THE Republic opens with a request for a definition of δικαιοσύνη, and the first book clarifies the nature of this request. The definition of justice as “telling the truth and paying one’s debts” is rejected, not only because it may sometimes be right to withhold the truth or not to return what one has borrowed, but because no list of types of action could supply what Plato is demanding. What he wants to know is what it is about an action or class of actions which leads us to call it just. He wants not a list of just actions, but a criterion for inclusion in or exclusion from such a list. Again, a definition of justice as “doing good to one’s friends and harm to one’s enemies” is rejected not just because of the argument that to harm someone would be to make him worse -that is, more unjust-with the consequence that the just man would be involved in making men less just, but because any definition of justice in terms of “doing good” and the like is bound to be unilluminating. When Thrasymachus comes upon the scene he tells Socrates that Socrates is not to offer him a definition which tells him that justice is “the same as what is obligatory or useful or advantageous or profitable or expedient.” Socrates retorts that this is like asking what 12 is and refusing to accept any answer of the form that it is twice 6, or 3 times 4, or 6 times 2, or 4 times 3. But Socrates does accept the task of offering a quite different kind of elucidation; it would be a mistake to suppose that when Socrates does offer us a formula, namely that justice is that state of affairs in which everyone has regard to his own concerns, this is in itself the answer that was being sought. This formula is unintelligible apart from the rest of the Republic, and Thrasymachus is right to suppose that the search for expressions synonymous with δικαιοσύνη would not be to the point. For to be puzzled about a concept is not like being puzzled about the meaning of an expression in a foreign language. To offer a verbal equivalent to an expression about whose meaning we are conceptually puzzled will not help us, for if what we are offered is a genuinely synonymous expression, then all that puzzled us originally will puzzle us in the translation. To understand a concept, to grasp the meaning of an expression, is partially, but crucially to grasp its functions, to understand what can and cannot be done with and through it. Moreover, we cannot decide what words are to mean, or what role concepts are to play, by fiat. We may on occasion wish to introduce a new concept and so legislate as to the meaning of a new expression; but what we can say in a given situation is limited by the common stock of concepts and the common grasp of their functions. No objection to the Republic is therefore more misconceived than that which would have been made by Humpty Dumpty (“When I use a word . it means just what I choose it to mean-neither more nor less”)14 and which was in fact made by Professor Karl Popper when he wrote, “But was Plato perhaps right? Does ‘justice’ perhaps mean what he says? I do not intend to discuss such a question. . I believe that nothing depends upon words, and everything upon our practical demands or decisions.”15 The point I have tried to make is that only those demands and decisions are open to us which there are concepts available to express, and that therefore the investigation of what concepts we either must or may use is crucial.

Thrasymachus’ own elucidation of the concept of justice is as follows. He does not believe that “just” means “What is to the interest of the stronger”; but he does believe that, as a matter of historical fact, rulers and ruling classes invented the concept and the standards of justice for their own purposes, and that it is in fact more profitable to do what is unjust rather than just. Socrates’ initial probing of Thrasymachus’ position is highly reminiscent of the Gorgias. He questions the concept of “the stronger” just as he did before, and he argues that the τέχνη of ruling, on the analogy of the τέχνη of medicine, must, if it is true art, be practiced for the benefit of those upon whom it is exercised. Medicine is for the benefit of patients, not for that of doctors, and so ruling must be for the benefit of the people, not of the rulers. But this thoroughly ineffectual analogy only belongs to the preliminary, sparring. The position that Socrates had finally restated in the Gorgias is one which can be strikingly attacked from Thrasymachus’ premises, and is so attacked by Socrates’ own disciples Glaucon and Adeimantus. But before Socrates completes the reiteration of his earlier attack on unlimited self-assertion-that restraint within the personality and between people is a condition of their well-being-he invokes the concept of ἀρετή, and the notion that there is a specifically human virtue, to exercise which will be to be in a state of well-being or happiness. Ἀρετή belongs now not to a man’s specific social function, but to his function as a man. The connection between virtue and happiness is written into this concept in what initially must seem an arbitrary way; the rest of the argument of the Republic is an attempt to remove this arbitrariness.

