¹ In literature only the prose writers and, with due reservations, comedy are considered as giving evidence for living speech. Most forms of late prose are of little use because frequently archaistic, though it is admitted that still living words are sometimes found for example in the New Testament or the Septuagint which are fortuitously absent from our older prose records.

In addition to LSJ I have consulted lexica or indexes to Herodotus, Thucydides (von Essen), Aristophanes, the orators, Menander, Plato (Ast), Xenophon (Sturz) except for rare words like the many cpds. of θλάω

I am grateful to Professor A. P. Treweek for criticisms and suggestions.

² References will be to Schwyzer, E.Griechische Grammatik 1 (München, 1939);Google ScholarLejeune, M.Traité de phonétique grecque (Paris, 1947);Google ScholarBrandenstein, W.Griechische Sprachwissenschaft 1 (Berlin, 1954);Google ScholarFrisk, H.Griechisches etymologisches Wörterbuch (Heidelberg, 1954);Google ScholarChantraine, P.Grammaire homérique 1 (Paris, 1942);Google ScholarWalde-Hofmann, Lateinisches etymologisches Wörterbuch (Heidelberg, 1938–52);Google ScholarSommer, F.Handbuch der lateinischen Laut- und Formenlehre (Heidelberg, 1914.Google Scholar For Greek detailed references will be given only to Schwyzer except for special reasons.

³ Lejeune, p. 65, n. 3 also speaks of differentiation.

⁴ It is an interesting coincidence that is used by Demosthenes only in xviii 205, in praising the Athenians under Themistocles, cf. Isoc. (Also in [D.] lx 19.)

⁵ Cf. with the other form Herod, ii 83 There seems little evidence for LSJ’s translation ‘masturbate’. Aristophanes’ Frg. 391, keeps the other form in a technical term.

⁶ Forms of modern Greek have etc., with the two possible ways of avoiding dental ₊ I by dissimilation.

⁷ LSJ and Sandys on D., loc. cit., are to this extent wrong in saying that the word is found only in in prose. See Sandys and also Büchsenschütz on X. HG vi 5.1 and the index to SIG for collections of examples of the phrase, which was obviously in international use, e.g. GDI 3749.55 Rhodes and Hierapytna, 5024.9, Cretan towns.

⁸ Cf. Schwyzer, 330 f., where the comparison with Romance is in part anticipated. There is much to be said for Brandenstein’s refusal to admit the actual existence of the dz which is usually postulated (p. 94). Compare what is said of μν representing μδ p.8.

⁹ This also with mistaken sense, ‘fall in battle’, instead of imitating the sound of falling; Leumann, Homerische Wörter, p. 217.

¹⁰ Suggested by Professor Treweek.

¹¹ A Companion to Homer, p. 442.

¹² Shipp, G.P.Essays in Mycenaean and Homeric Greek, pp.14 f.Google Scholar

¹³ A simile in P 757 (), a digression on Tydeus, E 801 and a description of a Cretan harbour in γ 296. For the pair, and , see A. C. Moorhouse, CQ xli 31 ff.: ‘Briefly, the difference is that generally has affective or emotional connotations, and that nearly always has not.’ M. finds the affective value in E 801 and P 757, leaving γ 296 an open question.

Homer also uses with no obvious difference of meaning from .

¹⁴ Brandenstein, p. 94, explains the double treatment of sm- as due to a mixture of languages (Sprachmischung), the sm- in Greek being due to the influence of some other Indo-European language. This has some support in the fact that several of the words in sm are borrowed, e.g. . The difficulty is that we cannot know what IE language could be involved. Brandenstein’s argument that a phonetic change would not at the same time leave the old form in existence is too absolute, as is shown by early itself, and compare the two pronunciations of neither.

¹⁵ The principle will be the same if we suppose the t to have become a fricative, as in Eng. thin, before the assimilation took place (Brandenstein, p. 95). In fact the clash would then be more severe, between two fricatives as well as two dental sounds.