A Grammar of Tamashek (Tuareg of Mali)

150 3 Phonology

For those nouns that undergo no audible change in Prefix Reduction
contexts (i.e. those with no vocalic prefix and those with a short-voweled
prefix), there are occasional parsing difficulties at the sentential level. The two
sentences in (128.a-b) are identical in form but have different senses.

The ambiguity is due to the following circumstances, in combination.
First, edi 'dog' does not undergo (audible) Prefix Reduction in the dependent
state, and so in postverbal position it can be taken as either object or subject.
Second, Tamashek has obligatory subject agreement on the verb, whether or
not the coindexed subject NP is also present, so the 3MaSgS prefix i- could
index edi as subject, or could index some other 3MaSg entity ('he') as subject.
Third, 1P1 clitics have identical forms for dative and object function (this is
also true for lSg). In most cases, transitive sentences are clearer, due to either
audible Prefix Reduction (distinguishing subject from other function),
pronominal incompatibility of subject affix on verb with postverbal NP, or use
of a non-lst person clitic (since 2nd and 3rd person clitics distinguish object
from dative function).

In such cases as ϊ-ηγα edi 'he killed the dog' versus i-qY-\e edi 'the dog
killed him', the presence of 3MaSg object clitic -\e in the latter distinguishes
the two, though in allegro speech the difference may be wiped out by
contraction of the two adjacent V's.

3.5.2 Verbs after particles

3.5.2.1 Verbs after Future particles

The (non-imperative) Shlmpf (=short imperfective, §7.2.3.1) stem is always a
"dependent" form. It is used chiefly after Future particle ad (or variant,
§9.6.3), whether or not a Negative particle is also present. An example is ad
0-aeks 'he will eat'.
The Shlmpf is usually just the "basic" (i.e. minimal) form of the verb stem
(e.g. -vCCvC-) plus a melody (, , or ). The Shlmpf lacks the
special ablaut components of the long imperfective (accent and lengthening of
V's, gemination of C's, addition of prefixal -t-).
Since the Shlmpf is not obviously connected to any other specific stem, no
(morpho-)phonological derivation will be offered here to "derive" the Shlmpf

(128) a. i-qra-\0-naer
3MaSgS-kill.PerfP-\Dat-lPl
'He killed the dog for us.'

edi

dog

b. i-ijYa-\0-na2Y

3MaSgS-kill.PerfP-\0-lPl

'The dog killed us.'

edi

dog