jeff_l
(Jeff_L)
#1
10.4 Ordering of clitics 613
(686) i-qre-\n-\ha-hi
3MaSgS-kill.PerfP-\3MaP10-\Dat-lSg
'He killed them-Ma for me.' (T-md)
Here the 3MaPl object has a reduced, arguably V-initial allomorph (/-aen-/)
after the V-final verb (-sgra-), and we get VV-Contraction with /a-ae/
surfacing as e. It appears that this "fusion" of object clitic with stem-final V
permits an exception to the first person first rule. However, even here, some
speakers including my primary T-ka informant strongly prefer to allow the first
person first rule to apply, so instead of (686) they have i-qYCi-\0-hi-\taen, with
3MaP10 -\taen at the end.
In the combinations not involving a 1st person clitic, the common
sequence is dative-object. In other words, the object-2 slot is favored for
object clitics. One can account for the data by recognizing (684.b) as the basic
order, while allowing for subsequent "fronting" of a lSg or 1P1 form to initial
position within the clitic cluster. However, there are also some attestations of
object-dative order even when a 1st person clitic is not involved, so the
ordering is not rigid.
In the uncommon case where two objects and one dative co-occur, the two
object clitics flank the dative clitic, so we get object-dative-object order.
Some further examples each involving one dative clitic and one object
clitic are in (687). The verb glossed 'kill' can also mean 'hit hard, whack'.
(687.a-b) include a 1st person, while the others do not.
(687) a. waer-\hl-\tt i-ηγα
Neg-\lSg(Dat)-\3MaSgO 3MaSgS-kill.PerfN
'He did not kill it for me.'
b. i-waet-\a-hi-\tt
3MaSgS -hit.PerfPADat- lSg-\3MaSgO
'He hit it for me.'
(#i-w£et-\t-\ha-hi usually rejected)
c. i-waet-\t-\ha-s
3MaSgS-hit.PerfP-\3MaSgO-\Dat-3Sg
'He hit it for him/her.'
d. i-waet-\ha-s-\t
3MaSgS-hit.PerfP-\Dat-3Sg-\3MaSgO
[=c]
e. i-waet-\a-hi-\ha-s
3MaSgS-hit.PerfP-\0-lSg-\Dat-3Sg
'He hit me for him/her.'
(#i-waet-\a-s-\ha-hi usually rejected)