Agreement Configurations: Grammatical Relations in Modular Grammar

M. J. Alexander, 1990

for $29.95 x

This thesis presents an analysis of the relationship between grammatical agreement and so-called "non-configurationality." We argue that the properties of non-configurational languages should be made to follow from independently motivated principles of agreement. We explore the contrast, discussed in Hale (1989), between languages such as Navaho (Athapaskan) which permit co-occurrence of full noun phrases and agreement inflection and languages which show complementary distribution between these categories, such as Dogrib (Athapaskan) and Irish (Celtic). We argue that the study of these contrasts supports the analysis of non-configurationality presented in Jelinek (1984) where agreement morphemes have argumental, theta-marked status, in non-configurational languages.

We first introduce issues in the study of grammatical configurations such as word order as induced by Case and thematic-role assignment (Chapter 1). In Chapter 2, we review the literature on the issue of non-configurationality, and in Chapter 3, we follow Alexander (1986) in arguing that no bivalent (two-valued) configurationality parameter can account for observed configurational variance and we propose a four-valued opposition—the Case/Agreement distinction—which obviates the need for a separate configurationality parameter. In Chapter 4, we discuss the proper analysis of languages which show complementary distribution between nominal arguments (NPs) and agreement morphemes, arguing that the analysis in Hale (1989) cannot be supported and, further, that this analysis conflicts with the analyses of configurationality proposed in Jelinek (1984) and Hale (1984). We argue that features of theta-theory provide a superior analysis of these languages. In Chapter 5, we discuss two languages, Canela-Krahô and Hixkaryana, which cannot be subsumed under Hale’s analysis.

Thesis Supervisor:      Dr. Kenneth L. Hale, Ferrari P. Ward Professor of Linguistics



1 Introduction 1
1.1 Where does word order come from 1
  1.1.1 Some theoretical background 1
  1.1.2 Word order in this work 23
1.2 Another question: how does agreement work 36
  1.2.1 Some background 36
  1.2.2 The outline of this work 41
1.3 Theoretical assumptions 53
  1.3.1 D-structure representations 63 The lexicon 63 X-bar theory 70 Theta theory 76 Case theory 79
  1.3.2 D-structure to surface—Move α 81
2 Approaches to configurationality 85
2.1 “On the position of Warlpiri in a typology of the base” 85
2.2 “Preliminary remarks on configurationality” 95
2.3 “Warlpiri and the grammar of non-configurational languages” 109
2.4 “Empty categories, case and configurationality” 122
2.5 Two alternative conceptions 131
  2.5.1 “The configurationality of Slave” 131
  2.5.2 The saturation parameter: Speas, 1986 142
2.6 Summary of the analyses considered 148
3 A non-bivalent theory 150
3.1 On the goals of Case/agreement 154
3.2 Licensing theory and multiple licensers 175
3.3 Agreement languages: a sub-class 187
4 Treating the class of middle cases 204
4.1 A conflict between theories 204
4.2 Implications of inflection across categories 224
4.3 Inflection across categories 231
  4.3.1 Argument languages 231
  4.3.2 Adargument languages 1: Athapaskan 237
  4.3.3 Adargument languages 2: Hixkaryana 240
4.4 Hale’s analysis and the facts of Yagua 241
  4.4.1 Positional complementary distribution 243
  4.4.2 Implications of Yagua 248
4.5 A revised incorporation analysis 250
4.6 A Hale-type analysis of Yagua 275
4.7 Two alternative views 279
4.8 Multiple nominals in Italian and Hopi 283
4.9 Conclusions 292
5 Incorporation and locality 295
5.1 Path containment in Hixkaryana 295
5.2 Subject complementary distribution in Canela-Krahô 306
5.3 Conclusions: on non-configurationality 319