Studies in the Epic Technique of Oral Verse-Making: II. The Homeric Language as the Language of an Oral Poetry

[This article was originally published in 1932 in Harvard Studies in Classical Philology 43:1-50. The original pagination of the printed version is embedded within brackets in this electronic version: for example, {1|2} marks where p. 1 stops and p. 2 begins.]

i. The Homeric Language and the Homeric Diction: Older Theories of the Homeric Language (p. i); the Homeric Language as a Poetic Language (p. 4); the Homeric Language as an Oral Poetic Language (p. 5).— 2. The Traditional Poetic Language of Oral Poetry: the Formula (p. 6); the Archaic Element (p. 9); the Art of Traditional Poetry (p. 12); the Foreign Element (p. 17); the Artificial Element (p. 19). — 3. The Study of a Traditional Poetic Language (p. 20). — 4. The Homeric Language as a Traditional and Oral Poetic Language: the Ionic Recording (p. 23); Arcado-Cyprian, Aeolic, and Ionic (p. 25); the Arcado-Cyprian Element (p. 26); the Aeolic Element (p. 27); the Traditional Language of Lesbian Lyric Poetry (p. 29); the Artificial Element (p. 33); Equivalent Aeolic Forms (p. 35); the Ionic Element (p. 37).— 5. Conclusions: the Theory of an Aeolic Homer Rejected (p. 40) ; the Theory of an Aeolic Diction Accepted (p. 43) ; the History of the Greek Heroic Style (p. 46).

Within the last twenty years Homeric scholars have shown that the language of the Iliad and Odyssey is a poetic language made to suit the needs of the verse, and they have thereby done away with a whole number of hypotheses which were no longer needed. A brief account of these earlier theories of Homer’s language will serve to set forth the subject of the present pages. The reader should bear in mind that we are speaking here at the beginning about language, and not about diction or style. All three have to do with the sum of words, word-forms, and word-groupings used by a man. As language, however, we look at them as used by a certain people, at a certain time, and in a certain place; as diction, as the material by which thought is expressed; and as style, as the form of thought.

The common view of Homer’s language in antiquity was that which, while it seems the simplest, is likewise the farthest from the truth: {1|2} Homer himself chose various forms and words from the dialects which he had heard in his travels about Greece. [1] Such a view could of course be held only in the lack of any careful study of the Greek dialects and of Homer’s language, for we now know that some of the forms in the Iliad and Odyssey are much older than others, while some could never have been a part of the everyday speech of any Greek. Also Homer’s use of the forms and words of different dialects follows a fixed rule and no varying chance of memory. [2] The gravest fault of such a theory, however, is that it supposes that one man could all by himself create a poetic language. Such a thing has been seen nowhere. No single poet could ever have such powers; and a poetic language, it is clear, is poetic only by a convention shared by the poet and his hearers, so that the growth of a poetic language must be gradual.

The ancients, since they had no rigorous historical method of literary criticism, may be excused for such a mistake, but not so the authors of a recent theory who hold that Homer was the native of a city wherein a mixed population of Aeolians and lonians had come to speak a language having the same variety of forms as that found in the Homeric poems. [3] Such a view altogether overlooks the nature of Greek poetic diction as it is to be seen everywhere in the poetry of the historical period; by the same reasoning the population of Attica was partly Dorian. Nor has anyone anywhere found a spoken language which shows even faintly such a variety of forms current side by side.

A third theory, which found many believers at the end of the last {2|3} century, held that Homer’s language was altogether Ionic and that the variety of forms was due to a simple literary conservatism which kept the older forms from age to age for purely stylistic reasons. [4] This view, like the next which will be mentioned, is much better than the first two, since it has a part of the truth in it. Its authors, however, had insufficient linguistic knowledge when they held that all of the Aeolisms in Homer had at one time been used in earlier Ionic, since many of the forms in question are the creation of a later period than that of common Greek. Yet their greatest faults were those of giving no {3|4} good reason for so strong a conservatism and of failing to see that the different forms are used under fixed conditions.

The last of the earlier theories is one of the oldest of all, since it was already held by Zopyrus and Dicaearchus at the end of the fourth century B.C.: τὴν δὲ ποίησιν ἀναγιγνώσκεσθαι ἀξιοι̑ Ζώπυρος ὁ Μάγνης Αἰολίδι διαλέκτωι· τὸ δ᾽ αὐτὸ καὶ Δικαίαρχος. [5] This hypothesis as developed by A. Fick won both more favor and more scorn than it deserved. [6] The favor it should not have had because Fick in putting it into practice used a method far too arbitrary, and those who scorn it now do not see that it first brought into prominence the two facts on which the whole problem of Homeric language hinges, namely that omer s poetry can with no very great change be turned from Ionic into Aeolic, and that the non-Ionic forms are kept as a rule only when Ionic itself has no forms which could take their place. K. Witte, when he wrote that the Homeric language is the work of the Homeric verse, gave the better reason for this, but it was Fick nevertheless who made the needed if false step, and we shall see what a large amount of truth there was after all in his views.

Witte was able to show long lists of words from the Iliad and Odyssey in which Aeolic or older forms stood beside Ionic forms, always with a difference of metrical value, and he was further able to show that these different forms were suited for use in different places in the verse. [7] As we shall see, [8] he failed to consider Arcado-Cyprian, and had no notion of an Aeolic poetic language, and so was wrong in thinking that some of these Ionic forms were only Ionic and so could not have been used by Aeolic poets. But this misunderstanding in no way affects the soundness of the principle which he drew from his evidence: {4|5} the Homeric poems were composed in a poetic language wherein old and foreign forms had been kept and new forms brought in by reason of the help they gave the epic poets in making their hexameters. These poets ever sought a language which was easier to handle, and for that reason ever made use of the fact that the older or foreign form of a word was to a Greek, as Aristotle tells us, more poetic than the form used in everyday speech. [9]

In one way, however, the theory of Witte, even with the further work done on it by Meister, is unfinished: they have logically proved that the language of Homer is the work of the Homeric verse, but they have not at all shown how the verse in this case could have such power. It did not have it in the later Greek epic, nor in Roman hexameter verse, nor in short do we find elsewhere in ancient or modern literature (with the very notable exception, however, of the early poetry of the nations) any but the slightest traces of the verse-form acting on the language of the poetry. Clearly a special language for the hexameter could come into being only when poetry was of a very different sort from that which we ourselves write, and which we know to have been written throughout the history of European literature. To say that the Homeric language was the work of the Homeric verse thus implies a poetry which is, at least to our way of thinking, of a very special kind, so that while the theory may be proved it cannot really be understood until we know just what this poetry was.

It is my own view, as those who have read my studies on Homeric style know, that the nature of Homeric poetry can be grasped only when one has seen that it is composed in a diction which is oral, and so formulaic, and so traditional. [10] So it is for the language of the Iliad and Odyssey: if we know what an oral diction is we shall have the larger {5|6} background which the theory of a language made to fit the hexameter calls for. At the same time the Homeric language when thus explained by the diction will in turn give us the history of that diction.

In a society where there is no reading and writing, the poet, as we know from the study of such peoples in our own time, always makes his verse out of formulas. He can do it in no other way. Not having the device of pen and paper which, as he composed, would hold his partly formed thought in safe-keeping while his unhampered mind ranged where it would after other ideas and other words, he makes his verses by choosing from a vast number of fixed phrases which he has heard in the poems of other poets. [11] Each one of these phrases does this: it expresses a given idea in words which fit into a given length of the verse. Each one of these fixed phrases, or formulas, is an extraordinary creation in itself. [12] It gives the words which are best suited for the expression of the idea, and is made up of just those parts of speech which, in the place which it is to fill in the verse, will accord with the formulas which go before and after to make the sentence and the verse. Each formula is thus made in view of the other formulas with which it is to be joined; and the formulas taken all together make up a diction which is the material for a ompletely unified technique of verse-making. [13] Finally, the formulas of an oral poetry are not each one of them without {6|7} likeness to any other; in that case the technique would be far too unwieldy. They fall into smaller groups of phrases which have between them a likeness of idea and words, and these in turn fall into groups which have a larger pattern in common, until the whole diction is schematized in such a way that the poet, habituated to the scheme, hits without effort, as he composes, upon the type of formula and the particular formula which, at any point in his poem, he needs to carry on his verse and his sentence. [14]

A single man or even a whole group of men who set out in the most careful way could not make even a beginning at such an oral diction. It must be the work of many poets over many generations. When one singer (for such is the name these oral poets most often give themselves) [15] has hit upon a phrase which is pleasing and easily used, other singers will hear it, and then, when faced at the same point in the line with the need of expressing the same idea, they will recall it and use it. If the phrase is so good poetically and so useful metrically that it becomes in time the one best way to express a certain idea in a given length of the verse, and as such is passed on from one generation of poets to another, it has won a place for itself in the oral diction as a formula. But if it does not suit in every way, or if a better way of fitting the idea to the verse and the sentence is found, it is straightway forgotten, or lives only for a short time, since with each new poet and {7|8} with each new generation of poets it must undergo the twofold test of being found pleasing and useful. In time the needed number of such phrases is made up: each idea to be expressed in the poetry has its formula for each metrical need, and the poet, who would not think of trying to express ideas outside the traditional field of thought of the poetry, can make his verses easily by means of a diction which time has proved to be the best.

