Slipping through the cracks of e-lexicography: lessons from ColloCaid

Thanks to the corpus revolution which underpins e-lexicography, headword lists
and defining vocabularies can now be adjusted to better reflect current language use. Definitions can be enhanced with information that goes beyond introspection alone. Syntactic patterning, lexical collocations and phraseology in general can now be given much more comprehensive coverage dictionary examples can be more easily found. Yet despite these unparalleled and undeniable advantages, there are elements of word usage that appear to be slipping through the cracks of corpus analyses and corpus-based lexicographic resources. Issues encountered during the development of the ColloCaid project, particularly its base word list, provide examples of such slips and a valuable opportunity to reflect on why they come about and how they might be addressed in future e-lexicography projects.