Message-Id: <m0rinqN-00001pC@xiron.pc.helsinki.fi>
Date:         Sun, 26 Feb 1995 13:40:28 EST
Reply-To:     jorge@PHYAST.PITT.EDU
Sender:       Lojban list <LOJBAN%CUVMB.bitnet@FINHUTC.hut.fi>
From:         jorge@PHYAST.PITT.EDU
Subject:      Re: On {lo} and existence
To:         lojban@cuvmb.cc.columbia.edu
Content-Length: 3190
Lines: 84

And:
> >        mi senva lo ninmu no'u mi
> >        ije  mi nanmu
> >        i seni'ibo lo ninmu cu nanmu
>
> It can't be true in this world, but there might be some world in which
> this could be true.

It is true in this world. Each of the three sentences
(with your interpretation of {lo}) can be true in this world.

> The process is analogous to deciding whether {le nanmu cu ninmu}
> is true. First the hearer must ascertain who {le nanmu} refers to.

Exactly. What you suggest in practice is that {lo} is nonspecific
nonveridical. I think that goes against the canon.

> > But we are considering the case where U is not the same as R.
> > U is the dream where you are a ninmu. Then clearly in this
> > universe R, {ro ninmu cu ninmu} is false.
>
> With this U, it is indeed false. But the point is that not every
> duhu derived from this seduhu need be false, or true.

Well, I don't know if that is the point. I would like that {ro ninmu}
mean {ro da poi ninmu}, and be able to use this for ease of logical
manipulation. In your scheme, logical manipulation becomes a headache
(you can't export quantifiers of a {lo} expression to the prenex
without filling the sentence with {da'i}s).

> That is, in this real world I can truthfully say "I described
> my wings".

Meaning "I said that I have wings"? That smells of sumti raising.
Can you write "I described my wings" using predicate logic?
I bet you need to embed the quantification of "wings" in a
sub-clause. We are raising from:

        mi cusku le sedu'u da poi nalci zo'u  mi ponse da
to:
        mi skicu lo mi nalci

But in the second one, the embeded quantification jumped out to
the main bridi.

> > You can't say "I don't have wings, but
> > they are very pretty".
>
> That's right. It's only certain things like describees that don't
> have to exist in the same universe as the universe in which the
> main predication obtains.

I think that's sumti raising in disguise.

Does {mi skicu lo mi nelci} entail {mi skicu da}?
Then what is the answer to {mi skicu da poi mo}?

> > > If they can't both be true, then {lo nu} must denote something
> > > that really happens. That would be very inconvenient.
> > Unless {nu <bridi>} means "x1 is a potential event of <bridi>".
> > Potential in R, independently of whether it happens or not in some U.
>
> You'd have to explain to me how one ascertains whether something
> is potential.

For every <bridi>, {da poi nu <bridi>} is defined as a potential event.

A potential event can happen, in which case it is an actual event,
or never happen, in which case it remains a potential event only.
(I don't like to define {nu} this way, I'm just trying to justify
it's use for irrealis events).

> > But I agree that {lo nu} should denote something that really happens.
> > Unfortunately, usage probably will decide against that.
>
> This, you will realize, is why I, having originally taken the same
> position as you, have elected to support the opposing view.

My problem is with your extension of this inconvenience to objects.
To let {lo nu klama} be an event that never happens is bad enough,
but to let {lo mlatu} be something that is never a cat is too much
for me.

Jorge