Date:  Thu, 26 Feb 1998 07:12:04 -0500 (EST)
Message-Id: <199802261212.HAA26930@locke.ccil.org>
Reply-To: Chris Bogart <cbogart@QUETZAL.COM>
Sender: Lojban list <LOJBAN@CUVMB.BITNET>
From: Chris Bogart <cbogart@QUETZAL.COM>
Subject:      Meaning of BAI tags
X-To:         lojban <lojban@cuvmb.cc.columbia.edu>
To: John Cowan <cowan@LOCKE.CCIL.ORG>
X-UIDL: f57036219633ef0d7e1382b565de6ae2
Status: OR
X-Mozilla-Status: 8001
X-From-Space-Date: Mon Mar 02 13:25:38 1998
X-From-Space-Address: -

I wrote:
>>Hmmm, am I right about that?  I had thought that  "Proposition BAI sumti"
>>always entailed "Proposition", but now I'm not so sure...

And Jorge added:
> {cu'u} seems to cause the same type of interference.

Hmm, maybe I'm not as clear on the meaning of BAI tags as I should
be.  I could see two possible interpretations of cu'u:

           sy. tanru cu'u xorxes
           1) S is a tanru, spoken by Jorge
           2) S is a tanru, according to Jorge

I think that's more than a doubt about the precise definition of cu'u, it's a
 question about
BAI tags in general.  The first option leaves the truth of the proposition in
 place and adds
extra information; the second more fundamentally changes the meaning as if cu'u
 were
actually the main verb and tanru were part of an abstraction (i.e. lenu sy.
 tanru kei se
cusku xy.)

More examples:
          sy. tanru va'o my.
          1) S is a tanru, and the conditions surrounding the fact of its
 tanrueity are M
          2) S is a tanru, at least under conditions M

          la .erik limna bai la rabyn.
          1) Robin forces Erik to swim
          2) *Robin forces that Erik would swim (somehow implying that Robin
 forces,
                        but Erik might not swim)

          la selbarna cu mlatu du'u la .and.
          1) Spot is a cat; And knows that
          2) Spot is the cat that And knows

The choice in all these cases is whether the interpretation could be such that
 the main
bridi wouldn't be true without the BAI tag.  The one with {bai} itself seems
 unlikely, but
the others...?

I prefer the interpretation where {Prop BAI Sumti} implies {Prop}, but that
 would mean
we have to be very careful with sentences like:

       Sy. mlatu fi'o selje'u my.
 ...which under my proposal would mean:
      1) S is a cat, and that's true under epistemology M
 ...but not:
       2) S is considered a cat under epistemology M (but maybe not under my own
 epist).

The reason I prefer it is that I think it would make lojban text easier to parse
 and
logically analyze in software.  I think a rule that says:

          A djuno B C D bai E
                implies
          A djuno B C D

would be wicked useful, because we could apply rules concerning "djuno" and
 "bapli".
Without such a rule, it's hard to know how the sentence could be analyzed.

Other rules that would be useful, although maybe problematic:
       A djuno B C D bai E   -->  E bapli lenu A djuno B C D
       A djuno B C D bai E  --->  A djuno
       A djuno B C D bai E  --->  A djuno bai E
       A djuno B C D bai E  --->  B se djuno fo D

...and even a base of rules worked out by hand with things like:
       A badna ---> A grute
       (A is a banana --> A is a fruit)

Right now the best my computer can do with lojban posts is tell me if the
 sentences
parse and spew out canned definitions of words I click on.  It wants to learn
 the
language better, and it's getting impatient.  Just yesterday it locked up in
 protest
after I downloaded my email.  Things are getting ugl@#&FFNeFLEhfelf
9pf4ay48n83ur93#(&%&%#Nce3q823pa34ca43rcu4ar

--------- user session terminated ----------