The meaning of GENERIC in linguistics (one last word for now)

[Laurence Horn]

At 2:53 AM -0500 3/6/05, Benjamin Zimmer wrote:
>On Sat, 5 Mar 2005 18:38:31 EST, RonButters at AOL.COM wrote:
>
>>I did find two instances of "autohyponymy" through Google, one of which
>>defined it as the case where "the new sense of a term is a hyponym of the
>>original." I now see that this is a term of his own creation and that at
>>least two other linguists have also used it since 1984. It looks like a
>>useful term to me, so I'm glad he coined it and I am sure I will use it
>>from now on whenever I am in need of such a word, even though this use of
>><auto-> strikes me as somewhat eccentric compared to the use in, say,
>>"autoerotic" or "automobile" (just a matter of taste, of course).
>
>Those of a structuralist bent would understand what Larry describes in
>terms of "markedness".  An autohyponym is the "unmarked" of two items
>asymmetrically opposed in a markedness relationship.  I don't know if
>Larry's 1984 piece (NELS, "Ambiguity, negation, and the London School of
>Parsimony"?) relates autohyponymy to markedness, but here's something that
>does:
>
>-------------------
>http://amor.rz.hu-berlin.de/~h2816i3x/LexSemantik1.pdf
>
Interesting convergence.  Manfred Krifka, whose course material is
posted at the above website, is one of the principal authors in _The
Generic Book_, which I was just citing last night in my "final
posting" on the genericide debate.

Well, yes, I do relate autohyponymy to markedness, but it's a fairly
complex relation.  Some cases of markedness don't involve
autohyponymy, of course--the tall/short, wide/narrow classes of
marked/marked adjective pairs, for example.  The unmarked term has
wider distribution and is less informative in those cases (as with
the autohyponym like "finger" vs. its cohyponym "thumb"), but the
marked term doesn't really count as a hyponym of the unmarked in
those cases.   In other cases, as with "drink" ('imbibe') vs. "drink"
('imbibe alcoholically'), we have autohyponymy but no markedness
(since there's no marked member of a lexical opposition.  Then there
are the problems with the concept of markedness itself, the
difference between morphological/formal markedness and semantic
markedness--happy/sad illustrates the latter but not the former,
happy/unhappy illustrates both, and neither involves autohyponymy.
And then there are the uses in acquisition, lang. universals,
innateness, phonology/phonetics, etc.  (Some semanticists, like
Krifka's countryman Martin Haspelmath, have gone so far as to junk
"markedness" entirely, but I do find it useful, if it's handled with
care.)

Larry