From LOJBAN%CUVMB.BITNET@uga.cc.uga.edu Mon Nov 14 03:50:43 1994
Message-Id: <199411140850.AA06447@nfs1.digex.net>
Date: Mon Nov 14 03:50:43 1994
From: Gerald Koenig <jlk@NETCOM.COM>
Subject:      lo, da poi
Status: RO

I continue to believe that "da poi" is not a substitute fo "lo" in all
contexts.  Here is another illustration:

1). re lo ci gerku cu blabi
Exactly three dogs exist, two are white.

2). re da poi ci gerku cu blabi         (putting da poi for lo)

Parser output:

1'). ({<re BOI> <lo [(ci BOI) gerku] KU>} cu {blabi VAU})


2').({<[re BOI] da> <poi [({ci BOI} gerku KU) cu (blabi VAU)] KU'O>} VAU)

The first," lo" sentence I am able to express completely  in first order
predicate calculus with equality. If there is interest I will post it.
It translates cleanly into English too.

The second "da poi" sentence I am unable to express meaningfully in pc.
In English I come up with something like " there exists an x, there
exists an x, which three dogs are white"  Neither is there any
structural parallelism in these two parses.  Unless someone can show me
how to translate 2). into pc, and show that it is equivalent to the pc
version of 1), I am opposed to letting lo (everywhere)=da poi, and
especially opposed to tweaking the grammar to make it work.  The more
the grammar becomes detached from predicate calculus, the more lojban
will lose stability and usefulness as a language that can interface
cultures and computers in the 21st century.

djer