Re: Relative clauses and lambda calculus
|Date:||Monday, October 4, 2004, 5:31|
(No I'm not death but deaf :-)) and here is an example of how natlangs can
process subclauses. Is lambda calculus different?
The tall man who is John's father's friend writes a red book.
The man WHO he is tall END WHO he is the friend WHO the father WHO John has
him has him END writes a book WHICH it is red.
The book that the tall man who is John's father's friend wrote is red.
The book WHICH the man WHO he is tall END WHO he is the friend WHO the
father WHO John has him has him END writes it END is red.
Nota: WHO can be written "START he/it" as well. Him and it refer to WHO and
WHICH inside the subclause. END ends the subclause. You can add as many END
as there are WHO but when using it I realized it's no need. I also read in
some best-selling book that the "has has has" string breaches some
"universal", but you can find in Sumerian strings of "a(k) a(k) a(k) a"
manyfold nested subclauses so I think "universality" depends which languages
linguists take time to learn (or don't).
Pablo Barenbaum wrote:
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.
To create a real conlang, the apply operator should be grammaticalized
somehow (as an inflection, syntactically, etc.). I leave the rest as an
exercise to the conlang world.
>>>I am math death, blind and mute so would you be so kind as to examplify your
finding with dummy vocabulary? Does it deal with two-object predicates as
Until now, the best i've seen to deal nested stuff in natlangs is marking
the start and the end of the subclause and referring to its head back within
the subclause. What other device do you use here?
Thanks in advance.