The Socratic Elenchus
as a Search for Truth

Andrew N. Carpenter

What is the point of the Socratic elenchus? One answer is obvious: it serves to show Socrates' audience that they are confused about their own moral beliefs. It is less clear what constructive role, if any, it plays. Socrates himself seems to offer no positive advice for his audience; after elenctically reducing his interlocutors to confusion, he typically goes on raise additional questions for elenctic treatment. After elenchus, it seems, his audience would still be extremely confused and, although some may now see the need for radical reform of their moral beliefs, they may have become skeptical about the prospects of gaining moral knowledge. Doesn't this square with the image of a philosopher whose divine mission involves using the elenchus to expose the pretense of knowledge in others while all the time maintaining that he too knows nothing? Of course Socrates has positive moral beliefs of his own. The point is that he doesn't conceive of himself as having moral knowledge; his reasons for holding these beliefs, if any, do not arise from elenctic argumentation. Although the elenchus efficiently exposes inconsistencies, apparently it can justify no positive beliefs.

Although this picture of the elenchus is a powerful one, Gregory Vlastos recently has argued persuasively that the elenchus has a powerful positive function, viz. it amounts to a philosophical method for discovering moral truths. Surely, though, the standard picture is right on one account: if the Socratic elenchus can be seen as a method for discovering moral truths, then elenctic arguments must be able to establish more than mere consistency or inconsistency; they must be able to establish the truth or falsity of particular propositions. Yet because the elenchus logically seems capable of doing no more than displaying inconsistencies among beliefs, it is difficult understand how it could lead to moral knowledge. Vlastos has characterized this as "the question about the elenchus" (e.g., at Vlastos, 1983:73).

 Is Vlastos' question a non-starter?

Of course, Vlastos' problem is spurious if the elenchus has no positive epistemological ambitions. In a rough sketch, the elenchus has this form: (1) first, Socrates gets the interlocutor to express some belief, usually one which concerns the definition of a moral concept; (2) next Socrates gets the interlocutor to express other beliefs, many of which concern the interlocutor's intuitions about virtuous actions and his conception of the good; (3) third, Socrates goes on to show that these additional beliefs entail the negation of the first belief; (4) finally, from what has been shown so far--that these beliefs are jointly inconsistent--Socrates proceeds to draw a substantive moral conclusion. To give an example from the Laches, Socrates gets Laches to express his belief that "courage is a sort of endurance of the soul" (1921), after which he gets him to express beliefs about examples of courage, for example that a soldier who "endures in war" knowing that his army vastly outnumbers his opponents is not courageous (193a), and about Laches' conception of the good, for example that courage is "a very noble quality" (192a) and that nothing evil and hateful is noble (192d). Next, Socrates argues that these additional beliefs entail that foolish endurance is not courage, which both he and Laches take to entail the negation of the first belief (193b-193e). Finally, from this Socrates apparently draws a substantive moral conclusion, viz. that Laches' "principle of endurance" is "unsatisfactory"; Socrates' question "and is this condition of ours satisfactory" is immediately and enthusiastically answered by Laches "quite the reverse" (193e).

One large question about the elenchus is the status of the beliefs introduced in the second stage: what reasons, if any, does Socrates have for them and what role, if any, do these reasons play in the elenctic arguments? Notably, it seems that Socrates and his interlocutors agree to these beliefs without paying any heed to reasons for them. This issue is obviously relevant to Vlastos' question: if the second stage beliefs are asserted on a strong enough basis, then there will be some justification for removing the inconsistency by rejecting the initial belief. Perhaps also relevant are such features of the dialogues' "rhetorical context" as that the interlocutor often tries to amend his original belief and the elenchus can continue through a number of steps, though never to a satisfactory conclusion.

