The visitor to the library of the classical faculty at Harvard sees many photographs of past professors, almost all imposing, bearded figures; so imposing are they that it comes as a rude shock when one remarks that many of them have no very notable achievements to their credit. But among them one finds one youthful, strikingly handsome face, and it is the face of the most celebrated of all. No American classical scholar of the twentieth century has won higher praise than Milman Parry, who was born at Oakland, California, in 1902 and died in 1935.

By the common consent of readers over nearly three millennia, the Iliad and the Odyssey of Homer contain poetry of the highest order. But they have certain features which the taste of many later ages finds perplexing, if, not repugnant. They contain numerous repetitions, including the regular use of certain standard epithets to describe particular persons; readers of Proust will recall the embarrassment and annoyance caused by Bloch’s tiresome habit of imitating this kind of thing in conversation. They also contain a number of real or apparent inconsistencies, so that it is not surprising that the eminently rational eighteenth century asked some awkward questions about them. The French, in particular, asked awkward questions, and the most notable attempt to answer them came from a German. In 1795 the great scholar Friedrich August Wolf1 argued that Homer was an oral poet, ignorant of writing. He lived, Wolf thought, about the middle of the tenth century BCE, and his poems suffered many alterations and expansions before being written down as late as the sixth century. From then on many scholars analyzed the epics into the parts written by Homer and the parts supplied by others. Most, like Wolf, placed Homer at the beginning of the process, but some placed him in the middle. Others defended the essential unity of the two epics against the analysts, and this state of affairs continued until well inside the present century.

Milman Parry put forward a theory which placed the old battle between analysts and unitarians in a new perspective. Educated at the University of California at Berkeley, he obtained an MA there with a thesis containing the gem of a theory which he later worked out in detail at the Sorbonne, where he gained the degree of Docteur des Lettres with two treatises published in 1928 in French, and later elaborated in a series of articles published in America in English.2

The names accompanied by their regular epithets—Achilles swift of foot, Hector of the flashing helmet, Agamemnon king of men—are not the only repetitions that occur in Homer; many phrases, and sometimes whole sequences of lines, are repeated in the poems, sometimes in identical form and sometimes with minor variations. Various elements of his theory had been anticipated by other scholars, but Parry was the first to argue convincingly that these and other features of the style, meter, and dialect of the poems indicated that they belonged to a tradition that must, at least in its beginnings, have been one of oral poetry. The oral poet drew on many formulas and on many themes; every performance of a poem was an act of creation, and no two performances of the poem quite the same.

After his return from France Parry taught for a year at Drake University in Des Moines, Iowa, before being appointed to a position at Harvard, where he remained for the rest of his brief career. In 1933 and again in 1934 and 1935 he made long stays in Yugoslavia in order to study the only oral tradition surviving in Europe to which he could have access, and managed to record many specimens of its performances. In this work he was assisted by Albert Bates Lord, who later became Arthur Kingsley Porter Professor of Slavic and Comparative Literature at Harvard and went on to publish much of the material which he and Parry had recorded. Disastrously, Parry died young; in December of 1935, arriving at a hotel in Los Angeles, he deposited on the floor a suitcase that contained a loaded pistol, which went off, killing him instantly. But Lord lived until the 29th of July last year.

Parry originally defined a formula as “a group of words which is regularly employed under the same metrical conditions to express a given essential idea.” The system of formulas, he went on to argue, was applied with such “extension” and with such “economy”—these are important technical terms in Parry’s theory—that although “traces of originality remain, perhaps,” these are “traces of an originality that does no more than rearrange the words and expressions of the tradition without important modifications.” But from the time of an article published in 1930,3 Parry attached much weight to the concept of analogy, by which he meant the formulation of new expressions on the model of particular words and of the metrical patterns of old formulaic expressions. This modification allows the oral poet much greater freedom; it extends the concept of the formula not inconsiderably, and as we shall see presently other modifications have extended it still further since.


