Advances in Role and Reference Grammar

CLAUSE COMBINING IN NOOTKA^265

also on non-final verbs. We have seen an instance of this in Nootka (section
6.4) where, as generally, the consecutive (= succession) is unmarked vis-à-
vis the simultaneous (= overlap). In Swahili, beside the consecutive (also
called "subsecutive") marked ka- mentioned by Weimers (1973:365) and
Van Valin (1984:545), there is a form in ki- indicating simultaneity of the
action (Polomé) 1967:116-117, 164; Hinnebusch 1979:251-252); this form
may, however, also occur sequence-initially as a conditional. Lenakel has
sequential ep- and concurrent ak- markers, which will replace past im- or
perfect n- but may co-occur with future t- and continuous am- (Lynch
1983:211). Others of our cited cases (Vedic, Hebrew) seemingly lack this,
not having contrasting verb forms.
One thinks also of excluding the presence of an overt conjunction
from clause chaining, which would fit most of our progressive cases, such
as Nootka and Swahili, while excluding the Vedic and Hebrew construc­
tions in question; this is an unclear criterion, however, as illustrated by
the Lenakel case where a conjunction is optional or by the (Australian)
Pitjantjatjara case, where switch-reference is marked on the conjunction
itself (Austin 1980:26-27).
Thus the generalization of the concept of clause chaining to progressive
directionality, which is already implicit in the concept of cosubordination,
entails a weakening, but not the elimination, of several of its criteria: the
presence of a fully specified verb form at one end of the chain, marking for
switch-reference, marking for the consecutive/simultaneous opposition and
its extensions, and absence of overt conjunctions.^29 Longacre (1985:239)
also points out that for a language completely of the (regressive) chaining
variety, the subordinate/coordinate distinction becomes irrelevant, being
absorbed into the medial/final one. Many chaining languages do, neverthe­
less, make use of subordinate clauses, so that their presence cannot pre­
clude the recognition of clause chaining. An adequate assessment of differ­
ing degrees of commitment of languages to progressive clause chaining
must await the accumulation of text-based statistical information such as I
have tried to present for Nootka.

Notes

* Earlier versions of this paper were presented at a Clause Combining Workshop,

Rensselaerville, New York, November 2, 1985, and at a Role and Reference