Thrasymachus on Justice, Rulers, and Laws in Republic I

One issue of contention amongst scholars of the Republic is whether Thrasymachus initially espouses a conventionalist account of justice, according to which just actions are merely those which are lawful; required, or at least allowed, by the laws passed by the ruler of the state. A further question is then whether his initial conceptions of rulers and laws are positivist ones, such that to be a ruler or law of a state is simply determined by the state’s constitution (or indeed by the ruler’s ability to enforce obedience). At 340c Thrasymachus effectively rejects such positivism by placing a condition on being a ruler that one should be exercising the art of ruling and on being a law that it should work to the ruler’s interest. Some have maintained that this works to clarify his initial account and so shows that he was never a positivist about rulers and laws. In this paper I argue against such an understanding of Thrasymachus’ argument and explore the problems which beset the position he is, on my reading, forced into by Socrates’ objection to his opening claim that justice is what is in the interest of the stronger and the argument he first gives to support it.


Republic I
Stephen Everson 1. According to Thrasymachus at Republic I 338c, 'justice is nothing other than the advantage of the stronger'. At 338e, he then declares that 'justice is the same in all cities, the advantage of the established rule' (τὸ τῆς καθεστηκυίας ἀρχῆς συμφέρον, 338e6-339a2). Half-a-Stephanus-page further on, we find him endorsing as his own the claim that 'whatever laws [the rulers] make must be obeyed by their subjects, and this is justice' (339c) and then, after tussling with Socrates for a page and a half, he repeats his claim that it is just to do what is to the advantage of the stronger (341a).
Finally, at 343c, he maintains as if obviously true that 'justice is really the good of another' and this turns his attention from considering the relations between rulers and their subjects to those between people more generally.
That Plato should put such a swarm of different characterisations of justice into Thrasymachus' mouth in such short order hardly seems designed to give his reader confidence that they collectively manifest a single and coherent view of justice, let alone a compelling or even plausible one.
Nevertheless, whilst Plato's commentators have generally been quick to dismiss Socrates' first two interlocutors in Republic I as intellectually inadequate to the task of joining him in the investigation of justice, they have been much less willing to accept at face value Plato's portrait of Thrasymachus as careless or confused. Indeed, it has become something of a standing challenge to the ingenuity and philosophical acumen of Plato's interpreters to show that Thrasymachus' various claims can be given a reading that renders them consistent without making them intolerably vague.
In an earlier paper, 'The Incoherence of Thrasymachus' (hereafter 'IT'), I argued that that challenge is not one that the interpreter of the Republic can, or should even try to, meet. Not only are the claims about justice made by Thrasymachus in the first book of the Republic actually inconsistent, they are intended by Plato to be so. 1 In fact, so I argued, not only is there a fault-line between the claims that justice is what is advantageous the stronger and that it is the good of another, but the first of these is used to articulate two different positions which are also themselves inconsistent.
Amongst a number of attempts by commentators to defend Thrasymachus' philosophical honour against my charges of confusion and incoherence, one of the most sophisticated and textually attentive has come from C.D.C.
