Morpho-syntactic Correlations In Ergative Languages A Frequency-based Approach

This dissertation is a typological study of the morphology-syntax correlations in ergative languages, which is based on a frequency approach to a sample of 80 languages. The basic framework of the thesis is of WALS, and the tool for data processing is R. It has eight chapters, which could be grouped into four sections.INTRODUCTION:Chapter 1. Why bothers another thesis on the typology of ergativity? This introductive chapter briefly discusses some theoretical importance of doing morpho-syntactic correlations of ergative languages. A short history of ergativity study in typology is also found there, including a review of Dixon's monograph (1994).SAMPLE AND DATA ANALYSIS:Chapters 2-5.Chapter 2, Sample and data. In this chapter, a volume of ergativity terms are discussed, mainly by introducing the terms of WALS. A sample of 80 ERGATIVE languages, which includes those ergative/active/split case marking languages and ergative/active verbal person marking languages, is also introduced in this chapter. Among these 80 languages,77 are recorded ERGATIVE in WALS, while three others are not, i.e. Dyirbal, Kham, and Tibetan.Chapter 3, Clause structure. Four topics are covered in this chapter, namely, (i) case marking and alignment (which differs from WALS, in which'alignment' covers both case marking in noun/pronoun and verbal person marking), (ii) ditransitive constructions, (iii) passive and antipassive constructions, and (iv) interrogatives. Some universals, or, more commonly, tendencies, are present therein. For instance, the canonical marking strategy in an ergative language is to have ergative case marking on nominals (rather than pronominals, which prefers accusative marking), and have verb to agree with the absolutive constituent. I also propose that there is no close relation between passive/antipassive construction and ergativity.Chapter 4, Word order. It tries to dispel a myth in the beginning of the chapter, i.e. an ergative language is'free word order language'or'no basic word order language', mainly by a survey of previous studies. It followed by an analysis of the basic clause order, order of the modifiers of noun phrase, and the position of adposition to noun phrase. I found that the most common clause order in ergative language is SOV, while SVO is witnessed a remarkably low ratio (8.75%) in the sample. Postpositions, GN, DN, NumN, NAdj, and NR sequencings are also favored in ergative languages.Chapter 5, Morphology. It covers two topics of morphology, i.e. head/dependent marking (or rather,'locus of marking') and inflectional morphology. It is proposed in this chapter that head marking (in clause, possessive noun phrase, and whole language typology) and suffixes are favored in ergative languages.RELATED ISSUES:Chapters 6-7.Chapter 6, Some universals of ergativity in The Universals Archive (UA) revisited. In this chapter, some thirty universals of ergativity in UA are examined in my 80-languages-sample. It is found, however, some of the universals may not hold.Chapter 7, A comparison of ergative languages and languages in general. It covers three issues, those word order universals present by Greenberg (1966), those universals of interrogatives present by Ultan (1978), with also a comparison of 23 parameters in WALS. Some of Greenberg and Ultan's universals are found to be not hold in the 80-languages-sample, while the others are proved to be valid.CONCLUSION:Chapter 8. To sum up the thesis,72 universals (or, tendencies) of morpho-syntactic correlations of ergative languages are presented there (the universals in chapter 6 and 7 are not included).