A Grammar of Tamashek (Tuareg of Mali)

3.3 Accent 91

(66) Perfective Negative of 'stretch on tiptoes'

a. 3MaSg waer 0-ojja

b. VV-Contraction precedes Default Accentuation, e from

VV-Contraction

lSg waer ojje-r

2Sg waer t-ojje-d

c. VV-Contraction follows Default Accentuation, e ablauted from /ae/

3MaPl waer ojje-n

2MaPl waer t-ojje-m

This paradigm has been verified for several dialects (T-ka, K-d, R). In
effect, then, the stem-final V of the Perf stems is deleted outright before the
3MaPl and 2MaPl suffixes (66.c). However, the penultimate accent, audible in
(66.c) and latently present in (65.c), suggests that this V is present when the
accentuation rule applies, so VV-Contraction follows Default Accentuation.
On the other hand, the stem-final V contracts with the initial short V of the lSg
and 2Sg suffixes (66.b), but this VV-Contraction precedes Default
Accentuation.
In trisyllabic unaugmented V-final stems, the paradigms are simpler. The
PerfN ablaut element e is not applicable to these superheavy stems (it is
audibly present only in bisyllables). Moreover, the lSg and 2Sg suffixes are
treated just like the 3MaPl and 2MaPl suffixes, so the stem-final V is deleted
outright but leaves a trace behind in the form of surface penultimate accent.
(67) shows the PerfP forms of a sample verb; the perfective negatives are
identical (but are preceded by waer).

(67) Perfective Positive of 'be confused'

a. 3MaSg 0-aemtaella

b. lSg aemtaell-aer

2Sg t-aemtaellae-d

c. 3MaPl aemtaellae-n

2MaPl t-aemtaellae-m

The Shlmpf /-amtalli-/ shows similar accentual behavior, allowing for the
deletion of word-final /i/: (Future) 3MaSg adl-mtall 'it will be confused', lSg
ad smtall-aeY, 3MaPl ad smtalb-n. LoImpfP forms are 3MaSg i-t-lmtalli 'it
is (often) confused', lSg t-imtalli-r, 3MaPl t-imtslli-n.
A further example is 'go south', PerfP 3MaSg 1-jjussa but lSg sjjuss-aeY
and 3MaPl ajjüssae-η. A causative example is 1-ss-ijja 'he bent his head', but