BMCR 1996.04.08

1996.4.8, Irwin, Plato’s Ethics

, Plato's ethics. Oxford: Oxford University Press, 1995. xx, 436 pages ; 25 cm. ISBN 9780195086447. $22.00.

As Irwin recounts in his preface, Plato’s Ethics began as a revised edition of the author’s earlier Plato’s Moral Theory and grew into a separate work unto itself, somewhat fuller and intended to be more accessible than its predecessor. While Irwin acknowledges the influence from critical discussion of the previous book, he does not give an account of it here.

The book is organized into 20 chapters with a total of 234 subsections, moving through dialogues in an order that appears to be taken chronologically (for the most part), although there is little discussion of issues of dating. After an introductory chapter detailing method and approach, Irwin moves into a detailed interpretation of Socratic ethics (based primarily on the Laches, Charmides and Euthydemus), then to Plato, first as a defender of Socrates attempting to fill gaps in Socratic moral theory ( Protagoras, Gorgias, Meno), before coming into his own, departing at points from his mentor. A chapter on the forms, followed by seven devoted to the Republic form this central part, before Irwin concludes by turning to Symposium, Phaedrus, Philebus, and Laws. The later sections tend to include more exposition of the primary texts, with less space devoted to the development of interpretation and conclusions. Indeed the book ends rather abruptly at the last topic treated, no overall conclusions summarized. This is perhaps indicative that in some sense Irwin’s Plato appears as a transitional figure, bridging the span from Socrates to Aristotle.

Aristotle stands not only on the horizon, but also at the outset. Irwin takes Aristotle’s comments on Socrates and Plato as standards—if not definitive, at least decisive in questions deemed unclear from the Platonic corpus. Irwin defends the use of Aristotle at length in the first chapter, putting on notice the reader—should there be any—who suspects Aristotelian ethics make significant concessions to the standpoint of the sophists that Plato so systematically rejects. His evaluation of the texts leads him to assert as a generalization that Aristotle sides with Socrates on metaphysical issues, and with Plato on ethics. Socrates, the herald of an ontology of substance? Surely an odd thought to some, but this is not Irwin’s topic. One also wonders how Platonic metaphysics is to be plausibly detached from Platonic ethics, and Irwin himself does not attempt to do this, incorporating a discussion of the forms into the account of the mature Platonic ethical theory. However, the generalization does perhaps foreshadow Irwin’s view of Platonic development: the late Plato of the Laws is supposed to stand apart from Socrates and from the Republic in thinking non-moral goods relevant to happiness, and so virtue alone insufficient (p. 346), and closer to Aristotle also in a broad conception of wisdom, which is “less restrictive” than the Republic and “truer … to the spirit of Socratic moral inquiry.” (p. 353)

Beyond drawing on Aristotle, Irwin’s approach to Plato is informed by modern ethical theory, not so much in discussion of issues, which is focused on the dialogues mentioned, but in the questions that structure the interpretation. Over the Platonic formulation of the one guiding moral question, “What is the good life?” (as at Crito, 48b) Irwin favors a division into two questions, one normative, “How ought we to live?” and one epistemological, “How can we know how we ought to live?” (p.3) This silently projects the post-Kantian is/ought distinction onto a theory where it is not made. Later Irwin accepts the claim that Plato thinks virtue identical to knowledge, which would make this division of the question of morality suspect; yet the distinction of questions sets up the book’s analysis of Socrates, who, the reader learns, can answer the first question, but not the second, thus leaving a gap for Plato to fill. Socrates, in Irwin’s view, knows that virtue is “fine and good,” knows that the good is happiness, but fails to give an account of happiness.

Irwin sees Socratic and Platonic thought emerging in conflict with common beliefs about virtue, which tend to question the Homeric ideal, “self-assertive virtue,” in favor of a new paradigm, “other-regarding virtue.” If these particular terms almost suggest the development of an idea of altruism which seems out of place in the Classical age, they do point to a tension in the literature of the period, one about which Irwin expresses surprise to find Socrates silent. (p.35) Irwin characterizes the Socratic challenge to contemporary attitudes by stating that Socrates’ contemporaries had “doubts about the connexion between virtue and happiness,” (p.51) that is, there is general suspicion that acknowledged virtues may not always lead to the agent’s own benefit. This would not appear to be the Socratic silence. In the Apology and Crito, for example, never is it questioned that one should seek one’s own benefit; rather, at issue is just what does lie in one’s own genuine benefit. Allowing this contention, the distinction of “self-assertive” from “other-regarding” is at best a false dichotomy resting on a confusion that Plato’s Socrates points to often.

