/ammi(ya)-/, á-mi-

The phonological interpretation of this word can be deemed certain. Not only does its spelling provide copious evidence, it also belongs to the word family around /ammu/ ‘I’ with further cognates in the other Anatolian languages, e.g. Hittite ūk/ammuk ‘I’, Lyd. amu ‘I, me’ and many more.

The spelling of the word is remarkable in so far as it shows considerable variation. Some of them attest to linguistic change such as the contraction of /-iya-/ to /-i-/ and the rhotacism affecting intervocalic /d/, both representing common developments in HLuw. (see e.g. Melchert 2003b:179-183 and Yakubovich 2016b:478).

In contrast to that, the numerous instances of initial-a-final, in which 〈a〉 (sign *450) is found at the end instead of the beginning of the word and hence transliterated as 〈*a-〉 (see Hawkins 2003b:159-161), mark a change only in orthographic conventions from the 2^nd millennium to the early 9^th century BCE (see Melchert 2010a:147-149).

A related phenomenon becomes evident in the occurrences lacking graphic representation of the an initial /a/ altogether, whether spelt first or last, which make up for less than ten per cent of the total number of attestations. They frequently appear in texts that also contain forms with the initial /a-/ spelt out, as noted by Melchert 2010a:151f.

Finally, attestations from the mid-9^th century BCE and later are frequently spelt with 〈á〉 (*19) instead of 〈a〉 (often in locations that featured initial-a-final before, see Melchert 2010a:150f.), even though 〈a〉 and 〈á〉 are not usually mutually interchangeable. Altogether, the use of 〈á〉 has caused much scholarly discussion, not least because 〈á〉 appears to be limited to word-initial position by and large.Notably, it is never used for the initial-a-final practice. Thus, Kloekhorst 2004a (again Kloekhorst 2006a) argues that 〈á〉 represents a different sound than 〈a〉, viz. a vowel preceded by a glottal stop, i.e. /ʔa/. Differently, Rasmussen 2007a:165 assumes that 〈á〉 is used to reflect /e/, which is unlikelySee Simon 2013a:9f. for a comprehensive critique. since it is generally agreed that Luwian has a three-vowel system /i, a, u/ and distinguishes for length (cf. e.g. Melchert 2003b:185). Yet another hypothesis is put forward by Poetto 2010a:272, who suggests (following up on an earlier note in Poetto & Salvatori 1981a:19) that 〈á〉 might stand for /ya/, although he can draw on only tenuous onomastic evidence, and the present lexeme appears to be a counter example. Examining word-initial plene spellings across Luwian, Simon 2010b:263f. sees 〈á〉 in the role of representing long /ā/ as well as /ʔa/. Melchert 2010a:153 notes that 〈á〉 and 〈a〉 are no longer complementarily distributed from the middle of the 9^th century BCE onwards, arguing against 〈á〉 representing a glottal stop. This view is met with hesitation by Weeden 2011a:67f. in his review of Kloekhorst’s EDHIL. Reviewing the data yet again, Simon 2013a reaches the conclusion that it does represent /ʔa/ after all. No position is taken here.

[AH.B.]