Actually, of course, this birth of a diction is beyond observation, and unless it can really be shown that a people reverting from written to oral poetry created anew a formulaic diction we must suppose that it took place in a very distant past, since the poetry of an unlettered race has as much claim to age as have any of its other institutions. But if the birth of a formulaic diction is only to be described theoretically, we can see in living oral poetries how such a diction is passed on from one age to another, and how it gradually changes.

The young poet learns from some older singer not simply the general style of the poetry, but the whole formulaic diction. This he does by hearing and remembering many poems, until the diction has become for him the habitual mode of poetic thought. [16] He knows no other {8|9} style, and he is ever kept from quitting the traditional diction and using phrases of his own make because he could not find any as pleasing or as useful as the old ones, and moreover, since he is composing by word of mouth, he must go on without stopping from one phrase to the next. Since his poetry has being only in the course of his singing, and is not fixed on paper where it can show itself to him verse by verse, he never thinks of it critically phrase by phrase, but only faces the problem of its style when he is actually under the stress of singing. Thus whatever change the single poet makes in the traditional diction is slight, perhaps the change of an old formula, or the making of a new one on the pattern of an old, or the fusing of old formulas, or a new way of putting them together. [17] An oral style is thus highly conservative; [18] yet the causes for change are there, and sooner or later must come into play.

These causes for change have nothing to do with any wish on the part of the single poet for what is new or striking in style. They exist above the poets, and are two : the never-ceasing change in all spoken language, and the association between peoples of a single language but of different dialects.

As the spoken language changes, the traditional diction of an oral poetry likewise changes so long as there is no need of giving up any of the formulas. For example, a change in the sound of a vowel or consonant which calls for no change in the metrical value of a word soon {9|10} makes its way into the poetic language: the singer naturally pronounces the word as he usually does, and there is not the least thing to keep him from doing so. But when a change in the form of a word must also change its metrical value it is far otherwise, for the poet, if he then wished to keep up with the spoken language, would have to put up with a phrase which was metrically false, or give it up altogether and make himself a new one. But neither of these two choices is at all pleasing. The rhythm must be kept fairly regular, [19] and the oral verse-making makes it very hard for him to find new words; it is even doubtful if with all the good will and time in the world he could do so in any great number of cases. Each formula, as it was said above, is the long-proven choice of a long line of singers, and it is not possible that a phrase which is useful in oral composition could be made in any other way than by a singer who, making his verses through his sense of the scheme of the formulaic diction, created, in the stress of the moment, a new phrase more or less like an older one. For otherwise the new phrase would not fit into the scheme of the diction, and since it could be used only with an effort it would not be used at all. [20] Finally the change in the spoken language would very likely be such that a phrase to express the same idea in words of the same metrical pattern would be out of the question. The new phrase must be shorter or longer, or begin or end differently. Then the formulas to which it would be joined must also be {10|11} changed, and so on. Thus by no wilful choice, but by the constraint of his technique of verse-making, the singer keeps the formula though its language has become archaic. [21]

As it happens, this archaic language does not at all displease him. His style is thus lifted above the commonplace of daily speech and made distant and wondrous. But though the old words and forms are thus desirable, they are never wilfully sought after. When the formula can be changed it sooner or later will be, and the cleavage between the old and the new in the style depends on whether it is easy or hard to change the formula. [22] An oral diction may thus in time become very archaic, since even though a word has been lost altogether from the spoken language its context in the poetry will teach the poet and his public its meaning. In the case of words which are not a needed part of the thought, such as the ornamental epithets, the meaning of the word may even be lost altogether. [23] In time, however, a point must be reached in the case of each formula where its meaning, needed for the thought, is lost, and here an even heavier constraint than ease of {11|12} verse-making comes into play: the formula must be given up cost what it may, and the singers must do the best they can to find another one to take its place. Thus the language of oral poetry changes as a whole neither faster nor slower than the spoken language, but in its parts it changes readily when no loss of formulas is called for, belatedly when there must be such a loss, so that the traditional diction has in it words and forms of everyday use side by side with others that belong to earlier stages of the language. [24] The number of new words and old words varies, of course, from one oral poetry to another as different factors have force: a complex verse-form, a fondness for tales of an heroic past rather than of the present, and the practice of poetry by a class of professional singers all tend toward a greater conservatism, whereas a short verse without enjambement, a change in the way of living of a people, and the lack of a class whose gain it is to keep the best poetry of the past all allow a quicker change. But the principle of change and conservatism of language is the same in all cases.

I have written so far, in telling of how the language of oral poetry comes to be archaic, as if the formula were the unit of diction, and such it is in the end. But in practice the oral poet by no means limits his borrowing to the single formula; rather he uses whole passages which he has heard. This is, indeed, his whole art: to make a poem like the poems he has heard. [25] I know only too well that this is sure to suggest {12|13} the thought of plagiarism to those not familiar with oral poetry, but it must be understood above all that plagiarism is not possible in traditional literature. One oral poet is better than another not because he has by himself found a more striking way of expressing his own thought but because he has been better able to make use of the tradition. He strives not to create a new ideal in poetry but to achieve that which everyone knows to be the best. This is true even of the poetry which may tell of happenings of the singer’s own day: the event may be new, but it will be told in the traditional way on the pattern of passages from other poems, and in more or less the same phrases as were used in those passages, so that the only difference between the poem made about the present and that which tells of the past is that the former will be made from the memory of a larger number of different poems. [26] For if the tale {13|14} is old, and, as is usually the case, regarded as more or less true, the singer may tell it just about as he heard it.

Yet no graver mistake could be made than to think the art of the singer calls only for memory. Those who have sought to record oral poetry in lands where it still lives have straightway found that the same poem, that is to say, a poem on the same subject, could be sung badly or well, and that the people carefully set apart the poor singers from the good. [27] Still the fame of such a singer comes not from quitting the tradition but from putting it to the best use. The poorer singer will repeat a poem with the loss of its most pleasing lines or its most dramatic moments, but the good singer will keep what is striking, and even add, on the pattern of other poems, lines which he knows will please, and new incidents, or give a fuller tale with many such borrowings. He may even have heard the same tale told by a singer living at a distance who inherited from a different tradition; then he will fuse the poems, using the best in each. Thus the highest sort of oral verse-making achieves the new by the best and most varied and perhaps the fullest use of the old. This is the meaning of what Telemachus says: {14|15}

It is the same in all thriving oral poetries. The good singer wins his fame by his ease and versatility in handling a tradition which he knows more thoroughly than anyone else and of which his talent shows him the highest use, but his poetry remains throughout the sum of longer and shorter passages which he has heard. [29]

The formula thus is by no means the unit of the singer’s poetry, but it nevertheless ever tends to become so, for no singer ever tells the same tale twice in the same words. His poem will always follow the same general pattern, but this verse or that will be left out, or replaced by another verse or part of a verse, and he will leave out and add whole passages as the time and the mood of his hearers calls for a fuller or a briefer telling of a tale or of a given part of a tale. Thus the oral poem even in the mouth of the same singer is ever in a state of change; and it is the same when his poetry is sung by others. [30] His great name and the {15|16} fame of his verse may urge those who have learned from him to a more careful and more faithful use of memory than that which they would show for the poetry of a lesser singer. But the memory of the hearer depends after all on his being habituated to the diction as a whole, rather than on the learning of the poem word by word, so that he too must change and add and leave out verses and parts of verses, and this process will go on until all that is left of the poem are its single formulas and shorter passages which are the final units in the traditional diction. It should be added here that an oral poetry practiced by guilds of singers with masters and apprentices would tend to a more faithful {16|17} keeping of poems which had won fame, and that one singer might win such a name that his disciples would find their profit in keeping his poetry as nearly without change as they could; but then they are no longer singers but rhapsodes, their task is not that of creation but only of memory, and they are merely keeping from age to age the verse which was first composed by a singer who made his poetry, in the way that we have seen, by an ever varying use of what he had sung and heard others sing.

