Multivariate exploration of instantial variation in situational context: The powerful role of the individual instance of language use

Stella Neumann

Abstract

Register variation is usually discussed in terms of variation between groups of texts as instances of a given register, i.e. as between-group variation. This suggests that different contexts represent clear-cut categories, thus leaving open the question of how to conceptualise texts representing borderline instances and instances that show traits from two categories. In this paper, I argue that viewing registers as categories obfuscates continuities between textual instances across categories reflected in similar distributions of linguistic features. To demonstrate the effect, I will visually explore results of the quantitative multivariate analysis of data from the International Corpus of English. The representation of every instance in visualisations of the results yielded by multivariate analysis reveals widespread continuities and a substantial amount of overlap between register-based groups of texts. This ‘fuzzy’ character of registers is understated by empirical studies of register variation, which average out textual instances on the basis of register labels. The paper shows that systemic functional linguistics has a powerful conceptual toolkit to capture the fuzziness of register variation and that it can be explored with the help of Geometric Multivariate Analysis.