A Grammar of Tamashek (Tuareg of Mali)

358 7 Verbal morphology

that work in the unsuffixed Shlmpf also affect the Sg Imprt. For 'eat' the
imperatives are Sg seks, MaPl asks-aet, and FePI aksa-mast. For 'go' we get
(with T-ka accentuation) Sg ajal, MaPl aejl-aet, FePI ajla-mast. However, the
Imprt (unlike the inflectable Shlmpf) can be directly followed by clitics,
including V-initial clitics like 3MaSgO -\e. These V-initial clitics block
resyllabification: Imprt Sg /aerri/ 'read!' appears in isolation as arar, but with
a V-initial clitic as aerr-\e 'read it!'. The exception is that in several dialects
/-asswi-/ 'drink' remains resyllabified as -ssii- even before clitics, which in this
event must take postconsonantal allomorphs, hence asu-\tt 'drink it' alongside
the regular but less common £sw-\e 'drink it!'. For more detail on
resyllabification see §3.2.4 and §3.3.2.
As noted above, long imperfective stems of the α/ι subclass require a pre-
ablaut reconfiguration from -vPQu- to -PvQi>, whereupon the regular
processes of long imperfective ablaut will work. Since there is no u vowel, the
characteristic long imperfective melody is in T-ka. The consonantal
modification for this class is T-c2.
The regular LoImpfP for stems of the α/ι subclass that have a
nongeminate PQ cluster is /-PdQQA-/. /A/ is a deletable stem-final low V; see
(354), above, for its manifestations. A sample paradigm is in (362), using
-vbsi> 'vomit' (PerfP -sbsa-).

(362) LoImpfP Paradigm of 'vomit'

a. stem-final V is deleted word-finally

3FeSg 0-bdss, rarely ts-bdss

1P1 na-bdss

3MaSg i-bdss

b. stem-final contracts with suffix-initial /ae/ to give ae

ISg bdss-aer

2Sg 0-bassae-d, rarely ts-bdssas-d

2MaPl 0-bdssae-m, rarely ts-bassas-m

3MaPl bassae-n

c. Final V appears as se before C-initial subject suffix

2FePl bassas-maet, rarely ts-bassae-mast

3FePl bassas-naet

The /A/ is "counted" in Default Accentuation, in theory. This is moot in
the LoImpfP, with its marked accent. In e.g. sassae-n-\taet 'they-Ma drink it-
Fe', from /sdssae-aen-\taet/, arguably Default Accentuation produces
/sassae-aen-\taet/. If so, when /ae + ae/ contract, the accent shifts to the next
syllable to the left, so in the end there is no audible evidence here for counting
stem-final /A/. I am also unable to find a context where a Prohib stem like
-baess- (/-baessA-/ shows concrete evidence that the /A/ is counted. The Prohib