Homer was poor. His scholars live at ease,
Making as many Homers as you please.
And every Homer furnishes a book.
—J.V. Cunningham

The Iliad and the Odyssey, traditionally attributed to Homer, may be the greatest epics in Western literature, but no one really knows where they came from. The inquiry into this mystery—the effort to establish who composed the epics, and how, when, and where—is called the Homeric Question, and it remains one of the most actively cultivated fields in classical scholarship. Gregory Nagy of Harvard has now given the question a new answer.

The modern period of Homeric studies began with the publication of F.A. Wolf’s Prolegomena to Homer in 1795.1 Wolf proposed a revolutionary theory: the society in which the Homeric poems originated was, like the one that they describe, illiterate. The poems were composed as songs to be heard, not read, by an audience as illiterate as the poet himself. But, Wolf continued, the Iliad and the Odyssey as we have them are of a size and overall unity such as (1) could not have been created without the aid of writing, and (2) could not have been apprehended except through reading. Illiterate Homer, therefore, could not have given the poems their present shape: he must have composed episodes of two thousand lines or so, enough for a satisfying evening’s entertainment. The picture given in the Odyssey itself was thus in all essentials correct:

The herald came near, bringing with him the excellent singer/whom the Muse had loved greatly, and gave him both good and evil. She reft him of his eyes, but she gave him the sweet singing/art…. [The gathered nobles] put forth their hands to the good things that lay ready before them./But when they had put away their desire for eating and drinking,/the Muse stirred the singer to sing….

(Od. 8.62-73)2

The episodes thus performed were orally transmitted by Homer’s successors. Someone later, in an age of writing, noticed that with a little editorial patching and pasting the episodes could be put together to form a unified, large-scale structure, and that is the origin of our Iliad and Odyssey (about sixteen and twelve thousand lines, respectively).

Wolf’s was a seductive theory, but its elegance concealed a paradox that returned to haunt Homer studies in this century: the size and artistry of the Iliad and Odyssey were, according to Wolf, too great to be the work of an illiterate oral poet. The defects in that unity, on the other hand, were judged by him glaring enough to betray the poems’ origins in the stitching-together of previously autonomous, orally composed episodes. For Wolf’s theory to hold, in short, the poems had to be too good to be oral, but too bad to be literate.

Working out the distribution of good and bad occupied the next century of Homeric scholarship. For example, a typical assumption was that Homer, in prehistory, composed a first-rate poem about Achilles’ quarrel with Agamemnon.3 As this poem, much shorter than our present Iliad, was passed along down the generations, various meddlesome editors augmented it with inferior materials: irrelevant digressions (including the large-scale importation of materials from entirely different stories), re-arrangements of the original Homeric material, and innumerable pointless and distracting repetitions. The critic’s task was to remove the later accretions and get back to the hypothesized original poem. His method was to find the joints where the inferior material had been attached to the true body, and to cut. Each cutter cut differently.

This approach was called Analysis, since it analyzed the text into different layers according to date (which was conjectured) and authenticity (which was also conjectured). The Analysts’ overall interests and aims, like Wolf’s, were historical: they sought to recover the stages by which the text as we have it came into being. The opponents of Analysis, called the Unitarians, tried to demonstrate the artistic integrity of the texts as they stand (their “unity”). Where Analysts undertook to reconstruct the history of the text from the nature of the defects they perceived within it, Unitarians countered that the defects were not there to begin with.

In the late 1920s and early 1930s the American scholar Milman Parry undertook to investigate (among other things) the repeated “name-epithet” combinations that are such a conspicuous feature of Homer’s language: “Ox-eyed Hera,” “Menelaus of the loud war-cry,” and so on. Why does the poet decide to call the chief hero of the Iliad “swift-footed Achilles” in some places, but “brilliant Achilles” in others, and perhaps something more distinctive (“swift-footed brilliant Achilles”) in yet others? Parry demonstrated that the names and epithets combine into units having different metrical or syntactical values: one such unit is available in each common metrical length, for each grammatical case (the nominative or subject case, accusative or object case, etc.). Such a system seemed to be designed for practical use. Parry identified the use as oral composition, and he found corroboration for this hypothesis in the practice of contemporary Yugoslavian bards, who maintained a vital tradition of epic composed in performance and characterized by a high density of formulas.


