¹ Cf. J. La Roche, Hom. Unt. 1.128–45, 276–81 (Leipzig, 1869); A. Dyroff, Gesch. des pronomen reflexivum I (Beiträge z. hist. syntax d. gr. spr. 3.3, ed. M. Schanz; Würzburg, 1892); K. Brugmann, Ein problem d. hom. textkritik u. d. vgl. sprachwissenschaft (Leipzig, 1876). The last is discussed by Leaf, Appendix A, Iliad 1.559–65.

² When we speak of English or Latin reflexive pronouns we know what we mean to the extent that we can describe the uses of himself or of sē, etc. The extent to which this can be done for an Indo-European reflexive pronoun is much smaller.

³ Quite probably the line is not original. It recalls B [674], on which cf. External Evidence 75.

⁴ This is due partly to the competition offered by the middle voice (cf. Brugmann 104; Wackernagel, Syntax 2.90), partly to the psychological beliefs then current. In Homeric Greek one does not speak to himself, he speaks to his θυμός; one does not bid himself act, his θυμός bids him. Cf. Dyroff 40 n. 2, where other literature is cited.

⁵ Beitr. z. Lehre v. gr. Akzent 36–7. The examples that he gives are the following: λοέσσαί τε (not λοέσσαιτε) τ 320; γενέσθαί τε τραφέμεν τε (not τετραφέμεν) H 199; ‘Hφαίστοɩο πάρ‘ οἰσέμεν (not παροισέμεν) Σ 191, with Herodian's remark ὅτι δὲ καὶ ἐν ἅλλοɩς ἀναγνὀμασιν ὁ ἀνὴρ (Aristarchus) τοιου̑τος, προείπομεν ἤδη; and doubtingly no anastrophe of ἄνα (because = ἀράστηθι, or voc. of ἄναξ), nor δία (= Δία).

⁶ Griech. Gram.² 87 (1890), not repeated in later editions.

⁷ For economy I have modified his Greek typography.

⁸ The disputed examples in B 196, K 398 are interpolated.

⁹ Collected by Brugmann.

¹⁰ The construction is not reflexive, and that is true also of some other passages to be mentioned. But when the adjective = ἐμός or σός is precisely the time when its non-reflexive use is (cf. above) easiest.

¹¹ Ar. Ω του̑: τινές σφου̑. Wackernagel, SUH 150, prefers the latter.

¹² Ar. Ω ἄλλο τεὴ ἐπινύσσεν ἐφετμή; for the psilosis cf. Class. Phil. 41.232–3 (1946), and esp. the postscript to follow in 1947.

¹³ Ar. Ω ἀμφὶ σφοȋσ‘ ὀχέεσσ‘.

¹⁴ There is a gap in the A-scholia from T 126 to T 326, and this may well be the reason that the reading adopted is unattested.

¹⁵ For this problem it matters little whether the lines are interpolated or not. The author had the model needed; and whether he lived in the 5th century or a couple of centuries earlier is not very important.

¹⁶ A variant ἡσιν appears sporadically in seven out of 31 passages (the Odyssey included), but can be no more than a secondary blunder. Chantraine, Gram. Hom. 274, seems to overstate the evidence: ‘Partout où se trouve la formule φρεσὶ ση̑σιν, une variante φρετσὶν ἡσιν est attestée dans des manuscrits.‘ He cites only the seven known passages.

¹⁷ Cf. La Roche, Hom. Unt. 1.125–6.

¹⁸ The variation in o 89 (parallel to Z 221) is too slight to be regarded as significant.