ajo, and it seems to follow that the corresponding IE a of the non-present stems is to be recognized in infinitives like irhawwar 'completion': irhaizzi 'he completes'. The suffix ah can be brought into the picture only on the assumption of contamination; but that is not improbable (cf. especially newahh- 'renew': Lat. novāre). On the denominative suffix a(i), see below p. 156. 'Les désinences verbales du présent singulier apparaissent en hittite sous une double forme: a) -mi, -ši, -zi (= *-mi, *-si, *-ti); 'En retranchant dans la seconde série les i propres au présent, nous obtenons: -h, *-th, -zéro. Le droit de restituer un th se fonde 1° sur la circonstance qu'en indoeuropéen il y a bien des désinences de 2-ème personne singular à th-. . . mais qu'il n'y en a pas une à t-. 2° sur la circonstance qu'un t primitif serait passé à z devant un i suivant. . . 'Si, dans le singulier du parfait indoeuropéen comportant les désinences -a, -tha, -e, on retranche les -e finaux, tout en redonnant aux a leur valeur propre, c'est-à-dire 22e, on obtient -22, -th (c'est-à-dire *ta2...), -zéro. La seconde série de désinences hittites est donc au fond la même que celle du parfait indoeuropéen; il y a là un parallélisme qu'on ne saurait méconnaître. Seulement en hittite il y eut élargissement par l'-i du present, tandis que le parfait indoeuropéen adopte un -e. On sait que le parfait indoeuropéen est au fond un présent à désinences particulières. 'Du même coup, deux questions difficiles de phonétique indienne reçoivent une lumière inattendue: 1° la raison d'être de l'aspirée -th-. 2° l'opposition entre la 1-ère pers. (cakara) et la 3-ème pers. (cakā1a): c'est l'opposition entre la syllabe fermée (*quequora-e) et la syllabe ouverte (*quequor-e).' I agree with Kurylowicz in identifying the endings of the Hittite hiconjugation with those of the IE perfect,1o but I can find no justification for his amputation of the final vowels of Hittite and IE before he makes the comparison. It is much more likely that the Hittite third personal ending i is cognate with the IE e, and that Hittite ahi (e.g. šaggahhi 'I know': šakki 'he knows') is cognate with IE ai (Skt. tutude, Lat. vidi, Ch. Sl. vêdê). Kurylowicz makes a curious slip in his treatment of the Hittite second 19 See LANG. 2. 33f. (1926), 3. 161-8, 4. 165. person. He is right in identifying the consonant of Hittite ti with that of Skt. tha and Gk. a. The IH ending was tha, and this must have yielded Hittite ta; then the analogy of the first person (possibly strengthened by the corresponding ši of the mi-conjugation) changed ta to ti. But it is impossible to ascribe the aspiration to the former presence of 22 (th<tǝ2), if we hold that 22 survives in Hittite. Although he does not suggest anything of the sort, it may be that Kurylowicz assumes loss of such an h after t; but if so he must explain the presence of h after t in tethai, tethiškizzi 'it thunders', and tetheššar, tethimaš 'thunder',20 and in Gnathita (Keilschrift Urkunden aus Boghazkoi 10.92.6.15). The same combination appears in proper names, e.g. Tuthaliyas, the name of several Hittite kings. After all our criticism there remains an impressive body of material which fits Kurylowicz's theory, and it would be easy to make considerable additions to it; but all such material is equally well accounted for by the alternative theory, according to which IH had a consonantal h that could not be vocalized, a vowel a, and long vowels ē, ō, and ā, whose nil-grade was ǝ. Consequently a moderate number of highly probable etymologies which are inconsistant with the assumption that 22 Hittite h or with its corollaries are enough to disprove it. = We have already noted