The revival of Thrasymachus’ case by Glaucon and Adeimantus runs as follows. Men in a state of nature are moved entirely by self-interest; the origin of laws lies in the moment when men discovered and agreed that clashes of self-interest were so damaging that it was more to their interest to forego doing injury to others than to continue in their natural way of life, so risking any injury that others might do to them. And ever since, men have obeyed the law only from fear of consequences; if men could avoid suffering the ill consequences of their actions, unlimited self-love would manifest itself openly instead of in law-abiding disguises. Suppose two men, one man now apparently just, the other unjust, were given a magic ring such as Gyges had to make himself invisible, so that both had complete liberty of action; then both would behave in the same way. They would, like Gyges, who seduced his queen and murdered his king, pursue the path of complete self-aggrandizement. That is, everyone prefers injustice to justice if he can be unjust successfully.

This example depends on that fallacious portrait of the presocial, natural man which I have already criticized. For Plato’s suggestion is that perhaps Gyges with his ring is natural man. The superiority of the present case over that originally put into the mouth of Thrasymachus is that Plato now turns sharply toward the identification of self-interest as a trait in social and not merely in natural man. He makes Adeimantus stress that the conventionally virtuous and just citizen is on the side of Thrasymachus, not of Socrates. For the Greek equivalent of the bourgeois father teaches his children to pursue virtue and to flee vice precisely and only because virtue brings rewards and vice has unfortunate consequences both in this world and in the next. But if these are the only reasons for praising virtue, how can justice in itself and apart from any rewards, be more profitable than injustice?

Plato’s answer is to try to show what justice is, first in the state, and then in the soul. He outlines a state in which all basic needs are met. Three classes of citizen are required: artisans and laborers to produce the material needs of society; soldiers to defend the state; and rulers to organize its social life. The key transition here is from recognizing three functions which have to be discharged in social life to asserting that three distinct and separate classes of citizens are needed, one to discharge each function. Plato relies for this transition upon two beliefs, one not certainly true, and one certainly false. The belief which is not certainly true is that one man is better to stick to one job, that this form of the division of labor is under all possible circumstances the best form; the belief which is certainly false is that men are by nature divided up into men best suited for each of these functions. Of this belief it might just be noted that it is invoked most often by those who believe that people like themselves are well suited to rule, while others are not; and that it ignores the fact that most people have different capacities which do not exclude one another, let alone the fact that in existing societies most abilities of most people are unrealized. But Plato’s beliefs on this point were powerfully reinforced by his doctrine of the tripartite soul.

The arguments for the tripartite soul are independent of those for the tripartite state, but it is necessary for the doctrine of the tripartite state that at least something like the doctrine of the tripartite soul should be true. That the soul has parts is shown, according to Plato, by the fact that it has conflicts. If a given man desires to drink (because he is thirsty) and does not desire to drink (because he suspects the condition of the water) at one

and the same time, then, since the same predicate cannot both hold and not hold of the same subject in the same respect at one and the same time, there must be at least two different subjects, of one of which we are predicating the desire to drink and of the other of which we are predicating the desire not to drink. The assumption which underlies this argument is that a man cannot simultaneously desire to do something and desire not to do it, in the same sense in which a man cannot simultaneously move in a given direction and not move in that direction. But where desires are concerned a man may desire some end, envisaged under a particular description, and not desire it under some other description. So a man may want to drink the water, because he is thirsty, but want not to drink, in case he risks an illness. It might perhaps seem that the short way to escape Plato’s argument here is to say that the man just does not have incompatible desires. He desires to quench his thirst, and he desires not to be ill, and it is merely a contingent fact that this water would both quench his thirst and make him ill. But to this the rejoinder might be that his desires remain incompatible: for what the man desires is to drink this particular water and what he fears is to drink this same water. Yet Plato is only right in a sense about these being incompatible desires, and that they are does not have any of the consequences which he supposes to follow. And this is because the incompatibility belongs to the possibility of satisfying both desires, not of having both desires. This is important because Plato uses his bad argument to expound a distinction between that part of the soul which is the appetites and that part which is the reason, a distinction which exerts enormous pressure upon some subsequent moral philosophy.

