[In the following essay, Calvert summarizes the critical debate over whether Plato's ideal republic would include slaves or not, concluding that although the standard critical view supports the existence of slavery in the republic, there is an equally balanced argument opposing the existence of the practice.]

I.

For a number of years, in the not too distant past, there was a lively debate between Plato's defenders and critics over the question of whether his Republic contained slaves. However, since the appearance of an article by Gregory Vlastos¹ some twenty years ago, it seems to have been generally felt that the issue has been resolved, and the controversy has died down. Vlastos argued that the evidence admits of no doubt—Plato included slaves in his ideal state. In this paper, I wish to have the case reopened, and to revive interest in what I believe should continue to be a matter of debate. In opposition to what has become the standard view, I am inclined to think, on balance, that his Republic could not contain slaves.

Vlastos begins by reminding us that, on those occasions when Plato wants to propose a radical change from existing institutions, he argues for such a change. If he had intended to abolish slavery from his ideal society, this would have been a radical change, and we should have expected him to indicate clearly such an intention, and to justify it. Since no justification is forthcoming, we may presume that no change is envisaged. This line of argument I shall call the presumptive argument. One version of it is mentioned by R. B. Levinson,² namely that the rough and dirty work, carried on behind the scene by slaves, will take place as usual, and the continuance of slavery is assumed without question.

Secondly, Vlastos maintains that while there is nothing in the text worthy of the name of evidence to undermine his presumptive argument, there is textual evidence to support it. The decisive passage is to be found at 433d, where Plato says that it is difficult to decide whether anything would contribute more to the well-being of the city than ‘the presence [of justice] in child and woman, slave and free, artisan, ruler and ruled, namely that each performed his own task as one person and was not meddling with that of others’.

Among Plato's defenders, R. B. Levinson³ was the only one to appreciate the damaging force of this passage. He tried to soften its impact by suggesting that Plato may have temporarily forgotten that he is referring to inhabitants of an ideal society and is simply talking about the conditions of man in an ordinary Greek city. Against this Vlastos⁴ has argued convincingly that Levinson's suggestion cannot even stand as a remote possibility. On this point Vlastos is surely right. There is no plausible way to deny that 433d is referring to the ideal society, nor is it possible to dismiss or overlook the fact that it makes specific mention of slaves. In which case, given that his initial presumption is now combined with and confirmed by undeniable textual evidence, his position certainly looks very strong.⁵

Is there any way to undermine his account? I believe there is. That his presumption has support from the text does not mean that there is no evidence to count against it. When Vlastos claims that there is nothing worthy of the name of evidence to undercut his argument, he overstates his case—and this is the first point I wish to establish. Secondly, I shall attack the presumptive argument directly, and show that it is not as convincing as it might appear.

II

To begin, let us turn to the text. An important passage occurs at the very beginning of the construction of the ideal society, at 371d-e:

‘There are others to serve … who are not worth admitting into our company for their intelligence, but they have sufficient physical strength for heavy labour. They sell the use of their strength; their reward for this is a wage and they are, I think, called wage earners. Is that not so?’

‘Certainly.’

‘So the wage earners complete our city?’

‘I think so.’

The suggestion is sometimes made that Plato assumes that what Levinson calls the ‘rough and dirty work’ is carried on behind the scenes by slaves, and their existence is simply taken for granted. But surely this passage makes it clear that Plato has not forgotten a category of people whose only natural talent is for the drudge work of heavy labour. This passage is significant for two reasons. Firstly, Plato makes it explicit that these people will sell their labour. They are not slaves whose labour is simply extracted from them. They are wage earners, and one way in which Plato describes the third segment of his society is as the class of wage earners (cf. 434c), by contrast with the non wage-earning function of the guardians (463b). Secondly, when Plato says that they fill the city, and that it is complete, he surely means that people of this class are citizens; they are members of society and not mere inhabitants outside its membership. They are not slaves, but a part of the wage-earning segment of society. They may not have the knowledge and skill of expert craftsmen; none the less, it is to this group that they belong. In this respect, Plato differs from Aristotle; for the latter, those who can only offer the use of their bodies are by nature slaves.⁶

This passage certainly does not exclude the possibility of the presence of slaves. But what it does do is to undermine two reasons for slavery. Against Levinson, it denies that there is any category of work for slaves to perform, and against Aristotle, it denies that deficiency in intellect is sufficient justification for enslavement.

