For the most part, Proto-Anatolian has been reconstructed on the basis of
Hittite, the best-attested Anatolian language. However, the usage of
Hittite cuneiform writing system limits the enterprise of understanding and reconstructing Anatolian phonology, partly from the deficiency of the adopted Akkadian cuneiform syllabary to represent
Hittite phonemes and partly from Hittite scribal practices. It is especially pertinent to what appears to be confusion of voiceless and voiced dental stops, in which signs -dV- and -tV- are employed interchangeably in different attestations of the same word. Furthermore, in the syllables of the structure VC, only the signs with voiceless stops are usually used. Distribution of spellings with single and geminated consonants in the oldest extant monuments indicates that the reflexes of
Proto-Indo-European voiceless stops were spelled as double consonants and the reflexes of PIE voiced stops as single consonants. This regularity is the most consistent in the case of dental stops in older texts; that merger was a secondary shared innovation in Hittite,
Palaic and
Luwian, but not in
Lycian. Concordantly, Common Anatolian had the following short vowel segments: */i/, */u/, */e/, */o/ and */a/. Among the long vowels, */eː/ < PIE *ē is distinguished from */æː/ < PIE *eh₁, with the latter yielding
ā in Luwian,
Lydian and Lycian. Melchert (1994) had also earlier assumed a contrast between a closer mid front vowel */eː/ < PIE *ey (yielding Late Hittite
ī) and a more open */ɛː/ < PIE *ē (remaining Late Hittite
ē), but the examples are few and can be accounted for otherwise. PIE syllabic consonants are also preserved in Common Anatolian. The status of the opposition between long and short vowels is not entirely clear, but it is known for certain that it does not keep the PIE contrast intact, as Hittite spelling varies in a way that makes it very hard to establish whether vowels were inherently long or short. Even with older texts being apparently more conservative and consistent in notation, there are significant variations in vowel length in different forms of the same lexeme. It has been thus suggested by Carruba (1981) that the so-called
scriptio plena represents not long vowels but rather stressed vowels, reflecting the position of free
PIE accent. Carruba's interpretation is not universally accepted; according to Melchert, the only function of
scriptio plena is to indicate vowel quantity; according to him the Hittite
a/
ā contrasts inherits diphonemic Proto-Anatolian contrast, */ā/ reflecting PIE */o/, */a/ and */ā/, and Proto-Anatolian */a/ reflecting PIE */a/. According to Melchert, the lengthening of accented short vowels in open syllables cannot be Proto-Anatolian, and the same goes for lengthening in accented closed syllables. However, in Hoffner and Melchert (2008), it is stated that scriptio plena was rarely used to mark yes-no questions, since there was no question mark in Hittite script. This usage comes from scribes in Assyria and Babylonia who wrote
Akkadian in
cuneiform script.
Consonants One of the more characteristic phonological features common to all Anatolian languages is the gemination of the Proto-Indo-European voiceless consonants (including the sibilant *s and the laryngeal *ḫ) between unstressed syllables and following long vowels. The two can be considered together as a gemination rule between unstressed
morae, if long vowels are analyzed as a sequence of two vowels. All initial voiced stops in Anatolian eventually merge with the plain voiceless stops; Luwian, however, shows different treatment of voiced velar stops *G- and unvoiced velar stops *K- (initial *G being
palatalized to */j/ and then lost before /i/, unlike *K), showing that this was a late areal development, not a Proto-Anatolian one. It is not known what the placement of the Anatolian laryngeal was, one theory is it was pronounced as a pharyngeal fricative /ħ/, but it could have also been velar /x/, uvular /χ/, or glottal /h/ based on PA's descendents. The sequences *h₂w and *h₃w yield a labialized laryngeal *ḫʷ. The liquids and nasals are inherited intact from Proto-Indo-European, and so is the glide *w. No native Proto-Anatolian words begin with *r-. One possible explanation is that it was true in Proto-Indo-European as well. Another is that it is a feature of languages from the area in which the daughter languages of Proto-Anatolian were spoken. ==Morphology==