145 / 2018-09-28 09:38:38

Situational genre and legal branch variation in American legal language: investigating statistical interaction in corpus data

ESP,corpus,legal language,variation,multi-dimensional analysis

全体主题 > ESP与大数据

Big-data era introduces the data-intensive approaches and methodologies into linguistics, especially forensic linguistics, or legal linguistics, as one essential component of English for Specific Purposes. This study was corpus-based multi-dimensional analysis of describing linguistic variation in a corpus of American legal language across six registers, or two situational genres (statutes and trial transcripts) in three legal branches (civil, criminal and administrative branches). Such description is both quantitative, including the findings from the multi-dimensional analysis as the data foundation and a couple of ANOVAs and corresponding post hoc tests, and qualitative, including related interpretations that are based on situational differences among two situational genres and between three legal branches.
In this linguistic variation study, the Biber Tagger (named after the founder of MD analysis Douglas Biber) was used to grammatically each of the texts for every majority of the linguistic features, shown as sixty-four linguistic features. These features were analyzed as linguistic variables, which could be further reduced by factor analysis using the normed rates of co-occurrence for interpretable dimensions. Among all six registers, there were five underlying dimensions resulting in both factor analysis and dimension scores. Using factorial ANOVAs was to test significant interactions between situational genres and legal branches on each of the five linguistic dimensions. Four of five were discovered as statistical interactions, while situational genres and legal branches were interpreted for all five dimensions.
Since legal language has specific linguistic pattern as one macro register, different sub-genres and sub-patterns could be identified through several specific linguistic co-occurrence phenomena. This study highlights the importance of the linguistic variability in American legal language, using appropriate statistical techniques, in an effort to discover the underlying connection between different registers, which imply the potential representations and interpretations on ESP analysis.