Now I do not wish to overestimate its importance. It does not contain a specific denial of slavery in the way in which 433d makes a direct mention of it. Nevertheless, I believe it does have a limited value. Despite Vlastos' contention that there is nothing in the text to count against his presumption,⁷ it does suffice to provide some opposing evidence. It corrects the misleading impression that, though very little is actually said, all that is said lies clearly on the one side. In which case, if the text is at least partly ambiguous in its stance on slavery, we may begin to feel some doubt about the strength of the evidence for his initial presumption.

III

As I hope to show, there are even more powerful presumptive considerations on the other side. These counter-presumptive arguments share one main feature—namely, that even though Plato may say very little which is explicit, the structure of his ideal society precludes the possibility of slavery. One such argument is mentioned by Levinson:

Whoever wishes to speak of Plato's Republic as a ‘slave society’ must be prepared to face the paradox that nowhere in its construction is there any need of slaves; they would come dangerously near to being a leisure class! Plato has reserved the work for his citizens, who must be saved from idleness, profligacy, and flitting from one occupation to another, which Plato believed he had seen, and had disliked at Athens.⁸

Perhaps surprisingly, Levinson provides one obvious rejoinder to this argument a little earlier;⁹ not all the work is reserved for citizens. What he calls the ‘rough and dirty work’ is carried on behind the scenes and taken for granted. But to this, a further response can be made, one which has been given above in the discussion of 371e. Those who do the drudge work are a part of the citizen body; they are members of the wage-earning class. The talent they have may be very limited, but in so far as a person can fulfil his potential within the polis, he is not a slave, but a member of society. This kind of ‘counter-presumptive’ argument, we note, depends upon trying to prove that there is no place for slaves because there is no work for them to do.

A second version of the argument focuses on the ownership of slaves. Who would they belong to? Since the guardians are forbidden any private property, they have to be excluded. Vlastos' answer is to say that the slaves are the property of the third class of citizens.¹⁰ However, in an earlier essay Vlastos also argues that ‘it is an axiom of Plato's political theory that the only one fit to rule is he who possesses Logos’.¹¹ Nor is it the case, according to Vlastos, that Plato sees any essential difference between the ruling of people and the ownership of slaves; in fact, the two situations are exactly parallel to each other. ‘… in principle, there is no difference in Plato's political theory between the relation of a master to his slave and of a sovereign to his subjects.’¹² In which case, since the possession of Logos is limited to the philosophers in the Republic, it follows that only the philosophers are entitled to exercise absolute authority over other human beings, and the third class who lack Logos are disqualified.

In his later article Vlastos plainly seems to be aware of this difficulty. He writes ‘the title to absolute authority over persons is conferred by philosophic wisdom. This would legitimize slave ownership by philosophers. It could not also legitimize slave ownership where it would be most urgently needed: in the economic sector, which is wholly in the hands of non-philosophers.’¹³ Yet despite his clear recognition of the problem, Vlastos appears to overlook its impact when, in the same article, he ascribes ownership of slaves to the third class, and we are left with what certainly looks like a blatant incoherence in his account:¹⁴ the third class are not qualified to be slave owners, or, alternatively, the guardians are the only ones entitled to own slaves, but the guardians are prohibited from possessing any private property including slaves.

Nor does this difficulty merely attach to Vlastos' treatment. It is more general. The question remains—if there are slaves in the ideal society, who could own them? The answer appears to be—no one.

A final version of the argument depends upon the claim that slavery is incompatible with the theories of justice and of the tripartite soul. The general theme of the theory of justice may be stated as follows: each person should fulfil his individual personality and talents in such a way as to benefit both himself and society. In conjunction with the doctrine of the parts of the soul, this would have the implication that if there are to be slaves, there would have to be people with naturally slavish personalities and talents. Justice would then be served when a person born with a naturally slavish soul acted out his life in accordance with this particular natural characteristic. But these requirements for a doctrine of just slavery in a Platonic society are impossible to satisfy. We have already seen that there is no specific slavish talent, in the sense that there is no job or societal function which a slave is uniquely fitted to perform.

