(^406) MARY L. NUNES
direct argument, are two-argument ACT-based nominals able to override
the default ACM interpretation and retain an activity reading in which the
direct argument is interpretable as the CL-A. This would explain why most
of the two-argument ACT vNs/vN-ings in the data base (as well as the non-
result ACM vN(-ing)s used iteratively as ACTs) are unable to solicit an
activity CL-A interpretation from their direct arguments.^25
The strongest of the numerous implications which emerge from this
analysis of vN direct-argument linking in activity-based vNPs is that ACT
nominalizations occur relatively infrequently in English. In light of the con
flict between the V-like dynamic nature of ACT vNs and the static nature of
prototypical Ns, this does not seem implausible. Nonetheless, numerous
questions which time and space have prevented from being addressed in
this study need to be answered before the behavior of activity-verb
nominalizations can be analyzed in satisfactory detail. (For example, what
is the role played by contrasting action nominals in nominals referring to
unbounded states of affairs?) The analysis presented here should therefore
be seen as a first step toward understanding the complex issues involved in
the direct-argument selection of vNs derived from ACT verbs susceptible to
ACM interpretations. This first step may be summarized as follows.
The RRG prediction that activity vNs will select CL-As as direct argu
ments is verified with single-argument vNs (and plural result vNs derived
from accomplishments used iteratively as activities). The fact that the pre
diction is only sporadically borne out with two-argument ACT nominals is
attributed to the effects in the vNP of the susceptibility of ACT verbs to
ACM interpretations. The exaggerated susceptibility of the English vNP to
a phenomenon which is merely contextually-potential where not overtly sig
naled in the clause is, it is argued, brought about by the combined effects of
- the syntactically "intransitive" (i.e. single-direct-argument) status of the
vN, 2) the conflicting macrorole choices of ACT vNs (select CL-As) vs.
ACM vNs (select CL-Us) for the linking of arguments to that single direct-
argument position, and 3) the prototypically static nature of nouns as
"things." This last factor, it has been suggested, encourages the default
interpretation of the macrorole associated with the vN's single direct argu
ment to fall to the ACM class of V sources, which, in naming bounded
events, are more easily reified than are the dynamic states of affairs named
by ACTs. Thus, in general, only where a y argument is free to be preposi-
tionally marked as an indirect argument of the vN are two-argument ACT-
based vNPs able to override the default ACM interpretation and mark the
CL-A as an activity direct argument.