A Grammar of Tamashek (Tuareg of Mali)

7.3 Verb classes and irregular verbs 363

(366) -vCCu- verbs, α/u subclass

gloss PerfP Imprt LoImpfP

a. from Arabic

'create' -abna- abnu -binnu-

'have mercy' -3?fa- 3?fu -ftffu-

'not need' -ama- amu -rlnnu-

'benefit' -anfa- anfu -nlffu-

'consent' -srdaι ι ardu -riddu-

'complain' -aska- asku -sikku-

'err' -axda- axdu -xiddu-

'be dirty' -axla- axlu -xillu- (K)

b. PQ with Q more sonorous

'be split' -afla- aflu -fillu-

'be sick' -akma- akmu -kimmu-

'go' -agla- aglu -gillu- (K)

[elsewhere a/i subclass, e.g. T-ka -vjli>, LoImpfP -jail-]

'apply henna' -arna- srmu -Yimmu-

'inherit' -stra- stru -tlrru-

'tire easily' -azra- azru -zlrru- (R)

'wring' -azma- azmu -zlmmu-

c. other PQ

'be separated' -abda- sbdu -blddu-

'be complete' -amda- amdu -middu-

Quite a few of these are Arabic loans (366.a). Of the remaining stems,
most have PQ clusters with Q more sonorous than Ρ (366.b); I cannot say
whether this is significant in any way.
The common VblN pattern is the feminine nominal t-a-CaCCaw-t, hence
t-a-baddaw-t,II t-a-bannaw-t, t-a-fallaw-t, t-a-rammaw-t, and t-a-rannaw-t.
The final w here is possible evidence of an original stem-final w, but other
interpretations are possible, and synchronically it would make more sense to
say that the lexical u is mapped onto a C position as w.
There are some indications of leakage over time from this subclass into the
more conspicuous α/ι subclass. For -akta- (Imprt aktu or £kt, LoImpfP
-kittu-) 'remember', -arna-(arnu varying with sern, -rinnu-) 'triumph', and
-asla- (aslu or asal, -sillu-) 'distract', my data show variable Imprt (=ShImpf)
forms: either aCCu, or else aeCC (if necessary resyllabified to aCaC) as with
the α/ι subclass.
All verbs known to me with final u in both perfective and imperfective
stems are in (367).