Not only this. In order for there to be a naturally slavish personality, Plato would have to have distinguished a fourth part of the soul, the slavish. Suppose this fourth part to be ‘in’ the whole soul, together with the other three. His account should then read that each soul consists of four parts, and that justice demands that those people in whom the naturally slavish element predominates over the other three should live their lives in accordance with this predominance. Now Plato, of course, does not in fact distinguish this fourth part. Moreover, if we read what he says at 435b ff. (and especially at 441c), the overwhelming impression emerges that he regarded his tripartite analysis as exhaustive and complete. He does not allow for the possibility of a fourth part in addition to the other three.¹⁵ The three parts form a unity; they are not a part of a unity. Thus far, then, the distinction which would be necessary for the existence of a naturally slavish personality cannot be sustained.

The only other way to avoid this problem would be to postulate that the tripartite analysis is intended to apply solely to citizens. Slaves, who are not citizens, would have a quite different kind of soul, composed entirely of the fourth element, and with no admixture of the other three. In which case, the unity of the tripartite analysis would be preserved, and it would still be possible to maintain a consistency between slavery and the demands of Platonic justice. This alternative, however, may be safely ruled out. The nearest description which we have to the naturally slavish personality occurs at 371e, and such a person is a citizen. But, most importantly, if there were a category of naturally slavish people, it would require that there be some unique job or set of jobs which they should perform; and we have just seen that there are no unique tasks or jobs to be reserved exclusively for slaves.¹⁶ As a result, we are left with the conclusion that an analysis of the implications of the theories of justice and the soul cannot be reconciled with the institution of slavery.¹⁷

With this in mind, let us return for a moment to Vlastos' account. We have seen that his interpretation of 433d depends on the notion that there are slaves who are inhabitants in, but not citizens of the ideal society. We have already presented some evidence to counter-balance this claim. But in the present context there is an additional consideration. Plato is quite specific that the three parts of the soul correspond to the three classes of society. On Vlastos' account, the slaves form a fourth sector exterior to the three classes of citizens. But in order for the city-soul correspondence to hold, this would have to mean that there was a fourth element of the personality, outside the unity of the other three parts. This situation has just been eliminated as impossible. Vlastos might counter with the suggestion that here we have a point at which the city-soul correspondence breaks down. In order for this suggestion to be plausible, however, we must surely be in a position to say that the evidence in the Republic for the inclusion of slaves is stronger than the need to preserve the correspondence. The need in Book IV for the city-soul correspondence is so obvious and strong, and so important for the way in which he constructs his theory of justice, that it is difficult to believe that Plato would have been willing to admit a breaking of the correspondence.¹⁸

This final version of the argument also puts us in a position to assess the textual evidence of the previous section from a different perspective. Even if the critics of Plato were right in claiming that one or more passages, considered in isolation, show that Plato permitted or advocated slavery, to this the response is that he ought not to have done so, because to have done so would be inconsistent with his theories of justice and the tripartite soul. The worst accusation which could be made is that Plato does not always live up to the standards he set himself, and occasionally says things which are inconsistent with the requirements of his own theory. If slaves are mentioned at 433d as inhabitants of his ideal society, it is not because he has temporarily forgotten what he should have said. There is no loss of memory, but rather a lapse from the implications of the ideals which his theory demands. (If Plato had not simply mentioned slavery at 433, but had actually sought to defend it, he would not have been able to do so consistently.)

An analogy can be drawn with his treatment of women. In Book V, some women are to be admitted to the class of guardians, on the grounds that they have the requisite natural capabilities. Yet this does not prevent Plato from sometimes making generalised derogatory remarks about all women, and when he does this, he is falling short of the standards required by his theory of justice.

In retrospect, it is perhaps a pity that in the Republic Plato did not make his position on slavery as explicit as his views on female guardians. It is intriguing to speculate what form this position would have taken. However, if I am right, I believe it would have been extremely difficult for him (to say the least) to provide a sufficient justification for its inclusion. I have argued that if he had intended to include slaves, he would have had the problems of what role they were to perform and who was to own them, not to mention the difficulties with the state-soul analogy. But suppose we exclude them—is anything lost? Would the state cease to function, or function only at the cost of considerable inconvenience to its citizens? Not as far as I can see. Unlike the actual Athenian society of his day which did rely on slave labour, there is no position for slaves to fill in the ideal society of the Republic, no vital role for them to perform. In the words of The Mikado, they never would be missed.

Notes

1. Gregory Vlastos, ‘Does Slavery Exist in Plato's Republic?’, CP [Classical Philology] (1968), reprinted in Platonic Studies (Princeton U.P., 1973; henceforth P.S.).

