Reconstruction of the Master Argument of Diodorus Cronus,"  International Logic Review 21 (1980): 60–65. As Øhlstrom points out, these efforts were marred by the assumption that "follows from" meant temporal sequence  rather than logical implication. Premise 2 was taken to mean: "The possible cannot become impossible." Mates showed in Stoic Logic (39) that the term used was  Diodorus's usual term for entailment; and that was how Chrysippus understood it when he attacked the argument. Prior (in "Diodoran Modalities") interpreted the  Master Argument in terms of the idea that "When anything is the case, it has always been the case that it will be the case." Hintikka, "Aristotle and the 'Master  Argument,'" American Philosophical Quarterly 1 (1964): 101, objected that he could not see why Diodorus "would have taken a statement concerning the truth of  past predictions as being a statement concerning the past." But, as John Sutula points out in "Diodorus and the Master Argument," Southern Journal of Philosophy  14 (1976): 323–43, 327, it is hard to construe the Diodorean sentence, "It has been the case that the shell will never be seen" as anything but a statement about the  past. Somewhat less awkwardly phrased sentences of the form "It has long been known that no   will ever  ' are very much the stuff of common discourse. Frederick  S. Michael also tries his hand at a reconstruction, in "What was the Master Argument of Diodorus Cronus?" American Philosophical Quarterly 13 (1976): 229–35.  Øhlstrom objects to Michael's version on two grounds: first, that it adopts a propositional interpretation of the entities that are supposed to be true or false, necessary  or possible; second, that it "proves too much," since it negates the distinction given in the premises between 'possible', 'true', and 'necessary'. I do not find this second  objection telling, since the argument is dialectical. It might well start out from the assumption of distinctions that it will reject or redefine in the end; they are, after all,  the distinctions of Diodorus's adversaries. But Øhlstrom (62) seems right in saying that Mates's interpretation of the argument as addressing propositional functions "is  more relevant than Michael's." Øhlstrom seems to me to be right when he follows Prior in interpreting the argument as based on the assumption that "time is to be  measured only by means of integers." For Megarian analysis does tend to freeze time, as a series of discrete moments in which there is no blur or flux. This quantizing  of time is much in keeping with the general Diodorean approach to "indivisibles." Øhlstrom's own reconstruction seems to me to suffer from a defect that Rescher  identifies in his own, if it is to be read historically: "Too much modern machinery is deployed" to warrant any claims to full historical authenticity." See Nicholas  Rescher, "A Version of the 'Master Argument' of Diodorus," Journal of Philosophy 63

  

Page 34

      (1966): 438Ø45. Øhlstrom's version is replete with formalism. It also rests on an assumption that he calls "very much like one of the so­called paradoxes of the  material implication," namely, that a false proposition implies any proposition. But would Megarics or Aristotelians rely on such a premise here? The premises of the  Master Argument itself seem to ask us to assume that the implications of what is possible are different from those of what is not. This would not be so if all false  judgments had the same implications. Indeed, we know that the Skeptics invoked a version of the paradox of material implication against Diodorus; see Sextus,  Outlines of Pyrrhonism, II, 110–11. So it seems doubtful that Diodorus himself will have used it; cf. n. 46 below. One final point about Øhlstrom's treatment: to  reject Diodorus's Premise 2 on the grounds (65) that "a catastrophe in a certain atomic power plant would generally be considered to be possible although it might  never happen," is to beg the very question at issue; see n. 8 above. 17. The Megarian assumes time for the sake of the argument. This is a fine example of the dialectical status of the whole discussion between the rival schools. An  Eleatic might say that time here belongs to the Way of Seeming. The Megarian dialectician need only say that he makes use of his adversary's assumptions, which are,  in the end, found wanting. The Stoic version of monism is instructive in this regard. For the Stoic cosmos was causal, temporal, and interactive. That was the force of  the idea of sympatheia. But the understanding of that force led not just to recognition of the interconnectedness of all things but to the discovery of their ultimate  identity within the divine unity or natural plenum. The growth of our understanding, it seems, is meant to move us from determinism to monism, that is, from the Way of  Seeming to the Way of Truth. Such uses of dialectic are paralleled and developed in the thinking of Nagarjuna. 18. See Metaphysics VI.3 1027a29–1027b16. 19. See Metaphysics Theta 3. 20. Aristotle himself is not perfectly consistent on this point. Contrast De Interpretatione 9 and Metaphysics Theta 3 with Metaphysics Theta 4 and De Caelo I.12.  Cf. Gerhard Seel, "Diodore domine­t­il­Aristote?" Review de Métaphysique et de Morale 87 (1982): 293–313, following J. Vuillemin, "L'argument dominateur,"  Review de Métaphysique et de Morale 84 (1979): 225–57. Vuillemin explicates the Master Argument as deriving dialectical leverage from De Caelo 283b6–17,  read as a refutation of Plato's doctrine of a created but everlasting world. Al­Farabi here sides with Plato. His polemic with the Ash'arites motivates his preservation of  the distinction between the modal and the merely circumstantial, or as we would put it, the intensional versus the extensional. (The Ash'arites, by contrast, are  motivated in their behaviorism by the traumatic experience of the Mu'tazilite mihna or inquisition. Better, they reason, to let God judge human dispositions and confine  our own attention to human acts and utterances.) Al­Farabi, like most Neoplatonists, has rejected both temporal creation and the view that Plato ever advocated such  creation. He sees temporal creation as poetry, signaling the subtler truth of emanation. He is therefore comfortable with Plato's "eternal possibilities." See Sharh p.  100 ll. 5–13; cf. L. E. Goodman, "Al­Farabi's Modalities," (in Hebrew with English summary) Iyyun 23 (1972): 100–12. For Aristotle's waffling on real possibilities,  see Simon van den Bergh's commentary on Averroes's The Incoherence of the Incoherence vol. 2, 215, items 3, 4, 13, 15, 16. 21. Sextus Empiricus, Against the Logicians II 254. 22. The Stoics preference of "indications" to implications reflects their nominalist penchant and refusal to ground knowledge in natural kinds. For the Stoic suspicions  of class logic, see L. E. Goodman, "Skepticism," Review of Metaphysics 36 (1983): 819–48.

  

Page 35

23. The linkage is accomplished through the seemingly innocent Premise 2, which, as Michael White puts it, merely expresses "the principle of reductio ad  impossibile;" see his "The Necessity of the Past and Modal­Tense Logic Incompleteness," Notre Dame Journal of Formal Logic 25 (1984): 59–71, 61. For  Aristotle, crucially, this linkage is causal; for the Megarian the notion of entailment makes it a matter of logic. 24. Cf. Al­Farabi, Sharh p. 82 ll. 21–24. 25. G. E. M. Anscombe, "Aristotle and the Sea Battle," Mind 65 (1956): 1–15, appendix on Diodorus Cronus, now reprinted in Aristotle: A Collection of Critical  Essays, J. Moravcsik, ed. (Garden City, New York: Doubleday, 1967), 15–33. 26. See the discussion in n. 16 above. 27. See Cicero, De Fato, vi 12, vii 13; cf. Michael J. White's discussion in "Facets of Megarian Fatalism: Aristotelian Criticisms and the Stoic Doctrine of Eternal  Recurrence," Canadian Journal of Philosophy 10 (1980): 203–4. 28. For the reliance of the Master Argument on Aristotelian premises, see, Peter Øhlstrom, "A New Reconstruction," 61; and O. Becker, "Zur Rekonstruktion des  'Kyrieuon Logos' des Diodorus Kronos," in Festschrifit für Th. Litt," J. Derbolav and F. Nicolin, eds. (Düsseldorf, 1960). 29. Al­Farabi, Sharh, 84 ll. 12–3. For the Stoic persistence in grounding determinism in the law of the excluded middle, see Cicero De Fato 38. 30. See L. E. Goodman, "Skepticism," 833. 31. Consider, for example, our lack of words to express many of the nuances and complexities of our sensory, social, and emotional experience. 32. Al­Farabi, Sharh, Kutsch and Marrow, eds., 84; Zimmermann, tr.,78. 33. For Aristotle those givens are a primary datum of philosophy. See G. E. L. Owen, "Tithenai ta Phainomena," Symposium Aristotelicum (Louvain, 1961), 83– 103; reprinted in Moravcsik, 167–90. 34. "W. V. O. Quine, "Two Dogmas of Empiricism," in From a Logical Point of View (Cambridge: Harvard University Press, 1961; first ed. 1953). 35. The problem with naturalistic definitions of the good is that they inevitably omit some key element of what human beings intend by the idea—inevitably, because  the idea of the good will always rise higher than any of its merely empiric representations or factitious instantiations, and thus will always leave it an open question  whether its purported analyses have been successful. In the case of modality, the omitted intension is not that of a floating axiological paramountcy but that of a floating  modal paramountcy. Omit what is modal here and you have omitted everything, just as you have omitted everything from a moral notion if you omit all reference to the  good. 36. "Time and Modality in Diodorus Cronus," Theoria 47 (1981): 31–53, 47. 37. It is not the case, of course, that scientists invariably assume the sum total of "scientific knowledge"—although that is a kind of goal or regulative ideal for them.  Often they must lay aside certain assumptions in order to avoid circularity or contradictions in their reasoning or to question or test critically what on other occasions  might be fundamental assumptions. 38. See L. E. Goodman, "Context," East and West 38 (1988): 307–23. 39. It is widely assumed that the recognition of intensionality provides compelling support for subjectivism or cultural/linguistic relativism. But that inference is a non  sequitur. For the compatibility of modal realism with a perspectival account of modality, see L. E. Goodman, God of Abraham (New York: Oxford University Press,  1996), 64. What logic needs, however, is not modal realism but modal pluralism, simply not to prejudge the universe in favor of some particular modal construction.  As Allan Bäck remarks, the

  

Page 36

      categorical demands of logical necessity are modeled in the treatment of the backgrounding assumptions of logic as a null set—or perhaps we should say a minimal  set: the postulates needed to allow inference or implication, but none of the special postulates, say, of physics or philately. 40. I am sensitized to this point by the urging of Allan Bäack. 41. Aristotle himself makes the distinction between categorical and hypothetical necessity (De Partibus Animalium 642a1–11), and I find it striking that he does so in  the context of a biological work, where emergent causes are most vividly in play. Cf. Allan Bäck's discussion in "Aristotelian Necessities," History and of Logic 16  (1995): 89–106. 42. My colleague Jeffrey Tlumak notes: "Contemporary logicians often do the same thing, without acknowledging their metaphysical biases. E.g., although  nonMegarian, most wielders (for substantive purposes) of modal logic adopt system S5, which conflates logical and metaphysical possibility (and so necessity),  subsequently begging many important philosophical questions. S5 itself is definitively nonperspectival: in its semantics the alternativeness relation is an equivalence  relation." Accordingly, modalities seen from one "possible world" necessarily preserve their modal character in all possible worlds. 43. Metaphysics Theta 3, 1047a10–20, Wheelwright, tr. 44. Sextus Empiricus, Against the Physicists II 48, 85, 120; Against the Professors I 311. Note the Eleatic character of his argument, i.e., its dependence on  overstrict interpretation of the exclusion principle as applied to the verb to be. 45. Sextus Empiricus, Outlines of Pyrrhonism II 110; cf. Against the Professors I 311. 46. See Sextus Empiricus, Against the Logicians II 115 and Benson Mates, Stoic Logic 42–51; cf. his earlier version of this discussion in "Diodorean Implication,"  Philosophical Review 58 (1949): 234–42. Diodorus's version of implication as presented by Sextus seems intended to offer a corrective to the proposed definitions  of his pupil Philo. Philo defined a conditional as "true" when and only when it does not have a true antecedent and a false consequent. This, as Mates notes, following  C. S. Peirce, is "exactly the same as the modern so­called 'material implication'" Diodorus wanted something stronger, since he objected to the notion that a  conditional was "true" unless it held at all times. He did not accept what we today would call strict implication, an approach that some others (perhaps including  Chrysippus) took, but instead formulated what Mates identifies as a type of "formal implication.'' The plausibility of Diodorus's approach may have rested on infusing  what we call 'validity' into the Diodorean idea of truth. That is, the Diodorean true implication was true for all values of the time variable, past, present, or future. But  Diodorus's approach also reflects his bias against time, since it is time that invalidates Philo's schema. Mates remarks (48) that the example Sextus gives (Against the  Logicians II 110) of Diodorean implication shows that "the ancients were aware that Diodorean as well as Philonian implication had its paradox, namely, that a  proposition which is 'always false' implies anything, even its own negation." But I see the matter a little differently: Sextus cites the conditional "If atomic elements do  not exist, then atomic elements do exist," rubbing the Megarics' nose in Diodorus's atomism, in order to show that here is a hypothetical whose antecedent is always  false and whose consequent is always true, on their account, but that plainly cannot be "true," since it has not­p implyingp. The point was to show that a Skeptic could  lodge the same type of objection against Diodorus as he had lodged against Philo. For, although time has been "cancelled out" of the equation by being universalized,  Diodorus's "stronger" definition of entailment is still a variant of (Philo's) material implication and vulnerable to a critique based on the paradox of material implication.

  

Page 37

47. Cf. Benson Mates, Stoic Logic, 44–7; Nicholas Rescher, Studies in the History of Arabic Logic (Pittsburgh: University of Pittsburgh Press, 1963), 79–80; A.  N. Prior, Time and Modality, 8 ff. 48. Michael White ("Facets of Megarian Fatalism," 205) also suggests a connection with the Stoic doctrine of periodos, or eternal recurrence, citing a fragment from  Cicero's De Fato preserved in Servius ad Vergil. Aeneid. iii 376 (= Loeb, De Fato, 246), in which fate is defined: "Fate is the interconnection of events that alternate  continuously throughout eternity, varying according to a law and of its own and an order of its own, yet in such a manner that this variation is itself eternal." Here the  Stoics seem to seek, as they so often do, to accommodate varying appearances to an underlying unity. Their method, in this case, is in part the Aristotelian one, of  finding invariance in the causal patterns of nature. But there is also an appeal to the notion that change itself involves no real departure from the fixed nature of things:  Not only is there a fixed pattern but a fixed rhythm and indeed a constancy that negates the appearance of change. Somehow eternal recurrence, instantiating the  underlying pattern on infinite occasions throughout eternity, is seen as a more comforting or less jarring notion here than is Plato's thought that the moving image of  eternity allows just one, fleeting appearance to each wraith that eternity casts up upon the screen of time. See also White's, "An S5 Diodorean Modal System,"  Logique et Analyse 22 (1979): 477–87. One of the strengths of Rescher's account of the Master Argument, "A Version of the 'Master Argument' of Diodorus,"  Journal of Philosophy 63 (1966): 439, n. 4, is that the tense logic it invokes relies ultimately on "a realization operator.'' Such an operator is much in the spirit of the  Stoic notion of an eternal destiny that is, in one way or another, instantiated or brought to the fore moment by moment. 49. Jaakko Hintikka, "The Once and Future Sea Fight," Philosophical Review 73 (1964): 461–92; 466. 50. Sextus, Against the Logicians II 115–17, Loeb, vol. 2, 298–99. 51. See Nicholas Rescher, "An Interpretation of Aristotle's Doctrine of Future Contingency and Excluded Middle," Studies, ch. 5, which follows al­Farabi's reading. 52. See Parmenides, Fragment 8, ap. Simplicius, in Phys. 78.5.

  

Page 39

Chapter Three—  Dialectic and Method in Aristotle Robin Smith In his 1961 paper "Tithenai ta Phainomena," 1  G. E. L. Owen addressed the problem of the relationship between science as preached in the Analytics and the  practice of the Aristotelian treatises. However, he gave this venerable crux a novel twist by focusing on a different aspect of the issue. According to the Prior  Analytics, it appears that the first premises of scientific demonstrations must be obtained from collections (historiai) of facts derived from empirical observation.  However, many of the treatises seem to make little use of empirical inquiry and instead concern themselves more with 'conceptual analysis'. This is especially true in the  Metaphysics and the ethical treatises, but it is also very much characteristic of the Physics. How are these two kinds of inquiry related? Owen took his cue from Aristotle's customary procedure of beginning his account of a subject with a diaporia, a survey of the available data and the views of others  (including his predecessors) in which he notes the 'puzzles' (aporiai), i.e., inconsistencies and paradoxical consequences that result from them. Owen's unifying  proposal was that both empirical data and the opinions of others can be described as 'appearances' (phainomena). This is plausible once we recognize that there is a  crucial ambiguity in this term: not only 'what is apparent' (data of empirical observation) but also what 'appears to' people (the opinions of people, at least those with  some level of general acceptance or philosophical currency) can be called 'appearances'. It is then possible to describe Aristotle's overall method in philosophical  inquiry as beginning with the appearances and undertaking to resolve the puzzles, while retaining as many of those appearances as possible. Moreover—and  crucially—Owen was able to assimilate Aristotle's philosophical

  

Page 40

method to dialectic, on a certain understanding of that term. If, as he supposed, dialectic is "argument from endoxa," where endoxa in turn are commonly held  opinions, then it is to be expected that some form of dialectic will be the source of the first premises of scientific demonstrations. This article, like so much of Owen's work, initiated a wealth of further studies. It has now become a commonplace among many interpreters that Aristotle's method of  inquiry was dialectical, where that in turn means roughly what Owen took it to mean. Terence Irwin, in particular, has developed a modified form of Owen's view in  extensive detail. 2 To confront the problem how argument from 'common beliefs', which are not necessarily all true, can establish the most secure of principles, Irwin first distinguishes  between 'pure dialectic', which is simply argument from the opinions of 'fairly reflective people after some reflexion', and a more critical type of argument he calls  'strong dialectic'. The latter, according to Irwin, emerges in the Metaphysics and elsewhere as a way of securing the otherwise indemonstrable first principles on which  scientific demonstrations rest. Irwin maintains that Aristotle came to see the inadequacy of the appeal to 'intuition' for the justification of these first premises which he  had made in Posterior Analytics II.19 and sought its replacement in a reconsidered form of dialectical proof. Irwin's defense of this picture rests in part on detailed analyses of the argumentative procedures of the treatises. To challenge these would require a comparably  detailed study of these analyses. I do not claim to offer anything like that here. Instead, I want to concentrate on the more direct textual basis for the view that Aristotle  thought dialectic, defined as a technique of arguing from a special class of premises called endoxa, could provide a form of justification of the first principles of  sciences. Generally speaking, that evidence can be divided into two parts. First, there are several proof texts for certain critical theses. Second, there is a certain  picture of endoxa, the 'common beliefs' on which dialectical arguments rest. I would like to raise some problems for each of these. My claims are as follows. (1) If we  take account of their contexts and what they actually say, the proof texts turn out simply not to support the claims which have been built on them. (2) The interpretation  of the endoxa of dialectic rests on a serious misunderstanding both of what dialectical argument is and what the goal of the Topics is. In making my limited case, I will  return to the fons et origo, Owen's article, and in particular to the proof texts on which it rests. I—  'This in Itself Is Sufficient Proof' The fundamental supporting texts for Owen's view are Nicomachean Ethics VII.1 1145b6–7, and Physics IV.4 210b32–211a7, which by his account both contain  ringing declarations of a dialectical method for establishing principles:

  

Page 41 For if the difficulties are resolved and the endoxa are left standing', as Aristotle says in both the Physics and the Ethics, 'this in itself is a sufficient proof.' (244)

The clear implication of Owen's quoted paraphrase is that the Physics and the Ethics advocate a common method and that the method in question consists entirely of  "resolving the difficulties and preserving the endoxa." I do not think that either passage actually supports that interpretation. If we look carefully at them in their  contexts, we will find that the two are making distinct claims, each more modest than what Owen imputes to them. Let me begin with the Physics passage. Aristotle says: We must try to conduct our search so that the essence will be given in such a way as to solve the puzzles, and what appears to be true of place will be true of it, and moreover the  cause of resistance 3 (duskolia) and puzzles about it will be evident. For this is how anything might be most beautifully shown (houtô gar an kallista deiknuoito hekaston.4 (211a7–11)

Owen presumably is offering us paraphrase, not translation, but still there is a poor fit between his account and what this passage says. The former leads us to believe  that Aristotle is stating minimum sufficient conditions for establishing something: 'this in itself is sufficient proof. However, Aristotle says that he is giving conditions for  the 'finest' or 'most beautiful' way to prove anything. These are counsels of perfection, not minimal conditions of adequacy; they go far beyond mere sufficiency to tell  us what the best of all possible outcomes is. Moreover, he does not tell us that this is how we must conduct our search if it is to issue in proof­he says that this is how  we ought to try to conduct it. It is consistent with this demand that an adequate proof may fail to achieve some of these desiderata. He does not say that this is how  we may prove, or how we must prove, but rather that this is how we should try to conduct an inquiry. Aristotle's point is that if indeed we can not only solve the  puzzles and leave standing what is thought to be true, but also explain what causes the difficulties in the first place and why people have problems with them, then we  will have the finest proof imaginable. Nothing that he says implies that these are necessary conditions for an inquiry, or even that they are attainable in every case.  Since elsewhere (for instance Metaphysics A, NE X) he clearly thinks that sometimes our pre­philosophical opinions cannot be retained after philosophical inquiry,  we ought to suppose that the outcome envisaged here will only be possible in some most happy sets of circumstances. Aristotle also includes a third requirement: we should try to make evident 'the cause of the resistance and puzzles about it'. A passage from the Eudemian Ethics  elaborates this same requirement more fully. Since it has often been seen as providing an epistemological reason for attaching weight to the opinions of 'the many and  the wise', I will quote it in full: We must try, by argument, to reach a convincing conclusion (zêtein tên pistin) on all these questions, using, as testimony and by way of example, what appears to be the case.  For it would be best if everyone should turn out to agree with (phainesthai

  

Page 42 sunomologountas) what we are going to say; if not that, that they should all agree in a way and will agree after a change of mind (hoper metabibazomenoi poiêsousin); for each  man has something of his own to contribute to the finding of the truth (echei gar hekastos oikeion ti pros tên alêtheian); and it is from such [starting­points] that we must  demonstrate: beginning with things that are correctly said, but not clearly, as we proceed we shall come to express them clearly, with what is more perspicuous at each stage  superseding what is customarily expressed in a confused fashion. 5(1216b26–35)

I have cited Woods's translation6  here, though I shall quarrel with it in a moment. Now, there is a clear resemblance between this passage and Physics 211a7–11 on  several points. Again, Aristotle says that we should try to achieve a certain result, not that we must achieve it. Next, he indicates different degrees of success at which  we may aim: best of all if everyone 'turns out' to agree with our conclusions, second best if they agree 'in a way'. However, he now adds an explanation why we should  aim at this: everyone has 'something of his own in relation to the truth' (oikeion ti pros tên alêtheian), and it is from this that demonstrations must start. Thus, as it  seems, Aristotle holds that everyone has a certain built­in grasp of a little of the truth, and we should therefore treat the opinions of the many, as well as those of the  wise, with respect, in need of correction and clarification rather than refutation and rejection. Aristotle is not simply advocating a dialectical method in philosophy but  giving us a reason for doing so. However, appearances cannot always be trusted. There are clues in the language of the passage that he has something quite different in mind. One such clue is the  word metabibazomenoi: 'after a change of mind', in Woods's version. This is not a very common word. Aristotle uses it only here and in two places in the Topics,  each of which sheds important light on our present passage. First, in Topics VIII, he is discussing the use of false as well as true premises in arguing dialectically. For  Aristotle, one thing which distinguishes dialectical arguments from 'contentious' or 'sophistical' ones is that they must be valid. However, their premises may sometimes  be only apparently true. Indeed, in some cases they must be false: if the task at hand is to argue for a false conclusion, then false premises must be used (otherwise the  argument would be invalid and therefore not dialectical). Aristotle goes on to note that it can also be dialectically appropriate to use false premises in establishing a  truth or refuting a falsehood: Sometimes, even if a falsehood has been supposed, it should be refuted by means of falsehoods. For nothing prevents things which are not so seeming more so to some  individual than what is true, so that if the argument arises from what seems so to that person, he will be more effectively persuaded or benefited. And whoever changes minds  (metabibazonta) well must change them dialectically, not contentiously (just as the geometer must do so geometrically), no matter whether the conclusion drawn is false or true.7 (161a30–36)

The point is that dialectical arguments are always directed at someone and rely on that person's opinions. If my goal is to persuade you, it will do me no good to use

  

Page 43

true premises in which you do not believe; I would be better off, in fact, using false ones you did believe, so long as they led to the result I wanted. To 'change minds'  is to lead people to have different beliefs, and that can only be accomplished rationally by beginning with beliefs they actually do have. Aristotle says much the same thing about 'changing minds' and dialectical argument in his remaining use of metabibazein, near the beginning of the Topics: [Our dialectical method is useful] in connection with encounters, because if we have reckoned up the opinions of the many, we will speak to them not from foreign opinions but  from their own, changing their minds about anything they do not seem to us to have said well. 8(101a30–34)

I take 'encounters' here to mean simply any occasion for argument with the public. A compilation of the opinions of others, as we shall see below, is one of the  components of the dialectical method it is the purpose of the Topics to present. Again, Aristotle's point is that we can only change people's minds argumentatively from  opinions they actually accept. These passages should lead us to see the EE passage in a different light. Getting everyone to agree with our view after a 'change of mind' means leading each person,  from premises he accepts, to accept our view. This is intrinsic to the persuasive function of argument. In each of our individual collections of opinions there is bound  to be something true—'something of our own', oikeion ti, with some relation to the truth—and this is the starting­point from which others can persuade us to believe  the truth. (Woods's translation obscures this by paraphrasing oikeion ti pros tên alêtheian as 'something of his own to contribute to the finding of the truth'.  Aristotle says nothing about finding or contributing to a search for the truth.) In short, then, what Aristotle is talking about in the EE passage—and in the Physics as  well—is not discovering the truth but persuading others to believe it. EN VII. I comes much closer to presenting a general methodology for establishing first principles. Irwin9  translates it as follows: As in the other cases we must set out (tithentas) the appearances, and first of all go through the puzzles. In this way we must prove the common beliefs about these ways of being  affected—ideally, all the common beliefs, but if not all, then most of them, and the most important. For if the objections (ta duscherê) are solved, and the common beliefs are left, it  will be an adequate proof.10(1145b3–8)

In contrast to the Physics passage, Aristotle is clearly talking about what is sufficient as proof here. Even so, we do not find anything quite so strong as Owen's 'this in  itself is sufficient proof'. In fact, it is difficult to be certain that Aristotle is stating anything like a general condition of adequacy for all proofs. If instead he is pointing to  something specific about the case at hand, then his intent may be quite the opposite: he may be saying 'if we can accomplish this much, it will at any rate be enough'. The discussion above of the Physics and Eudemian Ethics passages gives this possibility more substance. To begin with, note where this passage occurs. It is not

  

Page 44

at the beginning of a treatise but rather prefaces a section devoted to an unusual topic: the possibility of weakness of will. One unusual element is that the conclusion  Aristotle will ultimately defend in this case includes a rejection of a commonly held belief: like Socrates, Aristotle holds that strictly speaking no one can possibly act  incontinently. It is striking that this passage, which has come to be a sort of locus classicus for finding a dialectical methodology in Aristotle, is the prelude to the  rejection of a very widely held belief. However, if instead Aristotle is concerned with 'changing minds', then its occurrence here is especially apt. It is prima facie  implausible that there is no such thing as weakness of will; if we are ever to persuade others of this, we must begin from their own views. However, if we can eliminate  the difficulties (duscherê) that stand in the way of their accepting it, then we will have shown them adequately (dedeigmenon an eiê hikanôs) that it is so. I do not want to claim that these considerations absolutely rule out Owen's interpretation of EN 1145b3–8. In isolation, the passage does strongly appear to be  advocating a general method of inquiry. However, as we have seen, elsewhere Aristotle uses similar language when his real concern is with persuasion rather than  proof. I conclude, therefore, that there is no necessity for Owen's interpretation, and good reasons to reject it. None of this denies that Aristotle regularly makes use of surveys of the puzzles inherent in the views of his predecessors and of people in general. Indeed, he uses  these surveys as a guide to the development of his own position: the puzzles set questions to be answered, and good answers are those which account for all the  puzzles. What is wanted, however, is some evidence that this type of resolution of puzzles constitutes a proof. II—  Dialectical Argument and the Art of Dialectic Underlying these attempts to find a dialectical method of proof in Aristotle is a conception of what dialectic is; and here, in my opinion, is where the real trouble lies.  Interpreters have generally accepted an account of dialectic which places all its emphasis on the curious compilation of opinions called endoxa—the views of 'the  many or the wise', to give it its briefest of several disjunctive formulations. Precisely why Aristotle should single out this collection of opinions for attention is as vexing a  question as how he thinks the principles of sciences are established. Following Owen, many interpreters have tried to link the two, not only finding in dialectic a source  for the principles but also finding in this function of dialectic a partial explanation of the attention Aristotle pays to the endoxa. Irwin has developed a particularly rich  interpretation along these lines. Since the picture of dialectic it includes has been influential, I will turn my attention to it. Irwin finds a definition of dialectic in the first sentence of the Topics:

  

Page 45 Dialectic is 'a method from which we will be able to syllogize from common beliefs (endoxa) about every topic proposed to us, and will say nothing conflicting when we give an  account ourselves'. (36)

But in 100a18–20, Aristotle does not say that he is giving us a definition of dialectic at all. Instead, the words Irwin quotes are prefaced by 'The goal of our treatise is  to find a method'. Is Aristotle then claiming that in the Topics he means to discover dialectic? The answer, as I have argued elsewhere, is that there is a difference  between dialectical argument and a dialectical method or art. Dialectical argument existed long before Aristotle; he himself credits Zeno of Elea with its discovery.  What he offers in the Topics is an art of dialectic, to stand in the same relationship to dialectical argument as does the art of rhetoric to orations. Failure to recognize this distinction is, I think, the major cause of misunderstanding Aristotle's remarks about endoxa. In the Topics, he defines a dialectical deduction  (sullogismos) as one with premises which are endoxa, as opposed to the 'true and primary' premises characteristic of demonstrations. If we concentrate only on this  point, we may wonder exactly what this class of endoxa is. However, in other places Aristotle notes another equally important difference between dialectical  arguments and demonstrations: the premises of dialectical arguments are questions (An. Pr. 24a25, Top. 104a8). What differentiates dialectical arguments from  demonstrations is that there are two parties to a dialectical argument, one of whom presents the argument to the other as a series of questions held out for acceptance  or rejection. 11 In demonstration, one chooses as premises the true and primary propositions which underlie all further truths about the subject matter at hand; no  audience is necessary, and it is the task of the learner not to question but to accept. However, in a dialectical argument, the questioner can only develop an argument  on the basis of an answerer's responses. Consequently, the questioner must take account of the opinions of that answerer and whether the premises needed are  acceptable (endoxos) to the answerer. In the Topics, it is Aristotle's purpose to develop a method that can be used in any such dialectical context. He pays particular attention, especially in Book VIII, to a  stylized form of debate that apparently was practiced in the Academy (he calls this 'gymnastic'). But he does not lose sight of the wider applicability of his method to  any context in which arguments must be developed that depend on the opinions (or at least responses) of others. This, for instance, is the reason why he tells us in the  Rhetoric that the rhetorical art is a kind of hybrid of the dialectical art (dialektikê) and ethics: orations are directed at individuals, and we must understand the opinions  they have in order to persuade them. Now, Aristotle makes it clear that a dialectical argument, as he understands it, must be a valid argument (sullogismos), that is, its conclusion must actually be entailed  by its premises. This sets one task for an art of dialectic: we need a method for discovering premises that imply a given conclusion. But since dialectical arguments must  be constructed from the concessions of a respondent, premises will be useful to us only if they are accepted by our interlocutor. This sets a second task

  

Page 46

for the dialectical art: we need a method for determining, or at least predicting, which premises our respondent will accept. The first task can be solved if we have  something like a theory of validity, or at least some general rules relating conclusions to premises. The second task would be solved if we had a systematic  classification of premises according to the type of person who will find them acceptable. This, in my opinion, is exactly what Aristotle means by the endoxa. The  dialectician is to collect views of types of person—the views of everyone, of the many, of the wise, of various celebrated wise men—and use them to gauge the  acceptability of premises to a particular opponent. The clearest statement of this project comes, not from the Topics, but from the Rhetoric: For since what is persuasive is persuasive to someone (and sometimes is directly persuasive and convincing through itself, sometimes because of appearing to be proved through  such things), but no art investigates the particular (e.g., medicine does not investigate what is healthful for Socrates or Callias, but rather what is so for this type or these types of  person—for this is artful, but the individual is infinite (apeiron) and not knowable (epistêton), then neither will rhetoric study what is individually acceptable (to kath' hekaston  endoxon), e.g., to Socrates or to Hippias, but rather what is so [sc. Endoxon] to such­and­such people (tois toioisdi), just like dialectic. 12(1356b28–35)

This is a familiar Aristotelian thesis: art and science are of the universal, not the individual. It is only accidentally that medicine studies what is healthful for Socrates.  Likewise, dialectic, if it is to be an art which studies what is apparent, must study what is apparent to types of person, not to this or that individual except incidentally.  Against this background, consider the statement often quoted as giving a definition of endoxa: The endoxa are what seems so to everyone, or to most people, or to the wise (and of them, to all, or to most, or to the most famous and best accepted).13(100b21–23)

The great majority of interpreters, regarding this as defining a single class of endoxa, then are much exercised to understand what that class might be and why Aristotle  assigns it such importance. But compare the definition of 'dialectical premise' Aristotle gives a little further on in Top. I.10: A dialectical premise is a question that is acceptable (endoxos) to everyone, or to most people, or to the wise (and to all of them, or most of them, or to the most famous).14(104a8– 10)

Here, Aristotle qualifies the term endoxos with almost the very same words used in 100b21–23. If that earlier passage is a definition of endoxos, then this one is, in  Aristotle's terms, an example of dis tauto legein. It seems much more likely that the first passage is not a definition but a clarifying enumeration. That is, Aristotle says:  the premises of a dialectical argument must be acceptable; by 'acceptable' I mean what seems so to everyone, or what seems so to most people, or what seems

  

Page 47

so to the wise. . . . We find some confirmation of this in the continuation of the passage: . . .  so long as it is not paradoxical: for someone will concede 15 (theitê) what seems so to the wise, if it is not contrary to the opinions of the many.16(104a10–12)

In other words, Aristotle gives as a reason for including the opinions of the wise among endoxa the fact that people will usually accept these (with an appropriate  citation of authority) if they are not 'paradoxical'. Once we realize that dialectical argument and the dialectical art are distinct, the mystery about endoxa completely disappears. Aristotle's entire purpose is to spell out  an art for arguing successfully with other people on the basis of their opinions. Part of what that art must include is a study of the opinions of various types of person.  The endoxa of the dialectical art are simply lists of opinions, categorized in this way. III—  The Starting Points of Dialectic: 'Not Everyone's Opinions Count Equally' This interpretation is at variance with the understanding Irwin and others have of the endoxa. Supposing that Aristotle is trying (for unclear reasons) to work out a  specialized kind of argumentative method, relying on a special class of opinions, Irwin seeks an account of just what those opinions are. He concludes that in fact,  Aristotle wants to restrict the endoxa to the opinions of 'fairly reflective people after some reflexion' (38). In defense of this, he adduces two passages. The first is  from On Sophistical Refutations: the common beliefs of dialectic must be apparent, but apparent not to just anyone, but to people of a certain sort (tois toioisde);  for it is an indefinitely long task to examine the things that make something apparent to just anyone17 (170b6–8). The phrase 'people of a certain sort' strongly suggests that the point here is to restrict the relevant opinions to the opinions of a certain class of people. However,  Aristotle's own explanation of this limitation already undercuts Irwin's interpretation. If his purpose were to say that dialectical arguments rest on the common opinions  of a specific class of persons, we would expect him to say something like 'for there is no point bothering with the opinions of certain types of people'. What he actually  says, however, is that the task of determining 'the things that make something apparent to just anyone' is too indeterminate to pursue—'an indefinitely long task', in  Irwin's rendering. And as we have just seen, almost exactly the same phrase (tois toioisdi) is found in Rhetoric 1356b35, where it clearly is intended to indicate types  of person as opposed to individuals. To find an explicit declaration that 'not everyone's opinions count equally', Irwin turns instead to Eudemian Ethics 1214b28–1215a2, where Aristotle says that we  should not waste our time examining the opinions of children, the ill, and the

  

Page 48

mad about happiness, nor indeed should we give any special weight to the opinions of 'the many', since 'they speak haphazardly about practically everything, and  especially about happiness'. Now, this is already a remarkable passage from which to seek support for a view of dialectical argument as resting on, and giving initial weight to, the opinions of 'the  many or the wise', since it amounts to a global dismissal of the opinions of the many concerning 'practically anything'. Furthermore, its entire weight, with respect to  Irwin's thesis, is negative, showing only that Aristotle thought it 'a waste of time' (periergon) to examine what most people think. If the passage is to provide evidence  that Aristotle thinks there is a certain class of persons whose opinions do count, we should expect it to say who they are. Instead, Aristotle says: But, as each inquiry has its own problems, so, evidently, does that concerning the best and highest life. It is these opinions, then, that it is right 18 (kalôs echei) for us to  investigate (exetazein); for the refutation of those who dispute a certain position is a demonstration of the opposing view.19(1215a3–7, Woods's translation)

Aristotle has said earlier that argument about ethical matters is not what is wanted in the case of the immature, the unsound, and the immoral: what they need instead  is experience, therapy, or punishment, as appropriate. His point here is quite precise. Argument serves to change the opinions of others by taking their own opinions as  its starting point and showing them that these opinions entail other views; if those further views are repugnant to them, they are thereby motivated to change their  opinions. Those whose opinions are subject to rational modification in this light are candidates for argumentative persuasion, and to that end it is useful to study their  opinions and the arguments that can be constructed from them. However, children, the insane, and the wicked lack the opinions from which rational persuasion might  begin; it is pointless to consider how to refute their views, since what causes people of these classes to have their opinions is not argument but some form of pathology.  Aristotle takes part of this to be obvious: we cannot make a child an adult by argument, nor can we heal the sick (or convert the wicked, for that matter). The  rhetorical strategy of the passage is a comparison of these cases with 'the many': we cannot change their views by argument because they do not hold their views for  reasons in the first place but only 'speak haphazardly' (eikêi legousi) about pretty much everything. Therefore, there is no more point in trying to argue with them than  with children or the insane, and we may forgo discussion of their views.20 In summary, the EE passage does indeed say that only the views of certain people count, but that passage is not making anything approaching a general claim about  either philosophical or dialectical argument. Aristotle is giving a narrowly focused reason for ignoring the views of the mass of humanity about what life is best: they do  not have reasons for their opinions, so we need not worry about refuting them. Nothing about the context licenses the elevation of this into a

  

Page 49

general methodological principle. There is no warrant for using it to explicate SE 170b6–8. Moreover, if we look carefully at the context of the SE passage, it becomes clear that Irwin's translation fails to capture an important implication of the term 'indefinite'.  Aristotle's purpose in the entire passage is to explain which refutations his treatise can legitimately study and which it cannot. He begins by stating a distinction, familiar  from other works (including the Rhetoric as well as the Topics), between arguments that fall under the purview of dialectic and arguments proper to one of the  sciences. Refutations are arguments (sullogismoi); therefore, if we know what kinds of things (par' hoposa) arguments arise from—i.e., what kinds of premises—then  we will thereby know what refutations arise from. But arguments are classified according to their premises: some are 'in accordance with a particular art', i.e., rest on  premises peculiar to that art, whereas others are general. It is only the latter that fall under dialectic. As a result, says Aristotle, to undertake a completely general study  of how all refutations come about would require having a science of everything, which is not the task of any single science: For the sciences are likely infinite in number (apeiroi), and consequently so are demonstrations. But these are refutations, and true ones: for whenever something can be  demonstrated, it is also possible to refute one who accepts the denial of this truth. For instance, if someone accepts that the diameter is commensurate, someone could refute him  with a demonstration that it is incommensurable. Thus, we will have to be scientists about everything. 21

If every refutation corresponded to a demonstration, and if every demonstration were proper to some science, then there could be no such thing as the study of  refutations, except in a Pickwickian sense: studying refutations in general would require scientific omniscience. However, Aristotle holds that there are some refutations  that depend only on certain 'common' premises which are not peculiar to any given science. He sees these as being the special province of dialectic and its cousin,  rhetoric. Aristotle continues: Clearly, then, it is not the topoi of all refutations that are to be grasped, only of those that arise from dialectic: for these are common to every art and faculty. Moreover, the study  of a refutation in accordance with a specific science is for the person who possesses that science, both as to whether it appears to be one but is not and, if it is one, why it is. But a  refutation from common [premises], which fall under no science, is for the dialectical to study. For if we have what [topoi] the accepted deductions (endaxoi sullogismoi) about  something are from, then we also have what [topoi] the refutations are from: for a refutation is a deduction of the contradictory.22

The point Aristotle makes here is crucial to his understanding of dialectical argument. In the logical works and the Rhetoric, (1358a2–35) he differentiates the  arguments, premises, and refutations proper to the individual sciences from the 'common' ones applicable to all sciences. The latter, precisely because they are

  

Page 50

common, cannot serve as the basis of any kind of scientific proof: each science is autonomous with respect to demonstrations about its subject matter, and there is no  single overarching science embracing them all. Here, Aristotle draws from this the corollary that, since a refutation is just a kind of deduction, the theory of refutations  that concern any subject falls under the science of that subject, not a general science of refutations. Having made all these distinctions, Aristotle then reaches the conclusion he wants: the general study of refutations applicable to all subject matters is part of the  dialectician's art. In the course of stating this conclusion, he makes in passing the remark Irwin takes to be restricting the starting points of dialectic: Thus, we possess the kinds of [premises] which such refutations are from. And if we have that, then we also have their solutions: for the objections to these are solutions. And we  also have what apparent refutations are from (but apparent not to just anyone, but to people of a certain sort: for they would be indefinite if someone were to inquire from how  many [kinds] they appear [to refute] to just anyone). Thus, it is evident that it is for the dialectician to be able to grasp from what sorts of premises either a real or an apparent  refutation arises through the common [topoi] ... 23

It should now be obvious that this phrase actually has a completely different purpose from the one Irwin ascribes to it. Aristotle is not talking about apparent  propositions but about apparent deductions, since his overall goal is to argue that there is a class of refutations (real and apparent) which fall under the scope of  dialectic rather than any special science. At the very least, it is a distortion of emphasis to wrest this qualification out of its context and see in it a defining characteristic  of dialectical argument itself. It appears, then, that neither of these passages will bear the interpretive weight Irwin requires of it. The SE passage simply does not mean what he takes it to, and the  EE passage is not about dialectic. Neither provides any evidence that the starting points of dialectical arguments are 'the opinions of fairly reflective people after some  reflexion'. IV—  Dialectic and the Route to the Principles Let me now turn to what is perhaps the most crucial text linking dialectic with the indemonstrable first premises of scientific demonstrations. In Topics I.2, Aristotle  gives three venues in which dialectic will be useful; 'gymnastic' exercises, 'encounters', and 'the philosophical sciences'. I have alluded to the first of these briefly above  and discussed his remarks about the second. The third appears to be of much more moment. Here is how Irwin translates it (36): It is useful for the philosophical sciences, because if we fully examine the puzzles on each side (diaporêsai), we will more easily see what is true or false. And it is also

  

Page 51 useful for [finding] the first principles of each science. For we cannot say anything about them from the proper first principles of the science in question, since the first principles  are prior to everything else. Hence it is necessary to discuss them through the common beliefs on each subject. And this is proper to dialectic alone, or to it more than to anything  else; for since it examines, it has a road towards the first principles of all disciplines. 24

Owen takes this to be an explicit declaration that dialectic 'establishes' the first principles of the sciences (244). Irwin, more circumspectly, repeats Aristotle's phrase  that dialectic 'has a road' to the first principles. In either case, it appears that something much more is being imputed to dialectic than could possibly be accomplished  by the art as I have described it. However, context once again turns out to be important. Is Aristotle talking about dialectical argument itself, or is he talking about the dialectical art? The opening of the  section (101a25–26) leaves no doubt that he means the latter: he refers expressly to the pragmateia, i.e., either to the treatise he is writing or to its contents. These  are uses, then of the dialectical art. Since that art, in turn, includes a study of logical consequence, it is evident at once how it will be useful in connection with  'examining the puzzles' concerning any issue. Skill in deducing the consequences of a position is a natural concomitant of skill in deducing conclusions from an  opponent's opinions. In any event, Aristotle does not say that this will provide us with a proof of anything but only says that it will help us to see what is true or false. It  is consistent with this claim that other means are also required for discovering the principles, and it is consistent with the business of the Topics that Aristotle should  refrain from discussing them here. Aristotle proceeds to give a second use of dialectic in connection with scientific principles, one which looks at first much more promising—especially in Irwin's  translation, which glosses Aristotle's simple 'in connection with' (pros) with 'finding' (though as he notes this is not in Aristotle). We seem to find here an allusion to the  doctrine of the Posterior Analytics that no demonstrative science can demonstrate its own principles. So construed, this passage promises to be of great value, since  it would be speaking to the same point as the enigmatic final chapter of that work. However, Aristotle simply does not say here that dialectic establishes the principles:  he says that since a science cannot 'say anything about' its own principles, we must 'discuss' them (dielthein: Aristotle often uses the word of his preliminary  discussions of the views on a subject) by means of endoxa. But there is a considerable distance between 'discussing' first principles and 'establishing' them. If that distance is to be bridged, the weight must fall on the last sentence of the section, which Irwin and others take to say that dialectic provides us with a road  whereby we reach the first principles. Now, in what sense might we be said to arrive at the first principles? Aristotle's own view on this is clear. The first principles of  the sciences are truths which have a certain objective priority to the conclusions that can be demonstrated from them. This priority does not depend on

  

Page 52

our own epistemic reactions in any way: it is not only possible, but usually true, that we all find these first principles unconvincing or even absurd before we have  acquired scientific understanding. On Aristotle's view, it is by transforming ourselves, so that the objective first principles seem to us to be primary, that we come to  have scientific understanding. There are, consequently, two components to reaching the first principles: (1) finding out what they are, and (2) coming to see them as the  principles. There is considerable evidence in the Analytics that Aristotle thought (1) could be accomplished in a way that sounds quite strange to us. He thought that he had  discovered a theory of inference (the syllogistic) which covered all valid arguments whatsoever, and thus all demonstrations. He also thought—and understandably so,  given the properties of his theory of inference—that there are some true propositions with the unusual property that they cannot be deduced from any other true  propositions, though they may serve as premises from which to derive others. Indeed, he appears to have held the even stronger claim that if we collect together all the  truths about any subject, we will find that there are certain truths among them which cannot be deduced from any combination of the others but from which all the  others can be deduced. These propositions must be the principles for the simple reason that they cannot be anything else: they cannot be demonstrated. On the other  hand, if they are included among the principles, then we need no other principles, for all others follow from them. The upshot is that Aristotle believed there was a way  to specify the principles without appealing to epistemic status. 25 According to the Analytics, then, dialectic is not the means of accomplishing (1). What of (2)? Here, I think, we do discover an important role of dialectic, but it is not  the one Owen supposes it to be. The property of dialectic to which Aristotle appeals here is that it 'examines' (exetastikê gar ousa). The word used here for  'examine' is closely connected with refutation, in particular refuting someone else's views by showing that they lead to contradictions (Socrates used it of his customary  style of questioning people about their opinions). A process of refutation is not a very likely candidate for establishing the first principles. What it might do, however, is bring about a considerable change in our own epistemic situation. The process of exploring the contradictions implicit in our naive  opinions eliminates the air of certainty that attaches to them and puts us in that unpleasant state of 'wonder', which, according to Aristotle, is the beginning of  philosophy. A continual process of exploring what follows from what could plausibly be essential to the kind of epistemic conversion required if we are to become  scientific. Aristotle tells us in Met.   that scientific education resembles moral education: we strive in each case to change our untutored reactions, making what is in  itself good (or familiar) come to seem good (or familiar) to us. In ethics, the agent of this transformation is habit, born of repeated action. Repeated examinations of  opinions and their consequences, and of the principles and their consequences, could be the agent of a similar epistemic transformation.

  

Page 53

It might be objected that Met.   offers precisely to 'demonstrate by refutation' (apodeixai elenktikôs) the principle of non­contradiction, the 'firmest' of all principles:  since that procedure, however we interpret it, clearly involves a dialogue between two interlocutors, we might think that this is an actual example of dialectical  argument establishing what demonstration cannot. An adequate response, I believe, is Alan Code's argument 26 that Aristotle's goal in Met.   is neither proving the  principle of non–contradiction itself nor providing arguments to persuade anyone of its truth, but rather proving things about the principle. In fact, this fits exactly with  Aristotle's claim about dialectical argument in Topics I.2. Because it is, as Aristotle puts it, 'impossible say anything about' the principles (adunaton eipein ti peri  autôn) on the basis of the various sciences that we must turn to dialectic in order to 'discuss' them (peri autôn dielthein). V—  Conclusions Most of what I have had to offer in this chapter is critical. My main goal has been to show that arguments for a probative role for dialectic in connection with  establishing first principles rest on misreadings, sometimes serious ones, of the text. The more valuable project, which I have barely touched in here, is understanding  just what positive roles dialectical argument might have for Aristotle in the scientific enterprise. However, that must wait for another occasion.27 Notes 1. In Aristote et les problèmes de méthode, S. Mansion, ed. (Louvain, 1961), 83­103. Reprinted in Owen, Logic, Science and Dialectic, Martha Nussbaum, ed.  (Ithaca: Cornell University Press, 1986). I cite page numbers from this edition. 2. Terence Irwin, Aristotle's First Principles (Oxford: Clarendon Press, 1988). 3.   has a technical sense, roughly 'impede the progress of the argument by improper refusals to concede premises'. It is a fault of answerers, not  questioners; in my translation, I render it as 'be cantankerous.' Cf. Menander's Duskolos ('The Crotchety Old Man' or 'The Curmudgeon'); for an example of what  Aristotle means, I suggest Thrasymachus in Rep. I.

  

Page 54

6. Aristotle's Eudemian Ethics Books I, II and VIII, tr, with commentary by Michael Woods (Oxford: Claredon Aristotle Series, 1982).

9. Aristotle, Nicomachean Ethics, tr. Terence Irwan (Hackett, 1985).

11. This is exactly what the term 

15. The use of Greek 

 'something held out', would suggest.

 to mean 'concede as a premise' is ubiquitous in the Topics.

18. Through i follow Woods here, I think his translates fails to capture an important element of this passage. The word he translates 'investigate' is   which can  carry the semi­tecnical sence 'subject to a Socratic examination'. It is not common in Aristotle, and when it does ocur it almost always is in a context implying  refutation. 'These are the opinions it is in order for us to refute' would be better: the point is that it is a waste of time to refute people who are incapable of reasoning.

20. The remark that 'the refutation of those who dispute a certain position is a demonstration of the opposing view' in no way supports the notion that there is a special  kind of dialectical method for establishing first principles. Aristotle says many times that a refutation of a proposition is simply a deduction (sullogismos) which has its  conclusion the denial of that proposition. It is a simple mater of definition that a refutation of p is a 

  

Page 55

demonstration of 'the opposite view', i.e., the negation of p: to refute is to prove false.

25. See my ''What Use Is Aristotle's Organon?" Proceedings of the Boston Area Colloquium in Acient Philosophy 9 (1993): 261­285. 26. "Aristotle's Investigation of a Basic logical Principle: Which Science Investigates the Principle of Non­Contradiction?" Canadian Journal of Philosophy 16  (1986): 341­357. 27. A Version of this paper was read at the 1996 APA Central Division meeting in Chicago on April 27,1996. I am grateful to Alan Code, my commentator on that  occasion, for helpful suggestions.

  

Page 57

Chapter Four—  The Epistemological Basis of Aristotelian Dialectic Robert Bolton Just as people judge the value of currencies by reference to the currency most familiar to themselves, so is it in other areas as well.  (History of Animals I.6 491a20–22) Dialectic is peirastic when it concerns matters where philosophy reaches knowledge. (Metaphysics IV.2 1004b25–26)

I—  Conflicting Views of Dialectic In recent accounts of the procedures that Aristotle recommends and uses in order to reach and to justify the results presented in his works, reference to his method of  dialectic looms large. This is in marked contrast to the approach to Aristotle's methodology offered by leading scholars only a short time ago. Ross, for example,  summed up his own account of Aristotle's view of the merits of dialectic in the following way: The discussion [of dialectic in the Topics] belongs to a bygone mode of thought; it is one of the last efforts of that movement of the Greek spirit towards a general culture, that  attempt to discuss all manner of subjects without studying their appropriate first principles, which we know as the sophistic movement. What distinguishes Aristotle from the  sophists, at any rate as they are depicted both by

  

Page 58 him and by Plato, is that his motive is to aid his hearers and readers not to win either gain or glory by a false appearance of wisdom, but to discuss questions as sensibly as they  can be discussed without special knowledge. But he has himself shown a better way, the way of science; it is his own Analytics that have made his Topics out of date. 1

Recent writers have taken a view of dialectic quite opposed to the one summarized by Ross. Finding little or no indication that the method for finding and laying out  demonstrations described in the Analytics is actually guiding Aristotle's thought in his philosophical and scientific works, scholars have turned to his descriptions of the  method of dialectic, in the Topics and elsewhere, to find the key to understanding his procedures. The following summary assessments express a widely influential  current estimate of Aristotle's attitude toward dialectic as a tool for not only philosophical but also scientific inquiry: [Aristotle] nowhere suggests that any other method will lead to results which conflict with or go beyond the results achieved by the method of endoxa [i.e., by dialectic]. [Aristotle] establishes science on the basis of the opinions of 'the majority' and 'the wise'. . . . He announces time and again that the way to the truth is through the study of  'reputable' opinions [i.e., through the endoxa and dialectic].2

So in place of the earlier view that the method of the Analytics supersedes and replaces the method of dialectic, the view now more dominant is that whatever other  methodological procedures Aristotle may introduce, none is intended in any way to supersede dialectic as the proper method of scientific or other inquiry, and in  particular, as the proper method to use to discover the first principles of the sciences. By contrast, the method of searching for and setting out demonstrations which is  discussed in the Analytics is commonly taken nowadays to have to do not with genuine discovery or the epistemic justification, which that may involve, but only with  what is required, after dialectical inquiry is completed, either to systematically display the results of inquiry or to impart these results to learners or to deeply  understand these results.3  Other views of dialectic and its relation to demonstration than these two sharply opposed views have been taken. Some have wanted to  distinguish those scientific works in which Aristotle's method is dialectical (including, typically, the scientific works we regard as most philosophical such as the  Physics, De Anima, and Metaphysics) from others in which he uses, or also uses, nondialectical empirical methods. But even the defenders of this approach have  commonly also claimed that for the discovery of first principles in science dialectic in some form is the method which Aristotle recommends and regards as sufficient.4   Generally speaking, then, the two sharply opposed views represent the dominant tendencies in recent scholarship, and it will be helpful to concentrate on them for the  exploration of the epistemological issues to be discussed here.

  

Page 59

II—  What Is Dialectic? A First Approach Before considering these two views of dialectic further, it will be useful to remind ourselves in a general way of what the method of dialectic is, which has been viewed  so differently by two different generations of scholars. A convenient place to begin is with the opening sentence of the Rhetoric. Rhetoric is a partner of dialectic. For both of them are concerned with the sorts of things which it is in a certain way common to everyone to know and which require no specialized  knowledge. Thus everyone in a certain way takes part in both dialectic and rhetoric. For everyone on a limited basis engages in examination and in submitting to argument; and  also in defending himself and accusing others. (1354a1–6)

Here Aristotle points out that dialectic, like rhetoric, is a procedure for argument or reasoning on the basis not of expert knowledge of a given science or other  discipline but rather on the basis of what it is "in a certain way common to everyone to know" (1354a2–3). The remark quoted at the head of this chapter, as well as  the present passage, makes it clear that, in Aristotle's view, reasoning on such a basis is a fundamental part of the rational life of people in general. In the Rhetoric he  explains how this is so. "Everyone on a limited basis engages in examination [of claims] and in submitting to argument [when under examination]" (1354a4–5). We all  regularly test the claims of others in deciding whether to accept them, and submit in turn to the testing of our own claims by others, in the ordinary course of life. But  this mode of reasoning which everyone engages in, and not just the learned, requires a basis for argument and a method of procedure which is suitable for use by and  with everyone. This is dialectic. So just as Aristotle's interest in how the rational life of people in general is best led motivates him to describe and codify the proper  technique of rhetoric, which everyone engages in (1354a5–6), it equally motivates him to do the same for dialectic, which everyone also engages in. It is important to keep in mind what this main social setting is in which dialectic is conceived by Aristotle to have its original home. Sometimes accounts of dialectic  concentrate on the undeniable fact that dialectic came to be practiced in the Academy, and elsewhere, as a quasi sport for the purpose of mental gymnastics or  training. 5  With this in view the dialectical method of the Topics has been described as, quite generally, involving no serious concern for truth. But in Topics I.2  Aristotle distinguishes the gymnastic use of dialectic from two other distinct uses: the universal use, described at the beginning of the Rhetoric, in ordinary discussion or  conversation in order to correct the mistakes of others and submit to the same process ourselves (101a27, 30–34; cf. Rhet. I.1 1355a24–29); and a pair of uses in  science—an aporetic use in raising genuine problems for both sides of an issue so that "we may more easily discern what is true and what is false" (101a34–36), and  an essential use in reaching the first principles of the sciences (101a36ff.). If

  

Page 60

dialectic involves "no serious concern for truth" 6  or, more precisely, if it does not in any of its forms reliably provide us with better grounds for believing what is true  than for believing what is false, then it is hard to see how it could be rational to use dialectic to perform the nongymnastic functions, whether everyday or scientific,  which Aristotle assigns to it. Of course, Aristotle needs to show us that the procedures of dialectic are ones which it is rational for us to employ for purposes other  than mental gymnastics. It will be useful, in our attempt to see how he does this, to begin with a rough description of dialectical procedure. At the beginning of his official treatise on dialectic, the Topics, Aristotle makes it clear that a leading feature of the method is a special procedure for arguing for claims  and for defending claims against objection. Roughly speaking, on this procedure a claim may be said to be "dialectically justified" just in case either it follows in an  appropriate way from items which belong to the existing set of noted or accredited beliefs (endoxa), or it is consistent with (i.e., its contradictory does not follow in an  appropriate way from) items in this set of beliefs (Topics I. 1 100a18–30).7  The noted or accredited beliefs, or endoxa, which make up this set Aristotle limits to "the  things which are accepted by everyone, or by most people; or by the wise—either by all of them, or by most, or by the most famous and distinguished" (Topics I. 1  100b21–3). The noted or accredited beliefs which belong in this collection include not only those which concern some particular subject matter such as physics or  ethics but also those which have to do with what Aristotle calls logic (logikê), that is with the canons or techniques for arguing about questions of physics or ethics or  anything whatever (I.14 105b19ff.). The Topics itself is, in fact, mainly concerned with setting out an accredited list of these canons and techniques of argument. This initial description of the method of dialectic is, intentionally, rough and in various ways imprecise and incomplete; and much will need to be done to improve it. But  the description will do well enough as a starter for us to begin to see how one might categorize the method of dialectic from the epistemological point of view so as to  raise some questions about its potential philosophical interest. As described so far, dialectic might well seem to embody a version of a coherence theory of justification  for claims and beliefs since it takes a claim to be justified just in case it is, in an appropriate way, consistent with or implied by certain standing noted or accredited  beliefs. If so, then if it is the case, as is now widely held, that the method of dialectic is on its own totally adequate and sufficient for Aristotle for the justification of  results in science, in philosophy and elsewhere, then he favors some form of coherence theory of justification in epistemology generally, a line which some people have  in fact recently taken.8  Or, more cautiously, to the extent that Aristotle regards this method as sufficient to that extent it might appear that he favors a coherence  theory of justification. If he does, or to the extent that he does, then it would be of interest to know what form of coherence theory he holds and why he does so; and  also to know how his reasons for adopting a coherence theory may be related to those which lie behind the recent resurgence of theories of this

  

Page 61

type in contemporary epistemology. Is Aristotle's theory, like some current theories, developed in response to the conviction that there are no incorrigible or self­ certifying beliefs, due perhaps to the theory­laden character of all beliefs? Or is it developed, like other current theories, out of the conviction that appeal to what lies  outside what is most fundamental in our customary scheme of accredited beliefs is somehow incoherent? Or does it have some other basis, different from any  contemporary one? III—  Objections to the Sufficiency of Dialectic There is a well­recognized difficulty which needs to be mentioned at the outset which stands in the way of attributing to Aristotle a commitment to the total sufficiency  of the method of dialectic in science, or in some sciences, and to the sort of coherence theory of justification that might be involved in it. Aristotle distinguishes in the  very passage in the Topics where he gives his clearest general description of dialectic a class of propositions—the first principles of the sciences—for which "it is not  necessary to inquire why they are so; each of the first principles is worthy of belief on its own" (Topics I.1 100b20). Justified belief in these propositions, it would  seem, does not ultimately depend on an apprehension of their coherence with any other beliefs or propositions at all. 9  However, some recent writers have argued that  there is a way of understanding this doctrine so that it is compatible with a commitment on Aristotle's part to the sufficiency of dialectic for justification. Others have  argued that Aristotle came to abandon this doctrine when he saw properly the value of a special kind of dialectic.10 In the latter case, it could be argued that Aristotle  at least came to favor a coherence approach in epistemology and the question as to what motivated his change of mind would be of special interest. This is a matter which deserves further discussion.11 But we do not need to settle it here. Even if it is the case that this traditional problem does not in the end stand in  the way of supposing that Aristotle regarded, or came to regard, the method of dialectic as sufficient for discovery and for the justification of belief, in all or in certain  sciences, there are at least two other considerations which make it clear that, for him, the proper practice of scientific method clearly can and does lead to results  which both conflict with and go beyond results reached by dialectic. First, Aristotle's method of discovery and of justification in natural science generally is empirical  in important respects in which dialectic cannot be. Aristotle requires that scientific theories are reached and confirmed ultimately by reference to the data of perceptual  observation by contrast with accredited beliefs or endoxa. (De Caelo III.4 303a20–23 with III.7 306a–17; both quoted shortly below.) In his actual practice in his  scientific works he often appeals to new perceptual data which, as he sometimes explicitly claims, contradict all standing opinion; and he uses reports of observations  which come from sources (such as "some experienced

  

Page 62

fishermen") which fall outside all of the sanctioned sources of endoxa. 12 Put simply, the database which science works and reasons from includes items which do not  count as endoxa. So since dialectic is restricted in its database to endoxa13 the method of science cannot simply be dialectic. This is brought out quite explicitly in a  famous passage from Generation of Animals III.10 (760b27–33) which concludes Aristotle's discussion of how bees reproduce. This is the way things appear to stand concerning the genesis of bees on the basis of reasoned argument and the things which are accepted (ta sumbainein dokounta) about  bees. But the facts have not been adequately ascertained, and if they ever are to be, credence must be given more to perception than to reasoned arguments, and to reasoned  arguments only if what they show is in agreement with the phainomena.

Here Aristotle explicitly cites a case where the agreement of some theory with what is "accepted," which would at least include agreement with the endoxa, is not  enough. Further data he says, are required. And the primary test for the credibility of further data, and of any theorizing based on them, is not whether they become  endoxa or how they fit with the endoxa but whether those data are perceptual phainomena. It cannot be claimed that this requirement is one which Aristotle invokes  only for certain (empirical) sciences and not for others. In the Analytics Aristotle explicitly invokes this requirement for "any art or science whatever" (APr. I.30  46a17ff., quoted below). Secondly, Aristotle makes it clear that proper scientific method must aim to secure explanations in a way that goes beyond what dialectic can achieve. This comes  out most clearly if we consider the requirements for adequate definition prescribed by dialectic. Dialectic requires, and provides a method for reaching, definitions  which are inferable from or at least consistent with the endoxa or an appropriate subset of the endoxa (Topics 100a18–21). The main lines of argument (topoi) for  dealing with definitions are given in Topics VI–VII. But adequate scientific definitions must satisfy a different requirement. They must be capable of explaining the  phainomena. (De Anima I.1 402b16–403a2, Prior Analytics I.30 46a17–22, both quoted below.) Among these data may be items which do not count as endoxa,  for reasons explained above, and the relation required between the data and the definition is stronger than mere derivability or consistency.14 For both of these  reasons the definitions required by science as first principles may conflict with or go beyond those required by or available to dialectic, and dialectic is not sufficient to  guarantee that a definition appropriate for scientific purposes has been discovered.15 Again, this is a point which holds quite generally and not for some sciences as  opposed to others. These two crucial features, each of which separates scientific method from dialectical method (either as a method of justification or as a method of discovery), are the  ones to which Aristotle standardly gives special attention throughout the corpus, in works early and late, in his summary descriptions of proper scientific procedure.  The scientist's task, he repeatedly says, is to collect the full range of

  

Page 63

empirical phainomena on some subject and then look for the theoretical principles (preeminently the definitions) which serve to explain these data. Consider the  following passages: Most of the principles in each science are unique to it. Therefore, it is the role of experience (empeiria) to yield the principles of each science. I mean, for instance, that experience  in astronomy yields the principles of astronomical science, since it was only when the phainomena were adequately grasped that the demonstrations in astronomy [and thus the  principles on which genuine demonstrations must be based] were discovered. The same is true of any art or science whatever. (Prior Analytics I.30 46a17–22; cf. APo. I.13 78b34– 79a6, II.19 100a6–b5) It does not make good sense for it to turn out [as the Platonists claim] that one element alone [earth] has no part in the transformation [of the elements into each other]. Neither is it  apparent on the basis of perception; rather [on this count] all [the elements] change equally into each other. As a result these theorists are offering accounts which concern the  phainomena while their accounts are not in agreement with the phainomena. The reason for this is that they have not proceeded in the proper manner in the positing of their first  principles because they were determined to bring everything into conformity with certain fixed ideas. For surely the first principles which concern sensible things must conform to  the sensible, those which concern eternal things to the eternal, those which concern perishable things to the perishable: in general, principles must be conformable to their  subjects. But because of their affection for their principles they have behaved like people who are set to defend their theses for the sake of argument. For holding their principles  fixed as true they abide any consequence [of them] not seeing that it is necessary to judge some types of principles in the light of their consequences, in particular of what is  ultimate. The ultimate thing in the case of practical knowledge is the product, in the case of natural science it is the always authoritative perceptual phainomenon. (De Caelo III.7  306a3ff.) Knowing the accidents of a thing contributes a great deal towards knowing what the thing is. For when we can give an accounting, at the phenomenal level, of the accidents— either all or most of them—we will be able to speak best about the essence. For a statement of what a thing is the starting point of every demonstration [about that thing], so that  insofar as definitions do not lead us to a knowledge of the accidents, nor even lead us easily to likely proposals about them it is clear that they have been offered dialectically and  are of no value. (De Anima I.1 402b21ff.)