Reeve. On Reeve's account of Republic I, not only is Thrasymachus not, as I had it, 'the sort of person who is not argumentatively careful and who is capable of holding what are in fact contradictory views without realising it' (IT,p. 130), he is actually a careful and indeed powerfully insightful political thinker. 2 His account of justice, according to Reeve, 'is a coherent and resourceful blend… of ethical realism and semantic conventionalism, which identifies justice in each city with what is advantageous to its stronger ruler, and the semantic content of "justice" with what its particular laws prescribe'. 3 Far from accepting that Plato uses Thrasymachus to suggest that the various positions he conflates are each likely to prove attractive to the kind of incautious thinker who would conflate them, Reeve argues that Plato takes Thrasymachus to hold consistently to an account of 1 S. Everson, 'The Incoherence of Thrasymachus', Oxford Studies in Ancient Philosophy XVI (1998), 3 justice that makes him a 'worthy opponent' of the Socrates of the Republic, and therefore, it would seem, of Plato himself. 4 Using Reeve's riposte to IT as a spring-board, I want here to explore in greater depth Thrasymachus' arguments at 338-341 for his claim that justice is what is to the advantage of the stronger and in particular the effect of his introduction of the 'precise' notion of a ruler as one incapable of Thrasymachus offers his answer to the question 'What is justice?', to 340c, when Socrates asks him whether he actually wanted to say not that justice is what is to the advantage of the stronger but merely that it is what the stronger believes to be to their advantage. The second-[B]-runs from 340a to 341a, during which Thrasymachus rejects that proposal and elaborates a view according to which rulers are to be understood as experts incapable of making errors in ruling and ends up claiming to reaffirm the answer he gave at the start. A key question to be addressed is whether the position articulated by Thrasymachus in [A] is the same as that which he elaborates in [B].
2. 'Listen, then, I say that justice is nothing other than the advantage of the stronger,' says Thrasymachus at 338a giving his 'fine answer' to the question of what justice is. As an answer to that question, however, it is far from perspicuous and Socrates duly professes himself unable to evaluate it.
To clarify what it is supposed to mean, Thrasymachus appeals to the behaviour of the rulers of states: Don't you know that some states are ruler by a tyranny, some by a democracy, and some by an aristocracy? Of course. And in each state this element is stronger, namely, the ruler. Certainly. And each makes laws to its own advantage. Democracy makes democratic laws, tyranny makes tyrannical laws, and so on with the others. And they declare what they have made-what is to their own advantage-to be just for their subjects, and they punish anyone who goes against this as lawless and unjust. This, then, is what I say justice is, the same in all states, the advantage of the established rule. Since the established rule is surely stronger, anyone who reasons correctly will conclude that the just is the same everywhere, namely, the advantage of the stronger. 5 (338d-339a) This is sufficient for Socrates to begin to examine Thrasymachus' opening claim to see whether it is correct. His first move is to make sure that Thrasymachus is, as he takes him to be, committed to the claim that it is just to obey the rulers, which then provides the ground for Socrates' attack: 5 I have generally stuck by Reeve's own translations of the Republic from J.M. Cooper, (ed.), Plato: Complete Works (Indianapolis/Cambridge, 1997). There have been, however, a few occasions when the urge to revise has been too great to resist.
Tell me, don't you also say that it is just to obey the rulers? I do.
And are the rulers in all cities infallible, or are they liable to error? No doubt they are liable to error. When they undertake to make laws, therefore, they make some correctly, others incorrectly? I suppose so. And a law is correct if it prescribes what is to the rulers' own advantage and incorrect if it prescribes what is to their disadvantage? Is that what you mean? It is. And whatever laws they make must be obeyed by their subjects, and this is justice? Of course. Then, according to your account, it is just to do not only what is to the advantage of the stronger, but also the opposite, what is not to their advantage. (339b-d) We can discern three relevant claims in Thrasymachus' opening presentation: (1) Justice is nothing other than what is to the advantage of the stronger.
(E) The rulers of states always enact legislation that is to their advantage.
(2) Justice is the advantage of the established rule.
To these we can add the claim which Thrasymachus does not initially put forward himself, but which his argument assumes (and which Socrates makes explicit): (3) Justice is obeying whatever laws are enacted by the ruler.
Socrates' objection works by calling (E) into doubt, so that Thrasymachus' commitment to (3) will then undercut his ability to maintain (1): if it is just to obey whatever laws are enacted by the ruler of a state, and some legally required actions are to the disadvantage of the ruler, just actions are as a class no more to the advantage of the ruler than they are to their disadvantage.

3.