Parry’s theory is persuasive because it is systematic. It is entirely obvious that there are a number of frequently repeated phrases serving to designate Achilles. Parry demonstrated that these phrases unite with one another in a function system, and in doing so transformed our understanding of what a “formula” is. From being a free-floating, pointlessly repeated phrase, “swift-footed Achilles” becomes, in effect, one of a group of co-workers who make up a corporate system “Achilles”: it is the function of this Achilles-system to provide the composer, whenever he needs it, with an Achilles-designation in whatever metrical size or shape and syntactic function the context demands.

But where does the context come from? This is where things become difficult again. Parry’s theory can explain a great deal. For example, repetition—of words, phrases, lines, speeches, even whole scenes—is no longer a problem needing to be solved, but an expected, built-in feature of the compositional technique. But is everything in Homer a formula in Parry’s sense? During the theory’s early days of acceptance, scholars tried to show that everything was. The continual inflation of the importance of prefabricated formulas, however, was at odds with most readers’ sense that the individual scenes and speeches of these poems have been uniquely tailored to the dramatic, psychological, aesthetic, ethical—in short, artistic—needs of their contexts. Much of Homeric interpretation since the 1950s and 1960s, when Parry’s theory became widely accepted, has been devoted to trying to strike the right balance between these two competing considerations.

Professor Nagy, a pupil of Parry’s pupil and collaborator Albert Lord, has explicitly sought to continue and extend the work of these predecessors in exploring the traditional nature of Greek epic composition.4 The radical new claims he now advances are best appreciated against the backdrop of current scholarly opinion.

An orthodoxy has been established for some time, but it is one that tolerates different chronologies: for scholars who accept Parry’s theory, the question is “How and when did poems that were composed orally get written down?” For those who reject the oralist doctrine, “When were they written down?” is simply “When were they written?” The range of dates proposed roughly spans the eighth century to the sixth century BCE, with most scholars favoring the late eighth and early seventh centuries as the period in which the poems can be said to have attained, one way or another, their (approximately) final form.

That is just the beginning, however. We still have to get the poems down to the Hellenistic period—from the third through the first centuries BCE—in Alexandria, during which scholars produced the editions that lie behind the modern texts. In order to reach the Hellenistic era, however, the Homeric poems had to traverse the fifth and fourth centuries—no less than the entire Classical Age of Ancient Greece. This was a period of enormous literary production, and we know next to nothing about how any of it was passed on through the generations. We know more about the text of Homer during this period than about that of most authors, simply because he was quoted more often. The evidence does not suggest a high standard of accuracy in reproduction. Plato and Aristotle (fourth century BCE) quote Homer frequently, often presenting texts that depart from our own. But they quoted from memory. More significant are the substantial papyrus fragments that survive from early Ptolemaic Egypt (late fourth century to mid-second century BCE). These present a strikingly large number of readings that diverge, mostly in trivial ways, from the text transmitted in the medieval manuscripts. After about 150 BCE the papyri (and there are many of them) lose this “wild” aspect, and begin to reflect what looks like a standardized text (precisely where it came from and how it was established are two of the major outstanding questions). From this vulgate text the medieval tradition takes shape and brings the poems to the modern era.

There has been, then, within these limits general agreement that by the end of the sixth century BCE it is reasonable to speak of such a thing as a text of Homer. This was the text from which the divergent texts diverge, the true text that the great Alexandrian scholar Aristarchus (second century BCE) and his predecessors sought to restore in the Hellenistic era, and that we wish we could restore now. Departures from this text are to be explained by reference to the processes of corruption that beset all scribal transmission: there is Homer on one side, and on the other there is a lot of variorum non-Homer. The goal is to get back as close to Homer as possible.


Nagy’s radical new proposal, set forth in Homeric Questions and Poetry as Performance, stands all this on its head. There is no Ur-text, he says; there is, in effect, no Homer, or any one “monumental composer.” There is a vital tradition of epic performance in which, before many audiences, many singers over many years (and in many places) compose, by drawing on the storehouse of traditional formulas and themes, many Iliads, many Odysseys, as well as many other versions of the traditional heroic tales. This tradition remains vigorous and pliant, with some local variations, until well into the first stages of the Alexandrian period. (Athens in particular suffers an early and influential case of what Nagy calls “sclerosis,” i.e., progressive fixation of the previously fluid songs.) It is this performance tradition, Nagy argues, not the carelessness of scribes, that is the source of most of the variants that we see in the early copies and quotations. It follows that such “variants” must not be condemned as corruptions of the one true text and then consigned to the outer darkness of footnotes and critical commentary; rather they are to be brought into the textual sunshine of the upper part of the page, as presented in a new “multitext edition” that will honor these former outcasts as the preservers of “authentic tradition.”