The reader of such remarks and others like them (e.g., GC I.2 316a5–13, HA I.6 491a7ff., GA III.10 760b27–33) might naturally be led, as Ross and his  contemporaries were, to a view of Aristotle's theory of justification in science quite at odds with the currently standard one. In collecting the appropriate range of  empirical phainomena and explaining them (in accordance with the canons of the theory of demonstration) why does the scientist need to pay any attention at all to  endoxa as such? What is important is to collect the proper empirical data and explain them. Whether some or all of these data are endoxa seems irrelevant; as

  

Page 64

does the matter of whether the theories one offers to explain the data cohere with the endoxa. Looking to the body of noted or accredited belief would seem to have,  at most, pragmatic utility, by serving to direct attention to possibly significant data or possibly promising theories; but no capacity to guarantee discovery and no value  as such for justification. What guarantees discovery and what counts for judging the correctness of theories, it would seem, is the genuine empirical status of the data  explained and the genuine explanatory power of the theories. As noted above, Aristotle invokes this requirement not simply for the natural sciences (as in De Caelo  III.7) but (in Prior Analytics I.30) for "any art or science whatever." This view of dialectic as, in the final analysis, inadequate to guarantee discovery and irrelevant for justification in science is, as we have noted, just the view taken by  many scholars in the past before the current trend acquired its prominence. But despite the difficulties for the current view this older view does not at all seem to  account for the role which Aristotle explicitly assigns to the use of endoxa in scientific inquiry. He thinks it is necessary to find theories which accommodate the  endoxa (or at least some appropriate subset of them) as well as perceptual phenomena. In the De Caelo (III.4 303a20–23) he objects to the physical theory of the  Atomists because: In claiming that there are indivisible bodies they cannot avoid conflict with the mathematical sciences nor the denial both of many endoxa and many perceptual phainomena.

So he regards it as a legitimate test of the adequacy of theories in physical science that they fit not only with the perceptual phainomena but also with the endoxa. This  is further confirmed by a well­known passage in the Physics (IV.4 211a7–11). It is necessary to try to conduct our investigation so that it is determined what place is in a manner such that the difficulties are resolved, and the things that are held to belong to  it (ta dokounta) do turn out to belong to it, and, further, it will be clear what the cause is of the trouble and the difficulties about it. For in this way is each thing best established.

Here Aristotle claims it is necessary (though not sufficient) for the best justification to find theoretical principles which account for the "the things that are held." These,  again, must at least include the endoxa, or at least some appropriate consistent subset of them. So even though he gives higher authority in justification to perceptual  phenomena as such, endoxa have value in justification as well. And Aristotle says in Topics I.2 that it is dialectic which is the proper method to use to determine how  theories square with the endoxa (101b1–3). He also claims in this chapter that it is "necessary," and not simply heuristically valuable, to work dialectically through the  endoxa to reach adequate theoretical first principles (101a36ff.). He does not restrict this requirement to some sciences as opposed to others but invokes it quite  generally for "each science" (10135). He does speak of the requirement as applying in "the philosophical sciences" (a34); but this does

  

Page 65

not, as some have suggested, restrict the requirement in a significant way for present purposes. Aristotle counts the scientific study of any real kind as a branch of  "philosophy" (Metaphysics IV.2 1004a2–9; cf. VI.1 1026a18ff.). One might perhaps suppose that the task of "dealing with the first principles," for which dialectic is  essential, is a very small or special part of the scientific enterprise. Even if this were so it would, of course, be important to understand how dialectic can play a  necessary role in this process. But various passages make it clear that this task is coextensive with the scientific enterprise as a whole. In the first chapter of the  Physics for instance, Aristotle identifies scientific inquiry (methodos) with the process of the acquisition of principles, and in Posterior Analytics II.1–2, he  elaborates this point in detail. Aristotle also says that it is "the role of experience to yield the principles ... [This] is true of any art or science whatever" (APr I.30  46a17ff.). That is, Aristotle conceives of the task of collecting and sifting the experiential data as the main element in the process of working toward the principles, for  the obvious reason that the principles are found and validated as the ultimate items that explain this body of data (De An I.1 402b21ff., De Caelo III.7 306a3ff.).  Since these two processes, collecting the empirical data and explaining them, are just the ones in which Aristotle claims to be chiefly engaged quite generally in his  scientific works, if dialectic is necessary for "dealing with the first principles" in every genuine science, then dialectic is a necessary ingredient in the scientific enterprise  quite generally. So even though the new view that dialectic is totally adequate has serious deficiencies we cannot simply go back to the old view that dialectic is or  becomes irrelevant. We need to explain why Aristotle should give special status in scientific method quite generally to endoxa and to the method of dialectic which  reasons from them, in view of the limitations of dialectic as a tool for discovery and justification in science. In particular, we need to understand in this connection  why Aristotle should require the scientist to care at all about what people in general credit in physics or biology, or other sciences. This is a question which has so far received surprisingly little attention, even from those who have regarded Aristotle's method of working through the endoxa as, for  him, totally adequate for justification in science or elsewhere. There are indications in the recent literature of several different approaches which those who view  dialectic as Aristotle's primary method would take in answering the question. But it will be useful, before investigating these approaches, to consider more carefully just  what Aristotle himself says about how to do dialectic. If we can see in more detail just what the basic rules of the method are we will be in a better position to consider  what the value may be of various recent accounts of the ultimate basis of the method.

  

Page 66

IV—  How to Do Dialectic: Some Initial Questions In the opening sentence of the Topics, to which we have already alluded, Aristotle gives a general characterization of the method of dialectic. The purpose of this study is to discover a method whereby we shall be able to reason syllogistically from endoxa about any problem which may present itself and shall be able to  submit to an argument ourselves without saying anything inconsistent. (Topics I.1 100a18–21)

In this opening passage Aristotle introduces certain paradigmatic features of dialectical method. He points out at the end of the chapter that his discussion here, and  later in the Topics as well, is a discussion ''in outline" (tupo), by which he means that it is not "an exact account but only ... one which will enable us to understand each  subject in some fashion" (101a21–4). This provides for us an important general cautionary note, and in particular it implies that the initial description may in some  respects focus only on certain typical or paradigm cases or features of dialectic. It will be necessary to bring certain of these respects to the fore shortly. In this passage Aristotle mentions two features of dialectic. It provides us with a method for "reasoning syllogistically" (sullogidzesthai) and also with a method for  "submitting to an argument ourselves (autoi logon hypechontes) without saying anything inconsistent." As his language suggests, Aristotle conceives of the latter  activity as carried out, in a dialectical question and answer discussion, by the answerer, the one who "submits to an argument," from endoxa, which is constructed on  the basis of premises which are granted by him in response to questions, and attempts to avoid concessions which lead to a conclusion "inconsistent" with what he is  set to defend. In harmony with this, Aristotle typically conceives of the former activity, "reasoning syllogistically," as successfully brought off in a dialectical discussion  by the questioner. 16 This is because success on the part of the answerer in such a discussion involves, where possible, avoiding the concessions which are necessary  for successful syllogistic reasoning, i.e., reasoning to a conclusion which is inconsistent with what the answerer is set to defend (or what he has granted). So in an  important sense such reasoning does not even occur in a dialectical discussion unless the questioner is successful. Dialectical reasoning is reasoning "to the  contradictory" of the answerer's position (Sophistical Refutations I.2 16563–4; cf. Top. VIII.11 161b19–26. Reasoning can still occur to conclusions undamaging to  the answerer, but these are not the aim of the questioner. See Top. VIII.6). There are then, according to the passage above, two primary requirements for the answerer in a dialectical discussion: (1) to submit to reasoning from endoxa and (2)  to avoid inconsistency. But these requirements are, of course, in potential conflict. Suppose the endoxa on the topic under discussion are inconsistent among  themselves. As we have noted, the endoxa include the views of the majority but

  

Page 67

also the views of the experts, and Aristotle is well aware that on topics worth discussion the views of these two groups tend to conflict, both with each other and  among themselves (Top. I.11). Given this, as Aristotle himself says, genuine dialectical arguments can yield inconsistent conclusions (Rhetoric I.1 1355a33–36). If the  endoxa were not inconsistent this would not be possible. So if the answerer is required to submit to any valid reasoning from any endoxa then his task of avoiding  inconsistency would seem to be impossibly difficult. By the same token, the task of the questioner would seem trivially easy. What procedure is to be followed, then, in  case of conflict among the endoxa? Is each party permitted to rely on whatever consistent subset of the endoxa he may choose? Is the maximal consistent set the  appropriate one for both parties? These questions, and others like them, are of special importance for determining what kind of epistemological justification for a given claim is provided (for  questioner or answerer or both) by a piece of satisfactory dialectical reasoning. If any valid reasoning from endoxa is satisfactory, even on a topic where the endoxa  are massively conflicting, then how can dialectical reasoning, whether in the form of a single argument or of an extended chain of connected arguments, reliably provide  much at all by way of genuine justification for the acceptance or rejection of belief? Moreover, how could the simple restriction of premises to a consistent subset of  the endoxa, whether the subset be maximal or chosen at will, improve matters on difficult and important topics where conflict among endoxa is quite substantial? These questions and other related ones are now basically ignored in the literature. In order to try to begin to deal with them it will be useful to simplify somewhat our  description of dialectic. As we have seen, Aristotle's remarks are typically geared to reflect aspects of a live discussion in a question and answer format between two  people. This, of course, conforms to the history of the subject. ("Do you call the one who knows how to ask and answer questions anything except a dialectician?"  Plato, Cratylus 390c.) But this is one of the places where it is important to recall that Aristotle's remarks are not "exact" but only "in outline" and as such not always  universally applicable to all cases but only to certain typical or paradigm cases. He points out himself in Topics VIII.14 (16363–4) that one can do dialectic by  oneself, that one person can in effect play both the role of the questioner and of the answerer and produce those results which would come about if there were two  parties each doing his job properly. He has in mind particularly there the use of dialectic for mental training (163a29–32; cf. I.2 101a28–30) and the use of dialectic  for the scientific purpose of seeing both sides of an issue and thereby being in a better position to discern truth (163b9–12; cf. I.2 101a35–6). But if one can properly  do solo dialectic for those purposes one can also do it for the other scientific purpose of dealing with the first principles, mentioned at I.2 101a36ff. So it will be  sufficient here, in our attempt to understand dialectical justification, to consider the epistemological value of results reached in a way that conforms to the joint  requirements which must be met in a two­party dialectical

  

Page 68

discussion, by both questioner and answerer, whether or not there are two actual persons responsible for those results. This is, of course, important for determining  the role and the value of dialectic in Aristotle's methodological procedures in his scientific treatises since there, though he (like Plato in his uses of dialectic in the  dialogues) is often engaged in debate with various actual or possible opponents, he (again like Plato) is alone responsible for the proper presentation both of his  examination of the opponent and of the presentation of the opponent's position, with whatever resources for defense the latter may involve. We can begin to see how to deal with the questions posed above concerning the nature and the merits of dialectical justification if we compare with the opening  passage of the Topics which provoked us to raise these questions, a passage toward the end of the Sophistical Refutations which alludes back to it. Our purpose was to discover a technique for reasoning about the problem before us starting from things which are as endoxa as possible, since this is the business of dialectic in  its own right and of peirastic. But since, on account of its close affinity to sophistry, it is set up so as to be able not merely to conduct testing dialectically but also as one who  knows, for this reason we not only undertook the just mentioned aim of this study, to discover how to obtain an argument on the basis of what is most endoxon, but also to  discover how we can defend a position, in the course of submitting to an argument, in a similar manner. We have already given the explanation for this; for this was why Socrates  used to ask questions but not to answer them, since he confessed that he did not know anything. (34 183a37ff.)

This new description is, clearly, one which could be of use in answering the main questions raised above. If dialectical reasoning, or a certain type of dialectical  reasoning, must be reasoning not simply from endoxa but from what is most endoxon, or as endoxon as possible, then such reasoning cannot be based on just any  consistent subset of the endoxa (even if these endoxa are agreeable to questioner and/or answerer). Nor can it simply be based on the maximal consistent subset.  Rather it must be based on those endoxa each of which is, in some sense, most endoxon, or as endoxon as possible, at least on the subject in question. This  suggests, for instance, that if the premises of an argument are endoxa, and even drawn from the maximal consistent subset of the endoxa, but they conflict with other  endoxa any one of which is more endoxon than they are then the argument based on them will not be a proper, or at least not the best, dialectical argument since it is  not based on what is most endoxon among the relevant endoxa on the subject in question. By a natural extension of this, one can easily conceive of a dialectical  review of all of the problems and questions in a given area as serving to collect together a consistent body of information which is most endoxon on a given subject  such that what belongs to or is supported in an appropriate way by reference to this information is maximally well justified dialectically. If it could be determined, then,  what it is for an item, or a body, of information to be most endoxon on a given topic it could be asked what the merits of such a justification are.

  

Page 69

V—  Different Forms of Dialectic As an introduction to the exploration of this line of inquiry, it is useful to note that in Sophistical Refutations 34 Aristotle's departure from his earlier description, in  Topics I.1 and again in Sophistical Refutations 2, of dialectical reasoning simply as reasoning from endoxa, and his new requirement that both questioner and  answerer restrict themselves to reasoning from what is most endoxon or as endoxon as possible, is coupled with a change in the description of the method from  "dialectic" (or "dialectical reasoning") to "dialectic in its own right and peirastic" (Sophistical Refutations 34 183a39ff.). This raises the possibility that reasoning from  what is most endoxon is not required in some kinds or some uses of dialectic but only in others, particularly in the special form of dialectic called peirastic. Peirastic  calls for special treatment and it will be studied shortly. But independently of that study it is clear enough from other considerations that not all proper uses of dialectic  do require reasoning from what is most endoxon, or as endoxon as possible. In Topics I.2, as we have seen, Aristotle distinguishes the use of dialectic for mental  training or gymnastics from other uses. Later in Topics VIII.3 he claims that it is not required that the answerer in a gymnastic dialectical discussion for training  purposes only grant as premises things which are "more intelligible" (gnorimoteron) than the conclusion (159a10–14). 17 All that is required is that the endoxa  conceded "appear to be true'' (159a12).18 Whether they are more or less intelligible than the conclusion does not matter. In Topics VIII.5, however, Aristotle says  that in dialectic used for "testing and inquiry" (159a33), it is required that the answerer concede only what is both "more endoxon" and "more intelligible" than the  conclusion. He adds that the point of conceding only what is more endoxon is that this is the way to guarantee that reasoning will only be from what is "more  intelligible" than the conclusion (15968–9, 13–15; 160a14–16). This shows both that dialectic for gymnastic purposes has different rules than dialectic for "testing and  inquiry" (since only the latter need always reason from what is "more intelligible"). Also that gymnastic dialectic need not always reason from what is "more  endoxon" (since such reasoning does, and is meant to, guarantee that conclusions are derived from what is "more intelligible," and the latter simply does not always  happen, or need to happen, in gymnastic dialectic). It is an obvious corollary of this that gymnastic dialectic does not need to proceed from what is "most endoxon" or  "as endoxon as possible." This is clear from the fact that gymnastic dialectic need only reason from "what appears to be true." It may ignore any distinctions within this  class between what is more and less endoxon and, thus, also what is most endoxon or as endoxon as possible. Reasoning from premises of the latter sort, then, is  restricted to a special type or types of dialectic. What exactly is involved in this procedure? Aristotle first explicitly introduces the notion that dialectical reasoning, or a type of dialectical reasoning, should be from premises that are each "as endoxon as  possible," in Topics VIII.11. There he remarks that "(dialectical) reasoning is not

  

Page 70

equally endoxon and persuasive (pithanon) on all problems." Inquiry on some subjects is more difficult than on others, "so if someone reaches a conclusion from  premises which are as endoxa as possible [with respect to the problem at hand], then he has done dialectic properly" (161b34–38). Aristotle makes this remark at the  conclusion of a discussion of the various ways in which one may criticize a dialectical argument "in itself,'' as opposed to criticizing the performance of those who have  produced the argument (161b19ff.; cf. b38ff.). Since we are here particularly interested in the merits of the arguments themselves this distinction is important for our  purposes. One of the grounds for criticism of an argument "in itself' which he emphasizes is, if the reasoning proceeds from any premises which are "less endoxon than  the conclusion" or "more adoxon and less credible than the conclusion" (161b28, 30–31). This ground for criticism is, clearly, based on the requirement mentioned  above which is laid out earlier in Topics VIII.5–6 that, in dialectical "testing and inquiry," correct reasoning must proceed from premises which are each "more  endoxon and more intelligible" than the conclusion (159b8–9, 13–15, 160a12–16). The above remark (161b34ff.) is introduced as a clarifying addendum to the  ground for criticism based on this requirement (161b30–33). In effect, Aristotle is saying that though correct dialectical reasoning, of a certain type, proceeds from  premises which are each "more endoxon" than the conclusion how endoxon they are, within the class of what is more endoxon, will depend on the subject at hand.  Those arguments are most commendable which proceed from premises which are each maximally endoxon without qualification and (thus) maximally persuasive. But  sometimes the best dialectical argument, or arguments, that can be produced for some conclusion, or on some subject, will be ones for which some person is to be  commended for having done the best that can be done on that subject because there are no more endoxa premises on that subject from which the conclusion follows,  and perhaps even no more endoxa premises on that subject which conflict with the premises used. But still the argument may be less good than some other argument  on some other subject which has premises which are, absolutely speaking, more endoxa. The interest of this, for present purposes, is twofold. First, it supports the suggestion that Aristotle's interest, in Sophistical Refutations 34, in dialectical argument  from what is "as endoxon as possible" is an interest in the special type(s) of dialectic (viz. nongymnastic dialectic) in which the premises must each be "more endoxon"  than the conclusion, since that is his interest in the passage where he introduces this notion. Secondly, and more importantly, it shows that Aristotle is prepared to  apply two different standards for judging the merits of such dialectical arguments—a personal standard and an impersonal standard. By the personal standard an  argument may be commendable because it is an argument from what is more endoxon than the conclusion, and as much more so as possible given the subject under  discussion. Nevertheless, it may be, judged simply as a dialectical argument by the impersonal standard, quite weak because its premises are, in an absolute sense,  not maximally or highly endoxa.

  

Page 71

It is in this context that Aristotle makes a remark that is often misunderstood: It is unjust to find fault with those who reach a true conclusion from false premises; for what is false must always be syllogistically inferred from false premises and even what is  true can sometimes be syllogistically inferred from false premises. (162a8–11; cf 161a24ff., 162b27)

This is not, in Aristotle's view, an unjust criticism of a dialectical argument in its own right. Dialectical arguments can be criticized, and need to be defused (luein),  when they have false premises (VIII.11 161b21; 10 160b5, 23–39); and a dialectical argument with false premises is a "bad argument" (VIII.11 161b7 with  161a24ff.). It is, however, sometimes, an unjust criticism of the performance of some person who has argued from premises that are each, though false, as endoxon  as possible on some difficult and problematic topic (VIII.11 161b6–8 with a24ff.). But the merits of the performance do not mitigate the demerits of the argument  (b7). So this remark cannot be used, as it often has been, to show that dialectic is an unsuitable procedure to use in the search for truth. To understand how dialectic  can be so used, however, our first concern must be to understand what it is for a dialectical argument to be, by the impersonal standard, as commendable as  possible. That is, we want to know what it would be for a dialectical argument to be based on premises that are each "most endoxon" in an absolute sense. As a  prelude to this, it will be useful to consider, in more detail than is usual, how a dialectical discussion works to achieve the result that reasoning proceeds from what is  most endoxon, or as endoxon as possible, at least on the subject in question. VI—  Give and Take in Dialectical Inquiry Sometimes it seems to be supposed that a dialectical discussion is a very formal affair where there is little or no room for give and take. The questioner asks whether  the answerer will concede certain premises and the answerer, unless perhaps he has a question about the meaning of a premise, must say simply "yes" or "no." 19  Questioning proceeds until enough concessions are extracted to produce an argument which is then evaluated in isolation from other arguments and claims. Aristotle  perhaps gives this impression in Topics VIII.7, but it is clear from the surrounding chapters that a dialectical discussion typically involves much more give and take than  this. From the consideration of how this give and take normally goes, we can see how a dialectical discussion draws on, or comes to draw on, premises which are  each "most endoxon," on the subject in question. To begin with, each of the premises necessary for the questioner's proof of his thesis is standardly introduced not simply on its own for acceptance or rejection but as  the conclusion of an argument employing induction or analogy or deduction (VIII.1, 156a3–11, VIII.8 16038–9). The main exception to this is in the case of universal  premises which "appear thoroughly to be so" (lian prophaneis, VIII.1

  

Page 72

155b37; cf. VIII.2 158a3–6). 20 If a premise is apparently so beyond any doubt or question then there can be no worries that an answerer will fail to concede it. In  other circumstances the situation is different. There an answerer may fail to concede a premise. But if he does so he is not simply expected to say "no." If a general  premise is reached by inductive argument based on instances which are "apparently correct" (phainomena), then if it is rejected by the answerer an objection  (enstasis) or counter instance must be provided by him (VIII.8 160b3–5; cf. VIII.2 157a34ff.; b31–3). However, if the counter instance is a good one, or at least  appears to be (160b2), the questioner is not required to simply abandon his general premise. Rather he should qualify it so as to exclude the cases covered by the  objection (VIII.2 157b9ff.); unless, possibly, he is able to produce other premises to defeat the objection or support his premise (VIII.1 156a38ff.; VIII.3 159a4ff.).  In either event, nothing prevents the answerer from then introducing new objections, which can be responded to in turn. Sometimes the questioner will aid the  answerer, in order to gain his confidence, by bringing objections against some of his own premises (VIII.1 156b18–20). In addition to offering counterexamples to defeat inductive arguments for a questioner's premises, an answerer can introduce independent arguments (antepicheirein)  against these premises (even after they have been granted and a damaging conclusion has been drawn from them). This is done in order to undermine (luein) the  conclusion based on them (VIII.8 160b5–10 with VIII.9–10). Moreover, even if this sort of lusis or undermining cannot be provided, an answerer may  straightforwardly argue against a conclusion in its own right, where it conflicts appropriately with accepted opinions (VIII.8 160b6–10). When he produces such  arguments, against premise or conclusion, the answerer is, of course, in effect operating as a reasoner, i.e., as a questioner. Given this, there is no reason why his  argument, in objection either to premise or conclusion, cannot in turn be objected to by the original questioner who at this point is operating as an answerer.21 What rules determine when an objection (enstasis) or an undermining (lusis) or an attempt at counterargument (antepicheirein) is a good one in such a give and take  discussion for the purpose of inquiry? Clearly enough, the main rule will be the same one which governs all such dialectical reasoning, namely that the argument, if  relevant to the matter at hand, must proceed from what is "more endoxon" than its conclusion (VIII.5–6, see the passages quoted in the next paragraph below; 11  161b30–1). This means that if a counterexample is offered to a universal premise it must be more endoxon than that premise; if the counterexample is, in turn,  objected to it must be by reference to what is more endoxon than it. If a premise or a conclusion is subjected to undermining or counterargument it must likewise be  on the basis of premises which are more endoxa than it, and any counter response to this must be based on what is more endoxon than these premises. Thus  conceived, a single dialectical inquiry involves an attempt to establish some claim by reference to endoxical premises which are each not only (1) more endoxon than  that claim, but are also (2) such that there is

  

Page 73

nothing more endoxon than these premises with which they conflict. Each such endoxon premise will be relatively more endoxon than certain other things but not less  endoxon (or more adoxon) than anything incompatible with it. In that respect these premises can be said to be dialectically undefeatable. As such they can be said  to belong to what is as endoxon as possible or what is most endoxon not simply relative to some particular conclusion drawn from them but also relative to other  endoxa on the topic which they concern, including any endoxical information in any way relevant to that topic. Given the discussion in Topics VIII, there seems little  doubt that Aristotle's interest in Sophistical Refutations 34 in a method for arguing from what is "as endoxon as possible" at least includes an interest in the method  for achieving arguments from such premises which he spells out earlier. Nevertheless, as we shall see, Aristotle also supposes that among the dialectically undefeatable  premises on different topics some will be in another respect more endoxon than others, without qualification, so that the most endoxa of all would be a subset of the  undefeatable propositions. To determine whether Aristotle has any special interest in these propositions in Sophistical Refutations 34, and to see further what the  epistemological status of this special class of propositions would be, it is necessary now to turn to the crucial question of what it is for one proposition to be more  endoxon than another. VII—  Reasoning from What Is Most Endoxon The main evidence for answering this question is found in Topics VIII.5–6. There Aristotle lays down the principles which govern dialectic when used for "testing and  inquiry" (159a32ff). His main general principle there is this: The one who reasons correctly [as a result of questioning] establishes his set thesis on the basis of things which are more endoxa and more intelligible [than the thesis itself].  (159b8–9)

Further on he adds, in the same vein: The things which are granted [by the answerer] must all be . . . more endoxa than the conclusion, if the less intelligible is to be reached through the more intelligible. (159b13–15)

And: The questioner shall complete his reasoning with all those things being conceded to him which are more endoxa than his conclusion. Those who try to reason from things which  are more adoxa (discredited) than the conclusion clearly do not reason properly. Therefore, such things should not be conceded to questioners. (160a12–16)

These passages all confirm that one endoxon can be more endoxon than another and they suggest what is later confirmed in VIII.11, that the more endoxon some  claim is the more weight it has in at least one type of dialectic. So the preferred

  

Page 74

claim, in case of conflict, in dialectic will always be the more endoxon claim; and the most preferred are, thus, the most endoxa. Which are these? Our passages here  enable us now to answer this question. As we have already noted, Aristotle couples here what is more endoxon with what is more intelligible (gnorimoteron); and he  claims that "if the less intelligible is to be reached through the more intelligible" then we must proceed to the less endoxon through the more endoxon. This shows that  a more endoxon belief will always be more intelligible than a less endoxon belief. But in what sense is it more intelligible? That depends, for one thing, on the type of  dialectical inquiry we are pursuing. If the discussion is ad hominem and concerns the position of some particular individual who is to be examined from his own point  of view, then we are dealing, as dialectic always is, with what is endoxon, but only endoxon to that individual. And our arguments must be from what is more  endoxon and hence more intelligible to that individual: If the position in question is endoxon or adoxon not simply but to the answerer then what ought to be conceded or not is judged by reference to what is held or not held by him.  Or if the answerer is set to defend the opinion of someone else, it is clear that each point should be conceded or refused with a view to that person's own mind. (159b25–29)

But dialectical argument need not be ad hominem in either of these ways. More often it is based not on the endoxa according to some individual but on the endoxa  without qualification—"the things which are accepted by everyone, or by the majority, or by the wise." Here what is more endoxon and thus more intelligible must be  what is so not in relation to any particular individual but somehow generally. Earlier in the Topics (as in various other places) Aristotle describes two general ways in  which one thing may be more intelligible than another, in neither of which is a relationship to some special individual (or group) in question. One thing may be more  intelligible than another "simply" or "to us." Thus, simply, the prior is more intelligible than the posterior; a point, for example, than a line, a line than a plane, and a plane than a solid; just as a unit is more intelligible than a  number—since it is prior to and the basis of every number. Similarly, a letter is more intelligible than a syllable. However, to us it sometimes turns out in reverse. For a solid falls  most of all under our perception and a plane more than a line and a line more than a point. For people in general come to know such things earlier, since they can be understood by  the ordinary sort of intelligence, the others only by an intelligence which is exact and uncommon. (Topics VI.4 141b5–14)

Does proper dialectical argument proceed from what is more intelligible (or what is most intelligible) simply or to us? Quite clearly it must be the latter. As we have  seen, dialectic is described in the opening sentence of the Rhetoric as like rhetoric in drawing on things "which it is in a particular way common to all to  understand" (1354a2–3). Things which are more intelligible simply, as described above, do not fit such a description. They are intelligible only to an "exact and

  

Page 75

uncommon intelligence." The things that are most intelligible simply include as paradigms the first principles of the sciences as such (cf. Physics I.1 184a16–23). But  these cannot be as such among the most authoritative premises for dialectic. For these need not even be endoxa, much less most endoxa. They need not be in every  case (or any case) known or accepted by appropriate parties. But endoxa must be things that are accepted. The implications of this last fact are most clearly brought  out in a passage in Posterior Analytics I.19: It is clear that those who are reasoning from the point of view of standing opinion (kata doxan), that is only dialectically, should only consider this, whether their argument  proceeds from the most endoxa premises possible. Thus, although a term is not in truth the middle term between A and B, if it is held (dokei) to be then one who reasons through  it has reasoned dialectically. But with a view to truth it is necessary to proceed from what actually obtains. (81b18–23)

As Aristotle makes clear here, proper dialectical reasoning, as reasoning from what is most endoxon, need not proceed (in a case where a deductive syllogistic  argument is being used), through what is "in truth the middle term." Rather, such reasoning need only proceed through what is held to be the middle term. 22 This is to  say, however, that the explanatory relations which actually hold among things need not be reflected in a proper dialectical argument. This is because an argument  through a true middle term is just one which gets these actual explanatory relations right (Posterior Analytics II.2 90a6–7). This shows that it cannot be a requirement  of dialectic that its arguments proceed from premises that are more or most intelligible simply. For arguments which follow such an arrangement cannot fail to reflect  the natural order of explanation, and so to exhibit true middles, when we are dealing with matters where there is such an order (Physics I.1 184a10–21). Given this,  when Aristotle characterizes dialectical reasoning as proceeding from what is more endoxon and thus more intelligible he must have in mind what is more endoxon  and more intelligible "to us." This is confirmed by a passage in Sophistical Refutations 33 where Aristotle directly equates what is "held most of all" (malista  dokounta) with what is "endoxon most of all'' (182b37–183a4).23 VIII—  Rank Ordering among Endoxa Generally speaking then, those endoxa which are most endoxa and which, thus, have the most weight or authority are those which have the greatest level of actual  (explicit or implicit)24 acceptance. In case of conflict, this is the rule to be used to describe which endoxa to retain and which to reject. Given this, it seems likely that  the order in which the different types of endoxa are introduced in Topics I.1 100621–3 is not accidental. The first type introduced is "the things which are accepted  by everyone." These are the things that have, in dialectic, the greatest weight. This is confirmed by the fact that Aristotle treats what is apparent to

  

Page 76

everyone as practically unchallengeable (Topics VIII.1 155b37 with VIII.2 158a1). 25 What everyone holds, by way of affirmation or denial, is thus most endoxon. Does this put what everyone holds absolutely beyond challenge? Aristotle comes close  to saying this but he does not ever quite say it; and clearly he should not say it. Things that everyone holds could be themselves inconsistent. Here a resolution could be  reached if among the things that are acceptable to everyone some are more intelligible to everyone than others by being more firmly held. These could count as the  more endoxa (the more noted or accredited) not in extent of acceptance only but in strength of acceptance as well. Should there be at any time conflicting beliefs  which are each maximally endoxon without qualification, then the conflict between them will be dialectically unresolvable at that time. Aristotle is aware of such a  possibility, but he seems to feel that in such a case proper inquiry over time will change the status of one of the conflicting beliefs (SE 33 182637–183a4). In any event, not all endoxa are accepted by everyone. The next class that Aristotle lists is "the things which are accepted by most people." These come next, quite  clearly, in the hierarchy of what is "more endoxon and more intelligible" to us. Here too there can be conflicts needing resolution among beliefs that are equally widely  accepted and, thus, here too strength of acceptance as well as extent of acceptance might need to be appealed to. Next in Aristotle's list of endoxa come "the things accepted by the wise" (sophoi). Who does he mean by "the wise" here? In some contexts he uses this term to refer  to those who actually have special know­how or scientific knowledge and understanding (Nichomachean Ethics VI.7 1141a9–20). But he does not use the term in  this objective sense here. He substitutes for 'wise' in his discussion 'someone famous as a philosopher' (Topics I. 11 104b19–20; cf. with 104a8–12 and b31–4); and  it is clear that the views of those designated by the latter description are not necessarily for him instances of actual knowledge. As examples he mentions Antisthenes's  view that it is impossible to contradict anyone and Melissus's view that being is one (b20–2). So 'the wise' here are those who have the recognized standing of  experts or savants but not necessarily those who are such. (We continue, of course, to use terms such as 'expert' in this way in referring, for example, to "what the  experts say.") This makes it clear how the views of 'the wise' can count in their own right as noted or esteemed views (endoxa) in a way in which the views of  individuals with actual wisdom but who are totally unknown could not. The views of recognized experts are those which acquire standing for us in the body of  recognized "wisdom." This is also important for understanding how the views of ''the wise" can be less authoritative in dialectic than the views of everyone or of most  people. That they are less authoritative is further confirmed by Aristotle's discussion in Topics I.10–11. Aristotle introduces there the notion of a dialectical premise  (protasis), as a proposition which someone could be prima facie expected to grant when the question of whether to accept it is raised in dialectical discussion. He  describes it as follows:

  

Page 77 A dialectical premise (protasis) is a proposition, introduced in the form of a question, which is ... esteemed (endoxos) by everyone or by most people; or by the wise—either all of  them, or most, or the most famous, assuming it is not paradoxical; 26 for anyone would grant what is accepted by the wise if it is not opposed to general opinion. (I.10 104a8–12)

Here Aristotle makes it clear that in dialectic it is not to be expected that an opinion of recognized experts will be granted when it conflicts with dominant majority  opinion (cf. Politics III. 11 1282a15ff.). Thus, such an opinion does not count as a premise for a dialectical argument. This fits with what we would expect given that  dialectical argument has to proceed from what is more credible to and intelligible to us. We readily credit the recognized views of experts on matters where we have  no independent opinion of our own. There the views of recognized experts are the most widely recognized and credited, and thus the most intelligible to us, of  available views. Where experts have views that conflict with our own preponderant view, however, their views are not the most widely credited and intelligible to us of  those available. Thus they have less authority in dialectic where this is the standard in play. Of course, the view of a famous expert still counts by definition as an  endoxon (10021–3, 101a11), even if it is contrary to the majority view. How then does it enter in as an item with some weight in dialectical discussion if it cannot  serve directly as a premise? Will not an answerer always be entitled to reject it out of hand since it conflicts with the more endoxon majority view? Aristotle's practice  does not fit with such an approach; and his theory in the Topics shows how the views of experts even when they conflict with and are on their own less endoxa and  less intelligible than majority views, can be introduced and even successfully defended in dialectic. An expert view that conflicts with majority views is called by  Aristotle a thesis. A thesis is a view which conflicts with received opinion, held by someone famous as a philosopher . . . or a view opposed to received opinions for which we have an argument; for  instance the view maintained by the sophists that not everything that exists has either come to be or is eternal. For the cultured person who is literate exists though he has not  come to be and is not eternal.... This view, even if it does not seem [on its own] true to someone, might well seem so because of the argument. (104b19–28)

This passage indicates how a view which is, on its own, contrary to general opinion can acquire standing in dialectical discussion. It can be introduced by use of an  argument that shows that it follows from what we would generally accept. Aristotle does not explicitly apply this point to the views of philosophers here. But the  context suggests that he has this in mind; and he typically thinks of philosophers, or equivalently "the wise" (sophoi, 104b33), as those who offer reasons and  explanations for what they claim (Met. I.1–2 981a24ff., 982b11ff.; SE 1 165a19–28). Certain passages in the Ethics confirm this connection.

  

Page 78 It is perhaps rather pointless to examine all received opinions; it is sufficient to examine those which are most prevalent or those which are taken to be supported by some  argument. (I.4 1095a28)

Of the two types of views that Aristotle counts as worth considering here the minority views held by experts would belong to the second. Later, at 1098b27–8 he  redescribes the two types of views which he has proceeded to consider as those either of many people or of the renowned. From this, it is easy to see how special  expert views contrary to more widely accepted or majority views can be taken seriously. They can even be established in dialectic without contravening the general  rule that what is credited by and more intelligible to most people takes precedence over the views of the wise. If the expert views can be seen to follow from universal  or majority views then they acquire the authority of those views and so can pose a challenge to other (less endoxa) majority views. Where there is no dominant majority view on some subject, however, the views of the wise are, as we have seen, the most authoritative just on their own. It is of  interest that Aristotle gives no special weight at this juncture to the opinions of a bare majority. The opinions of most people (hoi pleistoi) or of the general community  (hoi polloi) outweigh the contrary views of experts (104a8–13), but the opinions of a bare or small majority do not. This is intelligible in the light of Aristotle's general  scheme. Dominant community views will be widely credited by us, but where there is a close division among community views, Aristotle supposes we are likely to  credit the "wisdom" of the experts rather than the views of a small general majority. The views of the wise can also conflict with each other. In introducing the views of the wise in his list of the types of endoxa, Aristotle mentions first the views of all the  wise, then those of most, then those of the most famous and distinguished. Is this also a ranking in order of greater to lesser authority? Are we to understand Aristotle  as saying that if the experts do not all agree then we should rely on the view of most of the experts? And if there is no substantial majority view, only then should we  prefer the views of the most famous and distinguished to those of other experts? Though Aristotle says nothing explicit on the point, this would make sense given his  general approach. He says, as we have seen, that we follow the practice of relying on established expert views where we have developed no relevant common views  of our own. But if we have developed no views on our own and we are thus in the position of having come to rely on recognized expert views as the ones we most  credit, it makes sense that the view on a given topic held by most experts would be the one that has achieved the greatest renown and credit by us. If, however, there  is no dominant view of the experts (again a bare majority counts for nothing), then it makes sense that we should go with "the most famous and distinguished." That is,  we should go with the experts whose views have received the greatest recognition from us. (The word translated 'distinguished' is endoxoi, 100b23.) This again  seems the best we can do if we are relying on what is most well accredited by us. However, in all these cases as in the earlier ones, it

  

Page 79

would, of course, be possible to override the prima facie order of authority by appeal to a supporting argument for some claim where the premises of that argument  belong to a more authoritative category than the claim. If this account is correct then the initial order in which Aristotle lists the types of endoxa is not accidental. He moves in his list in order from the (prima facie) more to  the (prima facie) less authoritative types of endoxa. IX—  Dialectic and Peirastic in Scientific Inquiry We are in a position now to raise pointedly the main question raised earlier: Why does Aristotle think it is necessary in scientific inquiry to use dialectic? Consider the  recommendation he gives to the inquiring dialectician. As we can see now, in effect it comes to this: Base what you say, if you can, on what everyone accepts; failing  that, appeal to what the dominant majority think. If there is no dominant majority view appeal to the view of the experts. If the experts do not agree, appeal to the  dominant majority of the experts. If there is no dominant majority view among experts appeal to the view of the most famous and distinguished experts. This sounds  like appropriate advice for the adman trying to maximize sales or the politician trying to maximize support for his policies, but hardly for a scientist. That is to say it  seems to embody a procedure for justification which is designed to be maximally effective in producing conviction when used with people in general. This is, of course,  just what is to be expected since, as we have already noted, Aristotle thinks of dialectic as like rhetoric in focusing its appeal on what everyone is maximally ready to  understand and accept. But the aims of science are different. The scientist's results can, Aristotle says, only be grasped by "an intelligence which is exact and  uncommon" (Topics VI.4 141b13–14). So why should science have any interest at all in procedures for justification best designed to produce maximal understanding  and conviction in people in general? In attempting to answer this question it is important to try to determine just how it is, on Aristotle's own theory of dialectic in the Topics, that dialectic does play its  necessary role in scientific inquiry mentioned in Topics I (1.2 101a36ff.). Writers have standardly supposed that Aristotle simply does not say anything further about  this in the Topics. Some have even suggested that this is an indication that the method of the Topics itself is not one that Aristotle in fact conceives of as appropriate  for use for the scientific purposes mentioned. Rather the method of the Topics is only for use in school gymnastics. 27 It is, however, a mistake to think that Aristotle  says nothing further about this matter. In what we call the Topics he does not elaborate on his claim that it is necessary to use dialectic to reach the first principles of  the sciences. (He comments, very briefly, in Topics VIII. 14 163b9ff., only on the other scientific use mentioned in 1.2 101a34–36.) But in what we call the  Sophistical Refutations, which is clearly a part of the Topics,28 he is particularly

  

Page 80

interested to describe a special form of dialectic which he calls peirastic (peirastikê), or the art of testing. Peirastic is introduced as a method for reasoning from  "what is necessarily known by anyone who pretends to have epistêmê" (i.e., technical or scientific knowledge, SE 2 165b5–6). Later it is described as a procedure  which, though it does not require epistêmê and can be used even by the unlearned, nevertheless can be used to test the claim of all those who profess epistêmê in  whatever subject (SE 11 172a17–32). In Topics 1.2 Aristotle justifies his claim that the use of dialectic is necessary to reach scientific first principles on the specific  ground that dialectic is "capable of examination" (exetastikê, 101b3). The two words peirastikê and exetastikê, here translated 'testing' and 'examination', are  derived from roots which overlap in meaning in Greek. Though peirastikê is a technical term, various cognate forms of the two words are used interchangeably by  Aristotle (as they are by Plato) in his general descriptions of dialectical procedure (Rhet. I.1 1354a5, Top. VIII.5 159a25, 33; cf. Plato, Apol. 22e–23c, Prot. 348a).  If Aristotle had used peirastikê, or some cognate word, instead of exetastikê, in Topics 1.2, then the material in the Sophistical Refutations would long ago have  been thoroughly studied to understand Aristotle's famous doctrine in Topics 1.2. But it would appear that little attention has been paid to the apparent connection  between the Topics passage and the later material up to now. X—  Some Main Features of Peirastic What is peirastic and how does it serve to examine or test claims to scientific knowledge? Aristotle describes peirastic as "a part of dialectic" and "a type of  dialectic" (SE 2 169b25; 11 171b45; cf. 9 170b10–11). Thus, like all dialectical arguments, peirastic arguments employ endoxa as premises (SE 2 165b3–4). They  must, however, also satisfy certain stronger requirements. Aristotle initially defines peirastic arguments as those whose premises are "things which are believed by the  answerer and which are necessarily known by anyone who pretends to have scientific knowledge" (SE 2 165b4–6). Some have claimed that this explicitly  distinguishes peirastic arguments from dialectical arguments and contradicts Aristotle's later claim mentioned above, that peirastic is one part or type of dialectic. 29  As we shall shortly see, there is no inconsistency. As Aristotle understands these initial requirements for peirastic any premises which satisfy them are guaranteed to  be endoxa and, thus, premises for dialectical arguments, albeit of a special type. To better understand what peirastic arguments involve, it is worth looking closely now at the passages in Sophistical Refutations 11 where Aristotle introduces the  important claim that the dialectician who is doing peirastic does his testing on the basis of a certain kind of knowledge. He describes this knowledge there as  knowledge of what is "common" (koina). The dialectician who does peirastic is, in fact, identified at the beginning of that chapter as the one "who takes

  

Page 81

into account the common things as they relate to the matter at hand" (171b6). These common things are described later in the chapter as things such that "the one who  does not know them necessarily does not possess (scientific) knowledge" (172a26–7). This description clearly reintroduces a feature of one of the special  requirements for peirastic dialectical premises introduced in Chapter 2 (that they are necessarily known by the professing scientist), and thus it makes it clear that  these common things are the actual premises involved in doing peirastic. Aristotle further describes the common things which peirastic dialectic uses in the following  way: Peirastic is not the scientific knowledge (epistêmê) of any definite subject. Thus it may deal with every subject. For each of the arts employs some common things (koina) so that  everyone, even the unlearned, makes use of dialectic and peirastic in some fashion. For everyone attempts to test those who profess knowledge on a limited basis, and these  things [which they use] are the common things; because people know these things themselves no less than the scientist even if they seem to say quite inaccurate things [on the  scientific subject]. So they all practice refutation since they practice an unskilled form of the activity which dialectic practices with skill, while the dialectician is the one who does  peirastic by use of an art of reasoning. (172a27–36)

What are these "common" things which everyone knows? One possibility, accepted by some interpreters, is that they are what Aristotle elsewhere calls the "common  notions" or "common axioms"—the principles which hold true for all subject matters (APo. I.10 76b11ff. Met. II.2 996b26ff.). In fact, on one manuscript tradition, the  one preferred by recent editors, Aristotle goes on immediately to characterize these "common" things as ''identical principles which hold true of everything" (172a36– 7). This is a natural reading in view of the fact that Aristotle contrasts these things with others which are "unique" (idia) to a given subject matter and such as to  determine "a distinct nature and kind" (phusin tina kai genos, 172a37–8). This is just the kind of language that Aristotle uses elsewhere to contrast the common  axioms with the other first principles special to each science (APo. I.10 76a37ff., 76b11ff.). But this cannot be what Aristotle means here. For one thing, Aristotle  emphasizes the point that only "some common things" are employed by each art (172a29–30). So he does not mean that they all hold true of everything. In addition, if  he did mean this, we would be faced with the bizarre consequence that the peirastic dialectician only examines or refutes claims by testing whether they fit with the  common axioms—that is, principles such as the law of noncontradiction, the axiom of equals and the like. But Aristotle permits peirastic dialectical argument to  proceed from any relevant endoxa not just from common axioms. He gives various examples in the Sophistical Refutations of such arguments. Earlier in Chapter 11,  for example, he offers the following example of an argument that appeals to something "common."

  

Page 82 Take the case where someone denies that it is better to take a walk after dinner [rather than before dinner] and gives as his reason Zeno's argument [that motion is impossible].  This is not a medical argument since the argument employs what is common. (172a8–9)

The argument mentioned here is in Aristotle's view an eristical or sophistical argument, not a dialectical argument (See 172a2ff.). But it is like peirastic dialectical  arguments, he says, in appealing to what is common. And in this case the common things are clearly not true or applicable in all subjects, but rather include, in the case  of Zeno's argument, such common and well–known information as: To go a certain distance you must first go half­way. One might suppose that such information must  at least be applicable in a number of areas, but this would, again, exclude from the dialectician's premises numerous relevant endoxa. Aristotle himself calls it common  because it is not "based on first principles that are unique" to the specific subject at hand, i.e., medicine. But rather, the common is based on matters which can be  adapted for use "against people in general who do not know what is possible and impossible in the subject at hand; since it will apply with such people" (172a4–7).  He later explains why the common things can be so used with people in general: "because people know these things themselves no less than the scientist" (172a33–4).  Here Aristotle opposes what is common to what is used in argument from proper first principles (i.e., in demonstrations) with those who have special scientific  knowledge of those principles. He describes the common as what can be used specifically by and with all people who may lack such special knowledge, but  nevertheless know these things. A similar contrast to this one is drawn earlier in Sophistical Refutations 9. There Aristotle distinguishes scientists who can produce or  evaluate refutations based on the first principles (archai) of a given science, i.e., refutations by demonstration (170a20–34; 36–38), from dialecticians. Dialecticians  are those who produce refutations based on ''the common sources of argument for each art and discipline" (170a34–36) and examine refutations "which do not fall  under any particular art" (i.e., are not demonstrative in character) but are "based on the common things" (i.e., on the premises for proofs from endoxa [endoxoi  syllogismoi] on any given subject) (170a38–b1). These common premises which determine the dialectician's province are clearly not restricted to propositions which  hold good of all, or many, subject matters but include any appropriate and relevant endoxa. This use of the term 'common' is found in the opening lines of the Rhetoric, quoted earlier, where it is specifically used to describe dialectic in both its questioning and  answering modes. The passage is worth repeating. Rhetoric is a partner of dialectic. For both of them are concerned with the sorts of things which it is in a certain way common to everyone to know and which require no specialized  knowledge. Thus everyone in a certain way takes part in both dialectic and rhetoric. For everyone on a limited basis engages in examination and

  

Page 83 in submitting to argument; and also in defending himself and in accusing others. (1354a1–6)

Another passage, a few paragraphs later, develops this point: In dealing with certain people, even if we possessed the most accurate knowledge (epistêmê), it would not be easy to persuade them by arguments based on this knowledge. For  argument according to knowledge (epistêmê) is instruction (didaskalia), and this is impossible in this case. Rather it is necessary to construct our proofs and arguments by use of  the common things (koina), just as we said in the Topics about ordinary discussions (enteuxeis) with people in general. (1355a24–29)

In the discussion of enteuxeis to which Aristotle refers, in Topics 1.2, he does not use the term koina but rather, as an equivalent to it, "the opinions of people in  general" (tas tôn pollôn doxas, 101a31). So it is not just any endoxa but only these common opinions, whether they hold true of one or many or all subjects that  constitute the "common" things as they are understood here. The language of these passages in the Rhetoric so closely parallels that in the Sophistical Refutations that there seems little doubt that the thought of the two is  essentially the same. So the term 'common' in these passages does not designate what is metaphysically common, i.e., true of or applicable to all or many subjects, but  rather what is epistemically common, i.e., what is intelligible even to common people, those without specialized knowledge of principles. Given this, it seems clear that we should not follow recent editors in accepting the manuscript tradition on which Aristotle describes the common things which dialectic  reasons from as "identical principles that are true of everything." Rather, we should prefer the alternate manuscript tradition on which Aristotle says simply that "there  are many of these [common] things in each area ... so that it is possible on the basis of these things to engage in peirastic testing on every subject" (172a36–9). On  this reading, Aristotle makes it clear that peirastic will pay special attention to those commonly understood things that hold true in the particular area under  investigation. When he goes on to contrast the whole class of common things with things which are "proper'' (idia) to a given area (172a38) he does not mean that  particular common things do not specially concern some particular area but only that collectively the common things are not connected with "a distinct nature and  kind (genos)" (a37–38). This is how he argues that peirastic (and eristic) argument cannot be demonstration, from "first principles" (172b1–4), since this would  require that the common things do collectively determine a distinct kind. 30 It is easy enough to see now why it is that peirastic arguments always count as dialectical arguments, as those two types of arguments are defined in Sophistical  Refutations 2. Since the premises of peirastic arguments are the "common" things, things that everyone knows, both the learned and the unlearned (172a27–36),  they are things which everyone accepts. Thus, the premises of peirastic arguments are endoxa, by the definition of that term in Topics I.1. More than that, since  everyone

  

Page 84

accepts them, they are each "most endoxon," or "as endoxon as possible," in the absolute sense. This explains why Aristotle, in Sophistical Refutations 2, can say  that the questioners's premises in peirastic arguments must be "believed by the answerer," when this is not at all a requirement of dialectic in general. Everyone accepts  the premises of peirastic arguments. It also explains why, in Chapter 34, he can show a special interest in peirastic as a procedure in which both questioner and  answerer proceed from what is ''most endoxon" or "as endoxon as possible." Peirastic is, without qualification, a procedure of that sort. As we noted earlier, Aristotle takes it that one can legitimately criticize a dialectical argument not only where the premises are not endoxa, or where they are less  endoxa than the conclusion, but also where the premises are false. (Topics VIII.11 161b19ff. at b21; 12 162b14. The criticism can be overridden where the  argument is used as a reductio to establish the falsity of a premise 162b 16–22). It is clear that Aristotle understands all of these modes of criticism as evaluations of  an argument as a piece of dialectic, not from some external nondialectical perspective. But it is a general puzzle how dialectic, which Aristotle always characterizes as  proceeding from the point of view of standing opinion (kata doxan) rather than from the point of view of truth (kat' aletheian), can be in a position to evaluate  premises as true or false (Top. I.14 105b30–31, APo. I.19 81b18–23). Since Aristotle describes peirastic dialectic as proceeding from what everyone knows and  what even pretending scientists must know, it is clear that for peirastic at least there must be some way of guaranteeing that its premises are true. If we can  understand how this is accomplished this may contribute to the resolution of the general puzzle. There is one further important description in Sophistical Refutations 11 of the common things which are known and understood by all and so serve to make genuine  peirastic dialectical examination possible. To show why peirastic dialectic "is not the same sort of thing as geometry but is rather an art one may possess even without  having specific [scientific] knowledge" (172a22–3), Aristotle offers the following argument. It is possible even for someone who does not know the subject [scientifically] to test another who does not know the subject, providing that the latter grants things not based on  what he knows [scientifically] or, in particular, on the special principles of the subject, but on things which are consequent on [the special principles] which are such that, though  knowing them does not prevent one from not knowing the discipline, still one who does not know them necessarily does not know it. (172a23–27)

What are these things different from but "consequent on" (ta hepomena) the special principles of a science or other discipline such that in order to know the science or  discipline you must know them as well as the special principles? In the case of a science, the answer is clear; they must be the things explained by the special  principles, the objects of the "knowledge that" which precedes the

  

Page 85

"knowledge why" provided by the principles (APo. II.1–2). This is all that the scientist needs to know in addition to the principles to have full knowledge of the  science and these things are, of course, in a very straightforward sense "consequent on" the special principles. If this is correct then it is quite easy to see how peirastic dialectic can play a crucial role in the process of discovery and justification in science. If the "common" things  (the special set of generally known endoxa), from which peirastic dialectic argues, do make up at least some of the things which must be explained by the principles  of a science, then one can easily see how they can be used to rule out any proposed theories that are incompatible with them and how dialectical inquiry in a given area  could serve to collect them together as a necessary prelude to the search for correct explanatory theories. But, as we have seen, these common things, the things  which peirastic dialectic particularly argues from, are described not only as things consequent on the special principles of a science but also, later in Sophistical  Refutations 34, as things which are "most endoxa." How can it be that there is this correspondence among these classes? Why should what is most endoxon turn out  to constitute at least a part of what we must be able to explain by appropriate scientific principles? And why should either of these things be what is commonly known  (the koina) on the topic in question and thus true? XI—  Dialectic in Science: A Standard View At this point, it will be useful to consider the few attempts which have been made in the literature to say what it is that makes dialectic suitable as a procedure for  justification in science or elsewhere, to see what they may offer in answer to these questions. One approach is suggested in the influential work done by G. E. L. Owen  on Aristotle's method of dialectic. Owen did not attempt to work out in any general or detailed way what the epistemological value is of the appeal to endoxa. But he  did argue that an appeal to endoxa or legomena (the things which are said) "may be an appeal either to common belief about matters of fact (e.g., EN I.11 1101a22– 24) or to established forms of language (e.g., ibid. VII.1 1145b19–20; 2 1146b4–5) or to a philosophical thesis claiming the factual virtues of the first and the analytic  certainty of the second (e.g., ibid. I.8 1098b12–18)." 31 Owen did not elaborate on what the "factual virtues" of "common belief about matters of fact" might be for  Aristotle. But he suggests here that the epistemological value of the appeal to "established forms of language'' derives from the fact that they have "analytic certainty." In  the same vein, Owen argues that when, for example, Aristotle argues dialectically in the Ethics against the suggestion that practical wisdom can be coupled with  incontinence on the ground that "no one would say this" (1146a6), what he is asserting is "not that it happens to be false but that given the established use of the words  it is absurd." He claims further that, in his

  

Page 86

discussion of incontinence, Aristotle appeals to certain endoxa in order "to show a priori that there is no use for the expression 'doing what is wrong in the full  knowledge of what is right in the given circumstances.'" 32 Owen does not elaborate on these claims at any length; and it would be easy, in the light of current  doubts about the coherence of the analytic–synthetic distinction and its value for epistemology to ignore his proposals. After all, there is no direct evidence that  Aristotle recognized a class of analytic statements. He has no term for our 'analytic', and he never explicitly attempts to justify belief in any class of statements on the  ground that they hold simply in virtue of linguistic conventions which fix the meanings of terms and so can be known a priori. But there are various reasons why the  suggestion deserves consideration that appeal to endoxa is epistemically significant, in some cases at least, because the endoxa reflect established forms of language  and are thereby analytic, and are known a priori with certainty.33 If it were the case that peirastic dialectic in particular did argue from analytic premises that reflect a  priori conventions of language, that would in fact explain certain things Aristotle says about peirastic. It would explain why the premises of peirastic argument are  commonly known and accepted by everyone, and also why they are true and thus capable of being used to decisively refute proposed scientific principles which are  incompatible with them. However, it would not explain one other crucial claim that Aristotle makes about peirastic premises, namely that they are things that are  "consequent on" and explained by the relevant scientific principles. Why should the analytic truths have this status? It might be suggested in defense of Owen on this point that Aristotle wants to argue that the body of common or accredited opinion on a subject fixes the reference of  the name of the subject. This is done in such a way that it is a logical presupposition of successful reference to that subject that most, and the most intelligible parts to  us, of that body of opinion are not false of that subject or, more strongly, are explained by the basic principles of that subject.34 In his own discussions of reference  Aristotle does suppose that there is some most intelligible opinion which at least sometimes has this kind of reference fixing status, namely the opinion expressed in the  "account of what the name signifies."35 Interestingly enough, this type of opinion can play for Aristotle the role of being one which appropriate theoretical principles  must explain on pain of failure of reference and hence failure of truth for those principles. To take a familiar example from Posterior Analytics II.8–10, Aristotle  supposes that our belief that 'thunder is a certain noise in the clouds' serves as the belief by grasp of which our thought is fixed on to the real thing, thunder in such a  way that that fact must be true, and explained by the fundamental principles of meteorology if the term 'thunder' is to refer to anything (93a21–9 with 94a7–9). A  similar example from Parts of Animals II.3 explicitly employs semantic terminology. It is clear that blood is hot in the following way, namely as something in the being of blood. Blood is said to be hot in the same way as something would be if we were to signify by  its name that it is seething hot water. But blood is not hot in respect of

  

Page 87 its permanent substance (hupokeimenon). So in one way it is hot in itself (kath hauto) in another way it is not. For heat belongs in our account of it in the way in which white  belongs in our account of the white man; but insofar as blood takes on heat as an accidental affection (pathos), 36 it is not hot in itself. (649b21–27, cf. 649a13–20)

Here Aristotlem makes it clear again that in the account of what we signify by a name a certain group of the features which we take to belong to what the name  denotes will be included. He also makes it clear that the features that are included in the account of what we signify by a name may be features naturally present and  whose presence is explicable. For instance, the mention of heat is included in the account of what we signify by the name 'blood', even though heat is not a part of the  fundamental essence of blood. Nevertheless, blood is naturally heated up, particularly by the heart, in the course of its production for the sake of nourishment and  growth, and it is in looking for the explanation of this that we find the real nature of blood (650a2ff., Juv. et Sen. 469b12ff.). This provides a good illustration of how Aristotle's scientific inquiries standardly focus on looking for explanations for opinions with this kind of reference fixing status.  But it offers only limited support for Owen's proposal at best, even putting aside the question whether these opinions count for Aristotle as items of a priori knowledge.  Since Aristotle only gives such reference fixing status to single select opinions on a given subject, and never to the body of accredited opinion, or of most accredited  opinion, as a whole, the role in justification which he accords to peirastic premises in general cannot be explained along these lines. Related problems arise from the fact that the appeals to "established forms of language" which Owen himself cites from the Ethics, do not seem to introduce analytic  theses. The first of these is: "People are incontinent in respect of pride and honor and gain." The second is: "Some people are incontinent without qualification.'' In  neither case does Aristotle regard the fact that "we say" these things as guaranteeing without further argument that they are so. He takes it as necessary to justify this  pair of related theses (in Nicomachean Ethics VII.4), and his justification for them does not draw solely on premises which are based on established forms of  language. The justification depends crucially on the claim that some things that are a source of pleasure are necessary for life—such as food and drink; some are  desirable for their own sake but not necessities—such as honor (1147b23ff.). On the basis of this he justifies the view that incontinence in respect of honor occurs and  also the view that incontinence without any added qualification occurs (when a desire for some necessity overmasters a judgment of what is best). But this basis for  Aristotle's argument is hardly analytic. Nor does Aristotle try "to show a priori," as Owen claims, that no one can do what is wrong "in the full knowledge of what is  right in the given circumstances." In the familiar passage to which Owen refers Aristotle does claim that for the incontinent person to know in this way "would be  astonishing" (thaumaston, 1147a9) or "would be taken to be strange" (deimnon 1146b35). But this does not establish that Aristotle takes it to be

  

Page 88

evident a priori that the incontinent person does not know that he is doing what is wrong in this way. For one thing, Aristotle knows that strange and astonishing things  do turn out to be true (e.g., that the diagonal of a square is incommensurable with the side, Metaphysics I.2 983a15–17). 37 More importantly, he takes it as  necessary to argue for another account of the knowledge which the incontinent person has as an alternative to this one and, more significantly, to argue against  attempts to defend the astonishing result. In both cases the argument depends clearly on a posteriori matters of fact. In the former case (1147a11–18) Aristotle  appeals to what "we see" in the case of people who are asleep, mad, or drunk. In the latter case his argument continues as follows: The fact that someone [who is incontinent] makes claims from his own knowledge is no indication [that he is genuinely activating this knowledge]; for those who are in such  conditions [of being asleep, mad, or drunk—which are like incontinence] even recite scientific demonstrations and verses of Empedocles. (1147a18–20)

Here, Aristotle evidently uses a claim which he has just established—namely, that the effect of strong appetites on knowledge in the incontinent person is on a par with  the effect of strong drink on knowledge in those who are drunk—to undermine a powerful argument designed to show that incontinent people do sometimes fully well  know that what they are doing is wrong. This argument is that they sometimes openly and correctly say such things as: "I know what I'm doing is wrong."38 Aristotle's  reply is in effect this: "Saying correctly that one knows something, even with apparent conviction, is no sure indication of genuine activation of one's knowledge of that  fact in case one is asleep, mad, or drunk. In such a state saying something coherent with conviction is not even a sign of genuinely understanding what one is saying.  And incontinence is like these states." This argument is clearly not a priori. It is based on, among other things, an appeal to our common experiences when asleep or drunk. But the argument is a crucial part  of Aristotle's defense of the claim that the incontinent person does not operate "in the full knowledge of what is right in the given circumstances." This makes it quite  unlikely that Aristotle intended to defend this claim on a priori grounds. It is however true, as Owen claims, that Aristotle does dismiss the suggestion that practical wisdom can be coupled with incontinence with the retort: "That is absurd ...  not a single person would venture to say [that]" (1146a5–7). But does Aristotle think it absurd simply in view of linguistic conventions or "the established use of the  words," as Owen says? Aristotle's own appeal here is in effect to the fact that absolutely everyone would reject the claim with the strongest conviction. He introduces  no further qualifications. This would, however, be an adequate basis for rejecting the claim out of hand, as Aristotle does, if on the rules of the method that he is using  here the strongest basis for rejecting (or accepting) some thesis is simply and without further qualification that it would be unequivocally rejected (or accepted) by  everyone. This, as we have seen, is just what the rules of dialectic and

  

Page 89

peirastic require. Given this, it seems that Aristotle is here simply making a standard dialectical move without illuminating what it is that provides the ultimate epistemic  backing for such a move. XII—  Revision of the Standard View As noted above, the current philosophical climate is less hospitable now than it was when Owen wrote to appeals to what is analytic and known a priori on that  ground; and this has led to greater caution in attributing appeals to the analytic to historical figures. So it is perhaps unsurprising that more recent writers working within  Owen's framework have shifted ground somewhat. They have argued that while the force of dialectical argument does not depend ultimately on the appeal to what is  analytic and in that way absolutely a priori, it may involve an appeal to what is a priori in a different way. The suggestion has recently been made in various quarters that the value of the appeal to endoxa derives from the fact that the endoxa for Aristotle are not to be  understood as a set of convictions that may or may not correspond to some belief­independent objective reality where belief­independent evidence may be sought for  their truth or falsity. "Aristotle does not make the clean separation between evidence and opinion on which this (position) presumes." The endoxa are, rather, a set of  convictions which, at least when properly refined, "hang together to constitute a world" in such a way that, ultimately, the accord of any beliefs with these convictions  is "criterial of their truth." 39 A typical development of this approach begins with the suggestion that, for Aristotle, the endoxa are simply one subclass of the  phainomena (the appearances) where the phainomena are to be understood as constituting the world as it exists for us. The phainomena, so understood, may of  course conflict but in resolving any conflicts, on this view, we must square our results with the phainomena and show that these results preserve the phainomena as  true, or at least that the greatest number and the most basic are preserved. In deciding which are most basic or authoritative, on this proposal, Aristotle considers our  actual practices in justification to see, in different areas, what judges we in fact rely on. This reliance, however, does not itself require further justification for him. The  reason for this is that "appearances and truth are not opposed, as Plato believed they were. We can have truth only inside the circle of appearances, because only  there can we communicate, even refer, at all." This does not absolutely prohibit the introduction of new views, but it does require that any new view draw on and  preserve beliefs which are "deeper'' than the ones which the new view forces us to give up—deeper in that "the cost of giving them up would be greater, or one we are  less inclined to pay." Given this, we can see how certain endoxa, those that are deepest or most endoxa, can be in a way a priori, by being "unrevisable, relative to a  certain body of knowledge" or even so basic that they "cannot significantly be questioned at all from within the appearances, that is

  

Page 90

to say the lives and practices of human beings, as long as human beings are anything like us." 40 In considering the merits of this type of view, it is important to keep in mind the following. As developed in the literature, the view is not always explicitly offered as an  account of why the method of dialectic, or its peirastic part, is a useful and necessary tool to use in science or elsewhere but as an account of Aristotle's views on the  proper method for the justification of belief quite generally. If the proper method is taken to be dialectic then these will, of course, come to the same thing. But even if it  is not, in crucial respects the rules of the method as described here coincide with the rules of the method of dialectic and peirastic as we have seen it laid out by  Aristotle in the Topics and elsewhere. In particular, the rule that where the things which are apparent to us conflict we are to prefer the ones which reflect what we as  a group are more inclined, or more strongly inclined, to accept is just the main rule of dialectical method as a tool for inquiry. Peirastic in particular argues exclusively  from what is most endoxon in an absolute sense, i.e., what is most apparent to us without qualification. So if the justification offered for this procedure were indeed  Aristotle's then that would enable us to see what the basis was for his confidence in dialectical method and in peirastic in particular. This view has many of the same advantages in accounting for what Aristotle says about dialectic and, in particular, peirastic as those listed earlier for Owen's view. If  peirastic argues ultimately from items that are constitutive of reality and impossible for us to intelligibly question, then that would explain why its premises are true and  known and accepted (at least implicitly) by everyone. Moreover, these advantages are achieved on this view without the difficulties that come with the introduction of  the notion of analyticity in interpreting Aristotle. However, the view would still not account for how it is that the premises of peirastic in particular are consequent on,  and explained by, the first principles of the various sciences. In addition to this, there are reasons for doubting whether this is an adequate account of Aristotle's views on method quite generally. To begin with, there are the two  respects noted earlier in which Aristotle's views on method in science go counter to the main rule given above. Aristotle takes it that in science the data of perception  always take precedence over endoxa without reference to how widely held, or how deeply entrenched, those endoxa are.41 Secondly, he holds that there is a  procedure for the justification of first principles in science that goes beyond their ability to preserve, or their inferability from, what is most endoxon. In general, his  doctrine is that such principles must exhibit and conform to what is most intelligible simply or by nature by contrast with what is most intelligible to us. This requires,  as is reflected in his discussion of peirastic, that these principles must be capable of explaining the appropriate endoxa on some subject rather than simply preserving  them or being coherent with or derivable from them. In addition, these principles must be capable of explaining the relevant perceptual data whether these count as  endoxa, or what is most endoxon or apparent to us, or not.