To get from (3) to (2), and hence to (1), Thrasymachus appeals to the legislative behaviour of rulers. The ruling element in a state enacts legislation that is to its own advantage and so if one acts in a way that is required by the law one will act to the advantage of the rulers who enacted it. Socrates' objection is straightforward. Even if rulers always intend to enact laws that are to their own advantage, it does not follow that the laws they actually enact are to their advantage, since rulers can make mistakes when legislating. Thrasymachus' argument would only go through if rulers were infallible-and this is what, of course, Thrasymachus precisely does then assert to save his argument. 'Do you think I'd call someone who is in error stronger at the moment he errs,' he demands, and proceeds to distinguish a precise sense of 'ruler', according to which rulers are legislatively infallible. To be a ruler is to exercise the art, the techne of ruling, and no incorrectly enacted law will be the result of the exercise of that art; a ruler who makes a mistake is no ruler at all.
Strictly speaking, then, a ruler can never make mistakes, since in making a mistake they would not comply with the conditions for being a ruler: But the most precise answer is this: A ruler, insofar as he is a ruler, never makes errors and unerringly decrees what is best for himself, and this his subject must do. Thus as I said from the first, it is just to do what is to the advantage of the stronger. (340e-341a) Once again, Socrates is prompted to ask for clarity as to what Thrasymachus is saying: Define (διόρισαι) clearly whether it is the ruler and stronger in the ordinary sense or in the precise sense (τὸν ὡς ἔπος εἰπεῖν ἢ τὸν ἀκριβεῖ λόγῳ) whose advantage you said it is just for the weaker to promote as the advantage of the stronger. I mean the ruler in the most precise sense (τὸν τῷ ἀκριβεστάτῳ λόγῳ).
What this alerts the reader to, of course, is the possibility of ambiguity in Thrasymachus' definition of justice. Thrasymachus concedes that when he first admitted (at 339c) that rulers are not infallible, he was using the term 'ruler' loosely but now maintains that his statement that justice is the advantage of the stronger, that is the ruler, the term is to be understood in its precise usage.
Socrates is not being pedantic in checking this last point, since the upshot of Thrasymachus' appeal to a precise use of 'ruler' to secure his claim that rulers are infallible is that the truth-value of a sentence containing the term will depend on whether the term is used loosely or precisely and this will not be apparent on its face. 6 The potential for confusion that arises in the Greek is not lost in translation and it will help to have a device to mark the difference in senses that the word has in its uses. So as to avoid confusion, I shall follow Reeve and talk of 'E-rulers' and 'T-rulers', where the first conforms to Thrasymachus' loose and the second to his precise use of the term. 7 6 One generally needs to be very wary of translations of Plato and Aristotle which invoke the notion of sense but, even duly wary, we can accept the use of 'sense' in Reeve's translation above with a fairly clear conscience. 7 In IT, I followed Reeve's earlier discussion in his Philosopher-Kings in using 'ruler' and 'Ruler' to effect this distinction (see his Philosopher-Kings (Princeton/Oxford, 1998) p. 12f). In the later paper, Reeve changes his terminology to that which I have adopted here. One advantage of the new over the earlier terminology is that the new terms do not create ambiguity when used at the beginning of sentences.
Thrasymachus is clear that the claim he is committed to from 341a is that justice is to the advantage of the T-ruler. What is less clear is whether we can take him to have made the same claim at 339a when he appeals to the behaviour of those in power in states to clarify his opening definition of justice. In IT, I argued that we cannot. The account of justice Thrasymachus offers in [A] is what has come to be called a 'conventionalist' one: 'to say someone is required by justice to perform a certain action is only to say that he is required by the law to do so ' (p. 107). Laws are what those in power in a state institute to regulate the actions of the citizens or inhabitants of that state and since what those who have power will seek to do in legislating is to regulate actions so that their interests are served, when citizens act as the laws prescribe they will act to the advantage of the rulerand that will be, in the new jargon, the E-ruler. Although Thrasymachus seeks to defend that claim against Socrates' challenge of fallibility by introducing, or at least appealing to, the notion of a T-ruler, not only does he thereby change it, so that he is not defending the claim he pretends or at least takes himself to be, but he also shifts the account of justice itself away from being the conventionalist one he first articulated. 8 In now placing what is in effect a normative condition on what it is for someone to be a ruler-that they should be exercising expertise in instituting a law-to determine whether an action is just will now itself require making a normative judgement as to whether any candidate law by reference to which it is to be evaluated actually served the interests of the supposed ruler who instituted it.

