Message-Id: <199602082040.PAA14554@locke.ccil.org>
Date:         Thu, 8 Feb 1996 12:06:54 -0800
Reply-To: Gerald Koenig <jlk@NETCOM.COM>
Sender: Lojban list <LOJBAN@CUVMB.BITNET>
From: Gerald Koenig <jlk@NETCOM.COM>
Subject:      Re: TECH: fuzzy logic
To: John Cowan <cowan@LOCKE.CCIL.ORG>
Status: OR
Content-Length: 2347

>X-cc:         lojban@cuvmb.cc.columbia.edu
>To: Gerald Koenig <jlk@netcom.com>
>Status: R
>
>The argument that Cowan's proposal is "idiomatic" loses me.  Of course it
>is "idiomatic" - that is what "convention" means.  We have an expression
>set that is plausible in the language and which has no well-defined meaning,
>nor is it at all in use, so we define it to mean something useful to some
>people.
>
>As to whether "idiomatic" usage is appropriate in the logical portions of the
>language - well scalar negation isn't the strongly truth-functional
>logical portiuon of the language (though perhaps, Cowan intends that th
>convention be used on ja'a as well as je'a).
>
>But if anything, the recent discussion should have made clear that most of
>what we call logic is a matter of convention.  We have two contradictory
>assumptions as to what "ro" might mean, and djer has observed that math
>uses on and pc has observed that logicians use another.  Whatever gets
>decided is there a "convention" and hence in a sense an idiom.
>
>lojbab
>


I am using "idiomatic"  in the third sense of webster:

"3. a phrase, construction, or expression that is recognized as a unit
in the usage of a given language and either differs from the usual
syntactic patterns or has a meaning  that differs from the literal
meaning of its parts taken together. (Ex.: not a word did she say)."

What this means for lojban that it becomes impossible to translate
lojban using a dictionary alone.  It creates a requirement for a list
of idioms, just like slang.  For example, (if I remember this right,
and I don't think I do), "qu'est ce que c'est un horloge"  means in
french, "what is a clock?' but if you try to translate it word for word
it is doesn't compute.  I am trying to exclude idiomatic expressions
from lojban, especially in the quantifier area. I believe that in the
quantifier area words should not differ from their literal meaning, and
that it should be possible to understand the whole from the parts.  I
believe that by restricting language change by fiat, there will be a
proliferation of idioms, rendering the language far more difficult
to acquire.

I don't see "ro" with or without existential import as meeting the
definition of an idiom.  It's just a matter of definition, and as long
as this is clear, there should be no problem.


djer