Not by accident I’ve advanced two considerations with reference to the Apology and Crito. Perhaps this may help clarify Irwin’s depiction of Socrates. For some would regard the Apology and Crito as essential to any understanding of Socratic ethics: Socrates’ defense speech when on trial for his life, and an argument justifying his willingness to die with equanimity for reasons of justice. Irwin, however, makes little use of these two pieces, building instead his interpretation of Socratic ethics primarily on the Laches, Charmides and Euthydemus, and Lysis.

Irwin’s examination of Socratic method asserts Socrates pursues real, not nominal definitions, and does so by means of the Socratic elenchos. Since Socrates does not attain the knowledge sought, Irwin concludes the method is deficient. The interpreter’s job, then, is to pinpoint the source of Socrates’ failure. Lest the reader be too astonished at this scholarly assignment, Irwin does qualify it. He observes both that Socratic dialectic leads to caring about virtue, and that this concern for virtue is taken to be of highest importance. Given the weakness of human wisdom, the failure to obtain knowledge, hardly a shortcoming, may even be a boon. Nevertheless, the failure to acquire knowledge is made the starting point for the account of Socratic ethics as a whole. Throughout the volume, this kind of intellectual honesty is Irwin’s virtue: a conclusion is drawn, then qualified as inconclusive, only to become the basis for the subsequent argument.

The details in Irwin’s construction of Socratic ethics include, of course, a number which are subject to debate. Even more serious than the question whether some of those details actually fit the text, they are as a whole difficult to fit together coherently. Socrates is called a “eudaimonist,” meaning that, at least rationally, happiness is the ultimate aim of action. (Thus Irwin’s emphasis on the Euthydemus, for it raises the topic of happiness in general, not as a particular virtue.) Furthermore, Irwin refers to the Euthydemus and the Lysis to argue that Socrates believes all virtues to be purely instrumental to the end of happiness. He cites Lysis 220a, saying that Socrates “implies that if we choose x for the sake of y, we cannot also choose x for itself.” (p.66) Yet the passage cited implies nothing of the sort. It reads: “Whatever is dear only for the sake of something else that is dear we do not properly call dear.” The example used in the immediate passage is money. The cited text asserts that whatever has purely instrumental value is not valued for itself, and so does not imply either that only an ultimate end can be valued in itself, or that nothing valued as a means can also be valued as an end.

This reading of the Lysis—asserting a single ultimate end and nothing good as both means and end—is foreseen and supported by Irwin’s use of a passage from the Euthydemus. First, a summary of the passage central to Irwin’s interpretation. At 279a ff., Cleinias agrees that all of us wish to do well. Furthermore, one does well if one possesses many good things, including the states of being wealthy, healthy, handsome, well-born, honored, temperate, just, brave, and wise, and finally fortunate. But goods are good only if used, and wealth, health, beauty and position will be used to good results only if guided by wisdom, and are in fact evils if guided by ignorance. Among the virtues listed, the only one mentioned in the consideration that certain goods may in fact be evils if not used wisely is bravery, which has not been defined here, and which in the Laches, another of Irwin’s favored “Socratic” dialogues, is not bravery unless accompanied by knowledge. (Surely it would be absurd to say wisdom is an evil unless used wisely.) The conclusion at 281e states that among these other things none is good or bad in itself, but, of wisdom and ignorance, wisdom is good and ignorance bad.

Irwin takes this passage to reveal a Socratic doctrine that wisdom alone is a good, and nothing else at all is a good. Yet this interpretation is beset by multiple problems beyond simply not matching the text, which says not that wisdom is the only good at all, but that between wisdom and ignorance, only wisdom is good, and, of the things mentioned, material goods and certain skills and abilities, none is simply good in itself. Irwin’s interpretation, at this point, suggests that 1) only what is good in itself and not as a means is good, and that this is a set with one member, namely wisdom, 2) nothing is good both in itself and as a means, and 3) anything apart from wisdom that has any relevance to the good is neither good nor bad, but neutral, and its relevance to the good is purely instrumental. This means, further, that, strictly speaking, nothing is better or worse: the ” … comparative claim (‘greater good’) need not imply that health is a good; if two things fail to be large, one of them may nonetheless be larger than the other … In this case, then, Socrates may mean that neither health nor sickness is good, but health is a greater good because it is more of a good, closer to being a good.” (p.58) One might pause to wonder whether this isn’t equivocation followed by a weak analogy, if, that is, one isn’t already caught up wondering whether it isn’t a flat out contradiction. Even if we grant all these points and set aside concerns over agreement with text, it follows either that all virtues are identical to wisdom (a possibility Irwin is not averse to), but that virtue is then, contrary to Irwin’s other claim, not instrumental, since it is a good, or it follows that wisdom is not a virtue, and the other virtues, be they one or many, are, as instruments, not good but neutral and so on par with health, sickness, and vices. Irwin does not draw any further conclusions, but merely takes the interpretation to provide further evidence of Socrates’ “unreserved commitment to virtue.” (p.60)