Plato’s picture of the parts of the soul is not in fact coherent. Sometimes he speaks as though the rational part of the soul had one set of desires and the appetitive part another; at other times, as if the appetites were the desires, and reason essentially a check and restraint upon them. He speaks as though the desire to drink was a nonrational craving, the apprehension of danger from drinking an insight of reason. But in fact we do not first have desires and then afterwards reason about them; we learn-and we use our reason in learning-to desire certain things (Plato does not distinguish the biologically determined appetite from the conscious human desire), and the desire to quench one’s thirst is as rational as the desire not to be injured by poison in the water. It is just not true that only our restraint upon ourselves derives from reflection; it is often upon reflection that we decide that we need to drink. An irrational fear of being poisoned might be checked by a reasonable desire to quench one’s thirst, just as much as vice versa-an irrational desire to quench one’s thirst might be inhibited by a rational fear of poison. What makes a desire reasonable or unreasonable is its relation to our other purposes and choices, possible as well as actual. A man may behave unreasonably by not allowing his desires play, and desire may on occasion correct an agent’s would-be rational assessments. But these facts Plato, and a long tradition which is to follow him, rule out of court in order to maintain that rigid division between reason and the appetites in which reason is always to be in the right.

The original source of this distinction is of course not in Plato’s own arguments from the alleged facts of conflict, but in his inherited Pythagorean and Orphic beliefs in the separation of an immortal soul from a body that is a prison and a tomb. But later writers, who might have been unimpressed by the religious doctrine, have still been content with the philosophical distinction. Plato himself gives a far more interesting and positive account of desire in the Symposium; but even here desire leads us away from this world in the end.

Plato’s doctrinal allegiances lead him in the Republic not merely to draw false conclusions from the facts of conflict but also, as I have just suggested, to misdescribe them. The essence of conflict of desire is that it provides an occasion for choice on my part between my desires, even if I do not choose. But Plato’s division of the soul into parts makes conflict a tug of war, which could not be an occasion for choice. “I” am not confronted with my desires. “I” am split between two autonomous parties, reason and appetite; or else “I” am reason, struggling against appetite. Nor is Plato consistent here with his other writings. The Greek word for soul, ψοχή, means originally simply that which makes the difference between life and death, between a man and a corpse. Some early Greek thinkers identify the soul with a material substance; the Pythagoreans with a harmony between the elements of the body, a balance. Plato argues against both in the Phaedo that the soul is an immaterial simple substance; that to be destroyed is to be divided up into parts; and that since the soul has no parts it must be immortal. Appetite in the Phaedo belongs to the body, so that the distinction between reason and appetite remains a constant element, pointing to the continuity of the religious background. But the Phaedo offers us no grounds for believing in the division into parts of the soul.

The division of the soul in the Republic is not just between reason and appetite; there is also the “spirited” part, which is concerned neither with rational standards of behavior nor with bodily desires, but with standards of honorable behavior, and with anger and indignation. Plato tells the story of Leontius, who, overcome with desire, stares at the corpses of executed criminals, cursing himself as he does so. The Platonic moral is that anger and appetite can conflict. The spirited part of the soul acts, when “it is not corrupted by a bad upbringing,” as an agent of reason, being indignant when reason is overborne. So a man who has been wronged feels indignant, but a man who feels that he is in the wrong cannot find it in his nature to be indignant if he is made to suffer in turn. So Plato says.

