Message-Id: <9207301236.AA02948@relay1.UU.NET>
Date:         Thu, 30 Jul 1992 06:17:35 -0400
Reply-To: Logical Language Group <cbmvax!uunet!grebyn.com!lojbab>
Sender: Lojban list <cbmvax!uunet!pucc.princeton.edu!LOJBAN>
From: Logical Language Group <cbmvax!uunet!grebyn.com!lojbab>
Subject:      response on Cowan on rafsi proposals
To: John Cowan <cowan@snark.thyrsus.com>
Status: RO

Only one off-the-cuff comment.

I do not believe that assigning new rafsi after the dictionary is put out
(or adding gismu for that matter) is any more acceptable than deleting them.
They are both violations of baselines which will be more firmly established inn
the public mind than our release into public domain has caused.  A gismu or
a rafsi not in the dictionary, and not deducible from its contents (i.e. my
the lujvo making algorithm, or le'avla rules) is just as unsettling to a new
user as a word that appears in a context to mean something different from its
definition in the dictionary, which is the effect of a rafsi or gismu change.
Indeed, I would suspect that we are used to the latter process in natural
language, and the non-baselining of place structures will mean that we had
better remain used to it for Lojban.

(I think that the release into the public domain has NO implications re
baselining, based on the place structure example, among other reasons.
If I had thought that release into the public domain would imply any de
facto baselining, I would have opposed any such thing.  still, John, also
chose to reject this argument, so I won;t dwell on it - but please no one
else use it, or my attitude on a very important issue may suddenly get
rather sour.  I don't think we should be forced to choose between our ideals
on public access to the design and our principles of what we believe the
design should live up to.)

But I want to make clear that changes to the baselines will become MUCH more
difficult after a book is published.  My default vote on all issues will be
an automatic NO with high thresholds for considerations of change.  Actually,
I will say that this first book is rather less critical in this way, in that
we have called it a working- or proto- dictionary to be supplanted by a later
"real" dictionary that will coincide with the{ hard 5t(-year baseline.
But I still think that each publication of a baselined design item sets it more
in concrete.  I do not feel as firmly committed on things explicitly left
unbaselined, because otherwise we rigidify the language too early in the
aspects where human usage is most critical to making the language come alive.

lojbab