Vlastos' problem is so urgent that it suggests a short, swift response: elenctic arguments are only capable of establishing inconsistency. Although this interpretation accords the elenchus a purely negative role--dashing all hopes of finding in it a philosophical method for discovering moral truth-such a minimal elenchus would still have great therapeutic value. By highlighting inconsistency, the elenchus would force its participants to sharpen and refine their moral concepts. Furthermore, it would show the interlocutors the inadequacy of their ordinary moral training and it would teach them the extent to which their ordinary moral beliefs are unstable and are in need of radical revision. Even if elenctic argumentation cannot conclusively show how the interlocutors' moral beliefs should be corrected, the sheer ease elenchus demonstrates the inconsistencies in their beliefs forcefully exhibits exactly how much adjustment is required between, on the one hand, their conceptions of the good and particular virtues and, on the other, their intuitions about examples of virtuous ac- CA action. Even if it itself cannot supply fell-fledged Socratic definitions of the good, this weaker ~t elenchus would force individuals toward them, viz. by forcing them to reflect on and explain the relationship between these types of moral beliefs.

This interpretation is highly problematic for rather obvious reasons. In the Laches Socrates agrees that Nicias "only wants to avoid the appearance of inconsistency" (196b), suggesting that eristic arguments attempt to establish merely verbal or apparent inconsistency. If this is correct, then there would be a clear distinction between eristic and weak elenchus, which, in accordance with Vlastos' 'say what you believe' rule, concerns inconsistency in an interlocutor's beliefs. Certainly the evidence for Vlastos' strong interpretation of the elenchus is not overwhelming. In his favor, the elenchus clearly is at the core of the early Socratic dialogues; whatever philosophic method there is there is contained in the elenchus. Certainly much dramatic and rhetorical weight rests on the outcome of elenctic discussions; as Donald Davidson writes (in a slightly different context), a weak elenchus would provide "relatively small comfort in return for a considerable investment in time and ego" (1992:1). However, perhaps the therapeutic effect of even a weak elenchus explains its prominence in the early dialogues. Nor is it clear, in the light of Socrates' professions of ignorance, that we should look for more in these dialogues; it is difficult question to decide whether these professions are even compatible with the existence of a Socratic philosophical method for discovering truth.

Unfortunately, the textual evidence is highly mixed. In the argument from the Laches cited above, as in most of the others, it is extremely unclear exactly what sorts of conclusions Socrates hopes to draw. On the one hand, after getting r Laches to agree that his original proposal is "unsatisfactory", Socrates immediately turns to consider Niches' claim that "courage is a sort of wisdom" (194d), which is what we would expect him to do if he considered Laches' thesis refuted. On the other hand, he nowhere explicitly claims that Laches' original belief is refuted and, indeed, he implies that it is not, remarking that they should "admit the principle of which we are speaking to a certain extent" (193e). It is at precisely this point, before more elenctic progress is made on Laches' claim, that Niches enters and the discussion turns to his claim.

In his "Afterthoughts" to his first article on the Socratic elenchus, Vlastos admits that in most of the early dialogues there is no decisive evidence that Socrates intends to prove or refute moral claims (1983:71). To be sure, Socrates may always have thought that it could without seeing a need to dwell on this point or reflect on how this is possible. As we shall see, Vlastos holds that Plato's worries about this problem prompted him to supplement the elenchus in various ways.

I think that there are independent philosophical reasons for finding Vlastos' problem is interesting. First, obviously, it would be extremely interesting ethically if it were possible to parlay an argument that a set of propositions is inconsistent into substantive ethical conclusions. Second, the question of the relationship between consistency and truth has independent epistemological interest: quite often it is charged that philosophers (e.g., coherence theorists, Kant in the Postulates) illegitimately move from claims about beliefs' consistency to claims about their truth-value. Finally, Vlastos' problem is engaging just because a Socrates who propounded strong elenctic arguments would be more interesting than one who limited himself to weak ones, especially on the plausible presumption that whatever Socratic philosophical method there is founded on the elenchus.

The dilemma of the elenchus

For these reasons, and despite the underwhelming textual evidence, Vlastos' problem is interesting and legitimate. Before turning to Vlastos' answer to it, I will examine the problem more closely, with an eye to discerning what forms an adequate answer might take.