Men therefore fall into three classes depending upon which part of the soul is dominant; this division is that required by the tripartite state. Into which class a man falls may in part be a matter of his early training, but cannot fundamentally be so determined. Plato believes that there are born shoemakers and born rulers. Justice in the state is a matter of everyone knowing his place. Of the four traditional virtues, courage belongs to the class of auxiliary guardians whose function is defense, and wisdom to the ruling guardians. Temperance is a virtue not of a class, but of the society as a whole because “the desires of the inferior multitude will be controlled by the desires and wisdom of the superior few.” Justice belongs not to this nor to that class, nor to particular relationships between classes, but to the society’s functioning as a whole.

Justice in the soul is likewise a matter of each part of the soul performing its proper and allotted function. An individual is wise in virtue of reason ruling in him and brave in virtue of the spirited part playing its role; an individual is temperate if his inferior bodily appetites are ruled by his reason. But justice belongs not to this or that part or relationship of the soul, but to its total ordering. The two questions then arise, What sort of man will be just? and, How could the just state come about? These questions are asked and answered together, and this is no accident. When, later on, Plato comes to discuss the corruption of state and soul he treats them as belonging together. Moreover, the just man will rarely exist except in the just state, where at least some men-the future rulers-are systematically educated in justice. But the just state cannot possibly exist except where there are just men. So the questions of how the state can come to be and of how the just man is to be educated have to be posed together. And so we reach the point where Plato brings on stage the ideal of the philosopher-king.

Plato defines a philosopher by setting out an account of knowledge and belief and then contrasting the philosopher, who knows, with the nonphilosophical man, who at best has only true belief or opinion. The argument begins from considering the meaning of pairs of predicates, and the examples used are beautiful and ugly, just and unjust, and good and bad. Plato says that “Since beautiful and ugly are opposite, they are two; and so each of them is one.”16 But many things exhibit beauty and many things are ugly. So that there is a difference between those who are aware of this or that object as beautiful and those who grasp what “beautiful itself” is. I use this expression “beautiful itself” (“αὐτὸ τὸ καλόν”) to translate Plato’s innovating use of itself to convert an adjective into an expression that names what the adjective is supposed to mean or stand for. And I use the expression “mean or stand for” not because I want to mislead the reader into supposing that “meaning” and “standing for” are the same, but because Plato makes just this mistake. The identification comes about in this way. Plato contrasts the man who uses the word beautiful in an ordinary, confused way with the man who has really grasped what beautiful means, and he interprets this contrast as the contrast between the man who happens to be acquainted with a number of beautiful objects and the man who is acquainted with that which beautiful stands for. The former man is in possession only of “belief”; his judgments are not reinforced by a well-grounded understanding of the meaning of the expression which he uses. The latter is in possession of knowledge; for he really understands what he says.

Knowledge (ἐπιστήμη) and belief or opinion (δόξα) can then be defined in terms of contrasting classes of objects. Belief is concerned with the world of sense perception and of change. Of this fleeting and evanescent realm we can have at best only true opinion. Knowledge is concerned with unchanging objects, about which we can have secure, rationally founded views. Plato’s distinction between knowledge and belief is a complex one. In part it is a straightforward distinction between those convictions which, because they are acquired by reasoning and backed up by argument, are not at the mercy of clever orators (that is, knowledge) and those convictions which, being a matter of nonrational conditioning anyway, are liable to change whenever subject to the techniques of nonrational persuasion (that is, belief). But clearly this distinction has nothing to do with the subject matter of our beliefs. It concerns rather the different ways in which individuals may acquire and hold their beliefs. Why, then, should Plato suppose that his distinction is one of subject matter? The reason is that Plato thought himself to have independent grounds for believing that no secure, rationally grounded judgments could be made about the subject matter afforded by sense perception. Some of these grounds were derived from earlier philosophers. Both Heraclitus and Protagoras had emphasized the relativity of judgments of sense perception. But the point that Plato is concerned to make can be detached from the detail of their particular doctrines.