Irwin does not assess the possibility that, for Socrates, virtue is constitutive of happiness, that while virtue may well be good as a means, virtue is also an end in itself, explaining why it is possible for the just person to be happy though facing execution, explaining the Apology and the Crito, which Irwin can explain, if at all, only by the improbable assertion that physical well being is not a good. At most perhaps we can say that if Irwin’s depiction of Socrates does not seem entirely consistent, it is because he never expects it to be consistent. Indeed, a lack of consistency provides the final peculiar argument for claiming that Socrates thinks virtues to be purely instrumental: ” … we ought to be persuaded that he accepts an instrumental view. Even if it does not offer a very plausible defense of some of his convictions, that is not by itself a good enough reason for supposing that he does not accept the view; many philosophers have accepted theories that fail to justify some of their convictions.” (p. 77)

Irwin’s Socratic ethics leaves us with a set of virtues, the status of their unity uncertain, all means to an end called happiness, which Socrates is unable to define. Irwin suggested early that the Socratic method sought real, not nominal definitions, borrowing a distinction from Aristotle. The search for real definitions certainly fits the tenor of Socratic dialogues, but seems incongruous with the conclusion that virtues are purely instrumental, and, if it is in fact the case that the inquiry should have pursued a definition of happiness apart from virtue, Irwin’s Socrates does not seem to have coherently formulated the question. This Socrates appears to generate more heat than light, and one wonders what he has to teach us today, or that he had anything to teach Plato in his day. Nevertheless, Irwin sees Plato writing in the Protagoras, Gorgias and Meno to bolster Socratic ethics by defining the vague and filling the gaps. At this point, therefore, the reader gets a more complete picture of what Socratic ethics might be. The most prominent gap emerges in connection with Irwin’s view that Socrates thinks all virtues are instrumental. If all virtues are but devices to something, what is the product of their implementation? For the answer to this, Irwin turned to the Euthydemus, where there is an object for the instrumental virtues: happiness. This still leaves happiness undefined. Closure is found in the Protagoras : happiness might be pleasure.

Irwin neither assents to nor dissents from the view that Socrates is a hedonist—and so remains noncommittal in this debate that enjoys current scholarly fancy. He does think Socrates’ views compatible with hedonism and therefore hedonism, at least at one time, the obvious choice for the disciple who wants to rescue the theory that virtue is instrumental: “It would be wrong to claim that in the shorter dialogues Socrates implicitly accepts hedonism or that he is committed to it by his other views; each of these claims would go unjustifiably far beyond our evidence. On the other hand, the shorter dialogues make it easy to see why Socrates might intelligibly accept hedonism…a defender of Socrates might naturally and reasonably turn to hedonism.” (p. 91) Thus, while it is unclear whether Socrates accepts hedonism, Plato’s Socrates of the Protagoras does. Irwin first develops the claim that the Protagoras advocates hedonism, before considering three possible objections to it.

Irwin’s interpretation of the Protagoras begins with a curious observation that proves to be a significant working assumption: “The dialogue ends with a puzzle rather than a definite solution…It does not follow, however, that Plato believes neither speaker has defended the true answer … ” In effect, Irwin tags the Protagoras as an aporetic dialogue, but is nevertheless disinclined to regard the dialogue, especially the discussion of hedonism, as dialectical.

In Irwin’s presentation, the topic of hedonism emerges against the background of the initial arguments concerning the unity of the virtues. Socrates had proposed to Protagoras an exclusive alternative in the form of analogy. If the virtues are parts of a whole, as Protagoras asserts, are they like the parts of a face—eyes, ears and nose—each different, or are they like parts of a bar of gold, all the same? Neither Socrates nor Protagoras explicitly raises other possibilities. Irwin refers to the passage without explicitly discussing the particular analogies Plato uses. The latter alternative Irwin calls the “Unity Thesis.” This title is perhaps somewhat misleading, since it is not questioned that the virtues form a unity; the issue is what kind of unity. In any case, Irwin makes clear that ‘Unity Thesis’ means ‘identity thesis’: ” … the claim that all the names of the virtues are really names of one and the same thing, so that the virtues are identical … ” (p.79) This is said to be distinguished from the claim that the virtues are similar and from the Reciprocity Thesis.” (p.79 f.) Thus, it appears that Irwin spies a threefold distinction, although only the capitalized two seem to be of import. We might conclude, then, that the first alternative of the text is dubbed the “Reciprocity Thesis (the claim that they imply each other and are therefore inseparable) (p.80) Nevertheless, it doesn’t immediately jump to mind to call the eyebrows and the nose “reciprocal” or to say they imply one another, and certainly Plato’s Protagoras chooses the first alternative precisely to attempt to maintain the virtues are separable.

