Aristotle on the Disciplines of Argument: Rhetoric, Dialectic, Analytic
Author(s): James Allen
Source: Rhetorica: A Journal of the History of Rhetoric, Vol. 25, No. 1 (Winter 2007), pp. 87- 108 Published by: University of California Press on behalf of the International Society for the History of Rhetoric Stable URL: http://www.jstor.org/stable/10.1525/rh.2007.25.1.87 . Accessed: 21/09/2014 02:48 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp . JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org. . University of California Press and International Society for the History of Rhetoric are collaborating with JSTOR to digitize, preserve and extend access to Rhetorica: A Journal of the History of Rhetoric. http://www.jstor.org This content downloaded from 192.167.204.6 on Sun, 21 Sep 2014 02:48:48 AM All use subject to JSTOR Terms and Conditions James Allen 87 Rhetorica, Vol. XXV, Issue 1, pp. 87108, ISSN 0734-8584, electronic ISSN 1533- 8541. 2007 by The International Society for the History of Rhetoric. All rights re- served. Please direct all requests for permission to photocopy or reproduce article content through the University of California Presss Rights and Permissions website, at http://www.ucpressjournals.com/reprintInfo.asp. DOI: RH.2007.25.1.87. Aristotle on the Disciplines of Argument: Rhetoric, Dialectic, Analytic Abstract: According to an argument made by other authors, analytic the formal logical theory of the categorical syllogism expounded in the Prior Analyticsis a relatively late development in Aristotles thinking about argument. As a general theory of validity, it served as the master discipline of argument in Aristotles mature thought about the subject. The object of this paper is to explore his early conception of the relations between the argumentative disciplines. Its principal thesis, based chiey on evidence about the relation between dialectic and rhetoric, is that before the advent of analytic dialectic playeda double role. It was both the art or discipline of one practice of argumentation and the master discipline of argument to which other disciplines turned for their understanding of the fundamentals of argument. A s everyone knows, Aristotle invented logic. The Prior Analytics tackles the question when and in virtue of what an argument is valid in an entirely new way. To be more precise, Aristotle is concerned not with everything that might called an argument, but with the syllogism, i.e., an argument () in which, certain things being laid down, something different from I am grateful to the participants and organizers of the conference on Philosophy and Rhetoric in Classical Athens for an exceptionally stimulating gathering and for many helpful comments. I owe a special debt to Chloe Balla for valuable written comments on an earlier version of this paper. Papers related to this one were presented at the Humboldt Universitat Berlin, McGill University, the University of Toronto, and the Central division meeting APAin 2001. I learned a great deal fromthe audience on each of these occasions, and I am especially indebted to Gisela Striker, my commentator at the APA meeting. This content downloaded from 192.167.204.6 on Sun, 21 Sep 2014 02:48:48 AM All use subject to JSTOR Terms and Conditions RHETORI CA 88 them follows of necessity by their being so (Topics 1.1.100a257; Prior Analytics 1.1.24b1820). This denition marks off a class of valid arguments that might be of use to someone actually mak- ing a case for something, for example, by excluding arguments whose conclusion already gures among its premises. The prin- cipal thesis of Aristotles theory is that every syllogism, i.e., ev- ery argument satisfying this denition, is a categorical syllogism, meaning an argument that belongs, or consists of parts that be- long, to one of the moods of the categorical syllogism, which we know under their medieval names, Barbara, Celarent, Darii, and so on. To employ language that Aristotle does not use himself, his answer to the question is that syllogisms are valid if and only if they are formally valid, and they are formally valid if and only if they can be shown by analysis to be categorical syllogisms. It is from the operation of analyzing arguments into categorical form that the Analytics take their name (cf. Prior Analytics 1.32.46b3847a5). Aristotle has notoriously little to say about where this logical theory belongs in his classication of the sciences. At the beginning of the Prior Analytics, he emphasizes that the study of the syllogism on which he is about to embark is an essential preparation for the study of demonstrative syllogisms and the kind of knowledge or understanding that one has by grasping them, which will occupy him in the Posterior Analytics (1.1.24a12; cf. 4.25b2631). 1 In Prior Analytics 2.23, however, he asserts that: Not only are dialectical and demonstrative syllogisms brought about by means of the gures [of the categorical syllogism] but also rhetorical syllogisms and, quite generally, any attempt to produce conviction (|) of any kind whatever. 68b914 Andthis is only one of a number of passages inwhichAristotle insists on the universal application of the categorical theory to arguments of any and every kind wherever they may be found (Prior Analytics 1.23.40b20 ff., 41b15; 25, 42a301; 28.44b68; 29.45b3646a2; 30.46a3 4). Let us call the discipline to which the theory of the categorical syllogismbelongs analytic, even though this termhas, at best, only 1 J. Brunschwig, Lobjet et la structure des seconds analytiques dapres Aristote, in E. Berti, ed., Aristotle on Science: the Posterior Analytics (Padua: Antenore, 1981), 6196. This content downloaded from 192.167.204.6 on Sun, 21 Sep 2014 02:48:48 AM All use subject to JSTOR Terms and Conditions Aristotle on the Disciplines of Argument 89 a slender basis in Aristotles own usage. 2 If, as Aristotle maintains, every syllogism is a categorical syllogism, then every discipline in which argument plays a part relies on principles that it is the business of analytic to study; and, to the extent that a discipline requires an explicit understanding of the fundamentals of argument, it must turn to analytic. Aristotle has in effect made analytic the master discipline of argument. As such, it can be contrasted with two other kinds of discipline that concern themselves with argument. On the one hand, there are arts of argument like dialectic andrhetoric, whose aimis tofurnishthe corresponding practices of dialectical and rhetorical argument with a systemor method. The Topics contains a method for the practitioner of dialectic; the Rhetoric, a method for the orator. On the other hand, there are the special sciences. The material discussed in the Posterior Analytics, where the conditions that a syllogism must satisfy if it is to qualify as a demonstration proper to one of these sciences are tackled, is either an appendix to analytic, construed narrowly as the general logical theory of the syllogism, or a second part of analytic, which, however, applies not to all syllogisms, but only to demonstrations. The picture of a system of argumentative disciplines dependent on analytic that emerges in this way seems to receive conrmation from the structure of the Organon, where Aristotles ancient editors brought together the works they took to be about logic. 3 According to tradition, the rst two works of the Organon, the Categories and the De interpretatione, prepare the way for the categorical theory of the syllogismtackled in the Prior Analytics by studying terms and propo- sitions respectively. The Posterior Analytics and the Topics then apply the theory to the domains of demonstrative and dialectical argument in turn, and the Sophistical Refutations brings the study of argument to a close by examining fallacious argument. At the very end of the Sophistical Refutations, and therefore of the Organon itself, Aristotle observes that, because he had no predecessors in the study of syl- logizing, it was necessary for him both to found the discipline and to bring it to the level already attained by other disciplines (34.183b34 ff., 184b1 ff.). He compares the more typical case of rhetoric, which had reached its then present condition gradually as one rhetorician after another built on, and added to, the contributions of his prede- 2 It is found only at Rhet. 1.4.1359b10. Metaphysics .2.1005b25, which nds fault with those who are ignorant of analytics, may furnish a parallel. 