If I can say of quite different objects that they are beautiful, and of the same object at one time or from one point of view that it is beautiful and at another time or from another point of view that it is ugly, then the meaning of the predicates beautiful and ugly cannot be explained simply by referring to the objects to which they are applied. This is not just because, as Plato pointed out in the passage cited earlier, the objects are many and the

meaning single; it is also because the judgments are liable to variation, and to contradiction by other judgments, whereas the one thing that does not vary is the meaning. To put it in a much later mode of speech, Plato is engaged in elucidating what is involved in describing two or more uses of an expression as instances of the use of one and the same predicate. The difference from Socrates is that Socrates saw only that the use of ethical predicates must be governed by criteria; whereas Plato supposed that if this is to be so, that if there are to be objective standards for the use of such predicates, it must be the case that such predicates are used to refer to objects, and objects belonging not to the multifarious, changing world of sense, but to another, unchanging world, apprehended by the intellect precisely through its dialectical ascent, whereby it grasps the meaning of abstract nouns, and of other general terms. These objects are the Forms, through the imitation of which or participation in which the objects of sense perception have the characters that they have.

The philosopher is the man who has learned through a training in abstraction to acquaint himself with the Forms. He alone therefore really understands the meaning of predicates, and he alone has genuinely founded moral and political views. His training is primarily in geometry and in dialectic. By dialectic Plato understands a process of rational argument which is a development from the dialogue of the Socratic interrogation. Beginning from some proposition which has been advanced for consideration, one ascends in one’s search for justifications up a deductive ladder until one reaches the indubitable certainties of the Forms. What Plato presents in the Republic is a progress in rational argument, culminating in a vision of the Form of the Good (that is, in a vision of what the predicate good stands for-that in virtue of which it has meaning). In the Republic there is, too, a strongly religious attitude toward the supreme Form, the Form of the Good. The Form of the Good is not one among the other Forms we contemplate: they belong to the realm of unchanging existence-the Form of the Good dwells beyond existence. Just as it is that by virtue of the sun’s light we see everything else, but if we look into the sun itself we are dazzled, so it is that in the intellectual light given out by the Form of the Good we grasp the other Forms, but we cannot contemplate the Form of the Good itself.

Good, then, for Plato-at least in the Republic-is only used properly when it is used as the name of a transcendent entity or when it is used to express the relation of other things to that entity. The difficulties in Plato’s conception of Forms were first formulated by Plato himself in later dialogues; at the moment we need only note that to suppose that there are Forms does not in fact do anything to solve the problem which Plato is posing in the Republic-that of how a predicate with a single meaning can be applied in many different ways and to many different subjects. For to say that a predicate derives its meaning from one primary case leaves it entirely obscure as to how this predicate is then capable of being applied in other cases. But just this was what we wanted to know. Moreover, we are at once involved in logical oddity. For if this is how we answer our problem, we are involved in saying that the primary application of beautiful is to the Form of Beauty, that of high to the Form of Height, and so on. But to say that “Beauty is beautiful” or that “Height is high” is clearly not to speak with a clear meaning. This fact Plato himself brought out in later criticisms of his own position.

