A Grammar of Tamashek (Tuareg of Mali)

3.4 Ablaut 123

derivation (95), namely syncope of a second-syllable schwa accompanied by

switch of a stem-initial schwa to ae. However, to my knowledge, T-ka does not

show this output type in short imperfectives: ad awst-aer Ί will hit'.

In A-grm, Stem-Initial Syncope does not apply to the perfectives in (94),

but we get stem-wide <L> vocalism anyway. Thus -xaebabae-t, -taerara-,

-s-aemm-askae-t, -s-aennaem-aehael-, and so forth. In A-grm, therefore, there is

no connection between Leftward L-Spreading and Syncope, since the

spreading rule applies in the absence of Syncope. I will not venture here into

the question whether A-grm preserves an archaic, Proto-Tuareg system, or has

recently reshaped the perfectives (ultimately on the model of the Imprt, e.g.

xsbubs-t, tsrur).

Aside from the cases where (outside of A-grm) Stem-Initial Syncope and

Leftward L-Spreading are correlated, there are a few nonsyncopating verb

types that have stem-wide <L> melody in the perfective. These are light stems

beginning with a full V (-uCu-, -uCvC-, -uCCi>, etc.), and middleweight

stems of shape -CuCvC- with a full V after the first C. An example is in (96).

(96) Perfective Verb (Surface <L> Melody, no Syncope)

lexical form PerfP gloss

-dubvn- -aedobaen- 'marry'

This perfective has a medial mid-height V, here o. See §3.4.7, just below,
for more data, and an analysis involving V-Height Compromise. The initial ae
(instead of a) in the PerfP forms of 'marry', 'be open', and 'be held in
common' suggest that the melody is stem-wide.
I see no reasonable way to integrate the perfective melody seen in (96)
with the surface melody described above for verbs subject to Stem-Initial
Syncope. Historically, it is possible that the tripartite perfective <L Η L>
melody seen in non-T-ka dialects in perfectives whose second V is a full high
V, as in Resit -aexibabae-t '(hole) gape' with vocalic sequence «LHLL», is
archaic. Using "M" for mid-height V, the surface bisyllabic trisyllabic «LML»
vocalic sequences seen in -aedobaen- (96) might be an additional vestige of a
tripartite <L Η L> melody.

3.4.7 V-Height Compromise

As just mentioned, there are some full-V-initial light stems (-uCi>, -uCvC-,
-uCCi>, etc.) and some -CuCvC- stems with full (high) V after the first C,
whose perfectives have mid-height V's. Additional data are in (97).