2. CICERO ON PLATO ON GYGES

Cicero narrates the story of Gyges' ring as part of his project in Book 3 of De officiis to show that what is honourable and what is beneficial go hand in hand and cannot come apart. The immediate occasion for Cicero's presentation of the tale is as an illustration of his claim that one should not do wrong even if one could be completely successful in concealing one's wrongdoing. ‘That is why’, says Cicero, ‘that fellow Gyges is introduced by Plato’ (3.38).Footnote ¹ Cicero then tells the story in a way that sticks fairly closely to the Platonic original (some differences will be considered shortly), and concludes, ‘Therefore if a wise man were to possess this very ring, he would think it no more permitted for him to do wrong than if he did not possess it; for good men seek things that are honourable, not concealed’ (3.38).Footnote ²

At first glance Cicero has offered a decidedly odd account of why Plato uncorks the tale of the lowly Lydian shepherd who, having discovered a magic ring that can make him invisible, uses its power to seduce the queen of Lydia, murder the king and usurp the throne. Could Cicero be such a poor, or disingenuous, reader of the Republic that he would state with apparent earnestness that the story of Gyges is intended by Plato to show that one shouldn't do what Gyges did even if one can get away with it? Surely its purpose is, on the contrary, to make it seem how implausible that thesis is, thereby illustrating the scale of the challenge faced by Socrates in defending it. After all, Glaucon himself, who narrates the tale in the Republic, states that consideration of it will incline us to the view that ‘those who practice justice do so unwillingly, through lacking the power to do injustice’ (359b7–9). Give someone the power to do injustice – that is, to do it with impunity, as in Gyges' case – and as the story shows, they will do it.Footnote ³ Admittedly, that is compatible with the notion that one should in fact not do injustice, and that a genuinely wise person would not do it. But when Cicero tells us that Plato presents the Gyges story to show that ‘nothing should be done greedily, unjustly, licentiously or unrestrainedly’ (3.37),Footnote ⁴ it is hard to resist the conclusion that the immediate take-home message of the story in Plato – that no one in their right mind would be expected to act other than Gyges in a similar position (360d) – has been turned on its head.

Let us, despite appearances, hold off from the conclusion that, in Cicero's eyes, Plato allots Gyges the role of illustrating why injustice should not be pursued; and let us note in the meantime that there is no reason why Cicero should not himself use the story to make that point. Indeed the most significant overall difference between the two accounts is that Cicero stacks the deck against Gyges rather more than Plato does, and this is something that may easily add to the impression that he takes even Plato's narration of the tale to be a self-contained parable against injustice. Thus in 3.38 Cicero calls the Lydian monarch not simply Gyges' king but his ‘master’ (dominum), thus making it particularly heinous that Gyges should have murdered him. He talks, as Plato does not, of Gyges removing from the scene not just the king but all ‘those whom he thought stood in his way’ (quos obstare arbitrabatur). And he speaks of Gyges being able, thanks to the ring, to conceal his ‘crimes’ (facinoribus), and to inflict ‘disgrace’ (stuprum) on the queen, there being no equivalently severe descriptions in Plato's narration. In these ways Cicero's Gyges is presented more overtly than Plato's as a ruthless evildoer whom we should not want to be like.Footnote ⁵ But one should none the less still be reluctant to read him as claiming that Plato saw the tale's purpose as being to advocate abstinence from an unjust use of the ring, when this seems so at variance with the way the tale is presented in the Republic.

Plato's Gyges, in turn, is more recognizably an ordinary human deliberative agent than is Cicero's – and a more troubling figure in that regard for being decidedly one of us. Note in particular the wonderfully acute description of Gyges ‘experimenting’ (cf. ἀποπειρᾶσθαι, 360a5) with the ring in order to confirm its powers; and the way that, in spite of its powers, he takes the trouble to get himself appointed a king's messenger as a reasonable pretext to be in his presence (a8–b1). With these touches Plato shows himself more ready than Cicero to lull us into a sense of psychological realism about the scenario, the effect if anything enhanced by Glaucon's opting to make his tale fable rather than history when calling its subject an ‘ancestor’ (πρόγονος) of Gyges of Lydia (359d2),Footnote ⁶ though Socrates is happy enough to refer back to ‘Gyges’ ring' in Book 10 at a safe distance from the original narration.Footnote ⁷

Plato's Gyges, then, if not exactly a sympathetic figure, is not the stage villain of Cicero's telling. That Cicero manipulates the Gyges story fairly consistently in one direction suggests that if he is in error about Plato's strategy, he is calculatedly so.Footnote ⁸ But if that explains it hardly justifies a declaration that Plato tells the story to warn us off immorality rather than make it seem an appealing option in the right circumstances. A more charitable reading therefore suggests itself: Cicero means that Plato introduces the Gyges story not to show by its mere narration that justice is always to be preferred to injustice, but as a first, motivating step towards the construction of an account that will vindicate that thesis. At this point we can dispense with charity and call upon some concrete evidence in favour of the claim that Cicero has a perfectly good understanding of the way the Gyges tale functions in the Republic. For Cicero's own deployment of the story provides a broad but unmistakable structural match with its appearance in Plato's work.

