Studies in the Epic Technique of Oral Verse-Making: I. Homer and Homeric Style

[This article was originally published in 1930 in Harvard Studies in Classical Philology 41:73–148. The original page-numbers of the printed version will be indicated within braces (“{” and “}”). For example, “{73|74}” indicates where p. 73 of the printed version ends and p. 74 begins.]

1. The plan of the study (p. 77). —2. The formula (p. 80). —3. The traditional formula (p. 84). —4. The formula outside Homer (p. 90). —5. The formula in Homer (p. 117). —6. The traditional oral style (p. 134)·

In my study of the traditional epithet in Homer, [1] I dealt with those formulas in the Iliad and in the Odyssey which are made up of a noun and one or more fixed epithets, and showed that they were created to help the poet set the heroic tale to hexameters. The noun has a metrical value which allows little change, but by adding to it an ornamental epithet one can make a phrase of the needed length which, since the epithet has no bearing on the idea of the sentence, can be used as freely as the simple noun. I also showed that the technique of the use of the noun-epithet formulas is worked out to so fine a point that it could be only for the smallest part due to any one man. Unless the language itself stands in the way, the poet [2] or poets of the Homeric poems has—or have—a noun-epithet formula to meet every regularly recurring need. And what is equally striking, there is usually only one such formula. An artifice of composition of this variety and of this thrift must have called for the long efforts of many poets who all sought the best and easiest way of telling the same in of stories in the same verse-form. Now no reader of the study, so as I know, has failed to grant its main thesis, which I have just given. When fault has been found, it has rather been with what has seemed to be the bearing of the limited conclusions on the larger problem {73|74} of Homeric style. It has been objected that formulas are to be found in all poetry, where they come either from one writer’s copying another or from his using his own diction over once he has formed it. [3] Then others have thought that the noun-epithet formulas in Homer’s style are an unusual feature, and that they might well have become fixed while the poet was elsewhere left to choose for his ideas what words he would. [4] These remarks, I think, are not without reason. I must claim to have said myself that one could not hope to show by the method used in the study that Homer’s style is altogether traditional, [5] and I believe that the chapter on the epithet in Apollonius and Virgil shows that true noun-epithet formulas are absent from later Greek and Latin verse, if not from all written European literature. [6] But the statement that a certain part of Homer’s diction is almost entirely traditional is one which is sure to suggest larger conclusions; and formulas—or what looks something like them at any rate—would seem to be fairly common in Greek, Latin, and modern verse. The conclusion that Homer’s style is more or less formulaic will not be complete until we know just how large a place formulas have in the style of Homer and in that of later poets. No number of formulas found in later authors would disprove the fact that the fixed epithet in Homer is traditional; but they might keep us from saying that Homeric style is so formulaic that it can be understood only as a traditional and an oral style.

So it is that the criticisms which I have just mentioned seem to me to point out the next step which would naturally be taken in the study of the traditional element in Homer. Moreover, we must know the nature of Homeric style as a whole before it will be possible to go on to other studies which seem to me necessary for the understanding of the Iliad and the Odyssey,—such as the use of noun-epithet formulas in the Greek epic after Homer, which should tell us much about the Singers—ἀοιδοί—and about the making of the Homeric {74|75} poems; or the relation between the formulas, the dialectic forms, and the hexameter, wherein lies much knowledge of the early history of the epic; or the stylistic likenesses between the Greek epic and the oral epics of other nations, which must form the basis of any attempt to judge Homer by what we know about these other poems.

Since these topics have all to do with the style, and more closely with the diction of Homer, I think it may be well here at the outset to explain what seems to me to be just now the value of studies of this sort. There is first of all the sure promise of better knowing the poet’s thought as we note in just what way he has chosen to express it. But there is also the hope that we may thus save the question of the making of the Homeric poems from the danger of scepticism. The scholars of our time have proved the weakness of the attempts made for more than a century to show how the Iliad and Odyssey were pieced together, and though one would hesitate to say, as Lucian makes Homer say, [7] that all the condemned verses are his, one reads the poems in a way which amounts to just about that. Yet those who have thus well refuted the theories which broke up the poems have themselves given no very good explanation of just how they were made. When they have turned to the positive side of their argument, which is the art of Homer, they have often added much to our enjoyment of the poetry, but they have often been as willful in their judgment as ever were those whom they sought to refute. Moreover, they have refused to see the need of answering certain valid questions which had been raised by the ‘higher criticism.’ For example, what reasons have they had for passing over the fact pointed out by Wolf that a limited use of writing for literary purposes, which is the most one can suppose for Homer’s age, must have made for a poetry very unlike ours? [8] What source have they given for the tradition that Homer was recorded only at a later time? [9] How have they explained the unique number of good variant readings {75|76} in our text of Homer, and the need for the laborious editions of Aristarchus and of the other grammarians, and the extra lines, which grow in number as new papyri are found? [10] Finally, have they shown why the poems should be of such a sort as to lend themselves to the many attempts to show the parts of which they were made, and have they told why these attempts were often made by men of the best taste and judgment? Like those whom they were refuting they have, I think, failed, because they would not see that in style and form Homeric verse is unlike that to which they are used.

If we are to learn the true nature of the poems, and if we are to solve the question of their authorship, or know that it cannot be solved, or, as may well prove to be the case, if we are to find that its sure solution does not count for so much as one thinks, we must take another course. We must go back to the principle of Aristarchus of getting “the solution from the text,” but we must enlarge it until it covers not only the meaning of a verse or passage but the poems entire, and lets us know why the poet, or poets, of the Iliad and Odyssey made them as they are, or as they were at first. Whatever feature of poetic art we may study, we must follow it throughout the traditional text, [11] and try to see it clearly and fully; but our hope will not be to find places out of harmony with one another, but instead, after finding all the elements of the poems which bear upon that feature, to draw from them when we can, but from them only, a new idea of poetic artistry. This is, of course, in my own terms, nothing more than one use of the historical method of criticism, [12] which has been used by all good critics. What I wish to point out is not the need of a new method, but of a stricter use, in the supreme problem of Homer’s idea of style and poetic form, of the one good one. It is here, rather than in the study of religious, {76|77} or cultural, or social, or historical details that we must look for the answer to the question of how the poems were made, since the statement of a fact can only be rightly judged when we know how the statement came to be made. Yet it is precisely in the matter of literary form that we are most likely to apply without thinking the ideas which have been gradually formed in us by the writings of later times.

The first move in this attempt to rebuild the Homeric idea of epic poetry will be to show that the Iliad and the Odyssey are composed in a traditional style, and are composed orally, then to see just how such poetry differs from our own in style and form. When that is done, we shall have solid ground beneath us when we undertake the problem of unity in the poems, or judge a doubtful verse, or try to point out how one epic poem would differ from another, or how the greatness of a Singer would show itself. We shall find then, I think, that this failure to see the difference between written and oral verse was the greatest single obstacle to our understanding of Homer, we shall cease to be puzzled by much, we shall no longer look for much that Homer would never have thought of saying, and above all, we shall find that many, if not most of the questions we were asking, were not the right ones to ask.

The poet who composes with only the spoken word a poem of any length must be able to fit his words into the mould of his verse after a fixed pattern. Unlike the poet who writes out his lines,—or even dictates them,—he cannot think without hurry about his next word, nor change what he has made, nor, before going on, read over what he has just written. Even if one wished to imagine him making his verses alone, one could not suppose the slow finding of the next word, the pondering of the verses just made, the memorizing of each verse. Even though the poet have an unusual memory, he cannot, without paper, make of his own words a poem of any length. He must have for his use word-groups all made to fit his verse and tell what he has to tell. In composing he will do no more than put together for his needs phrases which he has often heard or used himself, and which, grouping themselves in accordance with a fixed pattern of thought, come naturally to make the sentence and the verse; and he will recall his {77|78} poem easily, when he wishes to say it over, because he will be guided anew by the same play of words and phrases as before. The style of such poetry is in many ways very unlike that to which we are used. The oral poet expresses only ideas for which he has a fixed means of expression. He is by no means the servant of his diction: he can put his phrases together in an endless number of ways; but still they set bounds and forbid him the search of a style which would be altogether his own. For the style which he uses is not his at all: it is the creation of a long line of poets or even of an entire people. No one man could get together any but the smallest part of the diction which is needed for making verses orally, and which is made of a really vast number of word-groups each of which serves two ends: it expresses a given idea in fitting terms and fills just the space in the verse which allows it to be joined to the phrases which go before and after and which, with it, make the sentence. As one poet finds a phrase which is both pleasing and easily used, the group takes it up, and its survival is a further proving of these two prime qualities. It is the sum of single phrases thus found, tried, and kept which makes up the diction. Finally, the poem which is a thing of sound and not of writing is known apart from its author only because it is composed in the same style which others use and so can remember. Writing may be known, and the poem may be dictated and recorded, and the knowledge of writing may thus have some bearing on the text of the poem. But it will not have any upon its style, nor upon its form, nor upon its life in the group of poets and the social group of which its author was a part.

Such in its broadest lines is the composition of oral poetry as it is practiced in our own times in Serbia, among the Tuaregs, in Afghanistan, and in many other places; [13] and it is clear that the best way of knowing whether a style is oral and traditional is to hear it in use, or, lacking that, to compare the recorded work of several poets who have made their verses out of the same formulas. But we cannot do either {78|79} of these things for the Greek epic. There is too little known about the making of the early poetry in hexameters for us to liken the Singers to the Serbian Guslars without more ado, or to make of Homer a Singer like any other. Moreover, we cannot date the works of this early time at all surely, and we have nothing to show us that any one of the poems we have was made by a Singer. Opinion generally grants a vague body of traditional epic formulas, and we have a certain amount of poetry composed in a style which is either entirely or nearly like that of the Iliad and Odyssey; but the notion is also current that Homeric phrases found outside these two poems are more or less due to the studied imitation of the style which one poet made. We should be well off if we knew for sure that Homer could not write, but writing may have been known in Ionia in his time, whatever were the uses it was put to. [14] If we are to draw any solid conclusions {79|80} about the style of Homer, we have only one course to follow. Seeking “the solution from the text” we must see whether the diction of the Iliad and of the Odyssey is of a sort which can be understood only as a traditional and oral technique of making verses by means of formulas. The reasoning will be as follows. First, we shall define the formula. Then we shall look to see what means there are of telling whether a formula is traditional or not. The nature of the formula will show us that the more formulas we find in a poet’s diction, the smaller is the portion of them which could be the work of that single poet. We shall then be led to a study of the verse of poets who we know wrote, that we may learn how often the formula can appear in written verse. Finally, having seen if the formulas in Homer’s verse are so much more common that they suffer no comparison with those of any written poetry, and having thus learned how much of the formulaic element is surely traditional, we shall be able to consider what reasons there are to say that Homer’s is an oral style.

The formula in the Homeric poems may be defined as a group of words which is regularly employed under the same metrical conditions to express a given essential idea. The essential part of the idea is that which remains after one has counted out everything in the expression which is purely for the sake of style. Thus, the essential idea in the word-group ἦμος δ᾽ ἠριγένεια φάνη ῥοδοδάκτυλος Ἠώς is ‘when it was morning,’ that in βῆ δ᾽ ἴμεν is ‘he went,’ that in τὸν δ᾽ αὖτε προσέειπε is ‘he said to him.’ The word-group is employed regularly when the poet uses it without second thought as the natural means of getting his idea into verse. The definition thus implies the metrical usefulness of the formula. It is not necessary that a poet use one certain formula when he has a given idea to express and a given space {80|81} of the verse to fill, since there can be formulas of like metrical value and meaning which can take the place of one another, though they are rare in Homer. But if a formula is to be used regularly there must be a steady need for it. For example, Homer uses θεὰ γλαυκῶπις Ἀθήνη fifty times to express in the last half of the verse, after the trochaic caesura of the third foot, the idea ‘Athena.’ The simple number of times the phrase appears is the direct measure of its usefulness, though if one wishes further proof a study of its uses shows it to be part of a fixed device for making hundreds of verses. [15] Kurt Witte’s remark [16] that the language of the Homeric poems is the work of the epic verse is by definition true also of the Homeric diction so far as it is made up of formulas. When one has added the factor of the story, since it is this which gives the poet his ideas, and that of the poetic merit of the expression, which also must have its share in the making and the keeping of it, one may state the principle as follows: the formulas in any poetry are due, so far as their ideas go, to the theme, their rhythm is fixed by the verse-form, but their art is that of the poets who made them and of the poets who kept them.

