A Grammar of Tamashek (Tuareg of Mali)

640 12 Extraction processes

(739) a. [a t-ajrasw-aed]

[Dem 2S-find.PerfP-2SgS]

t-3kfo-d-\a-hi-\tt

2S-give.ShImpf-2MaSgS-/Obj-lSgS-3MaSgO

'What(-ever) you-MaSg find (lit. "found"), you (will) give it to

me.' [T-md]

b. [a t-ajraew-aed]

[Dem 2S-find.Reslt-2SgS]

'what(-ever) you-MaSg find (lit. "have found")' [K-d]

c. [a 0-taettae-d], t-szta-d-\t

[Dem 2S-eat.LoImpfP-2SgS], 2S-'pay.ShImpf-2SgS-\3MaSgO

'Whatever you-Sg eat, pay for it!' [K-d]

In (739.c), the verb form t-azls-d (omitting the following clitic)

corresponds to t-szsl-aed in T-ka. In the same example, note that LoImpfP

0-tsettse-d 'you-Sg eat' shows the ablaut modifications diagnostic of definite

relatives, here both χ-pcl Erasure (130) and Rightward Accent Shift (132),

compare the usual main-clause form 0-tdttae-d. The Resit verb in (739.b)

likewise shows χ-pcl Erasure.

When the participle after ά is a high-frequency adjectival participle, it may

retain its shape even when including an apparent χ-pcl due to Resit ablaut. In

other words, the shortening of the accented V due to χ-pcl Erasure seen in

(739.b-c) above does not occur in these more or less frozen adjectival

participles. In laebas-aen (740.a), χ-pcl is moot since the PerfP is already

laebas- with full a, so Resit lasbds- has not undergone secondary

V-lengthening due to ablaut. However, 0-oldr-aen (740.b) with α does show

χ-pcl, cf. PerfP -olaer- 'be good', and if χ-pcl Erasure applied to it we would

expect #0-olaer-aen.

(740) a. [a laebas-aen]

[Dem be.bad.Reslt-Partpl.MaSg]

'something bad'

b. [a 0-oldY-aen]

[what 3MaSgS-be.good.Reslt-Partpl.MaSg]

'something good (=the best thing to do)'

12.1.6.4 'whenever ...' (ed)

There is another generalized relative ed glossed 'whenever ...' or more loosely

'when ...'. It may be equated morphemically with ed 'when/because'

(§13.1.1.1, §13.2.2), cf. e 'where ...' (§12.3.10) and noun e-daegg 'place'. For