Received: from access1.digex.net by nfs1.digex.net with SMTP id AA01332
  (5.67b8/IDA-1.5 for <lojbab@nfs1.digex.net>); Tue, 2 Aug 1994 05:48:20 -0400
Received: by access1.digex.net id AA06909
  (5.67b8/IDA-1.5 for lojbab); Tue, 2 Aug 1994 05:48:17 -0400
Date: Tue, 2 Aug 1994 05:48:17 -0400
From: Logical Language Group <lojbab>
Message-Id: <199408020948.AA06909@access1.digex.net>
To: jorge@PHYAST.PITT.EDU
Subject: Re: current cmene project
Cc: lojbab@access.digex.net, lojban@cuvmb.cc.columbia.edu
Status: RO
X-From-Space-Date: Tue Aug  2 05:48:22 1994
X-From-Space-Address: lojbab

JL> Is there any policy on this? How can you tell that {tcadrlondono} is a
JL> culture word? "tcadr-" suggests that it is a city. Isn't it better to
JL> assume that the place structure of a fu'ivla of this type is given by the
JL> place structure of the leading gismu? For the London "culture word" I
JL> think {britrlondono} is much better.
JL> 

The classifier word is that - a classifier.  With names of cities, you could
indeed do this, but "cmacr-" on mathematical terms should not give you the
place structure of cmacu.

My reasoning is based on my "types of fu'ivla" 

If a name is a type-1 fu'ivla, then turning it into a brivla using "me"
gets you the "pertains to ... in aspect ...
Likewise if it is a type 2 (a Lojbanized name)
So the default for type 3 would also seem to be the same, though we can
establish conventions for particular classifiers.  Type 4 fu'ivla, which
should be based on significant Lojban usage, should have individualized
place structures.

Your suggestion on final vowels sounds fine to me.

lojbab