What is important is that the theory of meaning has been decisively brought on the scene. The logician has entered moral philosophy for good. But even though, from now on, the systematic and self-conscious logical analysis of moral concepts will be at the heart of moral philosophy, it can nonetheless never be the whole of moral philosophy. For we have to understand not only the logical interrelations of moral concepts, rules, and the like but also the point and purpose such rules serve. This involves us both in the theory of human purposes and motives and in the theory of society, since different kinds of wants and needs are dominant in different social orders. We can see all three interests, the epistemological, the psychological, and the political, meeting in the central parts of the Republic. For Plato’s first theory, that we can understand what goodness is, what justice is, what courage is, and so on, by seeing them exhibited in a certain type of state and a certain type of soul, now has to be reconciled with Plato’s second theory, that we can only understand what goodness, justice, and the rest of them are if we become acquainted with the relevant Forms. However, not only is a reconciliation not difficult, but it enables Plato to make his earlier contentions more cogent. The rulers of the just state, in whom the rule of reason is present, are rational in virtue of an education which has enabled them to apprehend the Forms. In the just state the philosopher is king; only he can bring into being and maintain in being a state in which justice is embodied both in the political arrangements and in the soul. It follows that the class division of the just society can, as Plato had earlier suggested, be maintained by educating some to be rulers, others to be auxiliaries, most to be ruled; the use of eugenic controls and selection methods is to insure that those fit for the education of rulers receive it. To make the common people content they will be told a story about the metals in the soul: precious in the souls of rulers, base in the souls of the ruled. Plato does not believe in a correlation between intelligence and some merely accidental property, such as a color of the skin, in the way that racists in South Africa and Mississippi believe; he does, however, believe in the occurrence of inborn intelligence, or the lack of it, in the way that conservative educationalists do; and he believes that ingenious propaganda-the telling of what he calls “noble lies”-can insure that inferior people will accept the fact of their own inferiority.

Those of superior intelligence proceed to the vision of the Forms in ways that Plato delineates by means of two different parables, that of the Line and that of the Cave. The Line is divided horizontally; below the division lie the realms of imagining and perceiving, while above it lie those of mathematical entities-which for Plato are closely related to Forms-and of the Forms. The passages about the Cave picture men chained so that they cannot see the daylight; behind them a fire and a puppet show are so contrived that the prisoners see a procession of shadows on the wall. They believe that the words in their language refer to the shadows and that the shadows are the sole reality. A man who escaped from the Cave would slowly accustom himself to the light of the world outside. He would pick out, first, shadows and reflections; then physical objects; and finally, the heavenly bodies and the sun. This is for Plato a parable of the ascent to the Forms. The man who returns to the Cave will be unaccustomed to the darkness; he will not for some time identify the shadows in the Cave as well as do those of his former companions who never left the darkness; and he will cause great resentment by his subsequent claims that the shadows are devised and unreal and that the true reality lies outside the Cave. So great will the resentment be, that if the chained men could, they would kill this man from the outer world-precisely as the Athenians killed Socrates.

What, then, is the philosopher who has ascended to the Forms to do? It will only be at the rarest moments in history, and possibly it will never happen at all, that he will have the possibility of intervening to create the just state. Plato himself, first in his response to the Athenians’ treatment of Socrates, and then in his own disillusionment with the tyrannical rulers of Syracuse, has a deep pessimism about political life. But if the ideal state can never become real, what was the point of depicting it? Plato’s answer is that it provides a standard against which we can judge actual states. This is part of what Plato himself is doing when he pictures a series of stages of decline from the ideal state and from the just soul; in so doing he brings out further the intrinsic connection which he believes to hold between politics and psychology.

The first stage in decline is the timocratic state; here the military and the guardians have fallen out, and the state is based on the military values of honor with some infusion of the values of private property. The next stage, the oligarchical, is one in which the class structure is maintained only in the interest of the ruling class, and not at all in the interest of the whole state; the rich use the class structure to exploit the poor. In the third phase the poor revolt and create a democracy, in which every citizen is equally free to pursue his will and his personal aggrandizement, while finally the would-be despot is able to enlist from such a democracy enough dissatisfied malcontents to create a tyranny. Plato’s aim here is at least twofold; he has placed the actual forms of constitution of Greek city-states upon a moral scale, so that even if we cannot have the ideal, we know that timocracy (traditional Sparta) is best, oligarchy (Corinth) and democracy (Athens) worse, and tyranny (Syracuse) worst of all. But his argument also brings out that one reason why they can be morally evaluated is because to each type of constitution a type of personality corresponds. In timocracy the appetites are restrained and ordered, but not by reason. Honor instead has this role. In oligarchy they are still disciplined, but only by the love of wealth and a regard for stability which springs from a regard for property. In democracy every taste, every inclination has equal sway in the personality. And in tyranny-in the men with despotic souls-the baser appetites, that is the bodily ones, exercise absolute and irrational control. Plato now uses this classification of personality types in order to return to the question of the justification of justice in the form in which Glaucon and Adeimantus had raised it. To do this Plato compares the external and opposed positions of the just man and the despotic man, who now turns out to be the extreme personality type of the unjust man.

