sobriety, however, belongs not to the whole Iliad, but to the events and characters of the war. Such figures as Bellerophon, Nïobe, the Amazons, which are thought of as traditions from an earlier generation, show the marvellous element at work.
2. Certain persons and events in the story have a distinctly mythical stamp. Helen is a figure of this kind. There was another story according to which she was carried off by Theseus, and recovered by her brothers the Dioscuri. There are even traces of a third version, in which the Messenian twins, Idas and Lynceus, appear.
3. The analogy of the French epic, the Chanson de Roland, favours the belief that there was some nucleus of fact. The defeat of Roncevaux was really suffered by a part of Charlemagne’s army. But the Saracen army is purely mythical, the true enemy having been the Gascons. If similarly we leave, as historical, the plain of Troy, and the name Agamemnon, we shall perhaps not be far wrong.
(b) The dialect of Homer is an early or “primitive” form of the language which we know as that of Attica in the classical age of Greek literature. The proof of this proposition is to be obtained chiefly by comparing the grammatical formation and the syntax of Homer with those of Attic. The comparison of the vocabulary is in the nature of things less conclusive on the question of date. It would be impossible to give the evidence in full without writing a Homeric grammar, but a few specimens may be of interest.
1. The first aorist in Greek being a “weak” tense, i.e. formed by a suffix (-σᾰ), whereas the second aorist is a “strong” tense, distinguished by the form of the root-syllable, we expect to find a constant tendency to diminish the number of second aorists in use. No new second aorists, we may be sure, were formed any more than new “strong” tenses, such as came or sang, can be formed in English. Now in Homer there are upwards of 80 second aorists (not reckoning aorists of “Verbs in μι,” such as ἕστην, ἔβην), whereas in all Attic prose not more than 30 are found. In this point therefore the Homeric language is manifestly older. In Attic poets, it is true, the number of such aorists is much larger than in prose. But here again we find that they bear witness to Homer. Of the poetical aorists in Attic the larger part are also Homeric. Others are not really Attic at all, but borrowed from earlier Aeolic and Doric poetry. It is plain, in short, that the later poetical vocabulary was separated from that of prose mainly by the forms which the influence of Homer had saved from being forgotten.
2. While the whole class of “strong” aorists diminished, certain smaller groups in the class disappeared altogether. Thus we find in Homer, but not in the later language:—
(a) The second aorist middle without the “thematic” ε or ο: as ἕβλη-το, was struck; ἔφθι-το, perished; ᾶλ-το, leaped.
(b) The aorist formed by reduplication: as δέδαεν, taught; λελαβέσθαι, to seize. These constitute a distinct formation, generally with a “causative” meaning; the solitary Attic specimen is ἤγαγον.
3. It had long been known that the subjunctive in Homer often takes a short vowel (e.g. in the plural, -ομεν, -ετε instead of -ωμεν, -ητε, and in the Mid. -ομαι, &c. instead of -ωμαι, &c.). This was generally said to be done by “poetic licence,” or metri gratia. In fact, however, the Homeric subjunctive is almost quite “regular,” though the rule which it obeys is a different one from the Attic. It may be summed up by saying that the subjunctive takes ω or η when the indicative has ο or ε, and not otherwise. Thus Homer has ἴ-μεν, we go, ἴ-ο-μεν, let us go. The later ἴ-ω-μεν was at first a solecism, an attempt to conjugate a “verb in μι” like the “verbs in ω.” It will be evident that under this rule the perfect and first aorist subjunctive should always take a short vowel; and this accordingly is the case, with very few exceptions.
4. The article (ὁ, ἡ, τό) in Homer is chiefly used as an independent pronoun (he, she, it), a use which in Attic appears only in a few combinations (such as ὁ μὲν ... ὁ δέ, the one . . . the other). This difference is parallel to the relation between the Latin ille and the article of the Romance languages.
5. The prepositions offer several points of comparison. What the grammarians called “tmesis,” the separation of the preposition from the verb with which it is compounded, is peculiar to Homer. The true account of the matter is that in Homer the place of the preposition is not rigidly fixed, as it was afterwards. Again, “with” is in Homer σύν (with the dative), in Attic prose μετά with the genitive. Here Attic poetry is intermediate; the use of σύν is retained as a piece of poetical tradition.