At the very end of his career, Parry began to reduce the importance he assigned to the formula and to increase that which he assigned to the theme. The theme has been defined as “a sort of basic unit of narration in an oral poem,” which may be a unit of action, such as a single combat, the calling of an assembly, or an arrival, or it may be an account of an arming, or a chariot, or a feast. Not that Parry ever offered an explicit definition; but the oral poet’s composition could be more convincingly explained if he was imagined as not simply drawing on a particular stock of formulas but as producing new versions of particular established themes.

Even before Parry’s early death his work had begun to be given recognition, notably by the eminent Swedish historian of Greek religion, Martin Nilsson.4 But in 1938 a dissenting voice was raised; Samuel Eliot Bassett5 in his posthumously published book The Poetry of Homer, while accepting Parry’s contention that the tradition to which Homer’s poetry belonged had for a long period been oral, argued that Parry’s principles seemed to deny to Homer the least spark of originality, and could not explain how such great poetry as his could have come into being.

After the war, when classical studies once more got into their stride, English-speaking scholars for the most part acknowledged the importance of Parry’s work while the Germans and other European scholars ignored it. Not that all English-speaking scholars went as far as to agree that the Homeric poems were composed without the aid of writing. Two important books published in 1952 agreed that Homer wrote in the wake of an oral tradition, but denied that his poems were oral poems; these books were Sir Maurice Bowra’s Heroic Poetry6 and H. T. Wade-Gery’s The Poet of the Iliad.7 Bowra’s book is a learned survey of the heroic poetry of many peoples and of many times; Wade-Gery’s contains an ingenious suggestion about the origin of the Greek alphabet. This is generally agreed to have been modeled on the tables of the written signs for syllables used for writing by the Phoenicians, that Semitic people of whom the Canaanites formed part and with whom the Greeks traded extensively not only during the Bronze Age but also later, from the tenth century. Wade-Gery suggested that the alphabet had actually been invented in order that the poet might use it to compose poems longer and more complex than any that had been known before.

Albert Bates Lord replied to Wade-Gery and Bowra in an article8 which is reprinted, with minimal changes, as the second chapter in the collection of his essays which has just been published, Epic Singers and Oral Tradition. He seems to take it for granted that all oral poetry conforms to the rules obeyed by the Serbian guslars, or bards, whom he had studied in company with Parry; “these statements,” he assures us, “have been checked by field experiments.” In replying to the charge that Parry’s principles deprive the oral poet of almost any possibility of invention, he invokes the concept of analogy, and he lays special emphasis on the concept of the theme, defined now as “the reported narrative descriptive elements,” so that he can end by claiming that “a singer can show originality both in new phrases and in new combinations of themes.” Answering Wade-Gery, Lord argues that Homer must have dictated his oral poem, as some guslars had done, to someone who could write; since Lord held that literacy brought immediate death to any oral tradition, he could not allow that Homer himself might have learned the art of writing.

A year after the original publication of Lord’s article a leading German-speaking scholar took account of Parry’s work for the first time. This was the celebrated Viennese professor Albin Lesky,9 who, like Bowra and Wade-Gery, accepted Parry’s claim that the Greek epic tradition must originally have been oral, but could not believe that Homer himself did not use writing. But in the English-speaking world Parry’s theory in its most rigorous form found powerful support. Sir Denys Page, who from 1950 to 1973 was Regius Professor of Greek at Cambridge, was its most enthusiastic advocate. “Subtlety of soul, complexity of character, true portrayal of personality,” wrote Page in 1955,10 “for these we must wait until the practice of the art of writing affords the poet the necessary leisure and the necessary means for reflexion, for planning the future in some detail, and for correcting the past.” “Intricacy of design and subtleties of soul,” he added later, “wholly alien to the oral technique of composition have been sought (and found) in him.” These pronouncements run directly counter to the collective judgment of readers over nearly three thousand years.