When the element of usefulness is lacking, one does not have a formula but a repeated phrase which has been knowingly brought into the verse for some special effect. Thus, the definition excludes the refrain, as found in Aeschylus or Theocritus:

or in Shakspere or Marlowe:

The definition likewise excludes the echoed phrase. [21] I give examples from Theocritus and Shakspere:

Non-formulaic too is the verse which is borrowed because the poet’s public knows it and will recall its former use, as that in which Sophocles gives to the dying Clytemnestra the words which Agamemnon had spoken in the play by Aeschylus. I quote the verses of the older poet, then those in which they are imitated: [24]

Finally a poet will often repeat a phrase after an interval in order to obtain some special effect, as Sophocles does when Oedipus, fearing for the first time that he himself is the slayer of Laius, repeats in horror the words by which he had banished from the land the unknown murderer. [27]

But the repeated phrase, as distinct from the formula, is used more often in less outstanding ways than these. When Bacchylides, for instance, wrote λευκώλενος Ἥρα, [28] or ὑψιπύλου Τροίας, [29] he was not using the words because he had a certain space of verse to fill and a certain idea to express: he was working epic phrases into his poem. Similarly, Pope in his translation of the Odyssey [30] borrowed Milton’s phrase “thick as autumnal leaves” from the Vallombrosa simile in Paradise Lost. The fact that he had nothing like this in the verses he was translating shows clearly what is evident enough anyway, that he was using the phrase for its idea rather than for any help it gave him in expressing certain ideas which he was seeking to put into verse. We shall see later that no distinction counts more for us than this between the real formula and the phrase repeated for the sake of its poetic thought or wording. I have quoted these examples here because they bear on the definition of the formula.

There is one other thing to note before leaving this subject: the problem of the formula is not that of literary influence. This fact more than any other has been overlooked by those who have dealt with traditional style. When Pindar, for example, wrote:

he was without doubt recalling the Homeric phrase—πτερόεντες ὀιστοί [32] —and the Homeric influence is proved. But what was a formula to Homer was none to Pindar. The task of getting his words into his verse was quite the same as if he had been using an expression of his own making. The formula is useful only so far as it can be used without changing its metrical value. The change of endings is too easy to have {83|84} any measurable effect upon the usefulness of a phrase. One counts by the thousands in Homer such cases as the change of ἐυκνήμιδες Ἀχαιοί to ἐυκνήμιδας Ἀχαιούς, or of θνητῶν ἀνθρώπων to θνητοὺς ἀνθρώπους. And to these must be added the change of δέ to τε, as when φέρων τ᾽ ἀπερείσι᾽ ἄποινα [33] becomes φέρω δ᾽ ἀπερείσι᾽ ἄποινα, [34] or even the omission of these particles, or such a change as that of μου to σου. But any less simple alteration in the word-group supposes thought of some length on the part of the poet. [35]

The question of who made the formula does not enter into its definition, since it would be equally helpful to a poet whether it was his own work or that of another. What means then are there of knowing whether the formulas in Homer are borrowed or not, since we have no right to suppose, as the basis of our reasoning, that the Iliad and the Odyssey are necessarily due to more than one man? The solution lies in the schematization of the Homeric style, which does away altogether with the need of knowing how many poets worked at these poems.

Formulas are of two sorts. First, there are those which have no close likeness to any other, as, so far as we know, is the case for ὀνείαθ᾽ ἑτοῖμα προκείμενα in the following verse, which is found three times in the Iliad and eleven times in the Odyssey:

The other kind of formula is that which is like one or more which express a similar idea in more or less the same words, as, for example, ἱερὸν πτολίεθρον ἔπερσε [36] is like ἱερὸν πτολίεθρον ἑλόντες, [37] or as ὀλέκοντο δὲ λαοί [38] is like ἀρετῶσι δὲ λαοί [39] and δαινῦτό τε λαός. [40] We may say that any group of two or more such like formulas make up a system, and the system may be defined in turn as a group of phrases which have the same metrical value and which are enough alike in thought and words to leave no doubt that the poet who used them knew them not only as single formulas, but also as formulas of a certain type. For example, one finds in the Iliad and the Odyssey a group of phrases which all express between the beginning of the verse and the trochaic caesura of the third foot, in words which are much alike, the idea ‘but when he (we, they) had done so and so’:

This scheme shows not only that the poet or the poets of the Iliad and the Odyssey had a formula to express the idea ‘but when they had prayed,’ for instance, between the beginning of the verse and the trochaic cæsura of the third foot. [41] It shows also that he, or they, knew {85|86} a type of formula in which to αὐτὰρ ἐπεί was added an indicative verb-form of the measure ⏔ – ⏑ beginning with a single consonant; and also another type in which to αὐτὰρ ἐπεί was joined first ῥ᾽, one form of that helpful and many-shaped particle, then an indicative verb-form beginning with a vowel or diphthong and measured ⏔ – ⏑. Thirdly, there was a type where αὐτὰρ ἐπεί, lengthened by the addition of the syllable -δη, allowed the use of verb-forms of the value – ⏑ And lastly, there was a type of formula in which αὐτὰρ ἐπεί, changed to αὐτὰρ ἐπήν, made way for subjunctive forms of the verb. Each of these our groups of formulas may be called a system, since it is clear that the poet, or poets, who used them, felt the exact device, as I have taken care to analyze it, for fitting into the verse verb-forms of certain moods and measures. Finally, the four groups taken together form a larger system in which the common likeness, while less close, is no less real.

It is the system of formulas, as we shall see, which is the only true means by which we can come to see just how the Singer made his verse; but we are interested in it now solely as a means of measuring the schematization of a poet’s style. There are in such a measuring two factors, that of length and that of thrift. The length of a system consists very obviously of the number of formulas which make it up. The thrift of a system lies in the degree in which it is free of phrases, which, having the same metrical value and expressing the same idea, could replace one another. What the length and thrift of a system of formulas are can best be explained by describing one of the most striking cases in Homer, that of a system of noun-epithet formulas for gods and heroes in the nominative. All the chief characters of the of the Iliad and the Odyssey, if their names can be fitted into the last half of the verse along with an epithet, have a noun-epithet formula in the nominative, beginning with a simple consonant, which fills the verse between the trochaic caesura of the third foot and the verse end: for instance, πολύτλας δῖος Ὀδυσσεύς. It is the number of different formulas of this type, well above fifty, which makes the length of this system. But besides that there are, in only a very few cases, more than one such formula for a single character, though many of them are used very often, as πολύτλας δῖος Ὀδυσσεύς, which is found 38 times, θεὰ γλαυκῶπις Ἀθήνη 50 times, Ποσειδάων ἐνοσίχθων 23 times. To be exact, in a list of 37 characters who have formulas of this type, which includes {86|87} all those having any importance in the poems, there are only three names which have a second formula which could replace the first. [42]

In the case of this system, as in that of other formulas, such as those of the types πολύμητις Ὀδυσσεύς and δῖος Ὀδυσσεύς, the length and the thrift of the system are striking enough to be sure proof that only the very smallest part of it could be the work of one poet. [43] But for the greater number of systems which are found in the diction of the Homeric poems we cannot make such sure conclusions, since their length is rarely so great and their thrift never so striking. This does not mean that the proof by means of the length and thrift of the system is possible only in the case of the noun-epithet formulas. It is clear without need of further search that the greater part of the system quoted above must be traditional, and that the type of the formula and the words αὐτὰρ ἐπεί at the beginning of the verse are surely so. But one can see that an attempt made in this fashion to see just how much of Homer’s diction had been handed down to him could give only very partial results, even if the task were not of an impossible length, as it is. What we must look for is, more simply, for the degree to which the diction of poetry outside the epos can become schematized. If, having gauged the systematization of Homer’s verse and of that which we know to have been written in the individual style of single poets, we find a difference which forbids any comparison, we shall know that Homer’s poetry was not made in the same way as was that of later poets. We shall then see that we are faced with a problem which can be solved only by granting that Homer composed his verses entirely in a style that was traditional and adapted to oral verse-making.

Table Placeholder

In making this comparison of the systems in Homer with those in later poetry we shall not, as it happens, have much to do with the thrift {87|88} of the system, since we shall find it hard enough to get together outside of Homer any systems which show the first quality of length. We {88|89} shall seldom get any farther than the overwhelming difference in the number of repetitions. Since this is the case, it is well to point out beforehand that the number of repetitions in a style, and the frequency with which they are used, bear directly upon the thrift of the diction. One may ask why Homer uses the formula αὐτὰρ ἐπεί ῥ᾽ ἵκοντο three times. [44] That is one face of the coin. The other face is the question of why Homer did not use other words, of whatever sort they might be, for the expression of the idea ‘but when they came.’ That is, the repeated use of a phrase means not only that the poet is following a fixed pattern of words, it means equally that he is denying himself all other ways of expressing the idea. This may seem a very trivial point to make, if one has in mind only a few scattered formulas, none of them used more than a few times. But when one has even a single phrase used, for instance as is τὸν (τὴν) δ᾽ ἠμείβετ᾽ ἔπειτα, 48 times in the Iliad and 24 times in the Odyssey, it is as if Homer wished to tell us how little use he has for all other ways of expressing the idea, which we must suppose to be very numerous. Then, when one multiplies the case of the single formula by all those which are to be found in the two poems, and which require the 250 pages of C. E. Schmidt’s Parallel-Homer [45] for their listing, one has the statement of a thrift of expression which it is rather hard, perhaps, for us to understand. Yet we must remember, as in the following pages we seek for formulas in later verse, that the poet’s indifference to the new way of saying a thing is to be measured in the exact terms of the number of repetitions he uses and of the times he uses them. {89|90}

P. F. Kretschmer, in his study of repetitions [46] within the work of Hesiod, [47] found within the 1022 verses of the Theogony 338 repetitions of which the larger part are phrases found in Homer. This proportion does not come near that of the twenty-five or six thousand repetitions [48] in Homer’s 27853 or so verses, nor would one look for it in poems of such different lengths. Still it is far above anything we shall find for the poetry of times in which writing was beyond any doubt the usual means of composition. We find in Hesiod even more repetitions from poetry which could not be his. A. Rzach [49] notes 67 cases in the first 100 lines of the Theogony where a phrase is identical with one found in the Iliad or the Odyssey. The number of repetitions within the Works and Days is smaller—84 in a poem of 828 verses. [50] For the first hundred lines Rzach notes 55 Homeric parallels, but in the gnomic portion of the poem this number falls to 31 in a hundred lines. [51] But even this figure is far from any which is ever found outside the early hexametric poetry. It is not the place here to explain the varying degrees of repetition within the Hesiodic poems, nor the use of Homeric phrases. That will be possible only in a longer study in which one will throw aside the idea of imitation, which has weighed so heavily on the early poetry outside Homer, and take up the repetitions as part of a traditional technique of verse-making. One will then learn, I believe, a great deal about the nature of the epic diction, of its use for different subjects, and by different poets or schools of poets, and of its decline. Here we can only point out that the formula is used in Hesiod far more often than it ever is outside of the early epic; and the same thing is to be said for the Shield of Heracles, which in its 480 verses repeats itself 63 times, [52] and has in its first hundred {90|91} lines 74 Homeric phrases, [53] as well as for certain of the Hymns and for the fragments of the other early heroic poems. T. W. Allen, for instance, states that 20 verses of the Hymn to Aphrodite “are taken from Homer with little or no variation,” [54] and I find seven Homeric phrases in a ten-line fragment of the Thebais. [55]

N. Riedy found in Solon 48 phrases repeated without change from the Iliad or the Odyssey or the Hymns; [57] of these all but one are found in the 221 elegiac verses of this poet. There are none in his iambics. This makes about 21 epic phrases to a hundred verses, a figure fairly near that found for the gnomic part of the Works and Days. In the 932 verses of Theognis which Bergk thought genuine R. Küllenberg found 144 phrases repeated from Homer, Hesiod, or the Hymns, which would be about fifteen epic phrases to a hundred verses. [58] No one has studied the shorter repetitions within the elegiac poetry, but Küllenberg remarks that in the hexameter the elegy follows the epic. [59] So here too the formulaic element must be studied as a part of the traditional diction of the early verse in hexameters. But Küllenberg also states that the elegy follows itself in the pentameter. He quotes in proof 18 phrases, all found in the last half of the pentameter, which appear in the work of the elegiac poets a total of 99 times. [60] Moreover, certain of the systems into which these phrases fall are {91|92} long enough to show the traditional character of the greater part of the expressions which make them up, as in the following example:

Such a large number of formulas and systems of this sort are found outside the hexameter only in this one place, and, if we knew surely that writing was the regular means for the composition of verse in the sixth century, there would very likely be no need of carrying our search any farther. But we do not. The example of Serbian poetry shows that traditional dictions can exist side by side for different verse-forms and for different types of poetry, and the doubt which hangs over the sources of Theognis’s poem would point to anything but an originally written text. A study of the elegy, which kept in view the possibility that its style was in a larger or smaller measure oral and traditional, might explain the very many verses and passages in Theognis which some editors have given to Tyrtaeus, Mimnermus, and Solon, because they are found, in more or less the same form, in their work as well. [61] If this small amount of poetry we have is at all typical, the common element in the elegy was very large. The conclusions of such a study, however, could have only a limited value for our own problem: it is hard to see how it could prove that Solon and Theognis first wrote out their verses, and though it would doubtless confirm Küllenberg’s idea that there was a formulaic element in elegiac style, common to all poets, there are other ways of showing this to be true for Homer.