6. In addition to the particle ἄν, Homer has another, κεν, hardly distinguishable in meaning. The Homeric uses of ἄν and κεν are different in several respects from the Attic, the general result being that the Homeric syntax is more elastic. And yet it is perfectly definite and precise. Homer uses no constructions loosely or without corresponding differences of meaning. His rules are equally strict with those of the later language, but they are not the same rules. And they differ chiefly in this, that the less common combinations of the earlier period were disused altogether in the later.
7. In the vocabulary the most striking difference is that many words appear from the metre to have contained a sound which they afterwards lost, viz. that which is written in some Greek alphabets by the “digamma” ϝ Thus the words ἄναξ, ἄστυ, ἔργον, ἔπος, and many others must have been written at one time ϝάναξ, ϝάστυ, ϝέργον, ϝέπος. This letter, however, died out earlier in Ionic than in most dialects, and there is no proof that the Homeric poems were ever written with it.
These are not, speaking generally, the differences that are produced by the gradual divergence of dialects in a language. They are rather to be classed with those which we find between the earlier and the later stages of every language which has had a long history. The Homeric dialect has passed into New Ionic and Attic by gradual but ceaseless development of the same kind as that which brought about the change from Vedic to classical Sanskrit, or from old high German to the present dialects of Germany.
The points that have been mentioned, to which many others might be added, make it clear that the Homeric and Attic dialects are separated by differences which affect the whole structure of the language, and require a considerable time for their development. At the same time there is hardly one of these differences which cannot be accounted for by the natural growth of the language. It has been thought indeed that the Homeric dialect was a mixed one, mainly Ionic, but containing Aeolic and even Doric forms; this, however, is a mistaken view of the processes of language. There are doubtless many Homeric forms which were unknown to the later Ionic and Attic, and which are found in Aeolic or other dialects. In general, however, these are older forms, which must have existed in Ionic at one time, and may very well have belonged to the Ionic of Homer’s time. So too the digamma is called “Aeolic” by grammarians, and is found on Aeolic and Doric inscriptions. But the letter was one of the original alphabet, and was retained universally as a numeral. It can only have fallen into disuse by degrees, as the sound which it denoted ceased to be pronounced. The fact that there are so many traces of it in Homer is a strong proof of the antiquity of the poems, but no proof of admixture with Aeolic.
There is one sense, however, in which an admixture of dialects may be recognized. It is clear that the variety of forms in Homer is too great for any actual spoken dialect. To take a single instance: it is impossible that the genitives in -οιο and in -ου should both have been in everyday use together. The form in -οιο must have been poetical or literary, like the old English forms that survive in the language of the Bible. The origin of such double forms is not far to seek. The effect of dialect on style was always recognized in Greece, and the dialect which had once been adopted by a particular kind of poetry was ever afterwards adhered to. The Epic of Homer was doubtless formed originally from a spoken variety of Greek, but became literary and conventional with time. It is Homer himself who tells us, in a striking passage (Il. iv. 437) that all the Greeks spoke the same language—that is to say, that they understood one another, in spite of the inevitable local differences. Experience shows how some one dialect in a country gains a literary supremacy to which the whole nation yields. So Tuscan became the type of Italian, and Anglian of English. But as soon as the dialect is adopted, it begins to diverge from the colloquial form. Just as modern poetical Italian uses many older grammatical forms peculiar to itself, so the language of poetry, even in Homeric times, had formed a deposit (so to speak) of archaic grammar. There were doubtless poets before Homer, as well as brave men before Agamemnon; and indeed the formation of a poetical dialect such as the Homeric must have been the work of several generations. The use of that dialect (instead of Aeolic) by the Boeotian poet Hesiod, in a kind of poetry which was not of the Homeric type, tends to the conclusion that the literary ascendancy of the epic dialect was anterior to the Iliad and Odyssey, and independent of the influence exercised by these poems.
What then was the original language of Homer? Where and when was it spoken? [The answer given to this question by Aug. Fick (in 1883) and still held, with modifications, by some European scholars can no longer be maintained. Fick’s original statement was that in or about the 6th century B.C.