In the Republic, after the Gyges story is told in Book 2, and enhanced by some further dialectic from Glaucon and Adeimantus in favour of injustice, there famously follows a long defence of the preferability of justice, spanning (albeit with many a twist and turn) the bulk of the work, after which, in Book 10, the Gyges story is finally recalled, in conjunction with another magical device, and put to rest:

Compare now the De officiis: the Gyges story is narrated (3.38), and after a discussion by Cicero of its force (3.39; I shall return to this below) a long defence is mounted of the thesis that the honourable and the beneficial go together (3.40–77), marked out as separate from the Gyges material by Cicero's noting at the end of 3.39 that we are ‘now resuming’ (iam revertamur) the theme of how the honourable and the beneficial relate.Footnote ⁹ The Gyges story is then eventually recalled, in conjunction with another example of wickedness, to reassure us that this is not a type to be emulated:

This patterning by Cicero tracks, surely self-consciously, the footprints of the Gyges story in the Republic. The tale is narrated, prompting a defence of what the tale on the face of it denies (that justice is always better than injustice), thence allowing the speaker a trenchant recollection of the tale once its dangers have been thwarted. Even Cicero's flourish of coupling Gyges with another evildoer (the exploiter of inheritances) offers a very Roman parallel to Plato's pairing of the ring with a second invisibility device (the cap of Hades). All of this suggests that Cicero himself regards Gyges as threatening enough to require his further re-engagement, since that is not a feature one would expect if he thought his initial narration of the tale had already done the work of dismissing him. The re-engagement, moreover, is a clear sign of Cicero acknowledging that in Plato the initial narration was indeed just a starting point. Despite his own more loaded treatment, then, Cicero, far from blundering (or being disingenuous) in his assessment of Gyges' role for Plato, shows a sophisticated awareness of how the story plays out in the pages of the Republic.

3. GYGES IN CONTEXT

The interim conclusion that, with regard to the Gyges story, Cicero is a conscientious reader of Plato can be strengthened. Cicero is attentive not just to the structural role but to the context of the Gyges story as it appears in the Republic. Immediately after narrating the story, Cicero considers, at 3.39, the objection from ‘certain philosophers’ (philosophi quidam) that the tale is pure fiction and could never be a reality. Cicero's response is that this is beside the point: the narration of the Gyges tale asks the reader not to accept that the story could come true but to state, if there were such a ring and you had possession of it, whether you would use it wrongfully. He claims his opponents face a dilemma: if they say they would use it, they admit they are vicious; if they say that they would not use it, they admit that wrongdoing is to be avoided in and of itself.Footnote ¹⁰

Who are Cicero's opponents here? Plausibly they are (or include) the Epicureans,Footnote ¹¹ since the latter notoriously assert that injustice is indeed not a bad thing for the agent in its own right (καθ᾽ ἑαυτήν); its undesirability lies in the wrongdoer's fear of being found out (KD 34). Cicero presses a proponent of such a view to say how they would respond if there were no danger of being found out.

Should not Epicurean advocates simply reply that in the Gyges scenario they would do the wrong but expedient thing, given that injustice is not bad in itself? The content of KD 34 seems to affirm such a view, but why would they have a problem with that? Things are not quite so simple, however. Firstly, there is a quietist strand in Epicureanism that recommends not breaching social convention: SV 51, for example, has Epicurus advising his interlocutor to avoid indulging his sexual appetite if that means breaching established law or custom, a maxim that one assumes Gyges has overlooked in committing adultery with the queen. More centrally Epicurus also holds that the just life and the pleasant life go hand in hand (Ep. Men. 132, KD 5).Footnote ¹² So he would appear especially vulnerable to a case that seems to divorce pleasure from justice, and to that extent Cicero's attack is well targeted.

