A Minimalist Study Of Lexical Assembling And Merge

This dissertation is a study of lexical assembling and merge in the field of generative grammar. It is more than half a century since Noam Chomsky pioneered the new direction in the study of language by combining linguistics and the brain sciences. Insofar there have been extensive studies of how the Faculty of Language is designed and how it operates to generate derivations. With the major framework drawn from the current minimalist documents, and some inspiring ideas from the related studies achieved in psycholinguistics and semantics, the main goal of this dissertation is to focus on lexical assembling and merge in the minimalist sense.The hypothesized Language Faculty in generative grammar consists of two major components: the lexicon and the computational system. As the source of discrete building material for the computational system, it is a natural presupposition that lexical items are stored in some way in the lexicon. In this dissertation, lexical items are decomposed into phonetic, semantic and formal features, which are hypothesized to be stored in different rooms in the lexicon. Then the semantic and formal features of lexical items are identified as fuzzy in nature. In order to clarify the paradox that lexical items are fuzzy in the lexicon and accurate in the computational system, the system of thought from the performance system is proposed to provide directions in the process of assembling lexical items.An elementary fact about the language faculty is that it is a system of discrete infinity, thus merge is inevitably needed to produce hierarchically structured expressions. The idea that all kinds of merge are free in operation does not contradict with the thought that their operations should be motivated or triggered in some way. By extending the selector mechanism from set merge to pair merge, this dissertation finds a theoretical correspondence between the probe system and selector system, which helps to unify the mechanism of all kinds of merge. Semantic features should not be accessed in the narrow syntax. In order to justify selectors' computation in external merge, this dissertation proposes to extend language computation from the system of narrow syntax to both semantic system and phonetic system.On the basis of the general proposals made in Chapter Two and Chapter Three, this dissertation shifts to the systematic description of how substantive categories and functional categories are assembled and merged in Chapter Four and Five. Semantic selectors of substantive categories form the main mechanisms for external merge and probes of functional categories are the keys to trigger internal merge. External merge between substantive categories are reclassified according to whether the semantic relation behind the merge is theta-related or not. The most obvious modification made in this dissertation is about the set merge of adjectives and prepositions. The adverbs generated in their specifier positions are reclassified as pair merge. The formal features of functional categories are also explored in this part. On the basis of person and number agreement between T and its nominal goal, this dissertation hypothesizes the v-V agreement on argument features and C-T agreement on tense features.A conclusion can be safely drawn from the above discussion that substantive categories and functional categories possess no sharp distinctions. On the one hand, substantive categories are never short of formal features in narrow syntactic computation; on the other hand, abstract semantic features of functional categories are also explored in this dissertation to account for the external merge they get involved in. Thus from a theoretical point of view, functional items and substantive items do not form sharp contrast in terms of assembling and merge in language faculty.