THE SUGGESTION that asking and answering moral questions is one thing, and asking and answering philosophical questions about morality quite another thing, may conceal from us the fact that in asking moral questions of a certain kind with sufficient persistence we may discover that we cannot answer them until we have asked and answered certain philosophical questions. A discovery of this kind provided the initial impulse for philosophical ethics in Greek society. For at a certain period, when moral questions were asked, it became clear that the meaning of some of the key words involved in the framing of those questions was no longer clear and unambiguous. Social changes had not only made certain types of conduct, once socially accepted, problematic, but had also rendered problematic the concepts which had defined the moral framework of an earlier world. The social changes in question are those reflected in Greek literature in the transition from the Homeric writers through the Theognid corpus to the sophists.

The society reflected in the Homeric poems is one in which the most important judgments that can be passed upon a man concern the way in which he discharges his allotted social function. It is because certain qualities are necessary to discharge the function of a king or a warrior, a judge or a shepherd, that there is a use for such expressions as authoritative and courageous and just. The word ἀγαθός, ancestor of our good, is originally a predicate specifically attached to the role of a Homeric nobleman. “To be agathos,” says W. H. Adkins, “one must be brave, skilful and successful in war and in peace; and one must possess the wealth and (in peace) the leisure which are at once the necessary conditions for the development of these skills and the natural reward of their successful employment.”2 Ἀγαθός is not like our word good in many of its Homeric contexts, for it is not used to say that it is “good” to be kingly, courageous, and clever-that is, it is not used to commend these qualities in a man, as our word good might be used by a contemporary admirer of the Homeric ideal. It is rather that ἀγαθός is a commendatory word because it is interchangeable with the words which characterize the qualities of the Homeric ideal. So in our ordinary English use of good, “good, but not kingly, courageous, or cunning” makes perfectly good sense; but in Homer, “ἀγαθός, but not kingly, courageous, or clever” would not even be a morally eccentric form of judgment, but as it stands simply an unintelligible contradiction.

How do adjectives of appraisal, such as ἀγαθός and others, function in Homer? First of all, to ascribe the qualities for which they stand to someone is to make a factual statement, in the sense that whether what you have said is true or false is settled by the man’s performances and settled simply and solely by his performances. The question, Is he ἀγαθός? is the same as the question, Is he courageous, clever, and kingly? And this is answered by answering the question, Does he, and has he, fought, plotted, and ruled with success? The point of such ascriptions is in part predictive. To call a man ἀγαθός is to tell your hearers what sort of conduct they can expect from him. We ascribe dispositions to the agent in the light of his behavior in past episodes.

From this alone it is strikingly plain that the Homeric use of ἀγαθός does not square at all with what many recent philosophers have thought to be the characteristic properties of moral, and indeed of evaluative, predicates. For it has often been held3 to be an essential feature of such predicates that any judgments in which one is ascribed to a subject cannot follow logically as a conclusion from premises which are merely factual. No matter what factual conditions are satisfied, these by themselves can never provide sufficient conditions for asserting that an evaluative predicate holds of a subject. But in the Homeric poems, that a man has behaved in certain ways is sufficient to entitle him to be called ἀγαθός. Now, assertions as to how a man has behaved are certainly in the ordinary sense factual; and the Homeric use of ἀγαθός is certainly in the ordinary sense evaluative. The alleged logical gulf between fact and appraisal is not so much one that has been bridged in Homer. It has never been dug. Nor is it clear that there is any ground in which to dig.