When poems thus pass from one singer to another in the same region the language of the poetry undergoes no change other than that which time may set upon it. But when the poets of one locality hear the poetry of a singer who speaks another dialect of their language their own traditional poetic language may undergo a much more rapid change. One must suppose that the two dialects are enough alike for their speakers to understand each other fairly well, and that the poems from abroad are such as to please. The fame of some singer may have spread until other singers came from afar to hear him; or the way of life in one region may have brought about a great liking for poetry, so that it was practiced more intensely and carried to a higher point; or the singers may have made their living by carrying their songs abroad. [31] In some way, then, the foreign poems are heard by the local singers and repeated more or less as they have been heard, and just as they have brought into their poetic language new words and forms of their {17|18} spoken language, so do they make the foreign poetry fit their spoken language in so far as they can do so without any too great loss. The new poems thus take on straightway a local color, but they keep those foreign forms which cannot be changed without harm to the verse, as well as words whose meaning may be known only from the context or which may be meaningless. [32] In time these poems, by the unending process of change which has been told of above, become fused with the local poetry, yet even when they have been lost as poems they leave their mark upon the poetic language. Coming from a tradition which has developed separately, the foreign poems have in them many pleasing {18|19} and useful formulas which are kept even after the poems which brought them to the new land have been lost, and in these formulas live the forms and words of the foreign dialect. Then on the pattern of these formulas others will be made which, while they are foreign in their language, are nevertheless native work. Even as the poets kept archaic formulas and made new formulas with archaisms, so do they keep foreign formulas and make formulas with foreign forms, so that a foreign word or form may show that a passage in a given poem was made abroad, or may prove no more than a contact at one time between the poetry of two regions. The poet and his hearers, it should be noted, in no way think of these words and forms as the words of a certain locality: like the archaic elements, they simply serve to carry the style above the commonplace of everyday speech. [33]

Finally, to the archaic, the new, and the foreign elements must be added a fourth and last: the artificial. Since the language of an oral poet is already far removed from daily speech, his public will not wonder at him if he should use a form which has never really been used anywhere. A whole new word no poet could make, since no one could understand him if he did, but he may make a form like another. That is to say, he may make the artificial by analogy with the real. The reason for such a creation is of course the same which leads the singers to keep the old and foreign forms, namely the need of a formula of a certain length which can be gotten only by this means. One poet, driven by this need, and making his verses under the sense of analogy which binds together the whole technique of his diction, will hit upon such a phrase, another will take it up, and it too will win its own place in the traditional poetic language. [34] Another kind of artificial form is due to the only partial adaptation of old or foreign forms. In certain cases the meter will allow part of a word to be modernized but force the singers to keep the rest of the word unchanged. [35] {19|20}

Such is the making of the language of an oral poetry. That the omeric poems were oral is shown by their diction, which, being formulaic, can only be traditional and oral. Putting the two sets of acts together, we see that the variety of words and forms which so ong puzzled Homeric scholars is the natural and necessary condition of the Homeric diction. Being oral it must be traditional, and being traditional it must have in it old words and forms, and it could be without foreign words and forms only if the people among whom it was developed had been cut off from the rest of Hellas. Until very lately scholars have started with the study of the forms and words in in the Iliad and Odyssey, and have sought in them an answer which could not be true because, though they were not aware of it, they were basing their search on their belief that they were written in just the same way they themselves would write poetry. But the sounder way, it is now clear, is first to learn the oral nature of the Homeric poems — this is shown us by the diction quite apart from the language – and then to turn to other poetries of the same kind, where we learn, besides many other things, that the language of Greek heroic poetry could not have been any other than it is.

Even more, the knowledge that the Homeric diction is traditional gives us the method we need for the study of the Homeric language, and shows us what we may hope to learn from it.

I. – The spoken dialect of the author of an oral poem is shown by his poetic language, which will tend to be the same as his spoken language where he has no metrical reason to use an older or foreign word or form or construction. Many scholars, when forced to show why the language of an oral poem follows a given dialect wherever the meter will allow, have supposed the text to be due to a scribe who, in copying a poem from another dialect, changed it to his own language where could. To the bookish mind such a process may seem quite natural; yet it is hard to see why a scribe should have wanted to do such a thing. {20|21} If he was merely copying the poem for other readers, why should he think that they would find it harder to read the original than he found it? If it was in order that the poem might be read to local hearers, why should he change only single forms or small groups of words? The foreign words and forms where the metrical value was different from that of his own dialect would be quite as puzzling to his hearers, and since he had writing materials to aid him at his task he should not find it much harder to change the language of the poem from beginning to end. When one does away with the factor of oral verse-making and its ever-present check on change there is nothing left to show why the change of language in a poem should be only partial. Further, the change of dialect which one thus finds in an oral poetry is so regular in its smallest points that one would have to suppose that such a scribe kept card indexes. But all such theories, after one has grasped the notion of traditional oral poetry, are seen to be forced. [36]

II. — On the other hand an oral poet, composing in a diction which follows his own language where it can, may be using phrases and passages which are neither his own work nor that of other poets of the same dialect, whether of his own or of an earlier time, but borrowings from the poetry of another dialect. Thus one cannot say that a given phrase or passage in a poem is the work of the author, or of another poet of the author’s linguistic group, just because it has forms of the author’s language. Such a phrase or passage may have been taken from another dialect and changed only where the forms had the same metrical value. Thus the proof that a given phrase or passage is the original work of the author’s dialect calls for the same demonstration by metrical value as the proof that it is the work of a foreign dialect. [37]

III. — A given word, form, or group of words can be proved to be the original work of poets speaking a given dialect only when it can be shown that no other dialect which had had a part in the history of the poetry had, in either its spoken or its poetic language, the same word or form or group of words with the same metrical value. That the poetic as well as the spoken language of another dialect must be taken into account is a fact which is usually overlooked. But it is clear that the poetic language of one region is as subject to change under the poetry of its {21|22} neighbors as is that of another region, and just as likely to show the same variety of forms. Thus the form which may seem to be the work of the author’s dialect may instead be taken from abroad, where it was a foreign form taken from still a third dialect. [38]

IV. – Conversely, a word or form or group of words which is metrically false, or fails to make sense, must be the work of a dialect whose words and forms when used would make the verse correct, or give it meaning. In making use of this principle, however, the critic must be quite sure he has to do in a given case with a word or phrase which is really metrically false and meaningless. [39]

V (exception to I). – A foreign or older form may be kept in the poetic language even when the poet’s own language has a form which could take its place, but such a keeping, apart from metrical reasons, will be due to the regular use of the form along with other words which are always used as a group and which the poet feels as such, or to the poetic character of the word, or to some other such special reason. This is most apt to be so when the words or forms used with it are themselves foreign or especially poetic, thus making a larger word-group which the singer feels as a whole, so that he changes none of its parts. [40]

VI (exception to IV). – The working of a formulaic diction may itself be the cause of metrical faults. These will be of two sorts: those which are due to the joining of formulas which do not fit, and those {22|23} which come from changing a correct formula to fit the needs of a sentence. [41]

VII (exception to III). — A form which seems old or foreign may be a creation by analogy from forms which are really so. The form, however, still stands to show that the poetry was at one time influenced by another dialect or that the tradition of the language is old.

VIII. — A word, form, or group of words which is old or foreign is not in itself proof that the verse or passage in which it is found is the work of an older or foreign singer. One must ever be ready to admit that a given poem may be made by putting together anything from single traditional words or phrases to whole traditional passages. [42]

Any attempt to localize the traits of Homeric language must be largely balked by the conditions of the search: the lateness of the inscriptions, their small number which allows us to know only a part of the words and forms of any one dialect, and our complete, or almost complete, lack of them for many regions. The evidence quoted by the ancient grammarians is simply by itself untrustworthy because they had no sound linguistic or textual method, and so must be left aside unless it happens to agree with the evidence of the inscriptions. The manuscripts of the Ionic prose writers likewise give us little help, since they have suffered from the ignorance, and even more from the mistaken linguistic notions, of their editors and copiers. The remains of Ionic and Aeolic verse are more helpful, though they too have suffered {23|24} from copiers and mistaken theories of language, and their evidence, as we shall see, bears on the poetic rather than on the spoken language. Thus the study of Homeric language must be based above all on the inscriptions.

The language of the Iliad and Odyssey has at least one very common sound and one very common form which, as the prose inscriptions show us, were used only in Attic and Ionic speech, but which might have been replaced without harm to the meter by the sound and form of the other dialects: η for original ᾱ, and the third singular of the imperfect ἦν. Original ᾱ is found in inscriptions of all the other dialects, and ἦς is found where that form of the of the verb occurs outside of Attic and Ionic, namly in West Greek, Boeotian, Lesbian, Arcadian and Cyprian. [44] Neither η for ᾱ nor ἦν could be the archaism of another dialect. [45] These traits of Attic-Ionic, though only two, play such a part in the language that they are more than enough to show, in view of what was said above, [46] that the singer (or singers), or rhapsodes, who composed, or gave final form to, the Iliad and Odyssey, spoke either Ionic or Attic. These traits of language do not, however, necessarily show whether the recording was due to Ionic singers, whose verse-making was a constant creation, or to Ionic rhapsodes, who were mere reciters, although all the evidence of contemporary oral poetry which I know points to the singer, and none to the rhapsode. Nor do they show in what measure the diction of the poem—words, phrases, verses, or passages—was the original creation of Ionic rather than foreign singers.

That the spoken language in question was Ionic and not Attic is shown by the following sounds and forms: η for original ᾱ even after ε, ι, ρ, where Attic would have – ᾱ; -σσ-, ἤν (εἰ + ἄν), ἤνεικα, ἱστίη for Attic –ττ-, ἄ̄ν, ἤνεγκα, ἑστία. Of these Ionic traits at least three—η {24|25} after ε, ι, ρ, -σσ-, and ἤν— could not be Attic archaisms. [47] The various Ionic dialects are not well enough known to allow a more exact localization. [48]

The various elements of the Homeric language are drawn from three dialects — Ionic, Aeolic, and Arcado-Cyprian. As we have just seen, the language was last affected by Ionic, and we have the following evidence to show us that the Arcado-Cyprian element was not brought into the language by any direct contact between Arcado-Cyprian and Ionic poetry, but came in along with the Aeolic element. The Homeric poems have Aeolic αἰ for Ionic εἰ when the next word is κε, αἴ κε being felt more or less as one word because of the foreign κε. The Arcado-Cyprian form of the phrase, however, has εἰ (Cyprian ἔ̄ κε, Arcadian εἰκ ἄν), and had this phrase ever been known to the Ionic poets it must, because of the greater likeness to the Ionic form, have straightway taken the place of αἴ κε. [49] Likewise, the poems have Aeolic ἄμμες, ὔμμες, etc., whereas the knowledge of the Arcado-Cyprian forms ἁ̄μές, ὑ̄μές, etc., would have brought about their use because of their greater likeness to Ionic ἡμει̑ς, ὑ̄μει̑ς, etc., ἁ̄μές doubtless becoming *ἡμές. Δόμεναι and φήρ, where Arcado-Cyprian has δοϝέναι and θήρ, point the same way; otherwise one must be willing to grant the bare possibility that δόμεναι might have taken the place δοϝέναι of by analogy with other non-thematic infinitives in -μεναι, and that the legend of the Beasts (cf. below, p. 37) might not have been known in Arcado-Cyprian {25|26} poetry. The history of the Homeric language is thus seen to follow the order Arcado-Cyprian, Aeolic, Ionic, whatever may have been the influence back from Aeolic to Arcado-Cyprian and from Ionic to Aeolic.

