The present thesis focusses on the structure of PP's in English and Polish, and the occurrence of PP's in subject and object positions. The main theoretical references are the X-bar Theory of syntactic categories and the Government Binding framework (GB). A consideration of English data corroborates Jackendoff's and Emonds' claim that apart from an NP, prepositions can take a PP and an S′ complement or no complement at all, though details of Jackendoff's analysis are revised. Polish prepositions allow the same range of complements, including no complement, although,with a greater variety of complex prepositions and with intransitive prepositions modified by relative and appositive clauses, the P-PP and the P-S′ structures are less common in Polish than in English. Subject and object PP's have so far received little attention. Like PP objects of prepositions, they are used if the intended meaning cannot be expressed by a suitable NP. The appearance of subject PP's in raising and passive sentences poses a problem for classical Transformational Grammar, though not for a slightly revised version of GB – another category-based framework. The analysis proposed here involves a particular view of the representation of Case, and a revised Case Filter. The Case Filter rules out not merely any lexical NP with no Case but any lexical XP which requires Case but has not been assigned Case. Thus, the properties of being an NP and requiring Case are independent of each other. It emerges from the investigation (i) that prepositions in English and Polish are more alike than one might expect, given the obvious differences between the two languages; (ii) that prepositions and PP's are like verbs and VP's – as Jackendoff emphasizes – but in some respects they show greater resemblance to other categories; and (iii) that syntactic categories are less important for the distribution of phrases than is commonly assumed, and that the meaning of phrases is of central importance for their distribution.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Linguistics and Philosophy, 2007. / Includes bibliographical references (p. 203-212). / This dissertation develops a form of relativism in which propositions are treated as sets of world-time-individual triples, in contrast to standard views that treat them as sets of worlds or world-time pairs. This builds on existing proposals for predicates of personal taste such as fun and tasty, and has ties to approaches to de se attitudes involving centered worlds. I develop an accompanying pragmatic view in which the context set is similarly construed as a set of world-time-individual triples. The semantic and pragmatic systems together are used to account for the behavior of predicates of personal taste, epistemic modals, indicative conditionals, and a variety of attitude reports, including control constructions. I also explore ways that this account can help solve puzzles related to Moore's paradox. To give one concrete example, I propose that the proposition expressed by the sentence it might be raining is the set of world-time-individual triples <w,t,x> such that it's compatible with x's knowledge in w at t that it's raining. On the pragmatic side, a speaker is justified in asserting this sentence in a conversation if it is compatible with the speaker's own knowledge that it's raining; by asserting it, though, the speaker is making the stronger proposal to make it common ground that it is compatible with the knowledge of the entire group of conversational participants that it's raining. If this proposal is accepted by the other participants, then the group will have established that their knowledge states are aligned in a particular way. I introduce the core semantic and pragmatic proposals in Chapter 2, focusing on epistemic modals, predicates of personal taste, and belief reports. / (cont.) In Chapter 3, I extend the analysis to indicative conditionals, showing that this solves longstanding puzzles involving the relationship between conditionals and disjunction. In Chapter 4, I extend the approach to certain control constructions, with a special emphasis on capturing their de se interpretation. In Chapter 5, I look at two puzzles related to Moore's paradox, with special attention to the meaning of imagine. / by Tamina C. Stephenson. / Ph.D.