Text

theses

A crucial step in determining the meaning of a sentence is identifying the grammatical roles of the phrases and mapping thematic roles onto them. There are (at least) five types of cues that could potentially facilitate sentence processing: propositional content, discourse context, overt case-marking, word order, and prosody. We investigated whether the way people produce spoken utterances depends on the consistency, reliability and robustness of these cues. To date, most research on spoken language production and processing has been done on languages like English that have relatively strict word order and impoverished morphology with little research on languages with flexible word order and rich morphology.

The research presented in this thesis addresses this gap by investigating the production of spoken Turkish, a language with flexible word order and rich morphology. In Turkish, sentences that have scrambled word order or lack overt object case-marking are temporarily ambiguous (i.e., garden-path sentences). We had nine Turkish speakers read aloud SOV sentences (i.e., sentences with default word order) and OVS sentences (i.e., sentences with scrambled word order) that did or did not have overt object case-marking. We found that there were prosodic differences between casemarked and non-casemarked sentences and between scrambled and non-scrambled sentences. These findings suggest that Turkish speakers prosodically mark grammatical roles when morphosyntactic cues like word order and case-marking are absent. We discuss some possible linguistic, psycholinguistic and information theoretic reasons for the observed prosodic differences, and outline future studies that could distinguish among these accounts.

ETD_11224

M.S.

Includes bibliographical references

ETD graduate

doi:10.7282/t3-7tzw-8h65

The author owns the copyright to this work.