¹ On the appearance of r for expected dental stop in these and other HLuvian forms see Morpurgo-Davies, , KZ 96 (1982) 245 ffGoogle Scholar.

² For the difficult syntax of this form see Morpurgo-Davies, , KZ 94 (1980) 91Google Scholar, and compare the Appendix below.

³ The final -i of this word is a mere space-filler, as elsewhere in the Assur letters, and tells us nothing about the vocalism of the final syllable.

⁴ Since the “thorn” may indicate r alone, we should perhaps read ka-pá + ra/i-ra + à-na as /kapar(r)an/, permitting direct equation with ka-pá-ra/i-zi/kaparanzi/. If this interpretation is correct, it would be an additional argument for reading ra + à instead of à + ra/i.

⁵ The hostile sense of the HLuvian verb is undeniable. D. Hawkins (personal communication) has kindly provided me with an additional example of (*245)tanata- parallel to arha *248-nu- “destroy” (on the latter verb see below). In borowski 1, §§3–4 he reads: pa-sa-pa-wa/i-′ Locus(-)x-x-za / á-mi-i-s[a ^?] / x[…//…]…(*245^?)ta ^?-na-tà-tu ∣wa/i-tu-′∣^?(“malleus”)á-hu^?-li-na arha *248-nu-wa/i-tu. Despite the badly broken context, the gist is clear: “May my … [the god X^?] devastate his place and destroy his—.”

⁶ Oettinger, Stammbildung 240Google Scholar, claims that the hapax dannattauwanzi is an error, not evidence for a stem dannattā-(i), but this is unlikely in view of such pairs as kutruwaḫḫ/kutruwā (i)- “call to witness” and armaḫḫ-/armā(i)- “be pregnant”.

⁷ For the verb (longus)yariya- “stretch out, extend” (transitive and intransitive) compare sultanhan 6: tara/i-sa-zi-pa-wa/i / ya + ra/i-ti-i “It (a grapevine) will extend/put forth shoots”. HLuvian tarsa- is cognate with Hitt. ^gištarša “shoots” at KUB XII 58Google Scholar IV 18.1 cannot follow the reading and interpretation of carchemish A 11 b,2 offered by Hawkins, and Morpurgo-Davies, in Kaniššuwar 75Google Scholar.

⁸ D. Hawkins has furnished a further example of the i-stem adjective in a similar context (maraş 8, §3): ta ^?-pa-wa/i ∥ ku + ra/i-ku-ma-na(urbs) ∣flumen.regio;-ti-na (“*245”)ta-na-ti-i^?-na wa/i-mi-lituus-ha “And … I found the river valley (of) Gurgum desolate.” The reading urbs + mi-ni in the above quotation results from a recent collation of karahöyük by Mr. Hawkins.

⁹ Unfortunately, however, Meriggi lists together under tanata- occurrences of aedificium alone and of aedificium plus vertical stroke. Only the latter properly belong here.

¹⁰ Morpurgo-Davies, , Fs Szemerényi (1979) 590Google Scholar, also interprets aedificare-mi-i beside aedificare-mi-ri + i(-i) as standing for *tamari or *tamati.

¹¹ It is possible, however, that we do have a hi-verb meaning roughly “destroy” attested in topada 8: […]rel-i arha ha + ra/i-ri + i… wa/i-ta á-pa-sa-na vas-tara/i-i-na á-pa-sa–[ha] terra-rel‹-ra/i-›na (deus)tonitr[us]-h[u-sa (deus)sarma-ma-sa … arha ha + ra/i-tú-u “If/As [someone] smashes [this inscription/these words], … let the Storm-god and Sarruma … smash his form and his land.” We should probably restore the same verb in bolkarmaden 5: za-pa-wa/i-ta-′ / capere-ma-z[a re]l-sa arha-′/ [ha + ra/i]-ri + i “Whoever smashes this document(?) …”. Due to the lack of arha and particle -ta, it is hard to say whether we have the same verb later in the same line: (deus)luna-mi-pa-wa/i-na ha + ra/i-tu “Let the Moon-god smash(?) him.” HLuvian arha har(ra)- could equate directly to Hitt. arḫa ḫarra- “crush, smash”, which significantly is the term commonly applied to damaged cuneiform tablets: KUB XIII 7 IV 3 kī tuppu arḫa ḫarran ēš[ta] “This tablet was broken/smashed.” As noted by Oettinger, , Stammbildung 505 f.Google Scholar, the earliest attestations of Hitt. ḫarra- are middles with a fientive or passive sense. The later attested active ḫ-verb harrai could thus be old. One difficulty with the equation is that the only way to read HLuvian ha + ra/i-ri + i as a hi-verb is to read /har(r)i/. To my knowledge there are no other certain examples of a pres. 3rd sg. ending -i in Luvian (instead of usual -ai), and it would be an unexpected archaism.