3 J. Brunschwig, LOrganon: Tradition grecque, in R. Goulet, ed., Dictionnaire des philosophes grecques (Paris: CNRS, 1989) I.485502. This content downloaded from 192.167.204.6 on Sun, 21 Sep 2014 02:48:48 AM All use subject to JSTOR Terms and Conditions RHETORI CA 90 cessors (183b26 ff.). For a very long time, it was assumed that these celebrated remarks were about the categorical theory of the syllo- gism expounded in the Prior Analytics and were meant to emphasize the place of central importance that analytic occupies among the disciplines of argument. But scholarship has gradually made it plain that Aristotle was not talking about analytic at the end of Sophistical Refutations. The most important piece of evidence is the curious failure of both the Topics and the Sophistical Refutations to take account of the categorical theory of the syllogism or its technical vocabulary, even though the declared object of the former is to expound a method of syllogizing for use indialectic (1.1.100a1 ff.). What is more, the syllogisms that the Topics tells us howto formby and large do not conformto the rules of the categorical syllogistic. Surely, the argument runs, the Topics and the Sophistical Refutations would have betrayed the inuence of the Prior Analytics, or its theory, if the Prior Analytics, or its theory, had been there to inuence them. The Topics and the Sophistical Refutations forma unity 4 fromthis point, when I speak of the Topics I mean to include the Sophistical Refu- tationsand on closer inspection it transpires that Aristotles proud claimto have been the rst student of the syllogismconcludes a reca- pitulation of the inquiry that corresponds to the programof the Topics (SE 34.183a37184b8). It is this inquiry that he meant to describe as the rst investigation of syllogizing. 5 The curious omissions of the Topics and the fact that the arguments it tells us how to form are typically not categorical syllogisms are explained by the fact that, al- though Aristotle had the idea of the syllogismwhen he composed the Topics, the categorical theory of the syllogism was still in the future. This is an old story. I have rehearsed it here not because of its intrinsic interest, but in order to set the stage for the question with which I shall be chiey concerned. Suppose that Aristotles concep- tion of the relations between analytic and argumentative disciplines like dialectic and rhetoric was not, as long assumed, constant, but that analytic and the categorical syllogistic were relative latecomers to the scene. What consequences does this have for the picture of the argumentative disciplines that emerged above? More precisely how 4 T. Waitz, Organon Graece, (Leipzig: Hahn, 184446), II.5289; L.-A. Dorion, Aristote: Les refutations sophistiques (Paris: Vrin, 1995), 245. 5 C. Thurot, E
tudes sur Aristote: Politique, Dialectique, Rhetorique (Paris: Durand,
1860), 1957; L.-A. Dorion, Aristote et linvention de la dialectique in M. Canto- Sperber and P. Pellegrin, eds., Le Style de la pensee: Recueil de textes en hommage a` Jacques Brunschwig (Paris: Les Belles Lettres, 2002), 182220. This content downloaded from 192.167.204.6 on Sun, 21 Sep 2014 02:48:48 AM All use subject to JSTOR Terms and Conditions Aristotle on the Disciplines of Argument 91 did Aristotle conceive of these disciplines and their relations before analytic? It will help to sharpen this question if we consider an objection. The Topics and Prior Analytics have very different aims. The Topics is a manual whose purpose is to furnish a certain practice of argument, dialectic, with a method. The Prior Analytics expounds the worlds rst formal logic. Why then, runs the objection, should the Topics enter into the details of a logical theory that it is the business of a different kindof workto study? Andit concludes that the Topics could have been written in much the same way whether it was composed before or after the discovery of the categorical syllogism. 6 The strongest form of this objection would undermine the case for a development in Aristotles thinking of the kind I have just de- scribed. If it is on the right lines, the Topics silence about analytic and the categorical theory of the syllogism do not speak as loudly as proponents of the developmental account suppose. I believe that this form of the objection can be answered, so that the case for the relative chronology of the Topics and Prior Analytics and a develop- ment in Aristotles thinking about argument between them stands. Quite apart from other considerations, there is abundant evidence that Aristotle thought the theory of the categorical syllogism did have a contribution to make to argumentative disciplines like dialec- tic and rhetoric. After the passage about the universal applicability of the gures of the categorical syllogism from Prior Analytics 2.23 that I cited above, Aristotle goes on to analyze forms of argument especially prominent in rhetoric in the light of the categorical the- ory. This analysis seems to be cited and repeated in less technical form in the Rhetoric in a pair of passages that appear to be late in- sertions in otherwise earlier, Topics-oriented surroundings (Rhetoric 1.2.1357a2258a2; 2.25.1400b131403a6). 7 What is more, in the Prior Analytics Aristotle explains how the method of invention based on 6 R. Smith, Dialectic and the Syllogism, Ancient Philosophy 14 (1994): 13351 (p. 140). 7 So F. Solmsen, Die Entwicklung der aristotelischen Logik und Rhetorik. Neue philol- ogische Untersuchungen 4 (Berlin: Weidmann, 1929), 1331; followed by M. F. Burnyeat, Enthymeme: Aristotle on the logic of persuasion in D. J. Furley, A. Nehamas, eds., Essays on Aristotles Rhetoric (Princeton: Princeton University Press, 1994), 355 (pp. 315); J. Allen, Inference from Signs: Ancient Debates about the Nature of Evidence (Oxford: Oxford University Press, 2001), 2022; opposed by J. Barnes, Proof and the Syllo- gism in Berti, Aristotle on Science, cited in n. 1 above, 1759 (p. 52) (in the context of broader agreement with Solmsens developmental thesis); C. Rapp (trans.), Aristoteles: Rhetorik (Berlin: Akademie Verlag, 2002), II.2024. This content downloaded from 192.167.204.6 on Sun, 21 Sep 2014 02:48:48 AM All use subject to JSTOR Terms and Conditions RHETORI CA 92 the categorical theory of the syllogism will be of use to the dialecti- cian in his quest for syllogisms (1.27.43b1130; 30.46a210), and he sometimes pauses to give tactical advice based on the theory to the dialectician. 8 This shows that Aristotle was concerned not merely to demon- strate the universal application of his theory to every domain of argument, but that he also believed that understanding it, or parts of it, or at a minimumsome of its results will help the practitioner of ev- ery kind of argument (cf. Prior Analytics 1.27.43a204; 29.45b3646a2; 30.46a3 ff.; 31.46a29). Had the theory been available when Aristotle composedhis manual of dialectical argument, then some of its results would have been included there just as some of them were included in his revisedmanual of rhetoric. The fact that they were not stands as powerful evidence that the Topics andits viewof syllogistic argument antedate the Prior Analytics and its theory. But this evidence is about how the new discipline of analytic added to the stock of dialectics techniques. The objection can be raised in a different form, which retains considerable force. In viewof the very different aims of the two disciplines, why should Aristotles conception of the discipline of dialectic and the place it occupies in the system of argumentative disciplines have been different before the emergence of analytic? And it is chiey this question, and not the question of what changes in dialectical or rhetorical methods analytic may have required, that I shall pursue here. The view that analytic replaced dialectic is sometimes attributed to proponents of the developmental thesis that I sketched a moment ago, and this is clearly much too strong. 