When Riedy [62] remarks that in Solon epic formulas are about twice as frequent in the hexameter as they are in the pentameter, and when Küllenberg says that the last half of the pentameter has formulas {92|93} found only in elegy, they are dealing with a fact which has been strangely overlooked, namely, that the formula can be useful only in the smallest degree in any other sort of verse than that for which it was made: the nature of the hexameter is such that only a small part of the epic formulas are found in more than one place in the verse; likewise one will hardly hope to find many of them in a verse-form in which the sequence of longs and shorts and the length of the kola are only rarely those of the epic verse. The one case in Solon’s work of an Homeric phrase which is not found in elegy occurs in a skolion which may or may not be his: [63]

Here the strong dactylic movement of the verse gives to the participle, and to ἄνδρα ἕκαστον, something of the movement we find in the Homeric line:

It is surely not a very striking phrase, and one would be tempted to say it was only due to chance, if it were not for the hiatus which makes it certain that it was taken from the epic, for like μέλανος οἴνοιο in the fragment of the epic poet Antimachus, it shows a sense for the lost digamma of the epic phrase similar to that of the feeling of the French for ‘h-aspiré.’ [65] But such likeness in rhythm between epic and lyric can only rarely happen. H. Schultz [66] gives 52 cases in which Pindar has copied a phrase of the Iliad or the Odyssey, of which, it is well to note, 48 are made up of two words, and the remaining four of three words: the rhythm barred out all longer Homeric expressions. Yet of these 52 there are only 19 which Pindar could use as he found them. [67] In the {93|94} case of the others he had to change the order of his words, or use them in other forms which would give them a new rhythm. They even then show the influence of the epic upon Pindar, but they do not show that he was helped in any way, since these words were no easier to work into his verse than any others which he might find himself. The number of phrases which Bacchylides took from Homer without change is equally small: H. Buss found eleven, all of two words. [68]

There is no need of pointing out that so few formulas in the work of Pindar and Bacchylides could have had no measurable effect on the way in which they made their verses; but besides that it is only too clear that these repeated phrases are not formulas. They are all of them high-sounding expressions which the poet has been able to work into his verse, as for example φίλον ἦτορ in the first of the Olympian Odes:

Both the meaning and the movement of φίλον ἦτορ are here very far from those which Homer has made familiar to us:

Even in what may be the most Homeric of all the Pindaric imitations, that of a phrase of three words in an ode in dactylo-epitritic metre, the words which go before and after rob the phrase of much of its Homeric sound: {94|95}

In Homer we had:

Far from being formulas by which he would regularly express his idea under certain metrical conditions, these phrases were to him fine expressions which his mind had kept solely for their beauty, and which the chance of his verse now let him use. One would not deny all usefulness to them, since they did after all fit into his verse, but that is exactly the usefulness of any phrase which goes to make up any poem.

Since, then, it is not the epic at least which gave the choral poet what Wilamowitz calls “ein ganzer Apparat von konventionellen Wendungen und Schmuckstücken zu Gebote, so dass er noch leichter als der Rhapsode den einfachsten Gedanken nach Bedarf variieren und dehnen kann,” [75] this conventional element in the style must be the work of the choral poets themselves. But Wilamowitz is surely mistaken here, for how could there be such a body of phrases for a poetry in which the order of long and short syllables in the verse varies with every poem? In ten pages of the concordance to Pindar there is not one repeated phrase, whereas not a column of the Homeric concordances but teems with them. [76] A comparison of the diction of Bacchylides with that of Pindar gives the same results. W. K. Prentice [77] gives {95|96} 72 cases in which there is some sort of likeness between the words of the two poets, as in these verses of Pindar and of Bacchylides:

Such an example as this well shows the influence of the one poet on the other, or the use of ideas common to the poetry; but only in one of the 72 cases, in which the phrase is used without change—ταῖς ἱεραῖς Ἀθάναις in Pindar [80] and τᾶν ἱερᾶν Ἀθανᾶν in Bacchylides [81] —did the older poet spare the younger the trouble of making the expression over for his own needs.

Indeed, it seems to me that one gives a very wrong idea of the style of choral poetry in likening its conventional side to that of the epic. Homer is telling the old tales in words which his hearers scarcely heeded as they followed the story, for those words were to them the only ones which could be used, and they knew them far too well to think about them. But Pindar is moving alone in his own thought, choosing in a way that is his alone from the grand words of poetry. There is of course much that is traditional in his verse: he uses the old words, and follows a more or less fixed order of thought, and uses the old tales, and points the moral. Nor did he scorn the common devices, such as that of passing from one part of his theme to another by means of a relative clause, nor fail to use an epithet to fill his verse when that would help him, as Lucian charged the lyric poets with doing. [82] Tradition gave him these artifices, but it did not give him his phrases. These he must choose, and if he would use an epithet he must think and pick. [83] We shall find in later verse what may be, {96|97} perhaps, a small number of formulas, but it will be in verse in which the order of shorts and longs is fixed and recurring. Pindar, ever faced with a new metrical need, however often he might use his ideas again, could make good or bad sentences, but they must be his own. If we admire the epic style as a thing beyond the forces of a single man, we must wonder at the use that Pindar alone could make of words.

Since the verse of drama is dactylic even less often than that of the choral lyric, it contains still fewer Homeric phrases. Of the 112 passages in Aeschylus given by Susan B. Franklin as showing epic phrasing, there are only three in which the words have been left unchanged. [85] All three cases occur in the lyrics. Max Lechner found five examples in Sophocles of the unchanged Homeric phrase, likewise in lyric metres, [86] and in Euripides eight, [87] to which I can add two that he has overlooked. Of these ten expressions seven occur in lyrics, the other {97|98} three are found within a space of twelve verses, for a reason. Andromache, just before the first entry of the chorus in the play which bears her name, ceases to speak in trimeters, which so far have made up the play, and breaks into a lament in elegiac verse. In this passage of fourteen lines, which is the only example we have of elegy in tragedy, Euripides was seeking an epic tone, and to this end he used an unusually large number of Homer’s words, and half as many of his phrases as he did in all the rest of his dramas. Moreover, all three phrases occur in the hexameters of the distichs,—so friendly was this verse to the epic diction, whereas the iambic and trochaic lines rejected it altogether!

But though we have fewer Homeric phrases in tragedy than in choral poetry, we may well look for tragic formulas. Here are poets using more or less the same style, and the same kind of stories, and, finally, giving the first place in their plays to the same verse-form. The irregular rhythms which kept choral verse free from formulas have here a limited place. A. B. Cook, [88] without attempting completeness, but implying that he gives all the more evident cases, cites 23 passages in Euripides’s Trojan Women which recall the wording of the Hecuba which appeared ten years earlier; in nine of these parallels we find expressions which are repeated without change. F. Niedzballa [89] gives a list, which seems to be inclusive, of repetitions within the Prometheus of Aeschylus; of these fifteen are unaltered. F. Schroeder, however, furnishes by far the most ample evidence of repetition without change within the work of a tragic poet: [90] he gives 297 different cases within the plays of Euripides, all but six of them in the trimeters. Of these phrases 48 are used three times, 13 four times, 4 live times, one six times, and one seven times. This makes 392 cases in which an expression appears which the poet has employed before. The first appearance of course cannot be counted, since a phrase cannot become a formula until it has been used more than once; and since the greater number of expressions are repeated only once, we can be sure that all but a very few of them are really being used for the first time in the first of the extant plays in which they appear. To this {98|99} number of 392 may be added 91 other cases of borrowing from the work of Aeschylus and Sophocles, making a total of 483 repetitions. This may seem at first a very large proportion; certain scholars have cited far fewer repetitions in later verse as a final proof that Homer’s use of formulas was no different from that of modem poets. Yet this number straightway loses its importance when one computes the average number of lines between these repetitions, and between those which we find in Homer, for we then see that Euripides is repeating himself, or borrowing, in every fortieth iambic verse, [91] whereas Homer, if we discount likewise the first appearance of the phrase, is doing so in more than every other verse. Nor is the objection at all sound that most of Attic tragedy is lost, and that if we had it all the number of repetitions would be much larger; the 91 phrases which Euripides took from Aeschylus and Sophocles would have to be multiplied many times before they would change our conclusions, and one would also have to suppose other poems for Homer’s time. In a stylistic study small statistical differences have little value: one must use the strictest method of search, but the differences found must be large enough to be beyond the reach of any imaginable faults in method. Whether we suppose that Euripides used a repeated phrase in every thirtieth, or every twentieth verse, it could never be more than an expression put into his verses from time to time. There is not the least reason to suppose, as there is for Homer, that he made any considerable part of his poetry out of them. As for the possible conclusion that Homer could have made as many formulas by himself as Euripides, that is, one for every 36 or so that were handed down to him, [92] I do not think that that will please those who dislike the notion that Homer’s style is not more or less his own. But it is time at this point to remind ourselves that we are, in these calculations, supposing that the repeated phrases in Euripides are all formulas. We may very well find that they are not at all the regular means of expressing an idea under certain metrical conditions, but phrases which the poet brought into his work a second time because he could obtain some special effect by them. {99|100}

We pass now from the mere frequency of the repetitions in tragedy to their nature, and it may be well to say here that our purpose is much more than that of showing that the number of tragic formulas is much smaller even than the number of repetitions, which is already too small to have any bearing on the epic practice. The one thing which we are seeking to know is what the formula is: its higher frequency in Homer cannot in itself have any great value for us until it leads us to see that frequency is a quality of the formula. The study of the character of the repetitions in tragedy, by showing us just why they are not more frequent, will help us to this understanding.

We may begin with the five cases which Schroeder found of a phrase which appears unchanged in the work of all three tragic poets. One of these expressions is no more than a poetic locution: ὑπουγῆσαι χάριν [93] for the prosaic χαρίσασθαι. Δόμων ἐφέστιος [94] is also highly poetic, and πρὸς τὰς παρούσας συμφοράς [95] and τῆς νῦν παρούσης πημονῆς [96] have a more than usual dignity of statement, but besides this the three phrases express ideas which are more than usually striking. The idea of τὸν παρόντα δαίμονα, [97] used by Aeschylus of fate, and by the other two poets of a particular fate, that of death, is of a force which calls for no comment. We find, then, that the repeated phrases common to the three writers are either especially poetic in wording or highly dramatic in thought. To bring out the meaning of this last term we must have more examples. In passing, then, to the phrases found in only two of the poets, there is hardly need of saying that the rarity of the cases we have just quoted disproves the existence of any body of formulas common to tragic poetry, and makes the question of the phrases found in the work of more than one tragic poet purely that of the influence of a dramatist.

Schroeder [98] lists 29 different phrases found both in Aeschylus and {100|101} in Euripides, all of which appear once in the work of the older poet, and of which three appear twice in Euripides, two four times, and one six times. He finds 34 different phrases which Euripides took from Sophocles, of which three appear twice in Euripides, two three times, and one four times. These figures check with those of others. Thus F. Niedzballa [99] gives a list of phrases repeated from Aeschylus’s Prometheus, of which fourteen occur in the work of Sophocles, and fifteen in that of Euripides. M. L. Earle, [100] in a study in which he sought to prove the influence of the Alcestis of Euripides on the Women of Trachis of Sophocles, found one repetition: καὶ συνωφρυωμένος. [101] To this must be added another given by Schroeder: ἅλις γὰρ ἡ παροῦσα. [102] Here again the evidence is overwhelming: to judge from the plays which we possess, Euripides uses an expression from Aeschylus or Sophocles in every 215 or so iambic verses.

When one looks at the phrases Euripides has thus chosen, it is straightway clear that almost none of them belong to the more general level of the style: either their wording is more than usually poetic, or their thought highly dramatic. As examples of the first kind one notes ὥστε ναὸς κεδνὸς οἰακοστρόφος, [103] —Pindar seems to have been the first to use οἰακοστρόφος for οἰακονόμος; [104] πράσσε τἀπεσταλμένα; [105] πόλεμον αἴρεσθαι μέγαν; [106] συνάψας μηχανήν. [107] These expressions Euripides took from Aeschylus. From Sophocles he took μωρίαν ὀφλισκάνω, [108] and τοῦτο κηλητήριον, [109] —he is the only other author to use this poetic adjective. In the greater number of cases, however, the phrase is rather what must be called a specific dramatic device. We must consider here the essential difference between epic and drama. The epic contains a good {101|102} deal of speech, which, in its way, comes very near to drama in its direct imitation by action, and often, to make this imitation effective, the epic poet uses formulas made especially for this end. Such formulas are of various sorts. They may express indignation at some sight:

They may set forth the clash of opinions:

They sometimes imitate the tone of one threatening:

Less often they may be of a less purely emotional tone, giving the intention of the speaker, as in

Of this dramatic sort too are the verses which comment on a situation, and which Homer, refusing to let himself enter his poem, always gives as the opinion of a character in regard to some certain event: [114]

What all these formulas I have quoted have in common is that they are the character’s expression of what is going on in his mind, rather than the poet’s statement of what a character did, which is the nature of narrative. Now the epic being far more narrative than dramatic, the dramatic formulas have only a very small place beside those which tell the tale. But the dramatic poet, giving us characters who think and feel before us, needs expressions of this sort far more than any other.

