James Dee (henceforth Dee) continues his Herculean, perhaps specifically Augean, labors. Having produced his volumes on epithets of men (2000, see my review at BMCR 2001.01.14), gods (2001), things and places (2002), and mere vocables (2001, “four sets of word-frequency lists”), he now offers two volumes containing two sets of information. First, he provides texts, the Iliad and Odyssey verse by verse, showing to which of the 32 metrical patterns (such as the most frequent “five dactyls” verse) each verse belongs, along with a graphic presentation of each pattern. E.g., A 331 (two spondees, three dactyls, final spondee) appears as lllllsslsslsslx, pattern 4. Second, he provides a double “sort” divided into the order of his 32 schemata, each array repeating in alphabetical order the verse-initial word(s), whether they contain recognized formulae or not. Thus verses beginning with a dactyl and spondee (e.g., aien aristeuein) could occur in eight different sorts—but this formula occurs in only one. Non-initial cola (and their metric patterns) therefore remain buried, but one book cannot do everything. One quickly observes that the one pentadactylic pattern requires 95 pages for its nearly 20% of the verses of the two epics; while the seventeen least common patterns (5.5%; 16 with spondaic fifth foot) together occupy only 26 pages.
A chart which one is forbidden to copy but which will hypnotize all Homerists offers the 32 patterns of 12 to 17 syllable verses in two sets. The first part lists the patterns with the dactylic fifth foot, and the second part lists the patterns with spondaic fifth foot. Both parts descend from the most common pattern DDDDDS (5378 verses of the two epics, or over 19%) to the least common, the hexaspondaic (a word not likely to make the next edition of the OED) SSSSSS (5 verses or .02%). Dee prefers the more graphic presentation of the two metrical feet as lss and ll to D and S, and that form he provides in the texts, “versuum ordine.” For those who have forgotten, dactylic fifth foot lines outnumber the spondaic 19 to 1.
Dee refers to Homer’s (please picture quotation marks on that venerable name) “seemingly inexhaustible variety of patterns” of verse. There are the 32 verse possibilities or permutations, but then any combination of those 32 limited types may appear in a speech or narrative. It is hard for me to conceptualize Homer, Homers, any Homer singing along and wondering if one metrical verse pattern or another, or two of the same shape should trump what he thought he had to say next. Of course, I am not Homer or even a plain vanilla Montenegrin bard, so that puzzlement means nothing. As Martin West wrote (in I. Morris and B. Powell, eds., A New Companion to Homer [Leiden 1997] 232), “epic poets were more concerned with fluent and coherent utterance than with polished versification.”
On the other hand, depending on your cola preference to be sure, you can divide a particular hexameter differently: two cola, tricola, or Hermann Fränkel’s four cola of rhythm and sense. Yet even the distinguished Homerist Geoffrey Kirk has admitted uncertainty as to the definition of a colon or a break ( tome, caesura; see YCS 20 [1966] 104). Milman Parry’s performers seem to think in both cola and larger units yet.
Analysis is left to Dee’s grateful audience, but at xiii note 7 he cites several standout facts. There are five consecutive runs of repeat patterns at E881-92, one run of the most common, pentadactylic schema (1) in seven consecutive lines (H145-51) and another of six (
All collectors of the tidbits of Homer, from Aeschylus on, know that every form of analysis has its own history. This self-styled “strangest of all reference works published on the Homeric epics” acknowledges two important predecessors, Jacob La Roche “Zahlenverhältnisse im homerischen Vers,” WSt 20 (1898) 1-69, and Frank Jones and Florence Gray, “Hexameter Patterns, Statistical Inference, and the Homeric Question: An Analysis of the La Roche Data,” TAPA 103 (1972) 187-209. The latter statistical study was not the only article of that fat 600 page annual that I did not rush to read upon receipt. Nor was it the last one I would have read, had I been shut in a Siberian Gulag instead of relocating to bustling Philadelphia by the Schuylkill. My copy certainly, however, remained unread until Dee drew my attention to it.
Dee has produced his own “Brand X” text, admittedly more for licensing reasons than for the opportunity to create a new Homer. He borrows features for his Iliad collation from Ludwich, Monro and Allen, van Thiel, and West, and for his Odyssey from Ludwich, Allen, Be/rard, Von der Muehll, and van Thiel. Such features include no indentations, no quotation marks, no lunate sigma, and no adscript iota, but (arguably, on the positive side) offsetting vocatives with commas, capitalizing gods’ epithets when used for personal names, and providing dactyls where some editors read spondees. Although Dee eschews quotation marks, at the end of each line in direct speech or speech embedded within speech he supplies a wavy equal sign, two of them for the embeds. He includes about 180 verses thought by various editors to have been interpolated.
Volume II provides a repertory arranged by the metrical patterns. The Homeric epics number 27, 803 verses in the Dee edition. No Homeric hexameter is hypermetric, final hiatus between verses was never objectionable, and every line ends with a spondee (or trochee). All these facts promote the welding of the parts of each chanted line into one separate unit, very often also a complete sense unit (with enjambment as a notable variety).
