Robert Colter (Centre College)


            Any reader of Plato’s Republic sees significant differences between Book I and the rest of the Republic. Book I is dominated by Thrasymachus, perhaps the most memorable character in the platonic corpus other than Socrates himself, whereas II-X are dominated by the much less interesting Glaucon and Adeimantus. The role Socrates plays in the discussion is also importantly different. In Book I he engages in arguments designed to refute his interlocutors and denies having knowledge himself, but in II-X he engages in positive theorizing toward a definite conclusion. Finally, the philosophical methodology seems importantly different. All these differences can be summed up in the claim that Book I bears a striking similarity to the so-called “early” or “socratic” dialogues, and II-X are paradigmatically “middle.”

            I will argue that the different philosophical methods introduced at the beginning of Republic II and which continue for the remainder of the dialogue are signaled in Book I. That is, certain features of the exchanges between Socrates and Thrasymachus highlight two crises in the so-called “socratic method.” I will illustrate certain facts about the arguments in the text that indicate a certain awareness of, and preparation for, developments in later books of the Republic, in particular of a methodological shift often taken to be indicative of Plato’s transformation from an author who transcribed the conversations of his teacher Socrates into an independent philosopher in his own right. Regardless of whether one thinks there is any such shift, I submit that it does not occur in the transition between Republic I and the remainder of the work.

            Precisely what counts as distinctly “socratic” is a subject of controversy. There are, however, two features of “socratic method” on which there is general agreement: the use of the craft analogy and the use of the elenchus. In Republic I, I shall argue, both features of socratic method are present, but featured in a fashion designed to highlight Plato’s realization, or at least presentation, of weaknesses in both aspects of socratic investigation.




            One of Socrates’ favorite argumentative devices is the Craft Analogy (CA), in which he compares the virtues to crafts. The salient features of the crafts are that each craft has a determinate domain or object and a product. At Euthyphro 13a-14b, Socrates uses the features of a number of crafts to indicate to Euthyphro that his definition of piety as care for the gods must be mistaken. In the examples Socrates uses, the crafts have a determinate domain such as horses, dogs, or cattle, and a product, namely to make the objects or members of the domain better. If piety is also a craft, then it must have the same features of a domain and a product. The product in the case of Euthyphro’s definition of piety as care for the gods must be analogous in that piety must produce the improvement of the gods, just as horse-breeding produces the improvement of horses. This cannot be what Euthyphro means, because gods do not need improvement. Since Euthyphro’s definition is not analogous to a craft, it must be mistaken. In general, then, the fact that a virtue is not analogous to the crafts in general or the specific craft under consideration is enough to show that the interlocutor has not defined the virtue.

            This methodology has two major presuppositions. First, it presupposes that the virtues are analogous to the crafts in all salient aspects. The second assumption is that all crafts have the same structure and are analogous to one another. If these assumptions do not hold, then the use of the CA as a method of refutation is called into question because various crafts might have different structures and a definition of virtue need not make that virtue be analogous to every craft, and thus the CA is not a reliable method of investigation.

            Now, consider the use of the CA in Republic I. Socrates uses the CA to show that Thrasymachus cannot be right that Justice is the advantage of the stronger. Thrasymachus has just claimed that the ruler, insofar as he is actually ruling, seeks his own advantage. Socrates replies with the example of a doctor and ship’s captain, and their corresponding crafts. In these cases, Socrates claims, “the crafts rule and are stronger than the things of which they are the crafts … No kind of knowledge seeks or orders what is advantageous to itself, then, but what is advantageous to the weaker, which is subject to it.”[1] That is, in the case of crafts such as medicine, the benefit that they aim to produce is the improvement or advantage of the determinate domain over which they have control. Socrates then concludes from the analogy that “no one in any position of rule, insofar as he is a ruler, seeks or orders what is advantageous to himself, but what is advantageous to his subjects; the ones of whom he is himself the craftsman.”[2] That is, since ruling is a craft, ruling must have the feature that the other crafts have, namely, that it looks to produce the improvement or advantage of its domain.

