‘Dividing Homer: When and How were the Iliad and the Odyssey divided into songs?’, Symbolae Osloenses 74 (1999) 5-91

The respected Norwegian journal Symbolae Osloenses organised a useful debate between M.S. Jensen (henceforward J.) and other scholars as to the origin of the Homeric book-divisions, and invited a wide range of responses. Extreme overwork prevented me from participating at the time, but it may be useful to review the debate in this forum instead.

It is odd that both Homeric epics are each divided into 24 books according to the 24 letters of the Ionic alphabet. J. reviews the theories as to when this arose: some trust pseudo-Plutarch’s claim ( Vit. Hom. 2.4) that οἱ περὶ τὸν Ἀρίσταρχον were responsible, others think it must be earlier, either Peisistratid or even earlier. Again, some claim that the book-divisions are carefully and skilfully done, others that they are not; some, that they go back to the poet himself, others that they do not. Before reading the debate, to the extent that I considered the matter (which did not seem of much importance), I thought they were Alexandrian (and certainly later than the Argonautica), not particularly skilful, and not derived from the poet himself. After reading it, I am more inclined to accept that they date back to at least the sixth century, but certainly not that they began with the poet himself; nor have I been persuaded that they exhibit any particular artistry, different in kind from that which pervades the compositional techniques of both epics.

J. gives a good survey of scholarly opinions. Among the external evidence, she usefully highlights the fact that two early papyri of the Odyssey respect the book-division, even though others did not, and the arguments of S. West against an Alexandrian origin for the book-divisions, which are decidedly more cogent than I had appreciated. The review of internal evidence is, as J. recognises, bound to be subjective (it is effectively criticised by B. Heiden, p. 55), and I have never found the arguments for the unity of particular books particularly cogent. Above all, as M. Edwards and I. de Jong point out (pp. 52-3, 63), to extract cogent conclusions from the analysis one needs to consider all scene-changes within the epics, not just those between books.

From her conclusion that the book-divisions are as much a feature of the poems’ artistry as any other unit of epic style (p. 22), J. goes on to show how she is bound to interpret this in the light of oral-formulaic theory. As someone with actual experience of field-work (in Albania), she is courageous, and absolutely right, to insist that this approach must be the starting-point of her interpretation; it is a misunderstanding to suppose, as does Ballabriga, whom she cites, that anyone who adopts it becomes a ‘prisonnier de l’oralisme’. Instead, I would compare the theory to an axiom in mathematics: it cannot itself be proved, but without it little else can be explained. J. maintains the use of the term ‘oral theory’, accepting (but perhaps not clearly enough) Lord’s point that it is not theory to insist that we know how the recorded or dictated texts in the Parry Collection at Harvard were created, whereas it is a theory that the Iliad is another such text. However, she wishes to adopt the word ‘text’ to indicate oral compositions as well as written texts, and speaks of a mental ‘text’; this innovation is rightly questioned by [Oslash]. Andersen (pp. 37-8). J. later speaks of an ‘oral technique of memorizing’ (pp. 33-4), which the rhapsode used in performing his text for dictation. This too is confusing, since oral poets do not memorize verses but remember phrases and story-patterns; Andersen and Cantilena rightly object (pp. 37-8, 47). J. later states that she sees no distinction between a creative oral poet and a repetitive reproducer, but it is surely essential to distinguish between aoidoi and a person like Plato’s Ion.

J. next discusses how the poems came to be written down, making telling points against the hypothesis of some kind of gradual textualisation, above all noting that the oral theory made the enormous contribution of uniting at once unitarianism and analysis, as well as matching the observations of folklorists and anthropologists ever since the Singer of Tales appeared (pp. 27-8). J. is absolutely right to hold that the Homeric poems are dictated texts; this was, as I showed in CQ 48 (1998), Parry’s own position (M. Parry, The Making of Homeric Verse, Oxford 1971, p. 451). She advocates, however, a dictated text procured from the poet(s) by the Pisistratids, one book per day. The pace of dictation is perhaps about right, but such a level of detail and so romantically attractive a hypothesis will be held by her opponents to detract from the hypothesis of a dictated text, which is the only one that explains the nature of the Homeric poems. Her hypothesis obliges her to assume that the poems were dictated in Athens in the East Ionic alphabet with its 24 letters. I agree that the earliest texts of Homer were in this script, but it does not of course follow that they originated as late as the 6th. century.

Among the responses the most significant are those of Andersen, Edwards (see above), J.S. Clay, M.L. and S. West, and M. Cantilena.

Clay (pp. 50-2) is rightly troubled that many Homerists are not taking account of the arguments of historians for dating the Homeric poems even earlier. The Wests (pp. 68-73) rightly remark that the term ῥαψῳδία for a Homeric book was not likely to have been introduced by the Alexandrians and deduce that this is likely to reflect Panathenaic performances. They rehearse the strong epigraphic grounds for thinking that the poems were written down before the Peisistratids. However, they envisage the process of creation as successive written drafts (how easy would this have been, with the technology then available?), and in an ungallant manner challenge the originality of M. Parry, as if he were not aware of the antecedents to his work, and as if the existence of such antecedents detracted from the validity of his approach or the magnitude of his achievement (this is rightly rebutted by J. on p. 74).

M. Cantilena fully accepts J.’s contention that the poems must be oral dictated texts, and that the date of the dictation is the date of Homer (pp. 47-50). He remarks, however, that J.’s dating of Homer does not consider the evidence of the language of Homer when compared with that of the other poems in the tradition, and that in my book of 1982 ( Homer, Hesiod and the Hymns, Cambridge) I showed that the fixation of the text of Homer must antedate that of Hesiod. (I would add, lest it be thought that I ignore my predecessors, that in this I gave a sound statistical basis to conclusions which G.P. Edwards and A. Hoekstra had already reached on other linguistic grounds.) In reply, J. cites ‘two basic problems’ in my work, problems to which I pointed myself — ‘the question of a steady rate of change in rhapsodic diction, and the lack of chronological fix[ed] points’ (p. 82). However, to mention these simply will not suffice as an objection. Given that the relative chronology is so statistically secure, if the poems of Homer are 6th. century, then the poems of Hesiod are too, and are at least one day younger; and the Homeric Hymns are younger yet. If all these poets were contemporaries of Pisistratus, it is astonishing that Herodotus or Thucydides, not to mention Aristotle, do not know about it. It is also hard to explain how Hesiod’s Works and Days could have been imitated (as it is) by Semonides and Alcaeus. On the contrary, A. Ballabriga would even have us believe that Homer alludes to Hesiod and Stesichorus (p. 42), which is impossible if Hesiod is even a second later than Homer! To me, J.’s most serious lapse is her rejection, without giving reasons (p. 33 n. 39), of C.J. Ruijgh’s linguistic and epigraphic arguments for the early dating of Homer, even though he too believes in an oral dictated text. That a number of scholars have recently pronounced for a 6th.-century date proves nothing: authorities must be weighed, not counted.

In short, this debate has been useful, rendering it somewhat less likely that we should trust the statement of pseudo-Plutarch that the book-divisions were created by Aristarchus. I remain unconvinced that they have anything to do with the poet; Parry was right to dissociate the division and numeration from the composer ( op. cit. 454-5). However, the stacks of leather or papyrus rolls which contained each poem needed numbering, especially under the Panathenaic rule for rhapsodic performance. Even so, my original questions remain: if the system was pre-Alexandrian, why are Apollonius’ books as long as they are and how did the story about Aristarchus arise?