It is also far from gratuitous in relation to Cicero's reading of Plato. That is because the Epicurean theory of justice that Cicero appears to be reacting against has notable affinities with the theory of justice that Glaucon constructs in the Republic to motivate the Gyges story. For Glaucon, the background of the Gyges tale is the notion that people ‘decide it is profitable to make an agreement with one another neither to do nor to suffer injustice’ (δοκεῖ λυσιτελεῖν συνθέσθαι ἀλλήλοις μήτ᾿ ἀδικεῖν μήτ᾿ ἀδικεῖσθαι, 359a1–2). And Epicurus is of the view that justice is a matter of expediency (τὸ συμϕέρον, KD 31), whereby people make ‘an agreement for the sake of not harming or being harmed’ (συνθήκη τις ὑπὲρ τοῦ μὴ βλάπτειν μηδὲ βλάπτεσθαι, KD 33).

Glaucon thus states for dialectical purposes, as Epicurus will state in his own name, and perhaps, given the linguistic resonance, in conscious reference to the Republic,Footnote ¹³ that justice boils down to a contract or agreement enacted for the prevention of mutual harm. Glaucon reasons, and Cicero inherits this viewpoint, that if that is what justice is about then there could be no reason for the possessor of Gyges' ring to do justice.

Cicero therefore again proves himself an adept reader of Plato. The attacks on Epicureans in his works are numerous and wide-ranging, and the De officiis is certainly not the only place where Cicero puts pressure on Epicurus' claim about the inseparability of morality and pleasure.Footnote ¹⁴ But in renewing the attack with the telling of the Gyges tale he has chosen an especially suitable context, given the kind of view about justice that Glaucon presents as lying behind the story in the Republic.

4. JUSTICE AND MODALITY

Cicero has, I hope to have shown, a sound understanding of his archetype. So let us now turn to the question (applicable to both Plato and Cicero): how much difficulty, if any, does the Gyges story actually create for the sort of contractarian theory of justice that each of our authors associates with it?Footnote ¹⁵ For Plato, the Gyges story needs to serve as a kind of reductio ad absurdum of a contractarian theory, in order for it to motivate the very different approach that Socrates will take. If contractarianism is the only game in town, then Glaucon's challenge threatens to show that there is no reason for any agent with the power of a Gyges to sign up to the contract – with the implications for justice that Glaucon goes on to spell out with more than a hint of relish. Socrates must then attempt, in order to dissuade a potential Gyges, to give a very different rationale for justice than that offered by Glaucon's sketch. But if one can simply retort that the Gyges scenario is unrealizable, then it remains unclear why Glaucon's contractarian notion should actually provoke an alternate theory in its stead (as it is undoubtedly designed to do).

Plato, then, ought to be concerned with the question of the tale's realizability – with the modal issue, as I shall call it. But it is Cicero who engages with it directly. He has his opponents claim, as a way of protecting their theory, that in fact the ring story cannot begin to damage them because the Gyges scenario is impossible.Footnote ¹⁶ Cicero thus goes one step further than Plato in explicitly raising and discussing this response. Although he remarks that the Gyges scenario is possible (3.39),Footnote ¹⁷ he is willing for dialectical purposes to accept his opponents' denial of this and says that the denial misses the point: ‘but I am asking what they would do, if what they deny is possible were possible’.Footnote ¹⁸ What the Epicureans need to say is whether they would follow Gyges if it so happened that the scenario could be realized.

As I indicated above, Plato is less forthcoming than Cicero on the modal question. But as we saw, he does have reason to be concerned with it, since denial of the possibility of Gyges' ring might appear to undercut Socrates' motivation for constructing his own preferred theory of justice. Interestingly, Glaucon does anticipate the fictional character of the Gyges story when he previews it as a ‘thought experiment’ (εἰ τοιόνδε ποιήσαιμεν τῇ διανοίᾳ, 359b9–c1), but he does not indicate whether its remoteness from reality might affect any morals (in both senses of the term) that one may wish to draw from it.

Now the question of whether a narration speaks of a possible world or not is raised much later in the dialogue, with regard to the central story it has told, that of the construction of Kallipolis, with Socrates insisting that its creation is ‘difficult, but not impossible’ (χαλεπὰ μέν, δυνατὰ δέ πῃ, 540d2–3; cf. 499d4–6, 502c6–7). Without this assurance, the tale of Kallipolis will have dubious force as a response to the tale of Gyges, at least if its construction is thought to be a necessary condition for the formation of people who embody Socrates' idea of justice, agents with no motivation to do injustice whatever the circumstances. But how much force does the tale of Gyges itself have as motivation for Socrates' preferred theory if its magical ring places the circumstances beyond the realm of the possible?Footnote ¹⁹ By the time we have reached the end of Book 7 we are perhaps supposed to realize that Plato considers this a proper question to ask. Although Glaucon does not ask it, we might be struck by a contrast between the forward-looking possibility of Kallipolis and the tale of a ring whose location in the remote past seems to emphasize its purely mythic status.

