A Corpus-based Study of Latinate Words in Contemporary English

The English language has borrowed extensively from Latin. Despite their widely acknowledged importance, however, Latinate words are under-studied especially quantitatively according to their use across different linguistic settings. In this paper, we report a corpus-based survey of Latinate words and their use in contemporary English. The objective is to chart the use and distribution of Latinate words across a set of different text categories and subject domains in order to identify patterns of variation across such settings. The British National Corpus is chosen for its large size and wide variety of texts. Our results will show that the density of Latinate words, calculated as the proportion of Latinate words amongst all running tokens, successfully separates speech from writing and, within writing, academic prose from non-academic prose. The correlation between Latinate density and degrees of text formality is then compared with human ranking and also analyzed by the linear regression analysis. The results suggest a strong correlation between Latinate density and degrees of text formality. Our results also show that even different domains have their own preferences for the use of Latinate words. Our investigation is significant in two respects. Firstly, it is probably the first corpus-based, large scale survey of the use of Latinate words in contemporary English. Secondly, it measures the density of Latinate words both across different text categories as an important stylistic feature and across a set of different domains as a subject-specific differentia. Such an investigation lends itself to our understanding of the impact of Latin on contemporary English and the empirical findings will also contribute to practical applications such as automatic text classification and genre detection.