In 1960 Lord brought out his book The Singer of Tales,11 which bears the title of the book Parry had been planning to write at the time of his death. Part I of this work explains the theory of oral poetry, and Part II, with one chapter on the formula and another on the theme, explains its application. The formulas, Lord writes, “are not limited to the familiar epithets and oft-repeated lines, but are all-pervasive”; “the formula technique in the Homeric poems is indeed so perfect, the system of formulas is, as Parry showed, so thrifty, so lacking in alternative identical expressions, that one marvels that this perfection could be reached without the aid of writing.” In this book the theory that Homer must have dictated his poems to someone who could write is worked out in greater detail. Two years later G. S. Kirk in a solid and comprehensive study of the epics 12 contended that the poems were preserved orally more or less unchanged before being written down in the sixth century; but Lord was able to reply with much cogency that they were hardly likely to have been preserved orally for so long without considerable alteration.

A year earlier had appeared in Germany a posthumously published study of the Iliad by Karl Reinhardt, a learned man who possessed a gift of literary appreciation unequaled among the Greek scholars of his time. “I cannot begin this book,” Reinhardt wrote, “without pointing out that I do not share this conception [Parry’s theory]. If it is correct, then it would have been better that this book had never been written.”13

During the following years several broadsides were directed against the oral theory by scholars unable to understand how such great poetry could have been created by the kind of process it envisaged. The most effective were fired by M. W. M. Pope (1963), by the Scottish poet and scholar Douglas Young (1965, 1967), and by Milman Parry’s son Adam (1966), who not only resembled his father in appearance and in character, but inherited a great share of his intellectual powers, and by Joseph Russo (1968).14 Lord replied to some of these criticisms in an article in which he commented on the volume in which Adam Parry’s criticism had appeared, an article not reprinted in the new collection of his essays. The solemn warning against “the subjective interpretation and appreciation of the Homeric poems” which Lord delivered in this article brought upon him strong criticism from Adam Parry’s wife, Anne Amory Parry.15 She pointed out that the sharp distinction Lord had drawn between oral and written poetry was “surely too crude to be true,” since oral and written poetry had common features as well as differences, and she remarked that Lord had utterly ignored the evident differences between Homer and the poetry of the Yugoslav bards, “through which, and through which alone, apparently, he would have us approach Homer.” The bibliography appended to Lord’s book mentions only one of the works cited in this paragraph.

It was becoming clear even to firm upholders of the oral theory that the concept of the formula needed considerable further elucidation, and it was carefully examined by two very able scholars, A. Hoekstra (1965) and J. B. Hainsworth (1968).16 Neither can be called an unsympathetic critic of the theory, yet each ends by stressing, more strongly than any of his predecessors, the extreme flexibility and adaptability of this elusive concept.

The strange affinity between Adam Parry and his father seemed even more uncanny when in 1971 Adam and his wife died together in a road accident in Colmar, a loss from which Greek studies in this country have scarcely yet recovered. But Parry had completed a memorable work which was published in that same year, a collection of his father’s writings, in which the two important Paris theses are translated from the French and the whole is preceded by a masterly introduction by Adam Parry himself. Twenty years later this still seems to me the best introduction to the Homeric question.17 Adam Parry argued, as he had done in his article of 1966, that though the Homeric epics bore many signs of belonging to a tradition that had for many centuries been oral they themselves could not have been composed without the aid of writing. His father, he pointed out, had during his brief career been concerned not so much with the Homeric poems themselves as with the tradition of which they were the final product, and he pointed to several indications that had his father lived he would have arrived at a view of Homeric poetry very different from that of Sir Denys Page and other “hard-core Parryists.”

During the Seventies much progress was made with the detailed work needed to put the oral theory to the necessary tests. In 1975 Norman Austin18 called into question the dogma that the standard epithets accompanying the names of characters in the poems were necessarily almost meaningless in their context. In 1977 a learned and judicious survey of oral poetry of many kinds was published by Ruth Finnegan. 19 She showed that there are many different kinds of oral poetry, and many different relationships between oral and literary elements in the same tradition. “I distrust over-arching theory,” she wrote, “or large generalizations purporting to cover all oral poetry.” Several early English poems, Dr. Finnegan showed, which contain repeated formulas and other features associated with an oral technique are known to have been composed with the aid of writing; certain living Bantu poets, she added, write their poems in a formulaic style elaborated from their originally oral poetic tradition. “In practice,” she concludes, “interaction between oral and written forms is extremely common, and the idea that the use of writing automatically deals a death-blow to oral literary forms has nothing to support it.”

