Formulaic language in Early English Books Online : From computational linguistics to classical rhetoric

As Sinclair (2004: 30) points out, ‘so-called "fixed phrases" are not in fact fixed; there are very few invariable phrases in English.’ Consequently, any exploration of formulaic language needs to consider not only their structural forms but also their contextual functions. To this end, we analyse repeated clusters of words (n-grams) from Early English Books Online (EEBO) and subject our results to qualitative analysis using analytical frameworks from both classical rhetoric and stylistics. With regard to the former… With regard to the latter, we pay particular attention to discourse presentation (i.e. the techniques used to present the speech and thought of third parties). We extracted our clusters from the entire EEBO collection, as curated by the Oxford Text Archive. The corpus consists of 60,328 printed works, with approximately 1.5 billion words. Our study therefore constitutes what is likely to be the largest extraction of n-grams from a historical corpus of English to date. The n-gram frequency lists are all made freely available online via the Oxford Text Archive. Preliminary analysis of the most frequent 4-, 5-, and 6-word clusters reveals a preponderance of formulaic expressions from religious texts, often bible quotations or popular prayers and other ritual expressions. Our analysis of the rhetorical and stylistic functions of these clusters allows insight not only into the nature of formulaic expression in Early Modern English but also the nature of interpersonal communication in this period. In this respect, our study demonstrates the capacity of large corpora of written material to offer insights into the nature of communication across both writing and speech.