The following Homeric forms are found, as far as the evidence of the inscriptions goes, neither in Ionic nor in Aeolic, but in Arcado-Cyprian; if they are found elsewhere it is in West Greek: [50] the infinitive of contract verbs in –η̑ναι (Arcadian κατυφρονη̑ναι, Cyprian κυμερη̑ναι, Homeric φορη̑ναι, ἀη̑ναι, and also, therefore, βιω̑ναι); the declension of ηϝ-stems in –ης, -ην (Cyprian ἰjερής, γραφής, Arcadian hιερήν, Homeric ζαη̑ν, Ἄρην, Μέγην); the suffix -τερος in the sense of one of a pair of things (Arcadian τὠρρέντερον γένος, Homeric θηλύτεραι, θεώτεραι, ἀγρότερος, etc.); Arcadian δω̑μα, ἀέλιος, ἔσχεθον, ἄματα; Arcadian and Cyprian πτόλις (Homeric πτόλις, πτόλεμος, cf. Eustathius ὁ δὲ πτόλεμος Κυπρίων). [51]

The following Homeric words are found in neither Ionic nor Aeolic inscriptions, but in Arcado-Cyprian; they occur in the Greek literature we know only as poetic words:

In Arcadian and Cyprian αἶσα, ἕρπω, εὐχωλά, οἶϝος.

In Cyprian ϝάναξ, ἀνώγω, ἀρτύω, αὐτάρ, ἔλος, ἰδέ, πόσις, σπέος, χραύω.

In Arcadian ἀπύω, ἀσκηθής, δέαμαι, κέλευθος, λεύσσω. [52]

If we consider how small a part of Arcado-Cyprian vocabulary it is which we know from the inscriptions, this number of poetic Homer {26|27} words which we find in current usage in Arcadia and Cyprus is highly significant. It can be understood only by assuming that the Homeric diction comes, for a large part at least, from Arcado-Cyprian poetry. The most stable part of an oral diction is its vocabulary, since it is usually easier for a singer to change a form on the model of his own language than it is for him to give up one word and find another, and his art of verse-making is chiefly the art of using the traditional poetic words. The Aeolic element in the Homeric language seems indeed to have been more one of morphology than of vocabulary.

The belief was held at one time that the Aeolisms in Homer were really only older forms of Ionic, but this was due to a misunderstanding of the nature of linguistic change, since most of the forms in question are due to two separate treatments of one original form. Thus the dative in -εσσι was formed on the analogy of the εσ-stems (γένεσ-σι, βέλεσ-σι); the first aorist in -σσ- on the analogy of σ-stems (ἔθλασ-σα, ἐτέλεσ-σα); the perfect active participle in -ων, -οντα, was formed after the present participle; πίσυρες for Ionic τέσσαρες, and Φήρ for Ionic θήρ, show different treatments of *qṷ and *ĝhṷ; ἴα for Ionic μία shows a complete absence of the initial *sm; ἀργεννός and ἄμμες; show different treatments of *-σν- and *-σμ- from those which gave Ionic φαεινός and ἡμει̑ς. Ὄππως, ὄπποι, etc. is an innovation of Lesbian, seemingly made after ὄττι (original *ὅδτι). In view of the number of these certain Aeolisms it is clearly better to take also as Aeolisms those forms which might be earlier forms of Ionic, e.g. the genitives in -ᾱο and -ᾱ́ων and in -οιο. Indeed the number of Homeric forms which are not Ionic but are found in other dialects is such that it seems to outweigh that of the archaic and artificial forms.

The following traits of the Homeric language are Aeolic:

(1) In Lesbian (Aeolic of Asia Minor), Thessalian, and Boeotian, the dative in -εσσι; θερσ- instead of θαρσ- (Homeric Θερσίτης, Θερσίλοχος, cf. Thessalian Θερσίτας, Θερσιλόχειος, etc.); ἴα instead of Ionic μία; the patronymic adjective instead of the genitive of the fathers name (Homeric Τελαμώνιος, Νηλήιος, etc.); the treatment of labiovelars as labials even before front vowels (Lesbian πέσσυρες according to Hesychius, cf. Balbilla πέσυρα, Sappho πήλυι, Boeotian πέτταρες, {27|28} Thessalian πέμπε, πεφειράκοντες, Homeric πίσυρες, πεμπώβολα, περιπλόμενος, ἔπλε, ἔπλετο, πέλομαι, etc., πέλωρ, πελώριος, Φήρ, and the variant reading φλίψεται for Θλίψεται in p. 221); declension of ηϝ-stems in –η̑ος, -η̑ι (Lesbian βασίληος, Thessalian βασιλει̑ος, Boeotian βασιλει̑ι); ἄμμες, ὔμμες, etc. (Lesbian and Thessalian, no evidence for Boeotian); the apocopated forms of the prepositions; the dual (Thessalian δείμενε, ἀδ[?]άτοιν, Boeotian ἐποιεστάταν, ἀνεθέταν, Lesbian [ἄν]δρε, τὼ ἐτιστάτα). [53]

(2) In Lesbian and Thessalian the development of σ followed by a liquid or nasal into double liquids or double nasals respectively (Lesbian ἔμμεναι, Ζοννύσω, Ἄλληκτος, Thessalian ἔμμεν, Διοννύσοι, Homeric ἔμμεναι, ἔμμεν, φιλομμείδης, ἔμμορε, ἐρεβεννός, ἀργεννός, ἀγάννιφος, ἔννεον, ἄλληκτος, ἔλλαβε, ἔρρεον, ἐύρροος, καταρρέω); the change of *τι̯ and *θι̯ into σσ (Lesbian and Thessalian ὄσσος, etc., Lesbian μέσσος, Homeric ὄσσος, μέσσος, etc.); κάλλος instead of Ionic καλός in compounds, and in the comparative and superlative (Lesbian Καλλίκληι, etc., Thessalian Καλλιφρούντειος, etc., cf. Boeotian καλϝός, Homeric Καλλιάνασσα, κάλλιον, καλλιγύναικα, etc.); κε instead of ἄν.

(3) In Lesbian ὄττως, ὄπποι, ὄττι, etc.; the infinitive of non-thematic forms in –μεναι (Lesbian ἔμμναι, ἔδμεναι, θέμεναι, δόμεναι, Homeric ἔμμεναι, ἔδμεναι, θέμεναι, δόμεναι, etc.); ἀμβρότην (Homeric ἤμβροτον, βροτός, ἀμβρόσιος); ζα- from δια– (cf. Lesbian Ζοννύσω, Sappho ζὰ δ᾽ ἐλεξάμαν, Homeric ζάθεος, ζατρεφής, etc.).

(4) In Lesian and Boeotian the aorist in –σσ-.

(5) In Thessalian and Boeotian the thematic and non-thematic infinitive in –μεν (Thessalian ἔμμεν, θέμεν, δόμεν, Boeotian δόμεν, etc., Thessalian κρεννέμεν, πρασσέμεν, Boeotian φερέμεν, etc., Homeric ἔμμεν, θέμεν, δόμεν, etc., ἀγέμεν, φερέμεν, etc., and the variant readings πολεμιζέμεν II 834, ἀκουέμεν T 79, etc.); the genitive in –οιο (Thessalian of Pelasgiotis and Perrhaebia Παυσουνείοιο, πολέμοιο, etc.).

(6) In Boeotian the genitive in -ᾱο (Boeotian Ἀριστέαο, etc.); the genitive in –ᾱ́ων (Boeotian δραχμάων, etc., cf. Thessalian κοινάουν, etc.); τοί, ταί instead of οἱ, αἱ. [54] {28|29}

Of all these Aeolic traits the only ones which are found in other dialects are the following: in Cyprian κε; in Arcado-Cyprian the declension in –η̑ϝος, -η̑ϝι, etc.; in West Greek the non-thematic infinitive in -μεν, μέσσος, etc. τοί, ταί; and in Arcadian and West Greek apocope in the prepositions. Ἀπ, ἐπ, and ὑπ, however, are Thessalian only. [55] The only cases of apocope in Ionic inscriptions are two occurrences of παρ. The Homeric words which have Aeolic (and original) ᾱ where one would look for Ionic η are dealt with below (p. 36). The traces of the digamma in Homer are likewise Aeolic, and allow us to say from which of the three Aeolic groups the poetry passed to the Ionians, but before dealing with this sound we must understand the nature of the Lesbian poetic language.