  

Page 91

It may be responded to this that the main rule as described above can accommodate these points. 42 First, it can accommodate the idea that in science perceptual  phainomena always take precedence over endoxa if one of the things that is most apparent and endoxon to us just is the rule that in science perceptual data always  take precedence over endoxa. Against this suggestion, however, a number of things can be said. To begin with, when he says we must give the top preference to  perceptual data in science, Aristotle does not say or suggest that the reason for this is that people in general adopt this rule. Moreover, Aristotle nowhere says or  suggests that people in general in his day even held such a rule. This rule was almost certainly only held by a very small minority of philosophers and scientists at  best,43 and it seems unlikely that people in general held any clear views on scientific method in Aristotle's day. Science was after all a quite new affair, and the vast  majority of people were no doubt entirely without views on the matter of scientific method. Aristotle does claim, as we have seen, that people in general go with the  experts where they have no contrary views of their own. But if most experts on scientific method rejected Aristotle's principle in his own day it follows that from the  point of view of the judges we trust or what is acceptable "to us" Aristotle should have rejected it. Another, deeper, question is how such a rule could ever be one  which it was rational to adopt given Aristotle's general view on the proposal in question here. On the main rule proposed, Aristotle's view is not that we should follow  the rules of method which are apparent, or most apparent, to us, but that we should follow whatever is overall most apparent to us with no special standing given to  rules of method as such. Given this general principle how could we, or Aristotle, ever adopt the specific rule that in determining what should be accepted in science,  perceptual data always take precedence over other appearances? On the general rule, the alternate specific rule that we should accept what scientists say, including  their perceptual reports, where and only where it is compelling in the light of the total scheme of what is apparent to us overall, is clearly a preferable rule. But if so,  then there is no way on this proposal, of justifying the specific rule which Aristotle adopts—always to prefer perceptual data in determining what counts in science no  matter what trouble this creates for our system of beliefs overall. It is, if anything, more difficult on the proposed general rule to see how to accommodate the specific rule that when it comes to scientific principles what is more  intelligible simply, that is what explains the appropriate perceptual data (whether these are endoxa or not), takes priority over what is more intelligible to us. The  general rule is in effect that what takes precedence is always what is more intelligible to us. How could it be that on the basis of this we decide to adopt the rule that  for scientific principles what is less intelligible to us but more intelligible simply is to be preferred, or the rule that for scientific principles whether they are endoxa or not  is irrelevant (Posterior Analytics I.6 74b21–26)? It might be replied that this is in fact the way we do find it more intelligible to view scientific principles even when  these principles conflict with other things more intelligible

  

Page 92

to us. But this is not in fact Aristotle's own view. As we have seen, he supposes that people in general are prepared to find special scientific views acceptable only  when they do not conflict with other generally held views of whatever sort. Given this, there is no way for him, or for us as he describes us, to move from the principle  that we should always prefer what is more intelligible to us, to the rule that in our scientific principles we should always prefer what is less intelligible to us but more  intelligible absolutely. Even if the rule to prefer in science what is more intelligible simply to what is more intelligible to us is one acceptable "to us" if some specific  scientific principle turns out to be less acceptable to us overall than something which conflicts with it that we accept, we could not accept that principle, on the general  rule, even if we accepted the suggested rule of scientific method. Given these difficulties it is apparent that we cannot use the proposal offered to try to show what the value of dialectic, and in particular of peirastic, is for Aristotle.  On this proposal, Aristotle regards the appeal to what is most apparent or most endoxon as ultimately an appeal to what fixes the limits of what is intelligible for us and  thus cannot intelligibly be questioned or challenged. Peirastic is based on the appeal to what is most intelligible to us. But since we have seen that Aristotle does not in  fact view even what is most intelligible to us as ever, for that reason, beyond question or challenge by reference to perceptual data or to what is less intelligible to us  but more intelligible absolutely, this will not help us to see what the epistemological basis of dialectic, or its value for science, is. XIII—  Dialectic in Science: The Traditional View By contrast with these recent approaches, which see dialectic as drawing on the a priori, the most frequent traditional point of departure on this topic has been the  group of passages where Aristotle talks about the relation between endoxa, or common opinions in general, and truth. The capacity to grasp the truth is the same as the capacity to grasp what is like the truth. At the same time, people have an adequate natural inclination toward the truth and so  they turn out to attain truth for the most part. Thus, the one who can aim skillfully [in argument] in relation to the truth is equally able to aim skillfully [in argument] in relation to  the endoxa. (Rhetoric I.1 1355a14–18) The study of the truth is difficult in one way but easy in another. An indication of this is the fact that while no individual can adequately attain it still we do not all collectively fail.  Rather each person has something of value to say about the world and, though individually he contributes little or nothing to the truth, out of the assembly of all something  significant arises. Thus, insofar as the truth seems like the proverbial door which no one can miss, in this way the study will be easy. That we

  

Page 93 can have some grasp of the whole and not be able to grasp some part shows the difficulty of it. (Metaphysics II.1 993a30–b7) Some of these views [about happiness] are voiced by many and from long ago, others by a few renowned men; and it is reasonable that neither group is entirely mistaken, but  right on at least one point if not most. (Nichomachean Ethics I.8 1098b27–9) In all these matters we should try to get conviction through argument using the things which are apparently the case (ta phainomena) as evidence and as illustrations. For it is  best if everyone is seen to be in agreement with what we are going to say, or if not that all should agree in some way. (This they will do when their views are modified.) For  everyone has something of his own to contribute to the truth, and it is from this that we must give some proof of what we say. For from the things which are said correctly but not  clearly what is clear will be reached by those who advance, if at every stage they exchange things which are more intelligible for the jumble of customary assertions. (Eudemian  Ethics I.6 1216b26–35)

These passages have been the ones most often referred to in attempts to account for Aristotle's confidence in the method of dialectic. 44 In the first of these passages,  Aristotle offers the rather optimistic view that people in general have a natural propensity toward the truth that results for the most part in the attainment of truth.  Variants on this theme are found in the subsequent passages. These passages do not, of course, show (against what was argued earlier) that Aristotle thinks that the  method of dialectic or peirastic is on its own sufficient for reaching final results in science or elsewhere. In the first passage, Aristotle claims only that endoxa are for  the most part true. He explicitly points out in the second and third passages that while in some general way, or on some point, people may have made substantial  progress toward the truth, in some particular area or on various other points they may have got nowhere at all (cf. Metaphysics VIII.3 1029b3ff., Politics III.11  1281b15–21, 35–39). Nothing here conflicts with the point that new perceptual data that are not endoxa can have higher epistemic standing than endoxa. Still, if it is  the case that where people do have views those views are likely to amount to something, that might seem to offer a basis for using dialectic and peirastic as a part at  least of one's standard method of inquiry. But do these passages even provide an adequate justification for that? For one thing, since Aristotle allows that at some  times and in some areas the endoxa will not be reflective of truth, it seems that he must accept the conclusion that, sometimes, dialectic will not be useful in some  specific scientific or other inquiry. But, if not, then does he have a way of telling us when it is useful and when not? In addition, these passages indicate why Aristotle  might have thought it valuable to consider received views on some topic but what guidance do they offer on how to proceed when the received views conflict, as they  often do? In particular how can they explain why one might want to place more weight on views that are more endoxa and more familiar to us? Since

  

Page 94

this is the way dialectic and, in particular, peirastic proceeds if these passages do nothing to explain this they do not show why in science (or elsewhere) we should  make use of dialectic or, more specifically, of peirastic. The standard discussions of these passages do not address these problems and questions. Recent writers, in particular, have tried to deal only with the more limited  question of why it is, in Aristotle's view, that people more often than not come to hold true beliefs. The answer that they have offered is that, for Aristotle, (1) human  beings are by nature intelligent or inclined toward the truth and thus naturally capable of coming to true beliefs. In addition, for Aristotle (2) nature does nothing in  vain, i.e., natural capacities have a purpose that is for the most part realized—in this case the acquisition of truth. So (3) people for the most part come to true beliefs.  45 Whatever the merits of this argument, it does not tell us how to proceed when people's views conflict. So it does not help to justify the dialectical, or peirastical,  way of proceeding in such cases.46 But what can we learn from the argument itself? Those who have offered this argument on Aristotle's behalf have been careful to point out that Aristotle himself does not use this argument, and to claim only that he is  committed to its premises. The first question then is whether Aristotle is actually committed to the premises. He commits himself to the first premise of the argument, as  well as to the conclusion, in the passage from the Rhetoric quoted above. But what about the second premise? Does Aristotle believe that natural capacities or  inclinations are for the most part realized in the manner required here? That is not clear. Stones have the natural capacity to fall toward the center of the earth. But  does this mean that all or most stones at some time actually realize this capacity? No, since certain conditions must be met for the capacity to be realized (e.g., a stone  must be placed in a medium less dense than it to fall) and it is not clear that these conditions are at some time met, or that Aristotle thinks that they are, for all or most  stones (Physics II.8 199b18, 26; Metaphysics IX.5 1047b39ff., 7, 1048b37ff.). This point can be urged even more strongly for the case of intelligence since  intelligence is, for Aristotle, not merely a natural capacity but a "rational capacity," i.e., one which it is within the power of an agent to exercise or not depending on,  among many other types of conditions, the agent's overall motivational state (Metaphysics IX.5; cf. On Dreams 2 460b1ff.). So Aristotle cannot, and would not, accept the generalization, in premise (2) above, that natural capacities, including intelligence, are for the most part realized, in the  way required for reaching his conclusion. He would feel obliged to show that the appropriate conditions, motivational and otherwise, are met in a way adequate for  bridging the gap between the first premise and the conclusion. If we can understand how he would do this then we can perhaps understand how he would deal with  the questions and problems for his views and how he would defend his claim that dialectic and in particular peirastic is necessary in scientific inquiry.

  

Page 95

XIV—  Dialectical Justification in Science The beginnings of an approach to this matter are to be found in Aristotle's remarks in the passage from the Eudemian Ethics quoted above (1.6 1216b26–35). In  that passage, Aristotle introduces an elaborate justification for a certain methodological recommendation. The recommendation is that it is desirable to get conviction  (pistis) in a way that relies on appeal to the phainomena or what is apparent to us. His reason for that is that if we proceed in this way then our results will be in  agreement with what everyone holds, either simply or "when their views are modified." The reason that he offers to explain why it is desirable to be in agreement with  what everyone holds is that everyone has something to contribute to the truth to such an extent that "we must give some proof of what we say" by reference to  people's opinions, at least when they are properly modified. Aristotle does not say how he supposes this "modification" should take place. But the rarely used word  which he employs here (metabibadzein) is prominently used also in the Topics to describe the dialectical procedure of revising views by reference to what is ''more  endoxon" or "as endoxon as possible" (VIII.11 161a33–37, with b34–38 and VIII.5–6). Since the procedure that Aristotle is describing here also involves reasoning  from what enjoys the maximum consensus, it is likely that this is what he has in mind. Here, then, Aristotle explicitly introduces the claim we have been trying to understand, the claim that it is not simply a desirable or useful point of method to appeal to  the common opinions, or the most common opinions, on a topic but necessary. He goes on to give the reason why this is so. His reason is that by proceeding in this  way we guarantee that we move from what is correct though somehow unclear and jumbled up to what is clear and more intelligible and, so he must be supposing, it is  necessary that we proceed in this way. What exactly is this way of proceeding? Why is it necessary? And how does proof by reference to what everyone holds play a role in this procedure? The doctrine  that Aristotle develops here is expressed in quite similar language elsewhere, notably in the methodological passage with which he opens his Physics and in a related  passage in Metaphysics VII.3. The natural procedure is to go from things which are more intelligible and clearer to us to things which are clearer and more intelligible by nature. For what is intelligible to us is not  the same as what is intelligible simply. Therefore, it is necessary to proceed in this manner: from what is less clear by nature, though more clear to us, to what is clearer and more  intelligible by nature. The things which are first evident and clear to us are things which are rather jumbled up. Later, when these things are analyzed, from them the elements and  first principles become intelligible. That is why it is necessary to proceed from the comprehensive to specifics; for the whole is more intelligible in relation to perception and the  comprehensive is a sort of whole, since it embraces many things as parts. (Physics I.1 184a16–26)

  

Page 96 It is advantageous to progress in stages to what is more intelligible, since for everyone learning proceeds in this way—through things which are less intelligible by nature to  things which are more intelligible. Just as in the area of action our task is to start from what is good to each individual and make what is good altogether good to each, so it is our  task to start from what is more intelligible to oneself and make what is more intelligible by nature intelligible to oneself. Now what is intelligible and fundamental to individuals is  often barely intelligible and encompasses little or nothing of the reality of the thing in question. But nevertheless, one must try to understand the things which are altogether  intelligible by starting from things which are hardly intelligible but intelligible to oneself and progressing, as we have said, through these very things. (Metaphysics VII.3 1029b3– 12)

In these passages Aristotle begins by asserting, as he does in the Eudemian Ethics passage, that we must proceed by starting from what is more intelligible to us and  moving to what is more intelligible simply. The initial reason that he evidently offers for this, particularly in the Metaphysics passage, is the rather obvious one that we  cannot start anywhere else than with what we understand. This by itself does nothing to answer the question as to why we are required, as the Eudemian Ethics says,  to give some kind of proof of our final theoretical results, which may not be intelligible "to us," by reference to what is most intelligible to us. This question is  particularly pointed in view of the claim in the Metaphysics passage that what is intelligible to any given individual may have little or nothing to do with the reality of the  things in question. Aristotle does, however, go on to indicate in the Physics why we need to provide such a proof. He supposes that the elements and first principles  that are the things most intelligible by nature are the elements and principles of the wholes which are most intelligible to us. This is to say that the data that make up  what is most intelligible to us must on the whole be explained by the things which are more intelligible simply. This is the view we found also in the passage in the  Sophistical Refutations. There Aristotle characterizes the things in some area of scientific inquiry which are "common," i.e., commonly known and accepted, and thus  used by peirastic, as ''the things which are consequent on" the special principles of the subject (11 172a29–34 with a23–27). If this is so then it is easy to see why, as  the Eudemian Ethics says, there must be a kind of proof of the principles by reference to common opinions and why he characterizes these opinions, in the  Eudemian Ethics, as correct. But why do the principles have to be principles of and hence to explain such data. Why cannot most of our common opinions on some  subject, including most of the ones which are most intelligible to us (and thus most endoxa), simply be wrong and thus not in need of explanation by any theoretical  principles? The reason which Aristotle goes on to offer in the Physics passage quoted above in defense of his claim that what is most intelligible to us must be explained by  appropriate principles is that these opinions accord with "what is more intelligible in relation to perception."

  

Page 97

This fits with a claim that Aristotle repeats elsewhere: At different times different things are more intelligible to the same people—in the beginning the objects of perception, but as their knowledge becomes more accurate, the reverse.  (Topics VI. 142a2ff.) I mean by what is prior and more intelligible to us what is closer to perception, by what is prior and more intelligible simply I mean what is further from perception. The things  which are furthest from perception are, most of all, universals; the things which are closest are particulars; and these are opposite to each other. (APo. I.2 72a1ff.)

What is most intelligible to us, Aristotle thinks, always coincides with what is most closely related to what we perceive. Elsewhere he is quite explicit that the fact that  all or most people accept something "gives it credibility as something based on experience." 47 He specifically says in the Physics that this is why first principles must  account for what is more intelligible to us. If this is so then we can explain what the value of peirastic dialectic is for science in the following way. We have seen that peirastic dialectic is a procedure for the  testing of claims by reference to what is most endoxon and most intelligible to us. As such, in Aristotle's view, it is a procedure for the testing of claims by reference to  those beliefs which we have which are most closely connected with information which we have acquired by perception and for the rejection of whatever claims may  conflict with this information. So peirastic dialectic, in effect, turns out to be a procedure which draws on the information which we as a group are now warranted in  accepting on the basis of what is most obvious to us from perception. This explains how Aristotle can describe peirastic as drawing on what everyone knows and  how dialectic, in certain forms, can draw on what may reasonably be claimed to be true. This also provides a clear role for dialectic in scientific inquiry understood in  the way that Aristotle standardly understands it in passages such as Prior Analytics I.30, namely as a procedure which starts from the data of experience and  proceeds to find principles which conform to and explain such data. It enables us to coherently integrate Aristotle's remarks about the need for dialectic in science  with what he says in these passages. It also shows why dialectic has the limits that we have seen it has in scientific inquiry. Peirastic draws on what is in conformity  with the most empirically well­justified information that as a group we have up to now. But further observation by the individual scientist or others may yield new  empirical data that show that some theory which is most empirically well confirmed for us now, because it offers us the best explanation for the empirical data which  we have amassed up to now, is wrong. This may occur whether or not these now data are or come to be endoxa or generally known, i.e., a basis for further  dialectical or peirastic argument. This explains why Aristotle can say that in science we must try to account for the endoxa but also suppose that perceptual  observations take precedence over endoxa in validating theories.

  

Page 98

This account of the value of dialectic also explains why dialectic cannot by itself provide us with adequate theoretical principles. It serves to draw on and thus aid us in  collecting the empirically most well­confirmed data which theoretical principles must serve to explain. But by itself it cannot determine the principles which do the  explaining. This is why, nevertheless, it plays a necessary role in finding those principles. Since it serves to help collect the most empirically well­justified information  that we as a group have now it plays an essential role in our search here and now for first principles. For the best we can do here and now in our search for first  principles is to find those candidates for first principles which, among other things, do the best job of explaining the empirically most well­confirmed information that we  have now. XV—  The Merits of Peirastic Justification On the account which we have now reached of the epistemological basis of peirastic dialectic that method turns out not to embody any general appeal to what is a  priori. Nor does it embody, despite certain appearances to the contrary, a strong coherence procedure for the justification of belief. A belief is not justified in peirastic  dialectic simply when it coheres well with the maximal consistent body of existing belief or anything of the sort. Peirastic dialectical justification for a given belief does  depend on whether that belief can be ultimately defended by reference to those beliefs which are "most endoxa," i.e., most widely and firmly accepted by us. But even  here it is not because those beliefs are most widely and firmly accepted by us that the belief in question is justified by reference to them. Rather it is because those  beliefs bear a special relation to the data of experience. Peirastic dialectical justification, for Aristotle, like justification in science simply, turns out to give special  priority (though not precisely the same priority) to the evidence of experience. This does not mean, however, that perceptual judgments are, for Aristotle, incorrigible.  Though this is not the place for a full treatment of his views on this subject, it is clear enough that he does not take ordinary perceptual judgments to be in principle or  in general unrevisable (De Anima III.3 428b18–25). This is compatible, however, with his general assignment of final epistemic authority in science to perceptual over  theoretical beliefs (De Caelo III.7 306a13–7). In peirastic dialectic this result also holds. It follows from the doctrine that revision must always be made by reference  to what is most intelligible to us, together with the doctrine that what is most intelligible to us is always closest to perception. In this respect there is a kind of  foundationalist empiricist element in Aristotle's theory of justification both in science and in peirastic dialectic generally. 48 This result raises a number of interesting questions. Perhaps the most important are these: How would Aristotle try to show that the more widely and firmly accepted  something is the more likely it is to directly embody information acquired

  

Page 99

by perception? And, how would Aristotle justify the reliance which he places in "the testimony of the senses?" To answer these questions, and other related ones,  would require us to go into Aristotle's rather complex views on information acquisition; and this is not the place for that. Here it must suffice for us to see that, recent  claims to the contrary notwithstanding, the old cliche that Aristotle is fundamentally an empiricist is, after all, in certain respects basically right. 49 Notes 1. W. D. Ross, Aristotle, 5th ed. (1949), 59. Cf. F. Solmsen, Die Entwicklung der arist. Logic und Rhetorik (Berlin: 1929), 26: "analytics supercedes dialectic." 2. J. Barnes, "Aristotle and the Methods of Ethics," Rev. Int. Phil. (1980): 495; M. F. Burnyeat, "Good Repute," London Review of Books (Nov. 6, 1986). Barnes  prefers to refer to Aristotle's supreme method as "the method of endoxa," a description not used by Aristotle. But Barnes is, of course, clear that "endoxa form the  subject matter of dialectic.'' See also M. Frede, Essays in Ancient Philosophy (Minneapolis: University of Minnesota Press, 1987), 95 ("As with all other sciences,  one arrives at [the] starting points [of metaphysics] dialectically"). For further references, see R. Bolton, "Definition and Scientific Method in Aristotle's Posterior  Analytics and Generation of Animals," in Philosophical Issues in Aristotle's Biology, A. Gotthelf and J. Lennox, eds. (1987), n. 4. 3. The three views mentioned are defended, respectively, in G. E. L. Owen "Tithenai ta Phainomena," 1961 (in J. Barnes et al., Articles on Aristotle I, 1975), J.  Barnes "Aristotle's Theory of Demonstration" (in Articles on Aristotle I), and M. Burnyeat "Aristotle on Understanding Knowledge," (in E. Berti, ed., Aristotle on  Science, 1981). 4. See Owen 1961, 118; G. E. R. Lloyd, Magic Reason and Experience, 1979, 118 and T. Irwin, Aristotle's First Principles (Oxford: Clarendon Press, 1988),  ch. 2 et passim. 5. See G. Grote, Aristotle, 2nd ed. (London: Murray, 1880), 271; H. Cherniss, Aristotle's Criticism of Plato and the Academy I, (Baltimore Johns Hopkins  Press, 1944), 18; J. Brunschwig, "Aristotle on Arguments Without Winners or Losers," in Wissenschaftskolleg Jahrbuch 1984/5, (Berlin: 1986), 40. The views of  Grote and Cherniss are criticized by Owen in Aristotle on Dialectic (1986), 103ff. Brunschwig resuscitates a version of their account. 6. Brunschwig, "Aristotle on Arguments," 40. Brunschwig argues that there are "discrepancies" and "inconsistencies" in what Aristotle says about the use of dialectic  with ordinary people which "seem to show that he is uneasy" about the use of dialectic except for school gymnastics (34). He argues that Aristotle claims at 105a16– 19 and 157a18–21 that "dialectical deduction (sullogismos) should be employed only with trained dialecticians," while induction is to be used with ordinary people.  Since "the official subject of the Topics is dialectical sullogismos" the Topics teaches us how to proceed "with well­trained dialecticians not with ordinary people." Of  course, Aristotle flatly contradicts this conclusion at the beginning of Topics I.2. (The pragmateia of 101a26 is the one described at 100a23.) Is this inconsistent with  what he says later? At I.12 105a16–19 he says at most that induction is more effective with people in general, sullogismos is more effective with trained arguers, not  that one type of argument is exclusively to be used with either group. At VIII.2 157a18–21 he repeats the same point. (The reference here back to the earlier

  

Page 100

      passage shows that mallon means more not rather.) More importantly, he repeats the point in a context which clarifies its import. The contrast between the two  types of argument, inductive and syllogistic, is introduced in this context to describe two possible ways of obtaining the necessary premise for a sullogismos (VIII.1  155b20–22, b35–156a11). That is the topic under discussion. Later in VIII.8 Aristotle even describes a premise of an inductive argument which yields a premise for  the main syllogism as itself a "syllogistic premise," i.e., a premise concerned with the (main) syllogism (160a35ff.). Thus, at VIII.2 157a18ff. when Aristotle is  recommending induction for use with ordinary people it is for the purpose of obtaining the necessary premises for a sullogismos. (One reason for this  recommendation is given at 156a3–7, cf. 105a16–19.) He recommends sullogismos for the same purpose in dealing with dialecticians. (One reason is given at  156a7–11, with 156b26–28.) So Aristotle does not later contradict his claim in Topics I.2 that one use of the mastery of the dialectical sullogismos (100a22–4) is  for discussions with people in general (101a25–27). His later remarks only concern how best to obtain the necessary premises for such a sullogismos in discussion  with different groups. (See also the references to the securing of premises at VIII.2 157b28ff., 158a14ff.)       At VIII. 14 164b8ff., as Brunschwig notes, Aristotle says that dialectic should not be used with "everyone," and that gymnastic dialectic should not be used with  "just anyone." Some people are too competitive in argument for this, he says. But there is no suggestion that these hyper­competitive people are the ordinary people  rather than, say, mainly trained dialecticians, or simply some of each. 7. As we shall see in detail below, the successful syllogistic reasoning, from (appropriate) endoxa, for a certain conclusion, which Aristotle mentions in this passage,  he typically takes to be, in a two­party dialectical discussion, the work of the questioner (Top. VIII.6 160a12–13). In this case the proposition which is "dialectically  justified" is opposed to a claim which the answerer is set to defend. The successful submission to argument, from endoxa, without having to say anything inconsistent,  is, in a dialectical discussion, the work of the answerer (Top. VIII.5 159a38ff.). In this case the proposition which is consistent with the (appropriate) endoxa and  thereby "dialectically justified" is a claim which the answerer is set to defend. Just what the relation between and what the value of these two modes of justification is  will be explored further below. The term 'dialectical justification' is used here in a technical sense to refer to whatever kind or kinds of justification it may be that the  questioner or the answerer, or both, have in either of the two cases. 8. See, particularly, T. Irwin, "Aristotle's Methods of Ethics," in Studies in Aristotle, vol. 9, D. O'Meara, ed. (Washington D.C.: Catholic University of America  Press, 1981), 207–8. 9. The difficulty was discussed by J. Burnet, The Ethics of Aristotle (London: Methuen, 1900), intr. xli. Burnet solved it by treating dialectic as a method of  discovery only, not of justification. Justification was based, he thought, on intuitive awareness of the self­evidence of the propositions in question. Most recent writers  have solved the problem differently. 10. See M. Burnyeat, "Aristotle on Understanding Knowledge," 130–132; T. Irwin, "Aristotle's Discovery of Metaphysics," Rev. Meta. (1977), and Aristotle's First  Principles, (1988). Cf. Irwin, ''Methods of Ethics," 208. 11. I discuss this issue further in "Aristotle's Method in Natural Science," in Aristotle's Physics, L. Judson, ed. (Oxford, 1991). 12. See, for instance, HA 513a8–15, GA 742a16–8; HA 532b18–26. Aristotle also often

  

Page 101

      relies on his own new observations which clearly do not (yet) count as noted or accredited beliefs (endoxa). See, e.g., Mete. I.6 343b8–14. For fuller discussion  of these points, see Bolton 1987, sec. 2. 13. This needs certain qualifications which are, however, not important in the present context. See Bolton, 1987, n. 8. Aristotle himself usually states the point without  qualification, as at Met. 995b23–4. 14. Of course, the endoxa with which any dialectically adequate definition must be consistent include endoxa about definitions. And it might be an endoxon principle  about definitions—though it seems it was not in Aristotle's day—that definitions should serve to explain a certain range of empirical data. But even if this principle were  endoxon that would not permit dialectical argument to secure it that a given definition which does in fact explain such a range of data is the correct definition—even for  someone who knew that the definition in question satisfied this principle. It would have to also be endoxon that the definition satisfies the requirement before it could  be concluded, by dialectical argument, that the definition is the correct one. Only endoxa can figure as premises in dialectical argument.       That the principle mentioned above was not endoxon in Aristotle's day is confirmed by the fact that the Topics mentions no such principle in its list of standard  procedures (topoi) for the examination of proposed definitions. It does mention the doctrine that a definition should be given by reference to what is prior and more  intelligible absolutely (Topics VI.4 141b25ff.). This doctrine is used elsewhere by Aristotle in his construction of his own view that ultimate definitions should explain  the perceptual data on a given subject (Physics I.1 184a18ff.). But no such application of this doctrine appears in the Topics. The only use of the doctrine in the  Topics is to argue that definition must be by genus and differentia since these are prior and more intelligible absolutely than the species. This latter point obviously  reflects standing doctrine. 15. For fuller discussion of this point, see Bolton 1987, esp. sec. 5. 16. See above n. 6. I here translate sullogidzesthai in a traditional way as 'reason syllogistically'. For convenience below I shall often speak of syllogistic reasoning  simply as reasoning. Inductive reasoning will always be described as such. 17. Aristotle contrasts here procedures for training with those for "learning." Whether the latter is or is not understood here as a species of dialectical procedure  requires discussion. (See below n. 21.) This does not matter for present purposes; all that is relevant here is what is not required in gymnastic dialectic. The difficult  term gnorimoteron is usually translated here 'more intelligible'; the intelligible is not simply what can be understood but what somehow makes good sense. 18. This does not mean that they must be accepted as true by the answerer and/or the questioner. In Top. VIII.11 Aristotle points out that in dialectical discussion for  "training" one sometimes is required to argue from false premises. But still these premises should be things that "appear so" (to phainomenon, 161b4). 19. See Brunschwig (1984/5) 33. 20. Cf. lian periphanes, 158al. Here we are, no doubt, dealing with what is most endoxon in an absolute sense. For further discussion of this see below. 21. This elaborate procedure where premises are not simply accepted or rejected but subject to dialectical objection and scrutiny is one which Aristotle says is  inappropriate when one is doing dialectic for gymnastic or training purposes. Rather, he says, it is to be used by "one who is trying to learn" (VIII.3 159a4–14 et al.).  It has been supposed that in

  

Page 102

      saying the latter Aristotle is indicating that this procedure is not to be used in dialectic but only in (nondialectical) discussion for learning (see E. Poste, Aristotle on  Fallacies (London: Macmillan, 1866), 102). But this is impossible. In the passages referred to above, Aristotle clearly has in view the use of this elaborate procedure  in dialectic itself. This is one clear indication that Aristotle thinks that a certain kind of nongymnastic dialectical discussion can be used for learning. It is worth noting,  however, that this learning procedure is not the one in view at SE.2 165b1–3, which Aristotle distinguishes from dialectic, since there the answerer who learns never  objects to the questioner/teacher's premises but always accepts them (as Poste sees, 102). 22. For a term to be held to be a middle­term is presumably for it to figure in two accepted propositions in the appropriate way (as subject of one and predicate of  the other), and for the syllogistic conclusion from the two propositions to be accepted as holding by virtue of the truth of the two propositions (see APr. I.1 24b18– 22). 23. Elsewhere in the Topics Aristotle can simply equate what is more intelligible with what "we know better:" A proprium of a thing must be given by reference to what is more intelligible than the thing. For instance, someone who claims that being most like the soul is a proprium of fire  refers to the soul, and this is less intelligible than fire, because we know better what fire is than what the soul is. (V.2 129b9–12)

24. On implicit beliefs as endoxa, see Barnes, "Methods of Ethics," 501. 25. Certain passages in the Ethics, assuming the method there is dialectical, further confirm this: Some object that what everything aims at is not thereby good; but there is nothing at all in what they say. For what is accepted by everyone that, we claim, is so; and one whose  purpose is to undermine this conviction will hardly have anything more credible to say. (Nicomachean Ethics X.2 1172b35–73a2) Is it then practical wisdom which resists [and is overcome in incontinence], since it is the strongest state? That is absurd since the same person will be at once practically wise and  incontinent on this view; and not a single person would venture to say that someone who is practically wise is the sort of person who would do the worst things willingly.  (Nicomachean Ethics VII.2 1146a4–7)

26. At 104b24 enantion tais doxais is substituted for paradoxos (cf. b19). Translators have sometimes used 'contrary to general opinion' or the like to translate  paradoxos. This has the strong authority of 104a8–12 behind it, and that passage at least indicates that this describes a clear case of a paradoxos view for Aristotle  and the case which is relevant for the characterization of a dialectical protasis. But 104b31–34 suggests (with b19–28) that a paradoxos view of an expert may also  be one which conflicts with the received view of the (other) experts, where there is no conflict with majority opinion. 27. See Brunschwig (1984/5) 34. Cf. Ross, Aristotle, 5th edition (London: Methuen, 1949), 27. 28. The evidence for this is conveniently assembled by Brunschwig, Aristote: Topiques, I (1967), xix.

  

Page 103

29. W. D. Ross, Aristotle's Metaphysics (Oxford: Clarendon Press, 1924), v. 1 260ff., C. Kirwan, tr. and ed., Aristotle's Metaphysics IV–VI (Oxford: Clarendon  Press, 1971), 84–5, Moraux "La joute dialectique d'apres le huitieme livre des Topiques" in Aristotle on Dialectic, G. E .L. Owen, ed. (Oxford: Clarendon  Press, 1968), 288–9 n. 3, followed by Brunschwig (1984/5), 35. These writers also charge Aristotle with a further inconsistency, that of describing peirastic  sometimes not as a part of dialectic, or as distinct from dialectic but as identical to dialectic. This charge is based in part on Aristotle's use in Topics VIII.5 of the term  peira, trial or testing, in describing dialectic quite generally. But the term peira, like the verb peiran, is an ordinary Greek word used in Plato, Aristotle, and elsewhere  to denote any kind of trial or testing, including any type of dialectical examination. The term peirastikê, by contrast, is a philosopher's term of art, not used even by  Plato and introduced with a technical definition by Aristotle in SE 2. There is no reason to suppose that any example of dialectic which Aristotle would describe as  peira must fit the technical requirements of peirastikê. Secondly, these writers claim that at SE 11 172a21, Aristotle says that dialectic "is peirastikê." But this just  means, in context, that it is dialectic which "is capable of peirastic testing" not some special epistêmê. This does not mean that everything that dialectic does is  peirastic, only that peirastic is a capacity of dialectic. Cf. 171b9 (with b4–5) and Met. 1004b22–26, where the same point is made. 30. The popularity of the alternate manuscript reading is no doubt influenced by the fact that in certain passages, such as Rhetoric I.2 1358a1ff., Aristotle draws a  contrast between those dialectical or rhetorical arguments which draw on material which may be applied in common (koinoi, 1358a12) to all or many subjects and  those which draw on matters which are unique (idia, a17) to a particular subject. But this is not the same contrast as the one which Aristotle draws in the above  passages. In the above passages Aristotle uses the term koina ("the common things") to describe and delimit the materials for dialectical, or peirastic, reasoning quite  generally. In Rhetoric I.2 1358a1ff., by contrast, the term is used to refer only to one type of basis for a rhetorical or dialectical syllogism by contrast with another  (a10–21, cf. I.3 1359a16–26). In fact, Aristotle claims there that most rhetorical syllogisms are based not on the koina, as he uses that term there, but on things which  are special to the subject at hand (1358a26–28; cf. III.22 1395b30–96b18). In SE 9 and 11, however, all dialectical or peirastic examination is based on koina, as  the term is used there (170a34­b1, 171b4­7, 172a27­b1). So the two uses of the term are quite different. For further discussion of this distinction and of the nature of  peirastic in general, see R. Bolton, "The Problem of Dialectical Reasoning (Sullogismos) in Aristotle," Ancient Philosophy, Special Issue, 1994. 31. Owen, "Tithenai ta Phainomena," 117. 32. Owen, "Tithenai ta Phainomena," 115. 33. Owen also makes the claim that "endoxa rest on experience even if they misrepresent it. If they did not Aristotle could find no place for them in his  epistemology" (117). He does not explain, however, what it is for endoxa to "rest" on experience. If this is to be compatible with his claim that endoxa are often  analytic and known a priori, it presumably cannot mean that all endoxa depend evidentially on or are known by experience. 34. M. Nussbaum apparently has in mind a version of this line in saying that the ultimate justification for the appeal to appearances is that only within the circle of  appearances can we "refer" at all. The Fragility of Goodness (Cambridge: Cambridge University Press, 1986), 257. See below for further discussion of her views. 35. Various different accounts have been offered in the literature of what sort of form

  

Page 104

      and content this crucial reference fixing opinion has in Aristotle's theory. It is suggested below that 'thunder is a certain noise in the clouds' is an example of such an  opinion. Other accounts of the form these opinions take are currently being developed. The differences among these accounts do not affect the point in question here. 36. See GC I.4 319b8 ff. on the relation between hupokeimenon and pathos. 37. See the passages in Bonitz, Index, under thaumadzein. 38. Aristotle's phraseology here is unusual—legein tous logous tous apo tes epistêmês (1147a18–19). This means at least that the claims about good and bad  which incontinent people make may express knowledge which they actually have, as in the use of apo epistêmês at Metaphysics I.4 985a16. But the emphasis in the  added use of the article tes—"from his knowledge"—suggests that the understood verb with this phrase is not simply gignesthai (come) but rather a verb such as  lambanesthai (are taken), which would indicate that Aristotle has in mind the affirmation of certain things selected by the incontinent person as things which are  known. In this event Aristotle is addressing himself to a tougher objection to his own view than is usually supposed. 39. The quoted phrases are from M. Burnyeat, "Good Repute," J. McDowell, "The Role of Eudaimonia in Aristotle's Ethics" in Essays on Aristotle's Ethics, A.  Rorty, ed. (Berkeley and Los Angeles, California: University of California Press, 1980), 372 and M. Nussbaum, The Fragility of Goodness, 242. The description of  the development of this position is mainly drawn from Nussbaum (240–258). A similar, though not obviously identical, development is suggested by Burnyeat and by  McDowell's discussion (369–373). McDowell's remarks are, however, made in the course of reflections on Aristotle's method in ethics, and it is not clear that he  would take them to apply in other areas. Burnyeat and Nussbaum take the method to be universally applicable for Aristotle. In T. Irwin, Aristotle's First Principles,  it is argued that in his later scientific works (including the Metaphysics and De Anima) Aristotle introduces a new form of "strong dialectic" which draws on endoxa  which are known a priori in a Kantian sense. They are not constitutive of reality but necessary preconditions of the possibility of knowledge (or significant thought)  about reality. Since Irwin, unlike other writers, is not trying to account for the role of dialectic in science, as dialectic is understood and described in the Organon, but  only as dialectic is later used (but never described), his views require separate treatment. Reasons for doubting that Aristotle introduces in the Metaphysics a new  form of dialectic, or uses a form which appeals to this sort of a priori knowledge, are given in "Aristotle's Conception of Metaphysics as a Science," in Unity, Identity  and Explanation in Aristotle's Metaphysics, T. Scaltsas et al., eds. (Oxford: Clarendon Press, 1994). 40. Nussbaum, 254, 257–8. 41. Nussbaum (268, 274, n. 12) takes the perceptual phainomena to be one subset of the phainomena, understood in general as what we think or what we say. But  there is no reason to suppose that whatever careful observers perceive, as Aristotle understands this in passages such as Apr. I.30, is always something that we  believe or say. 42. Cf. Nussbaum, 248–50. 43. We do not know of any Pythagorean or Eleatic or Platonist, or of any Atomist or other post­eleatic pluralist who held this view. Certain medical writers may have  held a view like this, but they constituted a small minority among theorists on scientific method. 44. See, among others, G. Grote, Aristotle, 2nd ed., (1880), 273; W. F. R. Hardie, Aristotle's Ethical Theory (Oxford: Clarendon Press, 1980), 38; Barnes,  "Methods of Ethics," 506ff.

  

Page 105

45. See Barnes, "Methods of Ethics." 46. J. Cooper (review of The Fragility of Goodness, Philosophical Review (1988): 553–4) has argued that "it stands to reason" from Aristotle's perspective that  those who seem to be acute and who exercise their intelligence more, namely 'the wise', are more likely to come to the truth than ordinary people. Thus for Aristotle  their views are preferable in case of conflict (cf.551). However, this is not Aristotle's own rule in peirastic, or dialectic generally; if it were dialectic would lose its  required effectiveness with people in general (Top. I.10 104a11–12). Moreover, in fact, as Aristotle knows, the special views of "the wise" tend to be massively  conflicting. This means that the views of the wise cannot in fact be for the most part true but rather are mainly false. Since Aristotle never says without qualification that  they are mostly true but only says this about the views of people in general, we lack the necessary support for attributing to him the (implausible) doctrine that the  views of the wise are more often, or more nearly, true than the common views of people in general. 47. Div. In Som. 462b14–16. 48. There is not space here to consider the role of nonpeirastic forms of dialectic in science. Aristotle is clear that periastic is the form of dialectic which is especially  relevant in scientific, or other philosophical, inquiry (Met. IV.2 1004b25–26, quoted above at the beginning). The account developed here can be extended to cover  the other forms. 49. This is a corrected version of a paper first published in the collection, Biologie Logique et Métaphysique chez Aristote, D. Devereux and P. Pellegrin, eds.  (Paris 1990). That paper was itself an extensively revised version of the paper presented at the conference on the Ile d'Oleron whose proceedings appear in that  collection. The excellent penetrating comments of Jacques Brunschwig and Daniel Devereux on the original version served as the main stimulus for the revisions. They  led me to minimize discussion of controversial points whose defense is inessential to my main proposals. Brunschwig also kindly gave me at the conference a copy of  his valuable paper (see n. 5), which I had not seen before; so I was able to rectify my failure to consider it earlier.

  

Page 107

Chapter Five—  Choosing the Good in Aristotle's Topics Eugene Garver I Those of us who think about Aristotle's ethics often wish that he had not insisted that deliberation and choice were confined to things that promote ends as opposed to  ends themselves. This is not just a terminological discomfort. The division of labor between wishing for ends and deliberating about and choosing means threatens to  make the choice of means rational but amoral and to make the wish for ends moral but irrational, a fateful division reflected in modern distinctions between  considerations of rational prudence, which apply to an agent's own good, and considerations of morality, which are from the point of view of personal rationality,  irrational. 1  The model for deliberation in Nicomachean Ethics III.3 is geometrical construction, a paradigm of rational amorality—"Mathematical works do not have  an êthos because they do not show deliberative choice (prohairesis)" (Rhetoric III.15 1417a19–21). The discussion of deliberation and choice in the Ethics is  frustratingly brief. Choice is simply defined as deliberative desire (III.3 1113a9–12, see VI.2 1139b4), and the account of deliberation simply says that at each stage  we choose "the easiest or best means" (1112b17). To say that we select the best seems unhelpfully redundant, while saying that we pick the easiest seems often to be  false: there are many reasons to choose a more difficult course over an easier one. It might be easier to ask you to do something than to do it myself, or easier to do it  myself than more generously to defer to you. It might be easier to accomplish

  

Page 108

some end by choosing a means that is less noble than the route I should choose. Surely the Ethics does not advocate efficiency as a universal practical value. The brevity of the discussion in the Ethics of how ends are grasped is just as unsatisfying as the discussion of deliberation and choice. It offers only unhelpful talk  about perception and intuition. If an end transmits its goodness to the things that promote that end, why should it not by that token transmit its irrationality as well? If  the end does not transmit its goodness to the means, it is hard to see how Aristotle can legitimately distinguish phronêsis from cleverness. Confining the obviously  rational activity of deliberation to means toward a given end seems to drive a wedge between morality and rationality. If someone's end is to count the blades of grass  in my lawn, partitioning the area into manageable sections by a geometrical construction would hardly count as rational. 2 Since Aristotle's remarks in the Ethics about the nature of deliberation and choice are so brief, it might seem promising to turn to a more extended treatment in Topics  III, the whole of which is devoted to arguments about what is "more worthy of choice" ( ) (III.1 116a4). "The same topics are useful also for showing  that something is simply worthy of choice or avoidance" (III.4 119a2) as well as worthy in the comparative degree, so we could hope for enlightenment about the  nature of choice in general. The brief remarks in the Ethics beg for amplification, and the Topics might bring it. Topics III elaborates a large number of arguments  designed to establish which possibility is more worth choosing. There are two large difficulties in trying to use the Topics, though. First, Topics III is about hairesis, while Ethics III is concerned with prohairesis. The Topics talks  about hairesis because it is a term in ordinary usage, while the Ethics makes prohairesis into a technical term. Second, it offers arguments for something being worth  choosing, but it completely ignores the restriction of choice to means, the restriction that caused discomfort in the first place. The first sentence of Topics III does  announce that it will discuss, as I said, arguments about what is more worthy of choice, but what is more worthy of choice ( choice or better of two (or more) things, must be examined in the light of the following considerations" (116a4–5).3

): "[w]hich is more worthy of 

It is important to understand the relation between these two senses of choice, the one limited to what we can do to achieve a desired end, and the other simply  equivalent to what is better. I want to use the Topics better to understand the nature of deliberation and choice as they are employed in the Ethics, not to note an  inconsistency and put it down to Aristotle's intellectual development or a divergence in task between the two works. The two senses of deliberation and choice  articulate a central ethical, practical problem. To take just one example, Ethics X and Politics VII show that the life of theôria is better than and should be chosen  over the political life. It is not clear whether that is a practical

  

Page 109

conclusion, that one should decide to act in ways that make one's life as theoretical as possible, or whether it is a verdict about which life is best that does not have any  necessary practical consequences. 4  It may be that once we know that theôria is the best life, all we can do is pray that we come to be in a position to enjoy it. In  general, at issue in the restriction of deliberation to means is the relation between judging what is best and deciding what to do. II The equation in Topics III between what is more worthy of choice and what is better is no hyperbolic opening sentence simply meant to engage the reader. There are  three crucial ways in which the details of that discussion of choice, by equating the choiceworthy with the good, looks incompatible with the Ethics. (1) As we have  seen, the Ethics restricts choice to means as opposed to ends, yet the Topics talks about choosing some things for their own sakes, others as means to something  else. (2) Where the Ethics restricts choice to things we can do, the Topics talks about choosing things that exist by nature or through chance. (3) The Ethics insists  that we should choose what is best in particular circumstances, while the Topics claims that what is good haplôs, simply or absolutely, is better than, and hence more  to be chosen than, what is good for someone in some situation. Choice in the Topics is not relative to an end, not relative to actions, and not relative to agents and  their circumstances. To expand choice beyond practicable and appropriate means contradicts choice in the narrower sense. I want to argue, in spite of this apparent contradiction, that the topics of argument Aristotle presents in Topics III do in fact flesh out the restriction on choice in the  Ethics, so that choosing becomes an ethical activity, and not an amoral act in service of possibly moral ends, what Aristotle calls cleverness. The sense in which choice  becomes an ethical activity will preserve the distinction between means and ends, the one apprehended by phronêsis and the other by ethical virtue. If my argument  succeeds, Aristotle's restriction of deliberation and choice to means—pros ta telê—will become a strength rather than a defect in his account. First, though, I want to give some examples of these three expansions of choice in Topics III, to things chosen for their own sakes, to choices that are disconnected  from action, and to things good in the abstract as opposed to what is good for the agent in current circumstances. (Since many of the same arguments are found in  Rhetoric I.7, whose subject is also the comparative good in deliberation, I will give some examples from there as well.) Topics III begins with the proviso that  deliberation and choice are concerned not with "things which are widely separated ... from one another (for no one is at a loss to

  

Page 110

decide whether happiness or wealth is more worthy of choice), but it is concerned with things that are closely related and about which we ... cannot detect any  superiority of one over the other" (116a5–10). It is hard to see how my examples are not cases where the superiority is not so obvious that deliberation is impossible. (1) We are told, in opposition to the restriction of deliberation and choice to means in the Ethics, that That which is worthy of choice for its own sake ( ) is more worthy of choice than that that is so for some other reason; for example, health is more worthy of  choice than exercise, for the former is worthy of choice for its own sake, the latter for the sake of something else. (III.1 116a29–37; see also III.3 118b22–27) The end is usually regarded as more worthy of choice than the means to the end ( That which is in itself more noble ( worthy of choice than wealth, and justice than strength. (116b37–117a4)

) (116b23)

) and more valued and more praiseworthy is more worthy of choice; for example, friendship is more 

You must consider whether one thing is worthy of choice for its own sake and the other for the impression which it makes on others, for example, health as compared with beauty.  (III.3 118b20–27; see also Rhetoric I.7 1364a1–5, 1365b)

Since ends are better than means, we choose ends over means. But the narrower sense of choice in the Ethics seems to have a point: I see that health is better than  exercise, but how could I choose health over exercise? (2) Topics III presents arguments about what is more worth choosing when not all the alternatives are things we can do. The Ethics tells us that people deliberate  about things "attainable by their own actions," and not human affairs generally: "We wish (not only for results we can achieve), but also for results that are (possible,  but) not achievable through our own agency, e.g., victory for some actor or athlete. But what we decide to do ( ) is never anything of that sort, but  what we think would come about through our own agency" (III.2 1111b24–6); "No Lacedaemonian deliberates about the best form of government for Scythia" (III.3  1112a33–1112b1). The Topics, by contrast, tells us that "our friends should be just is a thing which we choose for its own sake, even if it is not going to affect us at all, and even though  they may be in India" (116a37–39). "That which is in itself the cause of good is more worthy of choice than that which is accidentally the cause of good; for example,  virtue is more worthy of choice than luck" (116b1–2). Compare Ethics III.3: "We do not deliberate about the results of chance, such as finding a hidden treasure. The  reason why we do not deliberate about these things is that none of them can be effected by our agency" (1112a27–30). In the

  

Page 111

Ethics we deliberate about things within our power. Choosing things within our power over the operations of luck in the Topics looks like we are deliberating whether  to deliberate and choosing whether to act on choice. ''Everything is preferable ( ) at the time when it has greater importance; for example, freedom from pain in old age is preferable to freedom from pain  in youth, for it is more important in old age" (2 117a26–29; see also Rhetoric I.7 1365a). It is hard to imagine facing such a choice: when would I face it? And  similarly with Aristotle's next argument: "That is preferable which is more useful on every occasion or on most occasions, for example, justice and self­control are  preferable to courage, for the first two are always useful, but courage only sometimes" (117a35–37). What practical consequences could that truth have? How could I  choose self­control over courage? These topics look quite similar to the arguments about choosing the life of contemplation over the political life, so seeing just how  these topics can be principles of choice will be important. (3) "That which is good absolutely is more worthy of choice than that which is good for an individual, e.g., the enjoyment of health than a surgical operation" (III.1  116b8–10). If   here means "better," then the claim is obviously true. But if it means something connected to choice, it is not clear how this could  guide action: since the enjoyment of health is better than a surgical operation, should I choose not to have the operation? Can I choose to enjoy health instead? That the good absolutely is more worthy of choice than that which is good for an individual contradicts the Ethics. Consider this argument from Book V: Since the unjust person is greedy, he will be concerned with goods—not with all goods, but only with those involved in good and bad fortune, goods which are, (considered)  unconditionally, always good, but for this or that person not always good. Though human beings pray for these and pursue them, they are wrong; the right thing is to pray that  what is good unconditionally will also be good for us, but to choose (only) what is good for us. (V.1 1129b1–7) What is useful to a particular person is better than what is generally useful (Rhetoric I.7 1365a35). (See too Politics VII.3 1325a34–40.)

Justice depends on distinguishing the choiceworthy from the better, and choosing what is best for me, relative to my circumstances, and not grasping for what is better  as such. The Topics in this way seems to advocate injustice! Here too it seems that the expansion of choice from practical means to whatever is better has produced a  sense of choice that contradicts the narrower, deliberative sense, and contradicts it not only logically but ethically.

  

Page 112

III In spite of the trouble it causes, it is important for Aristotle to separate means from ends. Without that distinction, we would also lose the distinction between the plural  moral virtues and the single practical intellectual virtue of phronêsis, and so between two kinds of goodness. The goodness and choiceworthiness of an action is not  simply inherited from its end, as it is in instrumental actions. The goodness and choiceworthiness of an action is not independent of its ends, either, as in the craft values  of technê and the "hired gun" conception of rationality, which is not found in Aristotle. Phronêsis occupies a difficult position, resisting reduction to either ethical virtue  or technê, because the goodness of action has to resist these two reductions. What kind of goodness is the goodness of intelligently chosen means? Why is that kind  of goodness, located in phronêsis, a necessary part of the good and happy life? At issue in the relation between the restricted and the broader senses of what is more  choiceworthy is the kind of goodness that phronêsis has. One way to raise that central ethical issue is by asking the textual question about the relation between the  principles of phronêsis and the topics of Topics III. While the Ethics seems to make trouble by restricting deliberation and choice to means, the Topics gives us a contrary problem from the other direction. Once  Aristotle identifies what is best and what is more worthy of choice, then choice is reduced to impractical evaluation or approval of what is best. Choice in that sense  might be connected to goodness, but not to an ethics tied to action. If the choiceworthy is simply identified with the good, then we will need some other way of  deciding what to do. The whole point of deliberation and choice is that knowing what is best does not by itself tell us what to do. That is why we have to deliberate.  There is good reason for the distinction between what is choiceworthy and what is best. In spite of these evident differences between the sense of choice limited to means that we can do and are appropriate to our circumstances in the Ethics, and the idea  of choice identified with the good in the Topics, I want to show how the topics Aristotle offers in Topics III can function as principles of deliberation and choice.  (Whether showing how deliberation and choice can be moralized in addition helps to rationalize the wish for ends is another question.) The impractical looking topics  of argument Aristotle presents are standards by which we make choices. Deliberation is a search for middle terms that connect what I want with what I can do. The  topics are reasons for choosing one possible means over another, and therefore reasons for choosing a means. 5  The basic presentation in Topics III is comparative,  in terms of what is more worth choosing: it is through making comparative judgments that we can understand the principles of choice. To talk about choice in the  comparative degree allows Aristotle to talk about choice and the good without having to define either.

  

Page 113

Phronêsis is the ability to find middle terms that connect what I want with what I can do, and the phronimos will use the topics to conduct his search. In the Ethics  Aristotle talks about choosing means to an end rather than choosing among alternatives. He recognizes that we do choose among alternatives, but all he says is that we  choose the easiest and best. In Aristotle's Rhetoric: An Art of Character, 6  I argue that the "topics" in the Rhetoric substitute for principles in domains or circumstances where scientific  reasoning from principles does not obtain. Topics are not principles in the scientific sense, immediately known premises for scientific demonstration. Instead, the topics  are lines of reasoning that promote good choices by aiding us in the search for middle terms, that is, for reasons. These principles are not premises, so that they are not  moral truths. They are substantive, and ethical, inference rules. Some contemporaries like to distinguish "rules" from "principles" because rules seem to have no  quantifiers; Aristotle frequently expresses these topics not as unquantified propositions but as imperatives, directions for searching.7  They are like maxims of  interpretation. ''Respect precedent" is not a premise, but a principle of interpretation.8 Like scientific principles, and unlike the archai of the ethical virtues, the topics are impersonal. For that reason, they are incomplete relative to the task of practical  reason, and still require the ethical virtues. When all is going well, the dialectical determination of what is best, the orectic determination of what I want, and the  deliberative decision about what I should do, will all coincide. When that is the case, then phronêsis is at the same time the orthos logos of the ethical virtues—their  formal principle—and the reasoned ability to choose what is best to do. Phronêsis will be equidistant from technê and the ethical virtues. It would be disastrous if these principles were used as premises. We need contrary principles of choice to function effectively as practical agents. Conflicting premises  are inadmissable, but contrary principles need not be troublesome. Take an example from the Rhetoric. Sometimes it is better to choose one means because it is  scarcer, and so using it requires more skill and attention, and deserves more praise. And sometimes one should choose a means because of its abundance, so, for  example, to use it now will not use up the resource (I.7 1364a20). There is no overarching meta­principle that will say when scarcity is a guiding value and when  abundance. There is another reason these principles cannot be premises. Premises authorize and justify their conclusions, but rhetorical and dialectical topics do not. If a particular  act instantiates the categorical imperative, I am obliged to perform it. But my using the topic that ends are better than means by itself cannot settle a question or close  off deliberation. The topics are means of reaching a conclusion, but they do not lend authority to their conclusions, since they have no authority to transmit. They  transmit reasonableness. They are principles of choice; they do not determine my choice for me.

  

Page 114

These principles would offer the wrong kind of guidance if used as premises. If we had such premises for practical reasoning, cleverness—a science of the good or a  theory of rational choice—would substitute for phronêsis, and rhetorical skill could take over from statesmanship. We would, as the Topics directs, choose ends over  means, the practicable over other kinds of cause, and what is good haplôs instead of what is good in the circumstances. (The injunction, for example, that we count  each person as one and no one as more than one is a modern variation on the last of these rules.) But none of those make sense as a rule of action. Such a science is  incompatible with the agent­centered and practical orientation of Aristotle's Ethics. It would simply reduce the restricted sense of choice to the more expansive one,  and say that somehow Aristotle did not mean, or should not have meant, what he said about the restriction of deliberation and phronêsis to means. The connection  between the topical principles and practical reason must be more indirect than that reduction. Just after the first sentence of Topics III announces that it will look at how to decide which is the more worthy of choice or the better of two (or more) things, Aristotle  adds the limitation that such inquiry is worthwhile only if the choice is not obvious to perception (see also III.2 117a5). When the choice of means is not obvious, it can  be regulated by reason. IV I said in the last section that phronêsis had to have a kind of goodness all its own if it is to escape reductions to either moral virtue or amoral cleverness. There I  suggested that we could formulate that issue as one of the relations between the topics, the substantive modes of argument, offered in Topics III and the principles of  phronêsis. The topics could not be premises; when they function as premises, they become the principles of cleverness. How then do they function? In order to  illustrate just how these principles of choice can function in deliberation by offering directions for searching for middle terms, in spite of their impractical look, I want to  return to my series of examples that broke beyond the restriction of choice to the practical by identifying the choiceworthy with the good: (1) Things chosen for their own sake are, we were told, more worth choosing than means. Health, thus, is more worthy of choice than exercise. How could one ever  choose health instead of exercise? How could there be a deliberative situation in which health and exercise were alternatives? The whole point of exercising being a  means to health is that I only exercise when I do not directly have health. Because exercise is a means not an end, I choose it, for the sake of health. Health is an end,  and therefore not something I choose. It is not just that the Topics contradicts the Ethics but that the rule of the Topics is incoherent.

  

Page 115

On the contrary, I think that the line of argument that tells us to choose ends over means helps us to deliberate. It is important to remember that we choose to exercise,  or so Aristotle thinks, not because exercise itself is good, but because it leads to health. We get things wrong when we try to be as healthy as we can in order to  exercise, as when we try to get rid of all body fat in order to run ultra­marathons. We start out making money in order to have material goods and security, and then  we come to sacrifice personal comforts for more money; this rule gives us grounds for criticizing that transvaluation of values. The rule of the Topics can, then, be a useful reminder and a maxim. But are "maxims" and "reminders" principles of choice? 9  They are if they help us decide what to  do, and not just what would be best. Something of the point of this topic appears in the discussion of friendship in the Ethics. "Most people wish for (boulesthai)  what is fine, but decide to do (prohaireisthai) what is beneficial; and while it is fine to do someone a good turn without the aim of receiving one in return, it is  beneficial to receive a good turn'' (VIII.14 1162634–1163al). The benefactor follows the Topics by doing what is fine, and so finds his actions satisfying. The  recipient follows the Ethics by choosing the useful, but is on that account morally inferior. To make things worse, the recipient follows the Topics in choosing not  something that he does but something someone else does for him. This discrepancy between agent and recipient can create problems. Aristotle notes: "[t]he  benefactor's action is fine for him, so that he finds enjoyment in the person he acts on; but the person acted on finds nothing fine in the agent, but only, at most, some  advantage, which is less pleasant and lovable" (IX.7 1168a10–12). We have to choose to do what is useful, but most people's wish for what is fine is not an idle wish.  This topic shows that as far as possible we should arrange our lives so that the noble and the useful coincide, and in that way we can choose to do things that are their  own end. I want to stay with that topic for a moment, because it shows something else about the nature of practical reasoning. Because we should choose things that are their  own end over means, someone who realized in the heat of battle that philosophy is better than military victory might therefore stop fighting. Such a man would use the  wrong principle of choice for that situation. As my example of friendship shows, it is better to give than to receive. It does not follow that regardless of circumstances, I  should give and not receive. Theôria is better than the political life. It does not follow that we should engage in politics only in order to make theôria possible. But it  does follow that it would be wrong to engage in theôria in order to improve the quality of political life. Dialectic, like rhetoric, is a faculty for proving opposites. The Topics exhibits contrary lines of reasoning for choosing, and there are no lines that escape this sort of  contextual test and become categorical. Phronêsisconsists in the intelligent employment of this faculty, and not in an escape from these principles to categorical ones.  It is, as Ethics VI. 12 reminds us, virtue and vice

  

Page 116

that apprehend the principles of action. As I said before, the topics are means of reaching a conclusion, but they do not lend authority to their conclusions. That way  lies cleverness. They are principles of choice; they do not determine my choice for me. It is not the fault of the faculty of reasoning that someone would use the wrong  principle of choice. As he puts it in the Rhetoric, It is a function of one and the same art to see the persuasive and (to see) the apparently persuasive, just as (it is) in dialectic (to recognize) a syllogism and (to recognize) an  apparent syllogism; for sophistry is not a matter of ability (dynamis) but of deliberate choice (prohairesis) (of specious arguments). In the case of rhetoric, however, there is the  difference that one person will be (called) rhêtôr on the basis of his knowledge and another on the basis of his deliberative choice, while in dialectic sophist refers to deliberative  choice (of specious arguments), dialectician not to deliberate choice, but to ability (at argument generally). (I.1 1355bl5–21)

There are, consequently, no lines of argument, no considerations, that only phronêsis and not cleverness will employ. All three of the expansions of choice, beyond  means to ends, beyond the practicable, and from the good I should choose to the good haplôs, present opportunities for abuse. At the Same time they represent ways  the phronimos aligns choice in the limited sense to the overall human good. We can develop the relation of phronêsis to dialectic by reviewing more of our examples. The next example I gave in the last section was similar: "the end is usually  regarded as more worthy of choice than the means to the end." This too could be true, but does not look useful in regulating our choices. But the sentence continued:  "and of two means that which is is nearer to the end" (see also Rhetoric 1.7 1365a). Although neither honor nor wealth is the final good, it is better to choose  something honorable than something wealth­generating. My goal is to understand the nature of choice. Studying the Topics is one means I use; writing some more  macros to help me use my word processing program more efficiently is another. Whenever I find myself writing more macros, I, could use the reminder that that  activity is less choiceworthy than studying the Topics. The same pattern of reasoning occurs in practical choices that are more serious than my deliberations about writing philosophy. Among the many lines of reasoning  advanced in debates about the exclusionary rule forbidding the use of evidence gathered in violation of the Fourth Amendment, both sides appeal to this Same topic.  Opponents of the exclusionary rule characterize it as a deterrent to police misconduct and recommend other sanctions as more directly preservative of the Fourth  Amendment. If it is a means distant from its end, then it could be replaced by any other method that better Secured the end. It should be replaced by any method,  such as making police misconduct a tort, that is closer to the end itself.