9 The evidence I have cited showing that Aristotle thought that analytic had something to contribute to the practice of dialectical argument does not show that he thought that analytic abolished the discipline of dialectic anymore than it abolished that of rhetoric. But obviously the view that analytic left dialectic entirely unaffected and the viewthat it replaced dialectic do not exhaust the eld. It is hardly surprising that generations of readers took Aristo- tles claim at the end of the Sophistical Refutations to have been the 8 On this point, see H. Maier, Die Syllogistik des Aristoteles, (Tu bingen: Laupp 18961900), II.b.78 n. 3. An especially striking example is furnished by the counsel to be on the lookout for recurring terms in an opponents argument because the middle term on which his syllogism will depend must occur twice (2.19.66a25). 9 Solmsen is sometimes said to have held this view, but though an occasional incautious formulation of his may suggest this (Entwicklung, 26, 195), he did not hold it in anything like this unqualied form. This content downloaded from 192.167.204.6 on Sun, 21 Sep 2014 02:48:48 AM All use subject to JSTOR Terms and Conditions Aristotle on the Disciplines of Argument 93 rst to study syllogizing as a reference to analytic and the Prior An- alytics, for, as we have seen, in his mature view the study of the syllogism in general is the function of analytic. This is clearly im- plied at Prior Analytics 1.4, where Aristotle characterizes the sub- ject of the inquiry underway as the syllogism in general by contrast with a species of syllogism, demonstration, that is to be tackled later (25b2631). And in a couple of passages in the Posterior Analytics, Aristotle refers back to the Prior Analytics as the work about the syl- logism (1.3.73a14, 1.11.77a33). So it is puzzling to nd a manual of dialectic, the Topics, presented as the fruit of an inquiry into syllo- gizing, rather than a kind of syllogizing, viz. the dialectical kind. To add to our puzzlement, the Sophistical Refutations elsewhere uses the term syllogistic to characterize dialectical arguments in a contrast with rhetorical arguments (5.167b21 ff.). 10 And the Rhetoric constantly treats the syllogism as the special concern of the discipline of dialec- tic by contrast with the enthymeme, the form of argument proper to rhetoric. 11 My thesis is that, in a way, analytic did replace dialectic. It re- placed dialectic, however, not by abolishing it or, necessarily, by rendering all of its counsels obsolete, but by supplanting it as the master discipline of argument whose responsibility it is to treat of the syllogism in general. As a result, dialectic was demoted to the rank of a subordinate discipline of argument, oriented toward one practice of argument and dependent on analytic for what it needs to know about the fundamentals of argument. But analytic was a discipline of a new kind, not a method oriented to one practice of argument or a special science that makes use of argument, but a logical theory. If you will, in the course of becoming the master disci- pline of argument, it changed what it was to be a master discipline of argument. It is relatively easy to see how analytic was able to be the master discipline of argument. As the general theory of syllogistic validity, it is the home of principles to which every valid syllogism must conform. 12 To be sure, the Prior Analytics contains a good deal more than a logical theory, and to the extent that it is intended as an exposi- 10 Cf. Rapp, Aristoteles: Rhetorik, cited in n. 7 above, II.97 ad Rhet.1.1.1355a30. 11 Rhet. 1.1.1355a30, b1517; 2.1356a36-b2, b1013, 1358a26, 15; 2.22.1395b224. 12 This claim has to be qualied, and is qualied by Aristotle in Prior Analytics 1.44, where he allows that so-called syllogisms on the basis of a hypothesis cannot be analyzed as categorical syllogisms. It is nonetheless surprising just how minor a qualication Aristotle takes this to be. Cf. G. Striker, Aristoteles u ber Syllogismen aufgrund einer Hypothese, Hermes 107 (1979): 3350. This content downloaded from 192.167.204.6 on Sun, 21 Sep 2014 02:48:48 AM All use subject to JSTOR Terms and Conditions RHETORI CA 94 tion of the discipline, so does analytic. 13 But most of the additional matter consists of applications of the theory to issues that arise in connection with argument and different forms of argument. As we have already observed, Aristotle applies the categorical theory to forms of argument that are prominent in rhetoric. He also uses the formal machinery of the theory to tackle fallacies (2.1618), to prove that it is possible to syllogize true conclusions from false premises (2.24), and to construct an extremely simple method of invention that can be applied mechanically to any desired categorical conclu- sion to yield syllogisms for it (1.278). It remains to be seen how the discipline of dialectic, which unlike analytic is a method oriented towards one practice of argument, could also have been the master discipline of argument. For help we can turn to the relations between the disciplines of dialectic and rhetoric as Aristotle describes them in the Rhetoric. The work begins with the famous declaration that rhetoric is the counterpart of dialectic (1.1.1354a1). Two disciplines are counter- parts when they stand in the same relation to different objects. In the Gorgias Plato had set up a system of correspondences between arts that care for the body and those that care for the soul (463a-466a). Justice, for instance, is the counterpart of medicine because it stands in the samecorrectiverelation to the soul that medicine stands in to the body. Rhetoric is the counterpart of cookery (o(), or so Socrates maintains, because it is a false likeness of justice in the way cookery is of medicine, aiming at pleasure rather than the good of its subject and relying on experience and conjecture rather than knowledge. Aristotle defends rhetoric by making it an art of argument like dialectic. The features that it shares with dialectic, and which they both owe to their status as arts of argument, are therefore no more to be held against rhetoric than they are against dialectic. At the same time, he rebukes contemporary rhetoricians who neglected argument in favor of appeals to the emotions and discussion of the parts of an oration (Rhetoric 1.1.1354a1116, b1622). As arts of argument, rhetoric anddialectic have no special subject matters of their own of the kind that distinguish the ordinary run of arts and sciences, but can in principle be applied to any and everysubject (Rhetoric 1.1. 1354a13, 1355b79; 2.1356a314, 1358a21 25; 4.1359b1116). The objects to which they are related are instead different practices of argument, and the relation in which they stand 13 G. Striker, Aristotle and the Uses of Logic, in J. Gentzler, ed., Method in Ancient Philosophy (Oxford: Oxford University Press, 1998), 20926. This content downloaded from 192.167.204.6 on Sun, 21 Sep 2014 02:48:48 AM All use subject to JSTOR Terms and Conditions Aristotle on the Disciplines of Argument 95 to them is that of supplying them with an art or method. 