It is common ground that Thrasymachus' argumentative strategy in [A]
is to derive (2) from (3) and (E), and then (1) from (2). 13 It is also common 11 My distinguishing between 'concept' and 'conception' conforms to one standard usage-as set out, for instance, by David Wiggins in his Sameness and Substance Renewed (Cambridge, 2001), pp. 10-11. 12 It is perhaps worth my acknowledging that, for ease of discussion, I am allowing myself to talk as if Socrates and Thrasymachus were using the English words that serve here to translate the Greek words they actually do use. (One is reminded of the Oxford Greats candidate who declared with some impatience that if Aristotle had intended 'εὐδαιµονία' to mean happiness, he would have said so.) Even if this is not entirely felicitous, it does not, I think, result in any confusion. This is, of course, not the only place in Plato where there is a question as to whether the progress of the discussion clarifies an existing concept or merely introduces a new one. A prime example of that is Plato's account of justice later in the Republic itself, where he has been charged with setting out to defend one concept of justice-'vulgar-justice'-and defending something quite different-'platonic-justice'. Here the defender of Plato needs to show that the conception of vulgar-justice and that of platonic-justice are 'E-ruler' is to have the sense that is conferred on 'ruler' when people are talking loosely-that is, without the constraint of the possession of expertise. In IT, I argued that the conception of an E-ruler must indeed be what we can call a positivist one, where a ruler is merely someone who, or some group that, has the power in a state to direct the behaviour of the citizens of the state. This goes with a similarly positivist understanding of the laws of a state: a law will be a directive passed by whoever is in power in whatever is the way specified by the state's constitution (which in the case of tyranny may be simply the tyrant's issuing decrees or orders).
14 Finding both (2a) and (2b) in the text involves a degree of smoothing out. What Thrasymachus actually claims at 339a is that justice is what is in the interest of 'the established rule' (τὸ τῆς καθεστηκυίας ἀρχῆς συµφέρον), but although this might allow a distinction to be made between a ruling group (say, the aristocracy) and the members of that group (those who are the aristocrats at any time), it is clear that Thrasymachus does not intend any such distinction, so that it is harmless to take him to be talking of the ruler or rulers here. At 341a he moves immediately from his attempt to show that rulers never make mistakes, so that they only pass laws that are to their advantage, to the claim that justice is what is to the advantage of the stronger and clearly that move relies on the implicit premise that justice is what is to the advantage of the ruler.
What's key is that none of these concepts are normatively constrained. One doesn't need to evaluate a person to determine whether they are a ruler or a directive to determine whether it is a law. And from the facts about what laws are passed in a state by those with the power to pass them-the Erulers-one can, given (3), determine whether an action is just or unjust and, again, without needing to appeal to any evaluation of the action. 15 Given (E)-the rulers of states always enact legislation that is to their advantage-one thing all just actions at least have in common is that they are to the advantage of the E-rulers who enacted the laws to which they Socrates, whose challenge, he thinks, actually requires that Thrasymachus' underlying account of what it is to be an E-ruler is not conventionalist in the way I've just characterised.
Once Thrasymachus has sufficiently clarified what he means by (1) for Socrates to be able to examine whether it is true and also confirmed that he accepts (3), Socrates poses the question as to whether (E)-rulers can make mistakes in enacting legislation: Tell me, don't you also say that it is just to obey the rulers? I do.
