Valency mismatches and the coding of reciprocity in ... - Linguistics

Valency and reciprocity in Australian languages 563
in case frame, from ergative-accusative (22a) to nominative (22b), plus the use
of the reciprocal particle galagala.
(22) a. Ngiya:la bu:rwaw ngi:na.
1pl.inc.erg paint.fut 2sg.acc
‘We will paint you.’
b. Ngiya: galagala bu:rwaw.
1pl.inc.nom part paint.fut
‘We will paint each other.’
Karrwa (Garawa) (Furby & Furby 1977: 62) uses a single free reciprocal/
reflexive pronoun without other constructional coding (23a). This has a special
form, distinct from standard pronouns, but it is difficult to say what its case
is since it does not participate in a case paradigm. However, evidence from
other nominals that may be apposed to it in the same clause suggest it takes
the nominative rather than the ergative/instrumental: Mushin (2000) gives example
(23b) in which the nominal kurrukurru ‘big mob’, seemingly forming a
discontinuous NP with the reciprocal pronoun and therefore presumably in the
same case, takes the zero form characteristic of the nominative (in this example
the subject is grammatically singular but semantically plural).
(23) a. jala daba yalu-ngk=i.
then hit 3pl-rr=pst
‘Then they hit each other.’
b. Daba na-ngk=i nanaba kurrukurru.
fight 3sg-rr=pst there big.mob(nom)
‘Big mobs of people fought each other there.’
The encoding of reflexive/reciprocal constructions with a single pronoun or
pronominal affix is also found in two languages that neighbour Karrwa: the
closely-related Wanyi (Laughren 2001), 26 which likewise uses a single freereciprocal/reflexive
pronoun without other constructional coding, and Yanyuwa,
just to the north of Karrwa, which is related to Warluwarra (see example (12)
above), but which has developed bound pronominal prefixes, probably under
26. As in Karrwa, the determination of case on the reflexive/reciprocal pronoun in Wanyi is tricky,
and depends on finding examples with apposed nominals, but made trickier by the fact that
for nouns the ergative is signalled by vowel-length, which is rather difficult to hear in Wanyi.
Laughren (2001) claimed that subject nouns in reflexive/reciprocal constructions lacked the
vowel lengthening characteristic of ergative marking in regular transitive clauses, and that
therefore reflexive/reciprocal constructions involved case reassignment to their subjects, from
ergative to nominative. However, she has since rechecked her transcriptions of this data, as
well as obtaining examples with demonstratives where the ergative suffix (-ni) is clearer, and
her current position (e-mail to NE, 5 May 2004) is that subjects of reflexive/reciprocals retain
the ergative marking found in straight transitives.