Irwin finds evidence for the Unity-Reciprocity distinction in shorter dialogues: “In order to show that every virtue is fine and beneficial, Socrates argues that no virtue can prescribe actions that are rejected by any of the other virtues.” (p.80) This is well stated, and certainly does suggest mutual implication of the virtues. On other evidence for “unity,” i.e., identity, Irwin adds Socrates ” … also argues that knowledge is sufficient for virtue and that therefore all the virtues can be identified with knowledge of the good.” (p.80) This I find much less convincing. Far from arguing that knowledge is sufficient, typically the dialogues—for example, the discussion of courage here in the Protagoras, where Socrates brings Protagoras to admit that bold action done without knowledge cannot be called courageous, and the discussion of the Laches that foreshadows it—argue that knowledge is a necessary, not a sufficient condition for virtue. No person has virtue without also having knowledge. The identity of knowledge with virtue does not follow. The passage in the Protagoras also argues that knowledge is a sufficient condition for courage. Even then, identity does not follow from knowledge being a sufficient or even both a sufficient and necessary condition for virtue. Remember, mutual implication is the kind of unity here called ‘reciprocity,’ not the kind of unity called ‘unity.’

In part, Irwin’s interpretation depicts the Protagoras as Plato’s clear argument for the identity of virtues and their identification with knowledge, thus refining vaguer Socratic conceptions. Consequently it seems that Socrates earlier and in the introductory passages of the Protagoras“distinguishes the Reciprocity Thesis from the Unity Thesis and believes both, but has not sorted out the arguments to support each thesis.” It is most curious that he believes both, given that the former implies that the virtues are many and the latter that virtue is one. Indeed, it seems not to have been sorted out.

Unity, i.e., identity, looms larger as the topic turns to hedonism. It seems to me that one reason for Irwin’s inclination to attribute to Socrates here a strong commitment to hedonism lies in his reading of the context in which the doctrine is raised. Socrates urges the hedonic thesis, which Protagoras himself rejects and says the many reject. If no one else wants it, it would appear to be Socrates’ own. Also once fit neatly with the doctrine of the denial of incontinence, we have a tight package: the good is pleasure, no one knowingly pursues pain, and virtue is the hedonic calculus, making virtue one and instrumental. The instrumentality of virtue is retained, and now clarified, since we know more definitely what it is an instrument to, and finally, the sense in which virtue is unified is resolved.

Unlike the account that was given of Socratic ethics, this set of doctrines is internally coherent. However, I have doubts that it fits the text any better. As noted, Irwin points out, but does not deem significant, the dialectical structure of the dialogue. Socrates had wanted to deny, Protagoras to affirm that virtue can be taught. In the examination of Protagoras’ views on courage, he seeks against Socrates’ argument to detach courage from other virtues and knowledge. This is inconsistent with his claims about teachability, and the reader is reminded of the hypothesis of the Meno, that, if virtue is knowledge, then it is teachable. In the Meno, the discussion ends in a muddle, and we are explicitly told the reason for the muddle: we have failed to define virtue. The Protagoras, too, ends in a muddle, giving good reason to suspect that it does not straightforwardly present any view of virtue or the good that can be taken at face value as Platonic ethical theory. (Irwin does not refer to the Meno within his discussion of the Protagoras. In the separate and later discussion of the Meno, he does not draw any major conclusions; rather, its significance appears to be transitional, as Irwin claims that the distinction between knowledge and belief is meant to endorse Socratic method, but that it nevertheless goes against Socrates’ account of the character of virtue.)

Hedonism is raised in the context of this confusion. Protagoras acknowledges that some live well, others live badly. (351b) Inquiry into living well is the fundamental ethical question from the early dialogues. Protagoras also acknowledges that living pleasantly to the end of one’s life is living well. But he then wants to assert that some pleasures are good, some evil. This is implicitly inconsistent with the previous agreement, but, it seems, Protagoras has nothing else in answer to the question about the good life. As one would expect, if that is not adequately answered, then virtue is not defined, and the further questions about it—are virtues unified and in what sense, is virtue knowledge, is virtue teachable—also do not get beyond the confusion. That Protagoras has no answer to the root question is underscored by his response to the initial questions about pleasure. Socrates asserts (and the assertion follows from the previous agreement) that pleasure itself is good. (351b) Protagoras, with the many, not only wants to deny this, although it is implied by their other beliefs, he also himself takes this to mean that pleasure is identical to the good. Socrates’ only assertion is that pleasure, insofar as it is pleasure and abstracted from context and consequences, is a good. Whether this claim is true or not, it is consistent with a denial of the hedonic equation of pleasure with the good, the very step that the Protagoras of the dialogue makes himself.

