On Fri, Nov 08, 2002 at 07:08:52AM -0500, John Cowan wrote:
> And Rosta scripsit:
> > The position supported by everybody except pc (= me, xorxes, Jordan,
> > Adam, Nick + probably xod & Robin -- everybody who's participated,
> > & probably the remainder of Lojbanists too) is this:
> >
> > A. ro broda cu brode = ro da poi broda cu brode
> > B. ro da poi broda cu broda = ro da ga na broda gi brode
> > C. ro broda cu brode = ro da ga na broda gi brode
> >
> > The position supported by pc is that C is definitely invalid, while
> > either one of A and B may be declared valid, with the other one
> > declared invalid (though his preference is for A to be valid and
> > B to be invalid).
Well, let me say that I would prefer if it could just be decided
that {ro} is nonimporting. Under those circumstances the logical
structure of {ro broda cu brode} is still A(broda(x)) (brode(x)),
which happens to have the same truth conditions as Ax(broda(x) ->
brode(x)).
Actually I'll go so far as to say that this method of resolving the
issue isn't consistent with relatively clauses in lojban in general,
so it's best to just stick with the point, which is whether {ro}
imports.
The consensus I'd like to see, (and it seems like it is certainly
within reach) is that {ro} is noniporting, and that C is only valid
in that they have the same truth conditions.
> Well, I must break consensus on this. The position here labeled pc's is
[...]
> > Each position is partially but not fully consistent with CLL (which
> > itself is not internally consistent).
>
> I still don't understand where the inconsistency is.
The inconsistency, which araizen pointed out, is that the naku
boundary rules don't work if you have the importing {ro}.
For example, assuming an importing ro, as the book says.
ro pavyseljirna cu blabi
is false, there's no pavyseljirna, so
naku ro pavyseljirna cu blabi
should be true, but:
naku ro pavyseljirna cu blabi ==
su'o pavyseljirna naku blabi
which claims that there a pavyseljirna.
Thus either "naku ro pavyseljirna cu blabi" is also false, or the
naku boundary rules in the book don't work.
In my view changing import of {ro} is far less disruptive than
changing the naku boundary rules, and nonimporting gives us the
ability to talk about lo'i pavyseljirna when there's none of them
(we could get this using AndR's trick of thinking about the *da*
as the thing which imports, but we'd still need to change the naku
rules in that case).
Furthermore, I don't think nonimporting ro is as abnormal as pc is
leading us to believe.
> > The debate about whether the universal quantifier and/or ro is
> > importing is pretty much a red-herring, because it boils down to
> > a question of the effect of an empty universe on truth values.
>
> I agree.
I don't: I think import of {ro} is the real issue here.
--
Jordan DeLong - fracture@allusion.net
lu zo'o loi censa bakni cu terzba le zaltapla poi xagrai li'u
sei la mark. tuen. cusku
Attachment:
pgp00251.pgp
Description: PGP signature