Advances in Role and Reference Grammar

(singke) #1

174 L. MICHELLE CUTRER


This alternation of actor and undergoer control, with verbs such as ask, is
explained in RRG by the change in the semantics of the verb, which, in
turn, changes the controller. In the first sentence, ask is a directive speech-
act verb. John tries to get Bill to do something by asking. The theory pre­
dicts that directives have undergoer control. However, in the second sen­
tence, ask is not a directive. John is not trying to get Bill to perform the
action encoded in the lower clause, but rather, John is asking a question.
As ask is not a directive in this instance, the theory correctly predicts actor
control.
This analysis of ask will be modified in my analysis. As we will see, the
change in verb semantics does not account for all cases. In fact, indirect
questions may involve arbitrary control, as in (11).
(11) John asked how to solve the problem using the theorem.
The controller is an arbitrary group or person, the "how to" being the same
for any solver of the problem. However, context or real world knowledge
might suggest an arbitrary group plus John, or John alone.
A similar situation, in which a single verb exhibits both subject and
object control, occurs with the complex verb plead with:
(12) a. John pleaded with Bill to leave.
b. John pleaded with Bill to be allowed to leave.
In the first example, John tries to get Bill to do something with his pleas. In
this example, plead with is a directive speech-act verb, and therefore under­
goer control is assigned. In the second example, John is making a request of
Bill, that Bill give him permission to leave. In this case, plead with is not a
directive speech-act verb. As the verb is not a directive, actor control is pre­
dicted. Note that the contents of the linked core affect the meaning attri­
buted to the matrix verb.
To sum up, FVV (1984) claims that obligatory control relations follow
from the semantics of the verb; causative verbs and directive speech-act
verbs are predicted to have undergoer control. Other verbs will have actor
control. Predictions are made on the basis of the semantics of the matrix
verb. However, the FVV analysis deals only with a limited subset of infinit­
ival complements. In the remaining sections of this paper, I will present an
analysis of a broader range of control phenomena, which follows the basic
intuition that control is semantically determined. I will appeal not only to
verb semantics, but also to clause linkage conditions and pragmatic factors
in my characterization of control phenomena.

Free download pdf