In 1974, three years before the publication of this book, Dr. Finnegan had taken part, together with another of Lord’s critics, Joseph Russo, in a conference on Oral Literature and the Formula at Ann Arbor, Michigan; their remarks and Lord’s replies may be studied in its proceedings.20 Speaking of the formula, Dr. Finnegan said,

If one takes a narrow definition—in terms say of identically repeated word patterns—then, as has often been pointed out, this does not always apply to literatures other than those in which it was first discovered by Parry and Lord. While if one takes a wider definition—one in effect implied not only in later extensions of the theory but also in earlier statements like Lord’s “every line and every part of a line in oral poetry is formulaic”—then the specificity of the concept seems to evaporate and one is left with nothing very precise, despite the statistics.

Faced with this formidable critic, one might have expected Lord to put up a vigorous defense, but in fact he declined the challenge, contenting himself with making “three or four short points.” One of these has special interest, for Lord disavowed the belief, expressed in The Singer of Tales, that once a singer learned to write he lost the capacity to compose orally. One might have expected this change of opinion to cause him to modify his theory that the composer of the Homeric epics dictated them to a writer, but Lord seems never to have done this.

In 1982 two studies of Homer were published by scholars who unlike so many learned men had an exceptional feeling for poetry. Colin Macleod in his memorable edition of the last book of the Iliad,21 still upheld the view that the Iliad is an oral poem, but he thought that oral composition did not exclude “premeditation, artistry or profundity,” and shows a fine sense of the fundamental unity of the work. Jasper Griffin, in his brilliant study of the gods in Homer,22 citing the work of Adam Parry and of Ruth Finnegan, rejects the notion that the criticism of the poems must proceed on an entirely different basis as a result of Parry’s work. In 1986 Griffin23 revealed significant differences between the style and language of the many speeches put into the mouths of the characters of the Iliad and the direct utterance of the poet. Parry’s claim that the generic epithets are meaningless was shown even more clearly to be untenable, and Homeric language in general was proved to be a great deal more varied and more adaptable to the needs of each individual context than the oralists had been willing to allow.

Finally, in 1987 a young American scholar, David M. Shive, published the results of a careful examination of the theories of extension and economy which are central to the oral theory.24 Parry, he showed, had made the mistake of working not directly from the text of Homer while compiling his statistics, but relying on the numerous lexica and compendia which assist the readers of his works. Going through the poet’s various ways of alluding to his principal character, Achilles, in each of the five cases, nominative, vocative, accusative, genitive, and dative, and not forgetting those instances in which the poet calls him not by his name but by a patronymic, Shive shows that far from being thrifty or economical in naming Achilles, the poet has been lavish and profuse. To take one example, Shive lists no fewer than thirty-two ways of naming Achilles in the dative case alone; in his words “the idea of datival Achilles is expressed with consummate complexity.”

Writing in 1981, I had protested25 against Kirk’s statement that “a large number of crystallised formulas are employed with an astonishing economy and lack of unnecessary variation” and that “this suggests strongly and indeed imperatively that the oral technique was used in full and undiminished degree for the main act of composition of each of the monumental poems”; now Shive seems to me to have demonstrated that this claim is totally untenable. Alfred Heubeck in his introduction to what is now the standard commentary on the Odyssey26 stated his belief that the great epics must have been composed with the aid of writing, but was willing to concede that this could not be proved. If he had lived to read Shive’s book, I believe he would have stated this conclusion firmly.

This being so, one turns with special interest to the fifth chapter of Lord’s book, which is entitled “Homer as an Oral-Traditional Poet.” This is based on a lecture given in March 1984, but not hitherto published, and apart from the second chapter, already discussed, it is the only systematic treatment of this subject in the book.

