KIF: new version of KIF-Core

[Chris Menzel]

> Adam wrote:
>    As far as a technical proposal against empty quantifier variable lists, 
> and allowing logical connectives to have arity 0, I believe the opinion 
> I've already stated is no more or less technical than the contrary opinions 
> offered.  To reiterate, I believe that allowing empty lists of this sort 
> means allowing vacuous statements which have no value in knowledge 
> representation.

Adam, vacuous quantification is already possible in the existing suo-kif
syntax.  Here's an example:  (forall (?x) (happy adam)).  Even though the
variable list is nonempty here, the quantifier is absolutely vacuous,
certainly no less so than (forall () (Foo)).  If you truly want to be
consistent in the design philosophy you advocate above, you should rule
out statements with vacuous quantifiers like the above as well.  By
disallowing quantifiers with empty variable lists and allowing vacuous
quantifiers like the above, suo-kif is actually stuck in a sort of limbo
between the kind of strict syntax you seem to want and the freer syntax
favored by Pat and Mark (and me).  A well-designed language should either
rule out vacuous quantification in all its forms, or should allow it fully
without arbitrary restrictions.  Note, however, that if you do try to
cleanse KIF of vacuous quantification entirely you incur a computational
cost, as you need to move to a context-sensitive, hence computationally
more expensive, language to do so.  

So here's my argument:  consistency in design demands either that all
forms of vacuous quantification be disallowed or that all forms be
allowed.  The former is computationally more expensive than the latter.
Moreoever vacuous quantification has no ill effects -- it can be safely
ignored by users, and presents no semantic difficulties.  Hence, all forms
of vacuous quantification should be allowed.

As for (and) and (or), note that these might in fact come in quite handy
in some KR contexts, as they provide a user with simple ways of expressing
truth and falsehood, respectively.

-chris