Construction and application of Bounded Virtual Corpora of British and American English
Bickerstaff, Garrison E.
MetadataShow full item record
This study proposes a systematic approach that capitalizes on the quantitative advantages of corpus linguistics and caters to researchers’ needs for extremely large amounts of text that can address regional American speech. The Bounded Virtual Corpus (BVC), an online virtual corpus is constructed within finite boundaries and contains a ten-year span (1998 to 2007) of 25 American newspapers distributed over five different regions. The BVC thus reflects regional frequency of word use and supports lexicographic decision making. A separate British component reflects lexical activity in British English and contains the full text of 5 newspapers over the same ten-year span, which enables direct comparisons between British and American English. The BVC’s virtual corpus methodology allows for calculation of rates of occurrence and other metrics for word use that cannot be achieved by using the Internet as a corpus. Estimated word counts for each component—5 billion words for the American regions and 1 billion words for the British section—provide a baseline for such calculations. The BVC is unusually large in comparison to other currently well-known corpus projects and provides more evidence than smaller corpora for the study of careers of words, especially lower frequency words. The LexisNexis Academic interface includes a robust downloading interface that allows BVC search results to be saved conveniently and uploaded to offline concordancing applications as needed for analyses. Through results from the BVC, this study shows some recent differences in use of American and British English in the phrase go missing as well how the BVC can be used to identify and describe the trajectory of a neologism, carb, before and after it has reached its peak in frequency. There is also a demonstration of how the BVC can be a useful lexicographic tool regarding potential revision of A Dictionary of Americanisms on Historical Principles. The BVC can be replicated using detailed instructions included in the dissertation, and can be used free of charge by any researcher who has legal access to LexisNexis Academic.