Moreover, I fail to be ἀγαθός if and only if I fail to bring off the requisite performances; and the function of expressions of praise and blame is to invoke and to justify the rewards of success and the penalties of failure. You cannot avoid blame and penalty by pointing out that you could not help doing what you did, that failure was unavoidable. You may, of course, certainly point this out; but if your performance failed to satisfy the appropriate criteria, then you simply cannot prevent the withdrawal of the ascription of kingliness, courage, and cleverness or cunning. And this is to say that Homeric moral predicates are not applied, as moral predicates have been applied in our society, only where the agent could have done other than he did. Excuses, praise, and blame must all play different parts. We cannot even inquire whether (in the Kantian sense) ought implies can for Homer, for in Homer we cannot find ought (in the Kantian sense). So Odysseus blames the suitors, when he returns to Ithaca, for having had a false belief: “Dogs, you did not think that I would return home from Troy; for you have consumed my possessions, lain with my maidservants by force, and wooed my wife while I was yet alive, fearing neither the gods who inhabit the broad heaven, nor yet that there would be any retaliation from men hereafter; but now the doom of death in upon you all.”4 The suitors are blamed precisely for having a false belief; but this is what in a modern sense we would feel we could not blame people for. For to believe is not to perform an avoidable action. And it is not that Homer thinks that beliefs are voluntary; he is engaged in an assessment to which what the agent could or could not have done otherwise is irrelevant.

It will be useful now to look at a cognate of ἀγαθός in Homer, the noun ἀρετή, usually and perhaps misleadingly translated virtue. A man who performs his socially allotted function possesses ἀρετή. The ἀρετή of one function or role is quite different from that of another. The ἀρετή of a king lies in ability to command, of a warrior in courage, of a wife in fidelity, and so on. A man is ἀγαθός if he has the ἀρετή of his particular and specific function. And this brings out the divorce of ἀγαθός in the Homeric poems from later uses of good (including later uses of ἀγαθός). When Agamemnon intends to steal the slave girl Briseis from Achilles, Nestor says to him, “Do not, ἀγαθός though you be, take the girl from him.”5 It is not that, being ἀγαθός, Agamemnon can be expected not to take the girl, nor that he will cease to be ἀγαθός if he does take her. He will be ἀγαθός whether he takes her or not. The way in which “ἀγαθός” is tied so completely to fulfillment of function is also brought out in its links with other concepts. Shame, αἰδώς, is what is felt by a man who fails to perform his allotted role. To feel shame is simply to be aware that you have entitled people to accuse you of having fallen short of that which the socially established description both you and others had applied to yourself had led them to expect. It is to be aware that one is liable to reproach.

This whole family of concepts, then, presupposes a certain sort of social order, characterized by a recognized hierarchy of functions. It is noteworthy that the value predicates can only be applied to those men who fall under the descriptions which taken together constitute the social vocabulary of the system. Those who fall outside the system fall outside the moral order. And this is indeed the fate of slaves; the slave becomes a chattel, a thing, rather than a person. It would miss the point to comment upon this that the Homeric poems are not a historically accurate picture of early Greek society or that no society as rigorously functional in fact existed. What we get in Homer is rather an idealization of one form of social life; we are presented with a social order and its concepts in a fairly pure form, rather than in the kind of admixture of several forms which a total society often presents. But for our conceptual purposes this is none the worse. For we have other literary documents in which we can see how the breakdown of a social hierarchy and of a system of recognized functions deprives the traditional moral terms and concepts of their social anchorage. In the body of poems which pass under the name of Theognis of Megara,6 and which were written in post-Homeric and preclassical Greece, we find startling changes in the uses of ἀγαθος and ἀρετή. They can no longer be defined in terms of the fulfillment in a recognized way of a recognized function; for there is no longer a single and unified society in which evaluation can depend on established criteria of this kind. Words like ἀγαθός and κακός (bad) become sometimes merely neutrally descriptive of social position. Or they may acquire an even more radical extension of meaning. Both processes are seen at once in a passage which runs: “Many κακοι are rich and many ἀγαθοί are poor, but we will not take the wealth in exchange for our ἀρετή for the one remains with a man always, but possessions pass from one man to another.” Here ἀγαθός and κακός seem to mean nobleborn and baseborn or some such equivalents. They have lost their old meaning and been transformed into one of the key identifying descriptions under which those to whom the terms applied in their old sense now fall. But they are no longer evaluative in the same way. Whereas in Homer one would have said of a chieftain that he was ἀγαθός if and only ff he exercised his true function, now ἀγαθός describes someone who comes of a chieftain’s line, whatever function he may exercise or fail to exercise or whatever his personal qualities may be. But the transformation of ἀρετή in the same passage is quite different. For ἀρετή now denotes not those qualities by means of which a particular function may be discharged, but certain human qualities which may be divorced from function altogether. A man’s ἀρετή is now personal to himself; it has become far more like what modern writers think of as a moral quality.