The same forces which created the poetic epic language of Homer created the poetic lyric language of Sappho and Alcaeus. The scant remains of these two poets do not allow us to show, as we can do for Homer, that their diction is formulaic, and so oral and traditional. We do know, however, that Solon and Theognis were still following an oral tradition of iambic poetry, [56] and that they lived at that time, always so precious for our own knowledge of oral poetries of the past and present, when verse-making was oral but writing known and used as a means of recording and keeping. [57] All that we know of the use of writing in Greece at the beginning of the sixth century points to the same thing for Sappho and Alcaeus. Yet while we may still feel some doubt as to the way in which they made their verses, there is not the least doubt that their poetic language was drawn from an oral tradition: {29|30} only in an oral poetry does one ever find such a variety of forms that nave each one its own metrical value. [58]

Thus Sappho and Alcaeus use the endings of the spoken language, that is of the incriptions -ω, – ᾱ, – ᾱν, -εσσι, -οισι, -οντες (perfect active participle), but they also use -οιο, – ᾱο, – ᾱ́ων, -σι, -οις, -ότες; although the forms with -σσ- were the current forms, they also have ὄσος, μέσος, ὐπίσω, and first aorists such as ἐσκέδασ᾽, ὤλεσαν; beside Lesbian πόλις, ἔδωκαν, ἔσχον, ἶρος, ἄλιος, δόμος, they have πτόλις, ἔδοσαν, ἔσκεθον, ἴερος, ἀέλιος, δω̑μα; beside παρ, which we always find in the inscriptions, they have the longer παρα, and where their speech ἀπυ used they shorten to ἀπ; they have the uncontracted νόος beside νω̑.

The foregoing forms are either archaic or found in other dialects than Lesbian; the following are artificial forms: ἄμμεσιν beside Aeolic ἄμμιν; the genitives ἔμεθεν, σέθεν, ἔθεν; Νηρήιδες beside Νηρέιδες; the transfer of endings from one declension or conjugation to another, as in πόληος, τεμένηος, ἔχηισθα; the use of ν–movable which, in verb-forms at least, was foreign to Lesbian; artificial forms such as ὠράνω, ποικιλόδειροι, ὤρεσι, πωλυανάκτιδα, ὠλομέναν, beside ὀράνω, δέραι, ὄρσει, πόλυ. In all these cases there is a difference in the metrical value of the forms; the language is the work of the verse.

It is my own view [59] that initial digamma had been altogether lost in Lesbian by the time of Alcaeus and Sappho. Where it still seems to be called for to prevent hiatus or make position, it is not the sound itself but its one-time presence, which is felt, much in the same way that the French feel the one-time presence of h-aspiré. It was likely that even in everyday speech certain word-groups that had had an initial digamma were long kept without elision, as the combination of the unelided article before h-aspiré keeps the traces of that sound in French (δέ οἱ may be such an instance), but the greater number of cases in Lesbian poetry, as in Homer, must have been due to the keeping of the poetic formula. The poets and their hearers, being used to these formulas, would feel no fault where there was hiatus or failure to make position, while on the other hand, if they were using newer phrases, they were free to treat the word as if it had never had the {30|31} sound. Thus we find in Sappho γλω̑σσα ἔαγε, φαίνεταί οἱ, [60] and in Alcaeus πνεύμονα οἴνωι, ὑπο ἔργον (beside ἀμύστιδος ἔργον), λυ̑σ᾽ ἄτερ ἔθεν ( – ⏑ – ⏑ ⏑ ); likewise the trace of the initial digamma is seen in Sappho’s ἔειπε, and in Alcaeus’ ἀ]πυείπη[ι and ἐάνασσε, where the meter kept the unshortened form beside εἶπον and ἤλ̣π̣[ in Sappho and εἶπε and ἤλπετο in Alcaeus. There are in the two poets some 33 places where an initial digamma would spoil the meter.

The keeping of ϝρ- in some form which is noted in our manuscripts as βρ- (Sappho βρόδον, βράδινος, βράκεα) — how it was sounded is doubtful — was a poetic device to keep for these words the power of lengthening the foregoing syllable. Had such a treatment ever been a part of the spoken language the poets would have used it when the second element of a word began originally with *ϝρ- and Herodian would have quoted from Lesbian poetry *ἐβράγη and *ἄβρηκτος (*ἐ-ϝράγη, *ἄ-ϝρηκτος), and not εὐράγη and αὔρηκτος, as he did. These two forms are beyond doubt, because they show the treatment which the spoken language gave to the group vowel-digamma-consonant-vowel, as in δεύω (*δέϝσω), ναυ̑ος (*να̑ϝσος), and Εὐρυσίλαος (*Ε-ϝρθσί-λαϝος) of the inscriptions. Nevertheless such forms as εὐράγη and αὔρηκτος could never have been a part of the spoken language, since -ϝρ- would then have been treated as an initial and not an intervocalic sound-group, and they would have become *ἐράγη and *ἄρηκτος. Εὐράγη and αὔρηκτος can only be understood as poetic forms made to keep the metrical value of *ἐϝράγη and *ἄϝρηκτος. The following forms in Homer show this Lesbian treatment of the digamma: δεύω beside Ionic δει̑; χεύω, ἔχευαν, etc., beside Ionic χέει, ἔχεαν, etc., ἀπούρας (*απόϝρας), ἀπηύρας (*ἄπεϝρας), ταλαύρινος (*ταλάϝρινος), καλαυ̑ροψ (*κάλαϝροψ). A like treatment was given in Aeolic to the group vowel-consonant-digamma-vowel. Εὔαδε (from *ἔσϝαδε) is cited by Choeroboscus as {31|32} Aeolic. Homer has εὔαδε beside ἅδε, αὐίαχοι (*ἀσϝίαχοι), [61] αὐερύοντα (*ἀνϝερύοντα). Of these words ἀπούρας, ἀπηύρας, ταλαύρινος, καλαυ̑ροψ, εὔαδε, αὐίαχοι, αὐερύοντα, could have come only from the poetry of the Aeolians of Asia Minor. They are the proof that the Aeolic in Homer was brought into the Ionic epic from the Lesbian epic language at a time when spoken Lesbian had lost the digamma. [62] Likewise poetic and Lesbian is the treatment of a short vowel followed by digamma when the short vowel is followed by two more short syllables. The glossographers quote as Lesbian καυάλεον, φαυόφοροι, etc.; the same treatment is seen in Homer’s ἀλευόμενοι, etc., beside ἀλέοντο, etc. To this same source belongs the artificial treatment of intervocalic -δϝ- as -δδ- in ἔδδεισας, ἔδδεισεν, etc. [63]

There is no sure case of a form borrowed by the Lesbian singers from some other Aeolic dialect: wherever the Aeolic form in Homer differs from that of the Lesbian inscriptions the form may be archaic Lesbian. Yet the readiness with which the Ionic singers took over the Aeolic forms would rather point to just such an exchange between the Aeolic groups before the migration to Asia Minor. The Lesbian lyric language offers the same difficulty of deciding between forms which are Thessalian or Boeotian but might also be archaic Lesbian; nor is the source of the non-Aeolic forms always altogether sure. Ἀέλιος, δω̑μα, ἔσκεθον, ἵερος, and πτόλις, however, are Arcado-Cyprian and not Ionic; moreover, the non-Aeolic endings of the dative –οις and -αις, which are the only ones found in Arcado-Cyprian inscriptions, are later in Ionic than -οισι, -ηισι, which were not altogether supplanted until well into the fourth century. Ν–movable, which is generally classed as an Ionic trait, is nevertheless found in the dative plural of consonant stems in Thessalian and in verb-forms in some Cyprian inscriptions. There is thus nothing to show that the foreign element of the Lesbian lyric language was not drawn wholly from Arcado-Cyprian. This is a point to remember when we seek for forms in Homer {32|33} which could only be Ionic: if the form is also Arcado-Cyprian, we must grant that it may have been a part of the Lesbian epic language.

There are in the Homeric language a number of artificial forms which can be understood only as Aeolisms which were changed by Ionic singers to forms nearer those of their spoken language, though they could not make them altogether Ionic. The change in each case was brought about by a purely oral process. Thus the Lesbian poems which the Ionic singers learned had in them a number of perfect active participles in -ων, -οντος. Where the forms of the Ionic participle had the same metrical value these were put in their place, save in a few words where the Aeolic ending and the meaning of the word led the Ionic singers to mistake the forms for presents, as in κεκλήγοντας, τετρίγοντας. [64] Usually, however, the Ionic singers, when the forms of the two dialects were metrically different, were drawn by their habit of using different endings for the present and perfect to the endings -ότος, -ότι, which the rhythm forced them to lengthen to –ω̑τος, -ω̑τι. Thus we find τεθνηω̑τα beside τεθνηότος, πεπτηω̑τες beside πεπτηότα, κεκμηω̑τα beside κεκμηότας, etc. This same struggle of the Ionic singer between the foreign form and the habit of his daily speech is likewise the source of ἐήνδανε. He had heard on the one hand the Aeolic poetic *ἐάνδανε, but all the usage of his speech tended to ἥνδανε, so that, speaking the two forms as it were at once, he made ἑήνδανε. This form, it should be noted, shows that the initial digamma had been lost in Ionic.