Plato has three arguments to show that the just life is happier than the unjust one. The first is that the unjust man sets no curb upon his desires, and so his desires are without limit. But, being limitless, his desires can never be satisfied, and so he will always be discontented. The second argument is that only the philosopher is in a position to contrast the pleasures of reason with those of limitless appetite and sensuality, for he alone knows both sides. Finally, it is argued that the pleasures of intellect are genuine, while what the man of appetite takes to be pleasure is often merely a cessation of pain or discomfort (as eating relieves hunger) and at best far less real (in terms of the notion of the real as the unchanging and immaterial) than what the intellect delights in. These are bad arguments. The third depends for part of what it seeks to prove upon the arguments about the Forms, and it in any case ignores-with Plato’s characteristic and utterly deplorable puritanism-the many genuine bodily pleasures; the second is simply false-even in Plato’s terms the philosopher is no more acquainted with the pleasures of limitless desire than the sensualist is with the delights of rational control; while the first argument fallaciously infers from the premise that the sensualist will always have appetities which have not yet been satisfied the conclusion that he will always be and feel unsatisfied and dissatisfied. But it is not the badness of the particular arguments that is so important. Given Plato’s psychology, only bad arguments were available to him. For the complete divorce of reason and desire in the soul entails that the contrast has to be between reason on the one hand and senseless and uncontrolled appetite on the other. These are the only alternatives available, given the Platonic psychology; but in fact they are not, of course, the only or even the most important alternatives. In order to vindicate justice against injustice, Plato accepts the criterion implied by Thrasymachus’ exaltation of successful worldly, and especially, successful political ambition-the vulgar criterion of pleasure-and argues that the unjust but successful tyrant has less pleasure, is more discontented, than the just man, even than the just man unjustly done to death. But to do this, he has to equate the unjust man with the man who pursues pleasure limitlessly and senselessly. And Plato has to make this equation, since reason, in the Platonic scheme, can only dominate, not inform or guide, appetite, and appetite of itself is essentially irrational. The man who in fact threatens the prestige of justice is not the senseless sensualist or the unchecked tyrant, but much more often l’homme moyen sensuel, the man who is everything, including unjust and vicious, in moderation, the man whose reason restrains his vice today in the interests, not of virtue, but of vice tomorrow. This is the man who praises virtue for what he can get out of it in the way of wealth, office, and reputation, and this is the man whom Glaucon and Adeimantus had in mind. This was why the case that Glaucon and Adeimantus propounded was so much more of a threat to Plato than the case put by Thrasymachus. But Plato’s conceptual scheme tempts him into considering this man, whom he observes and describes with tolerable accuracy in the oligarchical and democratic states, as merely a less extreme version of despotic man. But despotic man is drawn so extremely that what is described is no longer a possible moral type. I can be said to pursue pleasure only if I am pursuing identifiable goals and making choices between alternatives in terms of them. The man who can no longer make choices but passes on heedlessly and inevitably from one action to the next is not a possible normal human type, but rather a compulsive neurotic. And this may have been what

Plato wished to describe, for he very strikingly connects the behavior of this man with the pursuit of those fantasies of which most men are conscious only in dreams, thus strikingly anticipating Freud. But any classification which entails making the way of life of l’homme moyen sensuel merely a moderate version of the compulsive behavior of the neurotic, and lumps both together in contrast with rationality, is thereby condemned as a classification. Nor does the myth with which the Republic closes help Plato. For the suggestion that in the realm after death the just will be rewarded and the unjust punished appropriately depends for its force upon the notion that justice is indeed superior to injustice, that the just man deserves his reward and the unjust his punishment. So that the question of the justification of justice is still left without a clear answer. A very brief reconsideration of the central arguments of the Republic makes it clear why on Plato’s terms this has to be so.

