Clausal Complementation as a Compatibility Relation

It is commonly assumed that verbs that take clausal complements select the complement clause types that they can co-occur with. I am going to challenge that assumption. Some sorts of clause types (different ones in different languages) are rigid in their meanings, which is to say that the sorts of meanings that they introduce, e.g., deontic modality or conditionality, do not vary across the predicates they appear with. It could thus be said that these clausal types select the predicates that they can appear with, which reverses what is usually thought about the predicate/complement relation. I will demonstrate how certain kinds of clause types have rigid meanings (drawing from English, Medumba and Lubukusu) and I will propose that in addition to semantic compatibility, there are also binding relations that must hold between predicates and the clausal complements they co-occur with, based in part on locality effects and in part on exponents of agreement. The strongest position might be that every clause type is rigid, so that so-called semantic selection by a predicate is always a compatibility relation, in part a mutual selection. I will outline what challenges such a theory would face.

As part of the challenge to show the semantic force of complementation on verb meaning, I will explore the role of predicate selection by complements in languages where the meaning of clausal complement verbs shift according to the clausal complement they co-occur with. A rather novel sort of subordinate clause type found in Lubukusu introduces an actuality entailment, which can shift the meaning of a potentiality predicate to render it, in effect, an implicative. Also in Lubukusu, the verb best translated as “think” and the one best translated as “believe” both can mean “hope” when they take a subjunctive complement. In Medumba, a Grassfields Bantu language, there are very few propositional attitude verbs, so the variety of propositional attitudes that are lexicalized in English are composed in Medumba using a variety of semantically rigid clause types. This raises questions about what is semantically encoded as the lexical meaning of a predicate in those contexts where its meaning can shift according to what it is composed with. For example, when your favorite philosopher or semanticist proposes a semantics for “believe”, how Eurocentric is their point of departure?

Ken Safir is a Distinguished Professor of Linguistics at Rutgers University where he has taught for 35 years and where he founded the Department of Linguistics. Throughout his career, has been a frequent contributor to the literature in syntactic theory, the syntax-semantics interface, theoretically informed comparative syntax and morphosyntax. A great deal of his work has explored the syntactic representation of bound variables and patterns of anaphora, but in recent years, he has turned increasingly to the distribution of clausal complementation from a crosslinguistic perspective. He has written on many sorts of languages, but most of his work has been on Indo-Eurpoean and Niger-Congo languages. For the last 17 years he has been the director of the NSF-supported Afranaph Project, which explores the languages of Africa in collaboration with native-speaker linguist consultants. Prof. Safir is frequently invited to address forums on online data access, as all the data collected by Afranaph from the 40+ languages it covers is freely shared as part of an open access database that is used worldwide.