Re: Response to Steve Rice

["Arthur W. Protin Jr." (GC-ACCURATE) <cbmvax!uunet!PICA.ARMY.MIL!protin>]

Folks,
    (When I read the posting from Steve Rice, I so wanted to respond,
but restrained myself.  But Rice bashing seems like too much fun to
stay out of.  (Steve, the only comment I have specifically for you
and your actions is: Learn to put a subject on your postings!)
Seriously, I will take my shots at what Steve said, but the targets
will frequently be his colleagues.)

    Steve said
>     Identities:  You seem to be committing ther same mistake as some
> Transformational-Generative people I've encountered:  confusing a
> model of grammar with grammar itself.  Loglan is a real language, so
> it and its grammar are found not on some computer but in the minds of
> Loglanists.  You might as reasonably confuse a sqare and a rectangle:
> all square are rectangles, but not all rectangles are squares.  If an
> utterance is grammatical, it will parse; but not all that parses is
> grammatical in terms of The Grammar as known by speakers.  Otherwise,
> you don't have a language, just a linguistic computer game.  (Let me
> know whether you win or the Klingons do.)

    This sounds real bogus to me, like the sort of garbage that kept
the linguists from dealing with the problems introduced by computers,
so that mathematicians had to fill the void.  I am sure that some will
insist that the grammar of "human languages" is to be found in the
minds of its speakers as an excuse to avoid new and difficult work.
Anybody with any knowledge of child development will immediately realize
that while the grammar of a language is developed in the child, the
reference for the speaker is the reaction of the listeners and their
acknowledgement that what was said is grammatical.  As we get older
and have the language skills to bootstrap from, the literate speakers
will frequently rely on reference texts to determine the fine points
of the grammar.  While it may be accepted in linguistics to allow
the things that parse according to the grammar to still be ungrammatical,
in exact sciences, such as mathematics (in particular "Automita Theory"),
the real grammar of a language must allow all grammatical utterances
to be parsed and none of the ungrammatical utterances to be parsed!
Maybe Steve has been working with approxiamations for too long.
    One of the stated objectives of Loglan is to be "syntactically
unambiguous".  This is a very clear and precise requirement.  Any
such grammar must not only completely distinguish between what is
grammatical, each and every grammatical utterance must have one unique
parse.  A well established practice for proving that requirement is
to derive a parsing algorithm from the statement of the grammar and
if that algorithm is deterministic then the parse is unique and the
grammar is syntactically unambiguous.
    When such a complete and definitive source of reference is available,
it is perfectly reasonable for some human speakers to use it (the computer
implementation of the parsing algorithm) and to rely on it more so than
the reactions of their listeners (as many of us now rely on multiple,
conflicting texts).  This can not be a real obstacle for linguists
to study Loglan/Lojban, it can only be a very lame excuse.

    thank you all,
    Arthur Proin


Arthur Protin <protin@pica.army.mil>
These are my personal views and do not reflect those of my boss
or this installation.
Additional disclaimer: These views are known to be at least a little
askew (if not totally at odds with) the current official view of LLG.