This theory accords with other recent efforts to find in classical Greece an “oral culture” lacking, among other things, any attachment to the ideas of verbatim quotation, textual fidelity, definitive versions, and so on, all of which are seen as fetishes of literate cultures like our own. A poem (or better, song), according to this oral- culture hypothesis, is its performance; but its performance is not the performance of a “text” in the way in which we conceive a performance of, say, Don Giovanni as realizing the fixed musical and verbal texts by Mozart and da Ponte. If such fixed texts were known at all in the early Greek world, they would have been rather marginal. Authoritativeness, to the degree that it existed at all, was possessed by performances—“events” in which the specific time, place, audience, and occasion all continuously contribute to shaping the song as it is being produced. Thus there were different Iliads for different people.

Here, then, is a new model for the transmission of the Homeric poems. Nagy cites as a parallel case the oeuvre of the twelfth-century troubadour Jaufré Rudel, whose lyrics survive in different versions tailored by the poet for different audiences and occasions. The editorial doctrines of a “New Philology” in medieval studies have sought to discourage the old imperative to eliminate “false” readings; for such readings, as in the poems of Jaufré Rudel, may simply preserve authentic variants of the author’s.5 R.T. Pickens’s edition of Jaufré Rudel6 is cited by Nagy as a model for the sort of new “multitext” edition of Homer he himself now urges. Thus we resolve the old task of separating the true text of Homer from the false variants of not-Homer by adopting it as our working, or at least preliminary, hypothesis that all defensible variants are true “Homer,” the name now serving to designate the tradition as a whole. Authentic “Iliads” were being created in performance right down to the period of Alexandrian editing, when finally a text became fixed and standardized.

This obviously raises the question: Who was doing the creating? The orthodox view that Nagy seeks to replace makes use of two different concepts of epic performance: that of the aoidos, the master of the compositional techniques of the epic tradition, who creates in performance, as an artistic act, the poem eventually transcribed, and that of the “rhapsode,” who is uncreative, but is the master of a prodigious memory, and is able to memorize and reproduce on demand the creations of the aoidos. The usual view has been that after the poems were transcribed or frozen through other means, the performance of them was basically entrusted to rhapsodes.

It is reasonable to accept that a certain degree of mastery of the creative compositional technique—knowing what combinations can go where—would have been a great aid to memorization. To that extent at least, Nagy is justified in questioning the usual picture of the rhapsode as a mindless mechanical performer limited to two functions, “record” and “playback.” But mastery of creative technique to that limited degree is also consistent with the more usual belief that what the rhapsodes were mastering was an established text. The major problem for Nagy’s new theory is simply that the variant recordings that we know of from the papyri and the indirect sources mentioned earlier are for the most part too boring and insignificant to imply that they derived from a truly creative performance tradition.7 I assume that this is why Nagy postulates what are in effect varying degrees of creativity in a variety of local traditions. But we are still left wondering if the banal repetitions and expansions that we find in various papyrus scraps really require us to accept, in order to explain them, a full-blown oral-performance tradition. (Publishers who sold texts at a per-line rate have been suggested as another source of opportunistic expansion.) It is hard not to suspect that the theoretical tail is here wagging the evidentiary dog, and that sheer love of the oral-performance theory is leading Nagy to discover it in every possible nook and cranny of the ancient world.

But the variant readings are ultimately what must concern us. A multiform text showcasing variants is in the end going to be only as valuable as the readings it puts before us. It is notable, then, that Nagy scarcely enters into the discussion of particular readings. Or rather, he discusses one variant reading at terrific, even stupefying, length, and is mostly silent about the rest. He prefers propounding the theory of variants to discussing the variants themselves. He is like a Dostoevskyan political idealist who loves the oppressed masses in the abstract but not in the flesh.