            There are several important features of this argument. First, Thrasymachus has accepted the idea, implicit in Socrates’ questioning, that ruling is a craft of some sort. As a craft, ruling has to share the features of the other crafts.  The feature Socrates emphasizes here is the relation between the craftsperson and the object of that craft - the things that the craft rules over. So, medicine rules over bodies, and horse-breeding looks after horses, and shoemaking looks after shoes. In all these cases, the principal aim of the craft is to produce or modify the things of which it is the craft. So, the shoemaker rules over shoes and their materials, including leather, hobnails, etc., in order to make good shoes, the horse-breeder looks after horses in order to produce better horses, and so on.

            So, for Socrates’ argument to be effective, ruling must be a craft. It must have the important features of the other crafts. Also implicit in Socrates’ method is the idea that all the crafts are the same in these respects. This is where Thrasymachus begins his attack.


You think that shepherds and cowherds seek the good of their sheep and cattle, and fatten them and take care of them, looking to something other than their master’s good and their own. Moreover, you believe that rulers in cities – true rulers, that is – think about their subjects differently than one does about sheep, and that night and day they think about something besides their own advantage. You are so far from understanding about justice and what’s just, about injustice and what’s unjust, that you don’t realize that justice is really the good of another, the advantage of the stronger and the ruler and harmful to the one who obeys and serves.[3]


Thus, Thrasymachus posits the examples of shepherding and cowherding as counterexamples to Socrates’ assumption that all crafts are alike in seeking the good of whatever they rule over. Since shepherding and cowherding are widely accepted as crafts, this poses a serious challenge to Socrates’ argument.

            Socrates responds by distinguishing between the craft of shepherding and the craft of wage-earning. He claims that since each craft has its own benefit shared with no other craft, and since all craftsmen earn wages, then wage earning cannot be identical to the craft of shepherding. Thus when craftsmen earn wages, they practice an additional, auxiliary craft. “We say that the additional craft in question, which benefits the craftsmen by earning them wages, is the craft of wage-earning.”[4] Socrates concludes that the benefit that a shepherd or cowherd gets from the practice of his craft is not a feature of that craft, but of the other craft, wage-earning, that they pursue at the same time.

            But this raises the problem again at another level. If wage-earning is a craft - and Socrates calls it one at 346c - then wage-earning is itself a counterexample to the claim that no craft benefits its practitioner. The benefit provided by wage-earning is the wages that the practitioner receives. In the very next exchange, Socrates says,


Then this benefit, receiving wages, doesn’t result from their own craft, but rather, if we’re to examine this precisely, medicine provides health, and wage-earning provides wages; house-building provides a house, and wage-earning, which accompanies it, provides a wage; and so on with the other crafts.”[5]


So, Socrates has committed himself to two ideas: first that wage earning is a craft, and secondly that wage earning provides a benefit – wages - to its practitioner. This is, then, the counterexample Thrasymachus was seeking, but provided by Socrates himself.

            We are now faced with a dilemma. Either wage earning is a second craft, and Socrates is right that shepherding seeks the advantage of the ruled, or wage earning is part of shepherding and Thrasymachus’ counterexample holds. On the first horn, we have another craft, wage-earning, which is itself a counterexample because the benefit it provides is the advantage of its practitioner. Socrates has just claimed that NO craft benefits its practitioner, but rather benefits those over which the craft rules. Thus Socrates’ own response undermines the CA as a method by undermining its presupposition that all crafts are analogous in the requisite aspects. The second horn equally, but more directly, undermines the CA.[6]





The “elenchus” is Socrates’ method of refuting his interlocutors. Gregory Vlastos describes it thus: “Socratic elenchus is a search for moral truth by question-and-answer adversary argument in which a thesis is debated only if asserted as the answerer’s own belief and is regarded as refuted only if its negation is deduced from his own beliefs.”[7] Vlastos offers the following schema:



1. The interlocutor asserts a thesis, p, which Socrates considers false and targets for refutation

2. Socrates secures agreement to further premises, say q and r.

3. Socrates then argues, and the interlocutor agrees, that q and r entail not-p

4. Socrates then claims that he has shown that not-p is true, p false.


There has been quite a bit of debate about whether Socrates can legitimately claim to have shown that p is false. Vlastos is certainly correct in noting that Socrates often claims to have shown something, but it seems clear that he is not logically entitled to make such a claim.[8] What Socrates has shown is that his interlocutor has an inconsistent set of beliefs, but this does not require that any particular belief is false, merely that at least one of them must be false.

