The observation that all the seven examples of ὄρεσφι in the Iliad occur in similes prompted me to examine the similes of the poem for other linguistic peculiarities, which I soon found to be abundant. The obvious question of interest then was whether these peculiarities belonged to the one rather than to the other of the two main types of unusual features in the dialect: archaisms and neologisms. It was immediately apparent that similes were full of rare late forms, whereas archaisms were very few and mostly uncertain. (It is of course immaterial that similes contain such archaisms as the genitive plural in -άων or the dative plural in -εσσι, which are part of the stock-in-trade of all epic poetry.)

A related question is whether an undue proportion of late forms occurs in similes. For the consideration of this question I made, for Studies in the Language of Homer (1953), a list of such forms from the first volume of Chantraine's Grammaire homérique (Phonétique et Morphologie), and examined their localization. I found it to be a fact that many, especially of the most striking of these forms, did occur in similes. At the same time I noticed that of the remainder a large number occurred in other types of passages which in one way or another fell out of the main narrative.

Chapter I (Chantraine) — Structure of the sentence

Λέκτρονδε is the only example quoted by Chantraine of such an expression without a verb, helped out perhaps by the preceding δεῦρο, which it specifies. Such ellipse of a verb of motion is common in spoken language presumably at all periods and the old-established δεῦρο, felt as a verb, cf. δεῦτε, is common in Homer with or without a(nother) verb, but λέκτρονδε seems stylistically un-epic: θ 292 (Ares and Aphrodite).

Chapter II—Apposition and agreement

The possessive adjective is uncommon except from proper names, and in Aeolic it seems to have been restricted to patronymics. This use of adjectives from common personal nouns is rare in Homer, who has not, for example, the πάτριος ‘of one's father(s)’ used by Pi. and in tragedy. The only examples known to me in Homer are in Od.: λ 521 (nekyia), о 247 (story of Amphiaraus) γυναίων εΐνεκα δώρων, λ 437 (nekyia) γυναικείας δια βούλας, ο 397 ἅμ' ὕεσσιν άνακτορίησιν.

Γυναιος, for which see Schwyzer I 1583, is found out of Homer (and Moschus) only in the Attic dimin. γύναιον. Γυναικείος occurs in Hes. Op. 753 (‘Days’) and in Attic-Ionic (i.e. Ionic γυναικήϊος). For the form see p. 4.

Άνακτόριος is from άνάκτωρ, which is not epic, but known or deduced for Ionic and Doric, from one of these in tragedy. There is thus strong evidence for the lateness of the group in Homer.

(i)Faults of transcription from the early alphabet

‘Les exemples les plus clairs sont ceux qui supposent la confusion entre ε, о d'une part, et les fausses diphtongues ει, ου d'autre part.' The examples listed seem, however, to be genetically of different types.

(a) It is a mere possibility that ὁμοστιχάει is due to faulty transmission for ὁμοῦ στιχάει. The abnormal compound may be original. The verb is already abnormal as only here active in form (but so late epic) and only here without διέκτασις: contrast εστιχόωντο (p. 37): О 635 simile.

(b) Κρείσσων, μείζων, ἕννυσθαι are ‘Atticisms of the tradition’, and usual in Homer. It is curious that the exceptional Ψ 135 καταείνυον or καταείνυσαν is in a ‘late’ book, as if in some way it escaped the Atticizing process because of its lateness.

(c) Καιροσέων, from a rare technical term καῖρός which is not otherwise represented before late Greek, mostly in the lexicographers, was apparently not understood by the scribes and left much as it was. The two abnormal contractions in καιρουσσέων mark the form as late: η 107 (gardens of Alcinous).

(d) η 163 εΐσον for εσσον has a parallel in Hdt. ὑπείσας for ὑπέσας, and is therefore a matter for general textual criticism rather than an example of μεταχαρακτηρισμός.

(e) Άνηρείψαντο for ἀνηρέψαντο from the obsolete ερέπτω, which has close cognates in other languages, in all probability is due to confusion with έρείπω, despite the difference in meaning.

The following studies are closely connected in aim and method. Their original purpose was to examine in as much detail as possible the development of the language of the Iliad in some of its typical features, with careful attention to the spoken dialects involved and to the influence of metre. The views put forward, especially in chapters II and III, of some aspects of the development of the poem itself have arisen directly from the linguistic study. The Odyssey is included in chapter I and some parts of chapter IV. Otherwise it is referred to only incidentally.

It is a pleasure to acknowledge my indebtedness to three scholars. Professor Chantraine read an early draft of a considerable portion of these studies during my stay in Paris some three years ago, and made many valuable criticisms and suggestions. I wish I could express adequately to him how much his kindness stimulated and encouraged an obscure colleague from the other side of the globe. Professor Trendall has followed my work throughout its long years of preparation and has always found time, he himself best knows how, to read it at various stages and help me with advice in questions of scholarship and practical matters. I have also profited greatly by criticisms of many details made by Professor D. S. Robertson. It is hardly necessary for me to add that I am alone responsible for the views put forward.

This edition of my Studies has been greatly enlarged by a much more ambitious attempt to apply the results of linguistic examination to most parts of the two poems (Part B). It has been my aim to set forth for the consideration of scholars everything known to me that seemed to have any relevance to the kind of question discussed in the book. For this purpose I have drawn not only on the material provided by the systematic investigations of Part A, but also on Ebeling, Leaf and Ameis-Hentze, and on the books by Risch, Leumann and Wackernagel mentioned in the Introduction. A detailed examination of the work of the latter group was planned, but abandoned as adding too much to the bulk of an already swollen book, and replaced by mere references to their books or else to Frisk or Chantraine.

The special position of the language of similes, and also of digressions, descriptions and comments, a theme of my older book, seems to be confirmed. For the recurrence of various types of comments, to me an interesting aspect of Part B, the reader is asked to consider the relevant part of the Index an integral portion of the book. That speeches tend to be later in language and to have more other abnormalities than the narrative has been noted by readers of the first edition and has now been stressed in the analysis of several books. The subject would no doubt repay a systematic treatment.

In Part B the distribution of abnormal features is plotted, especially for the narrative. As specimens, similes are treated separately for a few books and parts of books. The material in the first book of the Iliad is set out somewhat differently from that of the other books of the poem, and the plan for the eleventh book of the Odyssey also diverges from the usual.

Κ, Ψ and Ω have been omitted from the discussion, on the assumption that their lateness is commonly accepted, and some books of the Odyssey have been treated lightly, to save labour and space.

In the ‘Parallels’ attached to the books of the Iliad I have examined briefly linguistic evidence which seems to me to point to secondary composition.

I should make the general remark that I admit even for many examples for which I have not specifically stated it the possibility of a common source, i.e. that the line showing a secondary feature may be modelled not on one or two quotable parallels but on traditional lines. However, it seems certain to me that the borrowing is often from the actual lines known to us from the Iliad or the Odyssey (or very occasionally from Hesiod, the hymns or the cyclic poems, the lines being then very late in the two great epics).