And are the rulers in all cities infallible, or are they liable to error? No doubt they are liable to error. When they undertake to make laws, therefore, they make some correctly, others incorrectly? 15 In the light of this, when Thrasymachus demands that Socrates should say what justice is at 336c it is not unreasonable to take his prohibition of appealing to various evaluative concepts as an indication of methodological principle rather than a denial in advance that justice is something valuable in whatever way: 'Give an answer yourself, and tell us what you say the just is. And don't tell me that it's the right, the beneficial, the profitable, the gainful, or the advantageous, but tell me clearly and exactly what you mean' (336c-d). So, a ruler will make a law incorrectly (μὴ ὀρθῶς) if the law is not in fact to their own advantage-and given (3), when a citizen or subject obeys that law, their action will be just but not to the ruler's advantage.
Socrates' argument here relies on being able to make sense of a ruler's making a mistake and it is this which Reeve takes to show that at least Socrates thinks that Thrasymachus is no conventionalist: In constructing this trap… Socrates recognizes from the beginning that Thrasymachus has a standard of correctness for E-laws in mind, namely, that they 'prescribe what is advantageous for the rulers themselves' (339 c7-8). This is an embarrassment for Everson's view that E is conventionalist, since if it were, there could be no such standard and we would have to wonder why Plato sends Socrates off on so wrong a foot. 16 Reeve takes it as being obvious here that a conventionalist or positivist conception of rulers and laws will not tolerate a standard of correctness for laws. And, taking that to be obvious, he doesn't provide any argument to confirm that appearance. One possible motivation, however, might be this: properly to secure whatever was the objective for its enactment.
Reeve might retort to this that whilst a standard of precise formulation could indeed by affirmed as applying to all laws of any state without going against a positivist account of law, this is because it applies a standard of a relevantly different kind to that which is applied by Thrasymachus at 339c.
Any candidate for being a law will need to be formulated and so it is unproblematic to describe ways in which its formulation should be better or worse. Consider, in contrast, Austin's statement of the central claim of his positivist account: The existence of law is one thing; its merit and demerit another.
Whether it be or be not is one enquiry; whether it be or be not conformable to an assumed standard, is a different enquiry. 18 What Austin has in mind here is a moral or evaluative standard. He is taking a stand against, for instance, natural-law conceptions according to which an unjust law is no law at all. If Thrasymachus' standard for legal correctness is that the law should be to the advantage of the ruler that is to impose an evaluative standard for legal correctness.
What is at least moot, however, is whether at 339c this is accepted as a of precise formulation acceptable to a positivist was not merely that it regulates a feature of any candidate law but that it works as a criterion for judging how good any law is as a law. What it does not provide is a standard for judging whether any command or instruction is a law at all. One can still make perfect sense of the possibility that a law is a bad law because it is badly drafted. So what Reeve needs to establish is not merely that Thrasymachus accepts that being to the advantage of the E-ruler is a standard for the correctness of a law, but that he accepts it as a standard for being a law at all. In Austin's terms, it needs to pertain to a law's existence and not merely to its merit or demerit. Now, although Thrasymachus does not re-state the standard once he has introduced the notion of the T-ruler, it is certainly an implication of what he says at 341a that he would accept this as an existential standard. The decrees that a subject must obey are those issued by a T-ruler who is exercising their expertise in issuing them and these will necessarily be ones that are to the advantage of that T-ruler. But obviously Reeve cannot appeal to this as support for taking the standard in the same way when it is stated at 399c without begging the question. So one needs to consider the dialectical context of that statement if one is to determine whether it is there given as an existential or a critical standard for laws.
Seen in its context, however, it is clear that Thrasymachus does not think that an incorrect law is no law at all. Until he understands that his easy acceptance of the claim that rulers-E-rulers-can make mistakes in legislating will subvert his claim about what justice is, he is entirely untroubled by the idea that sometimes a ruler may mistakenly enact a law that is to their disadvantage and assents to the claim that 'whatever laws the rulers make'-i.e. whether they are laws that are correct or incorrect-'they must be acted on (ποιητέον) by their subjects' (339b). If Thrasymachus were expressing loosely the precise notions of the T-ruler and of the T-law that he will go on to articulate in order to escape Socrates' trap, he would surely have been prompted by Socrates' questions themselves to have given different answers to them and not waited to have Socrates explain-twicewhy the answers he does give cause the problems for him that they do.
