This dissertation examines the notion of illocutionary force and whether it is embeddable by examining the syntax, semantics and pragmatic effects of a range of root-like embedded constructions. Though illocutionary force has long been considered a property exclusive to root clauses, the examination and analysis of English embedded inverted questions and other quasi-quotational constructions cross-linguistically show that this is not the case.
The contributions of this dissertation are three-fold: a refined definition of independent illocutionary force; a syntax for non-root complement clauses that carry independent illocutionary force; and a model for the discourse that captures the effects of these clauses. I also work towards understanding how the instantiation of independent illocutionary force in such constructions leads to their restricted distribution.
Illocutionary force may be represented both lexically and through syntactic processes such as verb movement. I argue that verb movement to Force° is an interface operation—it occurs in syntax but is directly linked to a specific discourse interpretation. Building on Krifka (2014), illocutionary force is the expression of who takes responsibility for asserting or responding to a proposition or set of propositions, according to a given modal base. When illocutionary force is independently expressed on an embedded clause, the perspective holder and responsibility taker(s) are unambiguous and not mixed. In contrast, standard embedded clauses may be ambiguous as to who takes responsibility and may contain multiple perspectives.
Clauses with independent illocutionary force have an expanded C-layer that is nonetheless smaller than that in true root clauses. An Illocutionary Act head selects for the embedded ForceP, determines illocutionary force and, obliquely, determines the restricted distribution of quasi-quotational constructions. A range of facts show that quasi-quotational constructions are truly embedded but are not direct objects of the matrix verb. Instead, they are in close apposition with a nominal direct object. This structure accounts for the properties of quasi-quotational constructions as entities that refer to a conversational move proffered in the relevant discourse, following Roberts's (1996, 2012) Question Under Discussion framework.
