jeff_l
(Jeff_L)
#1
108 3 Phonology
For some cases of T-c2 in plural ablaut of nouns, see the following section
and §4.1.2.23.
3.4.2.2 Plain/geminate alternations in nominal Sg/Pl
Phonetically, h and to a lesser extent (tap) r tend to resist audible gemination
(i.e. increased duration), though at least in careful speech a morphologically
expected gemination can be audible.
Variation between geminated and ungeminated C's occurs in certain
nominal singular/plural alternations. In some cases it is not clear that stem-
wide ablaut is at hand. The following patterns can be distinguished.
First, in the type Sg t-a-duf-t 'wool', PI t-l-duff-en, it would seem that the
FeSg suffix -t forces Degemination of the geminate, which would therefore
be present underlyingly even in the Sg (so no ablaut-induced gemination
process need be recognized). In the full set of known examples (given in (152)
in §4.1.2.8), the degeminated C is from the set [ywrm f}, i.e. sonorants and
fricatives. The phonology here is somewhat opaque, and there is some dialectal
variation in detail.
Second, there are some masculines with Sg -CVP (V a full high vowel)
and PI I-CaPP-asn, e.g. ά-ril 'right hand', PI l-rall-asn. This is a rather
specialized Sg/Pl pattern with a semi-templatic look. For examples and
discussion see (153) in §4.1.2.8.
Third, there are a small number of cases like e-raess 'bone', PI i-raes-an,
where the Sg ends in a geminate, but where the PI shows an ungeminated C
before the MaPl suffix. The MaPl allomorph -an instead of -asn, combined
with the accent shift in the PI and the absence of a PI ablaut melody, show that
the PI is of the underlying type /-raesV-aen/ with some stem-final V. Therefore
we have a phonologically obscure alternation between a geminate PP and PV.
For more examples and discussion see (155) in §4.1.2.8.
Fourth, there are some cases where a medial or final geminate in the Sg is
degeminated in order to fit the stem into the templatic shape -CsCC-, as
mentioned in §3.4.1.5, above. Example: a-hajjar 'acacia pod', PI l-hsjr-an.
See (180) in §4.1.2.14 for more details.
Finally, there are two unsuffixed ablaut PI patterns that also involve
gemination in the PI only. These are probably the best candidates for a true
ablaut formative T-c2 outside of verbal morphology.
One type ((205.a) in §4.1.2.23), applying to Sg type t-a-CCsC-t (a fairly
common deverbal nominalization, but including some underived nouns), is
exemplified by t-a-ftsq-q 'laying out', PI t-i-föttar. The vocalism and
V-length patterns of the PI are consistent with PI ablaut. There are a few
similar plurals for nouns of the Sg type t-a/e-CvCuC-t with a full vowel υ in
the final syllable (205.b). Note in particular the short "v" in the position
corresponding to the a in the PI. An example: t-a-jSzel-t 'kidney', PI
t-i-jszzal. With cases like 'kidney' in mind, one could argue that t-a-ftsq-q
