The Significance of Background Information in Acceptability Judgements of Korean Sentences

摘要

Linguists in the Generative Linguistics tradition typically rely on their own intuition as a native speaker for their core data. This practice has continually been challenged to be more rigorous in their data collection methodology, and Experimental Syntax is an effort to address the issue. The purpose of this presentation is to discuss whether some background information of the subjects might affect the acceptability judgments of the subjects, mostly naive native speakers. It is well established, following the sociolinguistic research, that native speakers' use of language might be influenced by some social factors like age, gender, class, and others. The data to be discussed in the presentation are drawn from acceptability judgments on sentences by Korean native speakers, and comprise 68,158 data points, gathered from 302 native speakers with 574 stimulus sentences (around 240 items per subject). The background information that was collected includes factors like sex, age, dialect, and previous exposure to linguistics classes ('familiarity'). We discuss in detail some statistical issues to be dealt with for any proper treatment of the data, focusing on their distributional characteristics. The results show that the 'familiarity' factor largely plays a marginally significant role in the distribution of the data.