that the connection of Hittite šanh- 'petere' with Gk. ȧvów, avapai 'accomplish' involves the assumption of a both in Hittite and in IE although the absence of h before the vowel in Hittite should, according to the theory, indicate that there was no 22 to account for a change of e to a. It cannot be urged that the 22 of the following syllable led to a dissimilative loss; for Hittite contains many such words as arhahari, ešhahru, harharaš, huhhaš, lahlahiman, tuhuhša. = Other certain etymologies which involve IH a, although Hittite shows no h are; appa(n) 'afterwards, then, behind, back' = Gk. áπó Lat. ab, από etc.; aruwa (i)- 'worship' Gk. ȧpáoμaι;21 waki 'he removes': Lith. vagiu 'steal', waggari 'is lacking': Lat. vagor 'wander';22 karp- 'get, take (up or away), muster (troops)':23 Lat. carpo 'pluck, seize', Gk. κаρжós 20 See Friedrich, Zeitschrift für Assyriologie NF 3. 196 f.; Götze, KƖF 1. 186. 21 See LANG. 5. 10. 22 See LANG. 5. 228. Since the above was written Friedrich, Staatsverträge 2. 1711, has shown that waggari is a hi-conjugation verb with stem waggar-. It probably has no connection either with Hitt. waki or with Lat. vagor. 23 See Friedrich, ZA NF 1. 185f.; Sommer and Ehelolf, Boghazköi-Studien 10. 73f.; Götze, Hattušiliš 98ff., Madd. 45; Forrer, Forschungen 1. 64. These scholars (with the exception of Forrer, who does not discuss the question) regard 'take up' as the primary meaning of the word; but the numerous passages cited by them indi 'harvest'; wa-al-lu-uš-ki-ši (i.e. walškiši?) 'you are strong',24 wal-kiššaraš 'strong-handed':25 Lat. valeo, Goth. waldan, OHG walten 'be strong'. With a little more hesitation I would connect ališ 'white',26 with Gk. áλpós, Lat. albus 'white', and Lith. alvas 'tin', etc.27 = As noted above the IE denominative suffix aio/e appears in Hittite as a(i) (e.g. tarmaizzi Lat. terminat, tar manzi = terminant). Here, then, we must recognize an IE long a in a suffix which in Hittite has no h. If it be urged that the length and character of the vowel in IE are both due to contamination with the denominative suffix which appears in Hittite as ah (see above, p. 153-4), it will still be necessary to account for the Hittite a(i)-verbs; for IH e remains in Hittite. It is equally necessary to assume IH a for Hitt. tayezzi 'he steals':28 Skt. (s)tāyus 'thief', Av. tāyu- 'thief', tāya- 'theft', Ch. Sl. tają ‘hide', Gk. Tηráw (Pind. тāτwμevos) 'deprive of'. Kurylowicz (952) cites Gk. iorāμɩ as a typical example of a from ea2). The root appears in at least three Hittite verbs, none of which contain h. For ištanta- 'tarry': Goth. standan ‘stand' and ištap- 'enclose, keep': Skt. sthāpayati 'causes to stand', see LANG. 4.3f. Of particular interest is Hitt. tittanu- 'cause to stand, place',29 which furnishes us the explanation of the curious reduplication of Skt. tişṭhāmi. The Hittite causative lacks the predeterminative s which appears in the two Hittite words just cited and regularly in IE (cf. Gk. Téyos σréyos, Skt. tāyus = stāyus, etc.). The Skt. tiṣṭhāmi, then, is a contamination of *titāmi and *sisthāmi, while Av. histaiti, Gk. iornμi, and Lat. sisto represent the normal reduplication of the root *sta. The a of tittanu-, like that of ištanta-, represents IH ǝ (note the variant titnu-); the only one of the three Hittite words that contains original ā is ištap-. = Other certain etymologies which involve IH a without a trace of a cate rather a primary meaning 'get'. Delaporte, Grammaire de la Langue Hittite 101, suggests the meaning 'play' (a musical instrument) for the occurrence in Keilschrifttexte aus Boghazköi 4. 9. 5. 