The one favored variant that he does discuss is quoted to us not in a pre-Hellenistic author or papyrus, but by the second- to third-century CE writer Aelian in a work devoted to remarkable facts about the animal kingdom (On Animals 5.38). The passage Aelian cites is Odyssey 19.521, occurring within a simile contrived by Penelope to illustrate the nature of the grieving fantasies that assail her nightly in Odysseus’ absence:

As when Pandareos’ daughter, the greenwood nightingale,/perching in the deep of the forest foliage sings out/her lovely song, when springtime has just begun; she, varying [troposa]/the manifold strains of her voice [polyechea phonen], pours out the melody, mourning/Itylos, son of the lord Zethos, her own beloved-child, whom she once killed with the bronze when the madness was on her….

(Od. 19.518-523)

Aelian reads in line 521 not polyechea, “[having] manifold strains,” but polydeukea—“the meaning of which is unclear,” as Nagy puts it. Defense of this variant occupies a substantial part of the book, and serves as a practical illustration of the kind of reward we can hope, with Nagy, to recover from the storehouse of neglected variant readings. If it should surprise some readers that Nagy would choose as the representative for the entire class of neglected variants a word of unknown meaning, it will not surprise veterans of his previous works to learn that there waits in the wings an etymological proposal to render the unknown known.

Nagy connects the stem deuk– seen in polydeukea with that also shown in the Latin ducere, “to lead,” defining it as “to draw continuously toward a definite goal.” In Penelope’s simile the important result is that, with Aelian’s reading polydeukea, the nightingale varies (troposa) her voice (thus her song) continuously. Nagy then points out that in one of Jaufré Rudel’s songs a nightingale is said to mover his song, with the Provençal word mover used in the same meaning as Greek troposa, “to vary.” This nightingale, he writes, can be taken to stand for the troubadour Jaufré Rudel himself—and for Homer, too; for tropao, the verb with which Homer tells us that Penelope’s nightingale varies (troposa) her polyechea or polydeukea voice, is etymologically cognate with the very word “troubadour.” And like Homer, the nightingale is engaged in producing variants (troposa again)—variants of the kind which the competing traditions of this very passage exemplify.

With such arguments, and ones even more imaginative, Nagy proceeds, over several chapters, to extract from Penelope’s simile further coded insights into the problem of Homeric composition and transmission: “The variant epithet of the Homeric nightingale’s voice in Odyssey 19.521, polydeukes, ‘patterning in many different ways,’ applies to Homer himself and—just as important—to those who perform Homer.”

On the basis of this observation Nagy argues for the legitimacy of variants throughout the centuries-long performance tradition of the Homeric poems: for him all the performers are, in effect, Homer. And he means this as no mere conceit. In Sappho’s “Hymn to Aphrodite” the poet represents the goddess as addressing her by name in direct speech, not, one might think, a difficult or particularly pregnant piece of dramatization. But for Nagy, the poet—or performer of the poem—who starts out the poem as “Sappho,” is now, in Aphrodite’s speech, “in effect re-enacting the goddess.”

At this point we might think: “All right: the performer ‘is’ Aphrodite here in the same sense that an actor who plays Agamemnon in Aeschylus’ Oresteia ‘is’ Agamemnon.” But that is not all that Nagy means. As a protracted comparison of Sappho’s poem with a Native American ritual song from the Navajo and the Apache Changing Woman ritual indicates, he is thinking of literal identification: “While the music of Sappho lasts, Aphrodite is present, and whoever performs the music is Sappho, is the music of Aphrodite.”

Nagy’s formulation here strikingly recalls Euripides’ Bacchae 115ff: “whoever leads the [Dionysian] bands is Bromios [i.e. Dionysus],” a text influentially used by E.R. Dodds to support his claim that identification with the god was a feature of Dionysiac ritual.8 The trouble is that that version of Euripides’ text, which was based on an emendation, has been thoroughly discredited. In a meticulous study Albert Henrichs has related the history of the earlier reading, which he says drew on “the dubious assumption, ultimately rooted in Romantic tradition, that followers of Dionysus identified themselves with their god at the height of the ritual.”9 In fact, the kind of ritual identification that Nagy seems to be claiming (“whoever performs the music is Sappho”) is hard to find in ancient Greek culture. Sappho’s poem offers no evidence for the phenomenon, and of course the Navajo and Apache Changing Woman ritual has no evidentiary value for the Greeks.