  

Page 117

Proponents of the exclusionary rule, on the other hand, argue that it is directly mandated by the Fourth Amendment and so should not be characterized as a deterrent.  Since it is itself a way of obeying and enforcing the Fourth Amendment, it is not a means to its realization, but a part of the Amendment itself. Therefore, even if other  methods should greater reduce the amount of police misconduct, the exclusionary rule is not optional. 10 And So Aristotle criticizes the practical virtues since they are not chosen for their own sakes ( crucial place in the ergon argument:

) (X.7 1177b18). He uses this topic at a 

A thing pursued as an end in itself is more final than one pursued as a means to something else, and a thing never chosen as a means to anything else is more final than things  chosen both as ends in themselves and as a means to that thing, and accordingly a thing chosen always as an end ( something else. (1.7 1097a31–b7)

) and never as a means to 

The final example of choosing something for its own Sake was: "[Y]ou must consider whether one thing is worthy of choice for its own sake and the other for the  impression which it makes on others, for example, health as compared with beauty" (III.3 118620–27). At this point I think we can see that imperative as regulating  choice, too. We can forget that something is only worth choosing and doing because of its effect on others, and that something else is better because it is worth  choosing for itself. If I try to become attractively Slim by purging myself after meals, I have deliberated toward a lesser end instead of a greater one, sacrificing health  for body. In general, making deliberation and choice expansive enough to include choosing an end over a means saves ethical deliberation from being reduced to instrumental  reasoning because it allows us to treat subordinate ends, that is, all ends other than the supreme end of happiness, as ends while still subordinate. Unless deliberation  and choice are themselves ethical acts, acts of a virtue called phronêsis, all things deliberately chosen must be means. But when we choose to act virtuously for the  sake of happiness, our virtuous actions are not means to happiness. It is through the topics of Topics III that we can see how ends can be items in deliberation while  still being ends. (2) The second divergence between the Topics and the Ethics was that the Topics abandoned the restriction of choice to things that we could do, implying that we  could, but should not, choose things we could not do. It told us that we should choose what we could do over what happens by chance. I see a useful principle of  choice here. If I should pay off a debt, it is better that I call the person to whom I owe money, rather than hope to run into her in the market. It is a principle of  practical rationality and phronêsis that we should choose things we can do over means that rely on natural forces or on chance. "That of which a

  

Page 118

man would prefer ( ) to that of which he would wish another to be the cause; for example, friends are preferable to money" (III.3 118b8–10). 11 We run  into this topic in the Ethics and Politics, in the arguments for the Superiority of theôria based on the fact that theôria is an activity more within our own power and  control than political activities. We then wonder whether we are supposed to do anything based on the fact that theôria is the best kind of happiness. It may be best,  but how is that a principle of choice? I gave a couple of other examples from Topics III.1 of Aristotle saying that something was more choiceworthy than something else, when choice between them  seemed impossible. It seemed impractical to prefer a painless old age to a painless youth. There is no end toward which a painless old age is a better means than a  painless youth. But it makes sense to say that attention to the place of pain in one's life should increase as one ages. Similarly, it might be better to choose security  when one is older than when one is young. I was twenty­nine years old when I began my first tenure­track job. My chairman was ten years older, and I remember a  conversation with him during my first term there. I was complaining about many features of the job: students were told to give as little attention as possible to their  philosophy courses so that they could concentrate on their vocational studies, and we had to spend more time avoiding student complaints and declining enrollments  than truly teaching. My chairman advised me to put these irritations in perspective: this place has a really good retirement plan, I was told. I thought that I was too  young to take the retirement plan seriously in my own decision about whether to stay at that school. I thought my chairman was too young to have it dominate his  thoughts too. I chose to pay attention to some features of my position and not others, and chose to try to find a better job. Along similar lines, Aristotle told us to prefer justice and self­control to courage in the Topics. I see two ways in which that principle of choice could be practically  useful. First, it is worth considering that justice and self­control are, as Aristotle says, always desirable, and not just desirable in peace, when martial virtue is not  needed. When I choose to act courageously, I should realize that I am not choosing to act courageously instead of justly and temperately. Second, when we choose to act courageously, we should remember that courage is virtuous only because certain undesirable conditions call for courageous action.  Under those conditions, courage is called for. We should not ignore those conditions and think that courage is better than it is. Where these lines of argument are useful  reminders, they are also useful in criticizing proposed courses of action, and seeing which choices withstand criticism. That is how topical reasoning is supposed to  work. (3) Finally, the Topics tells us that goods haplôs are preferable to what is good for the individual. In this case the Topics seems to contradict the Ethics and Politics,  which tell us that the individual should choose what is good for

  

Page 119

him. 12 I think that the distinction between what is best haplôs and what is best for me to do in the present circumstances is in fact the fundamental ethical distinction.  It underlies the other two, the means/end and impractical/practical goods distinction, and should occupy our attention more than the means/end relations that have  dominated the secondary literature. In deliberative situations, when choice is about what we can do to effect some end, I should choose what is best for me in current circumstances. Thus the Ethics.  When choice is identified with what is best, then we choose what is best haplôs. So says the Topics. This looks like a serious divergence between principles of  dialectical argument and the substantive doctrine of the practical sciences. However, in the Topics too Aristotle shows that he is aware that it would be a mistake  always simply to identify the more choiceworthy with the better, as he did in the opening sentence of Topics III: Superfluities are better than bare necessities, and sometimes also preferable. For living a good life is better than merely living; and a good life is a superfluity, while life itself is a  necessity. Sometimes better things are not always preferable; for it does not follow that, if they are better, they are also preferable. For example, to be a philosopher is better than  to make money, but it is not preferable for him who lacks the necessities for life. Superfluity exists, when being already in possession of the necessities of life, a man tries to  procure some noble accessories. We shall perhaps not be far wrong if we say that the necessary is preferable, while the superfluous is better. (III.2 118a7–15) ("Better" translates  beltion throughout and "preferable" hairetoteron.)

Similarly on the Ethics side. As in the lines I quoted from Ethics V, we should choose what is best in the circumstances, while hoping that what is best haplôs be  good for us. A man is good for whom the things good by nature are good. For the things men fight about and think the greatest, honor and wealth and bodily excellences and pieces of good  fortune and powers, are good by nature but may possibly be harmful to some men owing to their characters. (Eudemian Ethics VIII.3 1248b27–31; see also Rhetoric 1.9 1366635– 1367al; Metaphysics Z.3 1029b5–7.)

The difference between goods haplôs and choices of what is best relative to an agent's circumstances shows why we need both meanings of choice. It is wrong to  choose what is best haplôs instead of what is best in one's situation. It is equally wrong, though, to infer that because something is choiceworthy in particular  circumstances it is therefore choiceworthy. The first is the sort of cleverness that thinks it offers universal rules for practice: Always maximize wealth, or individual  freedom. The second is the opposite sort of cleverness, which abandons universality and continuity for decisions about what is best in circumstances, without regard  for how good those circumstances themselves are. Only phronêsis can make us safe from these two ways of substituting reasoning for character. The intellectual virtue  of phronêsis is dependable

  

Page 120

without being predictable or deriving its consistency from a universal rule, and it is flexible without being merely opportunistic. If I am confident that after death I will spend my time conversing with a better class of people than is now available to me, that does not mean that I should pull back  from the commitments and pleasures I have with my current friends and family, as would happen if I ignored the restricted sense of choice where I choose what is best  for me rather than what is best haplôs. Acting wholeheartedly toward those people is what I should do in my present circumstances. I should treat my friends as  friends, not as the best who happen to be around. I use this particular example to fend off charges of egoism against Aristotle. The idea that we should not do what is best as such but what is best for us, even if that is  an ethical best, is often thought to be reason to convict Aristotle of egoism. The idea that we should be friends with our current friends, and not always be in the  market for better ones, shows that the restriction of choice to the best for us in particular circumstances does not imply egoism. Ethics can be practical and agent­ centered without being egoistic. Now I want to turn to a more difficult example along the same lines. Assume, with Aristotle, that being a politician is a better, happier, way of life than being a professional musician. The fact that if I were a better person, I would want to  spend more time in political activity than in playing music does not necessarily imply that I, such as I am, should now engage in more political activity. I could be better  off spending my time doing what I do best and find most fulfilling. On the other hand, it is also possible that devoting myself more to politics might be a way I could  become a better person. Aristotle, and we, need both the restricted and expanded senses of choice and good, the one to keep it practical and the other to keep it nonrelative and hence moral.  I said earlier that the topics or principles of choice should not function as premises but in a more indirect way. These examples show something of what that more  indirect way is. The topics are principles by which we choose. They are not what we choose. They allow us to find actions that are related to our ends as means, as  practical, and as appropriate to circumstances. All the arguments I either invented or cited using the topics from Topics III are instances of deliberation about means, not choice in some sense abstracted from the  practical. Although the arguments for goodness extend beyond choice in the narrow sense, they are reasons for choosing a means toward some other end. They are  therefore ways connecting the good of good choices to goods as such, and so to the human good. Through them, phronêsis can be at once an intellectual and an  ethical virtue. Through them, phronêsis can lead to activities which are part of living well, but which do not usurp the role of the moral virtues.

  

Page 121

V The Topics, then, supplies forms of argument for choosing one thing over another. These modes of argument ignore the restrictions of choice to means, to things we  can do, and to what is best for the agent choosing in the particular situation calling for choice. Rather than contradicting the restricted sense of deliberation and choice  in the Ethics, though, we do and should deliberate according to these broader arguments. So long as we do not use them as premises but as the kind of topical  principles they are, there is no contradiction but harmony between Topics III and the picture of deliberation and choice in Ethics III. The two senses of choice need  each other. The Topics presents the faculty of dialectic, a rational dynamis. Proficiency in such a power will make no one any wiser about anything (Rhetoric 1.2 1356a32–34).  The Ethics and Politics use these argument forms, so there is no contradiction between the works. But the purpose of the Topics is not practical science. Rational  dynameis are, in that way, amoral because of their rationality: to be a rational power is to be able to cause, and prove, opposites. Excellence in dialectic and rhetoric  is not knowledge about anything because faculties are by their nature incomplete. Dialectic and rhetoric, as faculties, must be only about means, not ends. When  dialectic and rhetoric forget their indefinite nature, they cease being subordinate to ends outside themselves, and then they presume to practical wisdom. Cunning  emerges, and amorality becomes immorality, when Such calculative ability presumes to be self­sufficient. Craft values, such as efficiency, replace moral values.  "Rhetoric dresses itself up (hypoduetai) in the form (to schêma) of politics, as do those who pretend to a knowledge of it, partly from lack of education, partly from  boastfulness, and other human causes" (1356a27–30). Plato enacts that presumption in the sophists' idea that the power to persuade, and the political power that is  supposed to follow from it, are ends in themselves. The differences between the faculty of dialectic and the sciences of ethics and politics accounts for the difference between the restriction of choice to means, with ends  supplied by moral virtue in the Ethics, and the unrestricted identification of the choiceworthy with the good in the Topics. The topics that I have called the principles of  choice constitute the faculty (dynamis) of affirming and denying, choosing and avoiding. The energeia of that faculty is phronêsis. The faculty uses the broader  meaning of choice in which it is identified with the good, while its perfection and virtue uses the restriction of deliberation and choice to means, to what we do, and to  what is appropriate. Because dialectic is unrestricted and undefined, the principles of choice are indifferently about what is choiceworthy and what is good, even while the only things  chosen are means and actions. The principles of choice use the broad meaning of choice which identifies it with good, while choices themselves are

  

Page 122

of means. What is better should influence one's choices as much as possible. It is desirable, and a part of living well, to align the practical and the impractical good. To  effect that alignment directly is vicious, as we have seen in Aristotle's definition of injustice, but it is equally wrong to ignore what is best haplôs in choosing what is best  for oneself. Wishes are not totally idle. Perceiving and affirming sub specie bonum and sub specie faciendum are not identical nor completely independent of each  other. I began by saying that readers often regret the fact that Aristotle restricts deliberation, choice and phronêsis to means and the practicable, reserving wish and ethical  virtue for ends. The conjunction of Topics III with the Ethics shows that Aristotle had good reason to separate means and ends, phronêsis from ethical virtue. Only  under this restriction can, as he says, pursuit and avoidance stand to desire as assertion and denial are to thought (Ethics VI.2 1139a20; Topics 1.1 104b 1–8). That  is, only under this restriction can practical desire and choice obey the law of non­contradiction, transitivity, and some of the rules of inference. I can desire, wish, hope,  and pray for all kinds of contradictory things. I can wish for the impossible. The price of rationality is this restriction to the practicable. 13 Notes 1. Note that this is just the criticism which Kant makes of eudaimonism in the Critique of Practical Reason, 62: If our end is enjoyment, then "the end itself, the  enjoyment we seek, is not a good but only well­being, not a concept of reason but an empirical concept of an object of sensation. Only the means to it, i.e., action, is  called good (because reasonable deliberation is required for it)." On the wedge between morality and rationality, see, e.g., Michael Slote, "Reply to Commentators,"  Philosophy and Phenomenological Research 54 (1994): 711: "Common­sense rationality is agent­favoring just as common­sense morality is agent­sacrificing." 2. For the purposes of this essay, 'means' is simply a translation of pros ta telê, things that are "toward" an end, without any further commitments about the relation of  such things to ends. The problem set by Aristotle's restricting choice and deliberation to pros ta telê is not solved by reminders that that expression is not confined to  means as actions distinct from the ends themselves. I sharpen the knife so that it will cut better. I cut with the knife in order to practice surgery. Both are good relative  to an end. In that way, both are pros ta telê. If surgery were not good, neither would be my cutting open your abdomen. If cutting were not good, neither would  sharpening the knife. The problem remains: rationality and morality are not intrinsically connected. 3. See also I.11 104b 1–8 for a similar equation of the choiceworthy and the better. Brunschwig notices the problem of the narrower and wider senses of hairesis,  and translates the Topics with the wider meaning by using souhaitable for haireton and préférable for hairetoton. See his note on 154. Aristote: Topiques, v. I, J.  Brunschwig, tr. (Paris: Société D'Édition 'Les Belles Lettres,' 1967).       There is a similar problem in the Rhetoric. Deliberative rhetoric advises and

  

Page 123

      dissuades people concerning particular choices of actions, while epideictic rhetoric concerns praise and blame. Consequently the scope of epideixis is much wider  than that of deliberative rhetoric. "[N]ot only a man or a god is praised but inanimate objects and any random one of the other animals" (1.9 1366a28). And yet  "praise and deliberations are part of a common species in that what one might propose in deliberation becomes encomia when the form of expression is changed. Thus  when you want to praise, see what would be the underlying proposition; and when you want to set out proposals in deliberation, see what you would praise" (1.9  1367637; cf. 11.23 1399b32–1400a4). 4. I explore these problems about choosing theôria in "Living Politically and Living Rationally: Choosing Ends and Choosing Lives" (forthcoming). In this chapter, I  claim that in looking at the Topics one can ignore the differences between hairesis and prohairesis, between seeing something as preferable and choosing to do  something, while in the final chapter I insist on the difference between the two. 5. These principles of deliberation resemble what Henry Richardson, Practical Reasoning about Final Ends (Cambridge: Cambridge University Press, 1994), calls  "offstage principle(s) of selection among alternative means, directing one to pick the easiest and best." 6. University of Chicago Press, 1995. 7. Stocker contains a similar argument against a certain kind of "teleological" reading of eudaimonia, which simlarly misconstrues means/ends relations. Insofar as (acting courageously in defence of and for the sake of one's city) is understood teleologically—as being done for the sake of courage, virtue or eudaimonia—it too  easily seems an instance of posturing, self­indulgence, preciousness, self­servingness, or something else or similar unattractiveness.... But we can understand it nonteleologically:  the courageous action is done for the city and from and with courage. Courageous action here would not be for courage, but would be guided by its canons and criteria. It would  be done from an appreciation of what courage involves, why one should act courageously, why courage is here and now called for, and so on.

      Michael Stocker, Plural and Conflicting Values (Oxford: Clarendon Press, 1990), 69.1 take it that Stocker here shows that one cannot use "Act courageously"  as a major premise in a practical syllogism. Its function is as a rule, like my topics, rather than as a premise. 8.  I have explored the uses of such principles in constitutional interpretation in "Philosophy, Rhetoric and Pluralism," Rhetoric and Hermeneutics in Our Time,  Walter Jost and Michael Hyde, eds. (New Haven: Yale University Press, 1996), 171–195. 9. Casting the topics as reminders foreshadows the way, in the history of rhetoric, topics moves from being topics for finding arguments, to topics of arrangement or  judgment and topics of memory. 10. I take the example from Malcolm Richard Wilkey and Stephen H. Sachs, "The Exclusionary Rule: A Prosecutor's Defense," Criminal Justice Ethics 1 (1982),  abridged and reprinted in Readings in the Philosophy of Law, John Arthur and William H. Shaw, eds. (Englewood Cliffs, N.J.: Prentice Hall, 1984), whose  pagination I use. Wilkey claims that the "alleged value" of the exclusionary rule "has always hinged on the assumption, still unproved after seventy years, that it will  deter police illegality.... The

  

Page 124

      Constitution prohibits unreasonable searches and seizures: It may mandate a remedy to enforce that prohibition, but nowhere does the Constitution mandate the  exclusion of all other possible remedies" (262). Sachs's reply quotes Justice Day's majority opinion in the first case announcing the exclusionary rule, Weeks v. United  States: "Acts violating the Fourth Amendment "should find no sanction in the judgment of the courts, which are charged at all times with the support of the  Constitution." In other words, for illegally seized evidence to be admissible would make the judicial system complicit in constitutional violations. As the Court put it in  Mapp v. Ohio, "All evidence obtained by searches and seizures in violation of the Constitution is, by that same authority, inadmissible in a state court." (Both quoted  on 268.) 11. There is a problem with this topic. It is also possible to argue in the opposite direction, since dialectic is a faculty that proves opposites. There is no general rule  that what is done through one's own efforts is always better than things that happen naturally or through other causes. E.g., "We deny that we work hard in order that  we may be thought gifted" ( ) (III.2 118a23). "What is naturally good is more worthy of choice than that which is not so by nature, e.g., justice rather than  the just man; for the former is naturally good, whereas the goodness of the latter is acquired" (III. 116b 11–12). For a use of this topic in the Ethics, see, e.g., IV  1120b11–13: "Those who have not acquired their property by their own efforts, but have inherited it, seem to be more generous; for they have had no experience of  shortage, and, besides, everyone likes his own products more than (other people's), as parents and poets do." But see III.2 117b12, where "what is nearer to the  good is better and preferable, and also what is more like the good; for example, justice is preferable to the just man." 12. See also Politics VII.13 1332a21–25: "A truly good and happy man, as we have stated elsewhere in our arguments on ethics, is one who by the nature of his  goodness (which is absolute) has advantages at hand which are absolute advantages." 133200: "We pray that the composition of the state be lucky enough to be  supplied with the goods which depend on fortune, for we posit fortune as being sovereign over them. The virtue of the state, on the other hand, is the work not of  fortune, but of knowledge and intention." See also Ethics VIII.2 1155621–27. 13. Gilbert Harman, Change in View: Principles of Reasoning (Cambridge, Mass.: MIT Press, 1986), 108–9. It is important to distinguish ends from means. It would be bad if alternative means to an end were treated in the same way as alternative ends. In the case of alternative ends, what  one normally does is to try to obtain both ends through scheduling. Doing this would be silly for alternative means to a given end. It would be silly to pursue both means in  sequence. Clearly, the rational thing is to choose a particular means and disregard the other. Similarly, in more complex cases in which different courses of action have different advantages, it is a point in favor of a course of action that it would satisfy more desires than  another, at least if these desires are the same in strength and would be satisfied to the same extent. But this is true only if the desires are ultimate desires. It would be silly to prefer  B to A as a means of getting E just because B but not A achieves indirectly.

  

Page 125

Chapter Six—  The Normalization of Perplexity in Aristotle Gareth B. Matthews No doubt the most famous statement of Socratic perplexity (aporia) in philosophical literature is this speech of Meno's in the dialogue named after him: Socrates, before I even met you I used to hear that you are always being perplexed (aporeis) and that you make others perplexed (aporein), and now I think you are bewitching  and beguiling me, simply putting me under a spell, so that I am full of perplexity (aporias) . . . you seem . . . to be like the broad torpedo fish, for it too makes anyone who comes  close and touches it feel numb, and you now seem to have had that kind of effect on me, for both my mind and my tongue are numb, and I have no answer to give you. Yet I have  made many speeches about virtue before large audiences on a thousand occasions, very good speeches as I thought, but now I cannot say what (virtue) is. (Meno 79e–80b,  Grube trans., modified)

If we can call anything the canonical expression of philosophical perplexity, this passage is surely it. 1 Let's stop for a moment to reflect on the features of the state that  this passage delineates. A first point to mention is that the object of inquiry over which Meno has become perplexed is something everyone should understand; it is the nature of virtue. No  doubt there could be philosophical perplexity about some specialized or esoteric subject matter, say, black holes, or transfinite cardinals. But the paradigm cases of  philosophical perplexity are cases of bafflement over

  

Page 126

something at once basic and also, in Some ways, quite familiar to everyone, such as what virtue is, or what time is. Second, the person here who has become perplexed, Meno, is a perfectly intelligent adult with a good reputation for understanding the subject under discussion. Of  course almost anyone can become philosophically perplexed: a genius, a dullard, a child, an old­timer, a highly educated scholar, or an ordinary working person with  no academic Sophistication whatsoever. But the paradigm case of philosophical perplexity will be bafflement in someone, like Meno, with a somewhat better than  average claim to expertise in the matter under discussion who, at the beginning, thinks there will be no problem explaining whatever it is he is asked to explain. Third, once Meno has become philosophically perplexed, both his mind and his tongue become numb. That is, he becomes confused and cognitively disoriented; he  finds himself unable to use ordinary language in the normal and assured way that was previously characteristic of him. This feature of philosophical perplexity flows  naturally from the other two. Since the subject under discussion is so basic, finding that he does not know how to analyze it leads Meno to be unsure that he even  know how to use basic words, such as the word for virtue and the words for the several distinct virtues. The resulting disorientation is what Wittgenstein alludes to  when he writes, ''[a] philosophical problem has the form, I don't know my way about." 2 So far the features I have highlighted do not mark off anything peculiar to philosophical perplexity, as distinct from other kinds. A clever and confident questioner might  ask me questions about, say, music, or grammar, that would lead me into a state with these same features. The questions could be basic. I might be at least slightly  more intelligent and better informed about music and grammar than the average person off the street. And if I were shamed or embarrassed and thrown sufficiently off  balance, I might lose all confidence I know how to use ordinary words to talk about music, or grammar, words that I had confidently wielded only minutes before. What is special about Socratic perplexity is something that is brought out especially well in this Meno passage; it is also something that marks off philosophical  perplexity from other kinds of bafflement and disorientation. Socrates, the expert questioner, is himself perplexed. "Now if the torpedo fish is itself numb, and so  makes others numb," Socrates replies to Meno, "then I resemble it, but not otherwise, for I am not myself free of perplexity (euporôn) when I make others perplexed  (aporein); but I am more perplexed (aporôn) than anyone when I make others perplexed (aporein)" (80c–d). Concerning virtue Socrates is an expert at least at asking questions. He knows the traps and pitfalls well. And he knows how to exploit them. He doubtless knows  better than anyone else in Athens what makes virtue a problematic notion. He has been over this territory many times before. Yet he himself is still perplexed about  what virtue is! When he says, at 71b, that he

  

Page 127

does not know "at all" (tò parápan) what virtue is and adds that he has never met anyone else who knew what virtue is (71c), he must mean that he himself does not  know how to answer the perplexity­inducing questions about virtue he is about to put to Meno, and he does not know anyone else who knows how to answer them  either. When you think of it, that is a very strange feature of philosophy. Compare the situation a few pages later as Socrates asks the slave­boy about how to construct a  square with an area twice that of the given square. Socrates gets the slave­boy perplexed about that question, too, but Socrates never says that he is himself as  perplexed as the slave­boy is. In fact, he is not perplexed—not even a little bit—about the geometrical question. He may have been perplexed once, but, if so, he long  ago surmounted the perplexity. Could Socrates be a good teacher of geometry and still be perplexed about how to construct a square with an area twice that of a given square? I think not.  Remaining perplexed about that question would surely disqualify Socrates as a candidate for teaching geometry. This is not so with philosophy. Socrates can be a  paradigmatically good philosopher and philosophy teacher and still be perplexed about how to answer philosophical questions he puts to others. We can say something even stronger. There is a way in which Socrates would be a less good teacher of philosophy than he is if he were not able to get himself  perplexed over, for example, the question as to what virtue is. Let me try to bring out this point by reference to an imaginary example. Suppose that I am the chair of a philosophy department and I am trying to decide who should be asked to teach a course in the philosophy of art. Suppose I have two  candidates, A and B. Suppose A is extremely well informed in aesthetics and has clear and articulate answers for almost every philosophical question you can ask  about a work of art. But suppose that A is no longer perplexed about questions in aesthetics. He is smug. He thinks he has the answers and, in fact, his answers do  stand up remarkably well under even the most persistent cross­examination. Suppose now that B has no theories at all to promote in aesthetics, but that B has a  special gift for bringing out what is philosophically problematic about art. And suppose that one reason B is so good at inducing perplexity in others about, say,  whether a mounted conch shell, which had been found on a beach, or Marcel Duchamp's famous urinal, 'Fountain', could, or should, count as a work of art, is that B  herself is deeply perplexed about these matters. Whom should I ask to teach the aesthetics class? I think I would choose B. If I did, I could reasonably expect that my  colleagues would at least understand my choice and consider it defensible, even if they did not all agree with it. Imagine by contrast, the chair of a mathematics  department who chose someone to teach an elementary course in geometry who was still perplexed about how to construct a square with an area twice that of a given  square. A chair who chose such a person to teach geometry would be acting quixotically and irresponsibly.

  

Page 128

After Socrates in the Meno has insisted that he is a self­stinging stingray, that is, someone who gets himself just as perplexed as he gets others, he invites Meno (at  80d) to join him in trying again to find out what virtue is. Meno responds to the invitation with the famous Paradox of Inquiry. "How will you look for it, Socrates," he  asks, "when you do not know at all what it is? How will you aim to search for something you do not know at all? If you should meet with it, how will you know that  this is the thing that you did not know?" (80d, Grube trans.) The connection between Socratic perplexity and the Paradox of Inquiry is direct. One who does not know at all what virtue is will not be able to aim an inquiry  specifically at virtue, as distinct from all the things that resemble virtue enough to be easily confused with it. Similarly, one who does not know at all what virtue is will  not be able to recognize virtue, even if he should happen to encounter it. Meno's quite reasonable idea is that an inquiry into X that is not aimed specifically at X and  that cannot recognize success if X should happen to be found, is not really a search for X. I want now to ask what happens to Socratic perplexity in Aristotle. Does Aristotle ever give expression to Socratic perplexity? Or does he at least admit to having  once been Socratically perplexed? One of our best clues as to what happens to Socratic perplexity in Aristotle links perplexity with the Paradox of Inquiry. The passage I have in mind is this one from  the very beginning of Metaphysics B: We must, with a view to the science we are seeking, first recount the things that must be puzzled over (perì ôn aporêsai dei). These include both the other opinions that some  have held on certain points, and any points besides these that happen to have been overlooked. For those who wish to get clear of perplexity (euporêsai) it is advantageous to  state the perplexities well (diaporêsai kalôs); for the subsequent freedom from perplexity (euporía) implies the solution of the previous perplexities (lúsis tôn próteron  aporouménôn) and it is not possible to loose a fetter one is not even aware of. But the perplexity in our thinking (hê tês dianoías aporía) reveals a fetter concerning the thing  (under investigation); for insofar as our thought is in perplexity (hêi gar aporei) it resembles people who are tied up; in both cases it is impossible to go forward. Therefore one  should have surveyed all the perplexities (diaporêsai), both for reasons we have stated and because people who inquire without first stating the perplexities are like those who do  not know where they have to go; besides, one does not otherwise know even whether one has found what one is looking for, for the goal (télos) is not clear to such a person,  whereas to one who has first discussed the perplexities it is clear. (995a24–b2, my trans.)

The methodology Aristotle suggests in this passage is one that he mentions elsewhere. The most famous statement of it is, doubtless, this one from Book VII of the  Nicomachean Ethics:

  

Page 129 We must, as in all other cases, set out the phainomena and, after first running through the perplexities (diaporêsantas), go on to establish, if possible, all the common opinions (tà  éndoxa) about (the matter under investigation) 3, or, failing that, of the greater number and the most authoritative (kuriôtata); for if we both resolve the perplexities and leave the  common opinions undisturbed, we shall have proved the case sufficiently. (1145b2–7)

The passage from the Ethics does not allude to the Paradox of Inquiry, but the methodology outlined seems to be the same as the one suggested in Metaphysics B.  The idea seems to be this. Suppose we are investigating weakness of will (which is, in fact, Aristotle's topic in Book VII of the Nicomachean Ethics). We begin by  lining up the accepted opinions, the endoxa, on weakness of will. We then run through the perplexities concerning this matter. Now we try to say what weakness of  will is in a way that will resolve the perplexities and leave standing the common opinions—if not all them, at least most of them, and, in particular, the most  authoritative. How can this methodology be a response to the Paradox of Inquiry? Well, the main threats that the Paradox poses are these: a) Not knowing for sure what weakness of will (or whatever the matter under investigation might be) is, we will not know how to aim our investigation at the proper  target (instead of letting it become focused on something easily confused with weakness of will, but distinct from it). (Let's call this "the Targeting Objection.") b) Not knowing for sure what weakness of will (or whatever we are discussing) is, we will have no way to be sure that it is really weakness of will that we have  arrived at with our theory, or analysis; we will not be able to rule out the possibility that what we have analyzed is something quite different from, even if also quite  similar to, weakness of will. (Let's call this "the Recognition Objection.") Aristotle's methodology for inquiry responds to the Targeting Objection by requiring that we begin our inquiry by (1) running through the perplexities that afflict the  matter under discussion (weakness of will, or whatever it may be) and that we (2) keep in mind the received opinions, or endoxa, on this subject. The idea is that an  investigation that has resulted in resolving the perplexities concerning weakness of will and also preserved the endoxa concerning weakness of will, or at least most of  them, and the most authoritative, will have been genuinely targeted at weakness of will. The Recognition Objection is responded to in a similar way. The idea is that we need not worry about having, already at the beginning of the search, the object we are  looking for, so as to know when we have found it. We can be reassured that the object analyzed by our analysis, or accounted for by our theory, will indeed be the  right object if the theory or analysis manages to deal with the perplexities concerning that object and preserves most, if not, all common opinions about it, especially the  most authoritative ones.

  

Page 130

Could one not remain Skeptical about the possibility of genuine inquiry, even in the face of Aristotle's reassurances? Could one not worry about whether weakness of  will might be something different from what even the most authoritative opinions allowed for, and different from what would be targeted by resolving even all the  perplexities thought to plague it? I think Aristotle would suppose that such a worry is irrational. He could say that weakness of will just is that state our perplexities are perplexities about, and our  common opinions—at least most of them, and the most authoritative among them—are opinions about. Of course someone could call something else "weakness of  will," or tell us that there is a hitherto unrecognized state that is what is really responsible for behavior we have taken to express weakness of will. But that would not  mean that this other, hitherto unrecognized, state is what weakness of will "really is." It would mean only that weakness of will is not really as important as we thought it  was, or is not important in the way we thought it was. I myself think this is a very interesting response to Plato's Paradox of Inquiry. Although it is not Aristotle's only response to that Paradox, 4  it is, I think, his most  helpful one. And it is part of the story of what happens to Socratic perplexity in Aristotle. It is worth noting that this methodological response to Plato's Paradox of Inquiry involves a shift in the way perplexity (aporía) is thought about. In the Meno passage  we began with, the one I have already indicated I am inclined to treat as a canonical expression of Socratic erplexity, aporía is a state of bafflement and benumbed  confusion. In Aristotle an aporía is not a state of puzzlement at all; it is rather a puzzle or conundrum or difficulty that produces, or is likely to produce, a state of  puzzlement. Following Aristotle's usage, the Paradox of Inquiry should itself count as an aporía. Certainly the earlier puzzle in the Meno about how anyone can desire  bad things (given that bad things are things that are bad for you) would be, for Aristotle, an aporía. And Aristotle's idea that we begin our inquiry by running through  the perplexities (diaporêsai) is the idea of listing problems, difficulties, or puzzles, rather than the idea of reexperiencing states of bafflement. This shift in Aristotle to thinking about aporía as a conundrum that might cause philosophical perplexity, rather than the state of perplexity it might cause, is already  under way in late Plato. We can see it in the transition from Part One of the Theaetetus to Part Two. In Part One Socrates presents himself as a philosophical midwife  who induces the labor of philosophical perplexity. In Part Two there is no more talk of midwifery. Instead we have a series of puzzles and perplexities about false  belief as well as a series of attempts to say how it is we can have false beliefs. So, although it is Aristotle who first makes it natural for us to translate 'aporía' as  'puzzle', or 'difficulty', the shift away from concentrating on the state of philosophical bafflement to identifying sources of

  

Page 131

that bafflement is already well under way in the late dialogues of Plato, especially in the second part of the Theaetetus and in the Sophist. So one thing that happens to Socratic perplexity in Aristotle is that it comes to play a role in Aristotle's methodology—not as a state of baffled confusion and  benumbment, but rather as the problem or puzzle or difficulty that might lead to such a state. Is there in Aristotle's writings any vestige of that characteristic state of  Socratic perplexity? Or is perplexity so normalized in Aristotle, so regimented and brought under control, that there is no remnant of the disorienting feeling of, in  Anscombe's translation of Wittgenstein, "not knowing one's way about?" Here I would like to contrast Aristotle's application of his methodology to two different topics in his Physics; in the first application Aristotle is analyzing the notion of  place (tópos); in the second, a few chapters later, the subject is time. Concerning place, Aristotle lists these five endoxa, or common opinions. Since he thinks he can accommodate them all in his analysis, he does not have to decide  which are "the most authoritative." (1) The place of a thing contains it but is not part of it. (2) A thing is neither larger nor smaller than its primary place. (3) If something moves, it leaves its former place behind. (4) Place always has an up and a down. (5) Each body has a natural resting place. (210b34–11a5)

No doubt all of us today reject the last two items on Aristotle's list. Does that mean that our concept of place is different, or just that he (or we?) have some mistaken  beliefs about place? To answer such a question within Aristotle's methodology, we would need to decide whether all these endoxa are equally authoritative. If (4) and  (5) are less authoritative than the earlier three, then it could be the case that, although he and we have the same concept, either he or we have some mistaken beliefs  about place. In running through the perplexities that concern place, Aristotle comes up with a list of at least a half dozen. Among them are these two: (a1) If a body has a place and a position, then so does its surface. But a surface doesn't really have a place, since there could be no more distinction between a surface and its  place than there is between a point and its place.(209a7–13) (a2) If every thing there is, is in a place, then clearly there will be a place of (that) place, too, and so on ad infinitum. (209a24–25)

The definitional analysis of place that Aristotle eventually comes up with is this: The place (of x) is the first (that is, innermost) motionless limit of that which surrounds  (x) (212a20–21). Having settled on this analysis, Aristotle goes back and shows how each of the perplexities he has listed can be resolved by appeal to it and also  how the analysis conforms to the endoxa he has mentioned.

  

Page 132

Thus, consider the second perplexity above, (a2), which Aristotle identifies as Zeno's perplexity and which threatens an infinite regress of places. According to the  definitional analysis, a place is a certain kind of limit (péras). Is each limit in a place? Aristotle says that each limit will be somewhere (poú) all right, but not in a place.  A limit, he says, is in the thing limited, but that is not a place. We might think to catch Aristotle out here. If a limit is in the body limited, and the body is in a place, why, by the transitivity of 'in', would not the limit also be in a  place? Aristotle has a reply ready. It is only in a different sense of 'in', or way of saying 'in', that a limit is in the thing limited. A limit is in the thing limited, he says, in the  way or sense in which form is in matter (210b22, 210a20). Since the transitivity of 'in' is not guaranteed where there is an equivocation on 'in', we are not forced to  admit, by this argument, that each place is in a place. The resolution of the other perplexity I mentioned (a1) is similar. Surfaces have locations the way points do, but not by being in a place. The discussion of place in Book IV of the Physics is so orderly and so well under control that one is not prepared for the discussion, five chapters later, of time. To  the casual eye this discussion is also orderly and controlled. But a closer look reveals something quite different. Aristotle introduces three perplexities concerning time, which he highlights with the ominous suggestion that they might make one suspect that time exists, either not at  all or only "barely and faintly" (217b32–33). (To be sure, he had also asked whether place exists, at 208a28–27, and in what way; but there had been no similar  suggestion that place might exist only barely or faintly.) In any case, here are the three perplexities concerning time: a) Some of it has been and is not, some of it is to be and is not yet. From these, both infinite time and any arbitrary time are composed. But it would seem to be impossible that what  is composed of things that are not should participate in being. b) Further, it is necessary that, of everything that is resoluble into parts, if it is, either all the parts or some of them should be when it is. But of time, while it is resoluble into parts,  some (parts) have been, some are to be, and none is. The now is not a part, for a part measures (the whole), and the whole must be composed of the parts, but time is not thought  to be composed of nows. c) Again, it is not easy to see whether the now, which appears to be the boundary between past and future, remains always one and the same or is different from time to time.  (217b32–218a10, Hussey trans.)

We would expect that, after giving us this list of perplexities, Aristotle would move on to the common opinions about time, perhaps making a distinction between the  most authoritative opinions and the less authoritative ones. Instead, we get something quite different. Here are the transitional statements: "[t]his

  

Page 133

may serve as a statement of the difficulties (or: perplexities) about the attributes of time. As to what time is or what is its nature, the traditional accounts give us as little  light as the preliminary problems we have worked through" (218a30–33, Hardie and Gaye trans). Instead of thinking he can assemble received opinions concerning time that, together with the listed perplexities, can serve as constraints on what will count as a  satisfactory definitional analysis of time, Aristotle goes immediately into the "traditional accounts," which he considers problematic in much the way the perplexities are. As part of his own account of what time is ("the continuous number of motion with respect to before and after"—220a25–26) Aristotle does seek to resolve the third  aporía above. His resolution is characteristic of him. "The now is in a way the same, and in a way not the same: considered as being at different stages, it is different— that is what it is for it to be a now—but whatever it is that makes it a now is the same"(219b12–15). Whatever we may think of Aristotle's resolution of this third perplexity, it is especially important to note that he makes no attempt whatsoever to resolve the first two  aporíai. Later commentators have been quick to supply resolutions on Aristotle's behalf, but Aristotle himself seems never to have been able to bring himself, in the  Physics or elsewhere, to offer resolutions of those two perplexities. Aristotle's discussion of time leaves other loose ends as well. After having concluded that time is the measure of motion with respect to before and after, he adds that  time will also measure what is at rest (22167–8). Yet he leaves unexplained how something that is, by definition, the measure of motion, will also be the measure of  rest. About perplexity and Aristotle's discussion of time, then, we can say three things. First, Aristotle seems to be trying to apply a methodology to his investigation of time  that makes a central place for the identification and resolution of perplexities concerning the subject under investigation. But, second, the application of his  methodology to this subject is so incomplete and unsatisfactory one is left to surmise that time continued to perplex him. Third, although Aristotle gives ample evidence  of remaining perplexed about time, he never claims for himself anything like Socratic ignorance concerning the nature of time. The disarray of his discussion does  suggest that he should admit to having stung himself in his attempt to say what time is, much the way Socrates stung himself in his attempt to say what virtue is. But  Aristotle never owns up to having stung himself into perplexity on this question. Are there any occasions on which Aristotle admits to having been a self­stinging stingray? Surprisingly, he does seem to do something like that in his discussion of  honey bees, in the Generation of Animals. "There is much perplexity (pollên aporían)," Aristotle confesses at the beginning of his discussion, "about the generation  of bees" (759a8). He then goes on to make

  

Page 134

clear that he simply fails to understand how honey bees reproduce. For one thing, he does not understand what role in reproduction is played by the queen bee (called  by him, and ancient Greeks generally, "the king"—basileús). Again, he cannot identify the gender of either the workers or the drones. He thinks it unreasonable to  suppose that the worker bees are female and the drones male, since, as he puts it, Nature does not give weapons for fighting to any female, and while the drones are stingless all the bees have a sting. Nor (he continues) is the opposite view reasonable, that the  (worker) bees are male and the drones female, for no males are in the habit of working for their offspring, but as it is the (worker) bees do this.  (759b1–6, Platt trans.)

Aristotle is so baffled by bees that he realizes his theory may be distorting the facts his theory is supposed to explain. "The facts," he admits, "have not been sufficiently  grasped; if ever they are, then credit must be given rather to observation than to theories, and to the theories only if what they affirm agrees with the observed  facts" (760a29—34, Platt trans.). This is a rare, although certainly genuine, avowal of Socratic ignorance. Characteristically in his writings Aristotle gives the impression of being totally in command.  Although he thinks it vitally important to identify the perplexities associated with a topic of investigation, he makes clear that he expects to be able to resolve those  perplexities before he quits the topic. Yet surely even Aristotle realizes that he will not be able to deal with all the philosophical perplexities he can identify. What then? The final step in the normalization of perplexity in Aristotle is to eliminate from consideration any perplexity that can be reasonably considered irresolvable.  Memorably, Aristotle does that with the question about how I can know whether I am now dreaming. Here is the way Plato has Socrates raise that question in the  Theaetetus: Socrates. There's a question you must often have heard people ask—the question what evidence we could offer if we were asked whether in the present instance, at this moment,  we are asleep and dreaming all our thoughts, or awake and talking to each other in real life. Theaetetus. Yes, Socrates, it is certainly difficult (aporón) to find the proof we want here. The two states seem to correspond in all their characteristics. There is nothing to prevent  us from thinking when we are asleep that we are having the very same discussion that we have just had. And when we dream that we are telling the story of a dream, there is an  extraordinary likeness between the two experiences. (158bc, Levett trans.)

In the Theaetetus Plato lets Socrates drop this puzzle without resolution, and also without reflection on whether one's inability to resolve it undermines other claims to  know or perceive something. After all, in a style of argument quite familiar to us today, one could argue that, unless Socrates can rule out the

  

Page 135

possibility that he is now dreaming that he is talking to Theaetetus, he cannot be said to know, on the basis of his senses, that he is doing so. 5  One can only conclude  that Plato did not pursue this perplexity, or its implications, because he simply did not know how to resolve it or what else to do with it, given that he could not resolve  it. It is not unreasonable to suppose that this dream question was much discussed in Plato's Academy. In any case, Aristotle refers to it in the way one refers to familiar  philosophical chestnuts. Then he dismisses it! Here is the passage: Some ... are perplexed (aporoûsi) because they want to know who will judge who is healthy, and in general on each subject (who will judge) who will judge it correctly. Such  perplexities (aporêmata) are similar to the perplexing question (tô aporein) 'Are we now asleep or awake?' and they all have the same force. For those who pose them ask for an  argument for everything; for they seek a principle, and they seek to get it through demonstration . . . . Their trouble is just as we have stated: for they seek an argument for  something for which there is no argument, for a principle of demonstration is not a demonstration. (Metaphysics IV 1011a3–13, Kirwan trans.)

Aristotle seems to see in this dream question the threat of a regress. But what sort of regress? The idea must be this. Suppose I think I have a test for determining  whether I am now dreaming. It might be the simple one of pinching myself, or it might be something more scientifically up­to­date, like tracking my brain waves.  Suppose now that I apply that test, whatever it is. (Call the test "T.") I apply T to my present experience to see whether I am now really engaged in writing a paper, or  only dreaming that I am doing so. Suppose that, according to the test result, I am awake, and, therefore, really writing this paper. Does that settle matters? No, not at all, for I can ask the very same question about my experience of applying T. That is, I can ask whether I really applied T, or only dreamed that I did. In  response to this question I could, perhaps, apply T to my first application of T, that is, to the experience of testing myself to see whether I was dreaming. But the same  issue could arise again. For I could ask about that "meta­application" of T, whether it was real, or whether I only dreamed that I performed it. And now we are off on  an infinite regress. Aristotle puts the threat of regress that he sees in an attempt to resolve the dream question in the same bag as a question about a judge of health. Presumably the idea  is that we can ask concerning our health experts—those health gurus who tell us who is healthy and who is not—whether their judgment is to be accepted. We can call  in outside experts to determine whether our "in­house" experts are reliable judges of health. But the same question of reliability can then be asked about the outside  experts. Aristotle says that the trouble in all these cases is that we are trying to do something like demonstrating, or proving, a principle of demonstration. His idea

  

Page 136

seems to be the entirely reasonable one that every valid demonstration applies a principle of demonstration that is used in that demonstration, but not itself proved by  it. We might try to establish the principle of demonstration by a fresh argument. But that, in turn, would apply a principle of demonstration that would not be  established by the argument in which it is used. The general moral is 'You cannot establish anything without first assuming something'. Although the details are in contention, Descartes seems to have thought one could establish 'I exist' with some sort of reasoning, or quasi reasoning, that does not  require us to assume anything else first. Notoriously, not everyone has been convinced by Descartes. In any case, what Aristotle says in this passage suggests that he  would reject Descartes's foundational project as misguided, not just in detail, but in principle. That completes my picture of the normalization of Socratic perplexity in Aristotle. As we have seen, Aristotle seems to respond to the Paradox of Inquiry by  suggesting a methodology that includes identifying the perplexities associated with the matter to be investigated or theorized about. He seems to consider resolving  those perplexities a necessary condition for achieving a successful investigation. But sometimes, as in his discussion of time in the Physics, he himself is unable to  resolve all the perplexities he has identified. Even when that happens, however, Aristotle is not prone to claim for himself Socratic ignorance. It is only in the rare case  (I mentioned the discussion of honeybees in the Generation of Animals) that Aristotle admits defeat. Finally, where he thinks a philosophical perplexity is irresolvable  in principle, he simply excludes it from consideration on that very account. 6 Notes 1. A close runner­up might be this passage from Augustine: ''What then is time? If nobody asks me, I know. If I should want to explain it to someone who asks, I  don't know" (Confessions 11.14). 2. Ludwig Wittgenstein, Philosophical Investigations, G. E. M. Anscombe, tr., (Oxford: Blackwell 1953), 123. 3. In this passage, weakness of will (akrasía), effeminacy (malakía), self­control (egkráteia), and endurance (kartería). 4. In A.1 of the Posterior Analytics Aristotle speaks of the Paradox as an aporêma and suggests that it might be resolved by distinguishing two ways of  understanding 'know'. 5. See Barry Stroud, The Significance of Philosophical Scepticism (Oxford: Clarendon Press 1984), 20. 6. I owe thanks to Nicholas White for suggesting some changes in an earlier version of this paper.

  

Page 137

Chapter Seven—  Dialectic, Contradiction, and Paraconsistency in Aristotle J. D. G. Evans The original context of this paper was a Round Table at the 1993 World Congress of Philosophy in Moscow. The session was organized by Jindrich Zeleny (Prague); and he  brought together scholars from diverse philosophical cultures, all with expert interest in Aristotle and dialectic. 1 The occasion was ripe for a synergy which the International  Federation of Philosophical Societies (FISP) is best placed to effect, and I dedicate this paper to my friends and colleagues in this organization.

I Paraconsistent logic recognizes the occurrence of true contradictions. For example a 'Liar' statement—such as "this statement is false"—is, if true, false and also, if  false, true; it is therefore both true and not true, and that is itself a truth. Similarly, if we consider a train at the very moment when it starts from rest and begins its  journey, it is plausible to maintain that then it both is and is not at rest. Can these insights be handled in a rational way? The dialetheist maintains that they can, with the  help of paraconsistent logic. He may also give philosophical welcome to the enterprise, if he believes that some contradictions are not only true but also true as a  matter of fundamental epistemological fact.2 What has Aristotelian dialectic to do with paraconsistent logic? The quick

  

Page 138

and tempting answer is: surely very little. Aristotle was deeply committed to the avoidance of contradiction. This commitment is reflected in the realism of his  metaphysics and epistemology, and it also underpins his method in logic and dialectic. The hallmark of someone who espouses paraconsistent logic is a claim to  countenance true contradictions. How could Aristotle have any patience with such a view? Yet those who recognize the pervasive importance of dialectic in Aristotle's conception of philosophy and its method should linger awhile before they dismiss all  interest in paraconsistency. Paraconsistent logicians trace their intellectual pedigree to the dialectical theories of Hegel and Marx; that is the source of their conviction  that the phenomenon of contradiction is philosophically significant. These philosophers' theories of dialectic were in turn developed out of Kant's account of dialectic as  a method of reason; and Kant acknowledged the provenance in the ancient Greek philosophers for his statement of the nature and value of dialectic. 3 There is, then, a chain of historical and philosophical connection running from Aristotle through the German philosophers to the contemporary dialetheists with their  paraconsistent logic. There is a second reason for considering the possible collocation of Aristotle and the modern dialecticians. This reason is more directly  philosophical. Along with his commitment to the avoidance of contradiction, there runs through all of Aristotle's work a strong vein of realism. Contradictions are false  because the world is thus and therefore not also otherwise. This suggests that for Aristotle truth is an objective matter, something available to be discovered (or not, as  the case may be). How can this realism and objectivism be reconciled with his evident respect for dialectical method?4  Aristotle makes strong claims about the significance of the  inheritance of views—endoxa—which he regards as setting the agenda for the discussion of a problem. But there might be a massive and thorough failure of match  between the consensus of views and the objective truth. In that case concern for endoxa would be at best an irrelevance and, more likely, a positive hindrance to  effective enquiry. Is there confusion and inconsistency in Aristotle's valuation of dialectic, or can we reconcile the various elements in his position? These are the philosophical issues  which we shall consider as we reflect on Aristotle's discussion of contradiction and ponder how he might fare among the dialetheists. II As is well known, Aristotle established and clarified certain fundamental logical principles concerning truth and falsehood. He also placed a high value on dialectic as a  method of rational investigation.5  There is a potential tension

  

Page 139

between these phenomena that needs to be exhibited and, if possible, defused. In this section I shall consider the logical principles and, in the next, the apparently  countervailing claims of dialectic. The logical principles include the Principle of Non­Contradiction (PNC) as well as its close relations the Law of the Excluded Middle and the Principle of Bivalence.  Aristotle argues for these principles at length in Metaphysics  . In many other passages he makes use of PNC and its congeners; the principle is not only embraced  unequivocally, it is also treated as fundamental. There is no reason to regard the extended argument of Met.   as anything other than a thoroughly Aristotelian piece of  philosophy. The deep significance of PNC and the other principles is acknowledged by all philosophers of logic; but these principles are espoused particularly by the  advocates of so­called classical logic in the Frege­Russell­Quine style. The classic in the title is Aristotle: his legacy is a recognition of the need for a sharp and  fundamental distinction between truth and falsity in the organization of our insights concerning thought and language. As Katalin Havas has argued, 6  PNC in Aristotle is simultaneously a principle of (i) ontology, (ii) logic, and (iii) philosophical method, thus: (i) in ontology, ( F,  x) (­ [Fx & ­Fx]); (ii) in logic, for any statement that is true or is false, its negation is respectively false or true; (iii) in philosophy, acceptance of one arm of a contradiction rationally commits that person to rejection of the other arm. It is this last aspect that I want to concentrate on. Throughout his work Aristotle observes the methodological position which is expressed in (iii). In Met.  .3–6 he  provides a battery of arguments for the fundamental character of the principle; and one of these runs that if the principle is true, it cannot be doubted ( .3 1005b11– 34). The argument claims that doubt about PNC could only be based on a belief, for some P, that P and not­P. This belief is equivalent to the pair of beliefs, for that  particular P, both that P and that not­P. However someone who held such a pair of beliefs would herself constitute a self­contradictory subject; for the opposing  beliefs would be contradictory attributes of one and the same subject (the believer). This audacious argument has been challenged for taking too cavalier a way with quantification and the scope of negation where epistemic and psychological predicates  are involved.7  I believe that the argument can be defended against this line of objection, but will not do so here. What I do emphasize is how central this argument is to  Aristotle's defense of PNC. The principle is asserted to be at the core of any rational person's belief set. It is right to emphasize, in addition to its other functions in  philosophy, this pragmatic role of the principle in the regulation of our intellectual proceedings. In this defense of PNC Aristotle shows himself to be committed not only to the truth of the principle but also (in terms of the contemporary debate

  

Page 140

concerning realism and anti­realism) to a realist construal of what that truth consists in. The content of a belief—the particular p that A believes—is one of the most  contested of all the kinds of item for inclusion among the furniture of reality. While anti­realists are ready to allow that it is true that A believes p, they will not admit the  content of this belief to the status of a real thing. By contrast Aristotle, as this argument in Met.   shows, counts as real this item which supports the status of PNC. His  commitment to the principle is thus based on ontological realism; and this fact serves to heighten the potential tension with his dialectic, which is the overall topic of my  chapter. Aristotle's commitment to PNC, as a fundamental and universally recognized premise in reasoning, seems to be firm. The support for this position is not, of course,  confined to Met.  . Further evidence comes from many other areas of his work; and I single out three elements in his philosophy of logic, having to do with his theories  of proof and refutation. First, in Prior Analytics one leading method for establishing the validity of syllogistic forms is the indirect one of positing the contradictory of the conclusion. If this  entails the contradictory of one of the premises, the validity of the syllogistic form is established. The inference is a version of modus tollens; from the falsehood of the  consequent in a hypothetical syllogism, the falsehood of one of the antecedents is deduced. Clearly the efficacy of the method depends on an unqualified commitment  to PNC and the other logical principles. 8 Second, in De Interpretatione the discussion proceeds as a systematic attempt at the correct location of contradiction, in quantified, modal, and temporal statements.  The exercise is necessary and insightful because the word order of colloquial ancient Greek left doubt as to the scope of the negation operator in a statement. If we  say, of some group of subjects, that they "are not white" or "have the capacity not to walk" or "are going not to reach maturity,'' the scope of the "not" leaves it unclear  whether all or only some of the subjects (or their modal and temporal counterparts) must satisfy the condition in order for the statement to be true.9  Again the whole  direction of the argument is regulated by a conviction that PNC applies to statements rigorously and universally. Third, in the analysis of false refutations in De Sophisticis Elenchis, Aristotle marks off a whole subdivision of arguments as faulty just for the reason that they fail to  respect PNC. These arguments fail to observe the niceties of PNC when cast in logically precise form, including such qualifications as time, respect, relation, etc.  which need to be observed if apparent contradiction is not to oust the genuine variety. The relevant qualifications are discovered and justified by the need to prevent  the assimilation of genuine contradiction to pseudo contradictions where both members of the pair may be true. Disputants who commit this kind of fault "do not know what refutation is." Throughout the discussion in SE Aristotle is considerably exercised to examine

  

Page 141

the integrity of his classification of false refutations; 10 and at one point he systematically explores the possibility that all false refutations might be brought under the  heading of ignorance of refutation (SE 6). Each of the headings of false refutation that have been distinguished in the previous chapters consists of a fault that can be  linked to some specific part of the definition of genuine reasoning. If we combine this suggestion with his earlier association of this fallacy and failure to understand and  observe PNC, we get the result that such failure provides something of a paradigm case for the analysis of unsatisfactory reasoning. Conversely PNC itself is the  guarantor of sound reasoning. The evidence supports the claim that Aristotle is a classical logician. In itself, that conclusion should not be surprising, given that classical logic effectively defines itself  through its derivation from Aristotle. But Aristotle gave a few hostages to fortune, which have been seized upon by those who wish to enlist his authority on behalf of  some nonclassical logic.11 The most notable examples are his discussion of statements of future contingency in De Int. 9, of predicate­negation where the subject fails  of reference in Cat. 10 13b12–35, and some complications in his account of negation in the square of opposition in De Int. 10 19b26–20a3 and An. Pr. A.46  51b22–35. But none of these alleged instances of compromise or backtracking is compelling. In the discussion of future contingents in De Int. 9 it is admittedly obscure what  exactly Aristotle wants to conclude about the truth­status of a particular statement about a future contingency; but he does not give up the view that a pair of such  contradictorily related statements are subject to PNC (19a27–30). Similarly the temporary dislocation that arises in Cat. 10 when Aristotle considers proper names  that fail of reference, causes a problem for the determination of which truth­value particular statements have; but the issue is resolved entirely within the framework of  PNC. Finally the explorations of the effects of negation in the square of opposition, so far from curtailing the application of PNC, serve instead to extend it into the  unexpected territory of negative subjects ("non­man" etc.). Therefore I am not persuaded that these particular problems are at all adequate to serve as a challenge to Aristotle's well­established position as a defender of PNC  and the other principles and as the quintessential classical logician. But as I indicated earlier, there are other internal difficulties which I take to be potentially more  serious. They arise in two main areas, one major and one minor. The major one is the theory and practice of dialectic, and the minor one is a brief comment which  Aristotle makes on the Liar Paradox. I shall consider these in each of the next two sections.

  

Page 142

III Aristotle's commitment to the importance of dialectic as a method of rational enquiry is widely recognized. But there is far less agreement as to the precise significance  of this. His conception of dialectic certainly involves a recognition of the necessity for productive philosophy to argue more than one side of a case. From there one  can progress—perhaps by a short step, perhaps by a longer route—to the practice of regarding the truth as consisting in the balance between two (or more) opposed  positions. The short step is verificationist: if the truth must be reached this way, then that must be how truth is constituted. The longer step depends on demonstration  that this is how truth is achieved. I shall now take the longer route. This attitude to philosophical truth is reflected in Aristotle's actual conduct of investigation, where an essential precondition for achieving sound insight is the  juxtaposition of antithetical endoxa. Those, like myself, who argue that there is a substantive connection—and not a merely homonymous one—between Aristotelian  dialectic and dialectic in Kant and Hegel, find in this aspect of Aristotle's method a major support for their claim. 12 But this gives rise to several questions. Does his  theory of dialectic compromise his rejection of the possibility of true contradictions—indeed, of contradiction as a fruitful methodological ingredient? Does his  commitment to PNC mean that any rapprochement between him and a Hegelian over dialectic is misconceived? Perhaps his own overall position is confused and  inconsistent? I shall answer no to all these questions. As regards dialectic, it is unquestionable that this method plays a fundamental role, both in theory and in practice, in Aristotle's philosophy. In Topics he emphasizes  the value of dialectic as an essential tool for the discovery of scientific first principles; and lest it might be supposed that that work presents a one­sided and eccentric  view of its own importance, we should bear in mind that in works other than Topics Aristotle consistently distances himself from the Platonic idea of dialectic as a  universal super­science.13 For the philosophical significance of dialectical method we can cite many passages in which Aristotle reflects on and justifies his habit of starting a discussion from a  juxtaposition of conflicting views. Prime examples of such passages are the introduction to the discussion of friendship in Eudemian Ethics H.2 (1245b13–18) and  the more extensive preface to the debate of difficulties in Metaphysics B.1 (995b4–6a17, esp. 995a27–33). Such passages supply essential insight into Aristotle's  conception of fruitful method in philosophy.14 In Met. B.1 (995a27–b2) he says: For those who wish to be free from difficulty, it is helpful to go through the difficulties well. To be free from difficulty later, is to untie what were difficulties earlier; but those who  are ignorant of the knot cannot untie it. The difficulty in

  

Page 143 our thought reveals the knot in the facts; insofar as our thought is in difficulty, its condition is like those who are tied up. So it is necessary first to have considered all the  perplexities, both for the reasons given and because those who search without having first gone through the difficulties are like those who are ignorant of where they should be  going—moreover, an ignorant inquirer would not know whether or not he has found what is sought. For such a person the goal is unclear; but for someone who has been in  difficulty, it is clear. (995a27–b2)

Two main ideas are connected in this remarkable passage. First there is the metaphor of intellectual bondage. Difficulties both constrain us and, when we have  diagnosed them, hold out the means for escaping from them. Secondly, the epistemological value of dialectic is indicated by the clear reference to Meno's puzzle in the  latter part of the passage. Aristotle took seriously the problem that enquiry might fail for lack of an adequate focus on its object; and unlike Plato who raised but did  not answer this conundrum, he used the theory of dialectic to address it. 15 Every investigation must be introduced by a rehearsal of the available plausible views (endoxa). The differences and disagreements between these views set the  agenda for Aristotle's positive discussion, which is to be understood as a resolution of the earlier difficulties. The positive account cannot stand independently of the  antecedent debate, since only to someone who has been gripped by the dilemmas in that debate can the resolution stand as significant. It follows that dialectical  exploration of conflicting views is no mere ornamental preface to real philosophizing, but instead is an ineliminable aspect to rational enquiry and intellectual advance. Here are some examples of the method. There is an apparently irreconcilable clash of views between philosophers who seek to explain things in terms of material  constituents and those who cite purposes and end results—between supporters of the material cause and of the final cause. Similar oppositions obtain between those  who advance a dualistic view of human personal identity, according to which each of us is both a body and a soul, and those who deny the separateness of the soul  and support a physicalist monism. A third case is the stand off between those who maintain that there is a common and universal goal in life, even though particular  people's mistakes about it can introduce diversity into its pursuit, and their opponents who understand the subjective differences in aim as fundamental and not to be  overridden by any conception of a common goal.16 In all these and many more discussions of particular philosophical problems, Aristotle demonstrates the fundamental importance of dialogue, discussion, and dialectic  for philosophy. Is he a dialectical philosopher in a more modern—Kantian, Hegelian, or Marxian—sense? To this I reply "no and yes." On the one hand, Aristotle's discussions of these problems move uncompromisingly beyond the difficulties which are constituted by the

  

Page 144

assemblage of inherited and conflicting views. Aristotle says that the views are all thereby shown to be wrong; and he attempts to provide an account of his own that  avoids their defects. He seeks to show that the initial sense of conflict is an illusion that can be corrected, and that without abandoning any of the insights of the warring  parties, we can reach a total and consistent view of the matter—a view that absolutely and without complication reflects the truth. His own account is true, and its  negation is false. On the other hand, this true account arises out of its rejected predecessors. It is epistemically related to them, and it must continue to be read and understood in this  way. Aristotle has a particular device for reconciling what, to their proponents, tends to seem irreconcilable in their theses. This device is the distinction between what  is so simply and without qualification and what is so in a qualified way. Thus while the goal which we aim for is simply the good life, the goal of some particular person  may be what appears good to him; and if appearances mislead him, these two specifications of the goal will not coincide. Similarly the unqualified explanation of  change is in terms of purposes and ends; but where the a thing's matter does not simply serve that end, a different, qualified explanation will be needed as a supplement  to the (unqualified) explanation. 17 In argument we need to be alert to the unqualified/qualified distinction. We learn from SE that we should be cautious about trying to accommodate this distinction to  the detailed provisions of PNC.18 The contradictory of being simply F is being simply not­F; and that is so even though it may be that something is simply F while  being not­F in a qualified way. I take this to mean that 'simply' is effectively a Special kind of operator—one that it would be wrong to assimilate to, for example,  'always', 'overall', or even 'generally'. There is a connection and a contrast between these two kinds of operator. They are connected since in both cases there is an opposition with what is so sometimes, in  a particular place etc. But there is also a contrast, since what is sometimes not­F defeats a claim that things are always F but not that they are simply F. This point can  be illustrated by an example which, I have argued elsewhere, well illustrates Aristotle's use of the 'simply' operator. This is the case of Taste—a concept which comes  both in an unqualified form and with qualification. There is taste simply, as when we say that someone or something shows taste; and there is the qualified taste that is  someone's taste or the taste of some group. The distinction gains philosophical point when it enables us to say that bad taste is not taste. For bad taste is still  someone's taste. Now consider how contradiction operates with respect to these two modes of taste. A particular claim about taste (simply) is contradicted by one that is similarly  about taste. Thus if it is tasteful to prefer Bach to the Beatles, the opposite preference shows lack of taste. At the same time if a particular person's or group's taste is  for either of these preferences—as it can be—then the opposite preference is not to that particular taste. So we have contradiction, both

  

Page 145

at the level of unqualified taste and in the case of any particularized taste. But there is no automatic way to relate the two contradictory pairs. A particular, qualified  taste may coincide with unqualified taste, or it may not. Aristotle's position on this matter would be that such coincidence occurs only in the case of the aesthetic  expert's taste; but other positions are possible. Thus the situation is quite unlike those with quantified and modal statements. Here there is a single scheme of contradictory opposition which links statements governed  by universal and particular quantification, and also those which are covered by the modal operators necessary and possible. Preservation of the rule of PNC is a  clear­cut all­or­nothing matter here. But it is otherwise with statements containing the operator unqualified and those with qualifications. Here we can have apparent  conflicts and tensions among different statements which yet fall short of contradiction. A good example is supplied for the case of Taste by Kant's famous antinomy:  "everyone has his own taste" and "there may be a quarrel about taste." 19. Antinomy it may be; but it is not a contradiction, and according to Kant (and Aristotle) it is  resoluble. So I maintain that this particular logical device—namely, the distinction between qualified and unqualified predication—enables Aristotle to relate the elements in his  dialectical conflicts in a way that is otherwise than as a contradiction. This particular logical device is formally distinguished from contradiction in his theory of argument.  Consequently his total acceptance of PNC—and more generally his respect for argument as an instrument of rationality—can be reconciled with his theoretical  commitment to and his unrestricted use of dialectical method in philosophical enquiry. We can also supply an answer to one of our initial questions—namely how an apparent commitment to realism is well served by dialectical method. The conflict  seemed to present itself as follows: while realism demands that we pursue truth, as this is reflected in such principles as PNC, dialectic deals in opinion which may not  contain truth. Juggling such opinions in such a way as to eliminate the appearance of conflict may have no tendency to generate a true rather than a merely consistent  account of the matter. That is the objection of the realist; but in fact this realism is unacceptably extreme. There is a more moderate form of realism which is found in Aristotle and is founded  on the very theses about dialectic which we have been examining. Truth must be such that it is intelligible to expert human understanding; and what distinguishes the  expert is precisely—as in the case of unqualified taste—the fact that his particular understanding is attuned to this truth. Given this moderate realism, the role of  dialectic becomes one of facilitating progress from the mere exercise of the faculty of understanding, in an inexpert way, to the achievement of truth­delivering  expertise. The details of the process are expounded in Topics, and its operation is illustrated throughout Aristotle's actual investigations of problems.

