SEMANTIC AND SYNTACTIC FACTORS IN CONTROL 169
Distance Principle in terms of grammatical relations. For a detailed critique
of these approaches, see Cutrer (1987).
Semantic accounts of the control problem are of several different
types. The analyses of Jackendoff (1972), and Nishigauchi (1984) predict
the controller on the basis of thematic relations. Rûzicka (1983) treats the
property of subject and object control as a semantic feature. Williams
(1980) and Cullicover and Wilkins (1986) base their analyses of control
relations on a theory of predication. Comrie (1984) and Foley and Van
Valin (1984) predict control relations from the semantics of the matrix
verb. Comrie, however, includes pragmatic or real world knowledge in his
analysis.
Although theorists who approach the control problem from a syntactic
standpoint continue to refine their predictions about the occurrence of and
constraints upon empty categories, it is clear from the variety of construc
tions which a control theory is asked to account for that any syntactic pre
diction of control relations will never be more than a general rule, with
numerous lexically marked exceptions. A semantic approach may account
for the facts in a more general manner without further need to mark excep
tions in the lexicon.
It will be argued that an analysis where control is predicted from the
meaning of the complement-taking predicate is preferable to an analysis
based on general syntactic principles, with lexically marked exceptions. An
analysis which relies on either simply listing the control properties of verbs
or marking exceptions in the lexicon is unsatisfactory. Radford (1981:381)
puts the point well.
Firstly, arbitrary lists of properties associated with predicates have no pre
dictive or explanatory value: ask the question "How do you know this is a
verb of subject control?", and you get the non-answer "Because it's listed
as a verb of subject control in the Lexicon." Secondly, treating control as
a lexically governed phenomena implies that control properties are entirely
arbitrary, and hence will vary in random fashion from dialect to dialect, or
language to language: this would lead us to expect that the counterpart of:
John persuaded Bill to leave.
in some other dialect or language would have subject control rather than
nonsubject control, i.e. would have PRO interpreted as referring to John
rather than to Bill. But as far as we know, this is not the case. And we
would hope that an adequate Theory of Control should explain why this is
not the case It seems likely that the question of whether a control pred-