The search for hidden and surprising meanings, often with dramatic implications, is characteristic of Nagy’s writings. He reconstructs missing traditions—above all, oral-performance traditions—in order to interpret, or re-interpret, texts that survive—above all, those of Homer. In Homeric Questions he restates and defends a provocative proposal he made in an earlier book, The Best of the Achaeans.10 Nagy based much of his lengthy argument in this work on the hypothesis that there existed a rich epic tradition (now lost) about a quarrel between Achilles and Odysseus, the respective protagonists of the two surviving epics, on the question of how Troy was best to be taken, with Achilles advocating force, Odysseus trickery. Hidden traces of this other epic tradition, Nagy thought, could be detected in our surviving Iliad and Odyssey, where, on the surface at any rate, the two heroes appear to enjoy amicable relations.

Nagy saw in this theory a clue to solving one of the most intractable puzzles posed by the text of the Iliad, the infamous occurrences in Book 9 of the grammatical forms calls “duals,” and he returns to this issue in Homeric Questions. The dual is an archaic category of inflection in Greek: as singular forms are used for one entity, and plural forms for more than one, so the dual is used for two: “he walks,” “they walk,” “the two of them walk.” In the ninth book of the Iliad an embassy is sent by Agamemnon to Achilles to try to persuade him to return to bat-tle. The ambassadors are Odysseus and Ajax—both, obviously, extremely well-known figures, the latter, in some traditions, Achilles’ cousin—and an otherwise barely mentioned old man named Phoenix, purportedly Achilles’ tutor when the latter was a child. The three ambassadors set off, but some of the forms applied to them are duals: “the two walked along the beach,” and so on. Even more obtrusive are the duals in Achilles’ greeting: “Welcome, the two of you! Two friends indeed have arrived—the need must be great—friends who are the two of the Greeks dearest to me, even in my present state of resentment.”

The usual explanation is that the passages having to do with Phoenix have been inserted into a text that originally described an embassy of two, Odysseus and Ajax. Whoever added the third ambassador, Phoenix, did so without making the appropriate adjustment of the dual forms to plurals. But what medium preserved these now unwanted duals, against, as it were, the will of the context? If Nagy is right that until quite late the poem existed only, or primarily, in performance, then surely the performers would have made the obvious and easy adjustment?

Nagy therefore has reason to seek a different explanation for these duals. He proposes that in them we have a remnant of the tradition, otherwise largely concealed in both epics, in which Achilles and Odysseus are enemies: Achilles uses dual forms in order to exclude—to snub in the most blatantly hostile fashion—Odysseus, his established enemy, the exponent of guile and deceit, and to include only Phoenix and Ajax. Further evidence for this hostility comes after Odysseus makes his plea for Achilles’ return, citing, among other inducements, the numerous goods offered by Agamemnon by way of reparation of their quarrel. Agamemnon had earlier, out of Achilles’ hearing, concluded this catalog with the obnoxious words “Let him be tamed!… Let him yield place to me, inasmuch as I am the kinglier and inasmuch as I can call myself born the elder!” Odysseus omits these offensive lines from his otherwise almost verbatim repetition of Agamemnon’s offer.

Achilles perceives the deception, a convenient adjustment of reality so typical of Odysseus’ shifty character, and so perfectly representative of the ends-justifies-the-means approach to life that Achilles most detests about him. Achilles replies:

without consideration for you I must make my answer,
the way I think, and the way it will be accomplished, that you may
notcome one after another, and sit by me, and speak softly.
For as I detest the doorways of Death, I detest that man, who
hides one thing in the depths of his heart, and speaks forth
another.But I will speak to you the way it seems best to me.
(Iliad 9.308-314)11

With “the man who hides one thing in his heart and speaks forth another,” according to Nagy, Achilles obliquely, but unambiguously, refers to Odysseus, for whom he thereby declares his hatred.

This proposal of Nagy’s drastically alters our understanding of the spirit in which the embassy is conducted. Or rather, Nagy’s interpretation violently distorts the text: apart from the friendliness that characterizes the relations between Achilles and Odysseus in both epics, the idea that in this particular scene Achilles detects and resents Odysseus modification of Agamemnon’s words would make sense only if Odysseus had claimed to be reporting them, which he never does.12 He simply says that he will tell “how many things Agamemnon promised,” and then produces the catalog, identical to Agamemnon’s but with Agamemnon’s “I will give” (etc.) changed to “he will give” (etc.). Odysseus ends where the catalog ends, without ever having promised or implied that he was even so much as “giving the gist of Agamemnon’s position,” much less a complete verbatim repetition of his words. That Odysseus uses the same lines as Agamemnon had, is, as Nagy knows, simply an expected feature of oral-formulaic composition. It is of course paradoxical that Nagy, the tireless campaigner against textual fixity, should make so much out of Odysseus’ supposed fiddling with the verbatim text of Agamemnon’s speech.

