Message-Id: <199308221834.AA11737@eli.CS.YALE.EDU>
Date:         Sun, 22 Aug 1993 14:33:10 EDT
Reply-To: Logical Language Group <lojbab@GREBYN.COM>
Sender: Lojban list <LOJBAN%CUVMB.bitnet@YaleVM.YCC.YALE.EDU>
From: Logical Language Group <lojbab@GREBYN.COM>
Subject:      Re: TECH: QUERY on ZI & ZEhA
To: Erik Rauch <rauch>
Status: O

Your summary seems quite reasonable, Jorge, and may be correct (there are
some border conditions that can screw up most generalizations about
the termonators, but I can't think of any right off hand that affect what
you said.  The best bet when you aren't sure is to include the
terminator.  It is never wrong to include a termoinator, sometimes is
helpful to a reader less skilled in the grammar, and sometimes also
is necessary.  The basic "rule" covering all terminators is that, if
you omit the terminator, does the next word appear like a plausible
continuation under ANY grammatical interpretation as yet unterminated or
unexcluded.  This is only agrammatical, and not a semantic constraint -
oftentimes the word slopped in makes for perfect nonsense and the normal
human Lojbanist will really want to stick in the terminator that was
mistakenly omitted if only to make some sense out of it  (on the other hand,
some Lojbanists like me, seem to be willing/able to find more sense in such
nonsense, and hence be less tolerant of mistakenly omitted terminators).

Note also that since the grammar is LALR1, you cannot look further ahead
and see that later grammar makes it clear which was intended.  The best
example of this is the non-logical connectives, which almost always require
the previous sumti to be terminated with ku.  After "le broda", JOI means
that you are logically connecting onto
"broda" and not "le broda", and this remains true even if the second token
following is another "le", making that interpretation impossible.  If the
grammar was LALR2 with error correction, the KU after "le broda" would
be unnecessary if a "le" follows, but that isn;t how we designed the grammar.

lojbab