Date: Mon, 10 Apr 1995 18:47:21 -0400 (EDT)
From: jorge@PHYAST.PITT.EDU
Subject: Re: More about scopes
Sender: Lojban list <LOJBAN@CUVMB.BITNET>
To: Veijo Vilva <veion@XIRON.PC.HELSINKI.FI>
Reply-to: jorge@PHYAST.PITT.EDU
Message-id: <01HPID5C2F5C0000B9@FIPORT.FUNET.FI>
Content-transfer-encoding: 7BIT
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
Content-Length: 5648
Lines: 157

And:
> > > More precisely, {koha} is like {lei}.
> > I'd like that, but it's not what you are saying below. If it was like
> > {lei}, then it would always have a single referent. (Which might well
> > be an n-tuple, but it would be a single referent.)
> If it is like {lei} it has a specific referent, that's for sure. But
> it's not for sure that it would always have a single referent - that's
> what we're debating in response to your thought-provoking question

{lei} always has a single referent, a mass. The mass may be composed of
several individuals, but from the point of view of the grammar it
has only one referent. {le} can have many referents, but each of them
is an individual.

(Before anyone says anything, yes, {le} can refer to masses, e.g.
{le gunma}, but this is still some number of individual masses.
{le broda} is each of the indviduals that I refer to as broda.)

> [Having just read the screenplay of the film _Pulp Fiction_, I am
> inclined to call it a coolass question].

Well, thank you!  (I think...)

>    le ci nanmu cu bevri *re* tanxe goi koha
>    i koha tanxe remei
>    i koha na tanxe xamei

I think that it's either {ko'a tanxe pamei} or {ko'a tanxe xamei}
(assuming all boxes are different, for simplicity). This is because the
second sentence is already out of reach of the prenex of the first, and
I think it doesn't make sense to have three pairs of boxes as the
referents of {ko'a}.

>    i kuhi mi na djuno le duhu xohu ro da du lo jei koha tanxe pamei

For every x, you don't know that x is the truth value of {ko'a tanxe
pamei}?  What a strange thing to say. What about for {ko'e goi lo jei
ko'a tanxe pamei}? Are you sure you don't know that {ko'e du lo jei
ko'a tanxe pamei}?

>    to vahi mi na djuno le duhu xukau koha tanxe pamei

A much better way to put it. :)

I think it can't be, unless you had started with {le ci nanmu cu bevri
pa tanxe remei goi ko'a}. In that case ko'a would be three pamei of
pairs.

>   le ci nanmu cu bevri pa tanxe goi koha i koha    blanu
>   nanmu-1        bevri tanxe-x             tanxe-x blanu
>   nanmu-2        bevri tanxe-y             tanxe-y blanu
>   nanmu-3        bevri tanxe-z             tanxe-z blanu
>
> where tanxe-x, tanxe-y and tanxe-z may be the same or different.

In that case, {ko'a} has three referents, and a better way to put it
would be:

  nanmu-1        bevri tanxe-x
  nanmu-2        bevri tanxe-y
  nanmu-3        bevri tanxe-z

  tanxe-x blanu
  tanxe-y blanu
  tanxe-z blanu

You get the first three from the first sentence, and independently
of that, you get the next three form the other sentence. The second
sentence is not under the scope of the prenex of the first.

> I don't know if we agree. Does {pa tanxe} (in the example) refer to
> one singleton set of boxes or three singleton sets of boxes?

To three, one for each man. It doesn't really matter in this case,
because you can't really split it into three sentences directly, but
consider the analogous case:

        la djan e la sofis bevri pa tanxe

That expands to:

        la djan bevri pa tanxe ije la sofis bevri pa tanxe

To how many boxes does {pa tanxe} refer to in the first sentence?

> Whatever,
> koha refers to whatever {pa tanxe} does.

The difference is that {ko'a} keeps its referents when it appears in
following sentences, while {pa tanxe} doesn't.

> > Prenexes have scope over a single
> > sentence, not over following ones, unless appropriately bracketed.
> If this is correct, then everything I'm saying is wrong.
> It seems that either scope may cross sentence boundaries, or,
> as you have suggested, {goi} has wide scope regardless of word-order.

Scope doesn't cross sentence boundaries unless bracketed.

However, there are still the two possibilities for {goi}. It can still
have small scope, in which case the assignment will depend on the whole
sentence context, or wide scope, overriding sumti order.

There is also the question of whether {re tanxe goi ko'a} assigns two
referents to {ko'a}, or a single mass referent, a pair. Probably the
first, unfortunately.


> > I think this is the right expansion:
> >        le ci nanmu cu bevri pa tanxe goi ko'a
> >        i ko'a blanu
> > Expands to:
> >        i ci da voi nanmu pa de poi tanxe zi'e goi ko'a zo'u: da bevri de

It should have read:
      i ro da voi nanmu ku'o pa de poi tanxe zi'e goi ko'a zo'u: da bevri de

> >        i ro da poi du su'oko'a zo'u: da blanu
>
> If you think that works, then I don't understand why you originally
> said {goi} should get wide scope.

Because with wide scope, the assignment seemed much simpler. I am still
not sure which I prefer.

> But I don't see how your expansion
> works. The problem is that for each nanmu, {koha} gets assigned a new
> referent.

That's right. Why would that be a problem? Everything is done
simultaneously, it's not as if you had to have the referents in line
waiting to be assigned to ko'a, or some other physicalist picture.

> Perhaps what is needed is a variant of {goi} that *adds*
> extra referents to {koha}, instead of replacing existing referents
> of {koha}.

That's not needed in this case, because there is nothing to replace,
there is no order for each of the men's boxes to get into ko'a.
But I agree it sounds interesting. Perhaps the trick could be done
with something like:

        ko'a goi ko'a e pa tanxe

Wouldn't that add a box to {ko'a}'s previous referents?


> > Is it available for perusal?
> No. Sometime next millenium.

It's interesting that one can say that and really mean it literally.
To think that we all will have been born in the past millenium not
very long from now.

Jorge