Re: open Lojban issues

["John E. Clifford" <pcliffje@crl.com>]

Given my unfondness for YACC and the fact that I have not followed
developments therein since the creation of the umpteenth version of _e_,
I'll pass on all technical grammar questions.  (I hope that the forthcoming
refgram will contain a fully commented copy of the current grammar and
some clues as to what those odd names mean.  Most seem to be historic
but to have long since departed from their historical roots.)

So, I jump to B.  What I suggested, actually, was 1) that the
quantifier+variable forms (the basic forms in logic) be taken as basic in
Lojban too and that the quantifier + predicate or term be defined as
representing them.  Thus, for example, _ro lo broda_ would be the
universal without existence claim simply because it represented _roda
ganai broda gi_.  This is standard logico-linguistic stuff.  The only case
for a universal requiring existence of a broda would be _ro da poi broda_
or any expression taken as equivalent to that (none have yet been so
construed officially, I think).  2) I suggested that we eliminate the
implicit INTERNAL quantifiers in _lo broda_ and the like.  Thus, a set is
the size it is but we need not comment on its size, even implicitly, to
refer to it -- we may not know its size and we certainly may not care. 
The internal quantifiers would still be legal, in case we did want to
comment.  And the set could be empty without any conflicts arising with
unmentioned sizes, thus eliminating the problem from which all of this
started.  I think the current rules about implicit *external* quantifiers
is probably about right.  I am not sure just where "world-creating" comes
from in this discussion and would like to keep it out, since it has horror
of its own.  At most, I would say that someone who uses an
existence-implying form is in a universe of discourse which contains that
sort of thing, but that need not be world creating, even if someone else
is not in that universe of discourse but claims to be in the same
discourse.  These are pragmatic, not semantic, problems -- somebody has to
get with the program. 

On G.  I am less sure about context leapers and related items now than I
was a year ago, since I have spent the intervening time in the slough of
opaque contexts and nested quantifiers, where the notions SEEM at first
glance to fit as well as the usual cases (negations and conditionals,
usually).  But each of these cases requires a slightly different arrangement
and I am not sure that a single mark will deal with all of them -- or
whether I would want it to.  I say drop it for now and cope with the
problem by the other means we have for a while: forethought quantifiers
for the transparent cases, second order devices (e.g., introducing sets) for
the nested quantifier cases, and the stock prohibition on quantifiying into
opaque contexts, even when they seem to be talking about external things
(the things inside the opaque context may be so different as to make the
quantification meaningless in practice:  the Lojbab of my beliefs may have
NONE of the properties of the real Lojbab, if I am seriously confused).  I
do not recognize MY problem of leapers in what Lojbab says, but that
may just be terminological drift that has plagued Lo??an since day one.  I
particular, I don't see what lambda has to do with it.

But, speaking of lambda, I want to reiterate my problem with introducing
lambda only in _ka_ constructions before we have any rules for making
sense of it gnerally.  Or, maybe, I just want to caution against calling
the form used in _ka_ constructions a lambda form, with the implication
that it can be used elsewhere.