80 holger gzella
the nasals /n/ (dental) and /m/ (bilabial), and the glides (semivowels) /y/
(palatal) and /w/ (bilabial). the stops /b/, /g/, /d/, /k/, /p/, and /t/ were, as
far as one can tell, still unaspirated plosives in all positions;22 the pronun
ciation of the “emphatics” (which may once have been glottal ejectives),
the lateral /ś/ and the aramaic reflex of /ṣ́/, however, are controversial.
at least the latter seems to have been close to /q/, if indeed the choice
for the grapheme {q} results from a similar pronunciation; consequently,
it is at times understood as a voiced emphatic fricative.23 Some scholars
also suppose that earlier aramaic still preserved the distinction between
protoSemitic /ḥ/ and /ḫ/ (ipa /x/), both written with {ḥ}, on the one
hand and between /ʿ/ and /ġ/ (ipa /ɣ/), written with {ʿ}, on the other,24
but this is more difficult to demonstrate. aramaic phonemes without
a corresponding grapheme were spelled with the letter designating the
respective sound correspondence in canaanite (as with the interdentals)
or the most similar sound (as, presumably, with the reflex of /ṣ́/; the same
applies to {s} for /θ/, which is closer to /s/ than to /š/, in tell fekheriye,
at a greater distance from the phoenician sphere of influence). all con
sonants could be lengthened (“geminated”), but it would seem that even
lengthened consonants were only articulated once, like mamma, fatto,
spesso, etc., in italian.
at some stage after the old aramaic period, the merger of the inter
dentals with the corresponding dentals (/ð/ > /d/, /θ/ > /t/, /θ̣/ > /ṭ/) and
the slightly later merger of the reflex of /ṣ́/ with /‘/ had been completed.25
Yet orthography often lagged behind these phonetic developments, hence
historical spellings like {z} for old /ð/, then pronounced like /d/, lingered
on in a few highfrequency words such as the demonstrative pronouns
(znh /ðenā/ > /denā/ ‘this [m.sg.]’) and the relative marker (zy /ðī/ > /dī/),
especially in formal orthography (as opposed to substandard dnh and
dy). in part, these mergers may have been underway in the last stages
of old aramaic. the reason for this hypothesis is that at least the Bukan
inscription (Kai 320), which dates from ca. 700 B.c. and was discovered
in the western part of presentday iran, oscillates between traditional {š}
for /θ/ in šwrh /θawrā/ ‘cow’ (line 5) and {t} for the same etymological
sound in tnn /θanān/ > /tanān/ ‘smoke’ (line 8). a chronological gap of
22 Beyer 1984: 125–128.
23 See Beyer 2004: 45f, 51.
24 Beyer 1984: 101f.
25 Beyer 1984: 100 and id. 2004: 51.