78 3 PhonologyIn (55.a), note that the syncopated a leaves a trace behind, in that the /as/ of
3MaPl subject suffix -aen is realized as a. For K-d, I have also seen this with
Shlmpf verb stems of the shape -aCaC- before a V-initial subject suffix, as in
t-azl-ad 'you-Sg pay', cf. T-ka t-azal-asd. In these dialects, some additional
machinery must be recognized to account for the spreading of the Η vocalic
feature from the /a/ to the suffixal /as/. A derivation involving metathesis
/CaC-as/ to /CCa-as/ (equivalent to /CCi-as/) might be considered, but a more
revealing analysis would extract the Η feature from the schwa and attach it to
the /ae/. The details probably differ from dialect to dialect; for some Κ speakers
I recorded e.g. 2SgS -ad after many Shlmpf verbs whose final stem-syllable
had a, so for these speakers a more general harmony seems to be at work; an
example is 2Sg Shlmpf t-ajay-ad 'you-Sg tie'. In any event, my focal dialect is
T-ka, which does not show e.g. 2SgS -ad except due to VV-Contraction from
III plus /ae/ (37.c).
While at least some of the cases in (55.b), are pan-dialectal, they involve a
rather complex and archaic-looking PI formation, where Syncope is combined
with an additional aw added to the stem before the usual MaPl suffix -asn
(§4.1.2.10).
The great majority of stem-final CvC syllables in both nouns and verbs are
quite stable (i.e. unaffected by Syncope) even when followed by a V-initial
suffix or clitic. For example, PerfP -awaet- 'hit' and its Shlmpf counterpart
-awat- do not syncopate (in T-ka or most other dialects) when a V-initial clitic
or suffix is added: ad awat-asr Ί will hit'.
In summary, Syncope as interpreted here applies to a, in the environment
VCaCV including the relevant affixal segments, with various morphological
restrictions (some of them sensitive to the size of the stem). These restrictions
amount to a preference for allowing Syncope to apply to stem-intial CaC with
fewer cases of application to stem-final CaC. The rule is given informally as
(56), which incorporates by reference the morphological contexts just given.
(56) Syncope (asymmetrical version, applying to /a/ but not /ae/)
a —» 0 in the environment VC_CV
[in environments given in (52-55)]In dialects other than our focal T-ka, FePl prefix combination t-i- on
nouns, which reduces to ""t-a- by Prefix Reduction (as in T-ka) in syntactically
defined positions, further syncopates to "t-0- when followed by CV... (but not
by CCV...). Thus t-l-rajw-en 'stomach cavities', but daer "t-a-rajw-en 'in the
stomach cavities' (T-ka), dialectally syncopated to daer nt-0-rajw-en (e.g.
T-md, K-d). In this example, the -i- prefix is accented, so when it is zeroed in
-0- we get a phrasal accent on the preposition.