  

Page 146

IV I turn now to the Liar Paradox. Aristotle's brief comments on this, in SE 25 18062–7 and Rhet. B.24 1402a3–28, have won no supporters. Aristotle is regarded as a  classical logician who lacks any instinct for the significance of this conundrum. By contrast, certain modern paraconsistent logicians—most notably, Graham Priest and  the dialetheists—regard it as one of the great merits of paraconsistent logic that it can accommodate Liar­type Statements. 20 If I say "what I am saying is false", what  I say is true precisely because it is false; while this is intolerable for the classicist, the dialetheist welcomes it as strong confirmation of his general account of the  possibility of true contradiction. It is significant that, in his comments on the Liar Paradox, Aristotle uses the very tools that we have already encountered in his dialectical analyses and which we have  also exploited in order to try to reconcile his commitments to dialectic and to classical logic. Aristotle's brief solution of the Liar in SE 25 is as follows. The elements of  both truth and falsehood which appear to reside in the Liar's statement, can be reconciled if we draw a distinction between what is so simply and in an unqualified way  and what is so with qualification. Armed with this analytic tool, we can allow for the possibility that a person (a liar) should proclaim his own falsity—either as a general  proposition ("all my utterances are lies") or with reference to some particular case ("this utterance is a lie"). We can avoid the paradoxical results of self­reference by  saying that in general (without qualification) the man utters lies, even though some particular utterance of his (including one which speaks about his own lies) is not a lie. The same analytical idea is at work in Rhet. B.24. Various things may be claimed to be probable or to be improbable. In appropriate circumstances, as Sherlock  Holmes was aware, the improbable is probable. Therefore we have the possibility of arguing that the same statement is both probable and improbable. It is probable  that people lock their houses when they leave them, and it is improbable that they do not. Yet the person who wants to make a false insurance claim will exploit these  very facts of probability/improbability. It is probable that in these circumstances the improbable will happen. The apparent contradiction in such a claim is disarmed  when we add the necessary qualification to one part of it. There is an interesting further twist. The distinction between unqualified and qualified statements can be viewed as a distinction between unqualified and qualified truth.  Rhetoric deals in probability rather than truth; therefore in this area we have to handle unqualified and qualified probability. This last distinction reflects well enough  Aristotle's point in Rhet. B.24 about fallacious enthymemes; but consider the implication for the other end of the comparison—nonrhetorical syllogisms. We are being  invited to view truth itself as subject to qualification. Aristotle does not develop this idea here, and it does not appear elsewhere. But it is consistent with the thought,  which we explored in the

  

Page 147

previous section, that PNC does not generate a single square of opposition to cover statements which make use of the qualified/unqualified operator. But his main line of solution to the paradoxes involves applying the distinction to truthfulness rather than directly to truth; and I maintain that is effective and illuminating.  It will work, provided that we know the general disposition—to lying or truth telling—of the speaker. It will not work for those modern versions of the paradox which  restrict the self­reference to a single utterance. I agree with the classical logician, against the dialetheist, that these latter versions are insoluble. But the dialetheist is  wrong to suppose that he thereby wins the contest. On the contrary, as Aristotle argues in his ethical theory, characters and dispositions are more fundamental than  particular acts; 21 and so if Aristotle can solve the Liar for the lying disposition, he has what matters for the defense of classical logic and PNC. I conclude that the modern dialetheist charge that Aristotle ignores the paradoxes of truth is unfounded. In that case we have removed a main reason for seeing his  commitment to PNC as a stumbling block to correct logical insight; and my earlier reconciliation of classical logic and dialectical rationality provides the essential  materials for the characterization of Aristotle as a dialectical logician. V Our discussion started from a cross­historical or cross­cultural distinction between two modes of understanding logic and philosophy. We also noticed that the  problem which was indicated by this approach, also appears—or should appear—within any particular reading of Aristotle's assessment of dialectic and contradiction.  In the course of the argument I have sought to resolve both of these tensions; and if the argument works it can perhaps be judged doubly successful. For it is itself an  illustration of the thesis which it advances. The apparent conflict between classical logic and dialectical reasoning needs to be noticed, explored, and teased out; and it  can be resolved. Philosophy is dialectical. Philosophers revel in the sinuosities of conflict and contradiction; their activities most certainly do not progress in the linear way which analytic  scientific method prescribes. But respect for conflict and argument is quite compatible with commitment to a particular position and with rejection of those that are  opposed to it. These are the messages that I take from my reading of Aristotle, and I think that they are true.

  

Page 148

Notes 1. J. Zeleny, "Paraconsistency and Dialectical Consistency" (Draft) (Prague, 1993). This was subsequently published as "Paraconsistency and Dialectical  Consistency," From the Logical Point of View 3 (1994): 35–50. Other notable contributors included Katalin Havas (Budapest), Francisco Miro Quesada (Lima),  Enrico Berti (Padova), Ryszard Wójcicki (Warsaw), and Christopher McKnight (Belfast). I have also derived encouragement from the overlap in theme between the  topics in my essay and those mentioned by many of the other contributors. In addition to the papers by Eugene Garver, Gareth Matthews, and May Sim, referred to in  n. 14 and n. 19 below, those by Robert Bolton and Michael Ferejohn particularly touch on topics which have influenced my own approach. 2. See G. Priest, "Classical Logic aufgehoben," in Paraconsistent Logic: Essays on the Inconsistent, G. Priest, R. Routley, and J. Norman, eds. (München:  Philosophia Verlag, 1989), 131–48; also, G. Priest and R. Routley, "The Philosophical Significance and Inevitability of Paraconsistency," in Paraconsistent Logic,  483–539. 3. I. Kant, Critique of Pure Reason A60–2/1384–6. See J. D. G. Evans, "Ancient and Modern Dialectic," Philosophical Studies (Ireland) 33 (1991/2): 39–53. 4. This point has been insistently raised by T. H. Irwin, Aristotle's First Principles (Oxford: Clarendon Press, 1988). See also M. Nussbaum, The Fragility of  Goodness (Cambridge: Cambridge University Press, 1986), 240–63. I have reviewed Irwin's book in Hermathena 147 (1989): 73–4. 5. G. E. L. Owen, "Dialectic and Eristic in the Treatment of the Forms," in Aristotle on Dialectic, G. E. L. Owen, ed. (Oxford: Clarendon Press, 1968), 103–25. I  have argued the case in J. D. G. Evans, Aristotle's Concept of Dialectic (Cambridge: Cambridge University Press, 1977), esp. 49–52; see also J. D. G. Evans,  Aristotle, (Sussex: Harvester, 1987), 2–7. 6. K. Havas, "Differences in the Unity," Logique et Analyse 29 (1986): 149–60. 7. See C. Kirwan, Aristotle's Metaphysics G, D, E(Oxford: Clarendon Press, 1971), 89–90; J. Barnes, "The Law of Contradiction," Philosophical Quarterly 19  (1969): 302–9. 8. An. Pr. B.11, esp. 62a11–19. 9. See L. R. Horn, A Natural History of Negation (Chicago: Chicago University Press, 1989), 6–21. 10. J. D. G. Evans, "The Codification of False Refutations in Aristotle's De Sophisticis Elenchis," Proceedings of the Cambridge Philological Society 21 (1975):  42–52. 11. See G. H. von Wright, "Truth, Negation, and Contradiction," Synthese 66 (1986): 3–14; K. Havas, "Differences in the Unity," n. 6. 12. J. D. G. Evans, "Ancient and Modern Dialectic," n. 3. H. Williams, Hegel, Heraclitus and Marx's Dialectic (Hemel Hempstead: Harvester Wheatsheaf, 1989),  esp. chs. 3–4. 13. SE 9 and 11; An. Post. A.11 77a26–34; Met. A. 9 992b18–24; see G. E. L. Owen, "Logic and Metaphysics in Some Earlier Works of Aristotle," in Aristotle  and Plato in the Mid­fourth Century, I. Düring and G. E. L. Owen, eds. (Göteborg: Almqvist & Wicksell, 1960), 163–90, and J. D. G. Evans, Aristotle's  Concept, n. 5, 41–9.

  

Page 149

14. See J. D. G. Evans, "Aristotle on Learning," Annales 24 (1992): 33–48. The significance of this passage for Aristotle's conception of philosophy is also discussed  by Gareth Matthews (128) and May Sim (190 and 193). 15. The Meno paradox is discussed by G. Fine, "Inquiry in the Meno," in The Cambridge Companion to Plato, R. Kraut, ed. (Cambridge: Cambridge University  Press, 1992), 200–26; J. D. G. Evans, "Meno's Puzzle," in The Concept of Knowledge, R. S. Cohen & I. Kuçuradi, eds. (Dordrecht: Kluwer, 1995), 67–80. 16. The main discussions in Aristotle are: PA A.1 639b12–41a18; De An. B.1 412a1–22 (with A.4 407627–34); EN  .4. 17. EN  .4 1113a23–6; Phys. B.7 197b18–8a13. 18. SE 6 168b11–16; Top. B.11 115b11–35. 19. Critique of Judgement (Section 56). I have developed this example in J. D. G. Evans, "Souls, Attunements, and Variation in Degree: Phaedo 93–4,"  International Philosophical Quarterly 34 (1994): 277–87. I note with satisfaction Eugene Garver's view (118–119) that this distinction is of the first importance for  Aristotle's ethical theory. 20. G. Priest, "Can Contradictions Be True?" Aristotelian Society Supplementary Volume 67 (1993): 35–54. 21. EN B.4 1105b5–7.

  

Page 151

Chapter Eight—  Perception and Dialectic in Aristotle's De Anima Michael Ferejohn I This chapter is as much about method as it is about doctrine. Largely as a result of G. E. L. Owen's landmark article, ''Tithenia Ta Phainomena," 1  most students of  Aristotle are by now familiar with a certain "dialectical"2  pattern of investigation which runs through his work on a wide variety of subjects. There are numerous  variations upon this but the basic pattern, as I shall be understanding it, typically consists of four recognizable phases. The first of these is an (i) Aporetic Survey in  which a number of initially more or less plausible (and usually mutually opposing) views on some loosely formulated issue are subjected to critical scrutiny, and each of  these endoxa is found to involve conceptual difficulties (aporiai). It is not always easy to see just where this phase leaves off, but at some point Aristotle transforms  the discussion by translating the question he is investigating into his own semi­technical jargon, and thereby (ii) Superimposing onto the investigation his distinctive set  of analytical (and usually metaphysical) concepts and distinctions (e.g., form and matter, actuality and potentiality, per se and per accidens, to name but a few). But  since virtually every key Aristotelian philosophical term is viewed and used by him as systematically (and instructively) equivocal, this Superimposition provides the  matter for the next phase, a systematic (iii) Disambiguation of the

  

Page 152

disputed question wherein one construal is normally designated as "primary," or "strict." With this in hand, Aristotle then finally proceeds to (iv) Resolution of the issue  by arguing for his preferred answer to the ''strict" Disambiguation of the question. Often this also involves construing each of the endoxa refuted during the Survey as  an unsuccessful attempt to express all or part of the truth contained in Aristotle's own position which goes wrong because of some failure to respect the distinctions  introduced into the discussion by Aristotle himself during Superimposition. Thus, though all positions except Aristotle's are literally rejected during this procedure, there  is nonetheless a sense in which all are accommodated, and it is in this sense that the method is said to "save the phenomena." As Owen correctly emphasizes, Aristotle does not confine his use of this dialectical method to areas of study that we might now want to categorize as philosophical',  or 'metaphysical'. Quite to the contrary, he makes frequent use of it in treatises such as the Physics and De Generatione et Corruptione, whose subject matter  encompass the "physical," or "natural" phenomena that are to be found within the domains of modern empirical sciences such as physics and chemistry. This has led  many writers to accuse Aristotle of advocating "armchair science." Though there is some truth in this accusation, Owen is right to argue that it is essentially an  overreaction which emanates from a fundamental misunderstanding of the nature and purpose of the treatises in question. 3  Leaving aside questions concerning the  truth and plausibility of their outcomes, the bulk of these investigations are best understood as a priori conceptual inquiries into the very metaphysical possibilities of  various general sorts of natural phenomena, and not attempts to codify the laws according to which these phenomena actually occur. Accordingly, these so­called  physical treatises4  should not be thought of as pretenders or rivals to modern natural science, but rather as essays on metaphysics. I do not want to dwell on this tired, old, and (to my mind) misguided issue here, though it will be constantly lurking in the background. For even if this essentially ad  hominen charge against Aristotle were dismissed, there is a closely related and more philosophical one that would still need hearing. What will be called the charge of  inflexibility may be put in the form of a potentially embarrassing question about the Superimposition phase of the Aristotelian methodology sketched above. What  right or reason does Aristotle have to believe, as he moves from one field of investigation to another, that his preconceived system of analytical concepts—which seem  after all to be a priori­will in every case be adequate to capture the salient features of the natural phenomena he is trying to understand? How, in other words, does he  know in advance that it will always be possible to achieve the Superimposition his method requires? Or to put it the other way around, how can he be sure that by  insisting upon Superimposition he is not distorting the facts he is attempting to explain?

  

Page 153

I believe it is possible to defend Aristotle at least partly against this charge, and moreover to do so without having to resort to Owen's unlikely and rather desperate  suggestion that Aristotle's system of analytical concepts, despite appearances, really is empirically grounded. The key to this defense is the crucial but generally  unnoticed fact that in the Superimposition phase of his method Aristotle is at least as interested in (and on the lookout for) misfit as he is in fit. That is, in a substantial  number of instances, the important lesson conveyed by an Aristotelian dialectical examination is not that the phenomenon under study falls neatly into one or another of  the categories in the analytical scheme which the method attempts to place upon it, but rather that in important ways it crosses over the distinctions invoked in the  analysis. Thus, in the end the object of study is frequently seen finally to reside in the seams and cracks of Aristotle's analytical framework—to be classified under one  heading according to some criteria, and under an opposing one according to others. Now if this is right, and that is primarily what I intend to show here, that would  provide grounds for seeing a great deal more flexibility in the method than appears initially. Although there are quite a few examples of Aristotelian dialectic which could be used to make my point (e.g., the analysis of growth in De Generatione et  Corruptione, I.5), I have selected just one, Aristotle's analysis of perception in De Anima II.5, because of its relative clarity and compactness. II By far the most common sort of endoxa examined and discarded in Aristotle's aporetic surveys are views (or caricatures of views) associated with his philosophical  elders or contemporaries. However, on some occasions, such as the one before us, they are what he refers to simply as 'dialectical propositions' in the Topics. 5   These are presumably wise old sayings attached to no one in particular but which seem at some level or other to contain an element of truth, though it must be added  that they are also usually so vague and general as to be nearly devoid of cognitive content. In the case of De Anima II.5 we are introduced early on to the following  pair. (D1) Like affects like. (D2) Unlike affects unlike. On the reasonable (if quite minimal) understanding that perception will have to be analyzed finally as some kind of interaction in which a perceived object affects, or  acts upon, a perceiving subject, Aristotle begins his investigation by referring us to arguments he gives in De Generatione et Corruptione I.7, to

  

Page 154

show that neither (D1) nor (D2) can be the whole truth about acting and acting upon in general. It is argued there that (D1), taken in an unqualified way, leads to the  absurdity that everything is always affecting itself (or, as he applies it here, that sense organs are always sensing themselves), whereas (D2), again taken as unqualified,  spells the impossibility of affective interaction altogether. This of course is only the beginning of the story. However, on the basis of our earlier description of Aristotelian dialectic, we can already predict how it is going to turn  out. We will be told eventually that each of (D1) and (D2) is "salvageable"—that is, that each, under the proper interpretation, correctly describes some aspect of  perception. Moreover, we can also predict that the materials for these respective interpretations will be generated during the Superimposition phase of the discussion  from portions of Aristotle's own analytical, metaphysical apparatus. These predictions do come true, but the procedure takes on an added dimension because of a facet of Aristotelian metaphysics not mentioned so far. Heretofore I  have been speaking of Aristotle's system of concepts as if it were all of a piece, whereas the fact is it can and (for the sake of full understanding) must be divided into  two easily separable parts, or better yet, levels. One of these we can call early and shallow, and the other late and deep. In view of the fact that the central and namesake doctrine of Aristotle's Categories is a fundamental classification of "things" (onta) and not linguistic expressions  (legomena), it is easy to think of this work as a sort of metaphysical manifesto in which the main concern is just to trot out his preferred set of ultimate ontological  divisions and then to comment on various interesting differences and relations among these. As tempting as this view may be, it is nonetheless distorted in an important  way. The Categories is not a work on metaphysics for metaphysics' sake; it should rather be construed primarily as an attempt by Aristotle to present a theory of  predication. That is, the major aim of the exercise is to produce an informal semantics for simple, affirmative, subject­predicate (atomic) sentences which respects  the distinction between necessary truths like: (1) Man is (an) animal. (2) Socrates is (a) man. (3) White is a color. and accidental (or contingent) truths such as: (4) Socrates is white.

  

Page 155

Moreover, Aristotle's aim is to accomplish this without recourse to the Platonic metaphysics of 'separated' universals (Forms), which by this time he eschews. Of course, semantics by its very nature requires at least some very broad ontological commitments, and one can in fact discern what could be thought of as a  classificatory, or "taxonomic" metaphysical substructure in the Categories. Briefly, the theory of predication presented there proceeds by connecting terms in  sentences with appropriate parts of an ontological framework. This ontological framework consists of a number—more or less ten—of hierarchical, or inverted tree­ like, structures formed by a single relation, 'X is "said­of' Y'. Such a theory of predication is meant to cover both the relation between genus and subordinate species  (as in [1] and [3]), and that between an individual and its natural kind(s) (as in [2]). Further, one of these structures (or "categories")—Substance—is accorded  elevated status because it contains the entities that are most real (i.e., primary substances, or concrete physical things) as its bottommost nodes. Each of the others  (Quality, Quantity, etc.) on the other hand, incorporates a subtype of what might now be generally referred to as 'attributes'. These attributes are all said to possess  only a secondary, or derivative sort of existence insofar as they stand in a second metaphysical relation (inherence) with primary substances (as in sentence [4] above).  Thus, in capsule, necessary atomic truths express intracategorial said­of relations, while contingent atomic truths express intercategorial inherence relations, where  the subject of inherence is always a substance. This account of the Categories semantics is oversimplified, 6  but the complications can be set aside. The important point I want to make and emphasize here is that  what we find in the Categories is no more than a minimal metaphysics—just enough to do semantics, one might say. The most obvious symptom of this minimalism is  the fact that neither of the metaphysical relations in which the system is grounded is explored in any depth. Aristotle appears content just to throw out a couple of  essentially substitutional tests for the "said­of' relation,7  and to characterize inherence in a purely negative manner. But clearly there are many pressing questions about  both these relations right beneath Aristotle's nose. Consider a small sampler. In the first place, if the said­of relation is to do double­duty to capture the relation  between genus and species, and that between kind and member, does this mean that there is no significant difference between these two subtypes? And what is it  about the relation between Socrates and his most proximate kind that accounts for the necessity of (2)? As for the inherence relation, notice that the Categories  semantics, as set out above, is entirely insensitive to matters of time and change; it gives a representation of "freeze­frames" of the world but has no way to depict  relations between such time slices. However, the fact is that inherent attributes (as opposed to essences) are of the sort that can be gained and lost by their subjects,  which is a large part of what is meant by saying that they are

  

Page 156

possessed contingently. How then to account for respective truth and falsity of such sentences as: (6) The man becomes cultured. (7) The man becomes a corpse.? These and other similar questions are never pursued in Aristotle's earlier works, and this is the reason I characterize the "taxonomic" metaphysics found in those works  as "shallow." On the other hand, Aristotle's progress from the physical treatises to the middle books of the Metaphysics, is marked by a growing concern about such  topics as change (of various sorts) and the relation between proximate species and their members. This evidently influences him to rework the early metaphysics by  forging "deeper" philosophical concepts such as matter and form, actuality and potentiality, nature and essence, and kinêsis and energeia in order to give accounts in  these areas. There are, to be sure, serious questions about whether the reworked metaphysics of the later works can or should be viewed as a reversal rather than an elaboration  of the Categories, and many of these questions center on precisely how the matter/form distinction is introduced into the system. 8  But however that may be, it at least  seems clear that Aristotle himself believes that his later, deep metaphysical doctrines are consistent with the shallow "taxonomic" system surveyed above, and this fact  has a significant bearing on the nature of Aristotelian dialectical inquiry. For there are a number of passages which indicate that Aristotle believes that a single question  can be resolved by the Superimposition of either the shallow or the deep metaphysics. Thus, the method (in its full ramification) in effect generates both shallow  resolutions and deep resolutions, and Aristotle sometimes even hints that they are simply convergent (if not equally perspicacious) means of expressing the same  underlying truth. De Anima II.5 is a good case in point. After announcing the general presupposition of his investigation, that perception is some sort of change of state brought about in  a subject by its being moved or acted upon by another object, Aristotle reminds us at 417a2 of the general account he gives of agent/patient interactions in De  Generatione et Corruptione, 1.7: If X affects Y, then X and Y must be like in genus, but unlike in species (or in determinate characteristics).

It is important to notice that this analysis does achieve some sort of resolution between (D1) and (D2), but also that it is a shallow resolution. It is curious that in De  Generatione et Corruptione itself, Aristotle evidently sees a link between this principle and considerations drawn from his deep metaphysics. This is

  

Page 157

because he quite consciously equates the "taxonomic" requirement that agent and patient be like in genus, with what would seem to be something altogether different,  that they be alike in matter (cf. 324a5ff.). He supports this by arguing that true affecting requires touching, and things different in matter cannot truly touch (323a30ff). How Aristotle can move with such facility between the concepts of genus and matter (and their respective correlates, differentia and form) is a question that I believe  ultimately leads into the puzzling but crucial ways in which his metaphysics intertwines with his biological work. But we do not have to follow that twisting trail here.  For there is also in De Anima II.5 a deep resolution of the aporiai about perception which makes use of a different pair of Aristotelian concepts. Again, relying on the  minimal pre­theoretic assumption that a perception in some way involves an external sensible object acting upon a perceiver, Aristotle now recalls his famous general  analysis of change developed in Physics III.1 and 2 in terms of his modal­metaphysical notions of actuality and potentiality. "A change (into the state of being F) is a  movement which is the actuality of that which is potentially (F) which occurs by virtue of its being potentially (F)" (Physics III.1 201a10–1). A full explication of this analysis would necessitate a detour into the highly abstruse Aristotelian concepts of kinêsis, entelecheia, and energeia. But it will be enough  for present purposes simply to note that the gist of the analysis is that a change into some state is an actualization of an object's potential for being in that state, where  the actualization comes about by virtue of that very potential. What is especially pertinent to De Anima II.5 about the Physics III.1 analysis is Aristotle's specific application of it to the specific sort of change that perception  seems most to resemble, namely alteration: qualitative change brought about by the action of one object upon another. A paradigmatic example of this would be a  cold body becoming hot as a result of being touched by a body that was already hot. Now we moderns are inclined to locate the potential responsible for this change  in the preheated agent, but the Physics analysis dictates that it must be located in the changing subject. What distinguishes alterations from other sorts of change,  according to Aristotle, is that they always involve the replacement of one of a pair of contrary qualities with another, for instance red with green, and further that the  potential for becoming the latter, which is assigned metaphysical responsibility for the change, is seen as somehow grounded in the earlier possession of the former.  This application of the general analysis to alteration is no more than hinted at in the Physics (or for that matter, in De Generatione et Corruptione), but its presence  there has been well documented by Sarah Waterlow Broadie, 9  and in any case is very nearly explicit in our chapter. Thus, as a first, rough approximation, we may  say that one central question of De Anima II.5 is whether perception is alteration.

  

Page 158

But there is another player in the game. Notice that there has not so far been an appearance by Aristotle's philosophical signature, the exploitation of systematic  ambiguity in his key terminology in what I referred to above as the Disambiguation phase of the inquiry. However, the terms 'potential', 'actual', and their cognates are  among those in whose equivocal natures he claims to find some of his deepest conceptual insights. Indeed, these two highly complex notions are important enough to  deserve treatment throughout the entire ninth book of the Metaphysics. Consequently, it is not at all surprising that the Superimposition in De Anima II.5 involves  some untangling of different meanings of these very terms. Specifically, this occurs as part of the modal­metaphysical analysis of alteration, wherein the potential to  take on a quality is grounded in the prepossession of a contrary quality. This analysis of alteration is not presented by itself, but is instead set against another sort of change analyzed in terms of an altogether different conception of  potentiality, one that is wholly unconnected to such notions as contrariety or privation. I shall call this other sort of change 'realization'. The potential to become F in  'realization', is understood to be grounded, not in the previous absence (or privation) of F itself. Rather, it is grounded in the previous presence of some inchoate  adumbration of F. The actual example relied on in De Anima II.5 is of a person who passes from the inactive possession of some intellectual capacity (specifically  knowledge of grammar) to its actual exercise. To be sure, this choice is well­suited to Aristotle's specific aims in the chapter, since it is an example of nonsubstantial  change, and so brings realization and perception as close as possible together. However, other texts suggest that what he generally regards as central cases of this sort  are substantial or quasi­substantial changes which occur in biological processes such as maturation and reproduction, where he characterizes a seed or juvenile as  possessing the potential to develop into an adult specimen. With this in mind, we can now improve on our earlier rough overview of De Anima II.5. To be quite accurate, the central question of the chapter is not whether  perception is an alteration or a realization, but to what extent it approximates each. Admittedly, if Aristotle presumed from the outset that it must turn out to be one or  the other, then it would indeed be difficult to defend him against the charge of theoretical inflexibility. But my central claim here is precisely that he does not operate  under that presumption and that in fact he ends up by concluding that perception is neither alteration nor realization even though it shares important common features  with each. It is argued, to begin with (417b16ff.), that perception is like knowledge. Just as the initial impartation of knowledge of grammar is an alteration, and its subsequent  exercise is a realization, the initial development of the sensory equipment is also an alteration, and its subsequent use of the equipment (i.e., perception itself) seem also  to be a realization. Here, there seems to be a parallel between the replacement of ignorance by learning in the student by the action of

  

Page 159

the teacher, and the replacement of insentience with sentience by the action of the male parent upon the female's material principle at conception. This parallel seems to  extend to the exercise of knowledge and perception, so that just as the exercise of knowledge is a realization, the same seems to apply to the exercise of perception. This may just be an analogy, and if so it invites serious questions about its justificatory worth, but for now I want to press only the methodological issues. It is crucially  important to see that Aristotle does not conclude at this point that perception is a realization. Instead, he follows the above reasoning immediately (at 417b19ff.) with  the observation that perception is also unlike paradigmatic cases of realization because it requires, as they presumably do not, an impingement by some external  object upon the subject of the change. And to this we might add the point made earlier that paradigmatic realizations, unlike perceptions, are substantial changes.  (Incidentally, converse considerations also show that the initial impartation of sentience is not strictly an alteration because the act of conception is a substantial  change.) To summarize the results of De Anima II.5, then, Aristotle's final position is that perception, strictly speaking, is neither alteration nor realization, though it exhibits  important features of each. On the one hand, it is, like alteration, a sort of nonsubstantial change brought about by the impingement of an external object on the  changing subject. But on the other hand it is like a realization in that it is the exercise of a pre­existent capacity or faculty. Thus it is, so to speak, quasi alteration and  quasi realization, which is the occasion for Aristotle's apologetic remark near the end of the chapter (418a1ff.) that his standard terminology is inadequate to reflect  all of the distinctions his investigation has uncovered. We are now in a position to appreciate the sense in which this deeper analysis "saves" both (D1) and (132). Perception is analyzed as a process in which a perceiver  is "acted upon" (in a nonstandard sense) by an external object, and as a consequence, "internalizes" the sensible features of that object in a manner in some way  analogous to a cold body's taking on a hot body's heat. What exactly this means is of course a long­standing problem I do not propose to tackle here. 10 But because  perception is represented as an extended temporal process, Aristotle sees himself as licensed to say (418a5–6) the following in the closing lines of the chapter. While  perception is ongoing and incomplete (i.e., prior to "internalization") the external object is unlike the perceiver it affects, whereas once it is completed they are alike.  Thus, in perception, it is true both that like affects like and unlike affects unlike.

  

Page 160

III What then of the charge of inflexibility? Suppose that my case regarding De Anima II.5 is clinched, and suppose further that the points made here can be generalized  to a significant number of other instances of Aristotelian dialectic as well. Then I will have shown at the very least that there is a substantial degree of flexibility as to  unforeseen results built into this method. But is it enough? Certainly it cannot be denied that even if one's methodology is tolerant to the possibility of analyzed  phenomena falling into the cracks between its fundamental classificatory concepts, the fact remains that whatever a priori concepts one begins with necessarily place  some constraints on the range of possible outcomes. My reaction to this complaint, on Aristotle's behalf, is to admit its truth and assert its inevitability. For it is equally certain that virtually any method of investigating  natural phenomena, no matter how "empirical" its manner of testing, must at the very beginning have something to test, and these, it would seem, must be "constructed"  out of concepts and distinctions whose applicability and even intelligibility, could not (on pain of regress) have already been successfully tested. This observation  applies as much to modern experimental sciences as it does do Aristotelian dialectic. But this inevitability notwithstanding, it is nonetheless reasonable to demand of  any program of inquiry that it possess some sort of feedback mechanism by virtue of which the character of new discoveries will be allowed to violate (and thereby to  modify) the fundamental conceptions and distinctions in which earlier hypotheses had been framed. What I have tried to show here is that whatever its shortcomings in  other areas, Aristotle's dialectical method for investigating natural phenomena does not suffer from comparison with modern scientific method on the issue of sensitivity  to this demand. Notes 1. G. E. L. Owen, "Tithenai Ta Phainomena," in Aristotle: A Collection of Critical Essays, J. Moravcsik, ed. (Garden City, N.Y.: Doubleday & Company, Inc.,  1967), 167–190. 2. I use the term 'dialectical' advisedly here. Since reading some recent unpublished work by Daniel Devereux on this subject, I am no longer as confident as I once  was that one can make a straightforward identification between the method of inquiry to be discussed here (which Devereux calls "the method of endoxa," and which  has its locus classicus in Nicomachean Ethics VII.3) and the method of dialectic discussed in the Topics. 3. Cf. Owen, "Tithenai," especially 187–90. 4. Incidentally, a treatise on the soul might seem like an odd location to find an

  

Page 161

      example of Aristotle's method of physical investigation unless one remembers just how "physicalistic" his psychology actually is (compared with, say, Plato's). 5. Topics I.10 104a12–38; examples given there are (1) that knowledge and perception of contraries are the same, and (2) that one ought to help one's friends and  harm one's enemies. 6. In fact, Aristotle struggles heroically (and with mixed results) to force various sorts of apparent "misfits" into this system, most notably predications of differentiae  and propria, and intercategorial predications with nonsubstantial subjects like: (5) The white (thing) walks.

7. Discussion of these tests is to be found in M. Ferejohn, "Aristotle on Necessary Truth and Logical Priority," American Philosophical Quarterly 18, 4 (1981):  285–94. 8. On this, see B. Jones, "Aristotle's Introduction of Matter," Philosophical Review 83 (1974): 474–500; A. Code, "The Persistence of Aristotelian Matter,"  Philosophical Studies 29 (1976): 356–67. 9. S. Waterlow Broadie, Nature, Change, and Agency in Aristotle's Physics: A Philosophical Study (Oxford: Clarendon Press, 1982), ch. 3. 10. R. Sorabji, "Body and Soul in Aristotle," Philosophy 49 (1974): 63–89; J. Barnes, "Aristotle's Concept of Mind," Proceedings of the Aristotelian Society 72  (1971–1972): 101–114.

  

Page 163

Chapter Nine—  Aristotle's Discovery of First Principles Allan Bäck Our discussion will be adequate if it has as much precision as the subject matter admits of; for precision is not to be sought for alike in all discussion.  (Eth. Nic. 1094b11–3)

Here I shall deal mostly with the question, how Aristotle arrives at first principles of the sciences. I offer a simple solution: Aristotle arrives at first principles in the  tentative, fallibilist way in which modern scientists do. Following Posterior Analytics 100b3, we can call this way, loosely, 'induction' ( ). 1  'Induction'  here should not be taken to amount to inductive logic or to statistical inference, where sample size and standard deviation offer methods to measure the strength of the  inference. Rather, I understand   amounts to a very messy mixture of looking at the available observations, reports, and expert opinions, analyzing and  drawing inferences from this material, and then theorizing, testing the outcome, and thereupon revamping the theory, including its first principles. The result is a more or  less coherent scientific theory, continually in process of systemization, unification, and revision. I take it that this description does reflect our current scientific practice. Have I committed an anachronism in attributing it to Aristotle? I shall try to show not. My approach has the merit of solving, or, rather, dissolving, some of the puzzles currently raised about Aristotle's method. At times Aristotle seems to distinguish  philosophy and science from dialectic, phenomena from endoxa,

  

Page 164

scientific demonstration from scientific research, and types of dialectic, strong and weak. Yet then, at other times, he seems to conflate these distinctions. If we view  Aristotle as a working scientist, more fallibilist and pragmatic than dogmatic, many of these contradictions disappear. For in different contexts with different materials  Aristotle would sometimes emphasize observation and other times the views of others. So too we do today in scientific practice. I consider my comparison not too  anachronistic, given that our scientific practice comes from the Aristotelian tradition. Too, I have some sympathy for Hintikka's complaint: Those who have not themselves wrestled with serious conceptual issues tend to underestimate by orders of magnitude the extent to which a great philosopher has to struggle  against confusion, contradictions, and other difficulties. 3

To put his point positively, I suggest that we consider Aristotle as a working scientist, interested not in looking at his theory but in using it. One advantage medieval  commentators have here over contemporary ones, despite their linguistic poverty, lies in their having this same pragmatic interest: they seek to do Aristotelian science,  not merely talk about it. Puzzles I shall begin by rehearsing some of Aristotle's doctrines about how we come to know first principles. These doctrines have occasioned many puzzles. For different  texts seem to conflict. Indeed the very diversity of the views of scholars in the secondary literature reflects such puzzlement and conflict.4 In its most general sense, 'induction' for Aristotle involves getting at the universal principles from items more particular and less primary (e.g., Eth. Eud. 1248b25–6).  In the sciences the primary principles consist of those common to them all, like the principle of non­contradiction, and those peculiar to a particular science, such as the  parallel postulate in geometry. Normal scientific explanation then consists in demonstrating theorems via deductive syllogisms from these universal first principles.  Somehow, then, although we have no innate knowledge of universals, we come to grasp the universal from the particular and have knowledge of the first principles  (Top. 105a13–4). At the end of the Posterior Analytics Aristotle gives a picture of how we come to grasp the universal from sense perception. By sense perception we become directly  acquainted with particulars. Repeated sensory experience of similar particular things produces a single memory in the mind, a phantasm applying indifferently to all  these particulars. When we apprehend this phantasm as universal, sc., as applying to all similar cases, even beyond those actually experienced, we have apprehended  the universal. Aristotle calls this ability to grasp the universal 'noûs' ('comprehension' or 'intellect'). 'Noûs' then is ''a

  

Page 165

generalizing capacity or ability that is responsible for the fact that a universal point, something, that is, which goes beyond what is grasped in sense perception, may  come to be present to the mind." 5  Induction in general then consists in "acquiring insight into some universal point as a consequence of attending to particular cases."6 The theory of this process does not concern me here. I am more concerned with how this process works out in practice. In effect, Aristotle has offered a mechanical,  psychological explanation: sense perception plus memory plus noûs yields the universal. The universals thus generated look mostly to be single concepts. But principles  are propositional. Perhaps we could get propositions via relational universals: we perceive so many particular cats on particular mats, apprehend the universal concept  of cat­on­mat, and come to the universal judgment that every cat is on the mat ( ). But it is hard to get to the principle of non­contradiction or the  parallel postulate in this way.7  For terms like 'contradiction' and '(perfectly) parallel' hardly seem given to us via particular sense perceptions. Even if we could, and did, apprehend universal principles in this way, the mechanical nature of this process sets Aristotle another problem: given that we come to  apprehend the universal without volition, willy­nilly, why then do not all people with the relevant experience agree on these principles? Like Socrates, or Descartes for  that matter, Aristotle seems to present the picture that if we but attend clearly to our thoughts, we cannot but help acquire knowledge of the universal. But most of us  do not. Thus, taken as a psychological causal process, induction does not suffice for us to discover the first principles of the sciences without error. Aristotle routinely links induction to experience and observation. The scientist observes, he says, 'the phenomena', given by sense perception (Cael. 306a5–17). That  is, she observes the particulars, and then comes to comprehend the universal by abstraction. Scientific knowledge becomes possible through making these  observations (An. Pr. 46a17–22). Aristotle contrasts this inductive, 'natural' method with a 'logical' or 'mathematical' method that proceeds from universal to particular  (An. Po. 79a2–6). He seems to attribute the latter method to, and even name it from, the Pythagoreans (Cael. 293a23–30; 297a2–6). He also seems at times to  identify this mathematical method with a "dialectical" method (Phys. 204b4). For this logical method starts with theory and opinion couched in general terms, not with  observation of phenomena. And where might these general remarks come from, if not from the previous opinions of experts or the people? In De Caelo et al.,  Aristotle, though, rejects this logical, dialectical method in favor of the natural method that begins from phenomena acquired by direct observation. But, as Owen made well known, Aristotle complicates his conception of phenomena considerably.8  For in his ethics he begins his discussion of akrasia, or  incontinence, by rehearsing "the phenomena." Yet he does not make

  

Page 166

observations of human behavior. Rather, he identifies these phenomena with the endoxa, the reputable opinions (1145b2–6). On account of this, Owen claims that  Aristotle uses 'phenomena' ambiguously: on the one hand to designate the empirical, given by sense perception, and on the other to designate the dialectical, given by  widely accepted or reputable opinion. Nussbaum offers a way to reunify Aristotle's conception of phenomena. 9  If even sense perceptions are theory­laden parts of the fabric of our web of belief, then the  empirical and the dialectical both concern opinion. We have no hard facts, just our beliefs about the world. Some beliefs may have a stronger tie to the stimulus  meaning of sensation than others, and so be more empirical. Still scientific knowledge, like all other conversation, falls within the hermeneutical circle of our society and  its ideology. Still, Aristotle does not always take this view. As we have seen, he often contrasts the natural and the logical methods sharply. In some texts Aristotle opposes  phenomena and endoxa, and in others he assimilates them, even in his treatises on natural science. (E.g., Gen. An. 725b5–6; 729b9–10; 760b27–33; 648a20ff.) We  can see how these texts support a variety of interpretations. We may well wonder if they are consistent. Similar problems arise with Aristotle's attitudes toward endoxa. For example, at the beginning of the Topics, Aristotle contrasts demonstration and dialectic.  Demonstration starts from premises that are primary and true, sc. the first principles of the sciences that are known to be true from themselves, or from those that are  derived syllogistically from such premises. In contrast dialectic syllogizes from endoxa, the reputable opinions "which are accepted by everyone or by the majority or  by the wise—i.e., by all, or by the majority, or by the most notable and reputable of them [sc., the wise]" (Top. 100b21–3).10 The scientific statements are true, while  the endoxa seem to be true.11 The gulf between phenomena and endoxa looks large. Phenomena give a true picture of the world; endoxa give a picture of what the world seems to be to most of  us. We have a division here resembling the divided line in Plato: endoxa belong to the sensible world of seeming, while phenomena, with the sciences based on them,  belong to the intelligible world of being and truth. Even if the endoxa are true, they do not qualify as science. Science requires more than having a true opinion; it  requires an explanation or account (An. Po. II. 1–2; Eth. Eud. 1216b 26–40).12 Aristotle agrees with Plato that we can have endoxa—reputable opinion that  happen to be true—without knowledge. Still, of course, Aristotle departs radically from Plato in basing science on the sense perception of the sensible world. As with  astronomy, all the sciences must preserve the phenomena.13 However, Aristotle does not equate endoxa in dialectic with majority opinion. He weights expert opinion more.14 As common people themselves yield to the opinions  of experts, we might say that deferring to experts agrees with common opinion, especially when the subject concerns something about which

  

Page 167

common people have no firm opinion (the sex life of nematodes), but even when the expert and the common opinions diverge. This weighting narrows the gulf between  phenomena that are true and endoxa which seem to be true. Besides, 'phenomena' themselves are literally 'appearances', how the world appears to us. 15 As we  judge, categorize, and reflect upon our sensations, the phenomena may well be thought to end up amounting to give what seems to us to be true, with special reliance  on those who have observed carefully and repeatedly. Aristotle blurs the difference between endoxa and phenomena even more when he distinguishes endoxa from  'apparent endoxa' (

) (Topics 100b25).

So at times the contrast between phenomena and endoxa seems to disappear. Aristotle himself identifies phenomena with endoxa in his ethical investigation of  akrasia. Again, in his study of the first principles common to all sciences in his science of being qua being, Aristotle again examines endoxa dialectically in the course  of establishing them as first principles. Finally, in justifying particular concepts in his Physics, Aristotle also appeals to endoxa, e.g., for his concept of place, although  he has stated his preference for phenomena in natural science (211a4–11). Even more striking is how Aristotle begins his Physics. He says that we "must advance from universals to particulars" (Phys. 184a23–4). He speaks of a universal  containing the particulars indistinctly and of our need to analyze it into its parts. He even suggests that in sense perception too we see objects at first not as singulars  but generically. Only with training can you come to distinguish a particular woman as your particular mother from women and mothers in general. This account of sense  perception hardly fits with his account of induction, when we come to apprehend the universal via the grouping together in experience of different particular  perceptions. Rather, as Nussbaum holds, he seems here to hold science to begin from the experience of a particular culture. Now we have endoxa grounding the  sciences too. According to Aristotle, dialectic itself has varied functions or uses.16 It can be used as intellectual calisthenics, to win debates in private and in public, and for the  philosophical sciences ( ) (Top. 101a25–8).17 The first two uses agree with the opposition of dialectic and  demonstration. For there the premises need not be true, but only agreed upon by the parties discoursing. Dialectic here has the goal of winning by persuading the  opponent. Aristotle gives a lot of advice on how to win, without much regard to truth in Topics VIII. All this fits his contrast between demonstration and dialectic. The third use is philosophical. In describing this use, Aristotle does not say exactly that dialectic justifies first principles. He says only that it helps us to see distinctly the  true and the false and to discuss the principles common to all the sciences (101a34–b4). But, as we have seen, in practice Aristotle regularly goes through the  reputable opinions and standing puzzles in the various sciences

  

Page 168

dialectically. Once again, all this runs together dialectic with induction, endoxa with phenomena, persuasion with truth. What more does he think we can do? At times Aristotle says that we can do more: by noûs. Aristotle agrees with Plato that we can have a simple act of apprehension of first principles after a laborious  process involving dialectic. Noûs is that ability to apprehend them. Aristotle has said that we apprehend first principles by induction through noûs. But we have seen  that in induction noûs consists in the ability to grasp the universal, but gives little assurance of infallibility. Thus, the first principles of sciences, known to be true through  themselves, appear justified by endoxa and dialectical reasoning. An Irwin Excursus Although many have disagreed with his conclusions, Terry Irwin has given an influential account of Aristotle's dialectic, responsible for much of the current discussion  of Aristotle's dialectic. As my own view can also be illuminated in contrast to his, I shall discuss his account briefly. Irwin's book owes its great length to collating and explicating both the texts where Aristotle mentions and discusses dialectic, and those where he uses dialectic via  arguing dialectically for the first principles of various sciences. Irwin's results may be stated succinctly: Aristotle has two methods for arguing for the truth of his claims:  the empirical and the dialectical. The empirical is based on direct observation of individual events, the phenomena, and proceeds to the universal from the particular via  induction. Aristotle assumes, Irwin says, "naively," that our experience of the world mirrors the objective structure of the world. 18 The dialectical is based on  collecting common beliefs (endoxa) about a certain topic, and proceeds through systematizing them. The common beliefs on some topic are those opinions that are  and have been held by people about it, although Aristotle is inclined to give more weight to opinions of the educated and the expert rather than to those of the ordinary  person. In the natural sciences, Aristotle seems inclined to give more weight to the empirical or inductive, than to the dialectical; in the moral "sciences," like ethics and  politics, he favors the dialectical over the empirical.19 In arguing for first principles, Aristotle uses the dialectical method almost exclusively. After all, even if the empirical method provided any relevant materials, using it  would beg the question here.20 We might try to justify, e.g., the principle of non­contradiction by noting that we have never observed an instance of a thing with  contradictory attributes. But such observations, like the inferences made from them, assume the principle of non­contradiction.21 We are left only the dialectical  method with which to justify the axioms.22 We now approach the crux. In general dialectic begins with established opinions and works from there, by codifying and organizing them, by re­

  

Page 169

articulating and modifying them in the face of objections ( ). We have here a method quite like the one implicit in many current studies today in applied  ethics. 23 As we organize these beliefs, we then should reach conclusions about which beliefs are most central and fundamental for our experience and attitudes. Thus  we should end up with results where (most of) those beliefs that were most strongly held by people at the start will be sanctioned as the first principles of one or more  of Aristotle's sciences.24 This method Irwin calls pure dialectic. Perhaps Aristotle uses it in his metaphysics of substance, which captures the ordinary intuition that  individual physical objects are fundamentally real. Yet even here, Irwin admits, Aristotle's conclusions far outstrip the original ordinary beliefs.25 However, at other times, even at the start Aristotle seems to select and emphasize only certain of these fundamentally held beliefs, while rejecting and ignoring others.  Perhaps he does so most clearly in his ethics, where he ends up with the conclusion that the good for human beings consists in a contemplative, virtuous life—a  conclusion that is rejected by the behavior if not by the beliefs of most human beings.26 Here Aristotle has a principle of selection for his dialectic that enables him to  get results radically divergent from common beliefs.27 Irwin calls this method strong dialectic.28 Aristotle looks "naive." For surely Aristotle's principle of selection begs the question: only those common beliefs are admitted or emphasized that advocate or conform  to the principles of Aristotle's science. But then the selection procedure contains already Aristotle's first principles implicitly. So, inasmuch as Aristotle uses strong  dialectic to justify his first principles, he is naive. Even pure dialectic will not help matters much: at best Aristotle will embrace the fundamental beliefs or prejudices of  his culture—a result particularly clear to some today in regard to Aristotle's ethical and political views. Irwin tries to avoid this conclusion when he discusses Aristotle's discovery of metaphysics in Metaphysics IV. Here Aristotle discovers or invents a science of being  qua being which will validate the fundamental principles common to the special sciences in more than a purely dialectical way. According to Irwin, this science uses  strong and not pure dialectic. By selecting the "right subset of common beliefs," "we could perhaps find conclusions with a better claim to objective truth" than those  reached by pure dialectic.29 Irwin himself immediately raises the question, how we are to find the ''right subset" without begging the question? He says that the situation  is not hopeless: we may be able to find some reasons for the superior reliability of certain of our ordinary beliefs; e.g., for the truth of the principle of non­ contradiction.30 What are those reasons? In effect Irwin isolates two reasons. First and especially, that we cannot give up such a principle without giving up the pursuit  of science (—at any rate, Aristotle's science) and "rational discourse about how things are;"31 second, that

  

Page 170

any opponent to this principle will also beg the question in the opposing argument. 32 Neither of these reasons is compelling. First, someone who objects to Aristotle's approach would hardly be fazed by having to give up Aristotle's science and  discourse that is rational according to Aristotle's standards. Imagine Aristotle having a debate with a Zen master!33 Again, the second reason is open to the same  objection: an account constructed by Aristotle (in Met. IV) of how a radical relativist like Protagoras might reply is surely no substitute for an independent reply. If  not, then I alone should be writing a critique of this paper. Anyway, the second reason looks like a classic instance of an et tu quoque fallacy: if you can beg the  question, I can too.34 So Irwin does not offer a favorable picture of Aristotle's systematic enterprise. He claims that Aristotle's theory of perception according to which sensation strictly  speaking is infallible, is (1) not only naive, "as it does not occur to him that the sensory 'appearances' ... may themselves be the product of theory and  interpretation," (2) but also inconsistent with his metaphysics, for his metaphysics defends realism against the phenomenalist Protagoras, whereas Aristotle has to  support phenomenalism in order to explain sensory illusion.35 Irwin concludes that Aristotle does not follow his own metaphysical principles in his account of  perception, or, for that matter, elsewhere, in his distaste of meniality and extreme democracy, in his support of public moral education, and in his views about the  middle class and private property.36 (Is it a coincidence that Aristotle does not follow his principles in precisely those cases where a modern Western person would  disagree with his views?) Irwin's view, then, saves Aristotle's first principles at the cost of his being incompetent in applying them.37 On Irwin's view, Aristotle is naive—hopelessly so, it would appear. Note that if Irwin is right, we do not have much motivation to study Aristotle as a systematic  philosopher—except for Irwin's remark that we are not much better off.38 Maybe some of his particular points still have merit or offer insights to contemporary  philosophers. Certainly, his work has great historical importance. Perhaps we can learn from his mistakes, although given human history this suggestion does not have  too much support. But the general thrust of Irwin's book (not to Irwin himself so much as to philosophers in general) is to relegate Aristotle to merely antiquarian  interest.39 The texts are difficult anyway: even after millennia of commentary, specialists do not agree about such central issues as: whether or not Aristotle is a  Platonist; whether or not Aristotle holds substances to have individual essences. And, if on top of this, Aristotle has a silly method, or applies his method so poorly,  why devote a chunk of your life to reading him?

  

Page 171

An Aristotelian Solution I wish to argue against this conclusion on several grounds. I shall claim, contra Irwin, that Aristotle's method is not "naive," in a pejorative sense: first, by making some  historical observations; second, by offering a different view of Aristotle's dialectic; third, by considering how else first principles might be defended. The third point will  lead me to my conclusions about how Aristotle discovers first principles. 40 First, it is commonly asserted that Aristotle was quite naive about matters of epistemology. Just look at his account of perception, where the form of the substance is  mirrored faithfully in the soul of the perceiver!41 Yet consider Aristotle's tradition. He held himself to be a follower of Socrates, who did worry a lot about how it is  possible for us to know what we know, and whether we know anything. Aristotle himself had to advance his own views in the milieu of sophists and cynics. Even his  ethical views should not be considered parochial: Aristotle lived in the Macedonian Empire and was acquainted with a wide variety of customs and laws. Note that his  own ethics, which extols the life of an idealized, intellectualized Pericles, did not suit his own current historical conditions. Aristotle took great pains to collect  information about others' beliefs.42 In such a pluralistic environment, we might have doubts about how simpleminded Aristotle was. You might imagine someone being raised in such a society and being well aware of quibbles, caveats, and objections, and yet still proposing a simple method and  outlook. This may be naive, but not in a pejorative sense: rather in the sense of naïveté that Nietzsche attributes to the Greeks.43 Some modern parallels might make this approach more convincing. Despite our pluralistic culture, and the strong legacy of Cartesian doubts coupled with the linguistic  turn, we still manage to find philosophers today who advocate realism and accept a causal theory of perception.44 Certainly the vast majority of scientists are realists,  in practice if not in theory. Again, later in the ancient world, the doctrine of apprehension ( ), which appealed to the sophisticated Stoics, looks naive.  For what is apprehended looks already judged and categorized. But such a view is naive only if there is an alternative. The classic alternative lay in the apprehension of  simple sense data, from which a world was synthesized. Yet this has today been dismissed as the myth of the given; we live in a world of experience and interpreted  objects.45 In such a world, how naive is it for us to start in the middle of things, with interpreted facts? Rather, if we accept the contrived, abstract nature of the  simples of a foundational approach, it is more naive to talk confidently of sense data or whatever as the ultimate, atomic given facts known by direct acquaintance. So  Aristotle's approach, of starting in the middle of things, not only is not goofy but might be right.46

  

Page 172

If, on the other hand, such an approach is sick intellectually, it is not only philosophers who are thus infected. Scientists typically begin with observations and  explanations couched in terms of ordinary experience. They will also survey and review previous data and theories in their field, if available. They generally are  painfully aware of the inadequacies and shortcomings of their initial data and theories. But they have to begin somewhere. Then, by working through this initial 'given',  by organizing and analyzing it, by collecting further pertinent data, they may arrive at theories and make observations far removed from the conceptions of ordinary  experience. For instance, take the history of chemistry. The initial observations and theory were couched in terms of everyday macroscopic objects and their qualities that are "the  most obvious to us." Further analyses, observations, and some intuitive hunches led to a chemistry put in terms of microscopic objects whose fundamental properties  are not qualities like colors, textures, and tastes, but quantities, like charge, mass, and spin. Modern chemistry offers a view that not so much clashes with ordinary  views than is disjoint with it: its theoretical concepts and experiments deal with items that most laymen do not even dream of. It does not use ordinary concepts or  means of sensation to make its observations. Yet it began from ordinary beliefs and observations, and typically a chemist in beginning a research project will survey the  past literature and data in the field, or, if that is not available, the beliefs of common lay experience. But then Aristotle's dialectic itself does not look so naive either. For he too starts with ordinary beliefs and observations, and, if available, the work of previous experts  in the field. 47 He too analyzes the beliefs and attempts to collect more data. In the course of his analysis, Aristotle likewise tends to come up with new distinctions and  concepts. Thus, in the Categories, Aristotle brings in notions like paronymy, propria, and differentiae and invents new names for new phenomena (10a32–b2).  Irwin is right to point out that these notions go beyond the common consensus: indeed, these notions are not even part of the conceptual structure of ordinary belief  (although over the course of history they, like many other scientific conceptions, like atoms, may come to be so). So Aristotle moves away from, or, better, goes  beyond common beliefs. If this be his method, he has lots of company: Spinoza too talks of first using a rudimentary tool and then later discarding it; likewise  Wittgenstein has his ladder. This view of Aristotle's dialectic that I am proposing has one major difference from Irwin's: Aristotle's results do not have to be justified by agreeing with the ordinary  beliefs. To be sure, they begin from ordinary beliefs and observations, as well as from those of experts and sages when available, as with a survey of the literature in a  modern science.48 Moreover, they must take account of past beliefs; that is, they must give an account of how people come to believe them. Making an account does  not mean to agree with them, but only to explain them (Met. 1073b10–7). Hence I do not see dialectic and science as two separate methods, but rather as two stages  of a critical, 'scientific' enterprise.49

  

Page 173

For an example, let us take Aristotle's conclusion about the good life for a human being: contemplative and virtuous activity. He surveys past opinions, both of ordinary  people and of the sages (or experts). He organizes them, and concludes that people have found three types of lives good: sensual gratification, fame, and  contemplation. He then brings in his sociobiological scientific account of human functions, itself based on common beliefs and experiences, a survey of past expert  views, and additional observations. He then concludes that contemplation is the human good. 50 Now, as I said earlier, this result would be rejected by the majority of  the human species. Aristotle's account conflicts with the endoxa of the ordinary people, although perhaps not with those of past philosophers (themselves a rare  breed). He needs to resolve this conflict, and can do so by offering an account why people commonly hold erroneous views here, and thus (in Irwin's translation)  prefer the lives of grazing animals. Aristotle does so by giving an account of desire for the preference of those who are "not properly brought up" and so do not have  intellectual sensitivity,51 and an account of akrasia for those who know better but act badly. Again, we have a parallel with modern science and not fortuitously, given Aristotle's influence on modern Western culture. We may collect ordinary beliefs and  observations about the celestial bodies as well as past astronomical theories. We proceed through analysis and further observation to codify and improve on these.  We may end up rejecting much earlier material as inaccurate or as couched in inappropriate terms: a past sighting of the sun at a certain point in time might be  inaccurate; talk of Apollo might not be productive. Say we end up with the theory of Newton. We still need to explain the common perceptions: why the sun appears  very small, and moves around the sky; why we appear to be standing still while the sun moves.52 Newton's theory explains the cause of these perceptions and the  content of past theories based upon them, while rejecting their literal truth. I contend that Aristotle proceeds similarly in his investigations. So, I hold, if Aristotle is naive, he has lots of company. Finally, let me ask, what other ways are there of justifying or sanctioning first principles? At times Irwin  distinguishes Aristotle's realist approach from modern approaches, like the psychological one of Hume and the phenomenalist one of Kant. But these distinctions differ  not in how to justify first principles, but in what states of affairs first principles have to be based on. Psychological facts or perceptual facts are not any more self­ validating than physical facts: e.g., the notion of a sense datum is quite an abstract one, and becomes self­evident to people only after a period of philosophical training,  if not indoctrination. Once again: not self­evident to us, but self­evident in themselves? I am guessing here about how Irwin perceives alternative ways to justify first  principles, as he does not lay them out in much detail. Again, if we consider modern practice, Aristotle's procedure does not look too strange. In axiomatic theory construction, we first take as axioms candidates

  

Page 174

that look plausible initially. After we work out some of their consequences, and consider our pragmatic interests in the theory, we may well modify the axioms and the  theory so much as to make the beginning theory a remote and dissimilar ancestor. 53 Again, Aristotle seems to take a similar course: he begins with a science and its  current state of the art, wrestles with the problem, and proceeds by making one or more fresh starts (e.g., De An. I.2; II. 1). His result too may be unfamiliar to those  familiar with the initial state alone. For instance, Aristotle lives in a society relying on divination and praying to many gods. His own view hardly reflects the views of the  common people or even those of his reputable predecessors. What other ways are there to justify first principles? Divine revelation? But Aristotle like us appears wary of the divine and of the Socratic daimon. Intuition of the  principles themselves? Aristotle himself stresses the role of noûs, although Irwin claims that Aristotle has rejected that approach in Metaphysics IV.54 But noûs,  though guaranteeing generality, does not guarantee infallibility (Met. 1074a15–7). Direct acquaintance with certain atomic facts? But this approach has been generally  rejected in recent decades, by philosophers of many traditions. As I have noted, Irwin himself says that we do not have a better way than Aristotle's for justifying first  principles. Yet, so I have claimed, he has defended Aristotle's method only at the cost of having Aristotle be incompetent in applying it. In contrast, I have urged that  the inductive and the dialectical not be taken as two separate methods, but as two stages of the same method and that Aristotle's discussions be viewed as "naive" only  in the sophisticated sense. Conclusions If we grant Aristotle a pragmatic, fallibilist method, many of the puzzles about his approach toward first principles disappear. Aristotle stresses observation by direct  acquaintance when the phenomena are ready at hand. Yet even then, e.g., in biology, he does not fail to review the theories of his predecessors. When he lacks such  phenomena, he relies more on previous reputable opinions and custom. So too he operates in his ethical and political theory. Even in his Physics he has a more  dialectical discussion of first principles. For how can anyone verify the existence and nature of moving bodies or of place by direct observation without begging the  question? In an Aristotelian spirit, it helps to distinguish senses of the question: how do we come to knowledge of the first principles? We get at, or discover, candidates for first  principles through a messy process involving direct observation of phenomena and generalization therefrom, as well as through a critique of current views in the field.  Noûs makes it possible for us to formulate abstract general principles. We experience the activity of noûs as a simple act of intellectual insight. We also assume that if  we proceed carefully the

  

Page 175

generalization process gives reliable results. We have to assume this do to science: compare Popper. 55 Though pragmatic, i.e., willing to accommodate these  standards to what we can obtain and achieve in practice, he views science as a progression toward the absolute truth about reality. Still, this mental act of noûs does  not guarantee truth. We may not have made enough observations, we may have misidentified what we see, as in the Coriscus example; we may have reasoned  fallaciously, like Melissus. We then have to test the candidates for first principles thus obtained. We do this both by dialectical (including logical) analysis and by looking at the explanatory power  of those principles and their agreement with experience.56 Surviving dialectical examination becomes a necessary though not a sufficient condition for establishing first  principles.57 By itself dialectic does not prove or establish principles. Rather, it tests them, and passing this test partially justifies them (Soph. El. 172a21).58 Dialectic  has other uses: in helping to discover first principles from endoxa, as well as in logical debate and in intellectual exercise. Aristotle offers his theory of demonstration as a description of the structure of scientific theory so as to be able to test the explanatory power of proposed first  principles.59 It is no coincidence that Hempel's D­N model of scientific explanation looks so Aristotelian.60 Hempel gives four criteria for assessing a scientific theory:  clarity and precision; explanatory power; formal simplicity; confirmation by experience.61 These criteria look Aristotelian to me. Aristotle too would like his principles  to explain thunder, eclipses, and olive harvests. Yet, though tentative and not evident to us, first principles are absolute and evident in themselves. That is, we assume  the first principles that we in fact have at some time to be true, while at the same time conceding that we might not have apprehended the right ones. Like modern  scientists Aristotle acts as if his principles are true, even though he is willing to change them.62 In short, the model of demonstration in the Posterior Analytics  presents the ideal; the treatises on the sciences present actual practice. The scientific treatises are works in progress and are concerned with arriving at conclusions in  the special sciences. The logical treatises focus on justifying the results and fitting them into a single unified theory. So too with D­N explanation and actual scientific  work. So I find it better to compare Aristotle with Hempel or Quine (in his more objectivist moments), than with Putnam or Goodman, as Nussbaum tends to do.63 I do admit that at times Aristotle sounds as if he accords noûs an infallible status as a state of soul by which we grasp the truth and never make mistakes (Eth. Nic.  1141a2–6; cf. An. Po. 100b5–14). Perhaps Aristotle has not completely departed from his dogmatic Platonic origins. Yet in practice Aristotle looks much more  flexible in pursuing the truth. I do not have the space here to defend this large claim. But let me enumerate some of the evidence for it. First, Aristotle will abandon lines  of inquiry but not his conviction that he can move toward the truth: his penchant for making 'a fresh start'. Again, in his

  

Page 176

logical works Aristotle seems to favor somewhat the structure of the Porphyrian tree for the categories, whereas the relation of genera and species becomes much  messier in the biological works. 64 Again, when he defines 'demonstration', he insists that it must give the cause for ( ) where the effect 'explains' the cause (An.  Po. I.13). Again, Aristotle wants scientific knowledge to comprehend what is by necessity (Eth. Nic. 1140b31–2). Yet, even theoretically, he accepts demonstration  from premises that hold only for the most part (An. Pr. 32b3–6; 43b32–5; An. Po. 87b21–2).65 Given his recognition of spontaneity and monstrosities, where  imperfections in the matter frustrate the natural actualization of the essence, Aristotle has to be satisfied with that! (Phys. II.5–6). Again he speaks of making  demonstrations if the attributes of the object have been apprehended and if none of them have been omitted (An. Pr. 46a22–7). He speaks of those that belong to the  object "apparently" and those that do so "in truth" (43b8–10). He concludes that the truer the apprehension, the better the demonstration—as if demonstration were  a matter of degree. Surely this is a tentative approach. We can see Aristotle making similar allowances for a fallible grasp of first principles by noûs. He describes the first principles also as 'more convincing and better  known to us'—as opposed, perhaps, to 'better known in themselves' (An. Po. 72a31–2, 72b1; Top. 141b5–14). He speaks of "presupposing" the principle of non­ contradiction (71a12). He also allows for different methods of reaching first principles depending on circumstances (Met. 1098b1ff.). All this makes thinking fallible.  For Aristotle likens perceiving to thinking ( ) (427a19–20). As Aristotle allows for error in perceiving as well as in thinking, I find it plausible that  noûs can make mistakes (427b13; 428b25–30). Again, in the Rhetoric Aristotle says that the same ability enables us to know the true and the plausible, and surely  both of these are not infallible (1355a14–7). Above all, over and over Aristotle claims that none of his predecessors have grasped the whole truth, about causes, first principles, the good, etc.66 Like modern  scientists, he has the optimistic attitude that we all grasp some aspect of the truth, although perhaps through a glass darkly. But do not take this attitude to amount to  the claim that noûs is infallible. For equally well it implies that (nearly?) all (human) acts of noûs contain some portion of error!67 It is ironic that Aristotle is said to  have inherited the doctrine of infallible noûs from Plato. For Plato's ultimate use of noûs is the apprehension of the Form of the Good. But Aristotle thinks that Forms  are an illusion, that 'good' has no unity—in short, that Plato was dead wrong in that apprehension. In sum, Aristotle seems to view thinking like sense perception: both generally give reliable results, but both can be mistaken.68 Consequently I think that what Aristotle  says about noûs presents no obstacle to my account of how he says we come to grasp first principles. Indeed, many of the problems in

  

Page 177

current Aristotelian scholarship may arise precisely from not appreciating the tentative, fallibilist nature of Aristotle's thought: perhaps they are false dilemmas. As  Hintikka has remarked and the medievals have appreciated, when we try to use Aristotle's philosophy instead of merely talk about it, we may well find that we too  have to adjust the ideal to the messy material reality. 69 Irwin has then no need to judge Aristotle's dialectic—or even its applications—naive. Or, if it be naive, it is naive in the way of modern science and philosophy. Unless  someone can propose something better, I see no reason not to take Aristotle's method and philosophical system seriously—as philosophy, and not merely as a quaint,  historical artifact. Anyway, that is why I devote a chunk of my life to Aristotle. The investigation of the truth is in one way hard, in another easy. An indication of this is found in the fact that no one is able to attain the truth adequately, while, on the other  hand, no one fails entirely, but every one says something true about the nature of things, and, while individually they contribute little or nothing to the truth, by the union of all a  considerable amount is amassed. (Met. 995a24–7)

Notes 1. Most clearly in the first sense of  Cf. Alexander, in An. Pr., 43, 27. 2. Kurt von Fritz, Die 

 distinguished by T. Engberg­Pedersen, "More on Aristotelian Epagoge," Phronesis 24 (1979): 301–19, viz. 301–3. 

 bei Aristoteles (Munich: Sitzungsberichte der Bayerishchen Akademie der Wissenschaften, 1964).

3. Jaakko Hintikka, "The Development of Aristotle's Ideas of Scientific Method," in Aristotle's Philosophical Development, W. Wians, ed. (Lanham, Md.: Rowman  & Littlefield, 1996), 131–50. 4. Robert Bolton, "The Epistemological Basis of Aristotelian Dialectic," in Biologie, Logigue et Métaphysique chez Aristote, D. Devereux and P. Pellegrin, eds.  (Paris, 1990), 185–235, viz. 185–7, 194, notes how much views on the significance of dialectic for Aristotle have vacillated. 5. T. Engberg­Pedersen, "More on Aristotelian Epagoge," 308. 6. T. Engberg­Pedersen, "More on Aristotelian Epagoge," 305. Cf. Topics 105a13–4; 100a6–7. 7. At one point he seems to imply that the principle of non­contradiction needs only the apprehension of its existence (An Po. 71a13–4). Yet that principle does  require combination of simples into a sentence, and so admits of truth and falsity. So probably it is not a single concept. Or, perhaps, we might take Aristotle to say,  like Descartes, that in the case of logical tautologies no inference is needed but only "the light of reason." 8. G. E. L. Owen, "Tithenai ta Phainomena," in Logic, Science, and Dialectic, M. Nussbaum, ed. (Ithaca, N.Y.: Cornell University Press, 1986), 239–51, viz.  240. 9. Martha Nussbaum, The Fragility of Goodness (Cambridge: Cambridge University Press, 1986), 244, 274–5.

