Intentional "transitive" verbs and concealed complement clauses

A number of authors in recent semantics and philosophy of language literature have argued that the object of a propositional attitude is an interpreted logical form (ILF) - i.e. a syntactic representation (or "phrase- marker") combined with the semantic values that are assigned to various parts of that representation. One of the consequences of the ILF theory (in particular its formulation in Larson & Ludlow (1993)) is that the intension- Ality of various natural language constructions is argued to live off of an interpreted that-clause. This idea has strong consequences for constructions which are known to be intensional, but do not appear to have clausal constituents. Famous examples include: John wants a donkey John seeks a donkey. Facing a similar situation, Quine (1960) proposed that seeks might be reanalyzed, introducing a clause into the construction. The idea has not merely attracted philosophers, however. McCawley (1974), Karttunen (1976), and Ross (1976) all gave early arguments for a concealed complement clause to intensional verbs. One way of updating this idea is to suppose that there are "invisible" complement clauses containing phonologically unrealized verbal elements. For example, (i) might have the analysis in (i'): (i') John wants [PRO VERB2 a donkey]. In this paper we revisit that idea and offer some preliminary syntactic evidence in support of its plausibility in contemporary linguistic theory, drawing evidence for the implicit clause analysis from facts about passivization, and left-branch effects, and quantifier float.