¹ Actes du Première Congrès Internationale de Linguistes 155–64.

² See, in addition to Hrozný's references, Delaporte, Grammaire de la Langue Hittite 80 ff., and Sturtevant, Language 4. 166 ff., Transactions of the American Philological Association 60. 28–33.

³ Maddawattas 80–100, especially 93 f.; cf. Lang. 5. 8–14.

⁴ Die Sprache der Hithiter = Boghazköi-Studien 2. 174 f.

⁵ Cited from an unpublished text by Götze, Neue Bruchstücke zum Grossen Text des Hattusilis und den Paralleltexten 631.

⁶ It may be significant that four of these forms occur in a single document. Of course the non-thematic forms like mar-ri-it-ta and par-si-it-ta-ru do not belong here. On the other hand one must remember that the sign ya sometimes has the value ye (see Lang. 7. 11–2).

⁷ See TAPA 60. 28–33. Several of the details suggested there are inconsistent with suggestions in the present paper.

⁸ See American Journal of Philology 50. 360–9, TAPA 60. 33.

⁹ See Lang. 4. 168.

¹⁰ Cf. Hrozný, Congr. 1. 157.

¹¹ As we shall see presently other IE medio-passive endings are found in Hitt., but in Hitt. they are either plainly composite or not confined to a single person and number.

¹² This form and the equivalent lukkatta were interpreted as case-forms meaning ‘at dawn, tomorrow’ by Sommer in a masterly article in BoSt. 7. 22–32. There is no denying that the words may readily be so interpreted in many passages, as (Hittite Texts in the Cuneiform Character from the British Museum 1. 3. 21): lukkattimas É-ri penahhi ‘at dawn I take them to the temple’; (KBo. 4. 4. 3. 40): lukkattama INA ^URU Iyahrissa EGIR-pa uwanun 'the next day I returned to Iyahrissas.' Nevertheless the verbal character of lukkatta is clear in many sentences, and lukkatti must certainly be taken the same way; I refer to sentences which begin: mahhanma lukkatta nu . . . 'when day dawns, then . . . ', or the like. The best commentary on sentences of the first type is supplied by the asyndetic use of verbs of motion, which Friedrich has discussed in Vert. 1. 162 ff.; for example; paiweniwarankan kuennummeni 'so we'll go kill him'. At first such a sentence as lukkattimas É-ri penahhi meant 'day dawns, I bring them to the temple. At the time when our documents were written the two words may have been little more than adverbs, in spite of the frequent use of mahhan or man with lukkatta; they had been weakened in the same way as the verbs of motion.

^12a Kurylowicz calls my attention to Wackernagel's demonstration (KZ 41. 309–13) that Vedic aduha ‘he milks’ is a middle form. Undoubtedly its ending is identical with Hittite -a.

¹³ On the IE evidence see Brugmann, Grundriss der Vergleichenden Grammatik der indogermanischen Sprachen 2². 3. 571 ff.; Hirt, Indogermanische Grammatik 4. 112 f.

¹⁴ See Ungnad, Zeitschrift für Assyriologie NF 2.104₁; Friedrich, ZA NF 2. 289–96; Sturtevant, Journal of the American Oriental Society 47. 174–7.

¹⁵ The suggestion of the infinitival origin of the medio-passive ti, t, and ri is due to Hrozny, SH 75, Congr. 1.160. My treatment is more detailed than his, and departs from his views in some respects.

¹⁶ See Endzelin, Izvěstija Otdělenija Russkago Jazyka i Slovesnosti Imperatorskoj Akademii Naukъ 12.1.40 ff., Lietuvių Tauta 2.284 ff., Lettische Grammatik 709. f.

¹⁷ In TAPA 60.32 I undertook to apply to IH the traditional derivation of the medio-passive in r from the third pl. active forms in r. It now seems to me necessary to abandon the whole idea, plausible as it was before the discovery of Tocharian and Hittite. The active forms in r may also be of nominal origin; but in Hitt. they seem to be much more firmly embedded in the verbal system than is the medio-passive ri.

¹⁸ I am as far as ever from agreeing with Forrer (Zeitschrift der Deutschen Morgenländischen Gesellschaft 76. 212, Forschungen 1.130) in regarding these forms as indeclinable participles in Hitt.; but they probably were just that at a remote period in IH. In this connection one may note that in Hitt. the nt-participle, although formed on the present stem as in IE, has the functions which there belong to the stem in to. Perhaps the two formations were once more closely associated than they are in IE.