A Minimalist Approach to Scrambling

Abstract

This study addresses the problems scrambling langauges provide for the existing syntactic theories by analyzing the interaction of semantic and discourse functional factors with syntactic properties of word order in this type of languages, and by discussing the implications of this interaction for Universal Grammar. Three interrelated goals are carefully followed in this work. The first is to analyze the syntactic structure of Persian, a language which exhibits free word order. With this analysis, the author has accounted for the relative order of categorized expressions, the motivation for their possible rearrangements, and the grammatical results of those reorderings. In this respect, a broad range of major syntactic phenomena, including object shift, Case, Extended Projection Principle (EPP), binding, and scope interpretation of quantifiers, interrogative phrases, adverbial phrases, and negative elements are examined. This monograph is the first major theoretical work ever published on Persian, and therefore fills the existing gap by providing insight into the syntactic structure of this language. The second goal is to connect these insights to similar linguistic properties in languages in which scrambling occurs (e.g. German, Dutch, Hindi, Russian, Japanese, and Korean), and to provide a deeper understanding of this group of genetically diverse, but typologically related languages. The final and principal goal is to situate the results of this work within the framework of the Minimalist Program (MP). The investigations in this study indicate that scrambling is not an optional rule, and that certain principles of MP, such as the Minimal Link Condition, are only seemingly violated in these languages. Furthermore, it is shown that careful analysis of scrambling with respect to binding and scope relations, and a reanalysis of the properties of A and A' movements, cast some doubts on the relevance of a typology of movement in natural language.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 126 citations exist upstream.