Thus evaluative predicates come to refer to dispositions to behave in certain ways relatively independent of social function. With this change comes another. In Homeric society the dominant hierarchy of functional roles determines which are the dominant qualities; skill, cunning, and courage of various sorts. When this hierarchy collapses, the question can be opened in a far more general way of what the qualities are which we would wish to see in a man. “The whole of ἀρετή is summed up in δικαιοσύνη,” writes one Theognid author. “Every man, Cyrnus, is ἀγαθός if he has δικαιοσύνη.” Here anybody can be ἀγαθός by exercising the quality of justice, δικαιοσύνη. But what does this consist in? The pressures of the time not only make ἀγαθός unstable in meaning, they also raise doubts about the nature of δικαιοσύνη. For the idea of a single moral order has broken down.

It has broken down partly because of the breakdown of formerly unified social forms. These were reinforced by a mythology with the status of a sacred writing, including the Homeric poems, which suggested a single cosmic order. In Homer the order of necessity reigns over gods as well as over men. Ὓβρις, willful pride, is the sin of overstepping the moral order of the universe. Νέμεσις awaits whoever commits it. The moral order and the natural order are not sharply distinguished. “The sun will not go beyond his measures; otherwise the Ἐρινύες, the handmaids of justice, will find him out,” said Heraclitus. But this mythological assertion of order changes its function too, as Greek society changes. Anthropologists very commonly assert that myths express social structure. And myths can do this in more than one way. Between Homer and writers five centuries later there is a great change in Greek myths about the order in the universe. The Homeric myth does reflect, though with much distortion, the workings of an actual society in which a close form of functional organization is presupposed by the moral and evaluative forms of appraisal which are in use. The later assertions of order in the universe reflect not a structure that is, but one that was, or one that is struggling to survive. They are conservative protests against the disintegration of the older forms and the transition to the city-state. The myths themselves cannot but open up the question of the difference between the order of the universe and the order of society. But above all, this question is sharpened by a widening awareness of radically different social orders.

The impact of the Persian invasions, of colonization, of increase in trade and therefore in travel, all these bring home the fact of different cultures. The result is that the distinction between what holds good in Egypt but not in Persia, or in Athens but not in Megara, on the one hand, and what is the case universally as part of the order of things becomes overwhelmingly important. The question asked about any moral rule or social practice is, Is is part of the essentially local realm of νόμος (convention, custom) or of the essentially universal realm of ϕύσις (nature)? Linked to this is of course the question, Is it open to me to choose what rules I shall make my own or what restraints I shall observe (as it may be open to me to choose which city I shall live in and what therefore shall be the νομος by which I live)? or does the nature of the universe set limits upon what I may legitimately choose? To the task of answering these questions there came in the fifth century B.C. a new class of teachers and a new class of pupils. Books on moral philosophy commonly concentrated on the teachers, the sophists, whom we see mainly through the antagonistic eyes of Plato. But the activities of the sophists as suppliers are unintelligible apart from the demand which they met. So let us try to specify this demand still more precisely.