Lord replies to Austin by adopting the wider of the two concepts of the formula mentioned by Dr. Finnegan in the passage quoted above, and deals with the objection that Homer cannot have improvised his poems by pointing out that he prefers to speak of composition by theme rather than of improvisation. Next he quotes passages of Pound and Eliot as specimens of written poetry and specimens of Yugoslav verse as examples of oral poetry. While grateful to Lesky for having been the first leading German-speaking scholar to take note of Parry’s work, Lord complains that since Lesky knows the epic “The Wedding of Smailagić Meho,” performed for Parry by the guslar Avdo Mededović, only from the summary of its plot given by Bowra, he has no right to insist that “the gap between it and Homeric art leaves a decisive impression.” Lesky’s article was published as long ago as 1954; but of more recent criticisms of the oral theory, Lord says nothing.

Lesky clearly agreed with Anne Amory Parry that “the crucial difference between Homer and Yugoslavian poetry is that of quality.” Lord seems to feel that a person ignorant of Serbo-Croatian has no right to express an opinion on this topic, and a knowledge of that language would certainly make it easier. But from Lord’s own writings, particularly those contained in this book, and from certain other works, particularly an article published in 1971 by the German scholar Franz Dirlmeier,27 one can get a very fair notion, it seems to me, of the kind of thing with which we have to deal. Lord offers samples of the original texts in Serbo-Croatian, together with literal translations; I take a few specimens quite at random. The first comes from “Cevljanin Rade and the Captain of Spuz,” a poem dictated to Parry in 1935 by a seventy-year-old guslar.

Cevljanin Rade is drinking wine
In the midst of Cevo in the white tower.
An adorned mountain woman is serving the wine,
In her right hand a beaker and a golden cup.
When Rade had his fill of wine,
He began to talk of many things,
How many Turks he had cut down
Around Spuz, the bloody town.

The meter is a trochaic pentameter with a regular break after the fourth syllable, described in the fourth chapter of The Singer of Tales; I have the impression that it is less effective than the Homeric hexameter. Here is a passage from a South Slavic wedding song:

The maiden is separated from her father,
The maiden is separated from her mother,
The maiden is separated from her brothers,
The maiden is separated from her sisters,
The maiden is separated from her family,
Her family and her kin.

These repetitions are not at all like those found in Homer; these passages, and most of the specimens of Slavic poetry quoted by Lord, remind us not of Homer, but of ballad literature. Matthew Arnold in the second of his memorable lectures On Translating Homer was at pains to explain the differences between ballad literature and the Homeric epic, and to show why translations of Homer into the style and meter of ballad literature are bound to fail. “The style of Homer,” Arnold’s antagonist, F. W. Newman, had written, “is direct, popular, forcible, quaint, flowing, garrulous; in all these respects it is similar to an old English ballad.”28 Anybody who believes that, or believes that it is similar to the Yugoslav ballad, would do well to read the few pages in which Arnold shows what are the differences, pointing out that Homer is not “quaint” or “garrulous,” but that he is eminently noble, which the ballad style is not.

It may be objected that the many shorter Serbo-Croatian poems are less important for comparison than are the two epics dictated to Parry by Avdo Mededović, “The Wedding of Smailagić Meho,” which, in the version dictated to Parry had over 12,000 lines, and Osmanbeg Delibegovic i Pavicevic Luka, which has over 13,000 lines. Summaries and translated extracts give us a general notion of the character of both these works. Only a fairly long extract would help the reader much, and my space here is limited; but I may record my impression that the epics in question are hardly as distinguished as, say, Sir Walter Scott’s long poems, Marmion, The Lady of the Lake, The Lay of the Last Minstrel. It seems clear enough that they cannot be compared with Homer.

Apart from the three chapters which have been discussed, Lord’s new book contains ten chapters dealing not only with Serbo-Croatian poetry, but with other Balkan poetry, with the Anglo-Saxon Beowulf and Elene, the Finnish Kalevala, the Byzantine epic of Digenis Akritas, and the epics of central Asia and their relation to the epics of the Balkans. Lord is much concerned with the reappearance of the same themes in different poems and in different places, and he has collected some interesting material of this kind.

