Message-Id: <m0qlFhR-00005LC@xiron.pc.helsinki.fi>
Date:         Thu, 15 Sep 1994 13:14:00 BST
Reply-To:     i.alexander.bra0125@oasis.icl.co.uk
Sender:       Lojban list <LOJBAN%CUVMB.bitnet@FINHUTC.hut.fi>
From:         i.alexander.bra0125@OASIS.ICL.CO.UK
Subject:      TECH: Any old thing whatsoever (was RE: do djica loi ckafi je'i
              tcati)
To:           Veijo Vilva <veion@XIRON.PC.HELSINKI.FI>
Content-Length: 4692
Lines: 123

Subject:  TECH:  Any old thing whatsoever (was RE: do djica loi ckafi je'i
 tcati)

cu'u la'o gy. Randall Holmes .gy.
> This seems only to go to you via my reply function, and I don't know
> how to e-mail the whole list; I'll try to do this in the cc: line at
> the end of this message.

It sounds from Mark's comment as if it got through.

> The same problem arises in TLI Loglan; the paradigmatic example which
> caused a lot of discussion was "I am waiting for a taxi".  The
> difficulty seems to be that the logical form of the sentence is an
> illusion; there is no box referred to in "I need a box", and there is
> no taxi referred to in "I am waiting for a taxi" (there need not even
> exist any boxes or taxis meeting your requirements for the statement
> to be true).  The context is "referentially opaque", in Quine's
> terminology, and the object of the sentence, if it has one, is some
> kind of "intensional" object (something on the order of a concept of a
> box or taxi).

>                                         --Randall Holmes
>                                         ("logician in residence", TLI)

Yes, but *what* sort of "intensional object"?
This attemps to identify the problem, but I'm not sure how much
it helps us find the solution.


cu'u la mark. clsn.
> I recall we went through this discussion once before; in fact it was
> spurred on by a similar discussion regarding TLI Loglan regarding taxis
> (mentioned by Randall Holmes here, I see).  The answer there (our analogous
> version of JCB's I think, and I liked it) was "loi tanxe".  This works.  I
> need [some part of] the mass of things that are boxes.  Possibly "lei
> tanxe" if you want to admit something that isn't a box but turns out to be
> what I meant anyway.  I don't think we need a new quantifier for this one;
> massification works (unless massification was rethought and redefined since
> the last time this question came through and I missed it).  I'll try to
> find quotes from the last time.

(I'm with Jorge on this one.)
Sorry, Mark, I didn't really buy this the last time round,
and I think I understand better why now.

    mi nitcu loi tanxe

means

    There is some part of the mass of things that are boxes
    that I need.

In other words, it suffers from the same problem as the {lo} version.
It's more difficult to think of examples where you would actually
want to say this, but I firmly believe that it has to work this way.
Massification is irrelevant.


Jorge's {xe'e} = "any":

This strikes me as pretty dubious semantically (even more
problematical than {po'o} = only).  "Any" is kind of ambiguous
between "all" and "one".

    xe'eti ka'e se pilno
    Any of these will do.

might as well be

    roti ka'e se pilno
    All of these are usable.
    Each of these is capable of being used.

(I've previously on occasion advocated translating {ro} as
"each" rather than "all".  It means the same in the simple cases,
and helps demonstrate problems similar to the one we're discussing
here in the more complicated ones.)


The most promising candidate for an answer seems to be that
there is some sort of abstraction which has been elided.
(This is presumably something like what Quine has in mind
as an "intensional" object.  (Haven't read Quine - should get
to the top of the to-do list sometime in the next ten years.) :-) )
But is it a potentially different abstraction in each case,
or is there conceivably a single general purpose one which
would cover them all?

{tu'a} appears to serve in all the cases we've considered
(except perhaps {sisku}, and I seem to remember John Cowan
having second thoughts about the use of {ka} for that one),
although it's just about as vague as it could possibly be.

Or should it perhaps be a Lojban abstraction,
{su'u <gismu> kei beleka broda} for some suitable {broda}?
A single one, or individually chosen?


cu'u la'o gy. Chris Bogart .gy.
> Reading this over, I'm unsure whether I really want to claim "ka" is just
> the right word; suppose there were a new abstractor with the same grammar
> that filled this function -- I have no idea how you'd define it in English,
> though.

Quite.

> But we shouldn't *require* nitce to take an abstraction because it's still
> useful to be able to say "mi nitce lo tanxe" if it is in fact a specific box
> you are referring to, rather than just a box in general.

Well, yes and no.
In theory, even if {nitce} (etc.) takes an abstraction, you can always
adjust the quantification to achieve that effect, e.g.

    da poi tanxe zo'u:  mi nitce tu'a da

In practice, however, that's admittedly not very convenient.


Sorry, no answers today, just questions.

Yours,
     Puzzled of Reading.