The so-called “distracted” forms were the work of singers who, torn between their desire to keep the metrical value of the genuine uncontracted forms on the one hand, and their habits of daily speech on the other, in which they used the contracted forms, made such artificial forms as μνωόμενοι, ὁρόω, ὁρόωντες, δρώωσι, etc. Thus the poet who had heard μνᾱόμενοι in verse, but said μνώμενοι in talk, would tend to begin the word with μνω-, whereupon the rhythm would force him to {33|34} keep without change the latter part of the poetic form and make μνωόμενοι. When faced with a verb of the Aeolic -ᾱ́ω conjugation, such as ἡβᾱ́ωσα, where he would usually say ἡβω̑σα, he would be drawn into using the first two syllables of the spoken word ἡβω-, and then when forced by the rhythm to supply two more syllables would use the ending of the spoken word -ωσα, thus using the ω twice and making ἡβώωσα. When faced with the poetic ὁράω beside the spoken ὁρω̑ the singer would be drawn to the spoken form, but the rhythm would force him to shorten the ω-syllable to ο, which would be felt only as a poetic sound to fill in the verse until the real last syllable could be given, making ὁρόω. The same thing happened in certain nouns: φάος, where the spoken form was φώς, became φόως; πρᾱ́ονες, where the spoken form was πρω̑νες, became πρώονες; σάος, where the spoken form was σω̑ς, became σόος; etc. Another way in which the uncontracted forms were brought nearer to the usual contracted forms was by lengthening from a short root-vowel and making τρωπα̑σθε from τροπάεσθε, πωτω̑ντο from ποτάοντο, etc. That this change was a purely oral process is shown by the fact that when the root had an α there was a tendency to keep the genuine uncontracted forms, as in ἀοιδιάει (the voice repeating in –άει the movement of ἀοι-), κραδάων, ναιετάουσι, etc.

A like oral creation of artificial forms is found in such forms as Ἰλῑ́ου, ἀγρῑ́ου, ὁμοιῑ́ου, ὅου, etc., found always before a double consonant, which can have come only from *Ἰλίοο, *ἀγρίοο, *ὁμοιίοο, *ὅο, etc. In another case the loss of the ending *-οο led to the making of the poetic form ὀκρυοέις in the phrases ἐπιδημίου ὀκρυόεντος (I 64) and κακομήχανου ὀκρυοέσσης (Z 344) for *ἐπιδημίοο κρυόεντος and *κακομηχάνοο κρυοέσσης. Van Leeuwen is doubtless right in thinking the poets were guided by the model of the adjective ὀκριόεις, [65] but of far more weight in each instance was the need of keeping the formula, and wherever the former presence of the *-οο ending is found there is a marked formulaic device. Thus the form Ἰλίου is found only in the phrase Ἰλίου προπάροιθε (3 times); ὅου is found only in a special type of clause, ὅου κλέος οὔ ποτ᾽ ὀλει̑ται (B 325, Hymn to Apollo 156), ὅου κράτος ἐστὶ μέγιστον (α 70). Ὁμοιίου, which is found only in the phrase ὀμοιίου πολέμοιο for *ὁμοιίοο πτολέμοιο, is used six times in the Iliad and twice in the Odyssey, and {34|35} here the Arcado-Cyprian πτόλεμος might show that the ending *-oo belonged to the Arcado-Cyprian poetic language. But we cannot be sure, since a Lesbian poet might have used the foreign πτόλεμος in a new phrase. [66] But whether *-oo is the original of Arcado-Cyprian -ω, or a middle stage between Aeolic -οιο and -ω, the creation of the artificial forms in question can have been due only to singers who had to keep the formulas.

It is likewise only the theory of oral verse-making which can show why certain Aeolic forms were kept when an Ionic form might have been used. In at least one case it was only because there was no such form in Ionic: θεᾱ́ was kept because Ionic used θεός for both masculine and feminine. But usually the Aeolic form was kept for less simple reasons, which could be only those of a poet who was drawn more towards the foreign form he found in poetry than towards the form which was habitual with him in his daily speech. An equivalent Aeolic form might be kept for one or more of the following reasons: the form was used along with other words, so that the poet felt the group as a unit and sought to change none of its parts; the form was known to him more from poetry than from speech, so that the habit of the poetic language was greater than that of his spoken language; the syntax of the form was foreign, and thus set it apart from the current form.

Aeolic αἰ for Ionic εἰ is regularly found in the phrases αἴ κε, αἰ γάρ, and αἴθε. In the phrase αἴ κε it has been kept because of the foreign κε with which it was felt more or less as a single word, like ἤν for εἰ ἄν. So soon as a single word is put between αἰ and κε, however, the motive for keeping the Aeolic form is lost and we have εἰ δέ κε, εἰ μέν κεν, εἰ γάρ κεν, etc. The use of αἰ γάρ, and its metrical variant αἴθε, is foreign to Ionic speech, being found only in Ionic and Attic prose in highly emotional passages where the author meaningly assumes the tone of poetry.

Ἐρι-, which seems to have come first from Arcado-Cyprian, [67] tended in the Ionic vocalization to become ἀρι-. The form ερι- was kept in {35|36} certain fixed phrases: e.g. ἐριαύχενες ἵπποι (5 times); ἐρίηρες ἑται̑ροι (22 times); ἐρίφδουπος πόσις Ἥρης (7 times); ἐρικυδέα δαι̑τα (5 times); ἐρικυδέα δω̑ρα (twice); ἐρικυδέα τέκνα (once). [68] It was these formulas that kept the prefix, but it was kept unchanged for another reason: where the radical of the word is a noun ἐρι- is used, while ἀρι- was brought in where the radical is a verb or adjective and thus properly calls for an adverbial prefix. Thus we find ἐριβω̑λαξ, ἐρίφδουπος, ἐρικυδής, ἐρίμυκος, ἐρισθενής, ἐριστάφυλος, ἐρίτιμος, ἐρίωπος, but ἀρίγνωτος, ἀριδείκετος, ἀρίζηλος, ἀριπρεπής, ἀριφραδής, ἀρισφαλής. In ἐριούνιος and ἐρίηρος the meaning of the radical is so vague that the prefix is scarce felt as such, and in ἐρίηρος it was the poetic θηλ-, for the prosaic θαλλ-, which kept the whole poetic word unchanged.

There is no need, if we would understand why λαός was not changed to *ληός whereas να̑ος became νηός, to argue that there was no word λεώς in Ionic. Indeed two names in the Iliad and Odyssey prove an Ionic λεώς: Λειώκριτος and Λειώδης are artificial forms which have been made from *Λαόκριτος and *Λαώδης (*Λαϝοϝάδης) after the spoken *Λεώκριτος and *Λεώδης. Ἀγέλεως in τοι̑ς δ᾽ Ἀγέλεως μετέειπε (χ 131, 247), where the Aeolic would be *τοι̑ς δ᾽ Ἀγέλαος ἔειπε, shows that there must have been many names of this type in early Ionic (cf. Ἀναξίλεο, Ἀρχέλεος, Θερσίλεω, λεώφορον, of the Ionic inscriptions), for the form ἔειπε was common enough in the Homeric poetic language and Ἀγέλαος is found in other verses where the change would not have been so simple. The foreign λαός was kept simply because it was more common in poetry than λεώς was in speech, so that the singers were more habituated to the poetic than to the prosaic word. Each time a singer met να̑ος in verse he would tend to modify it in the directions of the spoken νεώς. But λαός, which figured in so many poetic phrases—ποιμένα λαω̑ν (56 times), κοίρανε λαω̑ν (11 times), λαὸν Ἀχαιω̑ν (19 times), etc., in the epithet of gods λαόσσοος (6 times), in heroic names—and was moreover a word which by its meaning had a special dignity, won a place for itself far beyond the reach of λεώς. Where on the contrary a word had no such special place in the poetic language, but differed only in form from the common prose word, the singers would be drawn to the current form. Thus ἆος and τα̑ος became ἕως and τέως, contrary to the usual laws of {36|37} the heroic meter. [69] It is wrong, however, in the case either of these words or of the words which make hiatus through a lost digamma, or in such phrases as ἡμει̑ς δὲ δείσαντες, μάλα δήν, σμερδάλεα ἰάχοντες, to speak of metrical faults, for these irregularities of the rhythm are constant and accepted, and so are rightly to be felt as the correct rhythmic usage.

The Aeolic form φηρ- wherever it was used in the common sense of “hunting” or “game” appears as θηρ- in Ionic (θήρη, θήρ etc.), but when the word was used of the Centaurs the Ionic singers very naturally failed to translate the word into their own dialect, but left it to be a proper name, Φη̑ρες, ‘the Beasts,’ and so we find it in A 268, B 743.

Ἑρμείας was kept beside Ἑρμη̑ς, though Ἀθηναία became Ἀθηναίη, and Ῥεία became Ῥείη, because the non-Ionic ending –ειᾱς was well fixed in other heroic names such as Αἰνείας and Αὐτγείας. Ναυσικάα and Φεία were not changed because the names were not Ionic.

Πίσυρες kept its initial consonant because the whole word differed in form from τέσσαρες. Ἐρεβεννός and ἀργεννός were unchanged because there was no like word in -εινός in Ionic.

We may now turn to the forms of the Homeric language which are Ionic.