In this regard the helpful discussion by M.F. Burnyeat of Plato's use of fantasy (as it might have seemed to his readers) in the construction of Kallipolis,Footnote ²⁰ for example regarding the role of women, in order to abstract away from the concrete present and allow his readers to contemplate unrealized possibilities, would not straightforwardly apply to the Gyges case (which Burnyeat does not discuss). And it is instructive that Bernard Williams fails to get to grips with the issue in his complaint that, by purporting to draw conclusions about justice applicable to the actual world, the Gyges tale ‘compare[s] fantasy with reality’ which ‘hardly seems a helpful way to address the question of whether and why one wants the life of justice’.Footnote ²¹ Is that the problem with the efficacy of the Gyges story in the Republic? A fantasy might, after all, be realized in a hopeful (or alarming) future. Fantasies are not as such unrealizable; their use may be a perfectly proper way of thinking about how one wants life to be. The creative and strenuous exercise of human imagination can (and on innumerable occasions does) lead to the forging of new actualities. As Burnyeat argues, this idea is central to Plato's approach in the Republic. The problem of fitting Gyges' tale into the picture is that it projects itself not on to a new and as yet unrealized future, but an irrecoverable past.Footnote ²²

Let us, then, look more closely at Cicero's discussion of the modality of the Gyges story. He begins by having his opponents represent the Platonic tale as a ‘fictitious and imaginary fable’ (fictam et commenticiam fabulam, 3.39). On the assumption that Epicureans are among those whom Cicero has in mind, dialectical battle may already have been joined, since we are told, intriguingly, in Proclus' commentary on the Republic, that Epicurus' pupil Colotes attacks Plato for neglecting truth and knowledge and ‘spending time with falsehood, mythologizing like a poet’ (In R. 2.105.23–5 Kroll).Footnote ²³ Cicero's riposte is to fire off a Platonic myth with the purpose of demonstrating its efficacy as a tool of enquiry against those very opponents who would deny that its mythical character permitted it to function that way.Footnote ²⁴

But being fictional does not itself imply being unrealizable; and Cicero quickly moves on to the claim that neither Plato (quasi vero ille, 3.39) nor he himself place any weight on such a ring's being able to exist: the force of the parable is independent of its realizability. The contractarian theory suggests that, ultimately, one is just because one is better off being just than unjust, so long as there is reason for your neighbours to agree that they are better off being just too. The idea is that injustice is a zero sum game. Everyone loses out if it is committed; conversely, everyone wins when justice is practised. The Gyges tale presents us with a scenario in which one can commit injustice seemingly without losing out. The demand this places on Cicero's opponents is, as far as he is concerned, to choose: justice or expediency. And it is apparently of no avail for them to say that because the scenario has strayed beyond the possible the choice is not one they need (or intelligibly could) be responsive to.

We have already wondered why the opposition should not simply say ‘we choose expediency’. In the Republic, after all, the point that Glaucon takes the story to be making is precisely that not simply the unjust but also the just would exploit Gyges' ring (359c, 360b–c); as indeed, given the theory of justice adopted, whereby expediency is the ultimate guide to action, they presumably would and should. The potential embarrassment of the Epicureans' claim that justice and pleasure go hand in hand can then be read as a claim about how things are in the actual and any feasible alternative world. Not an impossible world in which magical rings of invisibility exist.Footnote ²⁵ There is no such thing as justice ‘in itself’ (καθ᾽ ἑαυτό), Epicurus tells us in KD 33, which, together with KD 36, stresses that justice is a matter of agreement and depends on the particular circumstances of time and place. The localism underlying this view is entitled to see itself as bolstered, not refuted, by the positing of a set of circumstances that could not obtain anywhere. If there is no such thing as justice in itself, these are circumstances in which the theory may, by its own lights, reply that there is no account of justice to be offered.

