Message-Id: <199512230550.HAA08298@xiron.pc.helsinki.fi>
Date:         Sat, 23 Dec 1995 00:49:08 -0500
Reply-To: Logical Language Group <lojbab@ACCESS.DIGEX.NET>
Sender: Lojban list <LOJBAN@CUVMB.BITNET>
From: Logical Language Group <lojbab@ACCESS.DIGEX.NET>
Subject:      Re: response to Steven Belknap on language baselines and
              stability (long)
To: Veijo Vilva <veion@XIRON.PC.HELSINKI.FI>
Content-Length: 2385
Lines: 46

>Logical Language Group writes:
>>Because the grammar of whatever is non-standard is as unrestricted as the
>>capability of non-standardness, there is no cmavo that could unfailingly
>>cover the territory.  za'e is pretty restricted and does not solve any
>>grammar problems.
>
>Sure, but I'd bet one could be devised which would cover *most* of the
>territory.  The rest could be managed with lo'u/le'u (which could take
>the subscript describing the variant, if needed).
>
>I'd also bet that the variants which *can't* be expressed with something
>like za'e, but in UI, would never be accepted into the language, anyway.


za'e is restricted to emphasizing the following word - and most importantly
doesn't say ANYTHING about what selma'o the following word is.  Thus, to
use it for anything other than a brivla (which is defined as being of a
particular selma'o by its morphology) means that the parser will not have
any idea what to do with the nonce word - it will have an unassigned selma'o.

There could perhaps be a metalinguistic technique deivsed that would assign
a nonce word to a selma'o (the parser would not handle this as is, but it
would certainly fall within the language definition parameters - say something
like za'e'e .ui xo'o would assign xo'o to the same selma'o as "ui" for nonce
use.  This would be a variation on "za'e" that had an extra word of some type
before the actual to-be-hioghlighted word.

This was cover any and all experimental cmavo that fit within existing
selma'o.  You can also do something like this now using "sei zo xo'o cu
cmavo be zo .ui", so we don;t actually need a za'e'e, though the
za'e'e would make it easier to flag such things for a parser to process
them.

But for anything that does NOT fit in an existing selma'o (as I presume
And's revolt against du'u would be), there simply is no way to tell the
parser how to process the grammaro fo the word, so you just
have to tell the parser to not process the sentence, and that is best done
with lo'u/le'u.

Itis unlikely that ALL proposals will fit solely within existing selma'o,
and I think the proliferation of recent proposals shows that grammar
experiments WILL be tried during the baseline period.  Of course you are
right that they have low chance of acceptance unless they can be made to
YACC (or unless we lose and they catch on in spite of it not being YACCable).

lojbab