¹ See Language 3. 216. Delaporte, Éléments de la Grammaire Hittite, treats my second class on page 73 and my third class on page 75. In this paper I assume the identity of the hi-conjugation with the Indo-European perfect system. See Kellogg, Some Indo-European Coincidences in Hittite 38 f.; Kurylowicz, Symbolae Grammaticae in Honorem Ioannis Rozwadowski 103 f.; Sturtevant, Lang. 2. 33 f., 3. 161–8, 215–25, 4. 165, 6. 32–5, 154.

² For the meaning and implication of this term, see Transactions of the American Philological Association 60. 25–37. For the development of the original diphthongs in Hitt., see Lang. 6. 25–35. As will be shown, ∗∗dhēhi became ∗∗dhehi before the close of the IH period.

³ Cf. Lang. 6. 34 f. I still hold to the etymology given there, and to most of the argument, but a number of details require correction to accord with the conclusions of this paper.

⁴ See Lang. 4. 159–70, 6. 149–58.

⁵ Indogermanische Grammatik 4. 265 ff.

⁶ See Lang. 4. 165 ff., TAPA 60. 28 ff.

⁷ Madduwattas 137. The word occurs in Keilschrift-Urkunden aus Boghazköi 21.19.3.6.

⁸ See Lang. 5. 9, 144 f.

⁹ I do not mean to say that this particular stem was necessarily inherited in the two ablaut forms. Ablaut of the type we : u was common in the Hitt. verb, and it may have spread beyond its original province.

¹⁰ So la-hu-ah-hi (KUB 7.1. 1. 30).

¹¹ Many of the forms adduced in favor of the hypothesis of contraction are proper names; but these are a notoriously unsafe foundation for phonology.

¹² See Lang. 4. 162. Cf. Walde-Pokorny, Vergleichendes Wörterbuch der Indogermanischen Sprachen 2. 441.

¹³ All of these reconstructions are tentative, particularly that of the plural. At any rate this is better than to assume contraction of Hitt. pahhuwar to pahhur, as I did in Lang. 4. 162. Cf. Walde-Pokorny, VWIS 2. 14 f.

¹⁴ See Lang. 4. 162 f.

¹⁵ I have proposed this in an article soon to appear in Revue Hittite et Asianique.

¹⁶ Symb. Gramm. 95–104. Cf. Lang. 6. 149–58.

¹⁷ See Kurylowicz, op. cit. 102; Sturtevant, Lang. 4.162.

¹⁸ See American Journal of Philology 50. 360–9; TAPA 60. 33; Lang. 6. 153.

¹⁹ Friedrich, Zeitschrift für Assyriologie NF 1. 16 f., showed that this suffix forms from adjectives verbs meaning ‘bring about the state denoted by the primitive’. Verbs showing this peculiar conjugation, most of them containing the suffix ah, are 3-yah- ‘triple’, 3-yah- ‘quadruple’, ÌR-nah- 'enslave, allapah- of unknown meaning, alwanzah- 'conjure', arawwah- ‘set free’, armah- ‘impregnate’, inarah- ‘make manly’, ishiullah- ‘make obligatory’, ishunah- ‘girdle’, isiyah-‘spy out, disclose’, iskunah- ‘place, appoint’, idalawah- ‘harm’, kappilah- ‘make angry’, katterah- ‘make inferior’, kururiyah- ‘make hostile’, lazziyah- ‘make favorable’, liliwah- ‘hasten’, luriyah- ‘make unlucky’, maniyah- ‘assign, govern’, marsah- ‘make bad’, nah- ‘fear’, nakkiyah- ‘make heavy’, newah- ‘renew’, paprah-‘make guilty’, sarazziyah- ‘make superior’, suppiyah- ‘make clean’, dannattah-‘make empty’, tan pedassah- ‘make of second rank’, dasuwah- ‘blind’, duddumiyah-‘deafen’, walkessarah-: ulkessarah- ‘make expert’.

²⁰ Götze, Madd., 81–100, has collected and analyzed the evidence on the Hitt. suffix, and I have shown, Lang. 5. 8–14, its identity with IE āį. Owing to the apparent absence of original ā-stem nouns in Hitt., I concluded that the suffix āį was not a conglomerate of stem-final ā and denominative suffix io, but a simple suffix, and that in IE times it gave rise to ā-stem nouns by retrograde derivation. Petersen, AJP 51. 252 and fn., has convinced me that I was mistaken on this latter point; I now believe that we must assume ā-stem nouns in IH, which gave rise to the conglomerate āį in the way hitherto generally assumed, and also to the suffix ah which is treated in this paper.

²¹ For the verbs in a(i), see Götze, Madd. 91–5.

²² Other examples in Götze, Madd. 97–100.

²³ It is necessary to assume the ending ai; for IH e i would yield Hitt. ai (earlier ēi). Apparently dative and locative were not differentiated in IH, and so one cannot say positively that the Skt. -aye must come from IE -eai; possibly IE specialized ei as the dative ending, since we have unequivocal evidence for ai only in infinitives.

²⁴ My reasons for assuming the ending mi in a thematic verb are (1) the agreement of Skt. and Hitt. in this matter, and (2) a conviction that the IE ending ō came from IH oho, a medio-passive ending. Cf. Lang. 4. 169.

²⁵ Since IE had verbs with iō in the first sing, and ī in the third sing, (see Brugmann, Grundriss 2². 3. 178–93 and references), the variation may have been inherited by Hitt.; but phonetic law would bring the io: ie verbs to this result, and it seems better not to go outside the denominative verbs for an explanation.

²⁶ See Brugmann, Grundriss 2². 3. 223.

²⁷ Add to the list: su-up-pí-ih-hi (KUB 20. 99.2.7): suppiyahhi, suppiyahmi.