language and script 101
instead of a prefixed, /t/. this form, exceptional in aramaic, may be a
vestige of an older stage of Semitic in which the /t/ was originally an
infix.94 the loss of intervocalic /h/, which presumably triggered or rein
forced the later shift of the causative prefix /ha/ to /ʾa/, seems to underlie
the nonstandard form of the “imperfect” yskr /yasker/ (instead of yhskr /
yahasker/ earlier in the same line) ‘may he deliver’ (Kai 224: 3) in Sfire.95
unlike all known varieties of aramaic, Samʾalian has allegedly preserved
a reflex of the protoSemitic nstem. the nstem reduces transitivity with
fientive verbs and thus acts as a mediopassive there, whereas it expresses
a fientive nuance (which often seems to coincide with an ingressive situ
ation type highlighting a change of state) with stative verbs.96 its form in
1stmillennium B.c. northwest Semitic can be reconstructed as follows
(assuming that /e/ is the reflex of */i/):
“perfect” “imperfect” imperative participle infinitive
n /naktab/ /yekkateb/97 /ʾekkateb/ /naktab/ /naktāb/the historical significance of this feature depends on one’s assumptions
about the position of Samʾalian within Semitic. Scholars who associate
the local idiom of ancient Zincirli with a variety of the northwest Semitic
branch preceding the split into canaanite and aramaic, view the nstem
as a retention from an earlier developmental stage of the language. those
who basically subsume Samʾalian under aramaic, conversely, would rather
classify the relevant examples as borrowings from phoenician, which has
preserved a productive nstem. one should at any rate emphasize that
evidence for an nstem in Samʾalian is very feeble indeed: it is restricted
to the alleged and partially reconstructed participle nḥ[š]b /naḥšab/
‘esteemed’ from the root ḥšb ‘to reckon’ in Kai 215: 10. if this reading can
94 alternatively, one could think of an akkadian influence, but the usual tendency of
infixes to turn into prefixes (rather than the other way round) would argue in favor of an
archaism in this most ancient aramaic text. cf. gzella 2009: 302. as a consequence, the
purported metathesis of /t/ with a rootinitial sibilant may actually be a remnant of an
older infixation that resisted the shift from infix to prefix. inconsistent evidence for this
phenomenon in Sefire (see the remark on the tstems) could also be adduced in support
of this idea but the distribution of forms with and without metathesis in old aramaic
remains unknown.
95 See Beyer 1984: 148 and fitzmyer 21995: 145. even if this is a scribal mistake (cf.
degen 1969: 19 n. 79) rather than a purposefully innovative spelling, it may have been
caused by a phonetic change that was already underway in contemporary pronunciation.
96 gzella 2009 provides a functional analysis and bibliographical references.
97 from /yinkatib/ with assimilation of /n/; this also applies to the imperative
/ʾinkatib/.