Corpora and historical linguistics Corpora e linguística histórica

, 
the Innsbruck Computer Archive of Machine-Readable
English Texts
,
the Lampeter Corpus, the Shakespeare Corpus, and EEBO texts
to quantify the level and development of spelling variation in the history of
English, and to identify spelling patterns across periods and genres (BARON;
RAYSON; ARCHER, 2009a, 2009b; BARON; RAYSON, 2009). Clearly,
tools such as VARD2 show the way to future development of software and
have great potential to enhance the searchability of historical texts.
Having access to normalised spelling versions of historical corpora
would thus facilitate the use of sophisticated statistical analyses. For instance,
keyword analyses can be used to study the various ways in which texts
function, their related semantic spaces and collocational patterns (WYNNE,
2008, p. 730-734; ARCHER, 2009). Similarly, n-gram analyses based on
multi-word sequences located by the computer can be used to study recurrent
phraseology across the history of a language (for the principle, see WYNNE,
2008, p. 734-735; on lexical bundles in Early Modern vs. Present-day English
trials and play texts, see CULPEPER; KYTÖ, 2010, chapter 5). Further, by
using a data-driven bottom-up clustering method Gries and Hilpert (2008)
identified historical stages in the data based on differing quantitative
distributions. The data, originally collected and exploited for Hilpert (2006),
had been drawn from the Penn-Helsinki Parsed Corpus of Early Modern
English and the Corpus of Late Modern English Texts, with the different
spelling variants harmonised to their present-day counterparts (Gries and
Hilpert, 2008: 65). The study showed that, for instance in the case of the
verbal complementation of ‘shall’, the three consecutive 140-year periods that
had been distinguished as a result of pooling together the original six successive
70-year periods in the corpora did not tally with the way in which the data
actually distributed, falling instead into two 180-year groups in quantitative
terms. Discoveries such as these are important in that they enable language
historians to gain fresh insights and approach language change from a novel
perspective. Clearly, developing such techniques, and providing versions of
historical corpus texts that enable their use, are among the top priorities in
historical corpus linguistics.

---

441
RBLA, Belo Horizonte, v. 11, n. 2, p. 417-457, 2011