WikiDiscuss

---

> The definition of {ce'u} suggests that it is syntactically like a quantifier,
> rather than like a quantifier-bound-variable all in one.

That's misleading, yes. It would be better to describe it as
"lambda-quantified variable" if lambda is to be mentioned at all.

> Now, it *ought* to
> function like a quantifier and take a bound vriable, but it does not. This
> creates problems for cases where more than one place is to be bound in the
> same way and also misleads about structure.

When more than one ce'u is used, they are taken as independently
bound variables. To bind more than one place together, we have to
use {ce'u goi ko'a ... ko'a} or other pronouns that point to the first
ce'u.

Another related problem is embedded properties. Just as with relative
clauses, the solution here is to use subindices (but I never remember
which way to count), or to put a {ce'u/ke'a goi ko'a} in the prenex
at the right level.

> On the other hand, the real
> lambda format is painfully lengthy (see LISP). Some compromise would be
> nice.

The current scheme for {ce'u} seems to work fine in practice,
since more than one ce'u is rare, and when we do need more than
one (for the x2 of {simxu} basically) we do want them as two
independent variables.

Embedded relative clauses do turn up from time to time, and
subindices are ugly. I don't have a better solution though.

mu'o mi'e xorxes

__
Do you Yahoo!?
New and Improved Yahoo! Mail - Send 10MB messages!
http://promotions.yahoo.com/new_mail