A Grammar of Tamashek (Tuareg of Mali)

(Jeff_L) #1
446 8 Verbal derivation

Therefore causative long imperfectives always have full V's in both the
initial and final syllables. For example, the LoImpfP of -s-vjvn- 'make kneel'
is -s-ajan-, where both V's have been lengthened by ablaut formatives.
A fifth idiosyncrasy is that when u-Spreading (119) (§3.4.9.3) spreads the
features of u in the first syllable (i.e. directly after the prefixal sibilant) to a full
V in the immediately following syllable, the first u loses its rounding and we
get a vocalic sequence «i u» instead of «u u». In other words, the back and
rounded features of u are not copied, they are picked up and transported,
leaving i as the unmarked high full V. Thus (bisyllabic) -s-urvl- 'give back',
LoImpfP -s-irul-, LoImpfN and Prohib -s-mil-. There is no exactly parallel
underived shape that one can contrast this with, since there are no underived
bisyllabic LoImpfP stems of the shape -t-uCuC- (because of clash avoidance).
Trisyllabic causative LoImpfP -s-Is-uduh- 'make pound', see (455), below,
allows u-Spreading to work exactly as in underived trisyllabic LoImpfP
-t-lCuCuC- (e.g. -t-ldubun- 'marry') with -t- prefix. So there is no real
phonological discrepancy between causatives and underived verbs with regard
to u-Spreading. However, causative LoImpfP -s-irul- with i...u does differ
from the other known bisyllabic u-Spreading case, the VblN type a-budar
'being rude', PI i-budur-aen with u...u (§8.6.1.4), unless the PI prefix i- in the
latter is taken as making the first u "medial."
Care must be taken in causatives to distinguish the LoImpfP from the
Resit, which begins with -s(s)-u... but does not lengthen the final syllable and
has (at least after the υ) the same melody as in the PerfP.
The absence of -t- in the LoImpfP in causatives is a fail-safe diagnostic
for distinguishing "causative" stems (including frozen ones with no associated
underived verb) from other heavy stems that happen to begin in a sibilant. For
example, the verb 'reply', Imprt s-üdmar, is shown to be a causative by its
LoImpfP -s-ldmur- (A-grm -s-udmir-). By contrast, -süssaba-t 'pound (millet)
to remove bran', which looks at first like it might be a causative
morphologically, has LoImpfP -t-isassabi-t with -t- prefix, and is therefore not
causative in form.
The VblN is basically regular, i.e. follows the same rules as apply to
VblN's of other heavy stems. However, the VblN of causatives shows the
same C,-Gemination pattern as was described above for short imperfectives.
As an example, consider the non-augment stem -s-vbvlvwlvw- 'make soggy'.
The Sg VblN is α-s-abbalawlaw with variant a-s-abbalawlaw. Aside from the
C,-Gemination (bb), these forms show the stem melody, a- Sg vocalic
prefix, penultimate accent (χ-pen), and optional replacement of final-syllable a
by α that we see in other heavy VblN's. Likewise, in the PI we get
i-s-abbala wliw-aen (with final-syllable schwa lengthened to i) or
i-s-abbalawlaw-asn. If the causative is augmented, we again get melody
and Ci-Gemination where applicable, but default rather than penultimate
accent. Thus -s-vfvqqu- (+ -t) 'hatch', VblN α-s-affaqqu. See §8.6.1.4-6 for
more on VblN ablaut of heavy stems.

Free download pdf