A Grammar of Tamashek (Tuareg of Mali)

7. 1 Augment verbs with -t- 295

(303) PerfP of 'fly away'

a. VV-Contraction (37.d) to e with V-initial suffix

lSg affurre-Y

2Sg t-sffurre-d

2MaPl t-affurre-m

3MaPl sffurre-n

b. Augment -t with no subject suffix

1P1 n-sffurrae-t

3MaSg l-ffurrae-t

3FeSg t-affiirrae-t

c. Augment -t- before C-initial suffix

2FePl t-affurrae-t-maet

3FePl affurrae-t-naet

In other dialects (A-grm Gao Κ R T-md), such stems begin with ae rather
than a in the perfective system (e.g. PerfP -aeffurrae-t). T-ka has harmonized
the stem-initial short V with the V of the following syllable (cf. §3.2.6),
but in T-ka there is no synchronic evidence that the initial V is underlying fx/.
The stem-final V is shortened before -t in ordinary inflected forms (PerfP,
Shlmpf), but when the final V is protected by a χ ablaut feature (as in the long
imperfectives) it does appear as a full V before -t. Thus 3MaSg PerfP
Ί-ffurrae-t (303.b) and 3MaSg Shlmpf l-ffurrs-t have short pre-augment V, but
3MaSg LoImpfP i-t-lfarru-t shows a full pre-augment V. See Pre-Augment
V-Shortening (115) (§3.4.9.2).
There are a few augment verbs, belonging to a broader adjectival class,
whose perfective inflected forms allow a full V to surface before the -t-,
violating Pre-Augment V-Shortening, and usually keep the -t- even before
V-initial suffixes. An example is 'be speckled' (372.d), with 3MaSg PerfP
kaesa-t (not #kaesae-t) and 3MaPl kaesa-t-aen (not #ksese-n in this sense,
though the latter is attested as a MaSg participle). The verbs in question have
PerfP CaeCa-t and Imprt QeCae-t, or PerfP CaCa-t and Imprt CaCae-t. The
Imprt forms suggest that these verbs have a basic shape -CVCu- (+ -t), but
have the peculiarity that the stem-final υ is unconditionally lengthened to α in
the perfective. There are other (C-final) adjectival verbs that have a similar
long α (or other long V) in the second perfective syllable, e.g. PerfP -aezzay-
'be heavy', bsehaw- 'be grey', and ssggcnr- 'be red'. While these adjectival
perfectives are lexically irregular, one could argue that they include a
lengthening formative χ (perhaps χ-f, i.e. applying to the final stem syllable),
and this χ blocks (or undoes) Pre-Augment V-Shortening in augment verbs
like CaeCa-t.
In (304), below, we see that the augment -t- is present in all imperative
forms. Based on the non-imperative data in (303), this is not surprising in the