Hedonism is opposed by the many and Protagoras, only to be shown in the dialogue to be at the core of their beliefs. There is no need to suppose Plato is advancing hedonism as a doctrine against common belief, when the argument indicates that he is showing the incoherence of common belief. This is a frequent theme, executed masterfully in the Protagoras, as in the Crito, Republic and elsewhere.

The “denial of incontinence,” involving the identification of virtue with knowledge and the claim that no person knowingly pursues evil, which Irwin assumes to be an unwavering Socratic doctrine, coheres with hedonism. Irwin presents the hedonism in the Protagoras as a doctrine that will support the presupposed denial of incontinence. Again, I believe this bends the text. In fact, it reverses the text by taking a consequence as a presupposition. In the discussion of the separability of courage from other virtues, Protagoras remained unwilling to identify wisdom with courage and reluctant to draw the conclusion that knowledge implies courage or courage knowledge. (350c ff.) Now, once it has been maintained that the good is identical to pleasure, it follows that no one knowingly pursues the bad, that is, no one seeks pain as an end, and that virtue is knowledge, namely the science of pleasurable pursuits, the hedonic calculus. (358c ff.) While the dialogues often argue that virtue implies knowledge, this is the only passage that would identify virtue with knowledge, and that only in tight connection with the acceptance of hedonism. The closest any other passage comes is the failed attempt in the Charmides to define temperance as self knowledge. (To claim Socrates identified virtue with knowledge seems, incidentally, to be an excellent example of an institutionally generated myth: several generations of students have heard the claim repeated from their first undergraduate philosophy course, and have gone on to repeat it to their own students without having looked at the texts freshly and carefully enough to notice it isn’t really there.)

Having raised some doubts about the accuracy of Irwin’s interpretation, I should mention three objections to Socratic/Platonic hedonism that Irwin discusses. The three are 1) that Socrates attempts to show how the many are committed to hedonism, 2) that hedonism is unnecessary for the rejection of incontinence and the unity of virtues, and 3) that no other Socratic dialogue affirms hedonism.

To the first Irwin responds: “This is a poor reason for supposing that he does not accept it himself.” (p. 86) One might add it is a poorer reason for supposing that he does accept it himself. Irwin notes that “the second-person character of the elenchos and its appeal to the belief of the interlocutor are pervasive … ” (p. 86) This, I think, is obvious and uncontroversial; however, a further claim about Socratic elenchos is open to question: “Socrates assumes that the conclusions reached by the interlocutor through the elenchos are true.” (p. 86) The elenchos is pervasive, as Irwin observes, but this assertion overlooks two equally pervasive features of Socratic dialogues: Socrates’ profession of ignorance, and the aporetic outcomes. The conclusions reached by the interlocutor in fact are often contradictory, implying that the initial assumptions must be revised, not that they are true. These features are common as well to the Protagoras, which ends, far from definite conclusions, with Socrates’ assertion that the argument has mocked them. (361a)

The second objection states that a rejection of “incontinence” and the “unity of virtues” is independent of hedonism. Irwin answers that the rejection of incontinence goes against common sense, thus demanding support, and that the argument for hedonism in the Protagoras is the only argument in the dialogues supporting the rejection of incontinence. This contention is dubious. First, it fails to take into account the dialectic of the Protagoras that issues in the rejection of “incontinence.” It assumes on the contrary that the rejection is an unshakable given, and hedonism is built around it as a support. However, the text argues from hedonism, as an assumption implicitly accepted—rightly or wrongly—by the many, to the rejection of “incontinence” as a consequence. If the issue amounts to asserting that, given that the better is more pleasant and this is what does and should drive people, then no one will choose the less pleasant as better knowing in fact that it is less pleasant, then it is trivial that no other dialogue offers an argument to support this, since no other Socratic dialogue undertakes to explicitly examine hedonism. On the other hand, if we take the rejection of incontinence to mean the claim that people do wrong out of ignorance, thinking mistakenly that the wrongful action is in their own best interest, then it is simply false that no relevant arguments are made elsewhere. Again, the Crito furnishes an excellent example: many people, thinking mistakenly the good of the body superior to the good of the soul for the purpose of living well will act unjustly and do so out of ignorance.