Accordingly we find that the repetitions in tragedy which are not stylistic are almost all special devices for supplying the dramatic {102|103} element. In one of the three cases in which Euripides has taken almost a whole line from Sophocles we have a means whereby a character expresses despair, as Electra does in the play of Sophocles:

So Hecuba speaks in Euripides:

In another case the repeated verse is one by which, in the rapid give and take of angry talk, one character bids another ask his question. Sophocles wrote:

Euripides changed only one word of this:

Both verses, of course, are found in stichomythy. An example of the phrase by which we know the speaker’s intention is this verse from Aeschylus:

In Euripides this becomes:

Very frequent among the phrases Euripides borrowed are those in which a character expresses himself by a nice use of language, as in the words which Euripides took from Sophocles: ἑκόντες οὐκ ἄκοντες, [122] and τά τ᾽ ὄντα καὶ μέλλοντα. [123] Sophocles in turn took a verse from Euripides and did not trouble to change it at all:

If the only repeated phrases in Homer were those of the dramatic sort which we have quoted, and if they did not appear too often, so that we could be sure that the poet’s hearers were always very much {103|104} struck by them, we should have no need to seek the difference between epic and tragic repetitions. But in Homer these formulas have only the smallest place beside those which make up the narrative, or even the speeches, and many of them are so frequent that it is doubtful whether their dramatic effect would ever have set them much apart from the more usual level of the style. It is otherwise for Euripides: almost all of his repeated expressions are especially forceful, and, rarely used more than once, they are always sure of their effect. They are, then, not a regular means of expressing the idea but a body of outstanding dramatic artifices. There is almost nothing in them to show that Euripides, in order to make the composition of his verses easier, was limiting his thought to the diction created by others.

Voltaire was doing very much as Euripides had done, when, in his Oedipe, he borrowed two verses from the play of that name by Corneille:

Voltaire felt called upon to give his reasons for thus using the lines of another: “Je n’ai point fait scrupule de voler ces deux vers, parce qu’ayant précisément la même chose à dire que Corneille, il m’était impossible de l’exprimer mieux; et j’ai mieux aimé donner deux bons vers de lui, que d’en donner deux mauvais de moi.” [125] This is the very reasoning whereby borrowing in the Greek orators was justified: τὸ γὰρ καλῶς εἰπεῖν φασιν ἅπαξ περιγίγνεται, δὶς δὲ οὐκ ἐνδέχεται. [126] These, however, are not the grounds for the use of the true formula: Voltaire does not say that the borrowed lines made his verse-making easier; he would have been ashamed to admit any other than purely artistic motives. For him, what comes before all else is the idea to be expressed, and which he has for his own reasons chosen to express. In this case he had found his ideas in Corneille, where they had struck him by their high emotional and dramatic quality. He used the ideas and the words from which he could not separate them; but we may {104|105} well suppose that he spent as much thought in borrowing these verses as Corneille did in making one of them out of the lines of Seneca:

One only has to think of the number of formulas in Homer, and of how closely they follow one another, to see that Homer’s use of borrowed phrases could have been nothing like this. A poet making verses with the greatest care, who sought to put into his poetry all that he had found best in the poetry of others, could never have thus stopped at every other verse to ponder some line he knew, whether that of another or his own. Virgil, it would seem, did this more than any poet we know of, yet he is far from such a practice. [128] The case of Virgil, indeed, bears very directly upon this distinction between the formula and the phrase which expresses an unusually striking idea: far from being led by any consideration of an easy verse-making he is quite willing to translate his striking ideas from Greek. Virgil is not a writer of plays, of course, to be brought in with regard to the effective dramatic phrase. But it is clear that, as the tragic poet is concerned with the forceful dramatic expression, so Virgil, writing heroic narrative, is seeking the salient epic phrase.

I do not think it should be said that the element of usefulness is absolutely lacking from the phrases which Euripides borrowed. One case I am very nearly tempted to class as a true formula. It happens to be the one which the poet used most often, namely ἀλλ᾽ εἰσορῶ γάρ. This is used once by Aeschylus and six times by Euripides. [129] The words are a means of turning the eyes of the audience towards an actor who is just coming on to the stage. But even here the device is more than ordinarily dramatic, and there is no other case which approaches this. Perhaps one should grant too that the poet was helped somewhat by the poetic locutions be borrowed. But the fact that they fall between groups of words which the poet was using for the first time, and not, as the true formula does, into a pattern of formulas {105|106} which were made to fit before and after it in the verse, brings them after all very close to the phrases which Pindar took from the epic and which, as we have seen, were no more or less helpful than any words which were being used for the first time.

The expressions which Euripides repeated from his own works are not very different from those he borrowed, except that they seem at times to be more particularly his own, such as certain forceful but prosaic phrases: μιλλῶμαι λόγοις, [130] ἐς τοσοῦτον ἀμαθίας, [131] οὐκ ἐς ἀμβολάς, [132] σὸν ἔργον ἤδηἁ, [133] and the like. But the poet shows his taste in borrowing others as well as in choosing his own words. The best way to show that the repeated phrases within his verses are of the sort we have already seen will be, I think, to take up all the cases of repetition in a certain play. It is of course the later pieces which contain the greater number; the Orestes with 36 has most. They may be classed under five headings.

I. Three phrases of a highly tragic force:

Ὦ Ζεῦ καὶ γᾶ καὶ φῶς. [134] This is one of the rare lyric formulas. It is used by the slave in the Orestes to tell his wonder at the vanishing of Helen. It opens the first chorus of the Medea, serving to deepen the effect of Medea’s lamentation off stage.

Ὥς μ᾽ ἀπώλεσας καὶ τόνδε. [135] In the Orestes Electra speaks thus in her outburst of hate for Helen. In the Phaethon the words are addressed to Helios by the mother as she leaves the stage following the body of her dead son.

Ἐγὼ δ᾽ ἐπεγκέλευσα. Euripides, dealing with the same characters as in an old play, is making second use of a dramatic play of dialogue. In the Electra the words had appeared in the lament between Orestes and his sister after they have slain their mother:

In the Orestes the phrase is spoken when the brother and sister and Pylades, having resolved the death of Helen, call Agamemnon’s spirit to their aid:

The grasping of the sword is here shifted from Electra to Pylades. One can see that Euripides was repeating a complex dramatic grouping of ideas in a different verse-form, though only in the case of three words could he keep the same language.

II. Five phrases which, though not so purely dramatic, contain an idea more than usually striking either in itself or in the way in which it is expressed:

Ὦ τλῆμον Ἑλένη. [138] This is used three times by Euripides. There is no need of explaining “unhappy Helen,” nor the thought of Helen unhappy in the sorrows she has caused, which is that found in two of the passages:

Ἅλις τὸ κείνης (μητρὸς) αἷμα. Orestes speaks thus in the Iphigenia in the Tauric Land and in the Orestes; both when he refuses to endanger the life of one sister for the sake of his escape, and when his {107|108} other sister asks for her death at his hands rather than at those of the Argives. [141]

Ταύτηι γέγηθα καπιλήθομαι κακῶν. [142] The verse is both very pathetic and very nicely put. The same thing may be said of δυστυχοῦντί σοι φίλος. [143]

Τοῖς μὲν λόγοις ηὔφρανε, τοῖς δ᾽ ἔργοισιν οὔ. [144] The verse upon which this was modelled is the following: τοῖς πράγμασιν τέθνηκα, τοῖς δ᾽ ἔργοισιν οὔ. [145] One thinks of the verse of the Hippolytus which Aristophanes ridiculed: ἡ γλῶσσ᾽ ὁμώμοχ᾽, ἡ δὲ φρὴν ἀνώμοτος. [146] Euripides was writing for the same Athens which, nineteen years before, had listened in grave wonder to the balanced style of Gorgias, in whose work we read such sentences as the following one from the Praise of Helen: τὸ γὰρ τοῖς εἰδόσιν ἃ ἴσασι λέγειν πίστιν μὲν ἔχει, τέρψιν δ᾽ οὐ φέρει; [147] and one is not surprised to find in this author’s Defence of Palamedes the same play of words as in the verse quoted from the Orestes: ὑμᾶς δὲ χρὴ μὴ τοῖς λόγοις μᾶλλον ἢ τοῖς ἔργοις προσέχειν τὸν νοῦν. [148]

Οὐκ ἔστιν οὐδὲν κρεῖσσον. [149] This is a phrase of purest gnomic tone.

III. Six phrases found in dialogue, either in stichomythy or at the beginning of a longer speech, always at the beginning of the verse, and with one exception the first words in the speech. They are a device by which a character who has just come upon the stage can begin his speech, or by which the thought can be carried back and forth between actors in the give and take of dramatic conversation:

Ὦ χαῖρε καὶ σύ. This is said in three cases by a character who returns the greeting made him as he enters. [150] In the Hippolytus they are put to a more dramatic use as the words of the dying hero to his father. [151]

Ἔα, τί χρῆμα. [152] This cry of surprise is uttered in three different plays by characters who, entering the stage, have come upon some startling sight. {108|109}

Ὦ φίλτατ᾽, εἰ γὰρ τοῦτο. [153] This emotional expression is used to begin a wish suggested by the previous speaker’s words.

The other three cases are found in stichomythy: κἀγὼ τοιοῦτος, [154] ‘I too’; ἐς ταὐτὸν ἥκεις· καὶ γὰρ οὐδέ, [155] ‘We agree, for …’ ; τί χρῆμα δρῶσαι, [156] ‘What must we do?’

IV. Thirteen phrases in which the diction is more than usually poetic. There are those in which the words themselves are of the sort not used in prose:

In others it is the way the words are used which is poetic:

In two cases the expression, by calling up the legend, brings into the style what may be called a romantic note:

These last two examples might have been put in the second class. {109|110}

V. There remain of the 36 repeated phrases nine which seem hardly striking enough in themselves to have been used for much else than their usefulness, though here too there are poetic words and word-order, and forceful ideas:

It is to this extent of nine short expressions in a play of almost sixteen hundred lines that Euripides used what would seem to be more or less true formulas. In all but one of these cases he was repeating himself for the first time, and in four of them he was repeating words which had been used earlier in the same play, but had not yet faded from his mind, which would be likely to hold for any length of time only the most remarkable. Moreover four of the six expressions whose first and second appearance occur within the play fall in this fifth class: the poet could not well use any very noticeable phrase twice in the same drama. Yet there remains a final reason why even these cases should not be classed unhesitatingly as formulas: there are quite as many repeated expressions in the prose of most writers, where the factor of the verse, essential to the formula, plays no part. Even here it is doubtful how often Euripides was guided by any other motive than that of the prose author who uses his words over purely for the sake of their thought or their fitness.

It is clear from this analysis that one would not be far wrong in saying that the formula does not exist in tragedy. The dramatic poet, working at ease into the mould of his verse those words he carefully chose for his very own thought, used from time to time some idea or poetic expression which had proved effective in the past and which he {110|111} remembered for that reason. Since he was using a regularly recurring verse-form, the metre did not prevent him, as it had prevented Pindar, from using the words which had already expressed the idea, or had even given it some of its value. In one sense the verse-form has influenced his style, in that it did not keep him from repetition, though it did not push him on to it as it had Homer. But in this last distinction lies all the difference between a traditional and an individual style.

We have learned the nature of the repetitions in tragedy. Looking now to see whether there are any traces of schematization in the style of the tragic poets we find a very few, such as the following:

The systems in Euripides are always made up as here of two or three expressions. There is hardly need of pointing out that they are of the same sort as the phrases repeated without change. The poet is usually modifying some striking idea or some forceful use of words to fit a new situation. One could rarely say that he was guided in any way by the wish for an easy versification. Yet this, we shall see, was the regular motive for the epic poet.

Since there are no systems of any length in tragic diction, there is, of course, no question of the thrift of the system. The lack of thrift in the diction is made clear, however, by the large number of expressions which could replace one another, that is to say, expressions in which the essential meaning and the metrical value are the same, but the words different. Such examples are the following:

Equivalent phrases of this kind are not lacking in Homer, but they are always due to the play of analogy which underlies the diction, and they are never phrases of more than a few words. [209] In tragedy, however, whole equivalent verses are very common, and we shall do well to consider some of them for what they show us of the working of the poet’s mind. Aeschylus wrote the following verse:

In Euripides this became:

We read in Sophocles:

This same idea had been expressed before by Aeschylus in a trochaic verse:

Rather than use the line of Sophocles, Euripides blended the two verses, and made:

Finally, the scholiast on this verse quotes another with identical thought from some unnamed poet:

Euripides varies the terms of his own statement:

Possibly the most curious case of equivalent verses is to be found in the following passages, one from the Helen, the other from the Iphigenia in the Tauric Land:

Not only do these equivalent verses show the lack in the poetry of any factor which would have urged the writer to a thrift of diction; they show clearly how the idea could lie in the mind of the poet without being bound to any certain words. Euripides, when he made verses, looked for terms to express his ideas, but the epic poet, we shall see, thought in terms of his formulas, and did not separate the idea from the words with which it went. It is not the place here to show this fully, but in passing I would quote certain Homeric lines and ask if one should not be much surprised to find the same ideas expressed in verses of different wording:

The first of these verses occurs 21 times, the second 17 times.