But what is even more disconcerting about Nagy’s interpretation is that it makes of Achilles precisely what he most objects to: a dissembler. When Achilles says, “I hate that man who thinks one thing and says another,” he uses the indefinite formulation (literally: “whoever thinks one thing…”) for precisely the reason that indefi-nite formulations are commonly used, namely, because he does not have any one person in mind, but is speaking of a class, a type from which he wishes to distinguish himself emphatically. His purpose in starting with this observation is to warn his audience—his dear friends—that he is going to speak bluntly, even harshly—which he does, rejecting not only Agamemnon’s gifts but also a desperate plea that he help the other Greeks in their mortal plight. The poet himself offers us some guidance at the speech’s end: “They [i.e., Odysseus et al.] were silent, stunned by his speech.” No more is needed to make sense of Achilles’ opening words.

Criticism of Nagy’s proposals about this scene, and the dual verb forms used in it, has been made by, among others, Jasper Griffin,13 of Balliol College, Oxford, whom Nagy answers at length in Homeric Questions. The disagreement between the two is significant—a continuation, in a sense, of the old war between Analysts and Unitarians. Even if one were to accept Nagy’s argument about Achilles’ use of the duals in his greeting, there would remain unexplained the duals in the narration: “The two walked upon the beach,” etc. Nagy argues that with these duals the narrator is adopting, in advance, Achilles’ point of view. Nagy cites some comparative evidence for such a procedure, concluding that “the perspective of the narrator of the Iliad becomes dramatically identified with the perspective of the main hero of the narration, Achilles.” That is, Achilles hates and excludes Odysseus, and therefore so does the narrator.

The problem here is that nowhere else does the Homeric narrator so spectacularly adopt the “perspective” of a character. We know and do not just speculate that Achilles hates Agamemnon, and in Book One he is about to kill him. Here, if anywhere, we might expect a narrator disposed to adopting his main character’s perspective to do so. But the narrator does not say, for example, “The great drunken dog-faced coward Agamemnon answered in reply”—things Achilles actually calls Agamemnon right then and there—but rather says respectfully, “Answered him then lord of men Agamemnon.” In short Nagy’s theory violates Homeric poetic practice, and it is this tendency that probably most concerns Griffin in his continuing polemics against Nagy. For it must be remembered that what is driving Nagy’s startling inversion of the text is a speculative theory about an antecedent tradition of a bitter enmity between Achilles and Odysseus. The imagined tradition that is not there is allowed to take charge of the poem in our hands, to subvert the painstakingly crafted work of art before us.

M.L. West concluded his review of Nagy’s The Best of the Achaeans with words that have been quoted on the back of the paperback edition since it appeared: “One is left with the uneasy feeling that curtains have parted in the wind, giving glimpses of unsuspected realities behind the apparently simple face of Greek heroic poetry.”14 Nagy’s penchant for this curtains-parting-in-the-wind effect, his desire to take a counterintuitive position and impose it through sheer brilliance of argumentation, has weakened his work. The decision to organize the argument of Poetry as Performance around the nightingale simile, speculatively explained and dumbfoundingly over-interpreted, is not uncharacteristic. When working, or overworking, this nightingale conceit Nagy’s writing is often hard to follow, for it fails to make logical relations clear. Transitions are often obscure, accomplished through exhortations by Nagy that seem arbitrary: “Let us return here once again…,” “Here the chapter comes to a halt. But the subject of the nightingale will continue…,” “Let us continue where we left off with the song of the nightingale….”

In the fifth chapter of the book, however, “Multiform epic and Aristarchus’ quest for the real Homer,” Nagy presents the ideas about the Alexandrian stages of the Homeric text that I have briefly touched upon above. The chapter—the longest of the book—is vigorously and lucidly argued. Coming upon it in the context of the rest of the book is somewhat startling but pleasing. Would that this gifted scholar give up his own quest for Homers that are not there, and instead hold himself to the standard and style of the perhaps not so exciting but undoubtedly more durable work suggesting how the text we call Homer’s might have taken shape in Hellenistic Egypt.

This Issue

November 20, 1997