failures of parallelism

[Laurence Horn]

At 2:05 PM -0700 7/9/04, Arnold M. Zwicky wrote:
>On Jul 5, 2004, at 1:23 PM, Laurence Horn wrote:
>
>>...How about these truly negative ones, just googled up for our
>>viewing pleasure:
>>
>>consider password protecting that directory so that anyone can't come
>>along
>>and drop your tables
>>
>>"A customer will feel safer knowing that anyone
>>can't just waltz into their place of business."
>>
>>But anyone can't solve that problem...
>>
>>The Reality as below: anyone can't do what they want to do/anyone
>>can't be what they want to be/anyone can't say what they want to
>>say/anyone can't feel what they want to feel
>>
>>Just anybody can't baptize anybody.
>>
>>People are looking for more substance in the music, but just anybody
>>can't give
>>it to them," Ice Cube told the Los Angeles Times
>>
>>But I still have to know the password so just anybody can't get on
>>my desktop and start loading things.
>>
>>With the fiscal problems we have in Maryland, people are beginning to
>>realize that
>>just anybody can't be governor
>
>"But anyone can't solve that problem" is a real baffler for me; i have
>to stop and work out what someone might have been trying to convey by
>it.
>
>the others are, to various degrees, better.  all except the first have
>a "just" in them (and i understand the first as if it had a "just"),

That's the trick; the "just" of course is a free-choice
disambiguator.  But I've never been sure what we mean when we say
there's an implied "just" (others have said it about (what I call)
metalinguistic negation in scalar contexts:  I didn't eat {just} some
of the cookies, I ate all of them).  My sense is that even if English
had no word "just" we could get such sentences--on the free-choice or
metalinguistic understandings respectively.

>which seems to improve things some, especially in the "just anybody"
>examples.
>
Well, for what it's worth, I get "ANYbody can't solve that problem"
in the same way--it doesn't really need the "just", but it does need
a  fall-rise contour , without or (preferably) with the "just"
signaling what's coming.  This is noted by Haspelmath (1993/1997),
who calls these (= "not just any X") anti-depreciatives; I call them
anti-indiscriminatives in a couple of papers I published in 2000.
What's being negated here is the "n'importe qu-", "it doesn't matter
wh-", or "any old" understanding associated with free-choice _any_.

Larry