30, 36, a meaning which would strengthen the above etymology, in view of OE hearpe, OIsl. harpa, etc., 'a pluck-instrument, a harp'. But Götze's (Hatt. 99) interpretation of the passage may after all be correct. 24 Forrer, Forschungen 2. 30. 25 Cf. Hrozný, Sprache der Hethiter 40; Holma, Journal de la Société FinnoOugrienne 33. 31f. 26 See Götze, Madd. 142. 27 See Walde-Pokorny, VWIS 1. 92 ff., and references. 28 Cf. LANG. 5. 145. 29 See Hrozný, SH 76,; Sommer and Ehelolf, BoSt. 10. 51; Götze, Madd. 136. contiguous h in Hittite are these: Hitt. hapatiš has IH long a in its second syllable, as appears from Gk. óŋôós and óráwv (see LANG. 4. 164). Hitt. man ‘äv' and man 'if' must be the same word as Gk. μáv. Hitt. mamahhi 'I say', memai 'he says' is identical with Gk. μéμvnμaι except that it is active (the etymologies which I suggested in LANG. 3.217 are to be rejected). It is one of the few hi-conjugation verbs to preserve the reduplication. The development of mn to Hittite m is seen also in larmai Lat. terminat (see above p. 156), ištamas- 'hear': ištamanaš 'ear', arummas 'of washing' (<*arumnaš): arrummar 'to wash', and in many other words.30 = Hitt. papra- 'do wrong, sin', papratar 'uncleanness, sin', papres- 'be guilty' are to be connected with Lat. prāvus 'bad'; but it might possibly be maintained that the vocalism is due to the form with suffixal h (in that case from 22) which appears in paprahhiškir 'they made impure' (KUB 7.53.2.16). Hitt. wašta- 'do injury, sin', waštul 'sin, injury' may plausibly be connected with vāstus, OE wēste 'desert', which WaldePokorny, VWIS, 1.219, compare with the root ya- ‘strike, wound'. Kurylowicz (97) makes much of that part of his argument by which he accounts for the initial vowels of Lat. aurora beside Skt. a-vasran 'they shone', Lat. augeo beside Skt. vavákşa 'he increased', and Gk. avôń 'speech' beside Skt. vadati 'he speaks'. While other scholars, following Hirt, assume for these pairs the ablaut bases **ayes-, **ayeg-, **ayed-, Kurylowicz accounts for the Latin or Greek a by assuming triconsonantal roots, 22-u-s, 22-y-g, and 22-u-d, so that *aus is from *22eus, *aug from *22eug, and *aud from *22eud. It happens that two of these three roots occur in Hittite, and neither of them has an initial h. I have recently treated the connection of Hittite au (s)- 'see' with Lat. aurora, etc. (LANG. 6.34-5), and I will not repeat. Hitt. uttar, gen. uttanaš 'word', certainly belongs to the root **ued 'speak';32 probably it contains the nil-grade seen in Gk. vôéw, vow 'celebrate', Skt. uditas 'spoken' etc., rather than the first full grade which Kurylowicz finds in Gk. avoń. The derivative verb watarnah- 'command' is certainly to be connected with uttar; it seems to imply a noun *watar with either o-grade or reduced grade of the root (**uótr or **u.tór). There can be no doubt, then, that the Hittite part of the theory proposed by Kurylowicz is impossible. If it is true that the long vowels 30 See Ehelolf, ZA NF 2. 313; Götze, Madd. 131. 31 See Götze, Hatt. 95f. 32 See Walde-Pokorny, VWIS 1. 251f. 33 See Götze, Hatt. 96. represent earlier diphthongs, ea1, e22, and eas, and if 22 is responsible for the vowels a and ā and for the voiceless aspirates, th, etc., all of the changes involved must have occurred in IH; for Hittite presupposes the same situation that appears in the IE languages. Hittite h cannot be equated with this 22; it represents an IH consonant which was incapable of vocalization, and which has vanished without leaving a trace in the IE languages. While Kurylowicz provides several additional etymologies for Hittite words containing h, I do not see that his article necessitates any essential modification of the conclusions regarding Hittite h which I reached in LANG. 4.159-170. |