The most serious competitor of Smyrna was Chios, the claim of which found favour with several of our earliest authorities. Especially important is the fact that in a fragment published among the remains of Simonides of Ceos (fragm. 85, Bergk), but now frequently attributed to Semonides of Samos, who lived about the middle of the seventh century, a verse of the Iliad (VI 146) is ascribed to the 'man of Chios.' In this island also there dwelt in later times a clan called Ὁμηρίδαι[333] who claimed descent from the poet. But here again there is a tradition that the population was at least in part Aeolic[334]. Moreover, though the language of the earliest extant inscriptions is Ionic in its main features, it possesses certain Aeolic characteristics, especially the change of ō̆n to ō̆i[335] before s (e.g. πρήξοισι, λάβωισι against Ion. πρήξουσι, λάβωσι). There is no record of an Ionic conquest of the island, as in the case of Smyrna, but we know that it did not enter the Ionic confederation until a comparatively late period, probably the seventh century[336].
It appears then that the peculiarities of Homeric language can be satisfactorily accounted for by the history of either Smyrna or Chios[337]. But now we are confronted with a very grave difficulty. The same scholars who hold that Smyrna was the birthplace of Homeric poetry yet insist that the language of the epics themselves was never anything else than Ionic, although they allow that the Iliad was nearly complete some considerable time before the conquest of Smyrna. This position is quite incomprehensible to me. The only explanation offered, so far as I am aware, is that the language of this district may have been of a mixed character[338]. It is no doubt true that when the coast was first occupied settlers may have come from many different quarters. But when Mimnermos speaks of the capture of 'Aeolic Smyrna' we are surely not justified in assuming that the city had become Ionicised before that time. On the other hand the supposition that the two dialects were not yet differentiated to any considerable extent appears to me to be irreconcilable with the evidence of the poems themselves as well as with all that we know of the history of the Greek dialects.
The question which we are discussing is one which concerns not only the Iliad and Odyssey, together perhaps with certain Hymns and other Homeric poems, but also the various works attributed to Hesiod. In particular we may note the Works and Days and the Theogony, both of which claim to be of Boeotian origin. Whether by chance or not they contain few forms[339] which are peculiar to Aeolic proper, i.e. the dialect of Asiatic Aeolis. But in all other respects their language is of the Homeric type, i.e. generally speaking Ionic, though with certain reservations, notably that ϝ is generally kept and ᾱ frequently occurs before ο, ω. Now it cannot be contended seriously that this extraordinary mixed dialect sprang up naturally on both sides of the Aegean. The only alternative however is to suppose that Boeotian poets borrowed it from Asia. But is this really probable at such a time?
Before leaving this subject we must notice briefly the alternative theory that the Homeric poems were originally composed in Aeolic. According to the form in which this theory has become most widely known they were translated into Ionic at a comparatively late date—towards the close of the sixth century[340]. It is rather a serious objection to this hypothesis that the poems contain no trace of late Aeolic characteristics, such as the change of n to i before s (e.g. τοίς, παῖσα). Further, if the poems had been known so long in Aeolic, though doubt might have prevailed as to Homer's birthplace, the fact that he was an Aeolian could never have been called in question. Above all it is difficult to see how the need of a translation could have arisen at such a date, for the Aeolic dialect was then well known throughout the Greek world through the poems of Alcaeus and Sappho.
On the other hand there can be no possible objection in principle to the idea that the poems have undergone a change of dialect. We have seen that a large proportion of Anglo-Saxon poetry has passed through a similar process, generally from one English dialect to another, but occasionally, as in the case of Genesis (vv. 235-851), from a continental to an English dialect. In some few cases we still have parallel texts preserved in different dialects[341]. Indeed, when the poetry of one community becomes current in another community, it would seem that under certain conditions such changes were not merely possible but even inevitable. This is a question to which we shall have to return in the course of the next few pages.
There is still one vexed question which we have not as yet touched upon, namely the relationship of the Homeric poems to the art of writing. Wolf and his immediate successors held that the art was unknown when these poems came into existence Among more recent scholars however the general tendency has been to regard this view as mistaken. Some leading authorities even hold that considerable portions of the poems were written down from the time of their composition.
The poems themselves contain only one reference to writing, namely in Il. VI 168 ff., where it is stated that Proitos sent Bellerophon to Lycia "and gave him baneful tokens, writing many deadly things in a folded tablet, which he bade him show to his (Proitos') father-in-law, with a view to his own destruction[342]." So long as no further evidence was forthcoming there was a natural inclination either to regard this passage as an interpolation or to interpret it as denoting something which could not properly be called writing. But of late years archaeological investigation has brought to light, especially in Crete, numerous inscriptions dating from very remote times, and there cannot now be any question as to the antiquity of writing in the southern Aegean. Moreover the Homeric poets themselves can hardly have been ignorant of the existence of such an art, for rock-hewn figures with inscriptions dating from pre-Homeric times—including one which has been identified with the figure of 'Niobe' mentioned in Il. XXIV 614 ff.—are to be found quite close to the Asiatic coast.
Yet in spite of all this it is a very remarkable fact that the whole 28,000 verses of the Iliad and Odyssey contain only one reference to writing. The significance of this may be appreciated by turning to modern Servian poetry, which abounds with allusions to letters and written orders, although the minstrels themselves are quite ignorant both of reading and writing. It is true that in Beowulf we find only one direct reference to writing. But the Iliad and Odyssey together contain nearly nine times as many verses as Beowulf.
On the whole it seems to me that the evidence for the antiquity of writing given above does not prove exactly what it is commonly supposed to do. The inscriptions on the rock-hewn figures are Hittite. The ancient Cretan inscriptions have not yet been deciphered, and in view of the later inscriptions found at Praisos the probability as yet is distinctly against their being in the Greek language. But in any case they date from ages long anterior to Homeric times, and there is nothing to prove their continuity with the writing of the historical period. With regard to the passage in Il. VI 168 ff. it is to be observed firstly that Proitos is one of the very earliest persons mentioned in the poems—some three generations removed from the characters of the Trojan story—and secondly that the curious phraseology seems rather to suggest that the poet was speaking of something which he did not clearly understand. On the whole then it is much to be doubted whether writing was a current and native practice during the period when the poems were composed.