14 In the passage that we are considering, these practices are described as activities common to a certain extent to all human beings, namely examining and upholding an argument in the case of dialectic, and accusing and defending in that of rhetoric. Later in the Rhetoric, Aristotle distinguishes other varieties of rhetoric with reference to their characteristic activities and the concerns of the audiences to which they are addressed. To accusation and defense, which belong to forensic oratory, he adds advocatingor opposinga course of action, which belong to deliberative oratory, and praise or blame, which belong to epideictic oratory (1.3.1358a36-b13). The Topics likewise distinguishes varieties of discussion with the difference that Aristotle seems to single out one of these as dialectic in the strictest sense (Topics 8.5.159a2536 cf. 11.161a25; Sophistic Refutations 2.165a38 ff.). This is discussion for the sake of | (exercise or training) and . (putting to the test), which he alsocharacterizes as for the sake of . andc (investigation) (159a33). These he contrasts with discussions between teachers and learners, where the aim is to impart knowledge, and competitive encounters, where there is no common task and each party strives for victory by any available means. Common to all of them are the presence of two parties, a questioner and an answerer, and the requirement that the argument advance by stages in which the questioner secures the assent of the answerer to premises put forward in the form of questions. Aristotle does not explain what is put to the test in dialectical argument. It is temptingtoconnect what he says here with(, which is introduced in the Sophistical Refutations as either a branch of dialectic or a closely related sister discipline, and whose function is to put an interlocutors claim to knowledge to the test by arguing in the Socratic manner fromhisthe interlocutorsadmittedopinions (Sophistical Refutations 2, 164b4; 8, 169b25; 11, 171b4; 172a217; 34, 183b1). 15 But it does not appear that the dialectical answerer Aristotle envisages in the Topics lays claim to knowledge. The defense of theses to which he is personally attached is only one possibility among others (8.5.159b1, 257). He will also, and perhaps more 14 J. Brunschwig, Rhetorique et Dialectique, Rhetorique et Topiques in Furley and Nehamas, Aristotles Rhetoric cited in n. 7 above, pp. 5796 (p. 59); C. Rapp, Aristoteles: Rhetorik, cited in n. 7 above, II.201. 15 P. Moraux, La joute dialectique dapre`s le huitie`me livre des Topiques in G. E. L. Owen, ed., Aristotle on Dialectic: The Topics (Oxford: Oxford University Press, 1968), 271311 (2889). This content downloaded from 192.167.204.6 on Sun, 21 Sep 2014 02:48:48 AM All use subject to JSTOR Terms and Conditions RHETORI CA 96 typically, uphold theses without regard for his own opinions; in this case, his task is to admit only premises which are reputable or more reputable than the thesis (8.5.159a38-b25). 16 In one form of dialectical discussion, the answerer undertakes to defend a thesis of a famous philosopher, and his task then becomes to admit only premises that would be acceptable to that philosopher (8.5.159b27 35; cf. 1.11.10461928). Though the answerers ability to uphold theses will undoubtedly be put to the test in dialectical discussions of these kinds, if he performs his task correctly, it will be chiey the thesis itself that is tested and made the object of investigation (cf. 8.4.159a204). And this seems to be what Aristotle has in mind in the Metaphysics when he observes that dialectic is peirastic, that is, that it probes or examines or puts to the test, where philosophy is gnoristic, i.e., has knowledge ( 2.1004b226). I shall tackle the question of what Aristotle means by exercise or training shortly. The aimof the discipline of dialectic, andtherefore Aristotles aim in the Topics, is to equip these practices of argument, and especially dialectical argument inthe strictest sense, witha method. The method had to unite those elements that are necessary to this practice, which serves, therefore, as the methods organizing principle. In treating rhetoric and dialectic as counterparts, each with its own sphere of operation, however, Aristotle is not, as it might at rst appear, treating them as completely autonomous and coequal disci- plines. It is plain fromthe Rhetoric that the discipline of rhetoric relies ondialectic for its understandingof the fundamentals of argument. In the rst, introductory chapter of the Rhetoric, Aristotle explains that the enthymeme is a syllogism of a kind (1355a810), 17 and that the orator who combines a grasp of the syllogismwith an understanding of the effects on argument of rhetorical subject matters will be best equipped to argue in rhetorical contexts. But for his understanding of how and from what a syllogism arises, Aristotle insists, the or- ator is reliant on dialectic since the consideration of every syllogism without distinction (oc) is the business of dialectic, either dialec- tic as a whole or a part of it (1355a810). 18 The specically rhetorical 16 M. Wlodarczyk, Aristotelian Dialectic and the Discovery of Truth, Oxford Studies in Ancient Philosophy 18 (2000): 153210. 17 Translated a syllogism of a kind in accordance with Burnyeats recom- mendation in order to leave open the possibility that an argument may qualify as an enthymeme without meeting the standards strictly necessary to be a syllogism (Burnyeat, Enthymeme, cited in n. 7 above, pp. 1315).. 18 Without distinction is the Oxford translator, Rhys Roberts, rendering. With the use of oc here compare De gen. et corr. 1.1, 314a2, where Aristotle proposes to This content downloaded from 192.167.204.6 on Sun, 21 Sep 2014 02:48:48 AM All use subject to JSTOR Terms and Conditions Aristotle on the Disciplines of Argument 97 knowledge that the orator must add to the grasp of the syllogism that he takes from dialectic is an understanding of the sort of matters with which enthymemes are concerned and how enthymemes differ from logikoi syllogismoi (a1014). The term logikos is an important clue about the nature of dialecti- cal argument proper. To tackle an issue logically, in this sense, is to do so on the basis of logoi, arguments or reasonings, viewed in one way or another, and to one degree or another, in abstraction from the special features of the subject matter at issue. These features are accessible only to those with a specialists substantive understanding of the subject matter, and in the treatises proceeding logically is typi- cally opposed to doing so on the basis of such an understanding. 19 Dialectic proceeds logically because when it tackles a question, it is on the basis of abilities available to the master of an art of argument without substantive specialized understanding. To the extent that one draws on such an understanding, one leaves dialectic behind (Rhetoric 1.2.1358a236; 4.1359b1216). The orators understanding of the enthymeme is inpart the result of adding to the grasp of the fundamentals of argument that he takes from dialectic knowledge of the kinds of matters with which rhetoric deals, but also of the kinds of occasions and audiences to which rhetorical argument is suited (Rhetoric 1.2.1357b26; 2.20.1394a247). One effect is a relaxation in the rigor or stringency permitted in en- thymemes by comparison with syllogisms proper. This is because the matters about which one argues in rhetoric, like those which are the concern of practical reason, rarely lend themselves to reso- lution by conclusive arguments (Rhetoric 1. 2.1357a17, 1315, 227; 2.25.1402b324). Typically it is possible only to present considera- tions of a certain weight, and orators arguments, though able to make a conclusion a reasonable thing to believe, cannot exclude the possibility that it is false. Often the best decision that jurors attend- ing conscientiously to the arguments on both sides of the case can reach is one that new evidence may show to be mistaken. And as- semblies that arrive at the decision that does the most justice to the treat generation and corruption oc v , i.e., without entering into the special features or peculiarities belonging to the generation and corruption of the different kinds of natural substance (cf. 2.9.335a258). 