Since Cicero is willing to concede, for the purposes of his argument, that the world of Gyges may not be a possible world, responses that point out that there are also real-world scenarios of untrammelled power themselves miss the point. One of the most important features of the Gyges case is that it presents a picture in which there is no question of one's impunity being at risk,Footnote ²⁶ unlike any possible human scenarios, however extreme, including those of would-be untrammelled power. Epicurus avers in KD 35 that no one who does wrong can ever be sure that they will not be found out, and that moral can be applied to those powerful enough to do wrong without necessarily hiding it, but who cannot, in any imaginably human world, regard their hold on power as secure beyond threat.Footnote ²⁷ It is only because the world of Gyges is a magical world that the invulnerability of the agent that makes it distinctive can be assumed. Indeed it is its disconnection from the realm of what could actually be, not its affinity to that realm, that surely gives the ring its special power as a parable about justice; and this is what Cicero recognizes in his claim that the possibility or otherwise of the story is an irrelevance.

But Cicero's basic point cannot be exactly that theories which look to expediency as the basis for justice get something fundamentally wrong about the concept of justice, although it is certainly true that one of his principal dialectical aims in De officiis 3 as a whole is to show that forms of reasoning that consider expediency in isolation from what is honourable are misguided – a point that underlies the dilemma he creates for his Epicurean opponents. But then the upshot for Cicero is not of course that expediency goes by the board but that, when properly assessed, justice is seen to be not contrary to the agent's benefit. So his engagement with the Gyges tale in its context indicates that for him justice has been misconceived unless one acknowledges that it is the expedient course even in counterfactual, hypothetical, maybe unrealizable circumstances, such as those in which one possessed the ring. Justice must, as it were, turn out to be expedient ‘in and of itself’.

As with Plato, then, the task for Cicero is to show that an agent is better off being just than unjust in the peculiarly strong sense that would have even Gyges, if sufficiently wise, reject injustice,Footnote ²⁸ aware, as Cicero puts it shortly before narrating the tale, that ‘the penalty of vice itself is most bitter’ (poenam … ipsius turpitudinis, quae acerbissima est, 3.36). But then we have to ask both authors: why should we think that this extremely strong condition is a requirement on a proper account of justice? It is tempting to reply that we have a firm intuition that justice must prevail come what may. Then in both Platonic and Ciceronian versions this will be exemplified in the thesis that justice must benefit the agent come what may. But why insist that it must do that in (what is conceded by Cicero might be) an impossible scenario, beyond the boundaries of whatever may come? Why should we need to extend the reach of justice that far?

One might approach the question by registering dissatisfaction with alternatives. We don't want debates about justice to be nothing more than engagements with Realpolitik – whether, say, in reasonably likely scenarios, perhaps historically based, an actual or would-be tyrant needs more deterrence from injustice than the thought that there is always a danger of being exposed or deposed. Despite the obvious importance of questions about what the best practical measures are to prevent the unjust exercise of power and (more generally) the mutual harm of unenlightened human interaction, to narrow down one's focus to that sort of issue is already to concede too much to the opponent's way of looking at things. Justice is about more than prevention and control. And whatever an Epicurean may say, no one should think that the concept of justice is exhausted by thoughts, for example, about the strategies of cooperation that various local or contingent social arrangements may permit or encourage – if for no other reason than that, even with the best of intentions, specific social structures are likely at times to produce forms of agreement whose consequences are wildly at odds with justice.Footnote ²⁹

Justice should triumph come what may, in any and every scenario. But if this is to include ‘even in impossible scenarios’ then we must ask again, why should that be? Perhaps the most basic question that Gyges' ring poses, and one that Cicero's discussion of the modal issue invites directly is: what is the principle underlying this requirement, from which it would follow that a defeat of justice even in a purely hypothetical situation must be ruled out? Anyone who senses that something is wrong with an account of justice that makes injustice the right choice for Gyges is faced with this question. The challenge is to spell out the underlying principle in a way that explains, without begging the question, why the strong requirement should be in place.

5. GYGES, JUSTICE AND NATURE

I shall not attempt to meet this challenge here, but in order to make it more concrete I want to offer some brief reflection on a particular aspect of both Plato's and Cicero's responses to Gyges. Plato's, as spelled out in the Republic, is that the initial debate about the behaviour of Gyges ignores the crucial dimension of the moral psychology of the agent – broadly put, and in a normative sense, it ignores the question of human nature. So Socrates famously argues that the constituents of the psyche of the unjust agent are in a state of disharmony, whereas the just agent's psyche is harmonious; and that the relation of these constituents in the just agent is established ‘in accordance with nature’ (κατὰ ϕύσιν, 444d8), whereas in the unjust agent it is set up ‘contrary to nature’ (παρὰ ϕύσιν, d9).

