Message-Id: <9202150503.AA00894@relay2.UU.NET>
Date:         Fri, 14 Feb 1992 23:52:59 -0500
Reply-To: Logical Language Group <cbmvax!uunet!cuvmb.cc.columbia.edu!lojbab>
Sender: Lojban list <cbmvax!uunet!cuvmb.cc.columbia.edu!LOJBAN>
From: Logical Language Group <cbmvax!uunet!cuvmb.cc.columbia.edu!lojbab>
Subject:      lerfu
To: John Cowan <cowan@snark.thyrsus.com>
Status: RO

For things like the *tsy.bu, I would favor instead making the sound a
name - indeed for most foreign sounds that have no obvious 'word', I
would favor using names.  Thus, to me, in preference to the
possibilities suggested for the Greek, I would choose omeg.bu/omegas.
bu, etc.

>Yeah.  It's silly to worry about the meanings of words like {.iobu}, as
>per my question earlier, but the fact is we have the
>{.uibu/zo'obu/joibu/denpa bu} precedents.

Specifically on this question, and also I think regarding other
questions that have arisen in this thread, remember that we are talking
about lerfu, which are part of a >character set<.  There is no
requirement that .iobu have the same meaning in the Cyrillic
character-set as in the Latin/Lojban one.

Mark asked a question or two about Hebrew, talking about alternate
words, for a lerfu depending on whether it is hard or soft or something
like that (???).  If there is a distinct symbol, there should be a
distinct lerfu.  The reflection should be of the visual symbol, not the
sound.

If the distinction is of a diacritical variety in the visual
representation, there is an alternate (possibly preferred but that is
what you guys oughta discuss) method that I don't think Cowan's proposal
shows any examples of.  This is the assembly of multiple lerfu/symbols
into a composite lerfu.  John will please give an example of actual
usage, but you use separate lerfu for each of the components and a
composite-marker cmavo to indicate the result is a single lerfu.  I
think we thus end up with something not too different from the Unicode
approach that Guy Steele mentioned, but in particular, we have a way to
deal with such writing systems as Korean.  Applied to more well-known
situations, the ~ might be given a lerfu of its own and n+~ would be
expressed as a composite.  Likewise the various diacritics that occur
over e in French, and the hats over Esperanto letters.  Again, the goal
is to label the symbol, not the sound the symbol stands for.  If this
can clearly be done to serve the latter as well, fine.  (But "c" in the
Lojban/Roman alphabet should be "cy", even if it is pronounced
differently in most languages.  English and Devanagari ligatures should
also be handled as composite lerfu, I opine.

Another option to bear in mind, for those languages that do not have an
upper-case/lower-case, is to apply the shifts corresponding to the cases
to get alternate values.  Even more useful, there is the cmavo "tau", I
believe that 'shifts' the following lerfu only to the alternate set.

This posting is a mess.  John could you do up a more cogent summary of
the varieties of lerfu expression and post it.  Since you said this is
all intended for the lerfu paper, the people discussing this really need
to know the rest of the lerfu rules.

(Someone is encouraged, of course, to tackle the IPA lerfu set, which I
think is another one on the list that John forgot to say was needed.
Are there others that we have a shift for John, that aren't under
discussion?)

Speaking off the top of my head,
lojbab