Re: TECH: AI Project Proposal

[nsn (Nick Nicholas)]

Thank you very much for your detailed and well-thought out comments! I will
certainly keep you up to date with what I get up to. Some brief comments:
at THIS stage, I intend to use what NLP researchers do to distinguish "a"
and "the", to handle veridicality; I'll fill in details later. I think I
may have to excise a couple of topics off the agenda, which may also need
some refinement (I did not say anything about what sorts of cross-indexing
I'd consider in anaphor). I agree with you on attitudinals, at least at a
first cut, and I've always been a support of "dikni BAI" interpretation.
Canonicalising the Lojban by a preprocessor to the Prolog may well be a job
for some of the shareware LLG has; I'll have to ask. I do have some confidence
in Prolog's Definite Clause Grammars. In any case, this work will necessitate
me a lot of reading of NLP work; I'm particularly thinking of McCawley's work,
_what linguists have always wanted to know about logic but were too afraid
to ask_ (or something of the sort), which does a lot of analysis as to how
Traditional Logic can be an inaccurate model of NL. (Tried to read through
it a lot of times, but never seem to have gone very far :( )

I didn't get time to respond to this at the time (It seems I have time for
very little these days :( ) --- I agree that we should emphasise disambiguators
like {posydji}, {ge'udji} etc, though I also have a lot of time for Mark's
position, that when a gismu will do, a lujvo is unneccesary (thus: bevri
instead of djabe'ipre for "waiter"); I'm struck, though, by the realisation
I discuss in the paper, and made perhaps six months ago, that belenu-lujvo
are really transformations from a "slightly less shallow structure" of
be-lujvo, and that your soidji really corresponds to nunsoikeidji, which I
think is to be preferred as a disambiguation (in the form nunsoidji) in
this case.

I intend, of course, to give you full credit in the paper for stimulating our 
awareness of lujvo place structures, and helping us systematise somewhat
their derivation. My next couple of sections of the paper (when I get around
to them :( ) may be a touch controversial, and I'd welcome your opinions
on them.

Momenton senpretende paseman mi retenis kaj # [Victor Sadler, _Memkritiko_ 90]
   kultis kvazaux                           &  (NICK NICHOLAS. Melbourne.
      senhorlogxan elizeon                  #   Australia. IRC: nicxjo.
         (Dume:                             &   nsn@munagin.ee.mu.oz.au .)