Discovering classes of attitude verbs using subcategorization frame distributions

In his classic squib on manner-of-speech predicates, Zwicky (1971) poses a cluster of questions foundational to the syntax-semantics literature:To what extent is it possible to predict certain properties of words (syntactic, semantic, or phonological), given others?[And] insofar as there are such dependencies among properties, what general principles explain them? Put another way: what sorts of word classes are there, and why these and not others?