The greater number of sounds and forms which are Ionic and not Aeolic are metrically equal to the Aeolic sounds and forms. They can in no way be looked on as an Ionic addition to the traditional diction, but are the work of Ionic singers changing the Aeolic epic language to suit the habit of their own speech. Thus, save for the few unusual forms just noted, Homer has η where Aeolic kept original ᾱ; ευ from εο where Aeolic had ε͜ο; ἦν for Aeolic ἦς; -ου, -ους, – ᾱς, -ε͜ω, -εω͜ν, in {37|38} the noun where Aeolic has -ω, -οις, -αις, – ᾱ, – ᾱν; -ειν, -ουσι, -ουσα, in the verb where Aeolic has -ην, -ωσι, -οισα; φᾱσί, Moυ̑σα, ξει̑νος, εἰμί, ἕτερος, θάρσος, etc., where Aeolic has φαι̑σι, Μοι̑σα, ξέννος, ἔμμι, ἄτερος, θέρσος, etc. Each one of these sounds and forms is a sign of the thorough Ionization of the traditional epic diction.

A number of Homeric forms which are Ionic and not Aeolic are found in the Lesbian lyric language. Some of these are also found in Arcado-Cyprian, which seems to be their source. These forms are: the noun-endings -οις, -αις, -σι; the first aorist in -σ-; the perfect participle in -ώς, -ότες; μέσος, ὅσος, etc.; and ν-movable in verb-forms.

The non-thematic infinitive in -ναι is not found in the remains of Lesbian poetry, but was used in Arcado-Cyprian (Arcadian ἀπειθη̑ναι, ἦναι, ἐξη̑ναι). We are thus unable to say surely that these infinitives are an Ionic addition to Homeric language. Likewise we are unable to claim an Ionic source for ἄν in Homer. Cyprian has κε, but Arcadian has ἄν, and for εἰ followed by ἄν it has εἴ κ᾽ ἄν; in one case in Arcadian we find εἴ κ᾽ used without ἄν followed by the subjunctive. [70] The tendency in Aeolic would have been to change ἄν in Arcado-Cyprian formulas to κε(ν), as it was in Ionic to change Aeolic κε(ν) to ἄν. There are in the Iliad and the Odyssey some 43 out of 156 cases where ἄν cannot be changed to κε(ν), and some 747 cases of κε(ν), but even this small portion of sure cases of ἄν may have come over from the Lesbian epic.

The Lesbian lyric language also has the artificial (or Arcado-Cyprian?) ἄμμεσιν, which is metrically equal to ἡμι̑ν scanned -́ -, and ἄμμες, [71] which can usually take the place of the Ionic accusative ἡμέ͜ας. Εο in Lesbian poetry can be scanned as a single syllable, and is thus the equal of Ionic ευ (Lesbian βέλε͜ος scanned ⏑ – ). Synizesis of ε with a diphthong, and so we may suppose with a long vowel, is common to Ionic, Aeolic, and Arcado-Cyprian verse (Lesbian θέο͜ισ᾽, ἀργάλε͜αι, Cyprian θε͜οι̑ς, 144, 2 in Hoffmann, op. cit., I, p. 76). [72]

{38|39} Once we have set aside the Ionisms cited in the foregoing paragraph we find that there are left almost no forms in the Iliad and Odyssey which have not equivalents in the Lesbian poetic language. We find δου̑ναι twice at the end of the verse, beside three places within the verse where δόμεναι could be used just as well; but we lack evidence on this form for Arcadian which might have a form *δω̑ναι (cf. Arcadian θη̑ναι).

Out of 25 uses of ἕως 23 call for the scansion of Aeolic ἆος; ἕως in P 727 is monosyllabic, leaving only B 78 where ἕως has a definitely Ionic scansion.

The Ionic forms of the first and second personal pronoun plural likewise give us almost no grounds for believing in Ionic changes in the diction. Ἄμμες is found in Homer 37 times within the verse, and never where the Ionic form could take its place. On the other hand ἡμει̑ς, in 73 of its 81 occurrences, could be replaced by ἄμμες, and of the 8 remaining cases 3 occur in the phrase ἡμει̑ς ἴδμεν, that is ἄμμες (ϝ)ἴδμεν. Beside ἄμμι (18 times) ἡμι̑ν is used 72 times: in 55 of these cases ἄμμιν could be used instead, in 16 where ἡμι̑ν is used at the head of the verse before a vowel ἄμμεσιν could take its place, and the single remaining case is ἡμι̑ν ἑκάεργον, that is *ἄμμιν (ϝ)ἐκάεργον. The accusative ἄμμε is used 13 times, always within the verse, while ἡμέας appears 3 times at the verse end, which shows that the Ionic singers felt ἡμέας as disyllabic. Where ἡμέας is found elsewhere it is 8 times followed by a consonant, including one case of the digamma, so that the poetic Lesbian could just as well be used, leaving only 8 cases for which we have no Lesbian equivalent.

Out of 35 occurrences of ὑμει̑ς there are only 4 where ὔμμες could not be used. Out of 38 occurrences of ὑμι̑ν there are only 5 where ὕμμιν could not take its place, and in 2 of these 5 places ὔμμεσιν could be read. Out of the 5 times that ὑμέας occurs, ὕμμες could be used in 3. {39|40}

There are thus only 21 verses of Homer where the meter seems to be warrant for the Ionic form of the personal pronouns: Γ 104 ἡμει̑ς, Δ 246 ὑμει̑ς, Η 194 ὑμει̑ς, I 528 ὑμι̑ν, I 649 ὑμει̑ς, K 211 ἡμέας, Λ 695 ἡμέας, Μ 223 ἡμει̑ς, Ξ 369 ἡμει̑ς, Ο 136 ἡμέας, Ψ 495 ὑμει̑ς, α 76 ἡμει̑ς, β 75 ὐμέας, 86 ἡμέας, 210 ὐμέας 244 ἡμέας, γ 81 ἡμει̑ς, π 387 ὑμι̑ν, χ 264 ἡμέας, ψ 138 ἡμέας, 224 ἡμέας. Even these few cases, however, are not sure. The larger number of accusatives might point towards an artificial (or originally Arcado-Cyprian?) *ἄμμεας in Lesbian verse, and in some cases the metrical fault caused by using the Lesbian form might be like the numerous other faults we see resulting from the formulaic technique. For instance, *ἄμμες ὀτρυνώμεθ’ in Ξ 369 might be due to the modification of a common *ἄμμες δ’ ὀτρυνώμεθ’; cf. at the beginning of the verse ἡμει̑ς δὲ φράζωμεν (Δ 14, P 712). [73]

We have seen in our study that both Arcado-Cyprian and Aeolic singers had used the epic diction and left the mark of their language upon its words and phrases in such a way that we know the epic diction was more or less altogether their creation; whereas we have found only very slight traces of Ionic work besides the mere change in pronunciation. There are two possible conclusions, either that the Iliad and Odyssey as we have them are, save for the Ionic pronunciation, the work of an Aeolic singer or singers, or that they are the work of an Ionic singer or singers who made their verses out of a traditional diction which had undergone almost no change from the time when the Ionians had learned it from the Aeolians.

According to the first explanation, this Aeolic Homer would have lived at the moment when the oral poetry was most creative — for course every oral poem has its moment of creation, however long it may have lived merely by recitation — and would have made his verses from an oral diction which was very ancient, which may even in some parts have gone back to the time before Greek broke up into the dialects in which we find it. But whatever may be the age of the diction, {40|41} it had, in the period before Homer’s time, been used and highly developed by Arcado-Cyprian and Aeolic singers. Whether there was a period of mutual exchange between Arcado-Cyprian and Aeolic, or whether Aeolic took over an Arcado-Cyprian tradition and greatly changed it, we cannot know. Homer would then have composed in this Arcado-Cyprian and Aeolic diction, and his poems would have won such fame that his followers found their profit in merely reciting them. Now recitation is not the natural practice of oral poetry, which, as was seen above (pp. 15-17), is ever in a state of change, so that one must suppose the formation of a guild, not of singers, but of rhapsodes, who made their living by faithfully keeping and reciting the poems which Homer had composed as a singer. We must then suppose further that this guild was fixed in a city which in Homer’s time was Aeolic but later became Ionic, so that the daily speech of the reciters changed from Aeolic to Ionic. This change in their spoken language would have brought about a change in their pronunciation of the Iliad and Odyssey; while the tradition of their craft may have kept the poems line for line, they would have changed the sounds and forms to suit the habits of their speech in so far as that called for no change in the words of the line. Then finally, at some moment well along in the Ionic period of the city, when the Ionic speech had had time to Ionize completely the pronunciation of the poems, they would have been recorded in writing.

