Percentage point--missed word in dictionaries

[Wendalyn Nichols]

In response to Victoria's comment:

>Bilingual and learners' dictionaries, especially, need to
>provide information about collocations.  Of course, their number is legion
>and no dictionary could possibly include them all in any way that would be
>meaningful to a user who didn't want to devote his entire life to reading
>it.

And to David's comment:


> This is surely a part--albeit a small part--of the art of lexicography.
>  Indeed, those publishers who think that computer programs or
> inexperienced editors can replace experienced lexicographers will end
> up with less artfully crafted lexicons.

If you're talking about machine translation, David, I certainly agree with
you--but not if you meant computer-based tools as a whole.

The art of lexicography can and should encompass the skilled use of computer
analysis, and the identification and inclusion of collocations is a great
case in point. Corpus analysis, routinely used outside of the US by houses
that specialize in bilingual and learners' dictionaries, is an invaluable
tool that enables lexicographers to spot the most common collocations and to
make decisions about including them (usually in example sentences) or
excluding them based on frequency of occurrence, which renders the choice
less subjective and less subject to gross omissions.. Allowing data to
overrule common sense becomes its own species of lexicographic error, of
course, which simply proves David's point about inexperienced
lexicographers.