Message-Id: <m0tGAeV-0000ZTC@xiron.pc.helsinki.fi>
Date:         Thu, 16 Nov 1995 20:01:07 +0000
Reply-To:     ucleaar <ucleaar@UCL.AC.UK>
Sender:       Lojban list <LOJBAN@CUVMB.BITNET>
From:         ucleaar <ucleaar@UCL.AC.UK>
Subject:      splitting PA into multiple selma'o
To:           Veijo Vilva <veion@XIRON.PC.HELSINKI.FI>
Content-Length: 1507
Lines: 34

Lojbab:
> And:
> >* Split PA up into several selmao.
> OK.  I'll bite.  The definition of a selma'o is that it have a distinct
> grammatical (not semantic) role.  We have kept PA as one selma'o because
> we truly could NOT figure out any substructure grammar that was
> applicable.  At one time we DID have an attempt at a substructure
> grammar - trying to build core numbers with prefixes and suffixes but 1)
> it wasn't even close to LALR(1), and 2) we kept finding things we wanted
> to say that violated the assumption that the numbers were central and
> the modifiers secondary.
> So even I am interested in what you would have in mind - not as a
> serious proposal, of course %^).

I have many lojbo hobbyhorses, but splitting PA isn't one of them.
What I had in mind is the following. For natural languages, if a
string of words is *utter* gobbledygook in every possible context, then
it is deemed ungrammatical, and the grammar must be written so as to
exclude such strings.

In Lojban, though, this standard hasn't been applied. The grammar
generates utter gobbledygook, and so is too lax by the standards of
natural language linguistics. PA is, I think, one area where we can get
utter gobbledygook.

If you turned a grammarian loose on Lojban they'd come up with a
different grammar from the official one, because Lojban has fiddled
the initial data. The current official grammar overgenerates.

Of course, what has been achieved so far in the lg description is
tremendously admirable.

---
And