Leaving tragedy to go on to later poetry one sees straightway that we have very little to learn about the formula outside the epic itself. We may even be charged with having followed thus far a very laborious course to prove what is clear enough anyway, namely, that the repeated phrase in poetry, unless it be poetry very different from our own, is an ornament of verse, not a means of making it. The repetitions in later verse have long been put to their proper use as a ready means of studying the influence of one author upon another, either upon his thought or upon some aspect of his style. As far as our understanding of Pindar or of Euripides goes, there is almost no value in the distinction we have been making between the phrase which, taken without change from another poet, might be helpful, and that which he took and changed for his own needs. The influence of ideas shown by the borrowed phrase is very real, but no more so than is that of the altered phrase, while its metrical help is too small to deserve note. {114|115} More than that, unless we consider the repetitions as showing this influence of ideas, we cannot know why they come more or less closely together in the verses of a poet. We took care to see just how many repetitions there were in tragedy, supposing that the exact difference in number between those in Attic poetry and those in Homer would have some bearing upon the problem of the formulas in the Iliad and in the Odyssey. But the truth is that only the absolute difference thus proved has any bearing on Homer’s practice. The contrast between a vast number of repetitions in Homer and a comparatively very small number in the work of the tragic poets at once suggested that repetition could not be due to the same causes in both cases. Then a study of the nature of the repetitions in tragedy showed that almost none of them, or even none of them at all, are true formulas, and so we reached that important point where we know surely that Homer’s poetry is governed by factors unknown to later Greek poetry. Just what those factors are we shall go on to see; yet our essay would not be complete if we did not pause here a moment to point out what sort of causes, special to a certain poet, or to the poetry of a certain period, have determined the frequency of repetitions in poetry outside the early epic.

If we find almost no Homeric formulas in Apollonius, for example, it does not at all mean that they would not have helped his verse-making, but that he wanted very much to avoid them. [221] If Theocritus, on the other hand, used twelve in his little epic The Infant Heracles, it means that he was seeking, in a rather amusing way, for the epic note, and that the use of the twelve formulas was in no way different from that of the ten which he changed metrically. [222] Indeed, one misses the point if one does not see the pains which the poet has spent upon them. Likewise the three hundred phrases of his own which, as we just saw, Euripides used over again, have nothing at all to do with the number of formulas which Homer might have made himself, but they do show the high point to which Euripides carried the artifices of his dramatic technique, and, if one will study them along with the other {115|116} cases in which he expresses the same ideas but changes his words, one will find them an excellent means of learning the nature of his tragic art. The number of repeated phrases in Virgil is high. The lists of E. Albrecht [223] give 372 cases from the Aeneid, as follows: 44 in which the expression had been used in the Eclogues or the Georgics; 248 in which the expression appears three times within the Aeneid; 47 in which it is found four times; 22 in which it is found five times; and 11 in which it is used even more often. This makes, not counting the first appearance, a repetition for every twentieth verse. The number is high, but it shows only two things. The phrases from the Eclogues and Georgics are a measure of the endless care which Virgil gave to his style, and of the need he felt of using again some of the best expressions he had made in his earlier years. The repetitions found in the Aeneid show this also, but far more they show that he was trying to make a poem like Homer’s. As in the case of the Homeric formulas in Theocritus we must see, if we are to understand the poet, how much toil these repetitions must have cost him. When we turn to our own poetry and to our own language, the nature of the borrowed phrase in written poetry becomes very clear. Charles Crawford, [224] after writing a concordance to the works of Kyd and the play Arden of Feversham, whose authorship is doubtful, was able to show 47 places where that play recalls Kyd’s Soliman and Perseda; of these there are thirteen where the same phrase is used, though never more than once. But it will do to quote a few of the repetitions: “to everlasting night,” “leave protestations now,’’ “vengeance light on me,” “this melancholy mood.” I give one more example because it should make clear how little effect the loose form of Elizabethan blank verse could have had upon the choice of a certain word-group. We read in Soliman and Perseda:

In Arden of Feversham we find:

To find repetitions which could be said to help the verse-making one must rather go to the tradition of Milton’s style. Here the strictness of the verse, and the demand for form in style, come much nearer to the practice of the Greek and Roman poets. Yet when one finds Pope copying “the glowing violet,” or “rough satyrs danced,” or “tufted trees,” or “dropt with gold,” one sees the utter vainness of thinking one will find a true formula in the remaining 51 pages of parallels to Milton which R. D. Havens collected from English verse. [227]

It would seem, indeed, that those who wished to show that Homer, in his formulas, was not really different from any other poet, were not altogether logical. Thinking that the use of formulas as a means of easy verse-making might damage Homer’s good name, they cited the examples of repetition in later verse, where they themselves would be the first to deny any but purely artistic motives. It may be, though, that Homer was not such a bad poet even if he did make verses in a way which some have found not quite right.

The easiest and best way of showing the place the formula holds in Homeric style will be to point out all the expressions occurring in a given passage which are found elsewhere in the Iliad or the Odyssey, in such a way that, as one reads, one may see how the poet has used them to express his thought. I have put a solid line beneath those word-groups which are found elsewhere in the poems unchanged, and a broken line under phrases which are of the same type as others. In this case I have limited the type to include only those in which not only the metre and the parts of speech are the same, but in which also at least one important word or group of words is identical, as in the first example: μῆνιν … Πηληιάδεω Ἀχιλῆος and μῆνιν … ἑκατηβόλου Ἀπόλλωνος. {117|118}

ΙΛΙΑΔΟΣ Α

ΟΔΥΣΣΕΙΑΣ Α

The expressions in the first twenty-five lines of the Iliad which are solidly underlined as being found unchanged elsewhere in Homer count up to 29, those in the passage from the Odyssey to 34. More than one out of every four of these is found again in eight or more places, whereas in all Euripides there was only one phrase which went so far as to appear seven times. If we had chosen our verses from the end of the Odyssey, one could not possibly have objected that the twelve expressions in the passage from the Iliad, which are repeated only once, are perhaps being used there for the first time. But there is no real need of judging this point. Without these expressions the difference between the repetitions in Homer and those in the work of later poets is very great; but more than that, we are looking for the difference not in repeated phrases but in formulas.

It is important at this point to remember that the formula in Homer is not necessarily a repetition, just as the repetitions of tragedy are not necessarily formulas. It is the nature of an expression which makes of it a formula, whereas its use a second time in Homer depends largely upon the hazard which led a poet, or a group of poets, to use it more than once in two given poems of a limited length. We are taking up the problem of the Homeric formulas from the side of the repetitions, but only because it is easier to recognize a formula if we find it used a second or third time, since we can then show more easily that it is used regularly, and that it helps the poet in his verse-making. We have found that formulas are to all purposes altogether lacking in verse which we know was written, and we are now undertaking the first step in showing the particular character of Homeric style, which is to prove that Homer’s verse, on the contrary, has many. We are establishing the difference between many formulas and none. But when that is done we shall still be left to decide the nature of the Homeric diction as a whole.

It is straightway clear that only a very few of the repeated expressions which are underlined in the two passages have anything either in thought or in style which could possibly set them apart in the poet’s mind as particular devices for making his verses effective. There is nothing unusual about ἀλλ᾽ ὅτε δή (107 times), γαῖαν ἱκέσθαι (18 times), οἶκόνδε νέεσθαι (10 times), ἔνθ᾽ ἄλλοι μὲν πάντες (10 times), nor in ἥνδανε {122|123} θυμῶι (4 times), κρατερὸν δ᾽ ἐπὶ μῦθον ἔτελλεν (3 times), ἔχων ἐν χερσίν (3 times). Nor are the expressions made up of a noun and an epithet or a patronymic more noticeable than those just quoted, though some may have wrongly thought them so. Πηληιάδεω Ἀχιλῆος might seem to one, who has not read much Homer, fully as forceful as the phrase Ὀρέστην παῖδα τὸν Ἀγαμέμνονος, which Euripides, we said, [314] used to call up the legend, and this is the way that students just beginning the Iliad in Greek read the word. But when one has read the two poems, and has met the expression seven times more, usually in a context which gives us not the least idea of why Homer wished to mention the father of Achilles, one becomes indifferent to the patronymic, and ceases to look for a special meaning in its use. Besides that, one has found Λαερτιάδεω Ὀδυσῆος eleven times, Ἀγαμέμνονος Ἀτρείδαο thirteen times, Νέστορ Νηληιάδη seven times, Ἕκτωρ Πριαμίδης eight times, Τυδέος υἱός thirty times, Ἀτρέος υἱός eleven times, not to speak of Πηλείδης or Πηλείων, patronymics which are used 93 times in the place of the hero’s name. After that it is very hard to remember, each time one begins the Iliad, to find in Πηληιάδεω the meaning which one gave it in its newness. What has just been said of the patronymic is likewise true of the epithet ἄναξ ἀνδρῶν, which is used not only 48 times elsewhere in the Iliad and the Odyssey of Agamemnon, but also of Anchises, Aeneas, Augeias, Euphetes, and Eumelos, none of whom is any more a king of men than is any other of the chief heroes; but all of them have names of the same metrical value as that of Agamemnon. Δῖος, the epithet of Achilles, appears again in A 145 with the name of Odysseus. Even here the beginner in Homer may still believe that these two heroes share the honor of being “divine,” whatever that may mean. But when he has found the word used for Nestor (B 57), for Agamemnon (B 221), for Paris (Γ 329), and, before he has finished reading the two poems, for thirty-two different characters, many of them of no very great legend, and when he has met it as an epithet of some noun once in about every seventy verses, he at length forms the habit of scarcely heeding the word as he reads. Finally, if he ever found a sinister meaning in ἑκηβόλου, the epithet of Apollo in A 14, he will have to make very much of an effort to find it again after he has heard the god called by that word, or by ἑκατηβελέταο, {123|124} ἑκάτοιο, ἑκάεργος, or ἑκατηβόλου, in twenty-nine other places. The fixed epithet in Homer is purely ornamental. It has been used with its noun until it has become fused with it into what is no more, so far as the essential idea goes, than another metrical form of the name. The reader knows the epithet and likes it, but it is the liking for what is familiar. He would be surprised if in a given passage the epithets were lacking, or were missing in certain known phrases; but when he does meet them he passes over them, scarcely heeding their meaning. The noun-epithet expressions are thus no more striking, if read rightly, than any other part of Homer’s diction. [315] The case of ἐπεκλώσαντο θεοί in α 17, with its metaphor of spinning, is similar. The same verb is used in seven other places in the poems of the lot assigned to a man, and like ἕρκος ὀδόντων, ‘the barrier of the teeth’ (10 times), πολέμοιο γεφύρας ‘the bridges of war’ (5 times), or ὀλέθρου πείρατ᾽ ἐφῆπται ‘the cords of ruin are fastened’ (4 times), and the fairly numerous other Homeric metaphors, its newness must have been lost long before Homer used it. This does not mean that the poetry has suffered either here or in the case of the fixed epithets; it is only a short-sighted judgment which would think of that. It means simply that the expression has found its place in the even level of this perfect narrative style, where no phrase, by its wording, stands out by itself to seize the attention of the hearers, and so stop the rapid movement of the thought, or, if one wishes, where every phrase has its perfection of style, so that the evenness of the diction comes not from its lack of what is striking, but {124|125} from its lack of any phrase which has not been accepted finally as the one best means for stating the idea.

It does not follow, because the style of Homer is even, that all the ideas of his poetry are equally forceful. There is hardly any need of pointing out the varying intensity of the thought within the fifty lines we are studying. Yet that intensity, where it appears, usually comes from the thought of the passage at that point, rather than from any certain expression. Very often, as one reads, the thought of some group of words will stand out, but it is usually the way in which they are used that makes them do so. The line which Homer uses in A 33 does not seem notable as one reads:

But when it appears again, in the scene between Priam and Achilles (Ω 571), it becomes one of the very pathetic verses in Homer. The words ἔνθ᾽ ἄλλοι μὲν πάντες in α 11 bring us to the moment when the Odyssey opens, and to the situation with which the poem begins, and does so with an ease which leaves us wondering; in A 22 this same expression is used for a more ordinary transition. Likewise the half-verse Διὸς δ᾽ ἐτελείετο βουλή is highly forceful in the prologue of the Iliad, but in λ 297, where it concludes Melampus’s adventure with the cattle of Iphicles, it is in no way remarkable. Besides this last repeated phrase there are six others, of the 63 found in the fifty verses we are considering, which express what seem to be more than ordinarily effective ideas: οὐλομένην ἥ (A 2), ἰφθίμους ψυχάς (A 3), ’Άιδι προίαψεν (A 3), ἔννεπε Μούσα (α Ι), πολύτροπον ὅς (α Ι), νήπιοι οἵ (α 8). Ἴφθιμος and πολύτροπος, it should be noted, are not ornamental epithets, but are used as an essential part of the thought. [316] It is then only to this extent of one out of every nine or ten that the repeated phrases of Homer are in any way like those which are found in later verse.

