PREPOSITION ASSIGNMENT IN ENGLISH 2992.6.2.2 Potential agents which Φ A
Section 1 identified the argument marked by with in a context such as (59)
as a co-performer of the action.
(59) a. John went to the party with Mary.
b. [DO (John/Mary, [go' (John/Mary)])] CAUSE [BECOME
be-at' (party, John/Mary)]
John is the agent/Actor, Mary is the non-actor agent, as marked by with,
and both John and Mary are arguments of go' and be-at'. This analysis pre
supposes an identical LS for sentence (60).
(60) John and Mary went to the party.
Here, in contrast to (59), John and Mary, as co-agents, are both posited as
Actors, since neither outranks the other for Actorhood. However, this
analysis of potential agents marked by with is less than satisfactory. Native
speakers interpret the volitional control of the event described in (59) dif
ferently from the volitional control in (60). While the scope of volitional
DO is clearly over both John and Mary in (60), native speakers differ in
their interpretations of (59).^15 Three interpretations are theoretically possi
ble, as specified in (59b, c, d).
(59) ?[DO (John, [go' (John/Mary)])] CAUSE [BECOME be-at'
(party, John/Mary)]
d.? [DO (Mary, [go' (John/Mary)])] CAUSE [BECOME be-at'
(party, John/Mary)]
Native speaker judgments of who is in control of the action in (59a) are
evenly split between John and Mary. If volitional DO is an operator over
Mary, rather than John, as specified in (59d), the situation is even more
problematic; since Mary is clearly the agent in this LS and, therefore, the
actor, Mary, should occur in the unmarked slot for actor, as syntactic sub
ject. One possible solution, suggested by Van Valin (personal communica
tion) is that it may be the end-of-clause, foregrounded position of the pre
positional phrase, with Mary, which suggests that she has volitional control
of the event. In other words, pragmatic inferences, having nothing to do
with role structure, may inform judgments about salience relations in the
clause.^16 If this is the case, the semantic content of with is not at issue in the
alternative interpretations. The kind of "conjunct splitting" illustrated in
(59a) is limited to volitional activity or accomplishment contexts (i.e., verbs
having a DO operator in LS), as illustrated by the contrasting "volitional