¹² On the form of the Hittite word (reduplicated pa-pr-) and a possible etymology see Melchert, , Studies in Hittite Historical Phonology (SHHP) (1984) 33, n68Google Scholar.

¹³ The interpretation of (loqui)marati- as “injunction, order” is now strongly supported by the existence of Lyc. mar- “order, command”, mara “law” and related forms: see Laroche, , Fouilles de Xanthos 6 (1979) 73 fGoogle Scholar. However, the proposal that (loqui)ma _x + ra/i-ta is the collective plural of the animate stem (loqui)marati- is brought into question by the existence of “loqui”-tà-zà-′ (kayseri 6), brought to my attention by D. Hawkins. While this example is singular and follows the curse formulas, one's first inclination would certainly be to take it as the neuter singular */m/paratan-za/ of the attested plural (loqui)*462 + ra/i-ta.

¹⁴ The reading ma _x instead of pá for sign 462 might also give a clue to understanding the phrase (annus)u-si (annus)*462 + ra/i-i in karatepe XLVIII 266–7 (the reading is that of Ho). It is clear from following references to ”(the time of) plowing” and “(the time of) harvesting” (see Meriggi, , Manuale II/1.83Google Scholar) that we are dealing with regularly scheduled sacrifices to a god. The sense of (annus)usi “annual(ly)” is clear enough, but the function of the further specification (annus)*462 + ra/i-i is obscure. However, the sequence of time specifications here recalls that of Hittite in KUB XXIV 1 II 3 ff.Google Scholar: ezen.Ḫi.a mu meanaš gimmantaš ḫamišḫandaš zenantaš “festivals of the year —, of winter, of spring, of autumn” (likewise XXIV 3 I 16 f., with a different sequence of seasons). Whether or not MU me(a)ni/a- means “new year” (see Friedrich, , HW 139 w/refs.Google Scholar), it is suggestive that in both Hittite and HLuvian we have a specification of the word “year” followed by various seasons. If we read (annus) *462 + ra/i-i as ma _x, + ra/i-i ( = /mari/), an equation of Hitt. me(a)ni- and HLuv. mari- is possible: for rhotacism of n compare maruha (sultanhan 7) with manuha elsewhere. Obviously, however, there are too many uncertainties involved to give the equation much weight.

¹⁵ The reading of the verb stem as marnu- is incompatible with the possible spelling “delere-⌈na⌉^?-nu-tu in carchemish A 4a, 3, for which see Hawkins, , AnSt 20 (1970) 105Google Scholar, and AnSt 31 (1981) 163Google Scholar. D. Hawkins informs me that he unfortunately has not yet been able to collate this passage, so the question remains unresolved.

¹⁶ In his table, AnSt 31.148, Hawkins lists mori as 386 + 381–2, but he now informs me that he takes the form of mori with a vertical bar as merely the relief form of the sign, not as a combination with sign 382, which he and I both read as “wood”, following an old suggestion of Gelb, , HH 3 (1942) 47, note 8Google Scholar. The logogram mori thus consists only of 386 + 381.

¹⁷ Compare Hawkins, , KZ 94.117Google Scholar. For rel-i “as (it were)” = “nearly” compare assur a,3: arha-wa/i-mu-u ∣REL-ri + i ∣mori-ha-na “We died as it were/nearly died” (for the verb ending see Morpurgo-Davies, , KZ 94.99Google Scholar).

Hawkins finds the verb ending -tati problematic, but a solution is suggested by the derivational chain in CLuvian kuwaya- “be afraid” >kuwayata- “fear” > kuwa(ya)tā(i)-“be afraid”. This set shows that Luvian formed verbal nouns in -ta- which served as the base for denominative verbs in -tā(i)- whose meaning was nearly identical to that of the original base verbs. We may therefore suppose that beside a reduplicated stem wawari(ya)-“die” there existed a virtually synonymous wawaritā(i)-. Once such a pattern became productive, it is not even necessary to suppose that the intervening verbal noun actually existed in all cases. Compare the similar situation of Palaic verbs in -nā(i)- which I have discussed in KZ 97 (1984) 37Google Scholar f. Other examples of extended verb stems in -ta-(i)- are attested in HLuvian: cf. pugnus-ri + i-ta ₅-wa/i “I seize”(?) in carchemish A 6, 7, beside the usual pugnus-ri(ya)-. The alternate reading by Hawkins, , AnSt 31.152Google Scholar, is by his own admission forced. Likewise, it is hard to avoid connecting á-ru-wa/i + ra/i-tu of şirzi 3 with Hitt. aruwā(i)- “bow down”. We may account for the extra syllable in HLuvian by assuming a stem /aruwarā(i)-/, rhotacized from *aruwatā(i)-, formed as per above.

¹⁸ Except where the vertical stroke is used to indicate the number “one”. This use of the vertical stroke is already properly distinguished by Laroche as sign 380. Like the oblique “thorn” 383, the vertical stroke 381 as a privative marker occurs only in combination with other signs, never by itself.