19 On the term see Waitz, Organon Graece, cited in n. 4 above, II.353 ff. (ad An. post. 82b35); A. Schwegler, ed., Die Metaphysik des Aristoteles (Tu bingen: Fues, 1847), IV.4851 (ad Metaph. Z 4.1029b13); H. Maier, Syllogistik, cited in n. 8 above, II.a.11 n. 3; M. F. Burnyeat, A Map of Metaphysics Z (Pittsburgh: Mathesis Publications, 2001), 1922 et passim. This content downloaded from 192.167.204.6 on Sun, 21 Sep 2014 02:48:48 AM All use subject to JSTOR Terms and Conditions RHETORI CA 98 considerations put before them may see that decision undone by future events. The syllogisms proper to dialectic, on the other hand, owe their greater stringency to dialectics nature as a purer art of argument. It is purer, however, not by abstracting from the concerns of actual practices of argument in the way analytic does, but by correspond- ing to a more purely argumentative practice, one, as we might say, less compromised by real world considerations than rhetoric and, in a different way, the special sciences are. The kind of logical under- standing of argument (in our sense of logical) which a master of the discipline of analytic will have, though possibly a helpto participants in argumentative practices like dialectic and rhetoric, cannot by itself equip themfor success in the way that the disciplines of dialectic and rhetoric must. The dialectician, as Aristotle conceives him, is able to argue on either side of any question without substantive knowledge of the eld to which the question belongs. Some of what he needs to know to this end corresponds to analytic, but he must also have a command of reputable opinions. One of the instruments (o) by means of which the dialectician is suppliedwith syllogisms is the col- lection and classication of premises (Topics 1.1314; cf. 2.101a3034). The boundaries of the logical inAristotles sense are drawnsoas toex- clude substantive understanding of a subject matter, but not content if that means familiarity with the opinions of the many and the wise. Before the invention of analytic, dialectic had either to borrow its teachings about the fundamentals of argument from another discipline or tackle them itself along with matters that are of little or no use outside specically dialectical forms of argument. According to the view that I am defending, it tackled them itself because it was, in Aristotles view, the primary discipline of argument, and it owes its primacy to the paradigmatic character of the practice of argument towards which it is oriented. As a result, every other discipline in which argument plays a part had to turn to dialectic for its understanding of argument, even if in the endits practitioners will never practice dialectic and have no need to master the discipline in all its many details. To make this view plausible it is essential to discover how the practice of argument to which dialectic was oriented could have been paradigmatic in a way that permitted dialectic also to serve as the master discipline of argument. In Topics 1.2, Aristotle distinguishes three uses for the pragmateia, i.e., the treatise itself or the method it contains. 20 In agreement with a long line of scholars I take the rst of 20 R. Smith, Aristotle on the Uses of Dialectic, Synthese 96 (1993): 33558. This content downloaded from 192.167.204.6 on Sun, 21 Sep 2014 02:48:48 AM All use subject to JSTOR Terms and Conditions Aristotle on the Disciplines of Argument 99 these, |, practice, in the sense of exercise or training, to be the practice, in the sense of activity or pursuit, of which the discipline of dialectic elaborated in the Topics is the art. 21 If this is right, then the second and third uses of the method, though to one degree or another dialectical, will not, strictly speaking, be forms or varieties of dialectic. But if the practice for which the discipline of dialectic set out in the Topics is practice in the sense of training, what is it training for? Aristotles idea seems to have been that dialectical argument was a form of intellectual practice or training, which gives rise to intellec- tual and argumentative facility in the way that a programof physical exercise performed under the direction of a gymnastic instructor produces physical strength and coordination (cf. Topics 1.11.105a9; 8.5.159a25; 8.11.161a25). Aristotle seems also to have believedthat di- alectic stoodinan especially close relation to philosophy (Metaphysics 4.2.1004b226; Topics 1.2.101a34-b4; 8.1.155b716, 14.163b916). It is not merely that the sample dialectical problems that are mentioned in the Topics tend to be philosophical. 22 The method it expounds seems to have in view philosophical discussions of the kind that we are led to believe took place in the Academy and which are mocked in the famous fragment of Epicrates in which Platos followers are pictured solemnly dening and classifying by species vegetables, not except- ing the pumpkin. It is concerned not only with the construction and evaluation of arguments that a predicate simply belongs to a subject (i.e., belongs to it as an accident), but gives equal attention to those in which the issue is whether it belongs as a genus, proprium or def- inition, and elaborates a separate method of inventing syllogisms for each of these, the four so-called predicables (1.6.102b27 ff.; 4.1.120b12 ff.; 6.1.139a24-b5; 7.5.155a1636). 23 As we shall see, Aristotles views about the relations between dialectic and philosophy are surprisingly complex, but if this conclu- sion is on the right lines, he regarded the practice of dialectic, or one favored form of it, as preparation or training or practice primarily for philosophy. Aristotles talk of | would then be another 21 E. Kapp, Greek Foundations of Traditional Logic (New York: Columbia University Press, 1942), 1213; P. Moraux, La joute dialectique, cited in n. 15 above, pp. 289 90; J. Brunschwig, Aristotle on arguments without Winners and Losers, Jahrbuch Wissenschaftskolleg zu Berlin 1984/5, 3140 (p. 34); O. Primavesi, Die aristotelische Topik: Ein Interpretationsmodell und seine Erprobung am Beispiel von Topik B (Munich: Beck, 1996), 2021, 49 ff., Rapp, Aristoteles: Rhetorik, cited in n. 7 above, I.2512. 22 I. Duhring, Aristotles use of examples in the Topics, in G. E. L. Owen, Aristotle on Dialectic, cited in n. 15 above, pp. 20229. 23 J. Brunschwig, Aristotle on arguments, cited in n. 21 above, p. 32. This content downloaded from 192.167.204.6 on Sun, 21 Sep 2014 02:48:48 AM All use subject to JSTOR Terms and Conditions RHETORI CA 100 instance of a widespread tendency to compare the training or prepa- ration of the soul to that of the body. We may compare Isocrates, who in his use of the idea of counterpart disciplines makes philosophy, in his sense of the term, the counterpart of (, of which a part is ( (Antidosis 18082). And though he was notoriously dismissive of the claims made by gures like Plato and Aristotle on behalf of what they call philosophy, Isocrates is also willing to concede some value to it as training (|) and preparation for philosophy properly so called (Antidosis 2669). There are a handful of clues in Plato which suggest that there was a practice of argument in the Academy that was conceived as a form of practice or preparation and was called |. Thus in the rst part of the Platonic dialogue that bears his name, Parmenides raises difculties for the theory of the forms, which is treated there as the view of the young Socrates. Socrates has gone astray, Par- menides maintains, by attempting to dene forms of beauty, jus- tice, goodness and the like before he had practiced or exercised properly. The practice he prescribes resembles Zenos manner of argument, which had been discussed earlier in the dialogue, and he insists that it is an essential preparation for grasping the truth (136ce). 