The argument begins from the need for an understanding of the meaning of ethical predicates apart from their particular applications. This starting point will recur in the history of philosophy in writers as different from Plato as St. Augustine and Wittgenstein. When we inquire about what it is for something to be just or red or equal, the rational first move is to offer examples, to try and give a list of just actions or red objects or cases of equality. But such a list misses the point of the inquiry. What we want to know is not which actions are just, but what it is in virtue of which actions are just. What is it that enables us to mark off those cases which genuinely belong on our list from those that do not? We need a criterion. Wittgenstein will suggest that the criterion is embodied in a rule, and the rule in a socially established practice. Augustine will suggest that the criterion is given by an interior illumination which is a gift of God. Plato finds his criterion in the knowledge of the Forms. But knowledge of the Forms is accessible only to a few, and only to those few who have either enjoyed the educational disciplines of the as yet nonexistent ideal state or are among the very rare natures which are both philosophically capable and inclined and also not corrupted by the social environment. It follows not only that only these few will be able to perform the task of justifying justice but also that only to them will the justification be intelligible and convincing. Thus the social order which the Platonic concept of justice enjoins could only be accepted by the majority of mankind as a result of the use of nonrational persuasion (or force).

Everything therefore turns for Plato on the possibility of establishing, first, that there are Forms and that knowledge of them has the role which he claims, and second, that only a minority are capable of this knowledge. The latter is merely asserted and never argued for. The former depends upon arguments about which Plato himself, as we shall see, came to have serious doubts. But behind all Plato’s explicit statements there lies a further assumption which must now be brought into the open.

We speak of justification in at least two radically different types of context. Within a discipline like geometry the justification of a theorem consists in showing how it follows validly from the axioms. There is no question here of what counts as a justification for one person not counting as a justification to another. Within the field of conduct, however, this is not so. To justify one course of action as against another is not only to show that it accords with some standard or conduces to some end but also to show this to someone who accepts the relevant standard or shares the particular end. In other words, justifications of this type are always justifications to somebody. Aristotle later tries to show how there are certain specifically human ends in the light of which policies of action can be justified to rational beings as such. But Plato restricts the class to whom his justifications can be addressed to those who have acquired knowledge of the Forms. When, subsequently, he discusses the justification of justice in terms independent of this knowledge, that is in the passages where he compares types of state and of soul, he in fact falls back on comparisons, half a priori, half empirical, which are bound on his own terms to founder-for they presumably belong to the world of opinion, of δόƷα, not of ἐπιστήμη, of knowledge-except against the background of a transcendental knowledge which has been pointed to but never brought on the scene.

One root of Plato’s mistake here is his confusion of the kind of justification which is in place in geometry with that which is in place in matters of conduct. To treat justice and good as the

names of Forms is to miss at once one essential feature of justice and goodness-namely, that they characterize not what is, but what ought to be. Sometimes what ought to be is, but more often not. And it always makes sense to ask of any existing object or state whether it is as it ought to be. But justice and goodness could not be objects or states of affairs about which it would make sense to inquire in this way. Aristotle was to make very much this criticism of Plato; Plato’s own blindness to it is one contributory factor to his curious combination of an apparent total certitude as to what goodness and justice are, and a willingness to impose his own certitudes upon others, with a use of profoundly unsatisfactory arguments to support his convictions.