Irwin answers the third objection, that no other Socratic dialogues affirms hedonism, by suggesting hedonism is compatible with and completes the notion Irwin has called Socratic eudaimonism. Again, I beg to differ. Not only does no other dialogue affirm hedonism, hedonism is incompatible with what Socratic dialogues do affirm. In the Crito, Socrates is willing to face death on the grounds of an argument. That argument states that, as health is the good of the body, so justice is the good of the soul, and, if living well is impossible with a corrupt body, it is even less possible with a corrupt soul. If, then, it is unjust to escape, then Socrates should not escape. Socrates does not propose to die in order to make himself or anyone else feel pleasure. This bit of Socratic ethics is blatantly inconsistent with hedonism and with an instrumental conception of virtue. The point need not be belabored.

Irwin turns to the Gorgias, which would appear to pose a problem for Irwin’s overall interpretation, since it contains a spirited refutation of hedonism. Irwin outlines the main points of the argument, including how Socrates faults rhetoric for attention to pleasure, not to the good, how Socrates argues Callicles’ assertion that the good maximizes pleasure is an incoherent assertion, since the attempt to fulfill all desires is unstable and inevitably bound with pain, and how Socrates claims that pleasure and pain can exist simultaneously, markedly unlike living well and living badly. The reader learns, nevertheless, this is no argument against hedonism. Irwin even acknowledges, “The thesis accepted in the Protagoras and the thesis rejected in the Gorgias are expressed in the very same words” (p.111) Nevertheless, Irwin thinks they are not the same, because the hedonic calculus, the “measuring science” of the Protagoras, provides an answer to the “unrestrained” hedonism advocated by Callicles. A reader might think on the contrary that the argument of the Gorgias is explicitly generalized beyond Callicles’ position, given the claim that pleasure and the fulfillment of desires, untempered by other considerations apart from pleasure and pain, is by its very nature unrestrained. However it may be, it seems fair to say Irwin’s account of the Gorgias is tied to his account of the Protagoras, so no further comment is needed.

Hedonism, then, in Irwin’s interpretation, fills the lacunae of early Socratic ethics. Still, he does not think the ethical theory of the Protagoras (and Gorgias, if one can accept with Irwin that it is compatible with the acceptance of hedonism) is the mature Platonic theory. As already mentioned, the Meno is construed as transitional. Written to bolster Socratic method, its distinction between knowledge and belief clashes with Socratic characterization of virtue: “The Socrates of the early dialogues turns out to have correct belief, but not knowledge, about the character of virtue.” (p. 147) Irwin asserts now that the importance Socrates places on virtue in the early dialogues does not cohere with the view that virtues are only instrumental; a pragmatic view of virtue fails to explain Socrates’ concern for virtue. To do this, it seems, requires some epistemology, which appears in the distinction between knowledge and belief, and is developed in the theory of forms. If this is so, then it would follow Socrates also does not have correct belief; rather, the account implies Socrates holds true belief about the importance of virtue, but no knowledge why it is important, as well as the false belief that the importance lies entirely in the utility of virtue. A reader should wonder, even on Socratic terms apart from the explicit distinctions of the Meno, that Socrates could have correct belief about something when he does not know what it is. It emerges in the book’s subsequent discussion of the theory of forms that the question of explaining Socratic concern for virtue is indeed related to the question of definition. The problem that Irwin identifies at the outset resurfaces: Socrates asks for definitions, but is unable to find them. (As before, Socratic professions of ignorance do not figure in the interpretation.) The theory of forms will now explain why.

With the Republic, the transition from Socratic to Platonic ethics is completed. The several chapters devoted to the Republic are introduced by a chapter on the forms. Irwin devotes most of this chapter to summarizing passages that bear on the theory of forms. He gives prominence to the argument about equals in the Phaedo, the discussion of relative sizes in Republic V, and the flux of the visible world along with Aristotle’s mention of Heracleitean influence on Plato. The salient point for ethics that emerges is “Plato’s claim that sensibles do not yield knowledge of the forms….” (p. 163) As formulated, this is ambiguous. Certainly in the passages cited, awareness of sensibles is a step leading toward knowledge of the forms, although the sensible world is not itself an object of knowledge. This, I take it, is Irwin’s meaning. In any case, the point is a salient one, because “Plato’s claim that sensibles do not yield knowledge of the forms helps explain why Socrates did not find definitions.” (p. 163) Socratic inquiry requires definitions that can formulate one property common to all instances. Plato argues that only a form can answer to this requirement. Irwin speculates that Socrates fails in seeking definitions because he is expecting a definition to specify an “observable property.” (p.163) I, at any rate, find this one of the most astonishing assertions of the book. It is astonishing for its sudden appearance at this point, no mention having been made of such a significant claim in the lengthy interpretation of Socratic method and Socratic ethics. (Irwin does refer to an earlier section on the Euthyphro (pp. 23-25) where he notes that Socrates asks for a definition that gives an account of one form, identifying “a property by reference to which we can say whether any action is pious.” But he does not remark that this is to be a sensible property.) The assertion is also astonishing for its content. Nothing is wrong with astonishing assertions, if they are supported by argument. Irwin admits there is none: “The Socratic dialogues and the Meno do not actually say that Socrates assumes that observable properties are needed for a definition.” (p. 163) He continues: “at least Socrates does not discourage the interlocutors from looking in this direction … ” with reference to the Euthyphro. This comes close to arguing that if Socrates does not deny the claim, he must accept it. Yet, if we look at the Euthyphro, the contention appears even weaker still. The dialogue asks for a definition of the holy. Euthyphro’s first response is by example: holy is what Euthyphro is doing. This is observable, if not a single property, and is rejected as not even addressing the question. The next response is a single property, namely the property of being dear to the gods. Is this property observable? By what organ of sense? If it were knowable, the property of being dear to the gods (or to all the gods) might be sufficient to distinguish holy from unholy acts. Even so, it would not be a definition, for it doesn’t state that in virtue of which the holy is dear to the gods. Socrates (as Irwin also says) is asking for an essence. If that does not dissuade the inquirer from looking to sensible properties, what does?