Nine patterns appear in more than a thousand examples each (all with dactylic fifth foot), eleven number in the hundreds (six with dactylic fifth foot), ten number in the tens (only one of them with the dactylic fifth foot). Finally, two types of spondaic fifth-footers number less than ten, five apiece for second foot dactyl and for verses with no dactyl. The Iliad and the Odyssey have high correlations of metrical types of verses (“almost … unity”); the statistical study of Jones and Gray gives these as percentages with chi-squares, product-moment correlations, etc., etc. (and other aspects of statistics that I do not understand). The results suggest common “authorship” of the two epics. The negligible differences in metrical patterns presumably signify one genius at work, rather than two or twenty (192, n.16).
For one interesting comparison, dactyls found only in feet 1 and 5 (DSSSDS) the commonest pattern in the Latin of Vergil’s Aeneid, ranks only eighteenth in the Iliad (Jones and Gray 198; 111 examples in this poem, or 0.71%, by Dee’s count). The high proportion of short to long syllables in Greek, in Homer’s Greek especially, renders this verse form infrequent.
Neither stylometric nor versometric (?, my neologism) data nor detailed comparisons with Hesiod, the Homeric hymns, and Apollonius’ epic can prove the hypothesis of single authorship. Dee’s array of data, nevertheless, will likely promote the minimalist’s case that the two epics proceeded from the same bard(s) at the same time “combining an elaborate technique with a straightforward manner of telling a story.” (Thus wrote Maurice Bowra in “Metre,” A Companion to Homer, ed. A. J. Wace and F. Stubbings [London and New York 1962] 29, apparently before reading Richard Armour’s chapter in The Classics Reclassified, in which certain famous books are not so much digested as indigested, together with mercifully brief biographies of their authors, a few unnecessary footnotes, and questions which it might be helpful not to answer [New York 1960]. More recent “companions” for Homer diverge, as they emerge from the academic presses of Brill [1996] and Cambridge [2004]. The former, A New Companion to Homer, has Martin West’s good chapter mentioned above discussing “Homer’s Meter.” The Cambridge Companion to Homer, edited by Robert Fowler and expected any day now, does not contain a chapter on Homer’s unique metrics, perhaps because of its focus on literary (rather than musical and technical) matters.)
If one examines a group of identical verse-starts such as, e.g., speeches beginning Troes kai Lykioi kai, various facts leap out (see Dee II. 300). There are seven such verses, all in the Iliad, all but one identical to the end (8.173, 11.286, 13.150, 15.425 & 486, 17.184, and 16.564). All show “schema 6”: lllsslllsslsslx (which I find less graspable than SDSDDS). They all appear within seven books. All but one are correctly noted in the right margin as occurring in oratio recta, but you need to go back to the left margin of volume I in seven different places to see if a verse of its identical metrical pattern follows. Universally (seven of seven) and not surprisingly, such a verse does not follow, but generally (four of six, and still not surprisingly), a word-identical hortatory fight verse does follow specifying the “vocatived” allies (8.174=11.287=15.487=17.185). Similarly, you must return to the text in volume I in order to discover how many individual “voices” speak these same words (in fact, two: Hektor for the six speeches, “Homer” for the singleton).
The “headless” ( akephalos) verse can be studied here but not the “mouse-tailed” ( misouros) stichos, since the feet are indexed by the first words of the verses — not those of the fifth or sixth foot. Thus verses beginning with a foot lacking a mora such as
We thank both Professor Dee and the Olms Verlag for services no mortal classicist ever expected to be rendered, with all our due appreciation of such obsessive labors. My wife, although puzzled that such human ingenuity and resources would be devoted to such arcana, allowed herself to be persuaded that it was a valuable tool for a very limited group of people, hoi oligistoi (we might call them).
One mentions only with reluctance that the production of both volumes disappoints. These volumes are not modestly priced. The more indefensible problem is that the initial and final pages fall out with the first opening of these volumes that should have been bound for the ages. Not since reviewing the wretchedly bound and thus ominously misnamed Curse Tablets and Binding Spells from the Ancient World (Oxford 1992) have I encountered such a sloppy and inadequate job of securing pages between covers. No reference book should have a so-called “perfect” binding, especially when the publisher goes to the trouble of providing linen covers and acid-free paper. In an unprecedented oddity of production, my copy arrived with two extra one-sidedly printed copies of volume 1, p. 444, unbound but inserted. None of these irritations is the fault of Professor Dee, in whose immense debt we stand.
As I dig trial trenches for treasure in the arrays of part II, these varied thoughts twitter and squeak like bats around my head. More experienced metricians may see straighter and shorter routes for exhuming the golden apples of Dee’s fifth labor; I am confident there will be interesting results. Parerga for the future will include repertories for Vergil and Ovid. In the parlance of the just completed Baseball World Series, Dee has hit another Homer.