Recall the problem sketched above: for Socrates to refute p on the grounds that it and other beliefsq and r are inconsistent, it seems that he would have to know (or at least be able to justifiably assert) q and r; otherwise, the interlocutor could preserve p by rejecting one of the other beliefs. There seems to be no way for an elenctic argument, by itself, to settle how the inconsistent beliefs should be revised. This dilemma arises: Does Socrates know in advance which beliefs are true? If so, he can legitimately direct the interlocutor to reject what he knows to be the false beliefs, but there is no philosophical need for the elenchus. At most the elenchus would be useful for rhetorical or pedagogic effect, to help convince someone else of what, on other grounds, Socrates already knows. If not, however, then when the elenchus discovers an inconsistency, Socrates can not know that the belief he proposes jettisoning is not true. In either case, there seems to be no way in which the elenchus, taken by itself, can constitute a method for discovering truth.

 All this suggests several constraints that an adequate response to Vlastos' problem must meet. It seems, first, that a plausible answer won't supplement the elenchus with an independent method for discovering truth. Such a method would clash with Socrates' professions of ignorance and with the sense, very clear in the early dialogues, that the elenchus is the Socratic method. Although formally the elenchus supplemented by such a method could be used to refute an interlocutor's belief (not-p because p, q, and r are inconsistent and we already know that q and r), this is not a plausible route for an interpretation of Socrates. Second, that the elenchus repeatedly fails to reach Socratic definition suggests that a more plausible solution might hold that the elenchus can in principle discover moral truth, although no positive results may be forthcoming from the arguments that Socrates himself gives. That is, the failure to reach a Socratic definition might be construed as suggesting not that the elenchus is incapable of reaching the truth, but rather that it does so progressively, perhaps after a long period of elenctic argumentation. Although pursuing this thought would require a separate discussion of an alleged Socratic commitment to the principle of the priority of definition, what we've al-ready seen could help square the claim that the elenchus constitutes a method for discovering truth with both Socratic ignorance and the fact that elenctic arguments fail to find Socratic definitions.

Next I shall turn to Vlastos own answer to his problem, which appeals to three methodological assumptions to which he thinks Socrates is committed. I shall argue that Vlastos' solution gets caught up in the first horn of the dilemma I just sketched.

 Vlastos' methodological assumptions

These are Vlastos' assumptions:

Everyone's moral beliefs always include a subset of true beliefs which entail the negation of each of his false moral beliefs;

The set of moral beliefs held by Socrates at any given time is consistent;

and, entailed by the proceeding:

 The set of moral beliefs held by Socrates at any given time is true. (1983:52,55)

Before I criticize him, I want to point out the subtlety and elegance of Vlastos' proposal. We saw that Vlastos' problem arises because Socrates' case always rests on statements for which no argument is given. It seems that this cannot lead to a proof or refutation of any moral claim because the interlocutor could simply choose to withdraw support of one or more of these statements. Likewise, Socrates' method seems inconclusive because it always seems possible that a different interlocutor might fail to agree with them at all. Vlastos addresses these problems head on. His subtle move is to admit these possibilities but argue that assumptions A and B make the elenchus capable of finding truth nonetheless. After raising the question of why Socrates' interlocutors could not "hang on to their thesis by welshing on one of the conceded premises", he writes:

Surely Socrates would be aware of this ever present possibility. Why then is he not worried by it? Because, I submit, he believes that if that wrong choice were made, he would have the resources with which to recoup the loss in a further elenchus Is. (1983:49-50)

 Vlastos accepts that any interlocutor could fail to agree with Socrates. The import of his methodological assumptions, however, is that additional elenchus would again and again attack these "wrong choices". Eventually, truth would be reached; by the end of the elenctic road each false belief will have succumb to overwhelming elenctic pressure. That is, even though each argument itself is fallible (a "wrong choice" might be made at any time), when it is extended over time the process of elenchus eventually winnows out all falsehood. This picture promises to vanquish all charges that Socrates is dogmatic: on this model, any moral claim, no matter how obvious it seems or how widely it is accepted is a potential object of elenctic scrutiny; through extended elenchus any belief may be revised. On this picture, then, elenchus affects all parties; far from dogmatically bullying his interlocutors into accepting his propositions, together with them and for the betterment of all concerned Socrates engages in a dialectical process that must produce truth in the end. Unfortunately, Vlastos' assumptions ultimately undercut the appealing epistemic modesty of this picture.