Defensible or not, the discrepancy between virtue’s importance and its instrumentality is remedied as the new epistemology makes way for revisions in the account of virtue. Republic II explicitly describes the virtues as being both means to good ends and ends in themselves. This, as has been seen, is taken to be a break with Socratic ethics.

If the instrumental account of virtue no longer holds, then we would not expect virtue to be characterized as identical with knowledge in the way Irwin attributed to Socrates. Irwin sees evidence for a revision of this in the moral psychology of the Republic. Indeed the view that virtue is knowledge of the means to pleasure would be difficult to reconcile to a tripartite division of the soul. Irwin asserts that, in denying all desires respond to beliefs about the good, Plato associates them with different parts of the soul. Irwin describes the soul as “an asymmetrical relation that includes a symmetrical relation,” (p. 217) making an analogy to governing and opposing parties. He also suggests that each part of the soul has its own unity within itself, and therefore can act to some extent on its own. One would wonder whether this depiction is compatible with the claim that one function of the rational part is to unify the soul. For, although the different parts provide different incentives, there is no suggestion in the text that the soul ever functions with only one part active, only that the governing part may differ. The text also does not appear to support the separate unity of the appetitive part unto itself. As illustrated by the tyrannical soul, the appetites are expansive, changeable and contradictory, rendering the soul of the tyrant, which is not without reason, but is dominated by desire, not unified, not even a friend with itself.

On either Irwin’s or another interpretation of the tripartite soul, Irwin is correct that division would not agree with the view that all desires are a desire for the good, which is identical to the desire for pleasure. I already voiced doubts about this version of Socratic ethics. If the so-called Socratic denial of “incontinence” is closer to the less extravagant claim that knowledge of what is better is necessary for the choice of what is better, then the tripartite division of the soul is not necessarily at odds with the Socratic dialogues. A frequent theme of Socratic dialogues is the incoherence of popular morality. We are driven by desires, the desires that Plato in the Republic classifies with the appetitive part—for wealth, honor, office—and these are considered the criteria of success. At the same time, most people (save, perhaps, a Thrasymachus) stop short of pursuing those under any circumstances, as scruple prevents one from trying to acquire such worthy things if it is at a cost of doing so unjustly. For this is generally acknowledged to be shameful, and the society should deal angrily with those who do not abide by such norms. Such reactions Plato in the Republic ascribes to the thumos. The early dialogues testify to the tension and conflict of these criteria, and the Socrates of the Crito subordinates both to the single criterion of determining what is true and just by rational argument.

Nevertheless, Irwin affirms a clear break with Socratic ethics, fully realized in Republic IV. He acknowledges that other interpretations of the relation between the ethics of the Republic and the ethics of the shorter Socratic dialogues are defensible, and he notes that a key in this interpretive decision (as we have already seen) is the reading of the Protagoras. To lend support to this, as before Irwin looks outside of the Platonic corpus to Aristotle’s remarks on Socrates and Plato. (pp. 242 ff..)

