Re: Relative clauses and lambda calculus

[This accidentally went directly to Pablo, and not to the list.]

 Your message required far too much thought to go unanswered.
In answering, though, you should understand that I have a quarter's
worth of graduate formal semantics under my belt, and I never
fully got the hang of lambda extraction (i.e., the most important
thing we learned, according to the prof.  I do know fuzzy logic,
though!).

Anyway, what you've done seems really neat.  I don't understand
many of the technical details once you start talking about the identity
function, but before that I got it.  The "language" that you created
actually seems like a decent way to create a robustly head-initial
language--just by itself.  I have one VSO language, but it's not really
*robustly* VSO, so relative clauses are handled a little differently.
Nevertheless, what you propose is a neat idea.

I know that in my semantics class relative clauses were handled
differently (I believe even "and" was handled differently...).  I can't
for the life of me remember how it was done.  I think adjectives
were handled differently, as well.

Pablo wroet:

<<
This is the man [my girlfriend's father is a friend of John and him]
>>
I believe that the way this would actually work in English is with a
gap were "him" is.  And there's a special reason that it won't work.
It's because there's an island in here which prevents abstraction.  It's
not that it *has* to, it's just that in English it does.  There is no
structural reason why "This is the man my girlfriend's father is a
friend of John and" shouldn't work in *a* language.  In English,
though, the islands are defined in such a way that it doesn't work.
The particular island in question here is the conjunctive phrase, but
it has a particular name that's escaping me...

Anyway, regarding what you said here:

<<
These kilometric expressions make me believe this is probably violating
a language universal, in the sense that it is counter-intuitive. Anyway,
at least for me, it is an interesting theoretical exercise.
>>
It doesn't violate a universal, it just violates the grammar of English.
There's no special reason why it should violate *all* grammars, save,
maybe, processing concerns.  That is, some things are too imbedded
to be recovered.  It's up to the grammar of a language to determine
the grammatical cut-off point.  An example from the Simpsons was
this, said by Moe:

"You know what I blame this on the breakdown of?  Society."

That's ungrammatical in English, but it's certainly understandable,
otherwise it wouldn't have been a successful joke.  So here the
grammatical cut-off point and the processing cut-off point don't
match.  But then, as I recall, there's a particular Dilbert cartoon
where there's a sentence that's pretty much unprocessable.  It's
ungrammatical, as well, but is so nested that it can't even be
processed, even though you can sit down with a pen and paper
and figure out exactly it works.  (Anyone know the cartoon I'm
talking about?  Is there an archive?)

Anyway, though, interesting work!  I'd like to see you put it
into use.

-David
*******************************************************************
"sunly eleSkarez ygralleryf ydZZixelje je ox2mejze."
"No eternal reward will forgive us now for wasting the dawn."

-Jim Morrison

http://dedalvs.free.fr/