24 Of course, it is another question why the practice of dialectic arose and took the form that it did. A fuller attempt to answer this question would require considering the Academic milieu and the wider philosophical world of the fth and fourth centuries. The inquiry could be extended to the broader social and political conditions in the Greek speaking world that favored the growth of public practices of argumentationandso encouragedreectionabout proof, validity, refutation and the like. But to the extent that this kind of inquiry appeals to contingent historical factors, it tends to draw us away fromthe kind of answer Aristotle himself would most likely have given. I take it that he would have explained the practice of dialectical argument as an expression of human nature. I have already cited the Rhetorics claimthat all human beings share in a way in dialectic to the extent that they attempt to examine and uphold theses (1.1.1354a3 6). The same sentiment can be found in the Topics, where all human beings are said to participate without art in that with which dialectic is concerned as an art (Sophistic Refutations 11.172a34). Remarks like 24 On this passage from the Parmenides see Maier, Syllogistik, cited in n. 8 above, II b 51, n. 1. This content downloaded from 192.167.204.6 on Sun, 21 Sep 2014 02:48:48 AM All use subject to JSTOR Terms and Conditions Aristotle on the Disciplines of Argument 101 these suggest that Aristotle regarded the more specialized form of philosophical argument toward which the dialectical method set out in the Topics is oriented as the form par excellence of the universal practice of examining and upholding theses in argument by question and answer. We may compare Isocrates attitude toward the kind of speech or logos that he teaches his pupils how to compose; he regards it as the expression par excellence of the human capacity for logos, which encompasses all that is most distinctive of human beings (Nicocles 610). These considerations go some way towards explaining how Aristotle could have regarded dialectic as the paradigm practice of argument and the corresponding discipline as the primary dis- cipline of argument. It will also help to attend to a feature of dialec- tic that it owes specically to its gymnastic character and which complements its nature as a pure practice of argument, that is, a practice participation in which calls only for a logical under- standing, in Aristotles sense, of how to construct and evaluate arguments and not substantive knowledge of the matters under discussion. It is clear that the answerers task is, in a way, to make the ques- tioners work more difcult. He is to do this not by using any means to obstruct the questioners progress, but rather by holding the ques- tioners argument to the highest standards, so that when the an- swerers thesis is defeated it will be because of its weakness not the answerers (Topics 8.4.159a204). This is what is behind Aris- totles talk of a common task shared by questioner and answerer (8.11.161a20, 37). 25 In Topics 8.5 the focus is on which premises the answerer ought to allow. As we have seen, Aristotle holds that he ought only to give his approval to those which are reputable or more reputable than the conclusion for which the questioner is striving to construct an argument. But throughout the Topics Aristotle makes note of objections that the answerer can raise. Indeed he says the dialectician is not only a proposer of premises but also a raiser of objections (8.14.164b3). And this shows that the answerer is also charged to ensure that the questioners argument proceeds validly. The practice of gymnastic dialectic, then, puts a premium on care- ful step-by-step argument in which each step must be made explicit and pass muster with the answerer who is on the lookout for any misstep This focus on the argument as such, with its attendant em- phasis on rigor and explicitness, furnishes another reason why the 25 Brunschwig, Aristotle on arguments, cited in n. 21 above. This content downloaded from 192.167.204.6 on Sun, 21 Sep 2014 02:48:48 AM All use subject to JSTOR Terms and Conditions RHETORI CA 102 fundamentals of argument are a special concern of the dialectical discipline. By contrast rhetoric does not share dialectics interest in rigor and explicitness for their own sake. As Aristotle several times notes, rhetorical syllogisms or enthymemes need not state all the premises on which the conclusion depends. 26 He emphasizes that the kind of listeners whom it is rhetorics task to address are unable to follow long arguments, but also that a full statement of the argument is hardly necessary if the listeners can take the point without it (Rhetoric 1.2.1357a45, 1622; 2.22.1395b246). The orator aims topersuade and argument is his principal instrument. The point is not that this means that anything goes in rhetoric. A noble orator will want to convince his audience of the truth, or the conclusion that stands the best chance of being true, and have them accept it for sound reasons, but even so he lacks the dialecticians interest in the mechanics of valid argument as such. He is able to adopt this attitude towards argument, however, because rhetoric can turn to dialectic for its understanding of the fundamentals of argument, which it then modies and supplements to suit its own ends. So far I have argued that in Aristotles early thinking about argu- ment the discipline of dialectic played a double role. It was the art or discipline of a practice of argument, so-called|, whichserved as its organizing principle, and it was at the same time the master discipline of argument, to which disciplines corresponding to other practices of argument had to turn for an understanding of the funda- mentals of argument. I suggested that the discipline of dialectic was able to performboth these functions for Aristotle because in his view, though it corresponds to one practice of argument among others, that practice is the paradigm form of argument, argument par excellence. (I have left entirely out of account a separate question, which is nev- ertheless related to those pursued here, namely how early Aristotle recognized that demonstrative syllogisms would require a special inquiry of the kind pursued in the Posterior Analytics.) 27 The perspective that we have now achieved should throw light on the other uses for the method of dialectic that Aristotle puts beside |. A closer examination of these uses will, I believe, lend further support to the thesis that I have been advocating. The second use of the method is 26 Tradition turned this accidental attribute of the enthymeme into its essence. Cf. Burnyeat, Enthymeme, cited in n. 7 above, pp. 214, 3950. 27 On this issue, see J. Barnes, Proof and the Syllogism, cited in n. 7 above. This content downloaded from 192.167.204.6 on Sun, 21 Sep 2014 02:48:48 AM All use subject to JSTOR Terms and Conditions Aristotle on the Disciplines of Argument 103 for encounters because, having reckoned up the opinions held by the many, we shall speak to them not from the opinions of others, but from their own, changing their minds about anything they seem to us not to have stated well Topics 1.2.101a3034. 28 The third is for the philosophical sciences, because, if we have the ability to go through the difculties on either side, we shall more readily discern the true and the false in each matter Topics 101a346 There are two ways in which the method of dialectic may be useful to these ends, however. Is it useful because (a) the practice of gymnastic argument, of which it is in the rst instance the method, is in turn useful in relation to other ends, or (b) the method itself, or parts of it, are directly useful to practices of argument other than dialectic and without being put to use in the practice of dialectic? The position that I have been defending would lead us to expect the answer to be in way (b) or a combination of ways (a) and (b). And certainly in the case of the second use of the method, in encounters, (b) is very plausible. The command of a mass of reputable opinions, organized with reference to the type of person to whom they are likely to be persuasive, will be of the greatest assistance in conversations of a not specically dialectical character, as Aristotle also notes in the Rhetoric (1.1.1355a28; 2.1356b324; cf. 2.22.1396b211). The philosophical use of the method is more complicated and has been the object of much attention. Aristotle expands on the description of the philosophical use of the dialectical method in the immediate sequel to the remarks cited above. 