I would appreciate any information/leads on what methods are being used
to provide corpus-based confirmatory evidence for the
linguist/lexicographer's postulation of different senses/uses of a word.
Granted that we expect different uses of a word to be reflected in
different text patterns, how can we actually exploit this?
I can imagine pursing the use of translations (in parallel texts) or
substitutability of synonyms, or use of collocates that are categorized
in some way, but there are a variety of problems.
Any leads/references?
Michael Barlow