Up until [B], then, Thrasymachus sees no tension in the possibility of a law that is nevertheless incorrectly enacted in that its regulation of behaviour does not actually work to the E-ruler's benefit. One might still wonder, however, whether even if the standard for the correctness of laws is critical rather than existential this might be in tension with a positivist conception of laws and legislation. If there is a single standard for any law according to which one can determine whether it is defective, does that not itself require one to conceive of laws in terms of some necessary evaluative goal that is internal to the practice of law-giving? Should we not see Socrates' assumption that Thrasymachus posits as a standard for the correctness of law that it should benefit the E-ruler a recognition that Thrasymachus is not, even in [A], operating straightforwardly with a positivist conception of laws and then Thrasymachus' acceptance of that standard as confirmation that he indeed is not?
Socrates' challenge, of course, picks up on (E), Thrasymachus' claim at 338e that E-rulers always enact laws that are to their advantage. Whilst this might work as a statement of E-rulers' purpose, their fallibility means that they will not always succeed in securing their advantage and when they don't the law they enact will not achieve the purpose they had for it. Making law is, on any account, a purposeful activity and whenever a ruler enacts a law, they will have a goal in doing so. On a positivist reading of [A], according to which someone will be a ruler if they have the power to enact and to enforce legislation and a law will have force just if it has been enacted according to the correct constitutional procedures, Thrasymachus believes that those who have the power to enact laws always want to regulate the behaviour of their subjects to further the interests of the ruler; whenever that is their purpose (which happens to be always), any law which fails to do this will be defective. The standard for success is set just by the goal the legislator happens to have: of course, it is not accidental that legislators have, when legislating, the goal of furthering their interests, but that is not because they are legislators or because they are legislating but simply because they are human and humans always seek, if unconstrained, to further their own interests. There is nothing unstable, then, in thinking that every law which fails to further the interests of the ruler signals a failure of skill on the part of the legislator whilst denying that this stops them from enacting a law or from being a legislator. Even if one thinks that legislating is an essential activity of being a ruler, one can allow that legislating is an activity that requires skill without making possession of that skill essential to being a ruler.
5. So, that Thrasymachus thinks that a law will be badly made if it doesn't further the interests of the E-ruler is consistent with his holding a positivist conception of laws, but it does not in itself, of course, require that. It is not an embarrassment for a conventionalist interpretation of [A], but neither does it in itself secure it. The 'in itself 'is important there, since the considerations which restrict the standard so that it is critical rather than existential also present a difficulty for a non-conventionalist reading of Thrasymachus' argument. The possibility that an E-ruler may legislate badly and enact a law that is in fact not to their advantage, but which, being a law, must be obeyed by their subjects, entails that a defective law is nevertheless a law-and also, it would seem, entails that that defective Eruler is nevertheless an E-ruler. A similarly defective T-ruler, in contrast, is no T-ruler at all. This then allows, indeed requires, one to find that (E) itself is a claim about T-rulers. Reeve acknowledges that this might seem to be in tension with Thrasymachus' grounding it in empirical claims about the behaviour of those who actually have power in Greek states, but thinks that this too will be eased by understanding how Thrasymachus conceives of the relation

between E-rulers and T-rulers:
If E is about T-rulers, however, in what sense, if any, can it be an empirical claim about the behaviour of rulers in actual Greek cities? While E-and T-rulers, laws, and the rest are excellent expository devices, it is a mistake to treat E-rulers as the sort we find in actual cities and make empirical claims about, and T-rulers as creatures of Thrasymachean ideal theory or fantasy, about whom we make a priori or conceptual claims. For Thrasymachus makes it quite clear that Trulers are E-rulers at those times when they are actually practising the craft of ruling, and so not making errors. When E-rulers are ruling correctly, they are T-rulers, therefore, and so-tautologically-never make errors. 20 We can compare this with Thrasymachus' own explanation of his initial response to Socrates' objection and his contrast between speaking of rulers loosely and precisely: I think we do speak with the relevant word in this way (λέγομεν τῷ ῥήματι οὓτως), that a doctor makes a mistake, or an accountant, or a grammarian. But each of these, insofar as he is what we call him, never makes a mistake, so that according to the precise account (ὥστε κατὰ τὸν ἀκριβῆ λόγον), and you go in for speaking precisely (ἀκριβολογῇ), no craftsman ever makes a mistake. It's when his knowledge fails him that he makes an error, and in regard to that error he is no craftsman. No craftsman, wise man, or ruler makes a mistake at the moment when he is ruling, even though everyone will say that a doctor or a ruler makes mistakes. It's in this way that you must also take the answer I gave earlier.