29 And further the treatise or the method it contains [the pragmateia] is of use regarding the rst of the matters that concern each science. For it is impossible on the basis of the principles of the science at issue to say anything about them, since the principles are prior to everything else, but it is necessary instead to tackle them via the reputable opinions regarding each of them. And this is peculiar or most proper to dialectic 28 These translations and that of 101a36b4 below are adapted from R. Smith (trans.), Aristotles Topics: Books I and VIII (Oxford: Oxford University Press, 1997). 29 Unless they are meant to introduce a fourth use in addition to the three promised at 101a26. On this question, see J. Brunschwig, ed. and trans., Aristote: Topiques, Livres IIV (Paris: Les Belles Lettres, 1967), xii, 116 n. 1. This content downloaded from 192.167.204.6 on Sun, 21 Sep 2014 02:48:48 AM All use subject to JSTOR Terms and Conditions RHETORI CA 104 since, because of its capacity to examine, it has a way to the principles of all the disciplines. Topics 101a36b4 Aristotle seems to mean that inquiry about principles belongs to the method of dialectic rather than to other methods, or in any case to it more than to them, and not that it is the most proper use of dialectic by comparison with the other uses to which it may be put. 30 It is noteworthy that on the rare occasions when he touches on the issue at all in the Topics, Aristotle seems to view the philosopher as a solitary inquirer whose characteristic activity is distinguished from dialectic in part by the fact that it does not involve relations to another (Topics 8.1.155b1016; Sophistic Refutations 16.175a912). Thus inTopics 8.1, Aristotle notes that what has gone before, viz. the discus- sion of the topoi, may be of use to the philosopher engaged in solitary inquiry as well as the dialectician, whereas the discussion of how to order and pose questions that is to follow will not be relevant to his concerns since it applies only to arguments conducted with others (155b716). This implies that philosophers can derive benets from the dialectical method, or parts of it, without engaging in the practice of dialectic and while remaining ignorant of some of its precepts. This seems to be conrmed in what is by far Aristotles most illuminating remark about dialectics usefulness to philosophy: Withregardtoknowledge andphilosophical wisdombeingable tograsp andto have graspedwhat follows fromeach hypothesis is an instrument (o) of nolittle value. For it remains tochoose one of themcorrectly, and for this one must have a good nature; this is the good nature that regards the truth, the ability to choose the true and ee the false. Topics 163b916 The context for this remark is Topics 8.14, which is about dialectical training (|). The training in question is not that in which the practice of dialectic consists, however, but training or preparation for this practice. Such training can be performed alone or in company (163b34), and it seems that the philosopher may prot from it without engaging in the practice of dialectic by going through the arguments that one might employ in an actual dialectical encounter. Of course, there is every reason to suppose that philosophers will also prot from such participation. The abilities and understanding developed by preparatory exercises will be further sharpened by the 30 Brunschwig, Topiques, 117 n. 2. This content downloaded from 192.167.204.6 on Sun, 21 Sep 2014 02:48:48 AM All use subject to JSTOR Terms and Conditions Aristotle on the Disciplines of Argument 105 actual practice of dialectic, andthere may be others which can only be acquired in this way. It is clear, then, that Aristotle takes the method of dialectic to be of use to philosophy in both ways, directly without being applied in the practice of dialectic and through being used in that practice, though it is the rst that he emphasizes in the Topics. The picture that emerges differs in some ways from the expecta- tions we might take away from, e.g., Nicomachean Ethics 1.1. There Aristotle tells us that every art and every method has its own end. But although it is clear that an arts end does not always tell us why anyone might care to practice itfor that we must look to the broader context and perhaps ultimately to practical wisdom or political sci- ence and the human good which is their objectit is implied that when an art or method is of use relative to goods other than its own, proper end, it is through achieving that end. As we have seen, however, more often than not the uses of the dialectical method that Aristotle has in mind are available to those who do not participate in the practice of dialectic. Although it answers our question about the dialectical methods usefulness to philosophy at a certain level of generality, this leaves unclear precisely how, as a form of |, the actual practice of dialectic is supposed to contribute to philosophy. One kind of interpretation sees a sharp distinction between dialectical activity and philosophy, i.e., between what one does as a dialectician and what one does as a philosopher. On this view, the practice of dialectic contributes skill or virtuosity in argument to philosophy rather than a deeper understanding of philosophical issues. 31 At the opposite extreme are interpretations which suppose that in practicing dialectic one is, at least much of the time, practicing philosophy and that a good part of the philosophers time, as a philosopher, will be occupied with the practice of dialectic. Adistinctionof the kindpositedbythe rst kindof interpretation is easiest to see when the skill and argumentative facility being developed by dialectic are exercised in relation to objects different from those tackled by philosophy. We may compare the way in which the participants in a Platonic dialogue will sometimes warm up, as it were, by dening an item or working through a division of little intrinsic interest before tackling a more serious question (e.g., Politicus, 285d-286a). In fact, the issues regularly cited as examples in the central books of the Topics are serious and not the triing matters that one would expect if the sole object in discussing them were to 31 Moraux, La joute dialectique, cited in n. 15 above, p. 308. This content downloaded from 192.167.204.6 on Sun, 21 Sep 2014 02:48:48 AM All use subject to JSTOR Terms and Conditions RHETORI CA 106 cultivate facility in argument. 