While the Republic smoothes some bumps in Socrates’ ethics, it exhibits tensions in its own account of justice. Irwin distinguishes between “p-justice” and “c-justice.” (p. 256) “P” stands for “psychic,” “c” for “common.” Psychic justice is “defined by the reference to the relation of the parts of the soul to each other,” while common justice is “another’s good,” defined “by reference to the agent’s relation to other people.” (p. 256) This is the reprise of a theme that Irwin signaled in the introductory sections. “Psychic justice” is a “self-regarding” virtue, concerned with one individual’s own good, while “common justice” is an “other-regarding” virtue, concerned with the good of other people apart from oneself. The challenge in the Republic, then, which Irwin thinks Plato struggles to answer, is to show that a person can never be “p-just,” that is, living with a well-ordered and unified soul, and commit an act that is “c-unjust,” that is, do an injustice to another person. Irwin thinks Plato fails to answer the problem in Republic IV. The later books of the Republic, then, are designed to show that deviant souls, ones obviously “c-unjust,” such as the tyrannical and democratic souls, are also “p-unjust.” (p. 283) On this, Irwin observes that they cannot be shown “p-unjust in the most obvious way, by acting incontinently or without any rational order in their choices.” (p. 284) He finally concludes that the just person is better in that the rational part of the soul alone is “impartial between the aims of the non-rational parts and comprehensive in its concern for the whole soul.” (p. 291) This coheres with Irwin’s initial view of the divisions of the soul: each part is a unity unto itself, reason is superior only because it is more balanced. If the virtues are no longer to be understood purely as means to an external end, this account does not go far beyond an instrumental account of reason. So Plato does, after a fashion, provide an answer, if not the full one that would require the absent account of the good. It is, nevertheless, very sketchy. To say that virtues are good as ends in themselves, not only as means, implies the existence of a rational end. Pleasure it cannot be, and, when Irwin considers whether it is demonstrated that the tyrant is “p-unjust” (see citations above), it appears to be suggested that the only kind of rationality conceivable is instrumental reason. Beyond this, Irwin only suggests that reason, as opposed to other parts of the soul, is more comprehensive. This leaves a gap that Irwin will fill with eros, which will give reason something in common with the irrational—perhaps pleasure is the key after all.

Above I suggested that the distinction between self-assertive and other-regarding virtues is an anachronism. This question, can one be “p-just” but “c-unjust,” or perhaps better, can one be happy but unjust, in fact grows out of Glaucon’s challenge. Glaucon supposes that justice is a purely instrumental good, good only because the person who appears just profits from the honor and acknowledgment bestowed by the society. He demands that Socrates argue against this that it is good both instrumentally and in itself. Whether or not the Republic accomplishes this task successfully, these are the proper terms of the argument: whether justice is good for the just person because of rewards it brings, or whether it is also good independently. For, to the extent that it brings good only to the others, not to the just person, neither Socrates nor Plato defend it. Holding to the anachronism, Irwin fails to pose seriously the question whether what is called “c-just” is in fact not the same as what is called “p-just.”

The interpretation of the Republic forms the centerpiece in Irwin’s account of mature Platonic ethics after the break with Socrates. Beyond the Republic, Irwin adds a discussion of Platonic love, resting on the Symposium, Phaedrus and Philebus, and finishes with very brief remarks on late dialogues including the Laws. The connection with eros is foreshadowed in the Republic, which shows that the proper object of self-concern is the “propagation of rational agency.” (p. 317) It appears that Irwin finds in eros the link to join more surely the self-regarding and other-regarding aspects of virtue. A kind of eros belongs to the rational part of the soul, which allows it to share something of the irrational appetites. It is ultimately, then, eros that allows Plato to step beyond a “narrowly instrumental attitude to practical reason,” (p. 306) to understand that justice is a “non-instrumental good.” (p. 305)

Of course, I have not been able to comment on every aspect of Irwin’s long and learned book. I have tried to bring together the most significant and striking claims, while offering observations on the quality of argumentation. Why do we read ancient Greek philosophy? Why do we read ancient Greek ethics? A text is worth reading as philosophy if and only if the philosophical questions are real, the conclusions defensible, the arguments plausibly informing to continuing philosophical discussion. Little is to be gained from time spent poking a corpse, speculating on past mental events from the minds of dead people. If an ancient philosophical text is worth reading, it should be because there is a richness of ideas, a depth and subtlety of reasoning that still speaks to us today. I find such a richness in the notion of the good life as a rational end—one that is not given but is achieved through excellence—and in the arguments over what it is and how obtained. Yet this is what I most miss in the product of Irwin’s considerable efforts. The question of the good life is obscured by the introduction of distinctions that make for a poor fit. We are then presented first, with a Socrates who thinks virtue is only a means, to what he doesn’t quite know, although it has something to do with pleasure, then, with a Plato who can reject the instrumentalism only by incorporating the irrational. Thus, in the end, we find a primitive utilitarianism, superseded by a version of emotivism. The concept of a rationality of ends disappears in an image of our recent philosophical history, fuzzy and partly occluded because projected onto the distant past.