Simple as such a theory may seem, — Fick, who had not the notion of an oral poetic language, gave himself needless trouble, since he need almost never have changed more than the form of the single word [74] — it is without any sound basis, and there is much to be said against it. First, such a close keeping of the words in a way quite foreign to the natural functioning of an oral diction must necessarily have made for a keeping of the older pronunciation; in any case the complete Ionization of the equivalent sounds and forms can be understood only by an utterly free handling of the diction by Ionian singers. {41|42} Secondly, we must then suppose that all the poems of the epic cycle, the greater Homeric Hymns to Apollo, Aphrodite, and indeed all the fragments we have of the older poetry, had likewise been kept by some guild from the Aeolic period, with just the same change in their own speech from Aeolic to Ionic; for the language of these other poems, as their diction, differs in no point from that of Homer. In the case of the Hymn to Apolloone would have to reason most ingeniously about the lines wherein the man who sang it makes mention of his hearers and of himself:

{42|43} Here we clearly have to do with a singer and no reciter; he is singing to Ionians; and he says that there are many other singers and that each of them has his own songs. One is forced to grant that this Hymn, in which we find the very same poetic language as in the Iliad and Odyssey, was the work of an Ionian, in the diction common to all other heroic poets of his time, at a moment when the tradition of epic poetry was still that of creation and not of recitation.

Moreover the theory of an Aeolic Iliad and Odyssey rests altogether on one supposition, which is false, namely that the ease with which the poems can be turned into Aeolic proves they must have been more or less entirely as they stand the work of an Aeolic poet: really it proves only that the formulaic diction was Aeolic. As was said above (p. 6), oral poetry is altogether made up of traditional formulas and series of formulas, each of which is an artifice for making the verse and the sentence. The singer has learned these formulas by hearing them in the mouths of older singers, and he makes his own poetry out of them from beginning to end, since the only way he can compose is by thinking in terms of the formulas. Thus while the poems of an oral poetry are ever each one of them in a never-ceasing state of change, the diction itself is fixed, and is passed on with little or no change from one generation of singers to another. This is why we find that even those Ionic words which in themselves are metrically different from the Aeolic words are used in the Iliad and Odyssey in such a way that the change to Aeolic is still possible: in each case the word is used in fixed formulas and types of formulas which are traditional artifices of verse-making. Ἡμει̑ς, for example, can be changed to ἄμμες in 73 out of the 81 times it occurs, which could by no means be due to hazard, and could not happen in the verse of any poet who was making each verse out of his own new words. Homer, however, was using the word in fixed phrases; so we find ἡμει̑ς δ(έ) 41 times as the device for beginning a sentence, and in 34 of these 41 it is a device for beginning the verse as well. Ἡμει̑ς μὲν accounts for 9 other occurrences of the word, falling 8 times at the beginning of the verse. Ἀλλ᾽ ἄγεθ᾽ ἡμει̑ς περ begins the verse twice. This, when we have set aside the 13 cases {43|44} of ἡμει̑ς at the verse-end, leaves ἡμει̑ς before a consonant in only 8 places where the phrase is not clearly fixed in the diction. It is of course largely hazard and the length of the Iliad and Odyssey which bring it about that we can observe the use of a fixed phrase in a number of places and so analyze the technique of its use.

Ἡμει̑ς δ(έ) is itself often used to make up longer formulas. One of the most needed artifices of the singer is that of ready phrases of different metrical values to join his sentences on to one another. Homer’s technique of the conjunctional formula is vast and complex, and, as in the case of the epithet, easily permits analysis into systems of great length and great simplicity. One such series of formulas is that for expressing the essential idea but we: ἡμει̑ς δ’ (20 times), ἡμει̑ς δέ (9 times), ἡμει̑ς δ᾽ αὖ (3 times), ἡμει̑ς δ᾽ αὖτ᾽ (once), ἡμείς δ’ αὖτε (3 times). Each of these phrases expresses the same essential idea but has a different metrical value. These formulas are then in turn used in other formulas. We find at the beginning of the verse ἡμει̑ς δ᾽ ἐμμεμαω̑τες (N 785, ψ 127), ἡμει̑ς δ᾽ ἐσταότες θαυμάζομεν (B 320, Ω 394), ἡμει̑ς δὲ φραζώμεθ᾽ ὅπως (Δ 14, Ξ 61, Ψ 177). Twice we have the pair of verses

(β 127-128, σ 288-289). We have the system ἡμει̑ς δὲ δείσαντες (ι 236, 396), ἡμει̑ς δὲ κλαίοντες (ι 294), ἡμει̑ς δὲ ἰάχοντες (δ 454). [76] Among the nine uses of ἡμει̑ς μὲν we find ἡμει̑ς μὲν τὰ ἕκαστα διείπομεν (Λ 706, μ 16), ἡμει̑ς μὲν γάρ (γ 262, 276) [77] . Nor are the 13 cases of ἡμει̑ς at the verse-end due wholly to chance: the word is regularly used there as a means of filling in the last foot of a verse in which the fifth foot has ended with -ομεν, -ομεν ἡμει̑ς, making, as it were, merely a longer personal ending. Thus we find ἐπέφνομεν ἡμει̑ς (Κ 478), εἴπομεν ἡμει̑ς {44|45} (α 37), κατελείπομεν ἡμει̑ς (λ 53, 447), ἔκταμεν ἡμει̑ς (μ 375), ἐπιέσσαμεν ἡμει̑ς (υ 143). We also find ἡμει̑ς at the end of the verse in the formulaic passage

(Δ 48-49, β 69-70). We are sometimes unable to analyze so exactly the artifice of verse-making which carries in itself the possibility of changing Ionic ἡμει̑ς to Aeolic ἄμμες, but we may be sure, nevertheless, that it is only because the small remnant of Greek heroic poetry which we still have does not let us follow everywhere the vastly varied technique of the diction.

Such an analysis as we have just made for the use of ἡμει̑ς in Homer could be made for any other form or word in the traditional diction which is used at all often, and since each form and word, save for the few Ionisms we have noted, is either Aeolic, or equal in metrical value to Aeolic, we should thus be analyzing an Aeolic technique of verse-making. There thus ceases to be anything surprising in the fact that the Iliad and Odyssey can be turned into Aeolic almost word for word: the formulaic diction was learned by the Ionians from the Aeolians, and though under the stress of habit of their own speech they made it Ionic wherever that could be done without harm to the technique of its use, they otherwise kept it almost without change, since the way in which verse is orally made forced them to do so. The few Ionic forms which we were able to point out above are, for all they are few and some of them doubtful, a precious proof that the Ionic singers had made at least some slight changes in the diction. Far from showing, by their rarity, that Homer could not have been an Ionian, they are, in view of the little change made in an oral diction from generation to generation, just about what one might look for in a tradition of Aeolic heroic poetry which was being carried on by Ionic singers. {45|46}

The study of the Homeric language has thus given us the outlines of the history of the heroic style. From a high antiquity it was carried on in the Greek peninsula by peoples who spoke Arcado-Cyprian and Aeolic. Just what part each of these two peoples had in keeping and developing heroic poetry is not clear. If it was the work of Arcado-Cyprian singers, they must have made an ample use of Aeolic poetry with its Aeolic forms; if it was the work of Aeolic singers, they drew largely on Arcado-Cyprian words and phrases. A long period with mutual exchange of poetry seems most likely. Why the Ionians while in Greece proper had nothing to do, so far as we know, with the formation of this heroic diction, and why when they migrated to Asia they brought with them no heroic poetry, must remain matters for conjecture. It was only in Asia Minor that they met with peoples of Aeolic speech and learned from them to practice the epic. Just how they learned must likewise stay in doubt. The art of the Greek heroic poetry is so far above that of any other oral narrative verse that one might, perhaps, conclude that it was the work of a more highly professional class than that which usually practices oral poetry. In that case we might suppose, as others have, that Ionic heroic poetry was due to the tradition of poetry in some city or cities where the speech of the people had once been Aeolic, and then, because of their defeat at the hands of Ionians, Ionic. However that may be, the new way of life which the Ionians took up in their new land gave a great impulse to the practice of their new poetry, and brought forth all the countless poems which were heard and forgotten, as well as the Iliad and Odyssey. Here again we are in doubt. Were the two poems put in writing during the lifetime of their author, and kept by some group who recited them? Or were they kept by some such guild of recruiters as that which was described above? Or were they passed in manuscript among many singers who, while they still practiced creative oral poetry, found this way of getting for their repertory the poems which had the greatest fame? [78] One thing is plain: our manuscripts cannot {46|47} all go back to a manuscript of Homer’s time; for their variant readings, while some are due to copyists, are for the greater part the variants of an oral tradition, which means that the manuscripts which the Alexandrians used came from different oral traditions.

Nor, because of any evidence which the language gives us, may we say that the Iliad and Odyssey are single poems, nor show how the singer, or singers, of the poems have put smaller poems and whole passages together. The answer to these problems is not to be found in the language, which, like the diction of which it was a part, merely shows that the whole of the two poems, with perhaps a few rare verses excepted, are the work of one or a number of Ionic singers using, at about the same time, the same traditional style, which was itself an Arcado-Cyprian and Aeolic creation. For on the one hand the same diction is common to all singers, and on the other its tradition is so conservative that even the complete unity of language which we find in the poems and in some of the Homeric Hymns might have been kept over a fairly long period. To prove that there were one or many poets, and to show what passages were taken whole from the tradition and which were made anew out of single formulas or verses, we must turn to the study of other oral poetries where the processes of composition can be studied in actual practice and in a greater body of poetry than we have for the Greek epic. When, by the exact analysis of oral poems in reference to their tradition, we have grasped in detail just how the oral poet works, and what it is that makes a poem good or bad in the judgment of himself and his hearers, we shall then, but only then, be able to undertake to study the authorship of the Iliad and Odyssey, and to try to apportion that which is due to the tradition and that which is due to the author. {47|48}