¹ Il. 24.535–42.

² Among the translators and commentators who make παναώριος carry the meaning ‘doomed to die young’ are Alexander Pope, Leaf, Munro, Mazon, Rieu, Macleod, and Willcock. Lattimore is the only translator I have found to notice that there is no mention of fate in the Greek. However, it is not clear what his rendering ‘but a single all-untimely child he had’ is supposed to mean since ‘untimely’ is not used of persons in normal English.

³ Eustathius 1130.24, 1277.42, 1412.42.

⁴ The best introduction to the theory of the oral composition of the Homeric poems is a book edited by Latacz, J., Homer: Tradition und Neuerung (Darmstadt, 1979)Google Scholar. It is an anthology of essays and articles and includes a full bibliography.

⁵ On the oral Homer see Latacz (previous note); on Homer and the alphabet see Goold, G. P., TAPhA 91 (1960), 272–91Google Scholar, and Wade-Gery, H. T., The Poet of the Iliad (Cambridge, 1952)CrossRefGoogle Scholar.

⁶ Ule, L., Concordance to the Works of Christopher Marlowe (Hildesheim/New York, 1979)Google Scholar and Spevack, Marvin, A Complete and Systematic Concordance to the Works of Shakespeare (Hildesheim, 1968–1970)Google Scholar.

My estimates of hapaxes and of total vocabularies for all the authors I mention are calculated from samples, and though I have taken care to make the samples as unbiased as possible the results are necessarily approximate. To make this clear and to prevent any illusion of pseudo-scientific exactness I have presented them in very round figures.

⁷ It is of course possible for words that are first attested in a particular author not to have been invented by him but to be anonymous creations of the time. Who first said super-giant or interface? Nobody knows. However, words of this fashionable kind are likely to recur in more than one author. Thus some commonplace words (e.g. ⋯ναɸορά and ⋯ναπληρ⋯) and some technical ones (e.g. ⋯μαξοπληθής, ⋯ναίρω, ⋯ναμοχλεύω) are first found in Euripides, where they are hapaxes, but the fact that all of them are freely used afterwards, the former in all kinds of writers and the latter in technical ones (Aeneas Tacticus and Galen), makes it unlikely that the creation of any of these words was due to Euripides himself.

⁸ The transition from a quantitative to an intensive force of the παν- in these compounds is noticed by Leumann, Manu, Homerische Wörter (Basel, 1950), 105Google Scholar.

⁹ It is generally considered by etymologists that ἄωροι, used to describe Scylla's feet in Od. 12.89, is derived not from ⋯ρη ‘season', but from a different root altogether.

¹⁰ P. Mag. Par. 1.342 and 1.2867. In Apollodorus Comicus, fr. 4 (Kock iii. 289) it is likely that the notion of death was introduced in the previous dialogue and was not dependent on the bare word ἄωρος.

¹¹ See van Leeuwen's commentary at Il. 1.352. On our passage (24.540) van Leeuwen has a strange note — ‘Prorsus adhuc immaturum. Mente addendum: et mox moriturum’. This indicates that he is not comfortable with the normal interpretation, but his comment hardly helps. All babies are immature when they are born, which must be why he puts in the adhuc. But it is not credible that παναώριος could mean ‘who is still immature now that he is a young man’; nor, even if it could, would the remark be relevant in the context.

¹² Theophrastus devotes a whole character-sketch to the man who is ἄκαιρος. For ἄτοπος of persons see Plato, Symp. 215a, Rep. 493c; Isocrates 12.149; and (probably) Menander's Heniochos, where a character states that the main thing is to avoid the ἄτοπον. Unfortunately in this example we cannot be sure that the masculine τ⋯ν ἄτοπον really refers to a person and not to a thing.

¹³ Both can qualify persons by way of character description, as too can the positive ὡραῖος (e.g. Aratus 1075, the ‘punctual’ farmer). μ⋯ ⋯ρασι is used as if it were an adjective by Aristophanes in the Lysistrata (391), and again in the same play (1037) as an adverbial phrase. Whatever now-lost topical point may underlie this, it is clear from the locative form that the phrase itself must have been an old one; and if the phrase was old, so too must be the attitude to άωρία implied by it.

¹⁴ Our knowledge of how the plot is going to develop may make us see Achilles' passionate complaint to his mother after the quarrel (352) that he is μινυνθάδιος (‘short-lived’) as a presentiment of early death, but in the immediate context it is clear that μινυνθάδιος is pointing the contrast between the brief life of Achilles as a human being and the immortality of his goddess mother, not between that of Achilles and other men. Of course when shortly afterwards Thetis is pleading with Zeus she can legitimately be made to call Achilles ὠκυμορώτατος ἄλλων (‘bound to die earlier than anybody else’) because this is a conversation between gods.

¹⁵ Milman Parry of course knew that there were unique lines in Homer, , and tried (HSCP 41 [1930], 133–4 = Latacz 228)Google Scholar to reconcile them with his general theory by claiming that if we had tens of thousands more Homeric verses we should find that the unique words and phrases of our present text were really parts of established formulae. The more modern method of explaining them, pioneered in The Singer of Tales (Cambridge, Mass., 1960), 95ffGoogle Scholar. by Parry's pupil, Albert Lord, but much developed by others, is to redefine formula to mean not a fixed set of words but a fixed metrical or syntactical pattern into which words new or old can be fitted as the poet wishes. The first solution is bizarre, tantamount to saying that though the theory is not supported by the existing evidence it would be supported by new evidence if only it were to come to hand. The other is self-cancelling — it makes the alleged process of oral composition indistinguishable from any other: for the way that we all, whether literate or illiterate, talk and write is precisely by fitting words to pre-existing patterns of syntax and rhythm.