Sociolinguistic variation in English derivational productivity : Studies and methods in diachronic corpus linguistics

Abstract

This dissertation studies how the productivity of word-formation varies across social groups in the history of the English language. Previous research into variation and change within the morphological productivity of derivational affixes has been hampered by the lack of suitable methods for comparing productivity measures across subcorpora. A further problem has been how to assess the statistical significance of the differences observed. The latter issue is also present in comparisons of word frequencies in diachronic corpus linguistics: previous work has tended to use tests which make the invalid assumption that words occur randomly in texts. Moreover, the question often arises whether the change observed is linguistic, stylistic or an artefact of the corpus. The present work explores sociolinguistic variation and change in the morphological productivity of the nominal suffixes -ness and -ity from Early Modern English to Present-day English, using materials such as the Corpora of Early English Correspondence and the British National Corpus. To do this, it employs robust methods to compare item frequencies over time and across social categories. Developed in collaboration with computer scientists, the methods include non-parametric measures of statistical significance as well as visualisations revealing variability within (sub)corpora and facilitating exploration. In addition to research into individual linguistic features, the methods can be used to compare corpora and study genre continuity at the levels of vocabulary and parts of speech. Besides corpus-linguistic methodology, the work contributes to the theory and description of derivational productivity. Firstly, it shows that each of the social categories studied - gender, social rank, and register in terms of participant relations - may have an influence on productivity, gender being the most consistent factor in the case of -ity. Furthermore, it shows that while productivity measures based on the frequency of hapax legomena, or words occurring only once in the corpus, are unusable in small corpora, they do function as expected in large corpora and remain theoretically valid. These findings should be taken into account in future research, and it is to be hoped that future studies will be significantly facilitated by the methodological contributions presented in this dissertation.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 44 citations exist upstream.