In the passage following the one just cited, Vlastos explains exactly what assumptions A and B are supposed to contribute to the elenchus. He writes:

This, I am suggesting, is his general view: if you disputed him by denying q instead of p, he would be confident that he could start all over again and find other premises inside your belief system to show that you haven't got rid of the trouble--that if you keep p, it will go on making trouble for you, conflicting as much with the other premises as it did with q and r before. (1983:50)

As evidence for Socrates' acceptance of this view, Vlastos cites a passage in the Gorgiaswhere Callicles objects to Polus' agreement that doing injustice is more shameful than suffering it. Polus won't agree to this; he doesn't accept Socrates' elenctic argument against Polus. Socrates, of course, is able to get Callicles, on the basis of other premises, to admit that doing injustice is worse than suffering it (482a-b; cited in Vlastos, 1983:52). To Vlastos, this shows that Socrates accepts assumption A: he is always willing to abandon any of the claims used in defeating an interlocutor because he can always find other premises, believed by that interlocutor, to reach the same conclusion. As he puts it, the thought is that "regardless of which conceded premises they choose to retract, there will always be others in their belief system which entail the Socratic thesis" (ibid.).

Although I don't have time to discuss them here, I think there are many textual difficulties with Vlastos' interpretation. Instead, I want to point out some reasons why Assumption B is quite dubious. It seems wrongheaded in the extreme to attribute to Socrates the belief that all his moral beliefs were consistent. Although Vlastos doesn't want to dismiss his professions of ignorance as shamming irony, it is difficult to see how they are compatible with assumption B. Perhaps Vlastos' thought is that although Socrates is entitled to conclude that "the set of moral beliefs held by him at any given time is true", his ignorance shines through in virtue of his lacking beliefs about the nature of the virtues. If so, however, it is hard to see how the elenchus could constitute a method for discovering moral truths; although Socrates would be in a position to refute interlocutor's beliefs, if he doesn't already have in hand Socratic definitions, he would seemingly not have the appropriate positive beliefs of his own with which to replace them. Worse, in light of his being unable to produce Socratic definitions of the virtues, is seems implausible to attribute to Socrates the claim that those moral beliefs, concepts, and intuitions that he will affirm stand in no need of elenctic therapy. Put somewhat melodramatically, assumption B seems wrongheaded because it makes Socrates out to be uncharacteristically arrogant and overconfident.

It tells more decisively against Vlastos, I think, that at the end of the only dialogue where he sees a textual basis for his interpretation, Socrates explicitly denies assumption B! Socrates provides this pessimistic evaluation of his moral beliefs:

…[I]t seems to be shameful that, being what apparently at this moment we are, we should consider ourselves to be fine fellows, when we can never hold to the same views about the same questions-and these too the most vital of all--so deplorably uneducated are we! (527d)

That the elenctic arguments in theGorgias have disclosed that Socrates himself doesn't "hold to the same views about the same questions" seems tantamount to an admission that there are inconsistencies in Socrates' own moral beliefs. Such an admission is deadly for Vlastos' project: it would show that Socrates isn't entitled to assumption B and thus cannot reach assumption C .

 I will end with a brief reflection that might go some distance towards salvaging Vlastos' interpretative. project. Might not assumption A alone be plenty to justify the use of the elenchus , not only for its power to remove error, but also to increase the probability that one's beliefs are true? If carried out to infinity, the elenchus would generate all true beliefs, for inconsistency leads to false beliefs. Whether or not Socrates assumed it, I think that there are good philosophical reasons for thinking that something much like Assumption A is correct. That it is correct can be seen, I think, not on inductive or other evidence, but on a priori grounds. Donald Davidson's work, in particular, suggests how such an argument might go. A farfetched suggestion? I'll end simply by remarking that this renown contemporary philosopher of language wrote his doctoral dissertation on—the Socratic elenchus.