  

Page 178

10. See J. D. G. Evans, Aristotle's Concept of Dialectic (Cambridge: Cambridge University Press, 1977), 77–8; Jonathan Barnes, "Aristotle and the Methods of  Ethics," Revue Internationale de Philosophie 34 (1980): 490–511, viz. 498–9, on difficulties of translating 'endoxon'. I shall just use 'endoxon'. 11. Alexander, In Topica 21, 31–22, 1, says that endoxa as such are neither true nor false. 12. Cf. Alexander, In Topica 16, 8–31. 13. William Wians, "Saving Aristotle from Nussbaum's Phainomena," in Essays in Ancient Greek Philosophy V, A. Preus & J. Anton, eds. (Albany: SUNY,  1992), 133–49, viz. 135. 14. C. D. C. Reeve, Practices of Reason (Oxford: Oxford University Press, 1992), 35–6. Cf. n. 59. 15. Jonathan Barnes, "Aristotle and the Methods of Ethics," 491, n. 1. 16. Following Enrico Berti, "Does Aristotle's Dialectic Develop?" in Aristotle's Philosophical Development, W. Wians, ed. (Lanham, Md.: Rowman & Littlefield,  1996), 105–30, viz. 116, I shall assume that Aristotle has a single dialectic with many applications. 17. Robin Smith, "Aristotle on the Uses of Dialectic," Synthese 96, no. 3 (1993): 335–58. 350, notes that strictly speaking Aristotle states that these are the uses of  his treatise on dialectic, sc., the Topics, and concludes, 352–3, that Aristotle does not use dialectic to establish first principles. 18. Irwin, Aristotle's First Principles (Oxford, 1988), 26, 33–5. 19. Irwin, Aristotle's First Principles, 23, 30. 20. Irwin, Aristotle's First Principles, 70, 118, 138, 153. 21. Irwin, Aristotle's First Principles, 48. 22. Irwin, Aristotle's First Principles, 196. Irwin claims that Aristotle, opposing Plato, first held, in the Posterior Analytics and Physics that the first principles were  justified by intuition (noûs), rather like Descartes (ch. 7 n. 7, ch. 8 n. 8). However, Irwin says that Aristotle changes his mind in Metaphysics IV, as I shall discuss. 23. There is, to be sure, interplay between the empirical and dialectical methods, but apparently not enough for Irwin to prevent them from being separate and distinct  methods. 24. I am thinking of those articles that present a series of examples, often hypothetical, elicit certain judgments, and then systematize and generalize our reasons for  them. For example, Judith Jarvis Thomson, "A Defense of Abortion," Philosophy and Public Affairs 1, no. 1 (1971): 47–66. Indeed, Russell states the approach  nicely at the end of his The Problems of Philosophy (Oxford: Oxford University Press, 1959), ch. 14. 25. Irwin, Aristotle's First Principles, 38, describes Aristotle's dialectic to consider "theses" that seem at odds with common views. A necessary condition for  accepting such a thesis is that it be "more consistent than its denial with the bulk of the common beliefs." 26. Irwin, Aristotle's First Principles, 52, 66–7, 73, 87–8, 92–3, 109. 27. Irwin, Aristotle's First Principles, 350–2, 355, 358, 437–8. 28. At least, Aristotle can obtain more divergent results than he would via pure dialectic. Yet, even there we have seen that Irwin allows that Aristotle judiciously  selects and organizes common beliefs.

  

Page 179

29. I agree with Enrico Berti in his review of Irwin's book, in Elenchos 12 (1988): 116–25. Viz. 123, that Irwin has no need to make this distinction, not made  explicitly by Aristotle. Rather, pure dialectic corresponds to the first two uses of dialectic discussed above, and strong dialectic to the third, philosophical use. 30. Irwin, Aristotle's First Principles, 175. 31. Irwin, Aristotle's First Principles, 176. 32. Irwin, Aristotle's First Principles, 188, 176, 195. 33. Irwin, Aristotle's First Principles, 185–6. 34. See the criticisms of Aristotle in Toshihiko Izutsu, Towards a Philosophy of Zen Buddhism (Boulder, Colo.: Prajna Press, 1982), 23, 28–9. 35. Perhaps not: this reason might be taken as an instance of dialogic of the sort of Hintikka or Paul Lorenzen. But then we are not using dialectic at all but, it seems a  form of science. Indeed, it is puzzling on Irwin's account why Aristotle would call his first philosophy a ''science" at all. Irwin himself considers this objection,  Aristotle's First Principles, 168, 174; ch. 9 n. 13; ch. 9 n. 60. His solution seems to be that Aristotle is introducing strong dialectic, a scientific form of dialectic, 175.  We might then wonder whether Irwin was right to separate science and dialectic so sharply in the early treatises. 36. Irwin, Aristotle's First Principles, 315. Note that this criticism attacks Aristotle's empirical method too. 37. The passages in sequence are: Irwin, Aristotle's First Principles, 326–7, 414, 455, 420, 461, 465. 38. Irwin also claims, Aristotle's First Principles, 49, that Aristotle makes a mistake even in the basic distinction of the empirical and the dialectical methods: the  dialectical method is not as realist as the empirical method, but Aristotle thinks it is. 39. Irwin, Aristotle's First Principles, 484. That is, we have not found any better way to justify first principles. 40. Irwin himself has reservations, but endorses the serious study of Aristotle as a systematic philosopher, Aristotle's First Principles, 480, 483. Irwin finds the  Cartesian approach, which is similar to Aristotle's intuition by noûs, even worse, ch. 7 n. 7; ch. 8 n. 8. 41. See too John Cleary, "Phainomena in Aristotle's Methodology," International Journal of Philosophical Studies 2, no. I (1994): 61–97, viz. 76–81. 42. Irwin, Aristotle's First Principles, 26, 315. Cf. Richard Rorty, Philosophy and the Mirror of Nature (Princeton: Princeton University Press, 1979), 38–41. 43. Irwin belittles how difficult it is to collect these views. But, even by our standards, Aristotle went to a lot of trouble. Given the difficulties of communication in his  time, surely much more! 44. The Birth of Tragedy 3–4; cf. Friedrich Schiller, "Über Naive und Sentimentalische Dichtung." The naïveté here has its differences, but the basic stance of taking  a simple approach for sophisticated reasons remains the same. 45. E.g., Harold Brown, Observation and Objectivity (Oxford: Clarendon Press, 1987), and also Putnam, Laudan, et al. 46. Not merely by Wilfrid Sellars and Roderick Chisholm, but also by many others in a variety of traditions; e.g., Hans­Georg Gadamer in Truth and Method,  second edition, J. Weinsheimer & D. Marshall, trans. (New York: Crossroad Publishing, 1989). 47. Irwin might be alluding to this issue at Aristotle's First Principles, ch. 2 n. 2. As I have remarked, Irwin claims that we are not much better off than Aristotle in  having a

  

Page 180

      way to justify first principles. So far we agree. However, Irwin ends up with the view that Aristotle does not apply his own methods well and indeed is confused  about their structure and interrelationship. In contrast, although I do not take Aristotle to be an omniscient deity, I am urging that Aristotle is not so goofy but naive. 48. Cf. John Cleary, Aristotle and Mathematics (Leiden: E. J. Brill, 1995), 201–4, on Aristotle's use of aporiai. 49. So too Richard Kraut says in his review, in The Philosophical Review 101, no. 2 (1992): 365–71, viz. 367–8. Also cf. Robert Bolton, "Definition and Scientific  Method in Aristotle's Posterior Analytics" in Philosophical Issues in Aristotle's Biology, A. Gotthelf and J. Lennox, eds. (Cambridge: Cambridge University Press,  1987). 50. Robert Bolton, "The Epistemological Basis of Aristotelian Dialectic," 212–3. 51. Cf. Irwin, Aristotle's First Principles, 350–2, 355, 358. 52. As J. S. Mill notes in Utilitaranism. 53. Cf. Descartes, Meditations 6; CSM II.56–7; VII.82–3. John Cooper makes a similar point in his review of Martha Nussbaum's The Fragility of Goodness,  Philosophical Review 97, no. 4 (1988): 543–65, viz. 553. 54. This process has been modeled in expert systems where a computer left running often ends up with strategies and solutions quite unfamiliar to its initial  programmer. 55. We might be tempted however to claim that Aristotle does not reject this account; at any rate he never says that he does, and continues to use it in Eth. Nic. 56. "[W]hy is it reasonable to prefer nonfalsified statements to falsified ones? ... from a pragmatic point of view the question does not arise, since false theories often  serve well enough ... because we search for truth (even though we can never be sure we have found it)." Karl Popper, Conjectures and Refutations (New York:  Harper Torchbooks, 1962), 58. 57. John Cleary, "Phainomena in Aristotle's Methodology," 61. Cf. On the Heavens 270b1–6. 58. Robert Bolton, "Aristotle's Method in Natural Science in Physics I," in Aristotle's Physics, L. Judson, ed. (Oxford: Oxford University Press, 1991), 1–29, viz.  21. He holds, 18–9, that dialectic does not establish scientific principles inductively. Cf. Soph. El. 170a33–9. I would say, rather, that dialectic can contribute to their  discovery as well as to their establishment. 59. "Rather, the function of dialectic must be (a) to find supporting arguments for principles already grasped inductively and (b) to provide arguments against putative  principles." C. C. W. Taylor, "Aristotle's Epistemology," in Epistemology, S. Everson ed. (Cambridge: Cambridge University Press, 1990), 116–42, viz. 133. I agree  with Robin Smith, "Aristotle on the Uses of Dialectic," Synthese 96, no. 3 (1993): 354: "... not that dialectic includes a method for establishing objective starting points  in the sciences, but that it has some uses in connection with any examination of the common principles." Unlike him, 352, I hold that dialectic plays a similar role, albeit  a smaller one, in establishing the principles peculiar to a particular science, as we have seen in physics and ethics. 60. Robin Smith, "What Use Is Aristotle's Organon?" Proceedings of the Boston Area Colloquium in Ancient Philosophy 9 (1993): 261–85. 61. Carl Hempel, "Aspects of Scientific Explanation," in Aspects of Scientific Explanation and Other Essays (New York: Free Press, 1965), 331–496, viz.  336. 62. "Empiricist Criteria of Cognitive Significance," in Aspects of Scientific

  

Page 181

       Explanation and Other Essays, 117. 63. C. D. C. Reeve, Practices of Reason (Oxford: Oxford University Press, 1992), 44–5. 64. W. V. O. Quine, "Epistemology Naturalized," in Ontological Relativity and Other Essays (New York: Columbia University Press, 1969), 67–90, viz. 83–4;  Reeve, Practices of Reason, 46; C. C. W. Taylor, "Aristotle's Epistemology," in Epistemology, S. Everson, ed. (Cambridge: Cambridge University Press, 1990),  116–42. 116 n. 3 notes the resemblance but stresses the differences. 65. Cf. Pierre Pellegrin, Aristotle's Classification of Animals, A. Preus, tr. (Berkeley: University of California Press, 1986), 50ff. 66. Cf. Mario Mignucci, " ' 

 and in effect sides with Mignucci.

67. To be sure, Aristotle does say that in the apprehension of simples there can be no mistake, as there has been no judgment (430a26–7). But first principles are  statements and hence not simples. 68. Robert Bolton, "The Epistemological Basis of Aristotelian Dialectic," 230–1. 69. Kurt Prizl, "Opinions as Appearances: Endoxa in Aristotle," Ancient Philosophy 14, no. 1 (1994): 41–50, viz. 48; Jonathan Barnes, "Aristotle and the Methods  of Ethics," 508–9.

  

Page 183

Chapter Ten—  Dialectical Communities: From the One to the Many and Back May Sim I shall argue that Aristotle's dialectic captures being. The middle term between dialectic and being is definition. Substance is the heart of being. Definitions are,  fundamentally, formulae of essences (101b37, 1030a6–1031a14), and substance, or essence, is what is most definable. Dialectic is an interrogative procedure  (155b8) that employs likeness and difference (105a20–108b36) to secure definitions (108b8, 20). In short, these three things go together: dialectic, definition, being.  The task of my chapter is to explore this nexus and reveal the essence of Aristotle's philosophical dialectic. Section I focuses on book I of the Topics to show the intimate connections between dialectic, definition, and being and hence the contribution of the Topics to a  metaphysics that claims to objective reality. Section II shows that in general, the remaining seven books of the Topics support the thesis of section I. This will reveal the unity of the Topics which is significant to  my interpretation of Aristotle's dialectic that stresses the Topics as the central text dealing with dialectic. Section III takes further the connection between dialectic, definition, and being by showing how the topics are used to generate materials for arguments and how  dialectic provides not only a road to first principles but the vehicles we can use to arrive there. This discussion introduces two senses of 'community' 1  implied by  Aristotle's handling of dialectic. One sense of community refers to particular groups of people who share opinions (endoxa) from which dialectical

  

Page 184

investigations are said to proceed and have a rhetoric based in a shared use of topics. The other sense of community refers to a philosophical community possessing  the topics that are used to generate the materials for arguments about first principles. By sharing topics for discoursing and investigating reality and its aspects, different  communities holding different opinions can argue with each other and arrive at truth. In short, I show how dialectic: (a) starts with a common opinion; then (b) applies  topoi to generate variations, and so provides materials for arguments; (c) selectively retains propositions which stand for or against the relevant four predicables; and  (d) arrives at the answer with respect to genus, property, accident, and definition. Aristotle claims that dialectic is a movement from the one to the many and from the  many to the one (Topics VIII.14 164b3–7). An opinion in step (a) is a one. Step (b) uses the topics to move from this one to many arguments or materials for  arguments. Step (c) uses the topics to narrow this set of materials, moving from the many to the few. Finally, step (d) employs topics to arrive at the definition of the  thing with which we are concerned (be this truth about being or good) along with the other predicables. Hence we arrive at a one which is the first principle of our  investigation. Section IV addresses the problem of how rhetoric relates to dialectic, elaborating the theme of dialectical communities. One might argue against my claim that there is  a larger human community that shares in the topics by saying that it is not apparent to everyone that such a community exists or what topics and principles it shares.  Another objection fastens on my claim that sharing the topics allows different people holding different opinions to argue with each other and arrive at truth. For not all  people argue with each other in the relevant ways, or arrive at truth. I acknowledge that topics are used rhetorically as utensils of persuasion and suggest that different  uses of these topics define different communities. I claim further that only a certain kind of community can recognize the larger human community that shares the  relevant topics and can recognize the metaphysical import of the philosophical dialectic they make possible. This is the philosophical community which finds persuasive  the right topics rightly understood. I—  Dialectic, Definition, and Being If I am correct about the interconnection of dialectic, definition, and being, and about Aristotle's arguing for a realistic outcome, then Aristotle's dialectical methodology  has been misunderstood by the most prominent recent commentators. To assume with G. E. L. Owen in "Tithenai ta phainomena" 2  and T. Irwin in Aristotle's First  Principles,3  that dialectic, unlike demonstration, is not intrinsically connected to reality or observable facts, is only to raise problems and doubts about ever  discovering objective truth—since demonstration relies in crucial ways on the results of dialectic. Aristotle exhibits no such doubts, and we must explain why this

  

Page 185

is so, and why Aristotle is right. Others perhaps more sympathetic with Aristotle's dialectic stress its usefulness to philosophical understanding when the knot from  which it proceeds gets untied. 4  However, the claim that Aristotle's dialectic is only useful because it leads to a lusis or unraveling of an aporia, misdirects us from the  substantive contribution of dialectic to our knowledge of truth and being, and to our correctness of choice and avoidance in actions. Aristotle argues that dialectic is  able to reach truth and correctness (104b1–3). Commentators such as Gilbert Ryle5  and John J. Cleary6  stake their tents with such a camp of thinkers who focus on  the lusis of an aporia but fail to show how Aristotle's dialectic could ever achieve the objective truth it claims to. Aristotle asserts that "reasoning ... is dialectical, if it reasons from opinions that are generally accepted" (100a30–31). Since Aristotle here contrasts dialectical  reasoning that proceeds from common opinions with demonstration which proceeds from true premises, it seems easy to suppose that dialectic in this sense cannot  attain objective first principles. Irwin, despite an initial separation of dialectic from being, tries to explain how Aristotle's dialectic might attain to objective reality.  Irwin's way out is to separate a 'strong' dialectic which he attributes to Aristotle's later works such as the Metaphysics, and a 'weak' or 'pure' dialectic which he  attributes to Aristotle's earlier works such as the Topics. He holds that pure dialectic rests entirely on common opinions and aims at coherence (Irwin, 466), whilst  strong dialectic rests on premises which we are not free to reject (at least so far as they support the doctrines of the Metaphysics). We are not free to reject the  results of strong dialectic for it rests on premises which correspond to an independent reality such that our rejection of them would lead us to lose the explanation and  understanding that rest on Aristotle's Metaphysics (Irwin, 466–467). But Irwin's attribution of 'pure dialectic' to the Topics is untenable. I will show that neither the  propositions nor the arguments of the Topics can be rejected, on pain of giving up the bases of all argument and giving up the being and truth that are achieved through  dialectical method. Irwin, like the majority of contemporary commentators, fails to recognize the intimate connection between dialectic and being. Such a failure is exhibited in Irwin's  willingness to separate the opinions from which dialectic departs from being. The upshot is that Irwin's pure dialectic begins from opinions, but fails to reach objectivity;  his strong dialectic achieves objectivity but does not really rest upon the common beliefs. The divorce of dialectic from common beliefs comes when he bases the  objectivity of strong dialectic on results of the Metaphysics. As such, the arguments in strong dialectic are no longer supported by the common beliefs (ibid.). In this  way, to the extent that objectivity is attained for Irwin, it comes at the expense of giving up a beginning from endoxa. Consequently, Irwin never quite succeeds in  convincing us that Aristotle's dialectic both turns on what is common and also achieves objective knowledge of first principles. My approach locates ta koina—what is common—in topics concerning likeness and difference, part and whole, and others that are truly common to all humanity.

  

Page 186

A proper understanding of these common topics allows for a movement that achieves objective knowledge of first principles. These topics are derived from reality and  its aspects, and the community that grasps their true standing is for Aristotle, a community that actualizes what is common to humanity and makes available to all human  beings true principles for their arguments. That Aristotle believes in the objective outcome of dialectic is apparent from Topics I.18, where he speaks about the usefulness of examining the various meanings of a  term. Such an examination ensures that "our reasoning shall be in accordance with the actual thing and not addressed merely to the term used" (108a17–22,  emphasis added). In. I.11, he speaks of the contribution of dialectic to "truth and knowledge" as well as "choice and avoidance" (104b2). Last but not least, dialectical  propositions or problems are constituted by definitions and definitions capture a thing's essence (101b38). Hence, by procuring the right problems and the right  propositions through correct definitions, the outcome of dialectical reasoning would also be objective. Another sign of Aristotle's conviction about the objectivity of dialectic in the Topics is this. While setting up the materials for dialectical arguments, he starts with the  subjects about which we reason. He says that dialectical arguments are "equal in number" to and even "identical with" the subjects about which we reason (101b13– 15). 7  If the materials of dialectical reasoning are the same as the subjects of reasoning, then dialectical arguments deal with the same reality as other sorts of  reasoning. Just as demonstrative reasoning is true or false depending on whether it captures reality8  so too there is correctness or wrongness, truth or falsity in  dialectical reasoning. The standards of truth and correctness for dialectical arguments or reasoning are the same as those for reasoning in general. Hence, despite the  fact that dialectical reasoning "reasons from opinions that are generally accepted" (100a30–31), the objectivity of its subject matter and the standards of its rationality  ensure that the opinions from which it reasons are judged either true or false, correct or incorrect, depending on whether these opinions correctly capture the pertinent  reality. Though I hold that dialectic, like demonstration, rests on reality for its truth, a key difference should be borne in mind. Whilst the true premises of demonstration are  used as they are, the probable premises of dialectic typically require modification. Opinions common to all or most are expanded by using topics such as sameness and  difference, part and whole, better and worse, etc. In short, Aristotle generates materials for dialectical reasoning by expanding the initial set of opinions so the options  we examine are not limited to (though they remain related to) the opinions from which we depart. In this way, topical dialectic is not limited in the opinions it examines  to those circulating in discussions at the time. This sort of modification is quite usual in Aristotle's own works. Moreover the topics which generate this expanded  starter set are themselves "common" in a quite strong sense.9  Demonstration proceeds from true and primary premises that are definitions and first principles which  dialectic provides. So the most basic

  

Page 187

distinction between demonstration and dialectic is that demonstration presupposes dialectic. Both are modes of reasoning, and both reach the truth. How then does dialectic capture reality? Aristotle begins answering this crucial question in Topics I.4. He holds that dialectical arguments begin with propositions 10 or  problems (101b15, 104a3–4), and reasoning reason about problems. Propositions and problems are similar in that they are formulated by four rudiments: a property,  or a definition, or a genus, or an accident in some combination. Ultimately, specific definition is the focus of these propositions and problems, and directly or indirectly,  the other three tell us about definition (viz. Topics I.6) even when a specific definition is lacking. The four rudiments of definition are extremely important because every proposition and problem tells us about a genus, or a peculiarity, or an accident.11 What is  peculiar is in turn divisible into that which tells us something's essence and hence is a definition in the strict sense, and that which does not tell us the thing's essence  though it belongs to the thing alone. This latter is called a property.12 These two rudiments, definition and a property, are among the kind of materials from which  dialectical reasoning starts. For instance, the definition of man can constitute a dialectical proposition as follows: '"An animal that walks on two feet" is the definition of  man, is it not?' (101b30) Whether such a proposition is examined dialectically then depends on whether the thesis is one that is puzzling to those who need argument  (105a2–4, viz. 104a3ff. or chapter 10) and I shall say more at the end of this section about the contribution of puzzles to dialectic in first philosophy. A genus is another one of the four rudiments which contributes to our understanding of 'what something is'; it is a predicate in the category of substance or essence.  Aristotle's example is that by predicating 'animal' of man and of ox, we are arguing that they are in the same genus. Such a proposition contributes to dialectical  arguments because if we could show that 'animal' is "the genus of the one but not of the other, we shall have argued that these things are not in the same  genus" (102b1–3) and hence shall come to know better what each is. Finally, Aristotle defines an 'accident' as "something which may possibly either belong or not belong to any one and the self­same thing, as (e.g.) the 'sitting posture'  may belong or not belong to some self­same thing" (102b4–7). An accidental predicate contributes to dialectical reasoning no less than the other three rudiments,  because the recognition that such a predicate is not necessarily predicated of a thing allows one to see the falsity in an opinion that attempts to make a universal  proposition from such an accidental predication. And plainly, knowing that an accidental predicate does not belong to something also contributes to our knowledge of  what that thing is. Knowledge of the definition of a thing is crucial, and this is a knowledge to which all four rudiments contribute. Definition is crucial because Aristotle stresses time and  again that people cannot argue dialectically unless they are directed at the same thing, and they will argue in vain unless they are directed to the actual thing (kat'  auto to pragma) (Topics I.18, 108a17–23, 105a31–33). Given that knowing

  

Page 188

whether interlocutors are directed at the same thing depends upon knowledge provided by one or more of the four rudiments, Aristotle holds these four rudiments to  be the sources of all the materials of dialectical arguments or reasoning. It is clear how the definition helps us know the essence of a thing. But the other three  rudiments, namely, the genus, property, and accident, also contribute to our grasp of the definition because they can destroy the definition if the opinions at issue  confuse them (mistaking an accident for a property, or a property for an accident, for instance). Aristotle writes, We must not fail to observe that all remarks made in criticism of a 'property' and 'genus' and 'accident' will be applicable to 'definition' as well. For when we have shown that the  attribute in question fails to belong only to the term defined, as we do also in the case of a property, or that the genus rendered in the definition is not the true genus, or that any  of the things mentioned in the phrase used does not belong, as would be remarked also in the case of an accident, we shall have demolished the definition; so that, to use the  phrase previously employed, all the points we have enumerated might in a certain sense be called 'definitory'. (Topics I.6 102b27–35)

The lesson to be gathered from Aristotle's discussion of the contribution of the four rudiments of definition to dialectical reasoning is that dialectic gets its structure from  aspects of being. Aristotle's talk of a definition which captures the essence or 'what it is' of something is based in a conviction—one well worth sharing—that there are  things with definable essences in the world. Aristotle's talk of a dialectic that can attain to such definitions is based on the conviction—also worth sharing—that the  opinions that withstand dialectical scrutiny capture the aspects of things and merely await arrangement into a structure of predication (genus and definition, property  and accident) that reflects the structure of being. For instance, Aristotle tells us that the topic of sameness and difference helps us define things and to recognize what a particular thing is (viz. 108a36–108b25). Given  that the genus captures the sameness of different kinds of objects, it too contributes to the definition. By the same token, insofar as recognizing the differentia proper  to a thing (108b6) entails recognizing that other qualities claimants propose are instead peculiar but not essential (i.e., properties) or perhaps even accidental, again the  other three rudiments directly or indirectly assist with definition and with insight into being. By examining the predicates that tell us what something is and distinguishing  such predicates from those which might seem to tell us what something is but really just capture properties or accidents, dialectic distinguishes the essentials from the  inessentials. For Aristotle, we—with our various opinions—are always with the world. The task is not the impossible one of determining whether the senses of terms  or opinions are ever able to reach the world; the task is determining the respects in which opinions that withstand scrutiny are true. Commentators who wish to read  into Aristotle's dialectic such an epistemological chore doom it, and us, to the impossible.

  

Page 189

It is Aristotle's verdict that the common or reputable opinions will have something to say about (or at least will occasion the generation of materials that have something  to say about) a given issue. The task is to determine whether rival opinions are saying the same thing, and if so, in what respect this tells us about the thing itself.  Dialectic pursues this task, and a key component in this dialectical sorting is the set of four rudiments that mark at one and the same time important considerations of  arguments and decisive aspects of things. Aristotle is right to suppose that dialectic does not just deal with words or subjective beliefs or personal constructions  divorced from the world. This is evident when time and again he tells us to get clear about the various senses a term might be used so that we are clear about the  various aspects of reality, or what he calls the "discovery of differences" ( ) (105a25). 13 Dialectical use of the topic of sameness and  difference will ensure that we are directing our minds at "the same thing" ( —108a21). The whole of Topics I.15 consists in different ways  of examining whether a term has several meanings or one. His examples here attest to the fact that if a term has several meanings, it is mostly due to its application to  various realities, so that again, the dialectician's task is one of sorting, rather than bridging some supposed gap between appearances and things. For example, 'to have  sensation' has more than one meaning. It is easy for arguments to go on interminably about the issue(s) of 'sensation', but genuine inquiry requires that we direct our  minds at "the same thing and the "actual facts'' and so we need to discover that 'to have sensation' has different meanings when applied to body and soul (106b21–25).  Or consider that 'clear' and 'obscure' mean different things when applied to sound and color (106a24–28). The examples are legion. In all these cases, the dialectical  investigation of sameness and difference is one that gives us the "differences of things" and hence gives us definitions that are objective due to the objective sameness  and difference that we can recognize once we sort out—dialectically—the proper applications of the definitions. In short, dialectic for Aristotle proceeds in the way it does, examining sameness and difference of meaning, framing issues with the four rudiments, because being itself  is so structured that there is real sameness and difference exhibited by these four aspects.14 Hence the dialectical procedure looks at the various senses of a term to  ensure that disputants use the same sense (and hence direct their minds at the same thing) and focuses on definitions which capture essences, genera, properties, and  accidents. By showing the objectivity of the materials of Aristotelian dialectic, we see then that dialectic in the Topics already proceeds from common opinions that are  for the most part based in reality in some way or other (the job is to figure out how). Moreover, the rudiments of definition and the topics themselves cannot be  rejected without giving up intelligible argument and truth, since the structure of the predicables reflects the structure of being, and since the topics are common to all. If  dialectic seeks definitions and turns on topics and predicables, and if topics and predicables always already reflect being itself, then dialectic always already

  

Page 190

reflects being itself in its seeking of definitions—especially those primary definitions that form the starting points for the demonstrations of the special sciences. Having shown the interdependence of dialectic and being through analyzing Aristotle's view of definition in the Topics, it remains for me to show that dialectic is  intrinsically philosophical or metaphysical. Recall that Aristotle's claim is that "Not every problem, nor every thesis, should be examined, but only one which might  puzzle one of those who need argument" (105a2–4, cf. 1006a6–10). Given that we could get definitions capturing the essences not only in the category of substance,  but also in the other nine categories, 15 one of the puzzles that arises concerns what is primary amongst these essences. Questions regarding the primacy of substance  or essence and questions such as "What is being?" and "Is being one or many?" arise as puzzles not only from Aristotle's own dialectic, which leads to definitions in the  ten categories, but also from Aristotle's predecessors' arguments (viz. Met. I.3–10), which support the oneness and plurality of being. Given that dialectic investigates  puzzles through investigating sameness and difference; priority and posteriority; motion and rest; the sensible and nonsensible, etc. (viz. Met. III.1 995b5–26); it is  intrinsic to this method to deal with the subject matter of being qua being, i.e., the question of first philosophy (viz. Met. IV.2). The investigation of being qua being  leads to what primary substance or essence is because it is the primary that is knowable and upon which all else depends for their names (viz. Met. 1003b16–17, cf.  1003b31–33 taken with 1004a11–28). Dialectic then, is philosophical because it proceeds from common opinions about reality, which common opinions attempt to capture the definitions of reality. Such  definitions however go on to constitute propositions or problems that are puzzling due to the convincing and yet conflicting ways in which one's predecessors have  argued about what is real. Since it is the task of metaphysics to study being qua being, i.e., what is essentially being, and dialectic provides the materials and ways to  study being, dialectic is essentially metaphysical and philosophical in the sense that it studies first philosophy. In section I, I focused on book I of the Topics to show that dialectic aims at definitions and definitions capture the essences of things in the world. Such a task was  accomplished by looking at the connectedness between the four rudiments of a property, a definition, a genus, and an accident; especially the relevance of each of  these rudiments for the discovery of the definition. At this point, it is important that I show in section II that the remaining seven books of the Topics support the  significant role played by definition and being in Aristotle's dialectic. II—  Unity of the Topics At first glance, the other seven books of the Topics support my claim regarding the focus of this treatise on definition and being. For instance, Topics II talks about

  

Page 191

accidents. Not only does this chapter deal with ways of investigating whether an accident is correctly predicated of a subject, it starts out by discussing the difficulty of  conversions with respect to accidental predications and how such conversions must be "based on the definition, and the property, and the genus" (108b13). Such a  distinction between conversions in accidental predication and those predications of definition, property, and genus supports the differences Aristotle establishes  between accidents and the other three rudiments of definition. Topics III, on the other hand, does not seem at first sight to discuss any of the four rudiments or  definition at all, and so does not seem to support my reading. Topics III deals with comparisons between two or more closely related things and the showing of the  superiority of one over the other(s) and hence its choiceworthiness. Nevertheless, this book too is intimately bound up with definition and being for it lays out the  superiority of peculiarity over commonality (117b30), the superiority of things determined through the topic of the greater and the lesser (119b16ff.), the good and the  bad (119b30ff.) applied to genera and kindred topics. These comparisons, and the standards and rules for determining which of two (or more) things is in fact more  choiceworthy, allow us to understand why a particular quality is selected as the differentia and another selected as the genus—for these are the qualities that are more  important or more choiceworthy absolutely. Most significantly, Topics III, by providing standards and rules for determining which of two (or more) things is superior,  provides the foundation for the criteria Aristotle invokes in his ethical works and his Metaphysics (I mean criteria such as self­sufficiency, permanence, and the like),  since it reveals the way we determine a hierarchy of goods. Topics IV and V explicitly support the claim that definition is the focal point of Aristotle's discussions about these four rudiments, for he asserts in these books that a  genus and a property are rudiments which relate to definitions. As Aristotle puts it in the opening sentences of Topics IV, "The next questions which we must examine  are those which relate to genus and property. These are rudiments in questions relating to definitions, but in themselves are seldom the subject of inquiries by  disputants" (120b12–15). Topics VI and VII focus on the discussion of definitions and hence also support my reading that the Topics is directed at the investigations  of being and definitions. It is noteworthy that Aristotle devotes two books to an explicit discussion of definition. This is a good indication that definition plays an  important role in dialectic as in the Topics generally. Topics VIII at first sight seems not to support my claim that dialectic is already philosophical in its search for being, since it starts by contrasting the philosopher and  the dialectician. Aristotle says, (a)s far as the choice of ground (topou) goes, the philosopher and the dialectician are making a similar inquiry, but the subsequent arrangement of material and the framing of  questions are the peculiar province of the dialectician; for such a proceeding always involves a relation with another party. (155b7ff.)

  

Page 192

The philosopher, to the contrary, is not so interested in directing his materials at another. Put this together with Aristotle's discussions of the arrangement of materials  and roles played by the questioner and the answerer (chs. 1–8), and one can see why certain interpreters claim that the Topics is simply a handbook of strategies for  the dialectician, who is to be sharply distinguished from the philosopher. 16 Nonetheless, a careful reading reveals that Topics VIII deals significantly with how  philosophers proceed, with the role played by definitions and the definability of first principles, and ultimately, with the significance of such procedures for philosophical  knowledge and wisdom. Aristotle recognizes three uses of dialectic—for the sake of mental training (pros gumnasian), for the sake of conversations or public encounters (pros tas enteuseis)  and for the sake of philosophic sciences (pros tas kata philosophian). Since there are three uses of dialectic, it is not always clear that Aristotle refers to  philosophical dialectic when he says something about 'dialectic'. Hence, with respect to Topics VIII's contrast between the philosopher and the dialectician, Aristotle  could very well be contrasting those who employ philosophical dialectic with those who employ dialectic for public encounters or for mental training. These latter two  kinds of dialectician concern themselves with interlocutors, the philosopher dialectician need not. Interpreters should take Aristotle's own advice in this treatise when  reading the treatise, and should distinguish various senses of dialectic and appreciate that he might be distinguishing the nonphilosophical uses of dialectic from its  philosophical uses. To emphasize the difference then, Aristotle in a number of places avoids talk of 'dialectic' in reference to philosophy in this context.17 Positive evidence that Aristotle deals with philosophical dialectic in this book, whose first eight chapters concern the roles played by the good questioner and  answerer, is present in chapter five. Aristotle is clear when he speaks of how there are no rules for those who use dialectic for the sake of experiment (peiras) and  speculation or inquiry (skepseôs) whilst there are rules for those who discuss for the sake of didactic and competitions (159a25ff.). He continues by saying, "since  then, we have no traditions handed down by others, let us try to say something ourselves on the subject" (159a36–37). Thus he proceeds to provide rules for being a  good answerer in a dialectic that aims at experiment and inquiry which is—according to Aristotle's distinction—philosophical. The Stagirite, in talking about training in  dialectic in the last chapter of Topics VIII, emphasizes the importance of examining the arguments both for and against when dealing with any thesis. This pertains to  all three sorts of dialectician. The philosopher, however, once both sorts of arguments have been laid bare, must seek the solution immediately. Thus he says, "If we  have no one else with whom to argue, we must do so with ourselves" (163a3). This is Aristotle's own dialectical procedure—just consider Metaphysics Z.13 for a  stark example of arguing against oneself or playing one's own opponent.18

  

Page 193

All these points are consistent with Aristotle's talk about the usefulness of dialectic for philosophical sciences in book I.2 when he says, "(f)or the philosophic sciences  it is useful, because if we are able to raise difficulties on both sides, we shall more easily discern both truth and falsehood on every point" (101a35–36). Such a remark  also reminds us of Metaphysics B.1 when he spoke of the importance of knowing the knot and hearing both sides in order to be a good judge and one who is  prepared for the solution (995a24–995b4). And this is why Aristotle spends most of the Topics talking about ways of coming up with constructive and destructive  arguments. It is significant that Aristotle talks about both constructive and destructive arguments because it is only when one could argue for both sides that the mind  could become perplexed and thus more "easily discern both truth and falsehood on every point" (101a36). Again, it is for the discovery of truth, rather than for the  sake of just winning an argument, that Aristotle has in mind in coming up with these materials for arguments 19 (viz. Topics II). This aim distinguishes philosophical  dialectic from the other two. For these reasons I do not find the Topics to be simply a handbook of strategies, pure forms without content, irrelevant to truth or the search after first principles.  Rather, the Topics provides us with ways of coming up with arguments which make truth accessible to us. It provides a road to first principles because it teaches us to  use topoi to arrive at definitions. That definitions play the leading role in philosophical pursuits is evident when he says "(m)oreover, you should have a good supply of  definitions ( ) and have those of familiar and primary ideas ready to hand; for it is by means of these that reasonings are carried  on" (163b20–23). That Aristotle stresses the importance of amassing "an abundance of material" (163a5) for arguments is mandatory, since it is in this material that the  problems are found and in regard to the opposing arguments that choices must be made. That dialectical training is essential to the philosopher in this task is evident  when Aristotle says, Also to take and to have taken in at a glance the results of each of two hypotheses is no mean instrument for the cult of knowledge and philosophic wisdom; for then it only  remains to make a correct choice of one of them. For such a process one must possess a certain natural ability, and real natural ability consists in being able correctly to choose  the true and avoid the false. (VIII.15 163b9ff.)20

Even more striking than this directedness of dialectic toward truth and being is the importance of first principles and how dialectic can arrive at first things via  definitions. For the relation between definitions and first principles, see VIII.3 (158a30–159a14). Let me cite just one passage from chapter three regarding definitions  and first principles to show their place in dialectical arguments: The same hypotheses may be both difficult to attack and easy to defend. Both things which are by nature primary and things which are by nature ultimate are of this kind. For  things which are primary require definition and things which are ultimate are reached by many stages if one wishes to establish a continuous train of proof

  

Page 194 from primary principles, or else the arguments have the appearance of being sophistical; for it is impossible to demonstrate anything without starting from the appropriate first  principles and keeping up a connected argument until ultimate first principles are reached. Now those who are being questioned do not want to give definitions nor do they take  any notice if the questioner gives them; and yet it is difficult to argue if what is proposed is not made clear. This kind of thing is most likely to happen in the matter of first  principles; for, whereas it is through them that everything else is made clear, they cannot be made clear through anything else, but everything of that kind must be made known  by definition. (158a30–158b4 my italics)

It is clear from this passage that arguments regarding things that are primary and ultimate require first principles which are known by definitions and hence we need to  proceed from definitions. But stress on the universality of definition comes at the end where he talks about the importance of first principles for making everything  clear. Since it is the task of first philosophy to get at first principles, again, we see the need for a dialectic that is philosophical, that gets at these first principles via  definitions. 21 Dialectic provides not only a road to first principles but the vehicles for arriving there. Topics are those vehicles. Topics are not simply forms without content which  could only produce reasons consistent with common opinions or with appearances that need saving. Some topics such as likeness and difference, greater and lesser,  seem really empty. However, others involve assertions; for instance, permanence is better than transience, ends are better than means, etc. These are not empty  because they shaped our theoretical and practical choices. Dialectic (a) starts with common opinions, then (b) applies topoi to generate variants of these opinions and provide puzzles and materials for arguments in the  discrepancies between these opinions and variants,22 (c) selectively retains propositions for and against the four rudiments of definition, and (d) arrives at definitions  with respect to these four rudiments. The distinction between genus, property, accident, and definition allows a dialectician to save a severalty of assertions, arranged  in a way appropriate to the subject matter. Aristotle claims that dialectic is a movement from the one to the many and from the many to the one (VIII.14 164b3–7).  Steps (a) and (b) use the topics to move from one issue to many opinions. Step (c) moves from the many to the few and (d) to the one best answer for each sort of  predicable. In the first movement, dialectic generates possibilities; in the second, we move from possibilities to actuality to arrive at the definition of the thing with which  we are concerned. III—  Two Communities in Aristotle's Dialectic One problem with the ability of dialectic to arrive at truth and first principles lies in its starting point. Since dialectic begins from available opinions, what

  

Page 195

guarantee is there that viable candidates will be numbered among those available opinions? Another problem concerns the end of dialectic; even supposing the truth is  numbered among the options, how does one grasp the objective first principles and define them? My reading of Aristotle solves both problems by showing how the  topics allow dialectic to move from (a) extant opinions to (b) a fuller range of possible answers to a question, and then (c) from these many possible answers to the  one true answer on a given problem. This is Aristotle's version of the movement of Platonic dialectic from the one to the many and from the many to the one. My reading of dialectic's movement from the one to the many and from the many to the one solves the two problems of the limited character of the starting points of  dialectic and the end of objective truth about first principles. Most contemporary interpretations of Aristotle's dialectic suppose that it starts with common opinions and  remains with these opinions, ending in coherence at best. My reading shifts the emphasis to the topics which are responsible both for the generation of a more  adequate set of possibilities from these initial opinions and for the movement of dialectic from the many possibilities to the one true definition about the thing at stake. Where do these topics come from and how do they work? I will show that the topics are principles that are common to every art and science according to Aristotle.  They can perform the task of getting at objective first principles and definitions because, according to Aristotle, they reflect structures of being. They reflect structures  of being because they are derived from reality or its aspects. Because these topics are derived from reality and its aspect, they are accessible to everyone. There are,  in the real, sameness and difference, means and ends, parts and wholes, and greater and lesser. Every community of persuasion uses some subset of these topics, somehow understood. The philosophical community is one community of persuasion among others.  This philosophical community for which Aristotle speaks in his Topics (as elsewhere), and which he aims to further, recognizes that there is an important group of  topics that are common because they reflect structures of being. Different human communities, plainly enough, may understand these topics differently. Similarly,  topical assertions (such as 'ends are better than means') may vary across communities of conviction. It is Aristotle's view that there is an objective fact of the matter  about which construal is correct. This account has two striking consequences. First, according to members of the philosophical community, so far as any community of persuasion employs topics, it is  in touch with being. That remains true even when it fails to be aware of this due to its own misunderstanding of the metaphysical import of its discursive utensils.  Second, so far as all communities of persuasion employ some topics as means of discourse and argument, and since there is a single group of topics at least potentially  common to all, then there are grounds for all human communities to discourse and argue with each other. That follows even though some—perhaps

  

Page 196

many—of these communities may fail to be persuaded by Aristotle's account and may believe there are no common grounds for discourse at all. Third, and most  subtly, the same topics on which differences and disagreements pivot are—rightly understood—the keys to the achievement of knowledge about first principles. While  rhetoric and dialectic intersect in the topics, persuasion and truth remain quite different things. For this reason, the philosophical community for which Aristotle speaks  appears—to other groups—as merely one among many communities of persuasion, and Aristotle's account of topical dialectic may indeed fail to convince. The  philosophical community distinguishes itself from other communities of persuasion in that what it finds persuasive is the truth. By starting with one opinion, we can move to the others or combinations of the others, hence expanding the opinion from which we started. Such a process can be  accomplished by using the topic of part and whole that all human beings share in virtue of belonging to some topical community. For example, one can see that the  whole soul is divisible into the two rational parts and the two nonrational parts. These various parts of the soul also give us various candidates for the good life, e.g.,  pleasure, politics, or contemplation or any combination of two or three. Having expanded the possibilities of happiness, one can eliminate some of these possibilities.  One can do so by using the topic of sameness and difference along with those of essence and accident and peculiarity and commonality to determine what is essential  or unique to the human soul when compared to plants and animals. Having first expanded the possibilities of happiness and then eliminated certain of these possibilities,  one is then left with a narrower set and will ultimately arrive at the one definition of what is the good life. 23 Throughout Aristotle's works, the procedure is always one  which starts from endoxa, expands the possible candidates for the problem at hand, and finally eliminates the possibilities that fail to fit the problem so that one is left  with a positive answer. Aristotle is insufficiently clear about the standing of these common principles in the Topics. More useful is his discussion of peirastic arguments in the De Sophisticis  Elenchis (SE from here on). The SE is a counterpart to the Topics because topics are open to sophistical abuse and so we need to get clear about sophistical  refutation and reasoning. I have in mind especially 165a25–29 (cf. 170a39–b5), where Aristotle asserts that knowledge of the sources of proofs implies knowledge of  the sources of the refutations. Knowing these refutations also implies knowing the solutions, for the objections to these refutations are at least intermediate solutions  regarding the problem at hand. To unravel this Aristotelian claim requires first a quick look at Aristotle's discussion of peirastic arguments in the Topics. At Topics VIII.5, Aristotle contrasts  peirastic, or arguments for the sake of experiment (peiras) or inquiry (skepseôs) with arguments for the sake of teaching and competition. Whereas there are already  rules for the latter two kinds of arguments, none exists for the former. Aristotle then sets himself the task of coming up with some rules for peirastic. In book VIII,  peirastic arguments are for the purpose of establishing truth and

  

Page 197

falsehood (161a26), unlike competition arguments where the parties argue against each other without a shared goal. Those who engage in peirastic share a common  purpose (161a38), so that those who are contentious in peirastic are bad dialecticians. 24 The SE includes Aristotle's detailed discussion of features that (philosophical) dialectic and peirastic (dialectic) employ in common. In SE II, Aristotle states again the  four kinds of arguments: didactic, dialectical, examination, and contentious. Peirastikoi are arguments "based on opinions held by the answerer and necessarily known  to one who claims knowledge of the subject involved"25 (165b4–6, cf. 172a30–32). Again, at SE XI, Aristotle writes, Dialectic is at the same time an art of examination . . . . For even a man without knowledge of the subject can examine another who is without knowledge, if the latter makes  concessions based not on what he knows nor on the special principles of the subject but on the consequential facts, which are such that, though to know them does not prevent  him from being ignorant of the art in question, yet not to know them necessarily involves ignorance of it. (172a21–27)

What we learn from this passage is that peirastic arguments are employable by anyone to establish another's ignorance of an art or science. So what is it that allows  everyone to test others who claim to knowledge of a certain art, even though one is ignorant of the art at stake? In SE IX, Aristotle tells us that it is the task of the philosophical or peirastic dialectician to scrutinize refutation and examination arguments by employing common  principles (ta koina) instead of those that are peculiar to an art. This is because it is not possible for the dialectician to have scientific knowledge of everything (and  because it is possible that the principles peculiar to different sciences are infinite). So, Aristotle says, "It is clear, then, that we need not grasp the commonplaces  (topous) of all refutations but only those which concern dialectic; for these are common to every art and faculty." The topics which dialectic employs are common not  only to every art or science but also to everyone.26 The commonality of these principles is not a universality where everything falls under the same genus, for there is  no such genus. Nor does dialectic have a particular subject matter, for then it would not be applicable to everything. It turns out that these principles of dialectic are  common for they are the sources (ex ôn, 170b) and principles (archai) of all arguments. They are numbered among the principles anyone needs to know in order to  argue at all. This becomes clearer when Aristotle speaks of dialectic as an art of examination at SE XI 172a21–35. Here again he contrasts dialectic and peirastic  with scientific knowledge; again, he talks about the common principles that everyone, including the unscientific person, uses to test those who profess knowledge. He  says, "Now this is where the common principles come in; for they know these of themselves just as well as the scientists, even though their expression of them seems  to be very inaccurate" (172a33–34). The topics are these common principles. Hence, while

  

Page 198

they may be known more or less well, or used more or less well, the topics are bases of all arguments; they are common to all and used by all even though few can  identify them and fewer still can appreciate their true metaphysical stature. This gives an added meaning to Aristotle's definition of dialectical reasoning as the sort of reasoning which proceeds from opinions held by all, or the majority, or the  wise. In a nutshell, there is a dual sense in which dialectic reasons from what is common because there is a dual sense of 'common'. One sense of 'common' refers to  shared viewpoints about certain subject matters, e.g., that pleasure is the highest good or all knowledge is knowledge of opposites. The other sense refers to the  common principles which everyone, scientific and non­scientific alike, use in their examination arguments. These are the topics on which all discourse and argument  turn, and which the properly philosophical community aims to bring to light. How these common principles of dialectic are employed is clarified in Aristotle's discussion of peirastic. Recall that the claim at SE XI is Dialectic is at the same time an art of examination . . . . For even a man without knowledge of the subject can examine another who is without knowledge, if the latter makes  concessions based not on what he knows nor on the special principles of the subject but on the consequential facts, 27 which are such that, though to know them does not  prevent him from being ignorant of the art in question, yet not to know them necessarily involves ignorance of it. (172a23–27, my italics)

Let us look more closely at what Aristotle calls "consequential facts," for these show one way common principles come into play in peirastic arguments. Aristotle's  point here is that when one proclaims knowledge of a subject, certain consequences follow which make up necessary though not sufficient conditions for knowledge of  that subject. Hence if one finds that the person proclaiming knowledge of a subject is ignorant of such consequential facts, then he is ignorant of the subject. On the  other hand, successful assertion of some consequential fact need not imply knowledge of the subject. One consequence of knowing some subject is knowledge of the  relevant propria; if one fails to display this knowledge, one is exposed as falsely claiming to expertise in the subject. However, again, knowledge of the propria  provides no guarantee of expertise. One also does not know a subject if, when purporting to give a property of that subject, one ends up giving us a term that is  universally applicable. This is because a fact consequential to giving a property is that one gives a term which distinguishes the subject from other things (viz. 130b12– 23). Another "consequential fact" stems from the point that definition is convertible with the subject defined. So if one asserts the definition of a species as "pedestrian  biped animal four cubits high," given that ''four cubits high" does not belong to everything falling under this species, then the claimant must not really know this species.  This is because a fact consequential to knowing a species is that one can come up with a definition that belongs to everything which falls under that species (viz. Topics  140b15–27). This test of a

  

Page 199

claimant's knowledge relies on the topic that things falling under the same species have the same definition, hence our expectation that one with knowledge of a species  must possess this knowledge. Aristotle's point is straightforward. 28 Consider a couple of practical analogues. If a cabbie claims to know the street grid of New York City, then she should be able  to direct me to Times Square. If her instructions send me walking in the wrong direction, plainly she does not know what she claimed to know. On the other hand, the  goodness of that one instruction does not suffice to prove her original claim. Or, if a man in a white coat claims to be a physician and recommends an antibiotic for  strep throat, perhaps even cautions against the pertinent side effects, this does not suffice to show he really is a trained physician. However, one unable to make the  recommendation or know the side effects would plainly show up as no physician no matter how white the frock. In each of these cases, knowing the consequential facts is a necessary but not sufficient condition for knowledge, since one can provide the consequential facts and  remain ignorant of the proper facts essential to the subject. We can perform this sort of examination even without knowledge of the subject proper, aided by topics  such as the more or less comprehensible, sameness and difference, a subject's relation to its dependents or consequents, and the relation between genus and species.  The point of the test is to make propositions couched in terms of predicables (property, genus, and definition in the above examples—accidents are handled similarly  in the Topics). These predicables are framed with the help of the topics. Aristotle believes that predicables, like the topics, are common and neutral and so can in  justice be used to outline the consequential facts one may expect the claimant to know. Hence an inability to distinguish, for instance, between peculiar29 and essential  features suffices to show that the claimant does not know what he claims to know. Our ability to expose pretenders to knowledge contributes to our knowledge. One can err not only about some subject, but also about the common topics and predicables themselves; the one who knows what they are and how to use them in  pursuit of truth is the (philosophical) dialectician. The dialectician, in general, is one who can apply what is common to whatever matters are at hand. Aristotle writes,  "The man, then who views ta koina kata to pragma is a dialectician, while he who only apparently does this is a sophist" (viz. 171b7–8). The interesting point about  how these common principles function in such contexts is that they are not only used to construct arguments but also to examine knowledge claims in ways that  provide some knowledge of the subject. Moreover, according to Aristotle, the topics are known to virtually everyone, scientist and nonscientist alike—despite the fact  that the nonscientist's expressions of these common principles are imprecise (viz. 172a33–34).30 Given that the common principles used to test one's knowledge are  known to most everyone, it is not surprising that Aristotle never stops to give a list, or to explain what topics are; rather he simply illustrates their use in arguments.  Aristotle's illustrations serve more as reminders, or helps regarding

  

Page 200

what we are already supposed to know. These helps or reminders assist our doing better what we already do when we test people who profess to know something.  31

Having seen that common topics function in the peirastic branch of dialectic to help in the search after truth, let us consider how such testing exemplifies the four­stage  dialectical movement from the one to the many and then from the many back to the one. Notice that peirastic testing starts with someone's claim to knowledge about  a subject. The claimant might profess to be an expert or a wise man, or he might simply be claiming to a bit of common knowledge. Take this knowledge­claim as  stage one. Stage two would focus on one of the claims such as 'S is P' (where 'P' refers to a property in Aristotle's sense). From this one claim, one could generate  many materials for examination by using topics such as sameness and difference, the more or less comprehensible, and so on. For instance, one could object that the  proposed 'P' is less comprehensible than 'S' and so unable to illuminate it.32 One might also object to the claim by pointing out that 'P' is not unique to 'S' because 'R'  is 'P' too (and hence 'P' must not be a property but rather an accident if it is true of the subject at all). This second stage could go on for some time. The importance of the first two phases for the contemporary discussion about Aristotle's dialectic is that it shows the  dialectician is not limited to the opinions from which he starts, even though the constructed variations remain related to those opinions. The third stage of dialectic  proceeds by applying topics to the numerous materials provided by stage two so that those not needed for the investigation at hand are eliminated from consideration.  Supposing we are in search of the definition of a subject, then propositions that refer to other predicables should be left out of consideration. (Of course, sorting out  which is which is a task at which dialectic excels.) For instance, we might retain and even focus on features of the subject that are so close to the definition they might  be mistaken for it. We might distinguish peculiarities from accidents but then confuse properties with the definition; a fresh application of the common principles should  help. The fourth stage consists in picking out the proposition that corresponds to the definition sought. This proposition may or may not have been part of the original  set of opinions approached for examination (a simple and obvious point inexplicable by the great majority of interpreters of Aristotle's dialectic). True to Aristotle's  claim that dialectic applies ta koina kata to pragma (171b7), it is the latter that act as the standard for measuring what fits and what does not. Only a genuine  definition will satisfy the philosophical dialectician, a definition distinguished from genus and properties and accidents with the use of topics functioning as common  principles. So the truth at which the dialectician arrives in this four­stage method goes beyond its starting point in common opinions in two ways. First, it generates  more materials for examination than what it begins with, and second, it terminates not in perception or opinion (even coherent opinion) but in the truth about the subject  under investigation. That, at any rate, is Aristotle's view.

  

Page 201

IV—  Rhetoric and Dialectic Up until this point, I have stressed Aristotle's philosophical dialectic as a method for achieving definition and first principles. Why, however, if this method can arrive at  the first principles of various sciences, are these first principles not more readily available? Put otherwise, how is it possible that while most everyone has access to the  topics that Aristotle portrays as so helpful in arriving at first principles, they so far have not been able to arrive at principles? Perhaps the topics are not so universal as  Aristotle claims, or not so helpful in the pursuit of truth, or both. A second objection concerns persuasion. Why have those who know the first principles not been able to use the topics to persuade others to agree with them? If  topics are used in persuasion and conversion as well as examination and truth finding, why—presuming a truth be known—does that truth so often fail to persuade? It  seems that if my interpretation of Aristotle's view is correct, one should be able to persuade an opponent, since the topics are common. To begin, Aristotle's enumeration of three uses of dialectic suggests that dialectic can be used for the sake of mental training, or rhetorical persuasion, or philosophical  inquiry. These goals need not be served simultaneously in any particular exercise of dialectic. Just because philosophical dialectic can achieve definition of being and  first principles does not mean that it need at the same time persuade others of the truth. The most direct argument against those who reduce Aristotle's dialectic to  rhetoric is that they fail to recognize the other uses of dialectic that Aristotle himself enumerates. A similarly direct argument resists the reduction of persuasive to  philosophical dialectic. I would be guilty of the same were I to reduce rhetoric and mental training to philosophical dialectic. Truth need not be persuasive. The topics  that turn on true principles might not be the topics that a given group finds convincing. Some of the topics Aristotle catalogues—such as 'sameness and difference' are quite organic, and what they gain in scope of application they lose in determinacy and  critical bite. Other topics are more specific, inclining the arguments in which they appear to more definite conclusions. Examples of specific more decisive topics are  that wholes are prior to parts, that things permanent are better than things transient, that things of the soul are better than things of the body. As a matter of fact—and it  is a fact that Aristotle consistently recognizes across his writings—there are many who do not recognize these truths or find persuasive the arguments that turn on them.  These people, according to Aristotle are simply not well educated. Aristotle realizes that not everyone is sold on the priority of whole to part, or form to matter, and so  forth. Such topics according to Aristotle are no less 'common' for this lack of recognition. Those who fail to recognize their commonality or fail to find them persuasive  are just mistaken, Aristotle would say. That is why he considers decisive the opinions of the wise rather than those of children or the many, and why he does not try to  argue with those who lack

  

Page 202

sufficient experience. Despite the fact that a dialectical investigation, say of happiness, begins from endoxa, Aristotle does not think that a young person, or a person  who is mature in years but not in character, is a qualified student of ethics. He demands that a student of ethics have experience of the actions about which political  science argues, and that he be led not by his feelings but by reason (viz. NE 1.3 1095a3–12, cf. Topics 105a2–4 cited above and Met. 1006a6–10). Reasonableness  is in part defined by a judicious acceptance of the right topics, and so Aristotle demands that the student must have had a good upbringing in order to have the proper  archê of thinking and judgment. Such an origin takes the form of believing that something is true even though we do not have the reason why it is true (NE 1.4  1095b4–13). Again, Aristotle maintains the significance of having the right origins as follows: Just as in other cases everyone goes in search with something in hand, we must so conduct our search that we try to arrive at what is said truly and clearly through things said  truly but not clearly. At the moment we are placed as we should be if we knew that health was the best disposition of the body and that Coriscus was the swarthiest person in the  market­place; we do not know what either of these things is, but it is helpful, in order to know what each of them is, to be so placed.

It seems that we need certain sorts of presuppositions in order to proceed with dialectical investigations. Key among the presuppositions are topics tied to the way the  world is. Common topics are pervasive but variable. What we need are the right topics specified in the right ways. It turns out that good education and upbringing are  needed for finding the right topics persuasive (e.g., that means are for the sake of ends and not the reverse) and for recognizing the link between topics and being.  Aristotle's recognition of this point is already on the way to admitting the bond between dialectic and being. Aristotle distinguishes between dialectic used in persuasion and used in seeking truth. Once we find the heart of dialectic in the topics, this distinction appears as a  difference between different sorts of community. That is because all persuasive discourse and argument uses some topics or other, but people will differ in the topics  they find convincing. Those who find the truly common topics to be convincing are those Aristotle calls the 'wise'. Most specifically of all are those who are persuaded  by the right topics and who acknowledge the bond between those topics and principles of being. Call this group which recognizes the link between common topics  and being the community of philosophical dialecticians. This is also a community of persuasion since its members not only use topics in the search for truth, but also in  the effort to convince others of their truth. That dialectic may fail in this second, rhetorical task, does not at all imply that it must fail in the first, more philosophical task.  The reverse is also true. And if rhetorical failure occurs between two communities of persuasion, the proper diagnosis is to be sought in the divergent topics that define  these two communities' different means of persuasion.

