Elements of control
Massachusetts Institute of Technology. Dept. of Linguistics and Philosophy.
Noam Chomsky, Sabine Iatridou and David Pesetsky.
MetadataShow full item record
How many different types of control relations exist? Is the classical distinction between Obligatory Control (OC) and Non-Obligatory Control (NOC) well-founded? What semantic and syntactic properties of infinitives determine their place in the control typology? How is the "understood subject" PRO linked to the controller? This thesis investigates these questions in two steps: First, we establish a typology of control and characterize the empirical profile of each type; second, we propose mechanisms of derivation and interpretation to account for the different types . The OC category is shown to consist of two subtypes, Exhaustive Control (EC) and Partial Control (PC). Tense in EC complements is null, and PRO must be referentially identical to the controller; Tense in PC complements is contentful, and PRO need only include the controller (although matching in syntactic number is still required). OC establishes an Agree relation between a matrix functional head and either PRO (in EC) or the infinitival Agr (in PC). The latter is parasitic on T-to-C movement occurring in tensed complements. Control via Agr blocks the transmission of semantic number from the controller to PRO, giving rise to the PC effect. The OC/NOC distinction is traced to the position of the infinitive: VP-internal clauses fall under OC, extraposed and intraposed clauses under NOC - a corollary of the CED, which constrains Agree. Extraposed clauses may be interpreted and pronounced in different positions - a claim that is supported by asymmetries between psych and non-psych predicates in Super-Equi constructions. Converging evidence from extraction confirms that infinitives displaying OC and those displaying NOC occupy different positions at LF although the same position at PF. As for the interpretation of OC, the choice of controller is subject to complex semantic/pragmatic considerations, rather than some syntactic locality principle. We also argue that OC cannot be reduced to predication, at least in the domain of adjectival complementation. Systematic contrasts between subject-gap and object-gap infinitives show that the former may denote either propositions (when occurring as arguments) or predicates (when occurring as modifiers). Thus, two sources exist for subject gaps - PRO (universally available) or A-bar trace (language-particular).
Thesis (Ph.D.)--Massachusetts Institute of Technology, Dept. of Linguistics and Philosophy, 1999.Includes bibliographical references (leaves 234-240).
DepartmentMassachusetts Institute of Technology. Dept. of Linguistics and Philosophy.
Massachusetts Institute of Technology
Linguistics and Philosophy.