MARY L. NUNEScisely because it does not permit an ACM performance-object interpretation.
ACTs which do permit such an interpretation are represented in Table 1 with
Type performance objects, and are dealt with in the discussion of Table 2 per
formance-object nominals.
U's vN [of A]dir.arg is not included in Table 1, as that type of construction is not
possible in English. Although ACTs capable of eleciting a pure ACT interpreta
tion take an A direct argument, pure ACTs have no U to be preposed (cf.
2.3.2.4). Thus, for example, the locative or theme arguments of ACTs receiving a
strict ACT interpretation are not linked to a macrorole and cannot occur prenom-
inally: the murder's investigation [of SherlockCL.A].
The only other case in which the A is linked to the direct argument position
in the vN data is with the exceptional two-argument ACM nominalizations of
enter. As with the Type D nominals in Table 1, however, the locative CL-U argu
ment cannot be preposed (except where entry and entrance are used in their "reg
ular noun" sense such that they refer to an attribute possessed by the possessor
NP (cf. 1.1 and 2.3.1): the arena's entrance, the house's entry; but the arena's
entrance of/by the toreador, *the house's entry of/by the robbers.) This reflects the
fact that locative vN arguments are permitted to occur prenominally only if they
are linked to the A.
In terms of RRG's positionally-determined definitions of thematic relations, both
Type A EXPs and Type PATs are patients. Both are the y-argument partici
pants in the state or condition named in the BECOME [state-predicate' (y)] por
tion of the particular ACM LSs in which they occur. Thus, the thematic relation
of PAT may be used to identify both patients and affected experiencers. There
are, however, important differences between the two. For example, the experien
tial states in which EXPs are situated do not include an end point beyond which
the state named by the vN can continue. Rather, with experiential states like
amusement and excitement, the state named by the vN continues only for as long
as the entity or activity creating/inducing the state continues to create/induce the
state in the entity experiencing the state. Thus, for example, the experiential state
of amusement lasts only as long as the entity experiencing the amusement con
tinues to be amused by the entity or activity creating/inducing the amusement. By
definition, then, the participant situated in an experiential state must be animate
and is almost always human.
Result states like destruction and reclamation, on the other hand, do include
an end point beyond which the state named by the vN continues for an indefinite
period of time. The result state of destruction, for example, continues indefinitely
after the entity situated in the result state has ceased being destroyed. Moreover,
whether or not the entity situated in that state is capable of perceptually
experiencing the state is irrelevant — i.e. the affected participant need not be ani
mate. Because these distinctions can help to clarify certain aspects of argument
preposing, the patients of experiential states, which are necessarily experiencers,
will be distinguished from the patients of result states in the ensuing discussion.