The problem of the nature of Homeric transmission is closely linked with that of the creation of the Greek alphabet, on which several other alphabets, including the Roman, which we still use, are based. The Greek alphabet was certainly one of the great achievements of the ancient culture. It was adapted from the syllabary devised by Phoenicians to express their Semitic language; but the main feature differentiating it from this, and indeed making it not a syllabary but an alphabet, is that instead of leaving the vowels to be inferred by the reader it expresses them by actual signs. The Linear B syllabary—used during the Bronze Age, it would appear, solely for the recording of inventories—went out of use toward the end of the second millennium BCE.

The earliest specimens of Greek writing in the alphabet date from the second half of the eighth century BCE, and since the publication of an article by Rhys Carpenter in 193329 most classical scholars have believed that the alphabet came into existence between about 720 and 710. But for some time now Semitic scholars have been sending out signals that the letter forms indicate that the adaptation must have taken place earlier, perhaps considerably earlier, than that date. In 1973 Joseph Naveh30 argued for the eleventh century. In 1975 P. Kyle McCarter31 in a careful study worked out a compromise, placing the invention of the alphabet at about 800 BCE; in Homer and the Origin of the Greek Alphabet, Professor Barry B. Powell accepts his results. Powell offers tables showing the early forms of the Phoenician letters and also the letter forms of the earliest Greek inscriptions, and also gives texts of the latter with detailed comments; these features alone are enough to make this book very useful to professional scholars.

Powell argues that the invention took place not in Rhodes or Crete, the places first favored by Carpenter, nor Cyprus, favored by Carpenter later, nor the trading post of Al Mina on the Orontes, excavated during the Thirties by Sir Leonard Woolley, but the Aegean island of Euboea. During the ninth and eighth centuries the Euboean towns of Chalcis and Eretria played a central part in trade and colonization both in the eastern and in the western Mediterranean; much light upon their activities has been thrown by the excavations at Lefkandi, which may have been the earliest city of Eretria. Directly opposite Euboea on the eastern side of the Aegean was Chios, an island traditionally connected with Homer. Chalcis and Eretria had combined to found the western colony of Pithekoussai, “Monkey Island,” modern Ischia, on the Bay of Naples; here was found a cup dated to about the second half of the eighth century and bearing, as do a surprisingly large number of the earliest specimens of Greek writing, an inscription in hexameter verse. The inscription claims that the cup is the very cup of the hero Nestor, mentioned in the eleventh book of the Iliad, and that whatever is drunk from it will have aphrodisiacal properties.

Citing parallel cases from the history of other writing systems, Powell argues that there was one single invention by a single individual. He revives the conjecture of Wade-Gery in the book of 1952 mentioned above that the alphabet was invented for the specific purpose of recording the Homeric epics; this he combines with the conjecture of Lord that Homer dictated his poems to a person who could write.

This theory raises not inconsiderable difficulties. At the start the new invention would have been intelligible only to the inventor and his immediate circle. In the early period, the alphabets used by the various Greek communities were by no means identical with one another. Certain letters called “supplementals”—phi, chi, psi, omega—had different values in different places; in Crete and in its neighboring islands of Thera and Melos, they do not appear at all. Powell’s view that they belonged to the original alphabet, but were abandoned or given different values in certain places during the half century following the alphabet’s invention, is not convincing. Fascinating as is the bold conjecture that the alphabet was invented specially to record the great poems, the humdrum notion that it was invented by traders to help them in the conduct of their business seems a good deal more consonant with the way things usually happen in ordinary life.

