Message-Id: <199310100018.AA04736@eli.CS.YALE.EDU>
Date:         Sat, 9 Oct 1993 17:19:26 -0700
Reply-To: jimc@MATH.UCLA.EDU
Sender: Lojban list <LOJBAN%CUVMB.bitnet@YaleVM.YCC.YALE.EDU>
From: jimc@MATH.UCLA.EDU
Subject:      Re: deleting places
To: Erik Rauch <rauch>
In-Reply-To:  Your message of "Fri,               08 Oct 93 11:38:00 EDT." <9310081543.AA16152@julia.math.ucla.edu>
Status: RO

Art Protin <protin@USL.COM> writes:
> 4)      What about all the attachable places.  By the logic that every
>         place is fundamental to the understanding of the concept
>         that a gismu embodies, the attachment of another place once
>         means either that the gismu embodies two concepts or that
>         all uses of that gismu have that attached place.

By attachable I assume Art means attachable with <BAI>.  What is the
current thought on "predicate Hilbert space"?  There are potentially
infinitely many lujvo (even lujvo that are credibly relevant to a
prespecified selbri, and not to mention the numerous non-lujvo selbri),
and each one of them can be transformed to a
sumti tcita using fi'o, and so infinitely many BAI places can be added
to a selbri in this way.  It makes my head ache to think that every
(Lojban) relation has infinitely many places even if most are elided.
It makes my head ache enough to want to declare that BAI places are
of second class -- my favorite transformation is to say that they are
merely abbreviations for restrictive subordinate clauses.  What say the
people who do heavy logic?

> 5)      Given that lojban is to be a human language, defined by usage
>         and described by our texts (as verses prescribed by them),
>         I believe that the rigid place view will have to be abandoned.
>         I believe that a property of teaching language by example
>         is that the place structure will have to be somewhat looser
>         and people will learn gismu initially as having minimal
>         place structure.  The less commonly used places will be
>         learned later as enhancements of the base concept.
>         While this may violate the paradigm of predicate calculus,
>         its either that or limit lojban to a second language, as
>         mathematics is.

I don't think the situation is that bad.  The "real" definitions in the
dictionary can be logically self consistent, but certainly teachers will
not bother to teach every nitpick detail in the first lesson.  If the
student doesn't know about a place that the expert does, he may miss
nuances of meaning and may see logical inconsistencies that are unreal,
but that's not a blemish in the language; it's a reflection that the
particular student hasn't finished the lesson.

                -- jimc