2. R. B. Levinson, In Defense of Plato (Harvard U.P., 1953), 163.

3. Levinson, p. 171.

4. Vlastos, P.S. 145-6.

5. One might in fact say that so strong has Vlastos' argument appeared to subsequent commentators and scholars that it is simply accepted without question. In her recent book, An Introduction to Plato's Republic (Oxford, 1981), 171, Julia Annas, without mentioning Vlastos, refers to 433d as the one passage from which ‘we know that there are slaves’.

6. For instance, at Politics I, 1255b21ff. Aristotle tells us that the natural slave does not possess reason, that he is useful, like a domestic animal, for the provision of the necessities of life by bodily service, and that nature intends (though it does not always succeed) to make the body of the natural slave different from that of the free man. The body of the slave ought to be suited for heavy labour.

7. At the conclusion of his paper (p. 146), Vlastos is quite adamant on this point. He asserts that ‘there is no contrary evidence’, and it is this complete absence of evidence to the contrary which he sees as making his initial presumption acceptable.

8. Levinson, p. 171.

9. Levinson, p. 169.

10. Vlastos, P.S. 141 n. 6. He gives a similar account in ‘The Theory of Social Justice in the Republic’, in Interpretations of Plato (ed. H. North, Leiden, 1977), 25 n. 92.

11. Vlastos, ‘Slavery in Plato's Thought’, reprinted in P.S. 149.

12. Vlastos, P.S. 151.

13. Vlastos, P.S. 142.

14. I hesitate to accuse such a careful scholar as Vlastos of this inconsistency, and doubly so when he acknowledges its existence. My only explanation for this lapse is that in the context he is primarily concerned that a passage from Morrow should not be misinterpreted, but that in so doing he overlooks the damage he does to his own case. While he says (P.S. 142 n. 13) that he failed to reckon with the problem in ‘Slavery in Plato's Thought’, n. 12 seems to make it clear that he sticks by the main conclusions of that article, in particular the view that the kind of authority involved in ruling or owning people is restricted to those who have the necessary logos.

15. This is, perhaps, contradicted by 443d, when Plato compares the parts of the soul to a musical harmony, and refers to the possibility of other parts lying in between (εἰ ἄλλα ἄττα μεταξὺ τυγχάνει ὄντα). But even if we grant the possibility of there being more parts, they would not lie outside the range of the other three. They would be mixtures of or intermediates between existing parts. A slavish element could not exist in between the existing parts.

16. In his exposition of 433d, Vlastos maintains that slaves contribute to the excellence of the polis (P.S. 141, 125), and certainly this seems to be Plato's message at this point. However Plato tells us nothing about what the slaves would do, and is silent on the matter of how exactly they would contribute to the well-being of society. According to Vlastos, as possessions of the third class, they would presumably function as manual labourers, in which role he suggests they would be ‘most urgently wanted’ (P.S. 142). But in ‘The Theory of Social Justice in the Republic’ (n. 10 above), 28ff, and especially on p. 30, Vlastos recognises the position of the unskilled worker who is a citizen, and, as I have argued, it is this citizen who performs the traditional slave labour. Manual labourers are certainly needed in the Republic; otherwise, as Vlastos aptly puts it, ‘the philosopher … would then have to divert precious time and energy to meaningless drudgery’ (p. 30). But, to repeat, these manual labourers are citizens not slaves.

17. This conclusion is reinforced by what Plato says at 454b ff. There he makes the important point that if we treat people differently, we must be able to appeal to some relevant difference in order to justify such differential treatment. So if slavery is to be defended as an institution, it would be necessary to pick out some relevant difference between slave and non-slave. But no relevant differences, either in terms of personality or suitability to a particular kind of job, can be found. In that case, there is no justification for claiming that anyone ought to be a slave.

18. The nature of the city-soul correspondence should not be misunderstood. I do not intend the analogy to be so strict that, corresponding to the cobblers and carpenters in society, there must exist an identifiable cobbling or carpenting part of the soul. Rather, if a person is suited to do one or more of the range of jobs assigned to the money-making class, this will mean that the appetitive part of the soul predominates. Plato's formulation of the principles of justice has left open a certain flexibility within the third class for people to exchange jobs, or to do more than one job (434a-b). But this flexibility does not affect the correspondence between the appetitive part of the soul and the third class of citizens. The point remains that if there is to be a justifiable doctrine of slavery, this would require a fourth part of the soul corresponding to the slave class.