But the most precise answer (τὸ ἀκριβέστατον) is this. A ruler insofar as he is a ruler, never makes mistakes and unerringly decrees what is best for himself, and this his subject must do. Thus, as I said from the first, it is just to do what is to the advantage of the stronger. (340d-341a).
Thrasymachus thus contrasts the precise account of craftsmen, according to which it will not be true that a craftsman ever makes a mistake, with a looser way of using the terms for craftsmen where people will indeed say that a doctor or ruler or whoever does make mistakes. He acknowledges that his claims in [A] involved this looser usage, but that, now using the term 'ruler 'precisely, what he says will not allow the possibilities that Socrates has tried to deploy to block his argument. So, Reeve and Thrasymachus are certainly at one in their trying to explain away Thrasymachus' initial acceptance of rulers' fallibility as the result mere loose speaking and in denying that in moving to talk more precisely he doesn't thereby change what he is talking about. What is at issue is whether their defence can be made good-and indeed whether in its details it is quite the same defence.
What, though, does it mean to say that Thrasymachus is talking loosely about T-rulers at 339c? One might, it seems, effect a distinction between talking loosely and talking precisely about something in various ways. One way might be where one imprecisely applies what is a precise criterion for applying the relevant concept. So, one might describe a quickly drawn shape as a square, even though the angles in contains approximate to being right angles when in fact they aren't. Another might be where the criteria for applying a term loosely are sufficient for making a probable judgement that something falls under the concept but do not determine that it doesdoing that will require some other or further criterion to be employed. So, a doctor might diagnose influenza on the basis of a patient's symptoms, but a certain diagnosis would require PCR testing. Doctors will commonly use 'flu' in that loose way, but still acknowledge that those they diagnose with flu may not actually have it. It is not that they've improperly or carelessly applied the criteria for precise diagnosis of a patient's infection, but that they've applied other criteria which are adequate for most practical purposes but are accepted to be insufficient to determine the truth of the diagnosis. Someone may have flu and yet be asymptomatic, so will not be found to have flu on the basis of the loose criteria, or may easily satisfy the loose criteria but not have flu.
It is clear, I think, that it is not the first kind of looseness of talk that is in question for Thrasymachus. He is not supposing that ordinarily, say, we will judge someone to be a doctor unless they are making a mistake but don't worry too much about whether they have actually made an error of diagnosis or treatment. His point is precisely that when we speak loosely we will say, and intelligibly so, that a doctor does make mistakes. The looseness, then, would need to be of the second kind. The idea, then, would be something like this. When Thrasymachus responds to Socrates' by tightening up the conditions for being a ruler, he does not determine a new concept but in effect argues that the existing concept itself, when properly understood, makes demands that are often neglected in our ordinary talk.
Although ordinarily we are ready to talk as if rulers can make mistakes in legislating, more careful thought about what it is to be a ruler will show that this is in fact incoherent: when one grasps the conditions for being a ruler, one will see that they in fact preclude anyone who is making a legislative error from complying with it.
