Stylistic and historical studies

Some of the earliest applications of corpus-based analytical techniques were for the study of literary style. By the late 1960s, a number of literary scholars were applying computational techniques to on-line text collections, to investigate the styles of authors, genres, and historical periods. This early history is somewhat less surprising when we consider the fact that literary scholars have always been interested in the language of texts. It was thus a natural development to move those texts onto computer, allowing the use of concordancing software and other computational tools for analysis.

More recently, corpus-based analytical techniques have become popular for studies in historical linguistics. This, too, has been a natural development, given that historical linguists have always relied on text collections from earlier periods to trace historical change. However, a major problem for corpus-based historical investigations has been the absence of representative historical corpora. While literary scholars have often focused on the works of an individual author, historical linguists require corpora that represent a range of texts from multiple genres, across historical periods. Compiling such historical corpora has presented many challenges (see Methodology Box 2).

Diachronic text corpora enable a multitude of investigations. In fact, any of the kinds of research questions that have been discussed in the previous chapters of this book can also be studied from a historical perspective using diachronic text corpora.