Cicero, although less interested, in this context, in questions of moral psychology, none the less indicates in De officiis 3 an allegiance to a naturalistic defence of justice (and of virtue in general). He states, for example, that ‘nothing is so opposed to nature as vice, nature desiring what is right … and rejecting the contrary’ (3.35)Footnote ³⁰ and that cruelty cannot be beneficial ‘because it is utterly inimical to human nature, which we ought to follow’ (3.46).Footnote ³¹ In short, one fulfils one's nature in being virtuous and is thereby in the best condition one can be. Justice is defended on the grounds that it is a central element of this best condition.

But this response remains problematic. That is because magical scenarios seem to muddy the sense in which human nature is still at stake. At one level, perhaps, Gyges is human, all too human. And there is no denying that Plato's presentation of him as such is seductively artful. But can the inhabitants of Gyges' world still be regarded as human?Footnote ³² Can we conjure forth a magical place in which human nature is none the less allowed by stipulation to stand fixed?Footnote ³³ I think the question of whether creatures who seem to resemble us, but who live in a world in which some of the fundamental rules are different, should be regarded as human, is unlikely to have a straightforward answer.Footnote ³⁴ For this reason the naturalistic route cannot promise to provide a satisfying response to the challenge of explaining why justice should prevail in Gyges' world. To show (or claim) that justice rather than injustice accords with human nature is not even half the battle, because it leaves obscure whether Gyges falls into the category ‘possessor of human nature’ that would bring him under the theory's wing. To emphasize: my claim is not that Gyges is a creature who possesses a nature other than human, but that, given the world he inhabits, we have no principled way of determining whether he does or not. For this sort of case naturalistic accounts are in danger of losing their explanatory power just at the point where it is needed.Footnote ³⁵

If justice should prevail to the extent that it expresses human nature, it is moot whether the latter can be said to be operative in Gyges' world; and it is of no help that the point cuts both ways. Glaucon as agent provocateur says his tale will confirm ‘the nature of justice’ (ἡ … ϕύσις δικαιοσύνης, 359b6) and reveal that ‘everyone's nature’ (πᾶσα ϕύσις, 359c5) seeks more than its share once we subtract human convention (νόμος) from the picture. A world with some very basic and prominent unnatural features seems as oddly designed to shore up that conclusion as it does the thesis that nature shows we should all, even Gyges, be just. (It also, of course, illustrates the difficulty of separating cleanly nature from convention.) When it comes to Cicero's naturalism, the larger and more direct influence on De officiis is that of Stoicism rather than Plato. But that only intensifies the question of whether a tale premised on a magical ring can be answerable to a framework that is based on an account of what nature is; and this is especially so if Stoic naturalism, more than Plato's, sees human nature's fulfilment in terms of its conformity to nature at large.Footnote ³⁶

It is, then, worth stressing that, in spite of what Epicureans may say, one simple solution to the problem of Gyges is unavailable: simply leaving the whole scenario to one side. Its extravagant unreality can in an important respect mislead. Care is needed in contrasting, as Julia Annas does,Footnote ³⁷ Plato's approach with ‘common-sense intuitions about justice’. One may certainly doubt that Plato would have seen his solution in terms of the moral psychology of the agent as chiming in with his readers' intuitions. Still, what Annas calls the ‘realistic response’ – that the Gyges case is too far-fetched to have any claim on us, what Cicero disputes in his attack on the Epicureans – does not, it seems to me, line up neatly with our intuitions about justice. The tale of Gyges has force precisely because many who encounter it do see it as a challenge; do, that is, have a sense that something crucial has been left out of an account that is unable to show why Gyges himself should be just.

In this regard the dialectic of Glaucon's narration of the story is complex. It is triggered by the contractarian view of justice. But given that Plato presumably did not think his readers would regard Socrates' taking up of the cudgels against Gyges as merely whimsical, he must have assumed that the response of a sufficiently large proportion of his readership would be that a theory of justice that cannot tell Gyges why he should desist must be flawed. Cicero's more moralizing version of the story, his early insistence on its lesson being that one should not do wrong even with utter impunity – hidden from ‘all gods and men’ (omnes deos hominesque, 3.37) – need not, I have argued, be taken to imply that he thinks the story itself shows this to be true. But it is, at least, an indication that Cicero, in tune with but more explicitly than the Platonic text he draws upon, well understands the intuitive desire to refute Gyges rather than let him be. Cicero does not, I think, succeed in fully explaining this intuition; but his presentation of the debate reveals with new clarity what needs to be explained.