            Let’s consider a typical example. In the Euthyphro the following argument appears:[9]


1. X is pious just in case x is loved by all the gods.

2. X  is loved because it is pious.

3. It is not the case that x is pious because it is loved.

4. X is a loved thing and god-loved because it is loved by the gods.

5. It is not the case that x is loved by the gods because it is god-loved.

6. “So, the god-loved is not the pious … nor is the pious the god-loved, as you say, but the one is different from the other.”[10]


Thus, by exposing an inconsistency in Euthyphro’s beliefs, in particular between 4 and 5, Socrates concludes 6, that Euthyphro’s definition is false. Note that in this example, the interlocutor accepts without objection Socrates’ claim that the thesis they have proposed must be false, and given their other commitments it seems perfectly reasonable.

            But, as noted above, nothing in the logic of the argument requires that the interlocutor reject that claim. As far as logic goes, all we know is that at least one of the statements is false. As Vlastos puts it:

“how is it that Socrates claims to have proved a thesis false when, in point of logic, all he has proved is that the thesis is inconsistent with the conjunction of agreed-upon premises for which no reason has been given in that argument? Could he be blind to the fact that logic does not warrant that claim?”[11]


Vlastos calls this “the problem of the elenchus.” Mere inconsistency does not establish the falseness of any particular belief.

            Once again, let’s compare the exchange between Socrates and Thrasymachus beginning at 348b:

Come, then, Thrasymachus, I said, and answer us from the beginning. You say that complete injustice is more profitable that complete justice?

I certainly do say that, and I’ve told you why.

Well, then, what do you say about this? Do you call one of the two a virtue and the other a vice?

Of course.

That is to say, you call justice a virtue and injustice a vice?

That’s hardly likely, since I say that injustice is profitable and justice isn’t.

Then, what do you say exactly?

The opposite.

That justice is a vice?

No, just high-minded simplicity.

Then do you call being unjust being low-minded?

No, I call it good judgment.

You consider unjust people, then, Thrasymachus, to be clever and good?

Yes, those who are completely unjust, who can bring cities and whole communities under their power. …

I’m not unaware of what you want to say. But I wonder at this: Do you really include injustice with virtue and wisdom, and justice with their opposites?

I certainly do.

That’s harder, and it isn’t easy now to know what to say. If you had declared that injustice is more profitable, but agreed that it is a vice or shameful, as some others do, we could have discussed the matter on the basis of conventional beliefs. But now, obviously, you’ll say that injustice is fine and strong and apply to it all the attributes that we used to apply to justice, since you dare to include it with virtue and wisdom.

You’ve divined my view exactly. (348b-349a)


There are several feature of this exchange that require close scrutiny. Thrasymachus offers his definition, and his corollary that injustice is more profitable that justice. Socrates proceeds by getting Thrasymachus’ agreement to further claims, until Thrasymachus declines to agree that justice is a virtue and injustice a vice. Socrates must stop the discussion in order to get clear on what Thrasymachus has claimed, namely that justice is to be classified with vice and injustice with virtue.

            This raises an important crisis for Socrates’ method of elenchus. Note carefully Socrates’ response, which indicates that he does not quite know how to proceed. He says that his job is “harder” since Thrasymachus has reacted differently than any interlocutor had done before. Of course it is “harder,” since Thrasymachus has, in effect, not allowed Socrates to derive the usual conclusion of elenctic examination, namely that some definition or proposal is false. Note also that Socrates then gives a brief synopsis of what he intended to do. If Thrasymachus had agreed, we would have seen another typical elenchus, but as it is, Socrates cannot proceed!