In any case, it is very far from certain that the alphabet was invented as late as about 800 BCE. Herodotus,32 writing in the second half of the fifth century, ascribed its invention to the legendary founder of Thebes, Cadmus. This is usually taken to indicate simply that the Greeks were aware of the alphabet’s Phoenician origin, since Cadmus was supposed to have come from the Phoenician city of Tyre. But Martin Bernal in Cadmean Letters boldly argues that Herodotus was right in implying that the alphabet was invented as early as the Bronze Age. He assigns the invention of the variable letters, the supplementals, to the earlier period; the alphabet arose, he thinks, from a kind of Semitic writing which flourished in the Levant during the middle of the second millennium BCE. The alphabet was reorganized, he argues, as a result of the later expansions of Phoenician trade between the eleventh and the ninth centuries.

Indeed the eminent Semitic scholar Frank M. Cross has argued that in order to explain the distance between the earliest Greek scripts and any point in the sequence of the Proto-Canaanite and linear Phoenician types of script, we must postulate a considerable space of time between the invention of the script and its appearance in the earliest specimens. This is an exceedingly complicated question, and must be handled with great caution; but I must admit that the study of the letter forms seems to me to indicate that the alphabet cannot have been invented anywhere near as late as Powell, following McCarter, thinks.

The obvious objection to an earlier dating for the alphabet’s invention is that our earliest specimens of Greek writing, which are set out and admirably discussed in Powell’s book, are not earlier than the second half of the eighth century. If the alphabet was invented much earlier, why have we no specimens from the early period? Most of the writing materials used at this date, it is true, are by no means durable; wax tablets and leather rolls cannot be expected to have lasted long. But we have much Greek pottery earlier than the middle of the eighth century; why does not a single fragment carry a Greek graffito?

I suspect that the answer may be that for centuries after its invention the Greek alphabet, like the other systems of writing current in the Levant in early times, was used only by a very restricted number of persons. In the palaces of the Middle East writing was long restricted to the caste of scribes. Can it not be that, in a similar way, among the Greek traders who first used it writing was in the hands of a small number of specialists, perhaps one to each trading group? If the alphabet was invented, not necessarily as early as the Bronze Age, but say in the eleventh or tenth century BCE, that would leave time for the variations in the use of the supplementals to become established; Bernal points out with much reason that it is likelier that letters were added than that letters were lost, so that the alphabets of Crete, Thera, and Melos, which lack the supplementals, may have been earlier, not later, than the rest.

Suppose the Greek alphabet had in fact been in existence for a long time before the eighth century, in the hands of a limited number of writers. Then there would have been plenty of time for the poets to have perceived the advantages offered by the useful device invented by the tradesmen, and to have familiarized themselves with its use. The remarkable prevalence of hexameter verse among the earliest Greek inscriptions suggests to me that by the middle of the eighth century the epics were by that time well known in many parts of the Greek world. References to Homer in art, it is true, seem to begin only about 700 BCE; but how many mythological scenes are depicted before that era?

Bernal offers a detailed treatment of the various Semitic scripts, including the syllabaries used in Spain, and discusses their possible relation with other systems of writing, including the Etruscan. I am not competent to pronounce on the conclusions he derives from this, but I am not convinced that the Greek alphabet originated as early as the fourteenth century, as he believes. However, I think it highly probable that it dates from a good deal earlier than the eighth century; and I believe that Homer was familiar with its use.

Firm belief in the oral composition of the Homeric poems is still widespread; during the year 1990 two learned men, Martin West and Richard Janko,33 have reaffirmed their faith in it, though without adducing new arguments that might confirm that faith and without attempting to refute Shive’s arguments. Belief in the theory is fundamental to the approach to early poetry of Gregory Nagy, who in the words of an eminent scholar, the late Friedrich Solmsen,

expands the concept of traditional poetry far beyond formulas and typical scenes, comprehending under its larger themes their interconnections, situations, emotions, especially as attached to specific characters and much else.34

Here we find resurfacing the notion that arose during the eighteenth century as a result of Herder’s work, according to which the author of the Greek epics was not Homer but the Greek people itself. This idea is not congenial to those who believe that it is not the people, nor the tradition, but poets, aided, no doubt, by the tradition, that create poetry.

This Issue

March 5, 1992