  

Page 203

One can be mistaken if one considers an issue by using a too narrow selection of topics or the wrong ones. For example, in thinking about happiness, one could arrive  at pleasure as the best candidate if one does not take into account the topics of completeness and self­sufficiency, which proclaim that completeness is better than  incompleteness and self­sufficiency better than dependence. Such errors are something like the failure of persuasion between communities except that they occur  within a single community. Implied but not expressed in Aristotle's account of philosophical dialectic is a distinction between two types of community. Communities of persuasion and the  community of philosophical dialecticians both use topics, but in different ways. Given the inevitability of topics in every discussion and argument—in cases of rhetorical  failure no less than successful persuasion—there is a third sense of community implied. This is the universal human community, demarcated by its use of common  topics unavailable to other animals and unnecessary to the god. Since it is the philosophical community that recognizes this specifically human community, it is the  philosophical dialectician who represents the human community, who acknowledges it and speaks on its behalf. The dialectical community is not, however, identical  with the human community, any more than it is reducible to one community of persuasion among others (despite the fact that this is the way it appears to those others,  in public). Despite the fact that the universality of the human community is based on the commonality of the topics, one's upbringing can interfere with one's ability to  recognize the topics that are available to most all human beings. That means some particular actual community will speak for the universally human community while  others will not. (Others may not even admit the possibility of such a universal community or the possibility that some group of Aristotelians might represent it.)  Particular communities of opinion are sources of errors in dialectical investigations as well as failures of persuasion. Errors in search for truth may originate in a  particular community's failure to educate so that its members are inculcated with the wrong kinds of topics. In cases of persuasion failure as well as error in the search  for truth, conversion or reeducation remain ways of removing those obstacles to correct dialectical reasoning that come in ignorance of common topics or acceptance  of the wrong topics. Right use of the right topics, according to Aristotle, is the key to the achievement of definition of being and first principles, which achievement is the ultimate goal of  philosophical dialectic as presented in the Topics. Aristotle consistently recognizes that even though certain things are by nature—e.g., we aim at being intellectually  and morally good by nature, and we make up a universal community for searching for truth by sharing the topics which aid in our search—interference could prevent  our reaching the natural end (Physics 199b15–18). Consequently, though there is a sense in which every person belongs to the universal human community in virtue of  being human at all, and has in principle access to the topics that make possible the search for truth, particular persons are characteristically members of actual,  distinctive communities of persuasion—

  

Page 204

trained in some limited use of topics. Among these communities of persuasion (but not reducibly) is the community of philosophical dialecticians. Its members find  persuasive the right topics used in the right ways. Aristotle believed that unless one were already living in the proper community, with access to the proper first  principles, the proper use of the topics would not be available and so dialectic would decay into some 'merely' rhetorical use of these topics. The complement of this  Aristotelian conviction is the thesis that a particular community can function not only as a community of persuasion, but also at the same time as a representative of the  universal human community. Such a universal human community makes possible the search for truth and first principles and makes possible their definitions; a thesis  agreement with which presupposes that one is "well­educated," that one accept the pertinent topics. 33 Notes 1. 'Community' is not a term taken from Aristotle's Politics or NE. However, my use of this term in this context is quite consistent with Aristotle's views generally. I  argue for this in sections III and IV. 2. G. E. L. Owen, "Tithenai ta phainomena," in Logic, Science and Dialectic, M. Nussbaum, ed. (Ithaca, N.Y.: Cornell University Press, 1986), 239–251. Viz. R.  Smith's essay in this collection for detailed exposition and criticism of Owen's view. 3. T. Irwin, Aristotle's First Principles (Oxford: Clarendon Press, 1988). 4. See most notably, J. J. Cleary, "Working Through Puzzles with Aristotle," The Journal of Neoplatonic Studies 1, no. 2 (1993), reprinted in The Crossroads of  Norm and Nature: Essays on Aristotle's Ethics and Metaphysics, M. Sim, ed. (Lanham, Md.: Rowman and Littlefield, 1995). 5. Gilbert Ryle, "Dialectic in the Academy," in New Essays on Plato and Aristotle, R. Bambrough, ed. (New York: The Humanities Press, 1965). 6. John J. Cleary, "Working Through Puzzles." I agree with Cleary's recognition that Aristotle's account of dialectic in his Metaphysics corresponds to his account of it  in the Topics. However, our accounts differ in the following ways. Cleary stresses the dialectical procedure. See his four stages for an adequate dialectical proof and  how the solution of the aporia preserves the truth of the most reputable opinions. I stress the metaphysical foundation of the dialectical procedure and such a  foundation of the common opinions—i.e., the intimate relation between dialectic and being. On my view, the concern of dialectic is not simply with the preservation of  the most reputable opinions. 7. By "reasoning" Aristotle means "an argument in which, certain things being laid down, something other than these necessarily comes about through them" (100a25– 27). He then divides reasoning into demonstration and dialectic. 8. See Nicomachean Ethics VI.1 where reasoning is divided into speculative and deliberative. The former is true or false whilst the latter is correct or incorrect. 9. How dialectic expands the materials for arguments by using select topics, and how these topics are common to humanity will be explained in section III. 10. Aristotle says, "For arguments start with 'propositions', while the subjects on which

  

Page 205

      reasoning take place are 'problems' (101b15–16). Even though Aristotle distinguishes propositions from problems, the distinction is not relevant to the aspect of  dialectic I am discussing because I am only dealing with the materials of dialectic and Aristotle includes both propositions and problems as the materials of dialectical  inquiry (viz. 104a3–4). The difference between propositions and problems is simply a turn of phrase—the former proposes a thesis and immediately asks for assent or  dissent whilst the latter incorporates the thesis into the question. Aristotle's example of a proposition is as follows: '"An animal that walks on two feet" is the definition  of man, is it not?' (Notice the assertion followed immediately by the question here). His example of a problem about the same subject is: 'Is "an animal that walks on  two feet" a definition of man or no?' (Notice the incorporation of the assertion into the question from the beginning here.) 11. In this way, the four rudiments contribute to our understanding of 'what something is', be it substance or essence, or any of the other nine categories (quantity,  quality, relation, place, time, position, state, activity, and passivity). See Topics 1.9 103b23–24. 12. Both definitions and properties are predicated convertibly of the thing (viz. Posterior Analytics 1.22 83a23–24, Topics 102a18–20). An example of the  convertibility of the subject and predicate of a definition is this: a man is a rational animal and a rational animal is a man. An example of the convertibility of the subject  and predicate of a property is, "if A be a man, then he is capable of learning grammar, and if he be capable of learning grammar, he is a man" (102a20–21). 13. That Aristotle is talking about the differences of things instead of the different senses of an expression here is plausible because of two points. First, he has  already talked about distinguishing the different senses of an expression in the previous item. Second, he immediately followed this discussion of the discovery of  differences and the investigation of similarities by talking about how it is possible to make propositions out of these items, i.e., it is possible to make propositions that  capture the similarities and differences of things themselves. He offers the following as an example of a proposition which captures the difference between sensation  and knowledge: "Sensation differs from knowledge, because it is possible to recover the latter when one has lost it but not the former" (105a29–30). This shows that  Aristotle is talking about the difference between actual things (or in this case, actual faculties), such as sensation and knowledge. 14. Viz. Topics 1.7 where Aristotle discussed the various senses of 'sameness' which senses correspond to the structure of being, i.e., being that is categorized by  predicating of it its genus and species, etc. (cf. VI.4, VI.7 146a3–13, 21–33, VI.13–14, 150a15–22, VII.1 151b28–152b35). 15. Naturally, these would be definitions of the forms in the relevant nonsubstance categories, not essences of the thing. 16. Eleonore Stump, for instance, sees a topic in the Topics as a strategy for arguing so that the Topics is simply "a handbook on how to succeed at playing  Socrates" (173). So she claims that teaching one to be good at dialectical disputation is the technê that this handbook wants to teach. See Eleonore Stump,  Boethius's De topicis differentiis (Ithaca, N.Y.: Cornell University Press, 1978): 159–178.       J. D. G. Evans also believes that dialectic and ontology have to be sharply distinguished for Aristotle. See J. D. G. Evans, Aristotle's Concept of Dialectic  (Cambridge: Cambridge University Press, 1977), 36. M. Husain's chapter in this collection also pursues a radical separation between dialectic and metaphysics. Some  of her arguments for Aristotle's deprecation of dialectic are similar to J. D. G. Evans's in that they both invoke passages such

  

Page 206

      as 1061b4–11 (where Aristotle criticizes dialectic for not recognizing the primacy of being), and 1078b23ff. (where Aristotle is talking about Socrates's lack of  proficiency in dialectic). See n. 21 for my disagreements with Evans's handling of these passages which are also applicable to Husain's arguments.       Barnes also puts down Aristotle's dialectic by saying that his method of endoxa "has, in the last analysis, very little content." J. Barnes, "Aristotle and the Methods  of Ethics," Rev. Int. de Phil.34 (1981): 510. E. Weil in "The Place of Logic in Aristotle's Thought," in Articles on Aristotle, vol. 1, Barnes, Schofield, and Sorabji,  eds. (London: Duckworth, 1975): 88–112, surprisingly, also agrees with the lack of content in the Topics. Weil says, ''It goes without saying that even topics contains  no ultimate criterion of truth. Topics too is purely formal, in the sense that it applies its procedures indiscriminately to any affirmation: truth depends on immediate  intuition, either perceptual or intellectual" (94). Again, Weil interprets the place of the Topics in philosophy as "a procedure for discovering the problems—not the  solutions—which present themselves to the philosopher in the course of his daily life" (93). The reason I find such remarks surprising for Weil is that he seems to  recognize the value of the Topics for getting at truth, or extracting truth from common opinions (viz. 97, 99, 103, 107). Most insightful is his suggestion that one ought  to study the relations between topics and ontology. He says, "Now it is clear that topics and ontology are simply two aspects of one reality: Aristotle says as much  himself in one remarkable chapter (Top. IX.9 170a2ff.); and this is corroborated by the part played in both disciplines by such fundamental notions as substance,  accident, property, genus, and definition" (108).       R. Smith, "Dialectic and Method in Aristotle," in this collection, also argues that the dialectician's role is that of persuasion rather than discovering truth. In view of  this goal, Smith interprets endoxa as a categorization of opinions that people will usually accept. Since the dialectician's goal is to persuade, her task or art is to arrive  at a list of opinions that can function as premises that people will usually accept so that she can use these opinions to persuade them. I agree with Smith's arguments to  the extent that they support the role of dialectic for public encounters. However, to the extent that Smith reduces the other two functions of dialectic to persuasion, I  disagree. For details of my disagreement with Smith's dismissal of philosophical dialectic in the Topics, see n. 20. In general, I think that Smith does a good job  showing that the "proof texts for certain critical theses" developed by Owen and modified by Irwin, of how dialectic proceeds (40), e.g., passages in NE VII and  Physics IV, do not completely support the view that dialectic argues from endoxa to first principles.       Wes DeMarco, "Plato's Ghost: Consequences of Aristotelian Dialectic" in The Crossroads of Norm and Nature: Essays on Aristotle's Ethics and  Metaphysics, M. Sim, ed., argues well that Aristotle's dialectic is better understood by examining his practice of moving to first principles in the Ethics and  Metaphysics, rather than by looking at what Aristotle says about dialectic in various texts. DeMarco's view that a small number of orienting moves suffices to  determine the large­scale structure of the NE and Metaphysics is like my view that different philosophical positions follow on different handlings of decisive topics  (including Aristotle's position). It is also similar in its appreciation of the fascinating way topics oscillate between a variable many of positions and arguments, and a  truthful one (once the ontological standing of the topics is grasped). Too, it is similar in its claim that these position­generating devices are plausibly internal to or  definitive of dialectic. Nonetheless, I believe that the Topics, taken as a whole, tells us what this dialectic is all about. DeMarco errs, therefore, in calling the Topics  merely a 'rhetorical workbook'. It is the Topics above all which shows how topics define stances toward argument and

  

Page 207

      inquiry—including that properly philosophical stance able to achieve definitions and knowledge of principles. 17. Aristotle makes room for a philosophical dialectic that provides the road to first principles in every inquiry, but most passages where he writes "dialectic" without  qualification refer to nonphilosophical dialectic. Since the point of my chapter is to highlight the philosophical sense, when I write "dialectic," I standardly mean  philosophical dialectic. 18. E. Weil, "The Place of Logic," also recognizes that Aristotle is still dealing with philosophical dialectic in this book. As Weil puts it, The dialectical exercise, which has immense philosophical value (for in Aristotle's view the technique of formulating questions, of finding 'places' for attack and of arranging them  in their proper order is common to the philosopher and the dialectician: Top. VIII.1 155b3ff.) can only be brought to a successful conclusion by a thinker on his own if he plays  both parts: if he cannot find an interlocutor he must raise objections against himself. (99)

      Again Weil says, "True dialectic—topical, peirastic dialectic—is analytic technique used in the examination of a thesis that is commonly received or otherwise  celebrated, an examination that is undertaken as a common task by two or more men in search of the truth" (99). 19. The fact that Aristotle is concerned to provide both constructive and destructive arguments shows that J. D. G. Evans is mistaken when he claims that dialectic is  different from philosophy because whereas philosophy is scientific, dialectic is tentative. Dialectic is tentative for Evans because he says that "dialectic can demolish  claims to knowledge but positively it is unable itself to produce knowledge" (Evans, Aristotle's Concept of Dialectic, 12). That Aristotle is concerned to provide  constructive arguments shows dialectic is indeed helpful in discerning truth and not only falsehood. Aristotle deals explicitly with both constructive and destructive  arguments in book II.3–4, at the end of 7, and at the beginning of 8 and 9; book III.6; book IV.1–6; book V.2–9; book VI.2–13; book VII.1–5; just to name a few  cases. 20. This passage is highly reminiscent of the passage in Topics 1.2 100b34–37. R. Smith, "Dialectic and Method in Aristotle," in this collection, argues against a  philosophical role of dialectic by arguing against Irwin's and Owen's interpretations of this passage in Topics 1.2. Viz. n. 24 in Smith's essay for the Greek text of the  whole passage. Passing over the initial claim in this passage that a full examination of the puzzles on each side will lead us to see more easily what is true or false, Smith  argues against the reading of Owen and Irwin—both of which stress dialectics's 'finding' or 'establishing' the first principles. More specifically, Smith attacks Irwin's  gloss of dialectic's usefulness 'in connection with' (pros) the first principles with 'finding' the first principles, which is not in Aristotle, and raises doubt about Irwin's  optimism regarding the phrase that philosophical dialectic "has a road towards the first principles of all disciplines." Smith looks at the context of this passage and  points out that at most, the dialectical art includes "a study of logical consequences" and hence is useful to the examination of puzzles. Smith elaborates this aspect of  dialectic by saying, "skill in deducing the consequences of a position is a natural concomitant of skill in deducing conclusions from an opponent's opinions." However,  Smith says, "Aristotle does not say that this will provide us with a proof of anything but only says that it will help us to see what is true or false.'' Smith is correct to say  that Aristotle does not say that such

  

Page 208

      a skill of deducing logical consequences of arguments is a proof. But this is not an attack on Owen's or Irwin's view that Aristotle's philosophical dialectic  establishes or finds first principles because Aristotle's first principles are not subject to proofs but are rather defined (viz. 158a30–b4). In fact, Aristotle's claim that  such an examination of the puzzles on each side will help us "see what is true or false" (a claim that Smith downplays), is perfectly consistent with my interpretation of  how dialectic functions for Aristotle. Dialectic, for Aristotle, generates materials for arguments. It allows us to retain the relevant predicables and ultimately arrive at  definitions. In this case, dialectic allows us to arrive at definitions of first principles.       Arguing against Irwin's insertion of 'finding' for 'in connection with', Smith points out that Aristotle does not say that "dialectic establishes the principles," but rather,  all Aristotle said was that we must 'discuss' the first principles through the common beliefs on each subject. To bridge this gap between discussing and establishing,  Smith claims that we need to look at the claim that dialectic provides the road to the first principles. Smith argues against dialectic's role in the discovery of first  principles by giving an account of how Aristotle talks about the discovery of first principles in the Analytics. Smith says that Aristotle holds, the even stronger claim that if we collect together all the truths about any subject, we will find that there are certain truths among them which cannot be deduced from any  combination of the others but from which all the others can be deduced. These propositions must be the principles for the simple reason that they cannot be anything else: they  cannot be demonstrated. (52)

      Smith quickly dismisses that the Topics accomplishes what Aristotle maintains in the Analytics even though Smith's account of what Aristotle does in the  Analytics is very close to the Topics VIII.15 and 1.2 passages. The similarities in what Smith mentions in the Analytics and the Topics VIII.15 and 1.2 passages are  as follows. All three passages deal with: (1) the collection of materials (be these puzzles, or truths, or hypotheses); (2) the examination of these materials that lead to  certain results (such as seeing what is true or false, and/or seeing that certain truths cannot be deduced from others); and (3) the road to, or the discovery of first  principles, or the achievement of philosophic wisdom.       Is Smith's dismissal of the Topics, claiming that it does not find out what the first principles are because it does not accord with the Analytics's account, too  quick? After all, is not the procedure of amassing the materials for arguments in the Topics, which procedure lays bare the true and the false and lets us define the first  principles, a process of collecting the truths about any subject which Smith claims to be Aristotle's stronger claim about the discovery of first principles in the  Analytics? See 163b20–23, 163a5, 163b9ff. and 158a30–158b4 (these two later citations are quoted in the text). What Smith seems to have overlooked in his  account of the first principles from the Analytics is their connection with definition. Smith is right to note that first principles cannot be demonstrated. But he misses  Aristotle's claims in the Topics and Posterior Analytics that they are defined. Speaking about the significance of first principles for demonstrations and arguments,  Aristotle says of these first principles that "it is through them that everything else is made clear, they cannot be made clear through anything else, but everything of that  kind must be made known by definition" (158b3–4). Notice the similarity of this passage to Smith's account from the Analytics. The only difference is that Smith  missed the connection of first

  

Page 209

      principles with definitions, a connection Aristotle makes explicit in the Topics (viz. next paragraph in text), and Posterior Analytics II.       Smith's next reason for dismissing the ability of the Topics to establish first principles is based on his interpretation of what Aristotle means by 'examine' as a  property of dialectic—which claim comes at the end of the Topics 1.2 passage. Smith says that the word for 'examine' "is closely connected with refutation." He then  says that "(a) process of refutation is not a very likely candidate for establishing the first principles" (52). Instead of establishing first principles, Smith thinks that it is  more likely that refutation or continual examination of our opinions, changes our ''epistemic situation" by eliminating our certainty, and hence works better to educate us  by changing our minds. Smith is right that refutation alone will not get us to first principles and he is also right that dialectical examination or refutation educates us.  What Smith seems to be missing is that such refutations or examinations are only parts of dialectic. That is, refutation is only the negative or eliminative side of dialectic.  There is also the positive side of dialectic: construction and definition. Once Aristotle moves from an opinion to an expanded set of possibilities, he eliminates the ones  that do not fit, leaving the one definition or principle for which we are looking. Viz. Topics VIII.14 164b3–7 for how dialectic for Aristotle is a movement from the  one to the many and from the many to the one. By focusing on the refutations alone in abstraction from definitions, which I have shown to be central for Aristotle's  dialectic, Smith, like Evans (viz. previous note), misses the constructive and positive side of dialectic, and hence misses the true purpose and process of dialectic. The  next two sections will show more clearly how Aristotle's examination of puzzles and common beliefs can lead to the truth and first principles. 21. Evans, Aristotle's Concept of Dialectic, notices that the concepts of the same and other, like and unlike, contrariety, priority and posteriority which are  mentioned in the Metaphysics (995b21–2) are "prominent" in the Topics. For instance, he says, Questions about whether two things are the same or other are said to fall under the same heading as questions about definition and are treated in Top. H. 1–2; and in Top. A.7 we  are given an analysis of the senses of 'same'. The notions of similarity and contrariety provide topics in the discussions of accident, genus, property, and definition.... Priority and  Posteriority play an important part in the discussions of property and definition. (38–39)

      Nonetheless, Evans claims that "no special emphasis is placed in the Topics on the use of these concepts in dialectic" (39). The detailed discussion of these topics  in section III below should show how all­pervasive the use of these notions is in Aristotle's dialectic.       Another significant objection Evans raises is that dialectic fails to recognize that "the universal characters—same etc.—are attributes of Being qua being" and  hence fails to treat them in such a way that shows the primacy of substance (15). Evans cites Met. K.3 1061b4–11 saying that "dialectics and sophistry deal with the  attributes of existing things, but not of things qua Being, nor do they treat of Being itself in so far as it is Being" to substantiate his point. This reference, however, is not  helpful to Evans for Aristotle recognizes at least three uses of dialectic and it is not clear that he is talking about philosophical dialectic in this context (cf. 192 of this  chapter). Besides the "kai" linking dialectics and sophistry could mean "as" rather than "and" such that the sameness or likeness of the dialectical and the sophistic  practice is being emphasized (cf. Evans, Aristotle's Concept, 14 n. 27).

  

Page 210

      Evans cites another passage in Met. Mu.4 (1078b23–30) to support his claim that Aristotle's criticism of Socrates's dialectic is that it is too bound up with  definitions. Evans says that Aristotle holds that if Socrates "had practiced dialectic properly, he would not have attempted to make definition the basis of his  reasoning" (25). It is important to note that this passage which Evans cites lies in a context where Aristotle has just been talking about some of the presocratics and  adherents to the theory of forms. It is questionable if Aristotle meant to dissociate dialectic from definition since in this context, it seemed that he is praising Socrates.  For Aristotle says, "There are two innovations which may fairly be ascribed to Socrates: inductive reasoning and general definition. Both of these are associated with  the starting­point of scientific knowledge" (1078b27–30). So to say that Socrates is still seeking definitions and hence definitions should be put down and dissociated  from dialectic is not a logical conclusion. After all, how is one to get to first principles if one does not engage in definitions? How is one to construct arguments if one  does not define? For Aristotle says, apart from the passage I just cited from the Topics (above), ''it is through definitions that we get to know each particular  thing" (998b5). Even the passage Evans himself quoted speaks of the importance of definition to logical reasoning. Quoting 1078b24–5, Evans writes "for he was  trying to reason logically, and the starting­point of all logical reasoning is the essence."       More importantly, I think that Evans's impression that Aristotle is criticizing Socrates's dialectic stems from Aristotle's statement that "At that time there was as yet  no such proficiency in Dialectic that men could study contraries independently of the essence and consider whether both contraries come under the same  science" (1078b25–27). Evans took this statement to mean that true dialectic would study contraries independently of essences or definitions. But such a reading does  not seem consistent with Aristotle's immediate criticism of the Idealists who separated definitions and universals. As Aristotle puts it, "But whereas Socrates regarded  neither universals nor definitions as existing in separation, the Idealists gave them a separate existence, and to these universals and definitions of existing things they  gave the name of Ideas" (1078b30–34). Aristotle's continued criticisms against these adherents to the theory of forms who separate definitions and universals for the  rest of this chapter and the next chapter shows that he is against such separation. So when he said of Socrates and the other presocratics that they were not proficient  enough in dialectic to separate definitions from the essences of existing things, he was just being sarcastic. It was, according to Aristotle, a good thing that they did not  make such a separation.       Robert Bolton, "Definition and Scientific Method in Aristotle's Posterior Analytics and Generation of Animals," in Philosophical Issues in Aristotle's Biology,  Gotthelf and Lennox, eds. (Cambridge: Cambridge University Press, 1987), like Evans, also seeks to undermine the role of definition in Aristotle's Topics. He  examines three types of definition in the Posterior Analytics (143ff.) and claims that these three types of definitions are absent from the Topics (147). Bolton  recognizes though, that Aristotle talks about two kinds of definitions in the Topics, namely, definitions which are made by reference to what is more intelligible to us,  and those that are by reference to what is more intelligible absolutely (141b3ff.). But Bolton focuses on Aristotle's rejection of the former as definitions. By saying that  the former should not be the kind of definition at which one aims, however, need not mean a total rejection of this method. All Aristotle is saying is that the latter is the  true definition for which we are looking. So there is not a great distinction between his downplaying those definitions which are made by reference to what is more  intelligible to us, and the hierarchical ordering of the three types of definition Bolton himself mentions.

  

Page 211

      For the significance of Aristotle's Topics for the understanding of the Analytics, see E. Weil's "The Place of Logic in Aristotle's Thought." 22. Permutations of topics generate variations and provide materials for arguments. By 'permutations' I mean the ways in which these topics can be combined with the  subject and predicate to generate materials for examining if an accident, a property, or a definition belongs to a subject. E.g., consideration of the topic of greater and  less degree provides us with four possible arguments in Topics II.10 114b36–115a14. The first possible argument from considering greater and less degree with  respect to accidents is to see if an increase in the subject is followed by an increase in the accident. If this happens, then the accident really belongs to the subject, if  not, then it does not belong to the subject. The second argument considers the case where one accidental predicate is applied to two subjects. Aristotle tells us that if  this predicate does not belong to the subject it is more likely to belong, then it follows that it also does not belong to that to which it is less likely to belong. The third  argument considers two accidental predicates when applied to one subject, and the fourth two predicates when applied to two subjects. Similar permutations are used  to examine properties (137b14–138a29) and definitions (146a3ff. and 152b6). Thus, by using the topics and their permutations to test each of the predicables and  their contributions to definitions, Aristotle's dialectic expands the opinion from which one begins and hence is far from being restricted to one's common opinions. 23. See my "Senses of Being in Aristotle's Nicomachean Ethics," in The Crossroads of Norm and Nature: Essays on Aristotle's Ethics and Metaphysics, M. Sim,  ed., section IV for a detailed account of how Aristotle dialectically arrived at contemplation as the highest good of human beings. 24. It is also in this context that Aristotle speaks of the importance of having many definitions ( ) and arguments for and against  something, which abundance helps one reach knowledge and philosophic wisdom, since with materials in hand, one need only choose the true and avoid the false  (163b6–30). It is also in this context that Aristotle speaks of the importance of being "prompt about first principles" (163b28) and the ability of moving from the one to  the many and from the many back to the one (163b34–35 and 164b3–7). 25. Aristotle adds, "In what manner has been described elsewhere"—presumably this refers to Topics VIII.5 159a 25ff. 26. See also 172a12–15 where he claims that dialectical arguments are generally applicable to all things and that all things can come under the same principles (autas  archas, 172a15). 27. It is interesting to note that Robert Bolton, in "The Epistemological Basis of Aristotelian Dialectic" (updated for this collection) translates this passage as follows: It is possible even for someone who does not know the subject [scientifically] to test another who does not know the subject, providing that the latter grants things not based on  what he knows [scientifically] or, in particular, on the special principles of the subject, but on things which are consequent on [the special principles] which are such that, though  knowing them does not prevent one from knowing the discipline, still one who does not know them necessarily does not know it. (172a23–27)

      Notice that Bolton inserts "the special principles" after the talk of consequence. Bolton

  

Page 212

      does this because his interpretation is that the common things or what he calls "the special set of generally known endoxa" need to be explained by the principles  of a science. These scientific principles are for Bolton what are absolutely knowable whilst the endoxa are what are knowable to us. The latter turn out to be things of  perception accessible to everyone for Bolton. Though it seems that making the endoxa objects of perception solves the problem of commonality, to insist that they, in  turn, are to be explained by the special principles robs dialectic of anyway to substantial truths. Bolton's compromise is to say that dialectic by itself cannot determine  the principles, but "it plays a necessary role in finding those principles" (235) since it collects the materials that need to be explained by the first principles, it helps in  finding those candidates for first principles that do the best job of explaining these materials. Bolton's view hence presupposes that there is some other way of getting  the first principles, which is then used to explain the endoxa. It is not that endoxa help us get at these first principles.       Let us look at the Greek of this passage to see if Bolton's position is warranted. Aristotle is saying that the man without knowledge of the thing or subject can  examine someone "if the latter grants things not out of knowledge, nor out of the special principles, but out of what follows (ek tôn epomenôn), which are such that to  know them does not hinder him from being ignorant of the art, but not to know them is necessarily ignorant." Notice that nowhere is there an explicit claim that the one  who is being tested grants things that are out of what follows from the special principles here. There is however, talk of how he grants things not out of special  principles. So, for Bolton to insert "special principles" after "out of what follows" is to smuggle in an item that Aristotle has already ruled out. Out of what follows  seems to be 'knowledge' that is so general that it is available to both the scientific and to the non­scientific. For instance, Aristotle continues this discussion of how  everyone makes use of the common principles to test those who profess knowledge by saying, "Now this is where the common principles come in; for they know  these of themselves just as well as the scientists, even though their expression of them seems to be very inaccurate.'' Armed with these common principles (by these I  mean the topics whilst Bolton means the objects of perception), we are told that we can test those who profess to know by testing their knowledge of what follows  from the subject matter at stake. How this works in my view will become evident in my discussion of "consequential facts" (viz. 198). As regards Bolton's claim that  the concessions deal with things that are consequent on the special principles, we could ask: how is one to pick out things that are consequent on these special  principles if Aristotle had already ruled out the use of such special principles here?       Even if we were to grant Bolton that these special principles can or do explain the common things, to bring them in (which special principles are presumably the  sufficient conditions of knowledge) to determine the common things (which are the necessary but not sufficient conditions) is to invert Bolton's own claim that one  proceeds from what is knowable to us to what is knowable in itself. 28. Aristotle's point is at least more straightforward than Bolton takes it to be. 29. Aristotle says, "For it is for the sake of comprehension that we introduce the property; therefore it must be assigned in more comprehensible terms, for it will thus  be possible to understand it more adequately" (129b 7–10). For instance, Aristotle gives the example of how it is a mistake to assert that a property of fire is that it is  like the soul since "soul" is less comprehensible than fire, it does not help us to know what fire is by stating that it is like soul. Here, our testing of this individual who  asserts that he knows what fire is reveals that by assigning an obscure property, he could not tell us a property which will

  

Page 213

      help in the understanding of fire. Thus the consequential fact, namely, that if one knows something, then one can correctly assign a property, has not been achieved  and hence his ignorance is revealed. 30. For instance, we could imagine that in the above example dealing with species and genus, the nonscientist might not know the technical terms 'species' and 'genus'.  Further, when he asks for clarification concerning one's claim that the soul is in motion, he will actually be asking about what type of motion the soul has without  recognizing that motion is the genus and that the type of motion about which he is asking is a species. Nonetheless, his inquiry will have the same result as someone  who knows these technical terms. 31. My discussion, in section IV, of the different topical communities implied but not expressed in Aristotle's account will let us retain the truth­achieving status of  dialectic without insisting that all or most know or reason according to topics such as species and genus, matter and form, etc. 32. The example of fire and soul in n. 29 is this type. 33. This paper is an attempt to think together several projects I have undertaken on Aristotle's dialectic since 1994. "Dialectic and Being in Aristotle's Topics" was  presented at the 13th Annual Joint Meeting of the SAGP and the SSIPS (1994), and the Central Division APA SAGP meeting (1995). "Dialectic and Community in  Aristotle's Topics" was presented at the 14th Annual Joint Meeting of the SAGP and the SSIPS (1995), and the Metaphysical Society of America Meeting (1996).  Finally, "Ethics and Community in Aristotle" was presented at the Southwestern Philosophical Society Meeting (1996), the Central Division APA (1997), and the 20th  World Congress of Philosophy (1998). To the numerous participants of these meetings who have challenged my views and offered encouraging remarks, I am most  grateful. I am delighted to acknowledge Lenn Goodman, Ed Halper, Martha Husain, Rob Bolton, and Allan Bäck for their helpful comments and questions. I owe  thanks to Tom Tuozzo for his comments on ''Ethics and Community in Aristotle's Topics" at the Southwestern Philosophical Society Meeting, and to Iakovos Vasiliou  for his comments on the same paper at the Central APA. I am also grateful to Professor Alasdair MacIntyre for reading drafts of these projects and for offering  encouraging remarks. Most of all, I am indebted to Wes DeMarco for the countless, sometimes heated arguments, and the always helpful suggestions on each and  every one of these projects which spurred me on to clarifying and crystallizing my position.

  

Page 215

Chapter Eleven—  Poetry, History, and Dialectic Edward Halper Twice in the Poetics Aristotle contrasts poetry with history. Whatever its didactic value, the contrast has not seemed to readers of special philosophical interest. The  aim of this chapter is to show that this contrast is philosophically significant not just for our understanding of tragedy but also for the light it sheds on Aristotle's overall  methodology. It is often noticed that Aristotle's own philosophical works do not generally follow the logically demonstrative method that he sketches in his Analytics.  My concern here is to explore one methodological alternative that he does use. I shall show how he uses the method sketched in the Topics to define a tragedy and  explain why the same method will not define a history. I Aristotle claims that the art of dialectic sketched in the Topics contributes to philosophical knowledge because it can be used to find indemonstrable first principles  from common opinions: "for, being capable of examining, dialectic has a path to the principles of all disciplines" ( ) (I.2 101b3–4). Scientific  knowledge of a subject consists of grasping its principles and demonstrating its essential attributes from them. How does one come to know the first principles?  Obviously, they cannot be demonstrated from prior principles; they are first principles. As such, they are somehow determined by dialectic. Thus, dialectic transforms  what we can call, for lack of a better term, a

  

Page 216

'subject matter' into a science. What is the state of this subject matter before dialectic discovers its principles? It is clear from our Topics text that dialectic will examine  common opinions, and it is well recognized that Aristotle's actual inquiries often begin from common opinions (the case of akrasia is frequently cited—NE VII.1  1145b2–7). So the pre­scientific subject matter must contain common opinions about its subject matter. Aristotle has a name for such a setting out of common  opinions: in the Prior Analytics, he speaks of deriving the principles of each field from experience and he refers to the account of the phenomena of a field as a  "history" ( ) (46a17–27). Evidently, 'history' precedes 'science', and transition is effected by dialectic. Aristotle has much to say about how knowledge is  derived from sensation and experience, but he never explains how (or whether!) his many remarks fit together into a single process. I suggest the following: individual  universals come directly from sensation and experience (An. Po. II.19); they are combined by experience and common opinions about them formed; these common  opinions are examined by the method of the Topics to determine whether their components are always conjoined and to arrive at definitions; and these definitions are  finally confirmed or refuted by the method of An. Po. II.13. The middle stage, the formation of common opinions, is the 'history'. We can get some idea of what it  includes from the large collection known as the History of Animals. This is a preliminary collection of facts that is to be followed by a treatment of the same subject  through causes (HA I.6 491a7–14), that is, by the properly scientific treatment in the Parts of Animals (See. I.1, e.g. 639b5–10). Along these lines, think of the first  books of the Physics and De Anima and the first two books of the Politics: in each case, there is no definition until the second or third books. The first books cannot  quite count as science. Each is rather a preliminary sorting­through of common opinions and aporiai—a 'history', though quite different from the History of Animals.  In general, then, Aristotle takes history as a preliminary to science. What, though, about the discipline that is not the 'history of X' but just history, the discipline that the Poetics contrasts with poetry? Does it share anything in common  with other histories, aside from its name? Is it, perhaps, a preliminary to a science of history? Aristotle never mentions a science of history; rather, he claims that history  concerns the particular rather than the universal (1451b6–7), a sure indication that it is not a science, for science is of the universal. History would seem to be well­ named: it is indeed an   and unlike other histories it is never transformed into a science. Aristotle sees the discipline of history as a chronicle of human events  that are intrinsically particular and, consequently, never the subject of scientific knowledge. 1 With the evident failure of recent attempts at formulating historical laws, most contemporary readers are not likely to feel uncomfortable with Aristotle's understanding  of history. But it should have been problematic for him, for in Metaphysics A he speaks of successive discoveries of first causes as a necessary

  

Page 217

development and suggests that it will be followed by decline and later rediscoveries along the same lines. That is, when Aristotle does say something that resembles  history he advances a necessary and universal cycle. Of course, the particular details of these events, the personalities and the circumstances, are always different; but  these particulars fall under the universal cycle. How, we might ask, could there be particulars that are not instances of universals? Aristotle's insistence on the  particularity of history remains puzzling. 2 If history is peculiar in not being a preliminary for some science, then poetry is peculiar in not having a preliminary treatment of common opinions. The Poetics begins  by listing the topics that need to be treated in this "discipline" ( —1447a8–14), using the same term that the Topics had used for the result of dialectic (I.2  101b3–4). There is, to be sure, a discussion of the development of tragedy and comedy in chapters 4–5 that is usually labeled as 'history', but Aristotle offers it to  elucidate the differences between poetic genres discussed in the first few chapters. He maintains that the development of 'poetry' is 'natural'. And his sketch of the  necessary stages of this development bears more resemblance to the development of a biological organism than to what we can surmise—from the surveys of common  opinion which typically constitute the first books of his works—he takes 'history' to be.3  Lacking such a prior survey, the Poetics is apparently immediately a  discipline. Are its principles apparent at the start? How can the science of poetry begin with principles and forgo 'history'? II The Poetics begins by identifying tragedy and other arts as imitations (1447a13–16) and by distinguishing matter, object, and manner of imitation (a16–18). Imitation  is the genus, and these latter are ways of differentiating it. Aristotle proceeds in the first three chapters to characterize the distinctive matter ("in which"), object, and  manner of tragedy by contrasting them with those of other arts. Thus, tragedy's matter is rhythm, song, and meter, but, unlike other genres, it uses them at different  times; the object tragedy imitates is acts of good characters; and its manner is dramatic rather than narrative. These three differentiae appear, slightly altered, in the  six­part definition of tragedy in chapter 6. There the matter becomes song and speech; the object imitated somehow breaks into three parts–character, thought, and  plot; and the manner is the spectacle. I shall say more about this definition in the next section. After setting out the differentiae (chs. 1–3) and before collecting them into a definition (ch. 6), Aristotle speaks about the causes of the generation of imitative arts  (chs. 4–5). Thus, the Poetics opens immediately with a causal account of tragedy, presenting the formal (definition) and efficient causes. Such causes, especially the  definition, are the principles that dialectic should aim to

  

Page 218

uncover. Though they have seemingly been uncovered without dialectic, we can still inquire whether there is any hint of the techniques sketched in the Topics. We need to read between the lines here to seek what went into the construction of what is presented to us as a finished discipline, but there are hints in the text. As  noted, Aristotle simply announces that imitation is the genus under which tragedy and other poetic arts fall (1447a13–16). Yet, if we consider the text that follows, we  see that an alternative genus is, in effect, proposed and that Aristotle defends imitation. Shortly after introducing imitation as the genus, he remarks that there is no  name common to poetic works that use language alone with or without meter. Nor would there be a name common to mimes of Sophron and Xenarchus and Socratic  dialogues even if they were all written in a single type of meter (1147a28–b13). However, some people, "attaching the term 'poet' to the meter, speak of an 'elegiac  poet' or an 'epic poet', not terming them poets in respect of imitation but by what is common in respect of the meter" (1147b13–16). The consequence of their thinking  is that someone who writes a work of medicine or science in meter counts as a poet. However, Aristotle declares, there is nothing in common between Homer and  Empedocles except the meter (1447b16–20). Further, someone who mingled different meter forms would still be called a poet (1447b20–23). Those who use 'poet' to refer to a meter, do not use the term to refer to imitations. They have, in effect, proposed 'having meter' as an alternative to imitation as the  genus of poetry. If having meter makes a work poetry, then Empedocles's physics counts along with the Iliad as poetry because both have meter, and neither a  Socratic dialogue nor a mime could count as poetry unless they were put in meter. It need scarcely be said that, though we do restrict 'poetry' to what has meter, no Greek would accept such a restriction for poêsis, any more than we would so  restrict the application of the phrase 'fine arts'. Even so, we may well wonder whether Aristotle adequately addresses the objection to imitation as the genus by simply  insisting that Homer and Empedocles have nothing in common. And we may also wonder about the adequacy of his claim that mimes and Socratic dialogues would  not have the same name—read, belong to the same genus—even if they were both in meter. And what could be wrong with mixing meter if poetry were characterized  by having meter? His critique of meter is, however, readily intelligible if we look to the Topics. The first topic he mentions there under 'genus' is to consider whether the things that fall  under the genus do not also fall under what is proposed (IV.1 120b15–20); in our case, mimes, recognized as poetic works, do not fall under writings with meter.  Similarly, Aristotle's claim that Empedocles is more a physicist than a poet (1447b18–20) amounts to objecting to the proposed genus on the ground that it would  include something that does not belong in the genus, an example of the second topic of the Topics' 'genus' section (120b21–29). Works that combine meters would  seem unable to belong to any of the species

  

Page 219

of a genus of poetry defined by its meter, an instance of still another topic (121a27–39). That is, if the genus of poetry were characterized by having meter, then its  proper differentiae would have to be types of meter; and works that mix meters would fall under more than one species. If characterizing the genus as meter results in  some of its content not falling under any species of the genus, the characterization must be wrong. I could go on. Often a single argument can be seen as an instance of  more than one topic. But the point is clear: Aristotle relies upon distinctions and topics like those he expounds in the Topics to reject implicitly having meter as the  genus of poetry. Since meter is not the genus, it is reasonable to take it as a differentia of poetry. Some poetry is in meter, other in written or spoken language; still other in song or  dance: all these media are differentiae that mark species of poetry. The proposal that meter is the genus of poetry amounts to a confusion of a differentia for a genus.  Further, insisting that there would be no common name for mimes and dialogues even if they were written in the same meter, Aristotle is claiming, in effect, that the  species 'poetry that imitates in meter' cannot be further differentiated by distinguishing the various subjects treated. Distinctions of content cannot differentiate a type of  meter; for they are not proper differentiae, and the same differentiae would reappear under other species of poetry. In short, in the course of examining the various materials in which poetry could be made, Aristotle implicitly considers whether meter is not the best definition of this  genus and rejects it. In the process we come to understand what does constitute the genus of poetry and what some of its differentiae are, namely, differentiae that  are the matter "in which" the imitation is made. Because there is a metaphysical connection between genus and matter (Met. Z.12 1038a5–9), it is not so implausible to  suppose that the matter is the genus. On the contrary, it is surprising to see Aristotle treat the matter as a differentia of poetry here; we need to consider whether it  counts as a real differentia, a form, or whether it is a differentia only in the looser sense that it distinguishes one genre from another. Let us turn now to the second differentia, the object to be imitated. This is the action of an agent, 4  and Aristotle distinguishes between agents of good, bad, and  average characters according to their actions. Imitations of each of these three types can appear in different materials. That is, this second differentia divides several  of the species created by applying the first differentia to the genus. That the same differentia can differentiate coordinate species of poetry is contrary to  Metaphysics Z.12's doctrine of proper differentiation: each differentia belongs in one species or, anyway, in one line of differentiation and, therefore, implicitly  includes the single higher species that it differentiates. Obviously, this is impossible if the same differentia falls under different coordinate species or different genera.

  

Page 220

(In the Topics Aristotle repeats the conclusion of Z.12 that each differentia could only fall under a single genus, but then quickly reneges and acknowledges that one  differentia can fall under distinct genera (VI.6 144b12–30). His example is that 'biped' differentiates both 'walking­animal' and 'flying­animal'. It is hard to fathom why  Aristotle would state both views here: if a differentia could fall under more than one coordinate genus, then why mention the view that it could not? The example is  also troubling because in Parts of Animals (I.5) Aristotle makes a point of denying that the feet of humans and of birds are biped in the same way.) Whatever we are to make of the Topics example, that the same differentia (good objects of imitation) falls under different coordinate species of poetry is  understandable in light of Metaphysics H.2's recommendation to define artifacts and natural composites by giving both form and matter. In such cases, the form alone  does not suffice because it could be present in different matters, just as, in the present case, imitations of good characters can be present in various materials. To define  a species of poetry, it is necessary to mention both the object imitated and the kind of matter in which the imitation occurs. It is clear, then, that the first 'differentia' is  not part of the form of poetry but its matter, an important point for understanding the definition that Aristotle advances in ch. 6. Like other artifacts, species of poetry  are distinguished from each other through form and matter. The third differentia, the manner (narrative or dramatic), must also be added to distinguish some species of poetry, in particular, tragedy and epic. This differentia  belongs to neither matter nor form. It refers, rather, to the way in which the object imitated exists, in the matter. Narrative and dramatic manners also differentiate  coordinate species, but they lack the range of the second differentia. Though the last and narrowest differentia, manner is not the most characteristic and important,  as we might have expected, but relatively unimportant; for Aristotle declares that the poetic effect can be achieved with or without a performance (1450b15–20). In  contrast, other differentiae mark features without which tragedy could not occur. III These three differentiae form the basis for the definition of tragedy that Aristotle offers in chapter 6, as I Said. The definition is: Tragedy is the imitation of serious and complete action having magnitude in pleasing speech, with each of the kinds of speech used separately in its parts, dramatized rather than  narrated, and accomplishing through pity and fear the purgation of these emotions. (1449b24–28)

  

Page 221

Aristotle goes on to explain that "pleasing speech" refers to rhythm and song and that "the kinds of speech [being] used separately" refers to some parts having only  meter and others also song (1449b28–31; cf. b33–34). In other words, the "in which" is divided into the song and diction—the choral and spoken parts of the  tragedy. Further, the manner is now termed the spectacle. The object imitated is most properly a noble action and the imitation of it is the plot; the action stems from  the character and thoughts of the agents. Thus, the initial three differentiae are refined into six parts of the tragedy: spectacle, song, diction, thought, character, and  plot. What is particularly puzzling about these six parts and the complex definition of tragedy is their multiplicity. In the Topics Aristotle declares that "for each being the  essence ( ) is one" (VI.4 141a35). He uses this principle to exclude the possibility of multiple definitions of the same thing (141a35–b1), and  it is the basis for a lengthy presentation of topics pertinent to definitions that contain a plurality of parts (VI.13–14). One problem in particular with defining anything by  its parts is as follows: the contrary of any composition of parts is its decomposition, but an essence (in general) has no contrary (14 151a20–31). If, then, a tragedy is  defined as above, as a composition of some number of parts, its contrary would be these same parts disassociated from each other. But tragedy either has no contrary  or has a contrary different from this. Alternatively, if the parts do define the tragedy, then they do so whether united or separated, but the latter could hardly be called a  tragedy. It is just this sort of problem that Aristotle addresses in Metaphysics H.3. Recall that in the preceding chapter he had recommended defining certain entities, artifacts,  and natural composites by both form and matter; for example, a house by its function of providing shelter and also by its constituent boards and bricks. In H.3 he  claims that we cannot think of these, the boards and bricks and their position (or function), as if they were coordinate components (1043b4–6). If they were, there  would have to be something besides them to unify them—and this most important constituent would have been left out of the definition (1043b11–13). It is rather the  case that one component is form and the other matter (1043b30–32). The point is that the form is not like the material constituents but something of a different sort— the principle that unifies the material parts (cf. Z.17). An artifact is, thus, to be defined by its material constituents and the organization or function that serves as its  principle of unity. When we bring this lesson from the Topics and Metaphysics back to the text of Poetics 6, Aristotle's remarks about the definition suddenly make sense. First, his  claim that "plot is principle and soul of tragedy" (1450a38–39) is not the throw­away comment it seemed: soul is the form of a body, and plot is the principle of  tragedy in the sense that both are formal causes. As the form, plot can provide unity to the other five parts, and Aristotle seems intent on showing that is the case.  Thus, the arguments to show that plot ("the structure"—

  

Page 222

) is more important than character (1450a15–38) have the effect of assigning the job of unification to plot rather than character. Both character and  thought are justifications for the actions of the protagonists; the tragedian wants characters who habitually do what he wishes them to do in the tragedy or who can be  led to the desired action through their thoughts. In either case, it is the action to be performed that is primary, and the character or reasoning is formulated to justify the  action. Hence, plot unifies thought and character. Aristotle's more detailed remarks on these parts later in the Poetics serve the same end. Thus, the four features of  character that he recommends in chapter 15 all make character contribute toward making the action, the plot, plausible. Likewise, his brief remarks on thought in  chapter 19 also emphasize its role in explaining action. Thus, Aristotle accounts for the unity of the formal elements by showing how all contribute to the unity of plot. There is significantly less discussion of the material elements of tragedy in the Poetics, but at every turn Aristotle subordinates them to plot. He recommends that plots  be summarized first and only then brought to completion through speech (ch. 17); and he counsels against elaborate speech that obscures character or thought (24  1460b2–5). Likewise, the chorus (song) should be part of the whole and contribute to the performance (18 1456a25–27). Spectacle is artless and belongs partly to  the designer rather than the poet (6 1450b16–20). Though it could produce pity and fear, the tragic poet will be using his art if he makes them come from the plot (14  1453b7–14). In sum, plot is the formal principle of unity of the formal parts, thought and character; and the material parts are appropriately used when they contribute to the plot.  Plot itself ought to be a unity (ch. 7). This is best done not by making plot simple but by organizing many events. In brief, the plot is one if it imitates a single action, and  many events count as a single action when they share a single end. Determining what counts as a single action is a central problem of the Poetics. For present  purposes, the details of this discussion need not concern us; what is important for us to realize is the significance that unity of the plot has for Aristotle. Approaching the Poetics with the criterion of the Topics in mind enables us to see that the character that would allow tragedy to be defined, the unity of its parts, is a  major theme of the work. What might have seemed to be scattered remarks can now be understood to contribute toward the task of showing that tragedy admits of  definition and, consequently, of being known. Claims about the unity of tragedy's parts function normatively as well as descriptively. Aristotle aims, after all, not only to describe tragedies, but also to teach us how  to make good tragedies. In a word, the more a tragedy is one, the better it is as a tragedy. That is, it is not any conjunction of plot, character, etc. that makes a  tragedy; the tragedy exists only if the plot functions to unify the other parts. Only then is the tragic effect achieved. We need not consider the truth of this account here.  What we do need to notice is that the Poetics satisfies

  

Page 223

the unity criterion set out in the Topics with a move that parallels that of Metaphysics H.3. The dialectic of the Topics is not explicitly the concern of the Poetics, but it  is presupposed at every juncture. Despite the apparent lack of unity in the definition of tragedy, a tragedy can and should be one. IV Let us now return to the points of contrast between poetry and history. Recall that there are two. First, poetry represents universals and history particulars, a  difference that makes poetry more philosophical than history (9 1451a38–b7). Second, whereas poems (specifically, tragedies and epics) imitate a single, complete  and whole action, a history concerns all the events in a single time that happened to one or more people and that relate to each other at random and have no single end  (23 1459a17–30). Since a universal is a kind of unity (see Met.  .26 1023b29–32), both contrasts are ways of saying that poems are unities and histories are not or  are much less so. Aristotle compares the tragedy or epic to an animal (23 1459a20; also 7 1450b34–1451a6); he means to say that its parts have an organic unity in  contrast with the events of history. In the preceding section, I argued that the reason that a tragedy is one is that its parts are unified by plot. It is just this unity of the tragedy that allows it to be defined  and to be an object of a science. History has no such unity. It is intrinsically undefinable, intrinsically particular, intrinsically unphilosophical. Poetry is more  philosophical than history because it can be known. Yet both passages contrasting poetry and history are followed immediately by injunctions to tragedians to draw  upon history as their source material. Aristotle advises writers of tragedy to: keep to actual names. The reason is that what is possible is believable ... things which have happened are obviously possible.... [Yet] even if it turns out that he is representing  things that happened, he is no less a poet; for there is nothing to prevent some of the things that have happened from being the sorts of things that may happen according to  probability. (9 1451b15–32)

So, too, after denying that historical events constitute a single, complete action, Aristotle goes on to speak of the Trojan War as a source of plots for tragedies and  epics (23 1459a30–b7). To be sure, the poet can make up his plot entirely; this is common in comedies. Even so the events must follow with necessity or probability.  So the poet will need to appreciate the motives of human acts—and for this he needs some grasp of what people have done. That is to say, he will need to know  history either to draw his plots from it or to learn what counts as necessary or probable events. Moreover, history and poetry cover the same ground, human actions.  They differ in the amount of unity that they take these actions to have.

  

Page 224

We can appreciate why history is a preliminary for tragedy by recalling Aristotle's remarks on the origin of poetry in chapter 4. He points to both the natural tendency  to imitate and the delight that we derive from imitation (4 1448b4–9). 5  The reason for our pleasure is that imitation is a source of learning, and learning is pleasurable  (b9–17). And learning from representations does not require knowledge of the original (b17–19). In a representation it is possible to draw attention to just those details that are salient. This is the reason that it is easier to learn anatomical details by first studying  diagrams: under a high­powered microscope or in nature the salient details are present with too many other details to grasp them clearly. So, too, the tragedy or epic  does not present all the actions or characters that occurred in a time period but only those pertinent to a single sequence of events. From this sequence we can learn  the cause of the event; we can see clearly that human actions have consequences. Poetry represents manifold human actions described in history as a complete  sequence with a beginning, a middle, and an end that follows with necessity or probability from what has gone before. So, while both history and poetry deal with  human actions, it is poetry that gives them a scientific treatment. Apparently, poetry functions, in effect, like a science of human action, and the material from which this science is derived is history. Thus, history stands to poetry like  the   of the soul (De Anima I) stands to the rest of the De Anima. In each case, a history is the essential preliminary that occurs before there is a definition. Trouble is, the discipline of history is unlike these other cases because attaining a definable human action would transform history into an entirely different discipline, a  discipline that is fictional. The poet who writes a tragedy about a historical event essentially transforms that event into something of a very different character. Whatever  the poet has to say about that event has no immediate relevance to its historical treatment. The historian, for his part, aims to gather and to organize particulars. Yet, if  he were to organize them into necessary or probable causal sequences, as the poet does, he would no longer be doing history, but, Aristotle seems to think, writing  fiction. Why can there not be a scientific history that studies the true universal causes of human actions? In one sense, Aristotle's answer has been clear all along: there are no  universals in human actions; apparently, tragedy creates universals and causes. How could this discipline that imitates human action, tragedy, have more universality,  more unity, than what it imitates? Indeed, how could it imitate human actions when these actions themselves have no unity? And does not Aristotle presuppose the  existence of such historical causal sequences when he claims that tragedy imitates ''complete actions" (6 1449b24–25)? Or, to raise the same issue a different way, we  can note that Aristotle's counsel to would­be tragedians to look to history so that their plots are

  

Page 225

necessary or probable presupposes that there is some order to history. How, then, can he consistently deny the existence of a scientific history? There is no way to resolve the puzzles of the two preceding paragraphs entirely, but three points, all observations about Aristotelian science and definition, go some  way toward making his thought intelligible. First, 'history' as Aristotle understands it is not simply the setting out of facts; it also involves making connections and  drawing conclusions—we need only look at a 'history' that precedes an Aristotelian science to appreciate what it can achieve. Were there a scientific history, its  conclusions would be definitive. As suggested earlier, few people today would fault Aristotle for not endorsing such a history. Not a science, Aristotle's history arrives  at only probable conclusions. Poetry may deal with a probable sequence of events, but it presents that sequence as a definite causal sequence. Second, an Aristotelian science requires a genus. As we have seen, the genus of tragedy is imitation. In contrast, an  ), and treats motions as essential or  accidental attributes of this genus. So, too, the discipline of history is concerned with human action, and the latter falls under many genera: military actions under  strategy, political actions under politics, and so forth. The tragedian imitates history not only in constructing a plot that may be modeled upon a historical act but also in  making the events of the tragedy follow a probable sequence. Yet, though tragedy imitates human actions, its genus is not human actions but imitation. As an imitation,  tragic action, that is, plot, is wholly different from the action it imitates. Moreover, the preliminary study that Aristotle recommends to the tragedian is not the history of  the development of tragedy or paradoxes about the existence of productive sciences—that, we can surmise, would not help him to write better tragedies—but history  proper. It follows that the relation of the discipline of history to poetry is not the same as the relation of other histories to the sciences that develop from them: history is  preliminary to poetry because poetry is partially modeled on it, not because pursuing history leads to poetry. Third, an Aristotelian science requires an essence. The point of showing that tragedy is one is to show that it has an essence and is, thereby, definable. Interest in  history generally stems from the insight it is supposed to provide into humanity, but Aristotle has several sciences that treat human nature: biology, ethics, and politics.  The latter two disciplines treat human actions. Reflecting on

  

Page 226

them, we can see that Aristotle has, in effect, no room left for scientific history. Ethics is concerned with happiness and politics with the state, but Aristotle understands  happiness and the state as human activities, activities that are complete because they are their own ends. And ethics and politics study, as well, how these activities  develop and how they are destroyed. Since happiness and the state are 'by nature', their generation and destruction follow regular paths that depend upon the  condition realized or lost. The course of a state's development is no more 'history' than the course of a person's development. Of course, in each case, there are  particular details that are accidental to science; such as, the personalities involved in the state and the color of a person's hair. Such accidents are the bedrock of the  discipline of history. All else is science or potentially science. In short, by providing a scientific understanding of human actions, ethics and politics exclude the  possibility of scientific history. Indeed, tragedy draws upon history, but the objects it imitates are moral actions: its characters have flawed virtues and, in some sense,  get what they deserve. History, alas, does not work that way. To conclude, the definition of tragedy advanced in the Poetics presupposes dialectic, but, unlike other Aristotelian sciences, it does not emerge from a history. History  proper, the preliminary to poetics, does not admit of scientific treatment because it contains no essential unities. Poetics, on the other hand, does not consist of seeing  unity in human events but creating a unity that is not there. The method at work in the opening of the Poetics is dialectic, and the distinction between poetry and history  is best understood in terms of Aristotle's views of science and definitions. 6 Notes 1. In contrast, Stephen Halliwell writes, "It is not that life or history cannot furnish coherent structures of action, but just that they do not regularly do so, according to  Ar. (though most historical writing is based upon an implicit rejection of this premise)," The Poetics of Aristotle: Translation and Commentary (Chapel Hill:  University of North Carolina Press, 1987), 105. When does history ever "furnish coherent structures of action"? Halliwell may be thinking of 1251b29–33, for there  Aristotle suggests that historical events could have happened according to probability. More on the problem of scientific history later. It is worth noting that the  probability of the events does not diminish the role of the poet in representing them. Significant art is required whatever the status of the events. 2. G. E. M. de Ste. Croix, "Aristotle on History and Poetry," in Essays on Aristotle's Poetics, Amélie Oksenberg Rorty, ed. (Princeton: Princeton University Press  1992), 28–29, argues that Aristotle's dismissal of scientific history is, "even on [his] own premises,.. not fully justified." This conclusion is based upon Aristotle's  claiming that science is of the universal or what is for the most part and Thucydides's concern with the constancy of human nature and the resulting patterns of human  behavior that tend to

  

Page 227

      recur. de Ste. Croix thinks that Aristotle should have recognized Thucydides's works (with which he was familiar) as scientific. 3. The development of tragedy and comedy described in Poetics 4–5 is no more a history than Aristotle's account of the development of a state. Like the state,  imitation arises from nature (4 1448b4–24; cf. Pol. I.2 1252a24–31), and it had only to reach its pinnacle in tragedy (4 1449a14–15; 5 1449b17–20; cf. Pol.  1252b27–31). The progression of development is a function of the form that ultimately comes to be realized through it. Stephen Halliwell, Aristotle's Poetics (Chapel  Hill: University of North Carolina Press, 1986), 93–95, arrives at the same conclusion. 4. Aristotle designates the object imitated with a participle, "people acting" ( ) (1448a1). His emphasis is on the quality of the action, for through it the  character of the agent is decided. See L. Golden and O. B. Hardison Jr., Aristotle's Poetics: A Translation and Commentary for Students of Literature  (Tallahassee: University Presses of Florida, 1981), 82. 5. For an interesting discussion of this text, upon which my next paragraph draws, see David Gallop's, "Animals in the Poetics," Oxford Studies in Ancient  Philosophy, VIII (1990): 145–71. 6. I am grateful to Martha Husain and Eugene Garver for comments on an earlier draft of this chapter, and to Larry Jost for comments at an APA meeting.

  

Page 229

Selected Bibliography Primary Texts al­Ash'ari. Kitab al­Luma'. In The Theology of al­Ash'ari, edited and translated by R. J. McCarthy. Beirut: Catholic Press, 1953. al­Farabi. Sharh fi­l­Ibara (Commentary on De Interpretatione). Edited by W. Kutsch and S. Marrow. Beirut: Catholic Press, 1960. In Al­Farabi's Commentary  and Short Treatise on Aristotle's De Interpretatione. Translated by F. W. Zimmermann. London: Oxford University Press, 1981. al­Ghazali. Ihya' 'Ulum al­Din. Cairo: Aleppo Foundation and Publishing and Distribution Companies, 1312 A.H. Apostle, H., tr. Aristotle's Metaphysics. Bloomington: Indiana University Press, 1966. ———. Aristotle's On the Soul. Grinnell, Iowa: Peripatetic Press, 1981. Augustine, St. Confessions. F. J. Sheed, tr. Indianapolis, Ind.: Hackett Publishing Company, Inc., 1993. Barnes, J. Aristotle's Posterior Analytics, Translated with Notes. Clarendon Aristotle Series. Oxford: Oxford University Press, 1975. Barnes, J., ed. The Complete Works of Aristotle: The Revised Oxford Translation, 2 vols. Princeton: Princeton University Press, 1984. Bonitz, H., ed. Index Aristotelicus. Aristotelis Opera. 5 vols. Academia Regia Borussica (Prussian Academy), 1870. Cicero, Marcus Tullius. Brutus, On the Nature of the Gods, On Divination, On Duties. Translated by H. M. Poteat, with an introduction by R. Mckeon. Chicago:  University of Chicago Press, 1950. ———. De Senectute,. De Amicitia, De Divinatione. Translated by W. A. Falconer. Cambridge, Mass: Harvard University Press, 1964. Cooke, H. P., tr. Aristotle. On Interpretation. Cambridge, Mass.: Harvard University Press; London: W. Heinemann Ltd., 1938. Forster, E. S., tr. Aristotle. On Sophistical Refutations. Cambridge, Mass.: Harvard University Press; London: W. Heinemann Ltd., 1955. Freese, J. H., tr. Aristotle. The Art of Rhetoric. Cambridge, Mass.: Harvard University Press: London: W. Heinemann Ltd., 1926. Gallop, David, tr. Parmenides of Elea: Fragments: A Text and Translation. Toronto: University of Toronto Press, 1984. Grote, G. Aristotle. Second edition. London: Murray, 1880. Halliwell, S. The Poetics of Aristotle. Chapel Hill: University of North Carolina Press, 1987. Hardie, R. P., and Gaye, R. K., tr. Physica. In The Works of Aristotle, vol. ii, edited by W. D. Ross. Oxford: Oxford University Press, 1930. Hope, Richard., tr. Aristotle. Metaphysics. New York: Columbia University Press, 1952. Hussey, Edward. Aristotle's Physics, Books III and IV: Translated with Notes. Clarendon Aristotle Series. Oxford and New York: Oxford University Press,  1983. Irwin, T. H., tr. Aristotle. Nicomachean Ethics. Indianapolis, Ind.: Hackett Publishing Company, 1985.

  

Page 230

Kirwan, Christopher. Aristotle's Metaphysics, Books G, D, and E Translated with Notes. Clarendon Aristotle Series. Oxford: Oxford University Press, 1971. Lettinck, Paul, and J. O. Urmson, tr. On Aristotle's Physics 5–8 with On Aristotle on the Void, by Philoponus and Simplicius. Ithaca, N.Y.: Cornell University  Press, 1994. Peck, A. L., tr. Aristotle. Generation of Animals. Cambridge, Mass.: Harvard University Press; London: W. Heinemann Ltd., 1942. ———. Aristotle. Historia Animalium Books I–III. Cambridge, Mass.: Harvard University Press; London: w. Heinemann Ltd., 1965. Peck, A. L., and E. S. Forster, tr. Aristotle. Parts of Animals Cambridge, Mass.: Harvard University Press; London: W. Heinemann Ltd., 1937. Rackham, H., tr. Aristotle. The Nicomachean Ethics. Cambridge, Mass.: Harvard University Press; London: W. Heinemann Ltd., 1926. Ross, W. D. Aristotle's Prior and Posterior Analytics: A Revised Text with Introduction and Commentary. Oxford: Oxford University Press, 1949. Ross, W. D., ed. Aristotle's Metaphysics. Oxford: Clarendon Press, 1924. Sextus, Empiricus. Sextus Empiricus. Translated by R. G. Bury. London: Loeb Classical Library, 4 vols., 1917–1955. ———. Outlines of Scepticism. Sextus Empiricus. Translated by J. Annas and J. Barnes. Cambridge and New York: Cambridge University Press, 1994. ———. The Skeptic Way: Sextus Empiricus's Outlines of Pyrrhonism. Translated with introduction and commentary by Benson Mates. New York: Oxford  University Press, 1996. Sharples, R. W., tr. Cicero. On Fate. Boethius. Consolation of Philosophy. Warminster, England: Aris and Phillips, 1991. Tredennick, Hugh, tr. Aristotle. Metaphysics Books I–IX. Cambridge Mass.: Harvard University Press; London: W. Heinemann Ltd., 1933. ———. Aristotle. Metaphysics Books X–XIV. Cambridge, Mass.: Harvard University Press; London: W. Heinemann Ltd., 1935. Tredennick, Hugh, and E. S. Forster, tr. Aristotle. Posterior Analytics and Topica. Cambridge, Mass.: Harvard University Press; London: W. Heinemann Ltd.,  1960. White, N. P., tr. Plato. Sophist. Indianapolis, Ind.: Hackett Publishing Company, 1993. Wicksteed, P. H., and F. M. Cornford, tr. Aristotle. The Physics Books I–IV. Cambridge, Mass.: Harvard University Press; London: W. Heinemann Ltd., 1929. ———. Aristotle. The Physics Books V–VIII. Cambridge, Mass.: Harvard University Press; London: W. Heinemann Ltd., 1934. Williams, Bernard A. O., M. J. Levett, and M. Burnyeat, ed., tr., and rev. Plato. Theaetetus. Indianapolis, Ind.: Hackett Publishing Company, 1992. Wittgenstein, Ludwig. Philosophical Investigations. G. E. M. Anscombe, tr. Oxford: Blackwell, 1953. Woods, Michael, tr. Aristotle's Eudemian Ethics: Books I, II, and VIII. Oxford: Clarendon Press, 1982. Secondary Sources Anscombe, G. E. M. "Aristotle and the Sea Battle." Mind 65 (1956): 1–15. ———. "Aristotle and the Sea Battle." Reprinted (without the appendix on Diodorus Cronos) in Aristotle: A Collection of Critical Essays. Edited by J. Moravcsik.  Garden City, N.Y.: Doubleday, 1967, 15–33.

  

Page 231

Bäck, Allan. "Aristotelian Necessities." History and Philosophy of Logic 16 (1995): 89–106. Barnes, J. "The Law of Contradiction." Philosophical Quarterly 19 (1969): 302–9. ———. "Aristotle's Concept of Mind." Proceedings of the Aristotelian Society 72 (1971–1972): 101–114. ———. "Aristotle's Theory of Demonstration." In Articles on Aristotle I, edited by J. Barnes, M. Schofield, and R. Sorabji. London: Duckworth and Company  Limited, 1975, 65–87. ———. "Aristotle and the Methods of Ethics." Revue Internationale de Philosophie 34 (1980): 490–511. Becker, O. "Zur Rekonstruktion des 'Kyrieuon Logos' des Diodorus Kronos." In Erkenntnis und Verantwortung Festschrift für Th. Litt, edited by J. Derbolav  and F. Nicolin. Düsseldorf: Pädagogischer Verlag Schwann, 1960. Berti, Enrico. Review of T. H. Irwin, Aristotle's First Principles. Elenchos 12 (1988): 116–25. ———. "Does Aristotle's Dialectic Develop?" In Aristotle's Philosophical Development, edited by W. Wians. Lanham, Md.: Rowman and Littlefield, 1996, 105– 30. Bolton, Robert. "Definition and Scientific Method in Aristotle's Posterior Analytics and Generation of Animals." In Philosophical Issues in Aristotle's Biology,  edited by A. Gotthelf and J. Lennox. Cambridge: Cambridge University Press, 1987, 120–166. ———. "The Epistemological Basis of Aristotelian Dialectic." In Biologie, Logique et Métaphysique chez Aristote, edited by D. Devereux and P. Pellegrin. Paris:  Éditions de CNRS, 1990, 185–235. ———. "Aristotle's Method in Natural Science: Physics I." In Aristotle's Physics, edited by L. Judson. Oxford: Oxford University Press, 1991, 1–29. ———. "Aristotle's Conception of Metaphysics as a Science." In Unity, Identity and Explanation in Aristotle's metaphysics, edited by T. Scaltsas, D. Charles,  and M. L. Gill. Oxford: Clarendon Press, 1994, 321–354. ———. "The Problem of Dialectical Reasoning (sullogismos) in Aristotle." In Logic, Dialectic and Science in Aristotle, Ancient Philosophy (Special Issue),  edited by R. Bolton and R. Smith, 14 (1994): 99–131. Brown, H. Observation and Objectivity. Oxford: Clarendon Press, 1987. Brunschwig, J., tr. Aristote: Topiques, vol. I. Paris: Société D'Édition 'Lex Belles Lettres', 1967. ———. "Aristotle on Arguments Without Winners or Losers." In Wissenschaftskolleg Johrbuch (1984/5): 31–40. Burnet, J. The Ethics of Aristotle. London: Methuen, 1900. Burnyeat, M. "Aristotle on Understanding Knowledge." In Aristotle on Science, edited by E. Berti. Padua: Antenore, 1981, 97–139. ———. "Good Repute." London Review of Books, November 6, 1986. Cherniss, H. Aristotle's Criticism of Plato and the Academy I. Baltimore: Johns Hopkins, 1994. Cleary, J. J. Aristotle on the Many Senses of Priority. Carbondale: Southern Illinois University Press, 1988. ———. "Working Through Puzzles with Aristotle." The Journal of Neoplatonic Studies 1, no. 2 (1993): 83–179. ———. "Phainomena in Aristotle's Methodology." International Journal of Philosophical Studies 2, no. 1 (1994): 61–97.

  

Page 232

———. "Working Through Puzzles with Aristotle." Reprinted in The Crossroads of Norm and Nature: Essays on Aristotle's Ethics and Metaphysics, edited by  M. Sim. Lanham, Md.: Rowman and Littlefield, 1995. ———. Aristotle and Mathematics. Leiden: E. J. Brill, 1995. Code, A. "The Persistence of Aristotelian Matter." Philosophical Studies 29 (1976): 356–67. ——. "The Aporematic Approach to Primary Being in Metaphysics Z." Canadian Journal of Philosophy Supp. 10 (1984): 41–65. ——. "Aristotle's Investigation of a Basic Logical Principle: Which Science Investigates the Principle of Non­Contradiction?" Canadian Journal of Philosophy 16  (1986): 341–357. Cohen, R. S., and I. Kuçuradi, eds. The Concept of Knowledge. Dordrecht: Kluwer, 1995. Cooper, John. Review of Martha Nussbaum, The Fragility of Goodness. The Philosophical Review 97, no. 4 (1988): 543–65. DeMarco, C. W. "Plato's Ghost: Consequences of Aristotelian Dialectic." In The Crossroads of Norm and Nature: Essays on Aristotle's Ethics and Metaphysics,  edited by M. Sim. Lanham, Md.: Rowman and Littlefield, 1995. Denyer, Nicholas. "Time and Modality in Diodorus Cronus." Theoria 47 (1981): 31–53. de Ste. Croix, G. E. M. "Aristotle on History and Poetry." In Essay on Aristotle's Poetics, edited by A. O. Rorty. Princeton: Princeton University Press, 1992. Döring, Klaus. "Gab es eine Dialektische Schule?" Phronesis 34 (1989): 293–310. Düring, I., and G. E. L. Owen, eds. Aristotle and Plato in the Mid­Fourth Century. Göteborg: Almqvist & Wicksell, 1960. Engberg­Pedersen, T. "More on Aristotelian Epagoge." Phronesis 24 (1979): 301–19. Evans, J. D. G. "The Codification of False Refutations in Aristotle's De Sophisticis Elenchis." Proceedings of the Cambridge Philological Society 21 (1975): 42– 52. ———. Aristotle's Concept of Dialectic. Cambridge: Cambridge University Press, 1977. ———. Aristotle. Sussex: Harvester Press, 1987. ———. Review of T. H. Irwin, Aristotle's First Principles. In Hermathena 147 (1989): 73–4. ———. "Ancient and Modern Dialectic." Philosophical Studies (Ireland) 33 (1991): 39–53. ———. "Aristotle on Learning." Annales 24 (1992): 33–48. ———. "Souls, Attunements and Variation in Degree: Phaedo 93–4." International Philosophical Quarterly 34 (1994): 277–87. ———. "Meno's Puzzle." In The Concept of Knowledge, edited by R. S. Cohen and I. Kuçuradi. Dordrecht: Kluwer, 1995, 67–80. Ferejohn, M. "Aristotle on Necessary Truth and Logical Priority." American Philosophical Quarterly 18, no. 4 (1981): 285–94. ———. "The Definition of Generated Composites in Aristotle's Metaphysics." In Unity, Identity and Explanation in Aristotle's Metaphysics, edited by T.  Scaltsas, D. Charles, and M. L. Gill. Oxford: Clarendon Press, 1994. Fine, G. "Inquiry in the Meno." In The Cambridge Companion to Plato, edited by R. Kraut. Cambridge: Cambridge University Press, 1992, 200–26. Frede, M. Essays in Ancient Philosophy. Minneapolis: University of Minnesota Press, 1987. Gadamer, Hans­Georg. Truth and Method. Second edition, translated by J. Weinsheimer and D. Marshall. New York: Crossroad Publishing, 1989.

