Comparing syntactic persistence in written and spoken monologue

Abstract

Syntactic persistence, the tendency for speakers to repeat recently-used syntactic structures, has been well demonstrated in dialogue and in single-sentence monologue primed by reading aloud pre-prepared material. Models advanced to explain syntactic persistence assume that priming will also occur in extended monologue, but there is no clear evidence that this is so.

This thesis examines within-speaker syntactic persistence of the genitive alternation in spoken and written monologue from the QuakeBox corpus and the Press database, two New Zealand corpora selected for their close match of time period, geographic location, and topic. Two research questions are considered: is priming present in extended monologue, and does priming differ between speech and writing?

In order to address these questions, I use binomial mixed-effect models to find the relative contribution of factors predicted to affect genitive choice and priming, and compare the relative impact of these factors, and the overall effect of priming, on the two corpora.

The findings of my research indicate that syntactic priming is present in extended monologue, and that this priming occurs more frequently in speech than in writing. My results also support observations in the existing literature that genitive choice is affected by animacy, the presence of a sibilant sound, and the semantic relationship between possessor and possessum.

While this study was not able to offer conclusive insights into the differences between α- and β-priming, and the issue of priming in nested structures, my findings indicate that these would be promising areas for further research.