Newsgroups: alt.usage.english,sci.lang
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!uunet!in1.uu.net!eskimo!rickw
From: rickw@eskimo.com (Richard Wojcik)
Subject: Re: grammar that includes semantics
X-Nntp-Posting-Host: eskimo.com
Message-ID: <D7228s.Kyr@eskimo.com>
Followup-To: alt.usage.english
Sender: usenet@eskimo.com (News User Id)
Organization: Eskimo North (206) For-Ever
References: <3mm9cm$771@riscsm.scripps.edu>
Date: Sat, 15 Apr 1995 02:43:39 GMT
Lines: 62

In article <3mm9cm$771@riscsm.scripps.edu>,
Mark Israel <misrael@scripps.edu> wrote:
>In article <D701DH.H49@eskimo.com>, rickw@eskimo.com (Richard Wojcik) writes:
>
>} Actually, "Tom ate the window" is semantically well-formed.  It may be 
>} a stupid or a weird thing for Tom to do, but the sentence itself makes 
>} perfect sense.  Perhaps you are looking for something a little more 
>} difficult to imagine, such as "Tom spoke a banana".  It is certainly 
>} true that windows aren't normally conceived of as food, but you can eat 
>} any solid object, in principle.  The object of "speak", however, must be 
>} some kind of speech object.
>}
>} There was a time when generative linguists would have called the "speak
>} banana" construction a "selectional restriction violation", after Chomsky,
>} and called it a grammatical violation.  However, that was back in the
>} 1960's, and this kind of anomaly is no longer considered a "grammatical"
>} violation in anybody's theory, to my knowledge.
>
>   Richard Wojcik, meet Paul Ivsin.
>
>In article <1994Dec31.174426.2124@midway.uchicago.edu>,
>pjc1@woodlawn.uchicago.edu (Paul Ivsin) writes: 
>
>}>> And even if they didn't, the grammaticality of sentences such as 
>}>> "Fowler's facts are wrong" leaves my original point undisturbed.
>}>
>}> You use "grammaticality" where I would expect to see "semantic
>}> validity" or some such.  Have you defined a grammar that includes
>}> semantics, and if so, where can I read about it?
>}
>} I use "grammaticality" more or less where I would use "availability",
>} and avoid semantics (valid or invalid) as much as possible.  If you
>} wished to be really precise, why not "lexicographic appropriateness"?
>}
>} No one, I think, would maintain that one needs to have a fully-
>} defined grammar before one can make judgments on the grammaticality
>} of given sentences.  My usage is unexceptional.

Mark, you have a real propensity for quoting people out of context as if
they were making comments that were relevant to the given context.  This is
a particularly egregious example of this habit.  You juxtapose two passages
from different threads as if they somehow contradicted each other.  The
sentence "Fowler's facts are wrong" is an interesting sentence, but it is
not a case of semantic anomaly.  I do not know what the "original point"
was, because you clipped out most of the context.  So I don't really have
any judgments about the point that Ivsin was making to you.

I wonder if you are familiar with the difference between referential and
attributive senses of noun phrases.  A sentence like "John's mother is
his father" may appear to be a contradiction, but it can be used in
perfectly valid ways, e.g. to mean that the person who appears to be John's
mother (attributive sense) is really his father.  The sentence "Fowler's
facts are wrong" is not necessarily bad usage if you take "Fowler's facts"
to be an attributive usage.  That is, it could be used to express the
notion that the propositions identified in the given context as "Fowler's
facts" are incorrect propositions.  It is an amusing way of saying that
Fowler's facts are not really facts at all.  If you look around for this
phenomenon, you will probably see that people switch between attributive
and referential senses of NPs all the time.
-- 
Rick Wojcik  rickw@eskimo.com     Seattle (for locals: Bellevue), WA
             http://www.eskimo.com/~rickw/