Having shown that the repeated phrases in Homer are only for a very small part to be classed as striking phrases, we must now go on to see if they are useful, since utility was a quality of the formula as we defined it. {125|126}

The technique of the formulas in Homer is vastly complex, but its general principle can be stated briefly. The Singers found and kept those expressions which without change, or with slight change, fall into that part of the hexameter which is determined by the rοle they play in the sentence. Since the problem of the poet is not only that of making a verse of six dactylic feet, but of fitting his words between the pauses within the verse, the formulas which express the most common ideas fall exactly between one pause in the verse and another, or between a pause and one of the verse-ends. The ways in which these formulas fit into the parts of the verse and join on to one another to make the sentence and the hexameter are very many, and vary for each type of formula. A full description of the technique is not to be thought of, since its complexity, which is exactly that of the ideas in Homer, is altogether too great. One must either limit oneself to a certain category of formulas, and describe their more frequent uses, as I have done in my study of the noun-epithet formulas, or one must take a certain number of formulas of different sorts which can be considered typical. We must choose the latter course. I shall thus consider the metrical usefulness of the first five repeated expressions which appear in the Iliad and in the Odyssey.

I. Πηλιμάδεω Ἀχιλῆος is one of a series of noun-epithet formulas, in the genitive, for gods and heroes: Λαερτιάδεω Ὀδυσῆος, πατρὸς Διὸς αἰγιόχοιο, Διομήδεος ἱπποδάμοιο, Μενελάου κυδαλίμοιο, ἀγαπήνορος Ἰδoμεvῆoς, Ἀγαμέμνονος Ἀτρείδαο, Θηβαίου Τερεσίαο, ῾Υπερήνορος ἱπποδάμοιο, and so on. The usefulness of these formulas lies in the fact that they can finish with the verse a clause which is complete but for the genitive of a character’s name; or that, if they do not finish it, they can bring the poet at any rate to the beginning of the next line where he can use formulas which regularly begin the verse. In the two cases where ἑκατηβόλου Ἀπόλλωνος, a formula of this type, is used with μῆνιν, the sentence ends with the verse:

In λ 387 another formula of the same type brings the clause as far as the verse-end, so that the poet can use the common device of beginning the next verse with a middle form of the participle:

The formula in the first line of the Iliad renders a like service. It would be very hard at the best to put οὐλομένην ἥ in any other place in the verse than at the beginning, where it is found all of the five times it is used. [317]

II. Very common in Homer is the device of using an adjective followed by a relative clause to continue a sentence which might have come to an end with the preceding verse. This type of enjambement is found four times in the first hundred verses of the Odyssey: νήπιοι οἵ (v. 8), δυσμόρωι ὅς (v. 49), μακράs αἵ (v. 54), ἀμβρόσια χρύσεια τά (v. 97). [318] One should compare with A 2 two passages from the Odyssey:

The verse which is repeated in these two cases differs in meaning from A 2 only in the word ἀνθρώποις. The formula which follows the trithe-mimeral cæsura in E 876:

is of the same type as that found in ε 67:

Indeed the play of formulas in this device of the appositive adjective extending to the middle of the second foot, followed by a relative clause which finishes the verse, seems so easy that one is tempted to make verses for oneself. Thus the line ξ 289:

III. Ἄλγε᾽ ἔθηκεν in A 2 is of the very common type of formula which is made up of a verb and its direct object and falls after the bucolic diaeresis. To give only a few of the formulas which are directly like it there are τεύχε᾽ ἔθηκε, εὖνιν ἔθηκε, κῦδος ἔθηκε, on the one hand, and on the other ἄλγε᾽ ἔδωκε, ἄλγε᾽ ἔπασχον, ἄλγε᾽ ἔχουσιν, and the like.

The uses of these formulas are much more varied than those of the noun-epithet formulas which serve to expand the simple name to a certain length, or than those of the longer types of formulas, such as the one just discussed, which make up an entire clause. For a shorter group of words such as ἄλγε᾽ ἔθηκε expresses an idea which will be used with many kinds of formulas to make many different sentences. This formula thus belongs to the less obvious part of the technique; yet it would be false to suppose that it is any less helpful to the poet than the longer ones: it is chiefly in the formulas of these shorter types that lie the suppleness and the range of the diction, and their usefulness is to be measured by the many different kinds of other short formulas with which they combine, as in Κρονίδης Zεὺς ἄλγε᾽ ἔδωκεν (three times), ἀλώμενος ἄλγε᾽ πάσχων (twice), and so on. A 2, however, is not the only verse where this type of formula is preceded by a dative. We find ἐπ᾽ αὐτῶι κῦδος ἔθηκε (twice), and τῶι δευτέρωι ἵππον ἔθηκε. In X 422 ἄλγε᾽ ἔθηκε follows περὶ πάντων. This last expression falls very often before the bucolic diæresis: περὶ πάντων ἔμμεναι ἄλλων, περὶ πάντων τῖον ἑταίρων and the like. It should be noted that the poet has another formula of the same meaning as ἄλγε᾽ ἔθηκε, but beginning with a single consonant, so that it can be used after a final vowel: κήδε᾽ ἔθηκε (Φ 525, ψ 306). Such pairs of formulas are frequent: ἄλγεα θυμῶι and κήδεα θυμῶι, αἴσιμον ἦμαρ and νηλεὲς ἦμαρ, εὖχος ἀρέσθαι and κῦδος ἀρέσθαι, and so on.

IV. In Α 3 we have a formula which, but for the change of a word, fills a whole line and is itself a complete sentence. Verses of this kind are outdone in usefulness only by those used unchanged, and one would have such a line here if one wished to adopt ψυχάς, the variant reading to Λ 55, for κεφαλάς. But the difference after all is very slight, as one can judge by the many other verses in which the poet has shifted only aword, or two words such as:

V. Διὸς δ’ ἐτελείετο βουλή in A 5 is of the type of numerous other formulas which form a complete sentence in one half of the verse. It happens that the words which precede it here have no direct parallel, but θέσφατα πάντ’ εἰπόντα which goes before it in λ 297 is like ἀγγελίην εἰπόντα in π 467. The formula appears in an altered form in Διὸς δ’ ἐξείρετο βουλήν (twice). Other formulas which are used in this way as half verses are κρατερόν δ’ ἐπὶ μῦθον ἔτελλε (four times), παλάμηι δ᾽ ἔχε χάλκεον ἔγχος (twice), νεμεσσήθη δ’ ἐνὶ θυμῶι (three times), and so on.

VΙ. Μοι ἔννεπε Μοῦσα is one of the rare cases of a formula of any length which is found in more than one place in the verse. One can see how its place was determined by the play of the other formulas which have taken up their regular position in the line. It appears also in B 761:

The beginning of the verse is that of A 8, which is likewise addressed to the Muse. Ὄχ᾽ ἄριστος and the related μέγ᾽ ἄριστος fall regularly at this place, and ὄχ᾽ ἄριστος ἔην is found three times. In the first verse of the Odyssey μοι ἔννεπε Μοῦσα falls before πολύτροπος ὅς which begins a series of formulas each of which has its fixed position.

VII. Πολύτροπος ὅς, the first of these, appears again in κ 330:

Ὅν τέ μοι αἰεί appears in six other places at the verse-end. It is one of a numerous class of formulas made up of relative words, particles, pronouns, and adverbs, which begin a clause of which the principal words will be found in the next line. Examples are εἴ ποτε δὴ αὖτε, εἴ ποτε δή τι, οὐδέ νυ σοί περ, καί ἑ μάλιστα, and the like. In the verses just quoted the formula of this sort leads up to φάσκεν ἐλεύσεσθαι, which is of the same type as φῆισιν ἐλεύσεσθαι (α 168). In α Ι we find ὃς μάλα πολλά followed by πλάγχθη, which brings the sentence to the end of the clause. A like use of πλάζομαι, as a run-over word, occurs in ε 389:

The use of a simple verb at the beginning of the verse, measured – ⏑ and followed by ἐπεί, is found, for example, in σ 174: ἔρχευ ἐπεὶ … This brings us by an unbroken chain of formulas to our next case. {129|130}

VIII. Τροίης ἱερόν reappears in Π 100:

The line, after the first foot and a half, is no more than a variation of α 2, made necessary by the fact that λύωμεν, beginning with a consonant, cannot be joined to ἱερὸν πτολίεθρον. There is yet another variation of the verse in ν 388 where the metrical value of the verb does not allow it to be placed at the verse-end:

IX· In κ 458 the poet used the fixed epithet of πόντος rather than the intensifying ὃν κατὰ θυμόν which is found in α 4:

Πάθεν ἄλγεα ὃv κατὰ θυμόν is one of a long series of formulas, all of which express in the different persons, numbers, moods, tenses, and cases of the participle the essential idea ‘to suffer woes,’ but each of which has its unique metrical value. A list of the formulas of this kind which fall at the end of the verse will give us some idea of the extent to which Homer had a formula for each metrical need: [319]

After the fifth foot and a half:

πάθεν ἄλγεα (twice)

After the fourth foot:

ἄλγεα πάσχει (10 times)

ἄλγε᾽ ἔχουσιν (4 times)

πήματα πάσχει (7 times)

πῆμα πάθηισιν (3 times)

After the third foot and a half:

πάθον ἄλγεα θυμῶι (6 times)

χαλέπ᾽ ἄλγεα πάσχηι

κρατέρ᾽ ἄλγεα πάσχει (4 times)

κρατέρ᾽ ἄλγε᾽ ἔχοντα (4 times)

κακὰ κήδε ἔχουσιν

κακὰ πολλὰ παθόντα (4 times)

κακὰ πολλὰ μογήσας (4 times)

After the third foot: {130|131}

ἄλγεα πολλὰ μογήσας (4 times)

ἄλγεα πολλὰ πάθοιμεν

After the trochaic caesura of the third foot:

πάθ᾽ ἄλγεα ὃν κατὰ θυμόν

ἔχοντί περ ἄλγεα θυμῶι (twice)

κακῶς πάσχοντος ἐμεῖο

ὀιζύομεν κακὰ πολλά

After the second foot and a half:

πάθεν ἄλγεα ὃν κατὰ θυμόν

χαλεπὸν δέ τοι ἔσσεται ἄλγοςThe help given the poet by these formulas is that each of them completes his verse, leaving him free to continue his thought by the formulas that begin the verse. One should not judge from this that the technique of formulas aims altogether at bringing the thought to a close at the end of the verse. It does do this often enough to bring it about that the thought comes to end in three out of every four verses in Homer, whereas in Apollonius and in Virgil it does so in only two out of every four verses. [320] But the technique also has its formulas which run the clause over into the following line. We have just studied in the case of πολύτροπον ὃς μάλα πολλά and πολύτροπος ὅν τέ μοι αἰεί two formulas of this kind.

X. ἱέμενός περ is of the type of a large number of formulas: ἀχνύμενός περ, κηδομένη περ, ἐσσύμενόν περ, γιγνόμενόν περ, οὐτάμενοί περ, τειρόμενοί περ, and so on. These formulas can be joined on to the large number of clauses whose thought is brought to a point of completion at the bucolic diæresis, where they usually end with a verb, for the fourth foot of the hexameter is very well suited to the verb by its measure and its position. [321] One may see this in τ 324, which is a verse very much like α 6:

The verse ξ 142:

should be compared with X 424:

We have now found that there are formulas in Homer, one at least to every verse or so, for we have seen that the repeated expressions in the Iliad and the Odyssey are really formulas. They express only for a small part ideas which are more than usually striking, and they form a part of a highly developed technique for making hexameters. What we have done then is to prove that the style of Homer, so far as the repeated expressions go, is altogether unlike that of any verse which we know was written.

But we have also seen a difference between Homer and the later poets which is not confined to the repetitions. We found only the slightest traces of schematization in the diction of Euripides, but we have had it continually before us in our study of Homer. First, we have had one measure of it in the simple number of the repetitions, and in the large number of times many of them appear. For, as was said at the beginning of our search, whenever a poet uses his words over, he is limiting his thought to a fixed pattern on the one hand, and on the other he is casting aside all the other possible ways in which the idea could be expressed. Secondly, we have seen, by the broken line used in the passages taken from the Iliad and the Odyssey, the large number of expressions which, though not repeated, are related to others of the same type. Finally, in showing the usefulness of the repeated expressions we did nothing more than find the systems of formulas to which they belonged. One must not overlook this fact that the schematization of the diction is always due to the fact that the poet is using, to express an idea, the same device which he had used to express one more or less like it. The role played by analogy as a guide to the poet in his choice of terms is one which, we shall see, can be fully understood only when one sees the relation between the play of sound and the thought of the poet, but at no moment should one forget that the use of like formulas is a direct means of overcoming the difficulty of expressing ideas in hexameters.

