88 3 Phonologyis easier to see in connection with VV-Contraction (§3.3.1.3, below). This is
because (unsuffixed) Shlmpf -aeks from /-aeksi/ is monosyllabic, and the fact
that accent shifts to the final syllable of a preceding particle, as in Future ad
0-aeks 'he will eat', does not tell us whether -aeks itself is treated as having
one or two syllables. The (unsuffixed) LoImpfP of this verb is -tatt from
/-tattA-/, suppletive but showing the correct -CdCC shape for this class, but
since the LoImpfP has a grammatical accent, the issue of syllable count is
moot. We can, however, demonstrate that a stem-final V is counted for
purposes of Default Accentuation when we look at the two other stems of the
long imperfective family, which occur with preverbal particles. Compare
LoImpfP i-tdtt 'he eats' with the corresponding LoImpfN stem in war 1-tatt
'he does not eat'. The accent does not shift to the preverb, showing that the
LoImpfN stem is treated as bisyllabic /-tatti/ for purposes of Default
Accentuation. Another example not involving suppletion: -vbsu- 'vomit',
PerfP -absa-, Shlmpf -sebs, LoImpfP -bass-, LoImpfN -bass- (as in war l-bass
'he does not vomit'). Contrast e.g. war i-ggit 'he does not hit' (from C-final
stem -vwvt- 'hit'), where the Neg preverb does bear the (default) accent.
If we look at heavy V-final stems, which have one more syllable than
'eat' and 'vomit', it is easier to see that the stem-final V is "counted" even
when deleted word-finally. Take -fvyku- 'be searched' (PerfP -affayka-), for
example, which belongs to the middleweight subtype of heavy verbs
(§3.4.1.4). The (unsuffixed) Shlmpf is -aeffaeyk from /-aeffaeykA-/. The accent
is stable, as we see in Future ad 0-aeffaeyk 'it will be searched'. This makes
sense if Default Accentuation treats the stem-final /A/ as syllabic nucleus, so
/-aeffaeykA-/ get underlying antepenultimate accent as /-aeffaeykA-/.
The heavy V-final verbs also have VblN's that lose the final V (in this case
III rather than /A/, since the VblN has stem-wide
larger class of heavy VblN's of which this is a special case have grammatically
marked penultimate accent, as seen in α-bakbak 'dusting off (§8.6.1). The
V-final verbs have VblN's with the same grammatically conditioned
penultimate accent at underlying level, which becomes surface final-syllable
accent due to the deletion of the final V, as in α-fayk 'being searched' from
/a-fayki/.
3.3.1.3 Accentual implications of W-Contraction§3.3.1.2, just above, showed that the underspecified stem-final vocalic
segments III and /A/, while deletable word-finally, are recognized in the
syllable count necessary for Default Accentuation to apply. In this section I
consider what happens when III and /A/, or a full-fledged V, contracts with the
initial V of a subject suffix. This forces us to look at augment verbs (those with
Augment -t- in some forms) as well as at unaugmented stems like -vksu- 'eat'
and -fvyku- 'be searched'. It turns out that, just as the quality of the contracted