106 3 Phonology
(80) Templatic Ablaut-Suffixal Plurals (-CaCC-)
gloss Sg PI
a. masculine
'acacia pod'
'horse'
'river'
'tomb'
'vein'
'money' d-zrsf
a-hajjar
e-baeje
e-jrew
a-zskka ι
3-mijj
l-hajr-an
l-bajw-an
Ί-jsrw-an
l-zskw-an
Ί-majy-an
l-zsrf-an
b. feminine
'body'
'tick'
t-a-fskka
t-a-salluf-t
t-l-fakw-en
t-l-salf-en
Dialectally, I can cite some further wrinkles, e.g. ae-dad 'finger' (A-grm)
with ungeminated final C, ΡΠ-ctadw-an. For more examples, see §4.1.2.14.
A further issue in the "componential" versus "templatic" discussion is the
possibility of accounting for some template-like features via a pre-ablaut
reconfiguration of input stems as they go into certain ablauts. The
morphological patterns that I have in mind are the long imperfectives and
perhaps some nominalizations (including VblN) of certain verb classes,
namely -VPQD- and -vPQvC- stems (i.e. light stems whose first C position is a
nongeminate cluster). In this interpretation, the following initial
reconfigurations would be recognized (81).
(81) Pre-Ablaut Reconfiguration of Stem
Such reconfigurations are helpful in allowing us to simplify the rules for
associating ablaut formatives, notably in the long imperfectives. For other
verbs, the LoImpfP is characterized (among other things) by χ-pcl and χ-pcl,
which respectively accent and lengthen the "first postconsonantal C" (see
below. The VblN's and other nominals are less transparent since each stem-
shape class has its own patterns, but we do frequently find a vowel between the
Ρ and Q consonants.
Of course, given e.g. -vntu- versus -nvti>, we might flip them and take
-nvtu- as basic, or even take -vnvtu- as basic and chip away at the short V's
(e.g. with Syncope). However, taking 'destroy' as having a basic form -hvlvk-
is undesirable, since there is another class with lexical -CvCvC- shape that has
gloss basic stem reconfigured LoImpfP VblN
a. 'begin' -vntu- -nvtu- -natt
b. 'destroy' -vhlvk- -hvlvk- -hallaek-
e-naett
a-haelak