The systems which were given to show the utility of the repeated expressions in Homer were often made up of phrases found only once in the two poems. That these expressions were formulas, however, was clear. We could not observe them in different places, and thus prove their regular usage, but we saw that they belong to particular {132|133} artifices of versification which have a fixed place in the diction. We have thus brought into the category of formulas not only the repeated expressions, but those which are of the same type as others. In the two passages analyzed above I marked with a broken line only those formulas which were like others in rhythm, in parts of speech, and in one important word; but there are more general types of formulas, and one could make no greater mistake than to limit the formulaic element to what is underlined. Γιγνώσκω σε θεά in E 815 is like μῆνιν ἄειδε θεά in A Ι, since in both cases one has a complete clause of the same length, followed by the vocative θεά. The similarity between Ἀχαιοῖς ἄλγε᾽ ἔθηκε and ἐπ᾽ αὐτῶι κῦδος ἔθηκε has been noted. Πολλὰς δ᾽ ἰφθίμους ψυχάς in Α 3 is an accusative phrase of the same length as πολλὰς δὲ δρῦς ἀζαλέας (A 494), and πολλὰς δὲ στίχας ἡρώων (Ψ 326). If one excepts the change from accusative to nominative, the formula ψυχὰς Ἄιδι προίαψεν is paralleled by ψυχαὶ δ᾽ Ἄιδόσδε κατῆλθον (Η 330, κ 560, λ 65). The use of ἡρώων at the beginning of a verse, followed by a new clause, appears in I 525, and Ἀργείων is often used in the same way. Τεύχε κύνεσσιν is like δῶκεν ἑταίρωι (P 698, Ψ 612). Often one finds the same verse-pattern where the words are different.

Even in the very limited amount of poetry in which we are searching for like expressions there are, with the exception of those phrases used more or less often to express some special idea, as, for example, ἐπ᾽ ἑτοῖμα προκείμενα quoted above, [322] very few which do not a into some closer or some more general system; and one must never forget that the results of any analysis of this sort are conditioned by the hazard that has given us under the name of Homer not quite twenty-eight thousand verses. If we had a greater or a smaller number, we should have underlined either more or fewer expressions when we analyzed the first verses of the Iliad and of the Odyssey. If we had even twice as much of Homer’s poetry as we have, the proportions {133|134} between the repeated expressions, the closer types of formulas, and the more general types, would be much changed, and we should very often find that Homer was using a formula a second time where, as far as our evidence goes, he is only using a formula which is like another. But as it is we have verses enough to show us the vast difference between the style of Homer and that of poetry which we know was written: we have found that the schematization, of which there were only the faintest traces in later poetry, reaches almost everywhere, if not everywhere, in the diction of the Iliad and the Odyssey.

Having shown this difference we must now look for its causes. Did this style of Homer’s come into being through one poet or many, in a short time or over many years? And why should Homer have wished to use a formulaic diction?

The direct proof that the style of the Iliad and the Odyssey is traditional is, of course, the schematization of the diction itself, and the number of artifices of verse-making which make up this schematization. It is not possible, for example, that one man by himself could work out more than the smallest part of the series of formulas of the type Πηληιάδεω Ἀχιλῆος. We may make ourselves believe that the one poet who composed both the Iliad and the Odyssey first used οὐλομένην ἥ, which is found three times in the first poem and as often in the other, yet we cannot go on endlessly adding νήπιοι οἵ (5 times), δύσμορος ὅς (6 times), σχέτλιος ὅς (4 times), νηλεὲς ὅς (Π 204), and so on. One cannot grant the same poet ἱέμενός περ (10 times), ἀχνύμενός περ (13 times), κηδομένη περ (11 times), οὐτάμενοί περ (4 times), and yet more. Virgil, striving to do as Homer, was able to repeat in the Aeneid 92 verses. How many of the 1804 repeated verses in the Iliad and the Odyssey can we then give to one poet, for whom we shall have to find I know not what reasons to repeat himself, since he could scarcely have had those which led Virgil to do so? Finally, how could one man even have made a beginning of the technique of the diction as a whole in which the various types of formulas accord with one another so well? Indeed, the more one studies the formulas in Homer and the artifices of their use, the more one sees what efforts have gone to their making. One may well say that the single series of formulas πάθεν ἄλγεα, ἄλγεα {134|135} πάσχει, and so on, is by itself alone far beyond the power of any one man. For the formulas are not only too ingenious to be the work of the one poet of the Iliad and the Odyssey, they are also too good. The epithets, the metaphorical expressions, the phrases for the binding of clauses, the formulas for running the sentence over from one verse into another, the grouping of words and phrases within the clause and within the verse, all this is many times beyond whatever supreme creative genius for words one could imagine for the poet Homer.

Moreover, we know that the Homeric diction was centuries making. The linguists have shown us that the language of the Homeric poems, which was once given the mistaken name of Old Ionic, is an artificial language, made up of words and forms taken from the current Ionic, from Aeolic, even from Arcado-Cyprian dialects; and along with these are artificial forms which could never have existed in the speech of any people. [323] The epic poets kept the older or foreign forms and words, and adopted or created new ones, in order to have a language which would suit the hexameter. The scholars who have thus finally given us the answer to the ancient question of Homer’s dialect have, however, not seen clearly enough that the survival of the older forms is due not to their metrical value alone, but also to the fact that they occur in traditional formulas. There is no reason, for example, why the Aeolic prefix ἐρι- should not have been changed to the Ionic ἀρι-. [324] Yet beside ἀριδείκετε λαῶν (6 times), and ἀριδείκετος ἀνδρῶν (twice), and so on, we find: ἐρίηρες ἑταῖροι (20 times), ἐρίγδουπος πόσις Ἥρης (7 times), ἐριαύχενες ἵπποι (5 times), Τροίην ἐριβώλακα (5 times), Θρήκης ἐριβώλακος (twice), Τάρνης ἐριβώλακος and so on. Since the presence of ἀρι- in the poems shows that the epic poets were not consciously archaizing in their use of the Aeolic prefix—and they archaize knowingly only when the metre leads them to it—we know that the series of formulas just given goes back to a time before the Ionians had learned the traditional formulaic style from the Aeolic Greeks. More {135|136} usually, however, the older forms are kept because the epic poets would otherwise have had to give up the formula altogether. The presence in the Iliad and in the Odyssey of Κρόνου παῖς ἀγκυλομήτεω (8 times), and θεοὶ δόσαν ἀγλαὰ δῶρα (4 times), warns us not to seek to change Πηληιάδεω Ἀχιλῆος (8 times), with its Ionic ending, to Πηληιάδα᾽ Ἀχιλῆος, though there is no doubt that the Aeolic poets used it thus. But in μητίετα Zεύς (19 times), [325] Διὸς αἰγιόχοιο (19 times), Διὸς νεφεληγερέταο (6 times), μελαινάων ἀπὸ νηῶν (7 times), μακάρεσσι θεοῖσι (7 times), the Ionic poets had to keep the Aeolic endings or lose the formulas. Likewise we have Aeolic ἄλλυδις in ἄλλυδις ἄλληι (8 times), περιπλομένων ἐνιαυτῶν (3 times) beside its Ionic form περιτελλομένων ἐνιαυτῶν (3 times), ἀνὰ πτολέμοιο γεφύρας (3 times) beside ὁμοιίου πολέμοιο (8 times), which falls at the verse-end where the poetry did not allow a syllable to contain a long vowel and be followed at the same time by two consonants. [326] It is certain in this last case that the formula represents an older ὁμοιίοο πτολέμοιο. But we also have στυγέρου πολέμοιο (twice) which must remain. [327] Κακομηχάνου ὀκρυοέσσης (Ζ 344), however, can only stand for an earlier κακομηχάνοο κρυοέσσης. The epic poets preserved the formula by creating the strange but easily understood ὀκρυοέσσης. But the antiquity of certain parts of the formulaic diction goes back even before the time when the Aeolic Greeks either learned this diction for the first time, or fused the lays of another Greek people with their own. We find in Homer a number of words which, in historical times, occur only in the dialects spoken in Arcadia or in the island of Cyprus: αἶσα, ἆμαρ, ϝάναξ, ἀνώγω, εὐχολά, κέλευθος, οἶϝος, and so on. [328] To these Arcado-Cyprian dialects has been given the name of Achaean, since it would seem that they are the remnant of the language spoken by those Greeks who were powerful in Greece {136|137} and the Aegean before the Dorians came. If E. Forrer’s translations of the Hittite tablets found at Boghaz-Keui is correct, an Achaean chief of the thirteenth century went by the title of κοίρανος. [329] There is of course no need to suppose that all the formulas in Homer which contain Achaean words go back to a time before the coming of the Dorians, since it is very possible that the later poets may have used one of the old words in a new formula, but in many cases it is easier to accept the antiquity of the formula than explain it by such a hazard. Αἴσιμον ἦμαρ (4 times), for example, is made up of two such Achaean words: ἤματα πάντα (27 times) appears in fifth-century prose inscriptions from Mantinea and Tegea. One can only guess at the age of νηλεής, in which appears the prefix νη- which had disappeared from spoken Greek before the historical period: this word is found nine times in νηλεὲς ἦμαρ. Most important, perhaps, of all the Achaean words is αὐτάρ, found only in Cyprian. The use of this word, of which we have given one of the systems above, [330] is so far-reaching in Homeric style that we must either accept the high antiquity of many of the most common phrases for joining clauses in the hexameter, or say that the later Greeks just happened to sieze upon what was to them no more than a helpful poetic word to use in many of their most common formulaic devices. It is hard to believe in such a curious chance. Finally, the age of certain parts of the diction, as well as of the form of the hexameter, is shown by the great number of noun-epithet formulas in which the meaning of the epithet has been lost to us, as it must have to Homer also, for otherwise we must suppose a rapidity of change in spoken Ionic which would be without a linguistic parallel. [331]

It may still seem to some at this point that the schematization of the diction, and the age of parts of it, prove that most of Homer’s style is traditional, but still leave room for the creation of phrases by the single poet. One could answer simply that the expressions created within the systems, by following the fixed types, would have none of {137|138} the newness which the term ‘originality’ suggests to us, and that those created outside the systems, if there are any besides the special formulas, are too few to call for much thought. But in treating of the oral nature of the Homeric style we shall see that the question of a remnant of individuality in Homeric style disappears altogether.

It is of course the pattern of the diction which, as in the matter of the authorship of the style, proves by its very extent that the Homeric style is oral. It must have been for some good reason that the poet, or poets, of the Iliad and the Odyssey kept to the formulas even when he, or they, had to use some of them very frequently. What was this constraint that thus set Homer apart from the poets of a later time, and of our own time, whom we see in every phrase choosing those words which alone will match the color of their very own thought? The answer is not only the desire for an easy way of making verses, but the complete need of it. Whatever manner of composition we could suppose for Homer, it could be only one which barred him in every verse and in every phrase from the search for words that would be of his own finding. Whatever reason we may find for his following the scheme of the diction, it can be only one which quits the poet at no instant. There is only one need of this sort which can even be suggested the necessity of making verses by the spoken word. This is a need which can be lifted from the poet only by writing, which alone allows the poet to leave his unfinished idea in the safe keeping of the paper which lies before him, while with whole unhurried mind he seeks along the ranges of his thought for the new group of words which his idea calls for. Without writing, the poet can make his verses only if he has a formulaic diction which will give him his phrases all made, and made in such a way that, at the slightest bidding of the poet, they will link themselves in an unbroken pattern that will fill his verses and make his sentences.

This necessity which oral verse-making sets upon the poet shows its force most clearly, as has been said, in the simple number of formulas found in the Iliad and in the Odyssey, but there are also many cases in which that force can be measured by its effect upon the single phrase. The greater number of metrical irregularities in Homer come in the play of the formulas, either from a change within the formula, as when υἱὸς Πετεῶο (3 times) becomes υἱὲ Πετεῶο (Δ 338), or from the {138|139} grouping of formulas which will not join to one another without fault, as in the following example. Homer makes a large number of verses by joining to different predicates, which fill just one half of the verse, subjects of the type πατὴρ ἀνδρῶν τε θεῶν τε, ἄναξ ἀνδρῶν Ἀγαμέμνων, and so on. By this pattern he made τ 59 and τ 102:

ἔνθα καθέζετ᾽ ἔπειτα περίφρων Πηνελόπεια …

ἔνθα καθέζετ᾽ ἔπειτα πολύτλας δῖος Ὀδυσσεύς.He could have made just as many verses of this sort as he had noun-epithet formulas in the nominative, falling after the trochaic caesura of the third foot, and beginning with a simple consonant. But for Telemachus, whose name has a measure which bars it from forming in the latter part of the hexameter any formula save the little used type Τηλέμαχος θεοειδής (5 times), he had only Ὀδυσσῆος φίλος υἱός (9 times), which begins with a vowel. Nevertheless, the force of the formulas and the pattern of the verse was so strong upon him that he made π 48:

In the same way, on the type of formula found five times in the following verse:

and in Δ 350 = Ξ 83:

he made γ 230:

The type of formula found in the first verse of the Iliad has entered into the making of Τ 35:

It is not until we have read forty verses farther in the poem, however, that we find the direct model of this incorrect line:

This verse in turn belongs to the system in which falls E 444 = Π 711.