            Socrates’ response here, I am suggesting, indicates that Plato himself was aware of “the problem of the elenchus.” He sees that if an interlocutor were to react differently from the reactions interlocutors such as Euthyphro, the elenctic method would have no further resources from which to proceed. It is precisely this feature which is highlighted by several things that follow.

            Many scholars have noticed that, following this exchange, Thrasymachus seems to merely play along with the examination. He says things such as, “if you like,” “whatever you say.” And so on. Thrasymachus’ views are not refuted and he has lost interest in the conversation because Socrates cannot refute him.[12] This aspect is also underscored by Glaucon and Adeimantus’ demand, at the beginning of Book II, for further argument. They say, “Do you want to seem to have persuaded us that it is better in every way to be just than unjust, or do you want truly to convince us of this.”[13] This clearly indicates that not only was Thrasymachus not convinced that Socrates had won the argument, but neither were Socrates’ friendly, cooperative interlocutors.

            The fact that the socratic method has been exhuasted in the discussion is also highlighted by the nature of the discussion in the rest of the Republic. This difference is of course one of the central differences between these two parts of the Republic. But, the necessity of it has already been both exposed and explained, although without the connection being explicitly drawn for us, in the passages from Republic I that we have discussed.




            The Socratic method has, for good reason, been one of the most influential intellectual developments in history. Any careful examination of it, as delineated above, however, reveals that it is not a method that can bring us to positive results in any reliable, rational way. Scholars have used this fact to make all sorts of inferences about the nature of Plato’s development. What I hope to have shown is that certain features both of socratic method and of the text of Republic I indicate that Plato was well aware of the problems and weaknesses of the method, and showed that he knew he needed a further method to travel the philosophical territory he desired. This is not to say that the elenchus and CA have entirely lost their lustre. They continue to remain an important part of Plato’s philosophical toolbox. What we have seen is that these aspects of socratic method are incomplete.

            Thus, careful scrutiny of Republic I indicates that the differences between Republic I and II-X are perhaps not as great as they might seem. Through Socrates’ argument with Thrasymachus, Plato has highlighted the limitations of Socratic investigation. The new methods of philosophizing introduced in book II are, I suggest, Plato’s way of responding to these limitations. As I see it, Plato very clearly wrote the Republic as a unity with respect to both philosophical method and purpose.



[1] 342c

[2] 342e

[3] 343b-c

[4] 346c

[5] 346d

[6] At this point, one might say that wage-earning isn’t a craft at all. Wage-earning might just be something that accompanies many of the other crafts. Thus, wage-earning does not have a determinate domain, since it is related to all the crafts at once, and would fail to be a craft on that account. If this is right, then Socrates could have avoided Thrasymachus’ counterexample, but simply failed to do so.

                But this then raises the problem again at another level. Consider justice, which is of course the main topic of the Republic. Is it a craft, or analogous to the crafts in the requisite respects? It would seem not. The domain of justice, as it is developed in the rest of the Republic, is a relation between persons and parts of the soul. Surely this is far from a determinate domain, but seems as varied and general as the domain of wage-earning. Also, the entire argument of Republic II-X relies on the idea that justice is beneficial to its practitioner, given Glaucon’s challenge and Socrates’ response to it.[6] So, if wage-earning is not a craft for the reasons above, it still seems to call into question the relation between the crafts and the virtues, and thus the CA in general, because it is rather analogous to the virtue of justice, yet not a craft.

[7] Vlastos, Gregory, “The Socratic Elenchus: Method is All,” in Vlastos (1994). Originally in OSAP 1983.

[8] See Irwin, Terence, Plato's Ethics (Oxford, 1995), 18.

[9] 10d-e

[10] See Benson, Hugh H.,  Socratic Wisdom:  the Model of Knowledge in Plato's Early Dialogues.  (NY:  OUP, 2000).

[11] Vlastos (1994, 21)

[12] This is reinforced when, at 354a, Thrasymachus mocks Socrates’ “victory.”

[13] 357a