  

Page 233

Gallop, David. "Animals in the Poetics." Oxford Studies in Ancient Philosophy 8 (1990): 145–71. Garver, E. Aristotle's Rhetoric: An Art of Character. Chicago: University of Chicago Press, 1995. ———. "Rhetoric, Hermeneutics, and Prudence in the Interpretation of the Constitution." In Rhetoric and Hermeneutics in Our Time, edited by W. Jost and M.  Hyde. New Haven, Conn.: Yale University Press, 1996. ———. "Living Politically and Living Rationally: Choosing Ends and Choosing Lives." (Forthcoming) Gill, M. L. Aristotle on Substance. Princeton: Princeton University Press, 1989. Golden, L., and O. B. Hardison Jr. Aristotle's Poetics: A Translation and Commentary for Students of Literature. Tallahassee: University Presses of Florida,  1981. Goodman, L. E. "Al­Farabi's Modalities." (In Hebrew with English summary.) Iyyun 23 (1972): 100–12. ———. "Skepticism." Review of Metaphysics 36 (1983): 819–48. ———. "Context." Philosophy East and West 38 (1988): 307–23. ———. God of Abraham. New York: Oxford University Press, 1996. Gotthelf, A., and J. Lennox, eds. Philosophical Issues in Aristotle's Biology. Cambridge: Cambridge University Press, 1987. Halliwell, Stephen. Aristotle's Poetics. Chapel Hill: University of North Carolina Press, 1986. ———. The Poetics of Aristotle: Translation and Commentary. Chapel Hill: University of North Carolina Press, 1987. Halper, E. Form and Reason. Albany: SUNY Press, 1993. Hardie, W. F. R. Aristotle's Ethical Theory. 2nd edition. Oxford: Clarendon Press, 1980. Harman, G. Change in View: Principles of Reasoning. Cambridge, Mass.: MIT Press, 1986. Havas, K. "Differences in the Unity." Logique et Analyse 29 (1986): 149–60. Hempel, Carl. "Aspects of Scientific Explanation." In Aspects of Scientific Explanation and Other Essays. New York: The Free Press, 1965, 331–496. ———. "Empiricist Criteria of Cognitive Significance." In Aspects of Scientific Explanation and Other Essays. New York: The Free Press, 1965, 101–22. Hintikka, Jaakko. "Aristotle and the 'Master Argument'." American Philosophical Quarterly 1 (1964): 101–14. ———. "The Once and Future Sea Fight." Philosophical Review 73 (1964): 461–92. "The Development of Aristotle's Ideas of Scientific Method." In Aristotle's Philosophical Development, edited by W. Wians. Lanham, Md.: Rowman and Littlefield,  1996, 131–50. Horn, L. R. A Natural History of Negation. Chicago: Chicago University Press, 1989. Husain, M. "The Multiplicity in Unity of Being qua Being in Aristotle's pros hen Equivocity." New Scholasticism LV, no. 2 (1981): 208–218. Irwin, T. H. "Aristotle's Discovery of Metaphysics." Review of Metaphysics (1977): 210–229. ———. "Aristotle's Methods of Ethics." In Studies in Aristotle. vol. 9., edited by D. J. O'Meara. Washington D.C.: Catholic University of America Press, 1981. ———. Aristotle's First Principles. Oxford: Clarendon Press, 1988. Izutsu, Toshihiko. Towards a Philosophy of Zen Buddhism. Boulder, Colo.: Prajna Press, 1982. Jones, B. "Aristotle's Introduction of Matter." Philosophical Review 83 (1974): 474–500.

  

Page 234

Kant, I. Critique of Practical Reason. Translated by L. W. Beck. Indianapolis: Bobbs­Merrill, 1956. Kirwan, C. Aristotle's Metaphysics G, D, E. Oxford: Clarendon, Press, 1971. Kneale, William, and Martha Kneale. The Development of Logic. Oxford: Oxford University Press, 1962. Kraut, R. Review of T. H. Irwin, Aristotle's First Principles. The Philosophical Review, 101, no. 2 (1992): 365–71. Kraut, R., ed. The Cambridge Companion to Plato. Cambridge: Cambridge University Press, 1992. Lewis, F. A. "Aristotle on the Relation between a Thing and Its Matter." In Unity, Identity and Explanation in Aristotle's Metaphysics, edited by T. Scaltsas, D.  Charles, and M. L. Gill. Oxford: Clarendon Press, 1994. Lloyd, G. E. R. Magic, Reason and Experience. Cambridge: Cambridge University Press, 1979. Mates, Benson. "Diodorean Implication." Philosophical Review 58 (1949): 234–42. ———. Stoic Logic. Berkeley: University of California Press, 1961. McDowell, J. "The Role of Eudaimonia in Aristotle's Ethics." In Essays on Aristotle's Ethics, edited by A. Rorty. Berkeley: University of California Press, 1980,  359–376. Michael, Frederick S. "What Was the Master Argument of Diodorus Cronus?" American Philosophical Quarterly 13 (1976): 229–35. Mignucci, M. "  et Nécessaire." In Aristotle on Science: The Posterior Analytics, Proceedings of the Eighth Symposium Aristotelicum, edited  by E. Berti. Padua: Antenore, 1981. Moraux, P. "La joute dialectique d'apres le huitieme livre des Topiques." In Aristotle on Dialectic, edited by G. E. L. Owen. Oxford: Clarendon Press, 1968,  277–311. Nussbaum, M. The Fragility of Goodness. Cambridge: Cambridge University Press, 1986. Øhlstrom, Peter. "A New Reconstruction of the Master Argument of Diodorus Cronus." International Logic Review 21 (1980): 60–65. Owen, G. E. L. "Logic and Metaphysics in Some Earlier Works of Aristotle." In Aristotle and Plato in the Mid­Fourth Century, edited by I. Düring and G. E. L.  Owen. Göteborg: Almqvist & Wicksell, 1960, 163–90. ———. "Tithenai ta Phainomena." In Aristote et les problèmes de méthode, edited by S. Mansion. Louvain (1961): 83–103; reprinted in Aristotle: A  Collection of Critical Essays, edited by J. Moravcsik. Garden City, N.Y.: Doubleday, 1967, 167–90. ———. "Dialectic and Eristic in the Treatment of the Forms." In Aristotle on Dialectic, edited by G. E. L. Owen. Oxford: Clarendon Press, 1968, 103–25. ———. "Tithenai ta Phainomena." In Logic, Science, and Dialectic, edited by M. Nussbaum. Ithaca, N.Y.: Cornell University Press, 1986, 239–51. Owen, G. E. L., ed. Aristotle on Dialectic. Oxford: Clarendon Press, 1968. Owens, Joseph. The Doctrine of Being in the Aristotelian Metaphysics. Toronto, 3rd rev. ed., 1978. Pellegrin, Pierre. Aristotle's Classification of Animals. Translated by A. Preus. Berkeley: University of California Press, 1986. Popper, Karl. Conjectures and Refutations. New York: Harper Torchbooks, 1963. Poste, E. Aristotle on Fallacies. London: Macmillan, 1866. Priest, G. "Classical Logic Aufgehoben." In Paraconsistent Logic: Essays on the Inconsistent, edited by G. Priest, R. Routley, and J. Norman. München:  Philosophia Verlag, 1989, 131–48.

  

Page 235

———. ''Can Contradictions Be True?" Aristotelian Society Supplementary Volume 67 (1993): 35–54. Priest, G., and R. Routley. "The Philosophical Significance and Inevitability of Paraconsistency." In Paraconsistent Logic: Essays on the Inconsistent, edited by G.  Priest, R. Routley, and J. Norman. München: Philosophia Verlag, 1989, 483–539. Priest, G., R. Routley, and J. Norman, eds. Paraconsistent Logic: Essays on the Inconsistent. München: Philosophia Verlag, 1989. Prior, A. N. "Diodoran Modalities." Philosophical Quarterly V (1955): 205–13. ———. Time and Modality. Oxford: Oxford University Press, 1957. Prizl, Kurt. "Opinions as Appearances: Endoxa in Aristotle." Ancient Philosophy 14, no. 1 (Spring 1994): 41–50. Quine, W. V. O. "Two Dogmas of Empiricism." In From a Logical Point of View. Cambridge, Mass.: Harvard University Press, 1961; first ed., 1953. ———. "Epistemology Naturalized." In Ontological Relativity and Other Essays. New York: Columbia University Press, 1969, 67–90. Reeve, C. D. C. Practices of Reason. Oxford: Oxford University Press, 1992. Rescher, Nicholas. Studies in the History of Arabic Logic. Pittsburgh, Pa.: University of Pittsburgh Press, 1963. ———. "A Version of the 'Master Argument' of Diodorus." Journal of Philosophy 63 (1966): 438–45. Richardson, H. Practical Reasoning about Final Ends. Cambridge: Cambridge University Press, 1994. Rorty, A. O., ed. Essays on Aristotle's Poetics. Princeton: Princeton University Press, 1992. Rorty, R. Philosophy and the Mirror of Nature. Princeton: Princeton University Press, 1979. Ross, W. D. Aristotle. 5th edition. London: Methuen, 1949. Russell, B. The Problems of Philosophy. Oxford: Oxford University Press, 1959. Ryle, Gilbert. "It Was to Be." In Dilemmas. Cambridge: Cambridge University Press, 1954; reprinted 1962, 15–35. ———. "Dialectic in the Academy." In New Essays on Plato and Aristotle, edited by R. Bambrough. New York: The Humanities Press, 1965. Scaltsas, T. "Substantial Holism." In Unity, Identity and Explanation in Aristotle's Metaphysics, edited by T. Scaltsas, D. Charles, and M. L. Gill. Oxford:  Clarendon Press, 1994. Seel, Gerhard. "Diodore domine­t­il­Aristote?" Review de Métaphysique et de Morale 87 (1982): 293–313. Sim, May. "Senses of Being in Aristotle's Nicomachean Ethics." In The Crossroads of Norm and Nature: Essays on Aristotle's Ethics and Metaphysics, edited  by M. Sim. Lanham, Md.: Rowman and Littlefield, 1995. Slote, M. "Reply to Commentators." Philosophy and Phenomenological Research 54 (1994): 709–719. Smith, Robin. "Aristotle on the Uses of Dialectic." Synthese 96, no. 3 (1993): 335–58. ———. "What Use Is Aristotle's Organon?" Proceedings of the Boston Area Colloquium in Ancient Philosophy 9 (1993): 261–85. Solmsen, F. Die Entwicklung der aristotelischen Logic und Rhetorik. Berlin, 1929. Sorabji, R. "Body and Soul in Aristotle." Philosophy 49 (1974): 63–89. Sparshott, F. "Aristotle's World and Mine." In Aristotle Today, edited by M. Matthen. Edmonton: Academic Printing and Publishing, 1986. Stocker, M. Plural and Conflicting Values. Oxford: Clarendon Press, 1990.

  

Page 236

Stroud, Barry. The Significance of Philosophical Scepticism. Oxford: Clarendon Press, 1984. Stump, Eleonore. Boethius's De topicis differentiis. Ithaca, N.Y.: Cornell University Press, 1978. Sutula, John. "Diodorus and the Master Argument." Southern Journal of Philosophy 14 (1976): 323–43. Taylor, C. C. W. "Aristotle's Epistemology." In Epistemology, edited by S. Everson. Cambridge: Cambridge University Press, 1990, 116–42. Thomson, Judith Jarvis. "A Defense of Abortion." Philosophy and Public Affairs 1, no. 1 (1971): 47–66. Van den Bergh, Simon. Commentary on Averroes' The Incoherence of the Incoherence. In Averroes' Tahafut al­Tahafut. London: Luzac, 1954, vol. 2. von Fritz, Kurt. Die 

 bei Aristoteles. Munich: Sitzungsberichte der Bayerischen Akademie der Wissenschaften, 1964.

von Wright, G. H. "Truth, Negation, and Contradiction." lSynthese 66, (1986): 3–14. Vuillemin, J. "L'argument dominateur." Review de Métaphysique et de Morale 84 (1979): 225–57. Waterlow, S. Nature, Change, and Agency in Aristotle's Physics: A Philosophical Study. Oxford: Clarendon Press, 1982. Weil, E. "The Place of Logic in Aristotle's Thought." In Articles on Aristotle, vol. 1, edited by J. Barnes, M. Schofield, and R. Sorabji. London: Duckworth, 1975,  88–112. White, Michael J. "An S5 Diodorean Modal System." Logique et Analyse 22 (1979): 477–87. ———. "Facets of Megarian Fatalism: Aristotelian Criticisms and the Stoic Doctrine of Eternal Recurrence." Canadian Journal of Philosophy 10 (1980): 189– 206. ———. "The Necessity of the Past and Modal­Tense Logic Incompleteness." Notre Dame Journal of Formal Logic 25 (1984): 59–71. ———. "The Spatial Arrow Paradox." Pacific Philosophical Quarterly 68 (1987): 71–77. Wians, William. "Saving Aristotle from Nussbaum's Phainomena." In Essays in Ancient Greek Philosophy V, edited by A. Preus and J. Anton. Albany: SUNY  Press, 1992, 133–49. Wilkey, M. R., and S. H. Stephen. "The Exclusionary Rule: A Prosecutor's Defense." Criminal Justice Ethics 1, 1982. Abridged and reprinted in Readings in the  Philosophy of Law, edited by J. Arthur and W. H. Shaw. Englewood Cliffs, N.J.: Prentice Hall, 1984. Williams, H. Hegel, Heraclitus and Marx's Dialectic. Hemel Hempstead: Harvester Wheatsheaf, 1989. Winter, Michael. "Aristotle, hos epi to polu Relations, and a Demonstrative Science of Ethics." Phronesis 42, no. 2 (1997): 163–189. Witt, C. Substance and Essence in Aristotle. Ithaca, N.Y.: Cornell University Press, 1989. Zeleny, J. "Paraconsistency and Dialectical Consistency." From the Logical Point of View 3 (1994): 35–50.

  

Page 237

Index A absolute, 9, 70­71, 84, 90, 92, 101nn14,20, 109, 111, 117, 175, 191, 210, 212.  See also haplôs, simply, unqualified abstraction, 6, 10­11, 14n9, 15, 22, 27, 109, 120, 165, 171, 173­74 accident, 9­11, 14, 63, 75, 79, 87, 154, 184, 187­91, 194, 196, 199­200, 206n16, 209n21, 211n22, 225­26.  See also attributes act, x, xvi, 4, 8, 44, 109, 111, 113, 115, 117­18, 123n7, 124n10, 127, 153­54, 156­57, 159, 168, 173­76, 217, 223, 225, 227n4 action, xiii, 52, 96, 109­12, 114­18, 120, 122nn1,2­123nn3,7, 124n13, 157­59, 185, 202, 219­26n1, 227n4 activity, 108­9, 116, 118, 120, 147, 173­74, 226.  See also energeia actuality, 8­10, 16­17, 19, 21­23, 25, 28, 30, 151, 156­57, 186­87, 189, 194, 203, 205n13.  See also entelecheia affection, 63, 87.  See also pathos agent, 52, 94, 107, 109, 113­15, 119­22n1, 156­57, 219, 221, 227n4 akrasia, 85­88, 102, 165, 167, 173, 216.  See also weakness of will Alexander, 177n1, 178nn11­12 al­Farabi, 23, 32n11­33n12, 34n20, 37n51 al­Ghazali 32n2 alteration, xv, 157­59. See also change Analytics, 39, 52, 58, 62, 215 Anscombe, G. E. M., 21 Apology, 80 aporia, xiv­xv, 7­8, 13n6, 39, 59, 125­26, 128­30, 133­35, 136n4, 151, 153, 157, 185, 204n6, 216.  See also knot, lusis, perplexity, puzzle appearance, xviii, 8, 17, 19, 22­23, 28, 30­31, 37n48, 39, 41­43, 46, 49­50, 58, 89, 91, 103n34, 144­47, 153, 167, 170, 173, 189, 194, 196, 201­3.  See also phainomena, preserve, opinion apprehension, 116, 109, 164­65, 167­68, 171, 175­76, 177n7, 181n67 archê, 2­3, 7, 13n3, 82, 113, 197, 202, 211n26. See also principles, first, origin, source, cause argument:  constructive, 193, 207n19, 209n20; destructive, 6, 8­9, 193, 207n19.  See also logos, syllogism, sullogismos, reasoning, refutation, dialectic art, x, xii, xx­xxii, xxiv, 45­47, 49­51, 62­65, 80­82, 84, 103n29, 116, 127, 174, 195, 197­98, 206n16, 207n20, 212, 215, 217­18, 222, 226n1. S ee also technê, dialectic Atomist, 64, 104n43 attribute, 3­5, 8, 13n3, 133, 139, 155, 168, 176, 188, 209n21, 215, 225. See also accident authority, xviii, 63­64, 75­79, 89, 98, 102n26, 129­32.  See also expert, opinion, reputation

  

Page 238

axiom, 81, 168, 173­74 B Bäck, A., 35n39, 36nn40­41 Barnes, Jonathan, 99nn2,3, 206n16 being, xi, xvi, xx, 1­12n1, 13nn3,6­14nn7­9, 16­17, 19­20, 24­25, 29­32n8, 35n35, 76, 86, 90, 94, 111, 122n1, 132­33, 157, 166­67, 169, 173, 183­86, 188­ 91, 193, 195­96, 201­3, 221 Berti, E., 179n29 biology, 157­58, 173­74, 176, 217, 225 body, 115, 117, 131­32, 143, 157, 159, 189, 201­2, 221 Bolton, Robert, xi, xiii­xiv, 12, 101, 148n1, 177n4, 180n58, 210n21, 211­12nn27,28 Brunschwig, J., 99nn5­6, 100n6, 102n28­3n29, 122n3 Burnet, J., 100n9 Burnyeat, M., 104n39 C Categories, 6, 9, 141, 154­56, 172 cause, x, xviii, 2­3, 11­12n1, 14n7, 18­22, 24­27, 29, 33n12, 36n41, 64, 41, 45, 48, 110, 114, 118, 121, 124n11, 130, 143, 165, 171, 173, 176, 216­17, 221,  224­25;  final, 11, 143.  See also material, formal change, xii, xix, 7­8, 11, 16­18, 22, 27­28, 31, 32n8, 37n48, 42­44, 48, 52, 61 ,63, 69, 76, 144, 155­59, 175, 178n22; qualitative, 7, 11, 157.  See also alteration character, 119, 139, 147, 202, 217, 219­22, 224, 226, 227n4 chemistry, 152, 172 choice, xiii­xiv, xviii, xx­xxi, 107­21, 122n2­23n3, 124n11, 127, 185­86, 191, 193­94.  See also decision, prohairesis, deliberation, principles (of choice) choiceworthy, x, 109, 111­12, 114, 116, 118­19, 121, 122n3, 191 Cicero, 22, 26, 31, 35n39, 37n48 Cleary, J. J., 13, 185, 204n6 Code, A., 13n6 coherence, xvi, 15­17, 22, 26, 29, 60­61, 86, 88, 90, 97­98, 163, 185, 195, 200, 226n1.  See also opinion common, x, xii­xiv, xvi, xviii, xx­xiii, 2, 4, 7, 14n9, 40­41, 43­45, 47, 49, 50­51, 59, 74, 78, 80­86, 88, 92, 95­96, 103n30, 105n46, 129­32, 143, 153, 158, 164,  166­69, 172­74, 178nn25,28, 180n59, 184­86, 189­90, 191, 194­203, 204nn6,9, 206n16, 207n18, 208n20, 211n22, 212n27, 215­19, 223;  things, xiii, xxi, 81­85, 103n30.  See also opinion, ta koina, endoxa commonality, xx­xxi, 191, 196­97, 201, 203, 212n27 commonplaces, 197.  See also topics community, 78, 183­84, 186, 195­96, 198, 202­4n1; of persuasion, 195, 202­4; topical, 196, 213n31.  See also persuasion, rhetoric, philosophical composite, 220­21.  See also individual, particular concept, xv, 3, 14n9, 112, 122n1, 131, 138, 142­44, 149n14, 151­54, 156­58, 164­67, 172, 177n7;  analytical, 146, 151­54 conflict, 58, 61­62, 64, 66­68, 70, 72­78, 89­94, 97, 102n26, 105n46, 142­45, 147, 164, 173 consensus, 95, 138, 172 consequential facts (and consequent on), 84­86, 90, 96, 197­99, 212n27, 213n29. See also peirastic, examination, test contemplation, 169, 173, 196, 211n23;  life of, 111 contentious, xxii, 197. See also eristic contingency, 16, 18­20, 22, 24, 27, 31, 154­56 contradiction, xiv­xv, xxi­xxiii, 109, 111, 114, 118, 121­22, 137­42, 144­47, 164­65, 168­69, 176, 177n7 contrary, xii, xiv, xvi, 5­6, 8, 12n2, 20, 47, 77­78, 91, 102n26, 112­13, 115, 157­58, 160, 161n5, 209­10n21, 219, 221 Cooper, J., 105n46

  

Page 239

courage, 111, 118, 123n7 criterion, 153, 175, 191, 206n16, 222­23 D De Anima, xi, xv, 58, 62­63, 98, 104n39, 153,156­60, 174, 216, 224 De Caelo, 20, 34n20, 61, 63­65, 98, 165 De Generatione Animalium, 63, 100 De Generatione et Corruptione, 63, 104,152­53, 156­57 De Interpretatione, 140­41 De Sophisticis Elenchis, ix, xx­xxiii, 140­41, 144, 146, 148n13, 149n18, 196­98 de Ste. Croix, G. E. M., 226­27n2 decision, x, xiii, 109­10, 113­15, 118­19.  See also choice, prohairesis, deliberation deduction, 45, 49­50, 52, 54n20, 71, 99n6, 140.  See also sullogismos definition, xvi­xxi, xxiv, 1, 9, 14n9, 18­19, 23­27, 29, 33, 35n35, 36nn42,46, 37n48, 40, 44­46, 54n20, 62­63, 77, 83, 101n14, 103n29, 107, 112, 121­22, 131­ 33, 141, 183­84, 186­96, 198­211, 215­17, 219­26.  See also being, substance, ousia, essence, dialectic deliberation, xiii, 107, 122nn1­3, 123nn3,5, 204n8. See also principles (of deliberation), choice, decision, prohairesis DeMarco, C. W., 206n 16 demonstration, ix­xii, xiv­xx, 8, 12n2, 13n5, 39­40, 42, 45, 48­53, 54­55n20, 58, 63, 82­83, 88, 113, 135­36, 142­43, 164, 166­67, 175­76, 184­87, 190, 194,  204n7, 208n20, 215.  See also science, scientific, reasoning (demonstrative) Descartes, 136, 165, 177n7, 178n22, 180n53 desire, 87, 107­8, 122, 124n13, 130, 173 determinate, 9, 11, 13n6­14n7, 156 determinism, 16, 18, 20­21, 23­24, 26, 29, 31, 33n12, 34n17, 35n29 development, 217, 225­26, 227n3 Devereux, D., 160n2 dialectic, ix­xxv, 1­11, 12n2, 13nn5­6, 16­18, 20, 25, 28­29, 32n1, 33n16, 34nn17,20, 40, 42­53, 54n20, 57­86, 88­90, 92­95, 97­98, 113, 115­16, 119, 121,  124n11, 137­43, 145­47, 151­54, 156, 160n2, 163­69, 171­72, 174­75, 177n4, 178nn16­17,23,25,28­79nn29,35,38, 180nn58­59, 183­203, 204nn7,9­ 5nn10,16, 206n16­7nn17­20, 208­11nn20­23,26, 212n27, 213n31, 215­18, 223, 226; pure, 40, 169, 179n29, 185;. strong, 40, 104n39, 169, 179nn29,35, 185. See also reasoning (dialectical), art diction, 221 didactic, xxii­xxiii, 192, 197, 215 difference, 13n6, 112, 116, 119, 121, 123n4, 143, 167, 172, 179n44, 181n64, 183, 185­86,188­92, 194­96, 199­202, 205nn10,13, 208n20, 217, 223 differentia, 157, 161n6, 172, 188, 191, 217, 219­21 difficulty, xviii­xx, xxiii, 107­8, 112, 120, 130­31, 133­34, 141­43, 151, 164, 170, 178n10, 179n43, 191, 193­94 discourse, xv, xxiv, 169­70, 195­96, 198, 202 divine, 174 dynamis, xxi, 116, 121.  See also faculty, potential E education, 52, 121, 126, 168, 170, 201­4, 209n20.  See also upbringing egoism, 120 element, 63, 95­96 elenchus, xxiii.  See also dialectic Empedocles, 88, 218 empirical, xiii, 28, 39, 58, 61­65, 97­98, 101n14, 122n1, 152­53, 160, 166, 168, 178n23, 179nn36,38 empiricism, xvi empiricist, 98­99 end, ix­xi, xiii, xvii­xix, xxiii, 12n1, 107­10, 112­22nn1­2, 123n7, 124n13, 143­44, 194­95, 202­3,

  

Page 240

      222­24, 226.  See also goal, telos endoxa, xiii, xv, xvii, xxi­xxii, 13, 40­41, 44­47, 51, 58, 60­86, 89­93, 95­8, 99n2, 100n7, 101nn12,14, 102n24, 103n33, 104n39, 129,  131, 138, 142­43, 151­53, 160n2, 163, 166­69, 172­73, 175, 178n11, 183, 185, 196, 202, 206n16, 212n27. See also common, opinion, more (endoxa) energeia, 121, 156­57.  See also activity entelecheia, 10, 157.  See also actuality epagôgê, 163.  See also induction epic, 218, 220, 223­24 epistêmê, xi­xii, xxiii, 1­12nn1­2, 13nn3,5, 80­81, 83, 103n29, 104n38.  See also knowledge, science epistemic, x, xiii, xvii, 4­5, 7, 10, 52, 58, 83, 86, 89, 93, 98, 139, 144 epistemology, xi, 1, 17, 31, 41, 58, 60­61, 67,73, 85, 92, 98, 103n33, 137­38, 143, 171, 188 equivocation, 132, 151, 158 ergon. See function eristic, xiii, xxii­xxiii, 82­83.  See also contentious, dialectic eristikê. See eristic error, 4­7, 11, 170, 175­76, 203.  See also mistake essence, xviii, xxiii, 2, 5, 9, 13n7, 16, 22, 29, 41, 63, 87, 155­56, 170, 176, 183, 186­90, 196, 205nn11,15, 210n21, 221, 225.  See also unity, definition, substance, ousia, one eternal, 63, 77 ethics, xiii, 39, 45, 48, 52, 60, 104n39, 107, 112, 120­21, 124n12, 147, 149n19, 165, 167­69, 171, 174, 180, 191, 202, 225­26.  See also virtue, morality Eudemian Ethics, xi, xviii, 41, 43, 47­48, 50, 93, 95­96, 142, 164, 166 Evans, J. D. G., xi, xiv­xv, 12n2, 13n6, 205­6n16, 207n19, 209­10n21 exact, 66­67, 74, 79 examination, x­xi, xiii, xv­xvi, xviii, xxii, xxiv, 8, 47­48, 50­52, 54n18, 59, 68, 74, 78, 80­82, 84, 101n14, 103nn29­30, 140, 145, 153, 175, 180, 186­92, 197­ 201, 206n16, 207nn18,20, 208­9n20, 210n21­11n22, 212n27, 215­16, 219. See also peirastic, exetastikê, test, consequential facts exetastikê, 52, 80 existence, xv, 155, 159, 174, 177n7, 224­25.  See also being experience, xiv, 17­18, 22, 24­28, 30­31, 34n20, 35n31, 48, 61, 63, 65, 88, 97­98, 103n33, 134­35, 164­65, 167­69, 171­75, 202, 216 experiment, 192, 196 expert, x. xiii, 59, 67, 76­79, 91, 102n26, 126, 135, 145, 163, 165­68, 172­73, 180, 198, 200.  See also authority, opinion, reputable, opinions, reputation extension, 19, 24, 29, 34n20.  See also being extensional, 1­3, 7, 12n3.  See also being F faculty, xix, xxi­xxii, 49, 115­16, 121, 124n11, 145, 159, 197, 205n13.  See also dynamis fallibilism, xvi, 163­64, 168, 174, 177 false, x, xii­xiv, xix, xxii, 18­22, 29­30, 33­34n16, 36n46, 42­43, 50­51, 55n20, 58­60, 71, 84­86, 101n18, 105n46, 107, 130, 137­41, 144, 146, 167, 177­78,  180, 186, 193, 204n8, 197­98, 207nn19­20, 208n20, 211n24 fear, 220, 222 Ferejohn, M., xi, xiv­xv, 14n7, 148n1 first:  philosophy, xxiv, 187, 190, 194;  premises, 39­40, 50; principles, ix­xvi, 8, 40, 43, 51­53, 54n20, 57­59, 61­65, 67,75, 79­83, 90, 95­98, 142, 163­71, 173­77, 183­86, 192­96, 201­4, 206n16­207nn17,20, 208­ 9n20, 210n21, 211n24, 212n27, 215.  See also principles, primary focal, 191.  See also substance, sense form, xvii, 10, 15, 27, 121, 132, 140, 144­45, 151, 155­57, 171, 176, 193­94, 201, 205n15, 210n21,

  

Page 241

      213n31, 218­21, 227n3 formal:  cause, 11, 221; principle, 113, 222; formal­final, 2­3, 9, 11.  See also cause, principle foundation, 191, 204n6 foundationalist, 98 function, x, xvi, xxiv, 7­8, 43­44, 60, 112­14, 116, 120, 123, 139, 167, 173, 180, 199­200, 204, 206n16, 208n20, 221­22, 224, 227n3 G Garver, Eugene, xi, xiii, xiv, 148n1 generation, xiii, xvi, 8, 133, 145, 147, 154, 156, 183­84, 186, 189, 194­95, 200, 206n16, 208n20, 211n22, 217, 226 Generation of Animals, 133, 136 genos, 81, 83.  See also genus genre, 217, 219 genus, xx, 2­3, 101, 155­57, 176, 184, 187­91, 194, 197, 199­200, 205n14, 206n16, 209n21, 213n30, 217­20, 225.  See also genos goal, 40­42, 45, 50, 53, 116, 128, 143­44, 167, 197, 201, 203, 206n16.  See also end, telos good, x, xviii, xx, xxiii, 11, 52, 101n17, 102n25, 104n38, 107­16, 118­22nn1­2, 124nn11­12, 125­27, 144­45, 169, 173, 176, 184, 191­93, 196, 198­99, 202­3,  205­6n16, 210n21, 211n23, 217, 219­20, 222 Goodman, xi, xii, 175 greater, 111, 117, 129, 191, 194­95, 211n22.  See also lesser, contrary gymnastic, 45, 50, 59­60, 69­70, 79, 99­100n6, 101nn17,21, 102n21.  See also dialectic H Halliwell, Stephen, 226n1, 227n3 Halper, E, xi, xvi, xvii, 14n9 haplôs, 109, 114, 116, 118­20, 122.  See also simply, absolute, unqualified happiness, 48, 93, 110, 112, 117­18, 124n12, 196, 202­3, 226 Hegel, xv, xvii, 138, 142­43 Hempel, C., 175 hermeneutics, 166 hierarchy, 76, 155, 191, 210n21 Hintikka, J., 30, 32n8, 33n16, 164, 179n35 historia, 39, 216, 224­25 Historia Animalium, 57, 63, 216, 225 history, xv­xvii, 67, 170, 172, 215­17, 223­25, 226nn1­2, 227n3 Hume, 173 hupokeimenon, xx, 7­8, 87, 104n36 Husain, M., xi­xii, xiv, 205­6n16 hypothesis, xxi, 160, 193, 208n20 I idion, 81, 83, 103n3.  See also peculiarity ignorance, x, xiii, 18, 133­34, 136, 141­43, 158, 197­99, 203, 212n27, 213n29 imitation, 217­26, 227nn3­4 indefinite, 47, 49­50, 121 indemonstrable, 40, 50, 215 indeterminate, 9­10, 47 individual, 2­4, 8­11, 13nn4, 7, 26, 30, 42­43, 45­46, 49, 74, 76, 92, 96­97, 111, 118­19, 155, 168­70, 177, 216 induction, x, xv­xvii, 71­72, 99­100n6, 101n16, 163­68, 174, 180. See also epagôgê infallible, 170, 175­76 inherence, 155 intellectual, 108, 112, 119­20, 138­39, 143, 158, 167, 171, 173­75; virtues, 112, 119.  See also virtue, phronêsis, contemplation intelligible matter, 2, 9.  See also more (intelligible) intension, 24, 26­27, 34­35.  See also being intensional, 1­3, 9, 11­14nn3, 7.  See also being intuition, 25, 40, 108, 169, 172, 174, 178n22, 179n40, 206n16. See also noûs, mind Irwin, T., xviii, 13, 40, 43­45, 47­51, 100n10, 104n39, 148n4, 168­74, 178nn22­23, 25, 28, 179nn29,35­36,38­40, 43, 47, 180n47, 184­85,

  

Page 242

      206n16, 207­8n20 J justification, 40, 58, 60­68, 79, 85, 87, 89­90, 93, 95, 98, 100nn7, 9, 103n34 K Kant, I., 13n6, 104n39, 122n1, 138, 142­43, 145, 148n3, 173 kinêsis, 156­57. See also motion knot, 142­43, 185, 193.  See also aporia, lusis, perplexity, puzzle knowledge, x­xi, xiii­xxii, xxiv, 1­3, 5­8, 10­11, 12nn1­2, 13n6, 46, 49, 57­60, 63­64, 67­68, 71, 74­76, 80­90, 96­97, 102n23, 103n33, 104nn38­39, 43,  105n46, 109, 112­13, 116, 121, 124n12, 126­29, 131, 134­36n1, 138, 140, 143, 147, 152, 158­59, 160n5, 164­66, 168, 171, 173­74, 176, 185­88, 190, 192­ 94, 196­202, 205n13, 207nn16, 19, 208n20, 210n21, 211nn24, 27, 212nn27, 29, 213nn29­31, 215­16, 222­24. See also epistêmê, science, demonstration, truth L lesser, 117, 191, 194­95 likenesss, xvii, xix, 183, 185, 194, 209n21 limit, 131­32 logic, xiv­xv, 15­17, 19­20, 22­29, 31, 32n11, 33n16, 34n22, 35nn23, 39, 36nn39, 42, 37n48, 60, 86, 137­41, 145­47, 163, 165­66, 175­76 logos, xix, xxii­xxiii.  See also form, argument, reasoning, rationality, syllogism, deduction, sullogismos, definition lusis, 72, 185.  See also aporia, knot, perplexity, puzzle, M Marx, xv, 138, 143, 148 material, 29, 34n16, 36n46; cause, 14n7, 143.  See also matter, cause Mates, B., 33n16, 36n46 mathematical, 64, 165; sciences, 3­4, 64 mathematics, 3­7, 9­11, 13n5, 14n9, 127 matter, xviii, 1­2, 9­11, 14n8, 132, 143­44, 151, 156­57, 201, 213n31, 217, 219­21.  See also intelligible (matter), material Matthews, Gareth, xi, xiv, 125, 148n1  maxim, x, xiii, 113, 115.  See also reminder McDowell, J., 104n39 medievals, 164, 177 Megarian, 16­26, 28­32n1, 33nn12, 16, 34nn16­17, 35n23, 36nn42, 46 memory, 164­65 Meno, xiv, 125­26, 128, 130 Meno, 125­28, 130, 143, 149n15 Metaphysics, xi, xxiv, 1­11, 12n2, 13n6, 17, 32n2, 34nn18­20, 36n43, 39­41, 52­53, 57­58, 65, 88, 93­96, 104nn38­39, 119, 128­29, 135, 139­40, 142,  148n13, 156, 158, 169­70, 172, 174, 176­77, 178n22, 185, 190­93, 202, 204n6, 206n16, 209­10n21, 216, 219­21, 223, 225 metaphysics, xv­xvi, xxiv, 1­2, 4­5, 7­11, 12n2, 13nn5­6, 14n8, 15­16, 23­24, 26­29, 99n2, 138, 151­52, 154­58, 169­70, 183­84, 190, 195, 198, 204n6,  205n16, 219 meter, 217­19, 221 method, ix, xi, xiv­xvi, xxiv, 1, 3, 5, 7­8, 12n2, 13nn3, 6, 39­47, 49, 54n20, 57­62, 64­66, 69, 73, 79­80, 85, 88, 90­93, 95, 98, 99n2, 100n9, 102n25, 104nn39,  43, 116­17, 138­40, 142­43, 145, 147, 151­53, 156, 160nn2, 4, 163, 165­66, 168­72, 174, 176­77, 178n23, 179nn36, 38, 180nn47, 59, 185, 190, 200­201,  206n16, 210n21, 215­16, 226; analytic, 85­87, 89­90, 99n1, 103n33, 146­47.  See also dialectic, art, scientific methodology, 43­44, 57, 128, 129­31, 133, 136, 152, 159­60, 184, 215 methodos, 217 Michael, F. S., 33n16 middle term, 75, 112­14, 183 mind, x, xii, xiv­xxv, 15, 29, 31, 42­44, 125­26, 129, 164­65, 178n22, 189, 193, 209n2. See also noûs, intuition

  

Page 243

mistake, 59, 93, 111, 114­16, 119, 122, 143­44, 147, 152, 170, 175­76, 179n38, 181n67, 199­201, 203, 207n19, 212n29, 218­19. See also error modality, 16, 18­19, 21­24, 26­27, 29­30, 32n8, 34n20, 35nn35, 39, 36n42 model, 107, 175 monism, 16­17, 28­29, 31, 34n17 morality, 48, 52, 107­9, 112­15, 120­22nn1­2, 136, 168, 170, 226.  See also virtue, ethics more:  endoxon, 68­70, 72­77, 95; intelligible, xiii, 69­70, 73­76, 78, 91­93, 95­97, 101nn14, 17, 102n23 motion, 82, 131, 133, 190, 213n30, 225.  See also kinêsis N nature, ix, xi, xiii, xix­xx, 2­4, 8­9, 12, 13n3, 14n7, 18, 22­23, 26­27 32n2, 33n12, 34nn17, 22, 35n35, 37n48, 51, 61, 63­64, 68, 75, 81, 83, 87, 90, 92­96,  103n30, 108­9, 115­17, 119, 121, 124nn11­12, 125­26, 130­31, 133­34, 138, 152, 155­58, 165­68, 171, 174, 176­77, 193, 203, 217, 220­21, 224­25,  226n2­27n3; by nature, 90, 94­96, 109, 119, 124n11, 193, 203, 226. See also physis necessary (and necessity), xv, xvii, xx, xxii, xxiv, 16, 18­24, 26­30, 32n11, 33nn14, 16, 36nn39, 41­42, 40­41, 44­45, 51, 61, 63­66, 71, 73, 75­76, 79­81, 83­ 85, 87­88, 90, 94­95, 98, 109, 112, 119­20, 132, 136, 140, 142­43, 145­46, 154­55, 160, 175­76, 178n25, 187, 197­99, 203, 204n7, 211­12n27, 216­17,  220, 223­25 Newton, 173 Nicomachean Ethics, xi, xviii, xxiv, 20, 40­41, 43­44, 54, 76­77, 85, 87, 93, 96, 102n25, 107­15, 117­19, 121­22, 124nn11­12, 128­29, 149n16, 160n2, 163,  175­76, 180n55, 202, 204nn1, 8, 206n16, 211n23, 216 Nietzsche, 171 normative, 222 noûs, xvi, 164­65, 168, 174­76, 178n22, 179n40. See also mind, intuition, fallibilism, infallible Nussbaum, M., 103n34, 104nn39, 41­42, 148n4, 166­67, 175, 180n53 O objectivity, 51­52, 76, 89, 138, 168­69, 175, 180n59, 183­86, 189, 195 Øhlstrom, P., 33­34n16 one, xvi, xix­xxi, xxiii, 4, 12n2­13nn3, 5, 14nn7­8, 132, 184, 187­90, 193­96, 200, 203, 206n16, 209n20, 211nn22, 24, 219­23, 225.  See also unity, one and many, substance, ousia one and many, 3, 8­9, 13n5 onta, 2­3, 7­9, 12n2, 154 ontological, 1­2, 12n2, 14n9, 154­55, 206n16 ontology, 139­40, 205­6n16 operator, 140, 144­45, 147 opinion, ix, xii­xvi, xviii­xix, xxii, 39­45, 47­48, 50­52, 58, 61, 72, 74­75, 77­78, 83­84, 86­87, 89, 92, 95­96, 102n26, 104n35, 128­33, 145, 163, 165, 166­68,  173­74, 183­90, 194­98, 200­203, 204n6, 206n16, 207­9n20, 211n22, 215­17; coherent, 200 reputable, 166­67, 174, 189, 204n6.  See also endoxa, common, coherence, reputation, preserve organization, 168­69, 172­73, 178n28, 221­22, 224 origin, 202, 224.  See also archê, principles, source, cause orthos logos, 113.  See also logos, rationality ousia, xi, xii.  See also substance, essence, form, being Owen, G. E. L., xv, 35, 39­41, 43­44, 51­53, 85­90, 99n5, 103n33, 151­53, 165­66, 184, 204n2, 206n16, 207­8n20 P paradox, xiv, 39, 47, 77, 102n26, 141, 146­47, 149n15.  See also aporia,

  

Page 244

      knot, lusis, perplexity, puzzle Parmenides, 16­17, 24­25, 29, 31 32n4, 37n52 paronymy, 9, 11, 14nn7­8, 172 part, x­xi, xiii­xvi, xix­xx, 1­3, 5­7, 9, 11, 13n3, 59, 63, 65, 79­80, 85­87, 90, 93, 95, 112, 117, 120, 122­23n3, 131­32, 141, 146, 152, 154­55, 158, 166­67,  172, 185­86, 195­96, 200­202, 206n16, 207n18, 209nn20­21, 217, 220­23, 225 particular, x, xiii, xvi, xxii, 3, 23, 26, 30, 35, 46, 49, 60, 73­74, 82­83, 93, 97, 109, 111, 113, 119, 120­21, 123n3, 124n13, 139­41, 143­47, 163­65, 167­68,  180, 183, 188, 191, 197, 201, 203­4, 210­11, 216­17, 223­24, 226.  See also individual Parts of Animals, 36n44, 86, 216, 220 pathos, 87, 104n36.  See also affection peculiarity, 49, 126, 164, 180n59, 187­88, 191, 196­97, 199­200, 217.  See also idion peiras, xxii, 192, 196. See also peirastic, exetastikê, examination peirastic, x, xiii, xxii­xxiii, 7­8, 12n2, 57, 68, 80­87, 89­90, 92­94, 96­98, 103nn29­30, 105nn46, 48, 196­98, 200, 207n18 peirastikê. See peirastic pêras, 132.  See also limit perception, xiii, xv, 17, 19, 61­64, 74, 90­93, 95­99, 101n14, 104n41, 108, 114, 134, 153­54, 156­59, 164­67, 170­71, 173, 176, 200, 206n16, 212n27. See also sensible Pericles, 171 perishable, 63.  See also change permanence, xix, 87, 191, 194, 201 perplexity, x­xi, xiv, 125­36, 193.  See also aporia, knot, lusis, puzzle persuasion, ix­xiv, xvi, xxii­xxiv, 42­46, 48, 53, 70, 83, 116, 121, 167­68, 184, 195­96, 201­4, 206n16.  See also community, rhetoric, change, mind, dialectic phainomena, 39, 62­64, 72, 89, 91, 93, 95, 101n18, 104n41, 129.  See also appearance, preserve phantasm, 164 phenomena, xiii, xv, 64, 138­39, 152­53, 160, 163, 165­68, 172, 174, 216 phenomenalism, 170, 173 philosopher, xiv, 2, 5, 8­9, 76­77, 91, 103n29, 119, 127, 138­39, 143, 164, 170­74, 179n40, 191­93, 206n16, 207n18 philosophical:  community, 184, 195­96, 198, 203; sciences, 50, 64, 167, 193 phronêsis, xxi, 108­9, 112­14, 116­17, 119­22.  See also practical, wisdom, wise, choice physical, 2, 9, 64, 152, 155­56, 160n4, 169, 173.  See also material, matter Physics, xi, xviii, xxiv, 39­43, 58, 64­65, 75, 94­97, 101n14, 131­33, 136, 152, 157, 165, 167, 174, 176, 178n22, 180n58, 203, 206n16, 216, 224­25 physis, 2­3, 9, 11, 12n3, 13­14n7, 225.  See also nature pity, 220, 222 place, xviii, 41, 64, 131­32, 144, 167, 174, Plato, 16­17, 34n20, 37n48, 58, 67, 80, 89, 99, 103n29, 104n43, 121, 130­31, 134­35, 142­43, 166, 168, 170, 176, 178n22 Platonic, xviii, 155, 160, 170, 175, 195 pleasure, 87, 120, 196, 198, 203, 224 plot, 217, 221­25 plurality, 1, 3, 12n2, 190, 221 poêsis, 218 Poetics, xvi, 215­17, 221­23, 226, 227n3 poetry, xvi­xvii, 215­20, 223­26 poion, 1, 6.  See also quality political life, 108, 111, 115 Politics, 77, 93, 108, 111, 118, 121, 124n12, 204n1, 216 politics, xxi, 8, 115, 120­21, 168, 196, 225­26 Popper, K., 175, 180n56 Porphyry, 176 poson, 1, 3­7, 9­11, 13n5.  See also

  

Page 245

      quantity possibility, ix, xi­xii, xvii­xviii, xx­xxi, 4, 10, 13n6, 16, 18­26, 28­30, 32nn2,8, 33nn14,16, 34nn16,20, 36n42, 53, 67­71, 73, 75­76, 79, 81­84, 95, 100n6,  104n39, 108, 110, 112, 115, 122,124nn10­11, 128­30, 132, 135, 141­42, 145­46, 152, 159­60, 165, 171, 174, 184, 187­88, 194­97, 201, 203­5n13, 209n20,  211n22, 221, 223­24, 226 posterior, 12, 74, 190, 209n2l Posterior Analytics, 40, 51, 63, 65, 75, 81, 84­86, 91, 97, 136n4, 163­66, 175­76, 177n7, 178n22, 181, 205n12, 208­9n20, 210n2l, 216 potentiality, 2, 8­10, 14n9, 16­17, 20, 24­25, 28­29, 31, 32n2, 151­52, 156­58, 195, 226.  See also dynamis, faculty practicable, 109, 114, 116, 122 practical, x, 63, 85, 88, 102n25, 108­9, 111­17, 119­22, 123n7, 194, 199; reasoning, 114­15 predecessors, 39, 44, 144, 174, 176, 190 predicables, 184, 189, 194, 199­200, 208n20, 211n22 predicate, 187­88, 191, 205nn12,14, 211n22 predication, xiv, 139, 141, 145, 154­55, 161n6, 187­88, 191 preserve, xviii, 41, 89­90, 109, 129, 166, 204n6.  See also opinion, appearance, phainomena presupposition, 86, 156, 176, 187, 202, 204, 212n27, 223­26 Priest, Graham, 146 primary, ix, xii, xv­xvii, xx, 2­5, 9­12, 13n7, 45, 52, 62, 65­66, 131, 152, 155, 164, 166, 186, 190, 193­94, 222; principles, xvi, xx, 164, 194.  See also principles Principle of Non­Contradiction, 139­42, 144­45, 147 principles, ix­xxiv, 8, 21, 26, 28, 31, 35n23, 36n44, 40, 43­44, 49, 51­ 53, 57­59, 61­65, 67, 73, 75, 79, 81­86, 90­92, 95­98, 101n14, 111­ 21, 123nn5,8,  135­36, 138­42, 145, 156, 159, 163­71, 173­76, 177n7, 180nn58­59, 183­86, 192­204, 206n 16, 207nn 16­17, 20, 208­9n20, 210n2l, 211nn24, 26­27,  212n27, 215­17, 221­22; general, xvi, 163, 174; of choice, 111­16, 120­21; of deliberation, 112, 123n5; of interpretation, 113; of unity, 221­22; proper, xx.  See also first (principles), formal (principles), archê, dialectic, primary, special (principles) prior, xi, xvi, 4­5, 7­8, 12, 74, 97, 101nl4, 159, 201, 215, 217 Prior, A. N., 21, 32n11, 33n16 Prior Analytics, 39, 45, 62­65, 97, 102n22, 104n4l, 140­41, 165, 176­ 77nl, 216 priority, xi, 4­11, 12n2, 14n9, 51, 91, 98, 190, 201, 209n2l prohairesis, 107­8, 116, 123n4.  See also choice, decision, deliberation proof, xviii, xx, 40­41, 43­44, 46, 50­ 51, 55n2O, 71, 82­83, 93, 95­96, 129, 134, 136, 140, 193, 196, 204n6, 206n16, 207­8n20.  See also demonstration, science property, 2­4, 9­11, 13n3, 14nn8­9, 52, 170, 172, 184, 187, 188­91, 194, 198­200, 205n12, 206n16, 209nn2l­21, 211n22, 212­13n29.  See also propria proposition, xix­xx, xxiii, 45, 50, 52, 54n20, 61, 73, 76­77, 82, 100nn7,9, 102n22, 113, 123n3, 146, 153, 165, 184­87, 190, 194, 199­200, 204­5n10, 205n13,  208n20 propria, 172, 198.  See also property pros, xix, xxiii, 42­43, 51, 109, 122n2, 192, 207n20 pros hen, 2­11, 13n5, 14n7.  See also focal, one Protagoras, 80 Protagoras, 17, 170 Putnam, 175 puzzle, xi, xviii, xxiv, 39, 41, 43­44, 50­51, 84, 128, 130­31, 134, 143, 157, 163­64, 167, 174, 179n35,

  

Page 246

       187, 190, 194, 207­9n20, 217, 221, 225.  See also aporia, knot, lusis, perplexity Pythagorean, 165 Q qualification, xiv­xv, xviii, 50, 70, 73­74, 76, 84, 87­88, 90, 101n13, 105n46, 140, 144­46 quality, xvii, 1, 11, 13n7, 115, 155, 157­58, 188, 191, 205n11.  See also poion, change (qualitative) quantity, 1, 3, 5, 205n 11.  See also poson Quine, W. V. O., 175 R rationality, 15, 43, 48, 59­60, 91, 94, 107­8, 112, 114, 117, 121­22nn1­2, 124n13, 137­39, 142­43, 145, 147, 169­70, 186, 196, 205n12 real, 15­19, 21­28, 30­31, 34n20, 35n39, 37n48 reasoning:  demonstrative, ix, xvi­xvii, 186; dialectical, ix, xiv­xv, xvii, xix, 66­70, 72, 75, 147, 168, 185­88, 198, 203.  See also practical (reasoning), rationality, logos, argument, syllogism, deduction, sullogismos refutation, xii, xxi­xxii, 42, 48­50, 52­ 53, 54nn18,20, 55n20, 81­82, 86, 140­41, 152, 196­97, 209n20, 216. See also argument regress, 132, 135, 160 reminder, x­xi, xiii, 115­16, 118, 122n2, 123n9, 199­200.  See also maxim reputation, 58, 126, 166­67, 174, 189, 204n6.  See also opinion, expert, authority Rescher, N., 33n16, 37n48 rest, xxiv, 17, 131, 133, 137, 190 Rhetoric, xvii, xix­xxi, 20, 59, 67, 74, 82­83, 92, 94, 103n30, 107, 109­11, 113, 116, 119, 121, 122n3 rhetoric, x, xii, xiv, xvii, xix­xxi, xxiv, 45­46, 48, 49, 59, 74, 79, 82, 10n30, 113­16, 121, 122­23nn3, 9 146, 184, 196, 201­4, 206n16.  See also persuasion, opinion, community rhythm, 217, 221 Rorty, A., 226 Rorty, R., 179n42 Ross, W. D., 57­58, 63, 99n1, 103n29 rules, 46, 65, 69, 72, 88, 90­91, 113­14, 119, 122, 191­92, 196 Ryle, Gilbert, 185, 204n5 S sameness, 186, 188­90, 195­96, 199­201, 205nl4, 209n20 science, xiii, xv, xix­xxiii, 5, 12n2­ 13nn3,6, 14n9, 15, 18, 24, 26­27, 39­40, 44, 46, 49­52, 58­65, 67, 75­76, 79­82, 84­88, 90­94, 96­98, 99n2, 104nn39, 43,  105n48, 113­14, 119, 121, 128, 135, 142, 147, 152, 160, 163­80, 190, 192­93, 195, 197­99, 202, 207n19, 210n21, 211­12n27, 213n30, 216­ 18, 223­26nn1­ 2, 227n2.  See also knowledge, demonstration, special epistêmê, philosophical, scientific:  demonstration, 39­40, 50, 88; method, 61­62, 65, 91­92, 104n43, 147, 160 self­sufficiency, 121, 191, 203 semantics, 154­55 sensation, 122n1, 166­67, 170, 172, 189, 205n13, 216 sense (meaning), ix­xi, xiv, xvi, xviii, xxii­xxiii, 1­4, 8, 18­19, 21, 24, 27, 29­31, 49, 51, 53n3, 54n18, 63, 66, 68, 70­71, 74, 76, 78, 84­85, 90, 100n7,  101nn17,20, 104n39, 108­14, 116, 120­21, 122n3, 132, 143­44, 152, 159, 164, 171, 174, 177nl, 183, 185­90, 192, 198, 200, 203, 205nn13­14, 206n16­ 7n17, 209n21, 219, 221, 224, 226; common sense, 14n9, 122n1 sense (faculty), 135, 154, 164­67, 171, 173, 176 sensible, 63, 157, 159, 166, 190.  See also perception separate, 4, 109, 112, 122, 154­55, 172, 174, 178­79, 185, 205n16, 210n21, 220­21 Sim, May, xi, xvi, 27, 148n1, 211n23 simply, 3, 74­75, 88, 90­92, 95­97, 109, 144, 146.  See also haplôs,

  

Page 247

     absolute, unqualified skepseôs, 192, 196 Slote, Michael, 122 Smith, Robin, xi­xiii, xviii, 204n2, 206n16, 207­9n20 Socrates, xiv, 125­28, 130­31, 133­34, 136, 165, 171, 174, 218 song, 217, 219, 221­22 Sophist, 131 sophist, xxiii, 2, 57, 68, 77, 116, 121, 171, 199  sophistical, xiii, 5, 42, 82, 194, 196, 209n21 Sophisticis Elenchis, ix, xvii, xx­xxiii, 12, 47, 49­50, 66, 68­70, 73, 75­77, 79­85, 96, 102n21, 103n29, 140­41, 144, 146, 148n13, 149n18, 175, 180n58, 196­ 98 soul, xviii, 143, 171, 175, 189, 196, 201, 212n29­13nn30, 32, 221, 224 source, xxi­xxii, xxiv, 40, 44, 61­62, 82, 87, 130, 138, 188, 196­97, 203, 223­24.  See also archê, origin, cause, principles special:  principles, 84­85, 96, 197­98, 211­12n27; sciences, xx, xxii, 2­4, 13n3, 169, 175, 190.  See also science, principles species, xii, 2, 13n3, 155­56, 173, 176, 198­99, 205n14, 213nn30­31, 218­20 spectacle, 217, 221­22 speech, 125, 217, 220­22 Spinoza, 172 standard, xiv, 70­71, 77, 112, 159, 163, 170, 175, 179n43, 186, 191, 200 starting point, 48, 50, 63, 99n2, 190, 194­95, 200.  See also archê, origin, principles, source, cause Stoics, 171 structure, 168, 172, 175­76, 180, 188­ 89, 195, 205n14, 206n16, 221­22, 226 Stump, E., 205n16 subjective, 143, 189 substance, 2­11, 12n2, 13nn4, 7­14nn7­9, 87, 155, 158­61, 169­71, 183, 187, 190, 205nn11,15­206n16, 209n21.  See also ousia, definition, essence, form sufficient, x, 41, 43, 58, 60­62, 64, 67, 78, 93, 175, 198­99, 202, 212 sullogismos, 45, 54n20, 82, 99­100n6.  See also deduction, rationality, reasoning, syllogism Sutula, J., 33nl6 syllogism, xxi, 45, 52, 66, 71, 75, 116, 123n7, 140, 146, 164, 166 T to koina, xxi, xxiii, 185, 197, 199­200.  See also common things, common taxonomy, 155­57 technê, xxi, 7­8, 12n2, 112­13, 205n16.  See also art telos, 128, 122n2.  See also end, goal test, x, 59, 62, 64, 68­70, 73, 80­81, 83­84, 97, 115, 135, 155, 160, 161n7, 163, 175, 197­200, 211nn22, 27­212nn27,29.  See also peirastic, examination the majority, 58, 66, 74, 77, 166, 173, 185, 198 the many, xii, xvi, xviii, 41­44, 46­48, 184, 194­95, 200­201, 209n20, 211n24 the wise, xii, 41­42, 44, 46­48, 58, 60, 74, 76­78, 105n46, 166, 198, 201 Theaetetus, 130­31, 134 Theaetetus, 134­35 theôria, 108, 115, 118.  See also contemplation theory, xiii, xv, 7, 12nl, 46, 50, 52, 60­64, 77, 79, 85­86, 96­98, 104nn35,43, 109, 114, 127, 129, 134, 136, 138, 140­43, 145, 147, 149n19, 154­55, 157­58,  163­66, 170­76, 194, 210n21 time, 16­17, 19­22, 25­26, 28­31, 33nn14,16­34n17, 36n46, 37n48 Tlumak, J., 36 Topics, ix, xiii, xvi­xxiv, 40, 42­46, 49­54, 57­62, 64, 66­71, 73­77, 79­80, 83­84, 90, 95, 97, 99n6­100nn6­7, 101nnl4,18, 102n23, 103n29, 105n46, 107­21,  122­23nn3, 4, 142, 145, 153, 160, 164, 166­67, 176­78, 183­96, 198­99, 202­3, 204n6, 205nn11­12,14,16

  

Page 248

      206n16, 207­9n20, 209­11nn21­ 22, 25, 215­223 topics, x, xiii­xiv, xvi, xx­xxiv, 67, 73, 108­9, 111­14, 116­17, 120­21, 123nn7, 9, 131, 148n1, 156, 183­86, 189, 191, 194­203, 204n9, 206n16, 209n20,  211n22, 212n27, 213n31, 217, 219, 221 topoi, 49­50, 62, 101n14, 131, 184, 193­94, 197 to pragma, xxiii, 187, 189, 199­200 tradition, 61, 81, 83, 92, 101n16, 164, 171, 174, 179, 192 tragedy, xvi­xvii, 215, 217­18, 220­27 truth, ix­xx, xxii­xxiv, 1, 3, 6, 8­9, 12nn1­2, 15­16, 18­22, 24­26, 28­ 31, 33nn14, 16­34nn17, 20, 36n46, 39­43, 45, 49­53, 58­60, 63, 65, 67, 69, 71, 75, 77,  81­86, 88­90, 92­95, 97, 101n18, 102n22, 105n46, 111, 113, 116, 124nn12­ 13, 137­40, 142, 144­47, 152­53, 154­57, 159­60, 166­69, 173, 175­ 76,  177n7, 178n11, 180n56, 184­ 89, 193­96, 198­203, 204nn6, 8, 206n16, 207nn18­20, 208­10nn20­ 21, 211n24, 212n27, 213n31, 222, 224 U ultimate, xv, 63, 65, 89, 101n14, 103n34, 154, 171, 176, 193­94, 203, 206n16 unity, xii, xvii, 2, 4­5, 14nn7­8, 29, 31, 34n17, 37n48, 176, 183, 221­24, 226; organic, 223.  See also one, principles (of unity) substance, ousia universal, xx, 3, 8, 15, 20, 29­30, 36, 46, 59, 67, 71­72, 78, 97, 104n39, 108, 119­20, 140, 142­43, 145, 155, 164­65, 167­68, 187, 194, 197­98, 201, 203­4,  209­ 110n21, 216­17, 223­24, 226n2 unqualified, 140, 144­47, 154.  See also absolute, haplôs, simply upbringing, 202­3.  See also education V virtue, 85, 109­10, 112­15, 117­22, 123n7, 124n12, 125­28, 133, 169, 173, 226; ethical, 109, 112­13, 120; moral, 112, 114, 120­21.  See also ethics, morality, intellectual Vuillemin, J., 34n20 W weakness of will, 44, 129­30, 136n3.  See also akrasia Weil, E., 206n16, 207n18, 211n21 White, M., 30, 32n8, 35n23, 37n48 whole, ix, xvi, xxiii, 83, 93, 95­96, 132, 185­86, 189, 195­96, 201, 206n16, 222­23 Wians, William, 178 wisdom, ix, xix, xxi, xxiv, 2, 7, 58, 76, 78, 85, 88, 102n25, 121, 192­93, 208n20, 211n24 wise, 58, 60, 74, 76­78, 102n25, 105n46 wish, 107, 110, 112, 115, 118, 122, 128, 142 Witt, C., 14n8 Wittgenstein, L., 126, 131, 136n2, 172 wonder, xiv, 7, 52 Woods, M. J., 42­43, 48, 54n 18 wrong. See error, mistake Z Zeno, 45, 82, 132

  

Page 249

About the Contributors Allan Bäck is professor of philosophy at Kutztown University. He publishes widely in ancient and medieval philosophy and in the philosophy of logic. His publications  include On Reduplication (Leiden: E. J. Brill, 1996) and Aristotle's Theory of Predication (Leiden: E. J. Brill, 1999, forthcoming). He is a recipient of the Alexander  von Humboldt Stiftung Forschungspreis. Robert Bolton is professor of philosophy at Rutgers University. He was educated at Princeton, Oxford, and Michigan. He is author of Science, Dialectique et  Ethique chez Aristote (Paris: Klincksieck, 1999); co­editor with R. Smith of Logic, Dialectic and Science in Aristotle (Ancient Philosophy Special Issue, 1994),  and co­editor with F. Lewis of Form, Matter and Mixture in Aristotle (Oxford: Oxford University Press, 1996). He has also authored numerous articles on ancient  Greek philosophy and science. He is currently engaged in a multivolume project on Aristotle on the varieties of human knowledge. J. D. G. Evans has been professor of logic and metaphysics at the Queen's University of Belfast, Northern Ireland since 1978. His education and earlier career was  at Cambridge, England. He has published extensively in ancient philosophy (especially Plato and Aristotle), modern philosophy (especially Berkeley and Kant), moral  philosophy, and teaching philosophy. His current projects include a historical dictionary of ancient Greek philosophy (Scarecrow Press) and a project establishing an  outline curriculum for teaching introductory philosophy with special reference to Africa (supported by UNESCO). Michael Ferejohn is associate professor of philosophy at Duke University. His publications on Aristotle's logic and metaphysics, and Socratic ethics have appeared  in such journals as Phronesis, American Philosophical Quarterly, Journal of the History of Philosophy, Archiv für Geschichte der Philosophie,

  

Page 250

The Review of Metaphysics, and Philosophy and Phenomenological Research. He is author of The Origins of Aristotelian Science (New Haven: Yale University  Press, 1991). He is currently at work on a book on Aristotle's metaphysics. Eugene Garver is Regents professor of philosophy at Saint John's University. He is the author of Machiavelli and the History of Prudence (Madison: University of  Wisconsin Press, 1987), and Aristotle's Rhetoric: An Art of Character (Chicago: University of Chicago Press, 1994). He is currently engaged in showing the  implications of the Rhetoric for the Ethics, especially concerning the relation between thought and character. Lenn E. Goodman is professor of philosophy at Vanderbilt University. He is a philosopher and scholar of Jewish and Islamic philosophy, and has written  interpretative studies of many of the major Islamic and Jewish philosophers as well as original work in Jewish and general philosophy. Winner of the Baumgardt Prize  of the American Philosophical Association, he served as the Jewish philosophy editor for the 10­volume Routledge Encyclopedia of Philosophy and has published  over a hundred philosophical articles and reviews. Goodman has lectured in Israel, Britain, Portugal, France, and throughout the United States. His philosophical  books include On Justice (New Haven: Yale University Press, 1991) and God of Abraham (Oxford: Oxford University Press, 1996), which won the Centennial  Prize of Gratz College in Philadelphia. His latest book is Judaism, Human Rights and Human Values (Oxford: Oxford University Press, 1998). His new book  Jewish and Islamic Philosophy: Crosspollinations in the Classic Age is forthcoming in 1999 from Edinburgh University Press in Britain and Rutgers University  Press in the United States. He is presently at work on new books for Routledge and for Oxford University Press, and on a philosophic book on truth for Humanities  Press. Edward C. Halper is professor of philosophy at the University of Georgia. He is the author of One and Many in Aristotle's Metaphysics: The Central Books  (Columbus, Ohio: Ohio State University Press, 1989), Form and Reason: Essays in Metaphysics (Albany: SUNY Press, 1993), and numerous articles in  philosophy books and journals. He is currently working on a commentary on books B and G of Aristotle's Metaphysics. A prominent research interest of Halper's is  how Aristotle's metaphysical conceptions manifest themselves in other Aristotelian works. Martha Husain is associate professor of philosophy at Brock University. Her main area of interest and publication is Aristotelian metaphysics, with a particular focus  on his distinctive concept of being. In the last few years, this has been applied to Aristotle's theory of art, and a book­length study, An Approach

  

Page 251

to Aristotle's Poetics, is nearing completion. Husain is also co­editor of The Journal of Neoplatonic Studies. Gareth Matthews is professor of philosophy at the University of Massachusetts, Amherst. With S. Marc Cohen, he translated Ammonius: On Aristotle's  Categories in the series, Ancient Commentators on Aristotle, edited by Richard Sorabji (London: Duckworth, 1991). He is the author of Thought's Ego in  Augustine and Descartes (Ithaca, N.Y.: Cornell University Press, 1992) and the editor of The Augustinian Tradition (Berkeley: University of California Press,  1999). His most recent book is Socratic Perplexity and the Nature of Philosophy (Oxford: Oxford University Press, 1999). May Sim is associate professor of philosophy at Oklahoma State University. She was educated at the University of Iowa and Vanderbilt University. Her research  interests have focused on Aristotle's dialectic, metaphysics, and ethics, and the relation between the Metaphysics and Ethics. She is the editor of The Crossroads of  Norm and Nature: Essays on Aristotle's Ethics and Metaphysics (Lanham, Md.: Rowman and Littlefield Publishers, 1995). Her current research project is a  book­length comparison of Aristotle and Confucius on the ethical life. Robin Smith is professor and head of the Department of Philosophy at Texas A&M University, where he has been since 1994. His publications are principally  concerned with Aristotle's logic. Prior works include Aristotle, Topics I, VIII and Selections (Oxford: Clarendon Aristotle Series, 1997) and Aristotle's Prior  Analytics (Indianapolis, Ind.: Hackett Publishing Company, 1989).