| The three experiments in human sentence processing reported here were conducted in the Algonquian language of Odawa. They represent the first psycholinguistic experiments to be conducted in an indigenous North American language. This dissertation argues that Odawa is not in fact a "nonconfigurational" language by adopting and providing supporting Odawa evidence for a configurational account of Algonquian syntax; however, the syntactic properties of the language include several phenomena not found in those languages in which psycholinguists have traditionally focused their work. These phenomena include a verbal "direction" system (direct/inverse), an obviation system (proximate/obviative), very free word order, frequent pro-drop of all NP arguments, and the lack of an indigenous writing system.; The broad goals of this research were (a) to determine the extent to which psycholinguistic investigation of "exotic" languages like Odawa could inform theories of sentence processing, (b) the extent to which psycholinguistic investigation might illuminate issues in a language's syntax, and (c) the extent to which psycholinguistic investigation is feasible in an endangered, geographically isolated language.; Experiment 1 investigated sentence production in Odawa. Its goals were two-fold. First, since Odawa is not a written language, this production experiment was designed to elicit some basic frequency statistics with respect to verb form and word order. Second, the experimental design allowed for an investigation of the effect of conceptual accessibility on word order, topicality, and syntax. The results suggested that conceptual accessibility affects the syntactic planning of utterances. Experiment 2 investigated sentence comprehension. It was designed to elicit both on-line measures and an off-line measure in order to assess the effects of word order, verb form, and the distance between thematic assigners and assignees on sentence comprehension and interpretation. The results were broadly consistent with a configurational account of Odawa syntax. Furthermore, sentences in which thematic assigners and assignees were in close proximity were found in most word orders to be more difficult to comprehend than sentences in which they were separated by one intervening adverb, but not much more difficult than sentences in which they were separated by two adverbs. This pattern of results was interpreted as evidence for a "good enough" processing strategy based on a linear heuristic, which allowed participants to derive relatively accurate interpretations without undertaking complex syntactic revision. Experiment 3 investigated the interpretation of pros. It was designed to investigate how discourse factors and animacy interact when participants were asked to comprehend sentences lacking external context. The results suggested that when context is lacking, Odawa speakers posit features for null NPs based on the features of the overt NP, as a sort of sentence-internal context. (Abstract shortened by UMI.)... |
