| Summary: |
This study is concerned with the examination of word-order universals and the attempt to explain the diverse statistical patterns by means of a variety of semantic structures. Central to my study is a critique (a) of Hawkins's approach, the Universal Consistency Hypothesis, and its concentration upon apparently exceptionless universals, and (b) the Bartsch-Vennemann approach which breaks down structures generally into OPerator-Operand pairs. In Part I, I introduce a geographically and genetically representative sample of 75 languages, and show that not only are violations to be found to Hawkins's principles of consistency, but that they are statistically predictable. The statistical results are also useful in relating word-order patterns with fusional, agglutinative and isolating morphology: in particular fusion is associated with violations of statistical universals. In Part II, I concentrate on adjectives and verbs and here the patterns of non-exceptionless universals become important: instead of distinguishing intensional and extensional adjectives after Montague, I identify a semantic principle of Adjunction, which includes some intensional adjectives, over and above those adjectives for which an analysis as (intensional) operators is still appropriate. The analysis further suggests that the Bartsch-Vennemann identification of operators and operands is often the reverse of what it should be, especially for genitival and verb-object constructions, and I discuss the implications of this for the verb and its semantic relationship with the subject and object. Finally I consider determiners, tense and focus as operators upon operators, showing why their syntactic behaviour leads to asymmetric statistical patterns.
|