We have seen how the word ἀγαθός had become unstable in its attachments and so had its cognate words, especially ἀρετή. “Virtue” is what the good man possesses and exercises. It is his skill. But what virtue is and what constitutes a good man, these have become matters of conflicting opinion in which the Homeric concept of the ἀγαθός has been divided up between rival inheritances. There is, on the one hand, the good man conceived as the good citizen. The values of the conservative Athenian whom Aristophanes portrays are loyalty to the city and more especially to the older forms of social order. In this there is certainly an element of the Homeric ἀγαθός. But equally, the Homeric chieftain’s personal values, the values of the courageous, cunning, and aggressive king, are now, if exercised by the individual in the city-state, antisocial. Self-aggrandizement, the use of the state as something to be preyed upon, these are the only courses open to the individual who wants in the fifth century to behave like a Homeric hero. The social order in which his qualities were an essential part of a stable society has given way to one in which the same qualities are necessarily disruptive. So the relationship of the ἀγαθός to the social values and especially to justice has become a crucial issue. But δικαιοσύνη (which though very inadequately translated as justice, is as inadequately translated by any other word, for it has a flavor all its own and combines the notion of fairness in externals with that of personal integrity in a way that no English word does) is of all notions the one that appears most to be put in question by the discovery of rival social orders. Different cities observe different customs and different laws. Does and should justice differ from city to city? Does justice hold only within a given community between citizens? or should it hold also between cities? The Athenians condemn the character of Alcibiades because he did not observe the restraints of δικαιοσύνη in his behavior within the Athenian state. But their own envoys behave just like Alcibiades in their attitude to other states. That is, they equate what is morally permissible with what the agent has the power to do. Their envoys can say to the representative of Melos, an island which wished to remain neutral in the Peloponnesian War, “For of the gods we believe, and of men we know, that by a law of their nature whenever they can rule they will. This law was not made by us, and we are not the first who have acted upon it; we did but inherit it, and shall bequeath it to all time, and we know that you and all mankind, if you were as strong as we are, would do as we do.”7

Thus the redefinition of evaluative predicates creates a problem for those who wish to use them, even in formulating their own intentions. The terms ἀγαθός and ἀρετή have become genuinely problematic, as has their relation to καλός, the predicate characterizing what is well thought of, and as has their relationship to δικαιοσύνη. The moral and political conservative still feels able to give the words a fixed connotation. He uses texts and tags from Homer or Simonides to provide him with definitions. But, in general, slides in the meanings of words become appallingly easy and frequent. It is often impossible to distinguish two separate phenomena, moral uncertainty and uncertainty as to the meaning of evaluative predicates. In the moments of greatest perplexity in fifth-century Greece these two uncertainties are one. Thucydides has recorded the corruption of language in describing the revolution at Corfu: “The meaning of words no longer had the same relation to things, but was changed by them as they thought fit. Reckless doing was held to be loyal courage; prudent delay was the excuse of a coward; moderation was the disguise of unmanly weakness; to know everything was to do nothing.”8

How does Ayer’s conception of the distinction between moral philosophy and moral judgment or practice apply in this situation? Can we distinguish two separate activities, “the activity of a moralist, who sets out to elaborate a moral code, or to encourage its observance, and that of a moral philosopher, whose concern is not primarily to make moral judgments but to analyse their nature”?9 It must at once be conceded to Ayer that there are some questions involved in moral philosophy which are purely philosophical and others which are entirely independent of philosophy. There are many cases where what is important is to emphasize that to commit oneself to a particular philosophical analysis of moral judgments does not entail committing oneself to making a particular specified set of moral judgments. In a contemporary context two utilitarian philosophers might agree in their analysis of the expression wrong and without a shadow of inconsistency disagree as to whether it is the case that all wars are wrong or that only some wars are wrong. Equally, two pacifists might agree on this latter issue, but one might be a philosophical intuitionist, the other an emotivist. When Ayer argues that philosophical theories of moral concepts and judgments are neutral as regards conduct this is clearly the type of case he has in mind. But the point at which Greek moral philosophy begins suggests that there is also a quite different type of case.

In Ayer’s type of case the moral vocabulary is taken as given and determinate. There are then two problems, How shall I use it? (morals) and, How shall I understand it? (philosophy). Philosophy, it should be noted, becomes an essentially after-the-event activity. But in the cases where the meaning of the moral vocabulary is itself in doubt, the answer to the question, How shall I use the moral vocabulary? will consist in formulating rules, no doubt partly already implicit in the previous uses of the word, but partly perhaps also designed to avoid incoherences and ambiguities in previous uses. These rules will set the limits upon the possible uses of the moral predicates, and so philosophical elaboration of the concept will partly determine the moral uses of those predicates. Thus the question of the criteria which are to be employed in moral evaluation cannot be clearly demarcated as moral but not philosophical, or as philosophical but not moral. Of course, to clear up the conceptual problems is not of itself to determine completely how we ought to act or to judge, but it does determine the limits of moral possibility in part. The task of the moralist and the task of the philosopher are not identical; but they are not entirely distinct either.