32 But interpretations of the second kind are hard to square either with the Topics conception of dialectic as a form of training or the way in which Aristotle sometimes contrasts dialectic and philosophy in the treatises. It may help to bear in mind that some forms of practice or exercise are manifestly different from the activities for which they prepare or train those engaged in them, while others differ chiey by being performed as exercises or for the sake of practice and that there are many degrees in between. In Topics 1.11 Aristotle considers what is to count as a dialectical problem. The examples he cites, e.g., whether the universe is eternal or not (104b16), show that serious philosophical questions are not excluded. He does, however, insist that those engaged in dialectic will not tackle questions the demonstration of which is near to hand or those of which it is too distant; the former because they do not present a challenge, the latter because the challenge they present is more than accords with the purposes of gymnastic (more than is v () (105a79). Though some questions of interest to philosophy may be excluded, discussion of many others is plainly permitted. I suspect that we shall come closest to grasping the distinction between dialectic and philosophy if we suppose that the qualication so far as accords with the purposes of gymnastic that Aristotle uses to restrict the scope of dialectical discussion can also be used to characterize the way dialectic tackles questions that it shares with philosophy. An argument or discussion will be dialectical to the extent that it accords with the purposes of gymnastic, and this will depend in part on whether it connes itself to the resources available to the dialectician but also how far and in what spirit the discussion is pursued. Nothing prevents the practice of dialectic from deepening the participants understanding of the issues at the same time as it sharpens their argumentative skills. Putting philosophical theses to the test (.) and subjecting them to investigation (c) should serve both these ends. But when the understanding amounts to knowledge, dialectic is left behind. That the transition from the dialectical discussion of an issue to one grounded in knowledge and understanding of the eldto which it belongs will sometimes be hard 32 Moraux, La joute dialectique, acknowledges this, but takes it as further evidence for the old view that the peripheral books, 1 and 8, are more recent than the central books. On his view, book 8 marks the emergence of a new perspective, which no longer identied philosophical research with dialectical discussion, but instead viewed the latter as a form of intellectual training sharply distinct from the former. This content downloaded from 192.167.204.6 on Sun, 21 Sep 2014 02:48:48 AM All use subject to JSTOR Terms and Conditions Aristotle on the Disciplines of Argument 107 to mark is a point Aristotle makes himself, when he observes that it is possible to pass by degrees without noticing it from discussing a matter in the way proper to dialectic to discussing it in the way that belongs to the science under which the matter falls (Rhetoric 1.2.1358a236). The last question, about the second and third uses of the dialecti- cal method, has taken us some way from our main purpose, which was to discover how Aristotle could have regarded dialectic as the master discipline of argument, but it has helpedadvance that purpose by throwing further light on the complex unity that the Aristotelian discipline of dialectic possessed. We are now in a position to see how dialectic has been able to present such a variety of aspects to different observers. Viewed from one angle, it is a method tailored to the spe- cial demands of a certain form of philosophical discussion. Viewed from another, it is the proper home of general reections about ar- gument of interest to every pursuit of which argument is a part. This complex unity also helps explainother features of the dialec- tical method as it is set out in the Topics. As we have seen, Aristotle holds that the consideration of every syllogism without distinction (oc) is the business of dialectic, either dialectic as a whole or a part of it (Rhetoric 1.1.1355a810). Presumably he raises the possi- bility that general discussion of the syllogism belongs to a part of dialectic rather than the whole because, as I have already suggested, the whole of it will inevitably contain much that is of use only to prac- titioners of specically dialectical forms of argument. Yet though the Topics contains much that has a bearing on forms of argument other than the dialectical, Aristotle makes no effort to gather this material in one place corresponding to the part of dialectic that he envisages or to tackle it in a way that makes it readily available to other dis- ciplines. The dialectical method unites those elements necessary to equip practitioners for successful participation in gymnastic argu- ment, and it tackles facts about argument of wider interest when and in ways that are dictated by this purpose. In addition, dialectics standing as an art of argument means that neither its boundaries nor even those of the part of it concerned with the syllogism in general will coincide with those of analytic. The Topics does, to be sure, contain much informal discussion of matters that are tackled with the aid of formal logical theory in the Prior Analytics. It has something to say about the premises and conclu- sions of syllogisms, though in a way that is peculiarly adapted to the dialectical arguments where the differences between the predicables matter as they do not in rhetoric (Topics 1.4, 1011). But it cannot be taken for granted that the understanding of from what and how This content downloaded from 192.167.204.6 on Sun, 21 Sep 2014 02:48:48 AM All use subject to JSTOR Terms and Conditions RHETORI CA 108 a syllogism arises, that the orator derives from dialecticlanguage that is remarkably similar to that of the Prior Analyticsis conned to the formor structure of syllogistic premises and conclusions (Rhetoric 1.1.1355a1112; Prior Analytics 1.4.25b267). It may embrace the rep- utable opinions which will be of so much service in encounters with the many (Topics 1.2.101a3034; Rhetoric 1.1.1355a279, 2.1356b324, 2.22.1396b411). Andit verylikelyextends tothe topoi, the elements of the method of invention on which, as Aristotle emphasizes, both di- alectic and rhetoric rely (Rhetoric 1.2.1358a1032) and to which much the largest part of the Topics is dedicated. Here too, however, rhetoric cannot simply take from the method of dialectic what it needs. The Topics presentation of the topoi divides into four parts, one for each of the four predicables. The difference between the different ways in which a predicate belongs to a subject is immaterial in rhetoric, and though many of the topoi in the Rhetoric resemble dialectical topoi, all reference to the predicables disappears. Other topoi in the Rhetoric have no analogues in the Topics (Rhetoric 2.23). What is more, to judge by the Topics and Sophistical Refutations, a knowledge of the syllogismin general will require a thorough under- standing of how to construct fallacious arguments. To the surprise of present-day students of Aristotle, the fullest discussion of the syllo- gism is not found in the Topics proper, but in Sophistical Refutations 6. There Aristotle analyzes each of the seven forms of fallacy not due to language as instances of ignoratio elenchi by showing how each of them violates a part of the denition of the syllogism (168a23; cf. 8, 169b40). Some of the apparent oddness of this procedure disappears, however, if we keep in mind that Aristotle has the needs of the partic- ipants in a practice of argument constantly in view. One must know how fallacious arguments arise if one is to detect and solve them, that is, to reveal why an apparently valid syllogism is in fact invalid (Sophistical Refutations 24, 179b234). This ability will come into its own most obviously when one is faced with deliberately fallacious arguments put to one by others. Yet here too a double perspective is in evidence; Aristotle maintains that it will also help the philosopher avoid inadvertent errors in his own reasoning (Sophistical Refutations 16, 175a912). If the argument of this paper is on the right lines, then admiration for Aristotles invention of logic should not blind us to the existence of an earlier phase in his thinking about argument. In it the place of the master discipline of argument later to be occupiedbyanalytic was occupied instead by dialectic, a discipline that owed its priority not, as analytic was to do, to a concern with the formof valid argument in general, but to its special relation to a favored practice of argument. This content downloaded from 192.167.204.6 on Sun, 21 Sep 2014 02:48:48 AM All use subject to JSTOR Terms and Conditions