One should note especially that in this case as in that of π 48, just quoted, the incorrect verse occurs before its correct model. In neither place was the poet altering a line he had just used, but was composing after the pattern which he had in his mind. Now it is not possible that the metrical irregularities of the sort which have been given, and they are very numerous in the Iliad and in the Odyssey, [332] could occur in any but an oral poetry. The poet who makes verses at the speed he chooses will never be forced to leave a fault in his verse, but the Singer, who without stopping must follow the stream of formulas, will often be driven to make irregular lines. In such cases it is not the poet who is to blame, but his technique, which is not proof against all fault, and which, in the unhesitating speed of his composition, he cannot stop to change.

But there is more to show the oral nature of Homer’s diction than the need of explaining why he limited his thought to the formulas and made faulty verses. There is also the diction itself. First, there is a fairly large number of cases where the pure sound of one expression has suggested another which is altogether unlike it in meaning. Thus some Singer, whether Homer or another, when he had to express the idea ‘along with the clouds,’ thought of the words ὁμοῦ νεφέεσσιν (Ε 867), simply because his mind was guided by the echo of ὁμοῦ νεκύεσσιν (Ο 118), ‘among the dead’; or perhaps it was the latter phrase which was the model. The only examples of this sort that I have found in Euripides are ἀξένου πόρου, [333] made after Εὐξείνου πόρου, [334] and αὐτὸς ἀνταπωλόμην, [335] made after αὖθις ἀνταπώλετο. [336] One case at first sight seems to come near what we find in Homer: we read in the Andromeda: ἔα· τίν᾽ ὄχθον τόνδ᾽ ὁρῶ, [337] and in the Cyclops: ἔα· τίν᾽ ὄχλον τόνδ’ {140|141} ὁρῶ. [338] This would be no more than a faint parallel to the Homeric verses:

Here, by the change of four letters, the verse ‘to throw ajar the thick cloud, or set it to,’ becomes ‘to open the door of our shrewd ambush, or set it to.’ But the source of the Euripidean phrase becomes clear when we find Aristophanes using it in a ridiculous scene in his comedy The Women at the Thesmaphoria, [339] which means that Euripides was answering Aristophanes’ mockery by mocking himself. Thus only by a planned comic use of words does the Attic dramatist do what the epic poets did without thinking. [340] Other examples in {141|142} Homer in which the sound of words has suggested the terms of statement for an unlike idea are the following. The likeness of νήεσσι to νήσοισι has given us the verses K 214 and α 245 = π 122 = τ 130:

The likeness of ἠδέ to ἦλθε has suggested one or the other of these two verses:

The line Β 581:

was the model of δ Ι, in which the relative οἵ becomes a demonstrative:

Apollo and Athena both take the form of a man named Mentes:

Of shorter expressions we find ἀμφήλυθεν ἡδὺς ἀυτμή (μ 369) and ἀμφήλυθε θῆλυς ἀυτή (ζ 122), ἄναξ ἀνδρῶν Ἀγαμέμνων (37 times), and ἄναξ ἐνέρων Ἀιδωνεύς (Υ 61), and so on. [341] There is in most of these cases nothing to show us which of the expressions is the model and which the copy, nor do we know that it was Homer who was thus guided in his language by the play of sound, since it is more likely that he knew both original and copy as separate formulas. This, however, affects {142|143} our conclusions in no way; we are merely saying that the traditional style which Homer used was oral, and not that Homer’s style was so.

It is largely chance that has given us these expressions in which we find likeness of sound without likeness of idea; yet we would have known just as well without them that Homer’s style was oral. For there is a simple, almost too obvious, fact to show it: namely, that there is no memory of words save by the voice and the ear. We who have lived our lives with books, and have read much, often reach a point, at least for prose, where the words upon the printed page are more symbols for ideas than the record of speech; and it is our eyes which carry to our minds the author’s thought, rather than our ears. Yet if we would remember any sentence, even any phrase, we must say it to ourselves either aloud or beneath our breath, until the organs of our voice will repeat, at our bidding, the gesture of its utterance. There are some, they say, who can recall whole sentences or even passages because they can picture to themselves the way they look in print. In the same way school-boys remember the place of Greek accents, being unable to make with their voices any sound for which they would stand; and we also know that one can learn a foreign language in a way by learning to tell the idea from the printed image of the words and phrases, and that one can even write sentences of a sort by grouping together such images by a purely visual process. But as there is no real knowledge of a language thus learned, so is there no real memory without sound. As a rule we are unable to recall a single phrase of the book we have read silently. The poet who is repeating his own phrase, or that of another, is doing so by ear. To deny this for any poet is to suppose impossible things. The repeated phrases in Virgil, then, would become, as it were, labels which the poet fitted into his verse in the same way that one pieces together a puzzle, and not, as we know they were, expressions which were judged in every part by his sensitive feeling for sound. And would one dare to say that Pope had never heard the phrase he took from Milton: “thick as autumnal leaves”?

We know Virgil’s practice of dictation, and of reading his verses to his friends, but we do not have to suppose that he spoke aloud every phrase which he had used over; nor did Pope necessarily have read to him all the poems of Milton from which he borrowed. He may {143|144} have muttered the words, or have spoken them to himself beneath the hearing of any other person. But memorizing under one’s breath is possible only up to a certain point. Pope may thus have learned all the phrases of Milton which he knew by heart, and Virgil most of what he repeated, though it is hard to believe that. But when we come to Homer such a thing is beyond reason. Would he, by the conning of long manuscripts of epic poetry, have learned the thousands of whole verses, the thousands of verse-parts, which make up the traditional diction? And we must suppose that the authors of all those manuscripts in turn had learned just as faithfully in the same way the countless mass of formulas, and so on back. But the argument has reached the absurd, and we are trying to suppose an oral poetry in an undertone. Homer could only have learned his formulas by hearing them spoken in the full voice of those poets to whom he listened from his childhood.

Homer learning his formulas from manuscript is no harder to imagine than Homer using his formulas to write verses. He is then keeping his thought to them not because he has to, but because he wants to. His is a strange game in which he must fit into his written lines only those phrases which have been used by others. Each one of the eighty-eight times he uses τὸν δ’ αὖτε προσέειπε, followed by a noun-epithet formula, or the hundreds of times he uses αὐτὰρ ἐπεί at the beginning of the verse, or whenever he uses any formula whatsoever, he is showing his skill in choosing the old expression, his stern disregard of all the new groups of words which, since his writing materials gave him time to pause, must have crowded annoyingly about his head. And this is the way we must suppose him to have made almost all, if not all, of his poetry. If one wishes to think that Homer composed his poems orally, and then sat down and wrote them out, there is little that can be said in disproof, and little that needs to be said, since the question ceases to be one of the oral style, and becomes that of the way in which the spoken poetry was recorded.

So far we have said only that the formula itself must be a thing of sound and not of sight. We now come to the last of proofs that the diction of the Iliad and the Odyssey is oral, and to the one which is most precious for our understanding of Homer: the technique of the formulas is one which could only be created and used by oral poets. {144|145}

Each system of formulas comes, in the last analysis, from some single expression. The simple fact that two phrases are too closely alike to be due to chance implies that one of them imitates the other, or that they go back to a common model. There was one formula, what one we shall never know, from which comes all the system found in Homer: δῖος Ἀγήνωρ, δῖος Ἀλάστωρ, δῖος Ἀχιλλεύς, δῖος Ἐπειγεύς, δῖος Ἐπειός, δῖος Ἐχέφρων, δῖος Ὀδυσσεύς, δῖος ὑφορβός. More than that, there was one noun-epithet formula which was the beginning of all the larger system δῖος Ἀχιλλεύς, μητίετα Ζεύς, πότνια Ἥρη, φαίδιμος Ἕκτωρ, and so on. Likewise, all the formulas of the system αὐτὰρ ἐπεί ῥ᾽ ἵκοντο, αὐτὰρ ἐπεί ῥ᾽ ἤγερθεν, and the like, go back to one source, as does the type of verse we see in the two following lines:

When one multiplies these cases by the number of the systems in the Homeric diction, one sees that the formation of the style was of a very special sort. The Singers, ever seeking to reduce the terms of their expression to the simplest pattern, used for this end the means of analogy. [342] That is to say, wherever they could obtain a new formula by altering one which was already in use, they did so, and this they did up to the point where the complexity of the ideas which must be expressed in their poetry put a stop to this making of systems. This means of forming the system is quite different from that which would have been followed if it were the usefulness of the formula alone which led the poets to make it and keep it, for then we should find a diction in which there would be formulas, but few of them would ha\e the same words as another. Instead of ἄναξ ἀνδρῶν Ἀγαμέμνων and ἄναξ ἀνδρῶν Αἰνείας, we should have had formulas with different epithets. We should not find τὸν δ᾽ ἠμείβετ᾽ ἔπειτα joined with twenty-seven different noun-epithet formulas, but many different kinds of lines for announcing a character’s answer. But such is not the nature of the epic diction, which so much preferred to use the same words where it could that there are in all the Iliad and the Odyssey only forty fixed epithets that are used for single heroes, beside sixty-one that are used for two {145|146} or more. [343] Thus we have δῖος in the nominative used of twelve characters, θεοειδής of fourteen, ἥρως of ten, δουρικλυτός of eight, and so on. In these cases, and in all others, we see the sound of the words guiding the Singers in their formation of the diction. Nor is the factor of sound limited to the formulas where the same words appear; it appears equally in the more general types where the likeness of sound consists in the like rhythm. The sound of the words has not acted so willfully in the creation of the systems as it has in the case of those formulas which we noted above, in which it has gone so far as to give the poetry its ideas. Here it has followed the thought which the Singers wished to express, though it imposed rigorous limits for that thought; yet whereas in the one place it created only a certain number of isolated phrases, it here has had an influence as far-reaching as the schematization of the style.

This formation of the traditional diction belongs, of course, to a time far earlier than that in which Homer lived, but the making of the diction is in no way different from a single poet’s use of it. One can say that the Singer, in a recitation of a few hours, repeats the history of his style, for it is the play of sound which guides him in his grouping of the formulas, quite in the way that it had guided the poets of an older time in their making of them. As they had made for him οὐδέ με πείσεις, οὐδέ με λήσει, οὐδέ ἕ φημι, and so on, to be used at the verse-end, even so, when he has ended a sentence at the bucolic diaeresis, he is led by the habitual movement of his voice to these formulas. Or at the beginning of a verse, when he has another sort of transition to use and a certain act to tell, he will be guided by his feeling for what there is in common in the sound of such a system as νῦν δ’ ἐθέλω, νῦν δ᾽ ἔχομαι, νῦν δ᾽ ἦλθον, or οἱ δ᾽ ὅτε δή ῥ᾽ ἵκοντο, οἱ δ᾽ ὅτε δὴ σχεδὸν ἦσαν, and so on. And it is here, finally, that we can see why we should not seek in the Iliad and in the Odyssey for Homer’s own style. The poet is thinking in terms of the formulas. Unlike the poets who wrote, he can put into verse only those ideas which are to be found in the phrases which are on his tongue, or at the most he will express ideas so like those of the traditional formulas that he himself would not know them apart. At no time is he seeking words for an idea which has never before found expression, so that the question of originality in style {146|147} means nothing to him. It may here occur to some to ask how the diction was ever made if one thus grants the Singer no power to change it. It is to be answered that the years of its first making belong to a very dim past, and were also those of its least perfection; then, that we may well suppose for the single poet a very few cases where the play of words has suggested some new epithet, or phrase, or verse, which the other Singers found worth using and keeping, but that there could never be more than a few such creations for any one Singer, and they could win a place in the diction only as they were in accord with what was traditional, and fitted the habits of verse-making of the other poets. Indeed, in certain places in the poems we can see how certain very effective phrases or verses were made. The wondrously forceful line:

is made up of verse-parts found in other parts of the poems: κεῖτο μέγας (M 381); μέγας μεγαλωστί (Σ 18); λελασμένος ὅσσ’ ἐπέπονθεν (ν 92); λελάσμεθα θούριδος ἀλκής (Α 313). There is a striking play on the name of Odysseus in α 62: τί vύ οἱ τόσον ὠδύσαο, Ζεῦ; which is made after ἐπεὶ μέγας ὠδύσατο Zεύς, which is found in Σ 292. There are in all the poems only two other places where Ζεῦ is found at the verse-end: μητίετα Ζεῦ (A 508) and εὐρύοπα Ζεῦ (Π 241). That Homer might, by a like new play of formulas, have added to the great wealth of the traditional style is possible, but we shall never know, since if he did so he was guided by the same play of words and phrases as all those other poets who, bit by bit, and through the many years, had made this best of all styles.