4.1 Noun morphology 205(we can't tell whether this is true of the PI of 'well rope'). In (181.a) we see
that Sg gg degeminates to w. The noun 'day' in (181.b) is irregular in that the
stem V is short in the Sg but full (=long) in the PI. The opposite happens in
(181.d), where a full V in the Sg appears as schwa in the PI. In (181.e), the full
V following stem C 2 is syncopated, as usual, and the schwa in the PI could be
explained as due to Schwa Insertion (breaking up a triple CCC cluster).
In (182) I give some cases where the Sg already has a medial schwa.
(182) Plural Ablaut (Singular with Schwa)
singular plural glossa. Sg ά-CCsC, PI not resyllabified
ά-dhaj i-dhaj-an 'aardvark'
ά-lmaz l-lmaz-an 'twilight'b. Sg ά-CCsC, PI resyllabified as -CaCC-
ά-zraf l-zarf-αη 'money'
ά-drezII II "i-darz-an 'sauce'
ά-rram l-varm-an 'town'
ά-sbaj i-sabj-αη (T-ka) 'bracelet'
[PI alsol-sbag-aen (A-grm)]c. Sg ά-CaC
ά-yas l-yas-αη 'horse'd. Sg a-CaCC
a-haks Ί-hsks-an 'acacia tree'
ά-jatt l-jatt-αη 'tent stake'Because the Sg and PI stems (disregarding affixes) are identical in (182.a,
c-d) it is indeterminate whether the PI has undergone ablaut. In some cases
there may have been an original stem-final V that contracted with *-aen to
produce the attested -an. In (182.b), Sg -CCaC corresponds to PI stem -CaCC-,
so one could argue that transfer onto the template -CvCC- has occurred. The
fact that the MaPl suffix appears as -an with full a, and the unlikelihood at
least in (182.a-b) of a deletable stem-final V (historically or synchronically),
suggest that the plurals are of the same type as the ablautedl-CvCC-αη plurals
described earlier in this section.
The forms discussed so far in this section have MaPl -an instead of -aen.
There is no similar audible distinction between FePl suffix allomorphs, the
suffix being uniformly -en. However, there are some feminine nouns that show
the same Sg/Pl stem alternations seen above for masculines. In (183) I give
some masculine/feminine pairs that bring out the similarities especially well.