Now, on Reeve's construal of the argument, someone will satisfy the predicate 'is a T-ruler' when they are actually exercising the craft of ruling. 21 Anyone who is a T-ruler must also be an E-ruler: only those with the power for their laws to be enacted will be able to exercise that craft. So the notion of a T-ruler is certainly more restrictive than that of an E-ruler, but doesn't follow from this that it is a more precise one. It is not that the conditions for determining whether someone is an E-ruler are somehow indeterminate and in order to resolve that indeterminacy, one needs to appeal to the extra condition introduced with the concept of a T-ruler.
Working out who are and who are not T-rulers will not revise one's judgements as to who are and who aren't E-rulers-indeed they presuppose such judgements have already successfully been made. There is nothing wrong with introducing the concept of a T-ruler and it may well be that there is some utility in being able to talk just of those rulers who are 21 In fact, Reeve's claim, cited above, that Thrasymachus makes it clear that someone is only a T-ruler when they're exercising the craft of ruling is at best optimistic. What Thrasymachus says is that one wouldn't call someone a ruler when they're making a mistake. This would allow them to be a ruler even when they're not exercising the craft as long as they possess the knowledge that would be exercised in legislating. Nawar points to Thrasymachus' claim at 340e that the craftsman makes an error when their knowledge fails them to argue that for Thrasymachus error is incompatible not just with the exercise of the craft but its possession. If one accepts that then the concept of a T-ruler will be that of an Eruler who possesses the craft of ruling and not that of someone who is simply exercising the craft they possess. This merits a separate a more detailed discussion than I can offer here, but does not, I think, affect the structure of my discussion. Neither concept of T-ruler is one that is required to make something more precise of that of the E-ruler. move. This is the second, and more significant, correction that needs to be made to the argument of IT. I effectively assumed there that, given (3) what secures Thrasymachus' conventionalism up to his advancing the claim that justice is the good of another is his continued adherence to (3)that whatever laws a ruler enacts, justice is obeying these. Indeed, his appeal to the notions of T-rulers and T-laws to answer Socrates' objection to the argument of [A] is motivated by his desire not only to protect (1) but the conjunction of (1) with a version of (3). It would have been quite possible for Thrasymachus to have given up (3) as a ground for (1) and to have argued, for instance, that strength merits obedience, so that it is just to act in the interests of the stronger. Obeying laws that have been enacted by Trulers would then be a particular application of that more general principle of justice. 24 This, however, would allow justice to stand as a prior and independent source of norms for deciding whether to obey any candidate law. Thrasymachus does not take that line: certainly, he places normative constraints on being a ruler and being a law, but although these then determine which actions justice requires-or perhaps rather which actions rulers would have us believe justice requires-they are not themselves requirements of justice.
Even in [B], that is, although whenever someone acts justly by acting as the laws require them to they will thereby act in the interests of the ruler, they can act in the interests of the ruler without acting justly. Because it is only necessary and not sufficient that an action should be in the interest of the ruler that it should be required by law, which kinds of action will be legally prescribed is still a matter of what the ruler happens to decide. If the ruler has not bothered to legislate to further their interest in some way, then, there will be no requirement either of law or of justice for the subject to act as if they had. Which of the actions, or kinds of action, that will be to the ruler's interest are required by law is up to the decision, indeed even the whim, of the legislator. The central conventionalist claim-that whether an action is just or not depends on which laws happen to be enacted-is untouched. Thrasymachus may introduce technical conceptions of rulers and laws in order to save his commitment to (1), but even this is not enough to bring his legalistic conception of justice into line with what follows from 343c, when he quietly abandons (3) and allows that the just person's 24 So, this would seem to be the underlying position of those cited in the Laws who also adopt the slogan that 'justice is whatever is to the advantage of the stronger'. When the Athenian refers to them in Laws IV 714bff, he refers back to the discussion in Book III of what should ground authority in the state, and associates them with the view that 'the stronger should rule and the weaker should obey'-'a decree of nature, as Pindar of Thebes once remarked' (690b).
willingness to obey the law is but one way in which they subordinate their interests to those of other people.