[Lexicog] Criteria for example sentences
Patrick Hanks
hanks at BBAW.DE
Tue Mar 16 10:57:15 UTC 2004
Ron Moe said:
> Listing collocates works if there aren't too many. But for general words you
> might want to consider looking for prototypical examples. Even though 'bark'
> is not very general, it is still subject to prototype effects:
>
> bark v. (of a dog) to make a quick, loud cry. Used of dogs and other animals
> with a cry like a dog's, such as a seal or cheetah.
>
> In other words 'bark' is not a general animal cry. It is the cry of a dog,
> extended to the cries of animals that sound like the bark of a dog.
> Yesterday I was telling a story about seeing a cheetah kill a gazelle and
> barking to call her cubs to the kill. I described the cry of the cheetah as
> a 'bark'. I think I said something like, "The cheetah started to bark. A
> cheetah's cry isn't like other cats. It's more like the bark of a dog." The
> notion of prototypes is very useful in dealing with categories of all sorts.
> If we want our example sentences to be most useful, they should exemplify
> typical usage. So the collocates used in the sentence should be
> prototypical, rather than marginal. So I would prefer "an itchy sore" to "an
> itchy wool sweater"; "cows moo", but "bulls bellow".
Thanks, Ron. IMHO, your definition of bark just about hits the nail on the head. I'd
like to offer a couple of intervention re prototypicality and evidence:
1. Adam Kilgarriff's wonderful WASPBENCH tool (http://wasps.itri.bton.ac.uk
-- it's free!) shows that in British English (BNC, using the MI statistic) the most
significant noun modified by itchy is feet (!), followed by rash, scalp, lump, skin,
and patch. Itchy also collocates significantly with red. A quick glance at the
Associated Press Corpus of 1991-2 suggests that the idiom "itchy feet" may be
British rather than American (it does not occur in AP). The significant collocates
of itchy in AP are rash and eyes. (I do not find "itchy sore" in either the Brit or
the American sources that I'm using. So this is a plausible but not a supported
example.)
The data is quite sparse (92 occurences of itchy in BNC; 52 in AP). We are
currently hampered in making such comparisons by the absence of an American
National Corpus equivalent in size and composition to BNC. (We live in hopes:
keep up the good work, Nancy, Randy, and co.!) Maybe Americans do use the
idiom itchy feet, just not AP journalists... Anyway, the data currently available
is at least sufficient to show that a prototypical example for itchy that is common
to both BrE and AmE is an itchy rash.
[PS for pedants: I do know that in Latin data is a plural noun. In British English it
is an open choice between a plural noun and a mass noun. I prefer the latter.]
2. I too believe that we should look to evidence of actual usage for examples,
selecting examples that are typical. (Novice lexicographers, in my experience,
are irresistibly drawn to "boundary-case" examples -- a problem that was greatly
exacerbated in the days when they were free to make up examples without
reference to evidence.)
We need to make a clear distinction between what goes on in a language and
what goes on in the world. Thus, in English, prototypical (bovine) cows moo.
Female whales may be called "cows" but they are not prototypical cows and
they do not moo. In fact, as Fritz points out, whales "sing" (without differentiation
of gender).
Do (bovine, prototypical) bulls moo? I don't think so. I can't find any evidence
that they do. Do they make a sound that is qualitatively different from that made
by their female counterparts? I don't know, but I do know (from corpus evidence)
that in English cattle "low", and that cattle is a gender-neutral term.
3. How to identify prototypical usage? Well, relative syntagmatic freedom is one
criterion. For example, "We saw some cows from the train" is an unexceptional
sentence of English, but "we saw some cows [meaning female whales] from the
boat" is very odd.
Patrick
Patrick Hanks
DWDS Collocations Project, Berlin-Brandenburg Academy of Sciences
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/lexicography/attachments/20040316/1baea953/attachment.htm>
More information about the Lexicography
mailing list