Message-Id: <199311041408.AA03328@eli.CS.YALE.EDU>
Date:         Thu, 4 Nov 1993 14:09:49 GMT
Reply-To: Richard Kennaway <jrk@SYS.UEA.AC.UK>
Sender: Lojban list <LOJBAN%CUVMB.bitnet@YaleVM.YCC.YALE.EDU>
From: Richard Kennaway <jrk@SYS.UEA.AC.UK>
Subject:      Re: TECH: more thoughts on zi'o
To: Erik Rauch <rauch>
Status: RO

Arthur Protin writes:
>All travel has a destination,
>regardless of anybody's inability to express it.  The same is true
>of all the other places of klama.

When klama is used, there is a destination, but this is simply because the
destination is part of the definition of klama.  I don't know what people
mean when they talk about the essential parts of a concept, such as a
destination being essential to the concept of travelling.  It is either
part of the concept or it is not.  One concept relating to motion may
include a destination; another may not.  How one chooses and shapes the
concepts to make primitive words for is a difficult question, since the
relevant decisions are at a finer level of detail than sources like the
Eaton list give information about.

>I accept "P(a,b,c,d,zi'o)" as "there is no e such that P(a,b,c,d,e)"
>provided that zi'o is always required to be stated, never implied.

My understanding of zi'o is that zi'o-ing the destination place of klama
leaves intact the assertion that a certain mover, origin of movement,
route, and means of transport are related in a certain way, something which
is not implied by "there is no e such that P(a,b,c,d,e)".

>The construct "P(a,b,c,d,...)" says nothing about the existance or
>relevance of e unless the the relation P is such that e is guarenteed
>to exist, as is the case with klama.

My understanding of omitted places is that they are considered to be filled
by something which the speaker has not specified, and are semantically
identical to filling the place with zo'e.  All the places of a predicate
are asserted to exist, by definition, unless replaced by zi'o.

>Even when e is guarenteed to
>exist, when e is unspecified nothing more is implied about it.
>The construct "not P(a,b,c,d,...)" says to me that there exists
>a value of e such that "not P(a,b,c,d,e)" is true.

No, it asserts that for some particular but unspecified value of e,
P(a,b,c,d,e) is false.  Omitted or zo'e-d places are not existentially
quantified over.

>This is a much
>weaker claim than there is no e such that "P(a,b,c,d,e)" is true.
>The stronger claim is represented as "P(a,b,c,d,zi'o)" which is
>where this paragraph began.

This is also not correct.  That stronger claim would have to be represented
by using an explicit quantifier, not zi'o.  "P(a,b,c,d,zi'o)" asserts a
relationship among a, b, c, and d which, as Robert Chassell said in this
thread, is not determined in any algorithmic way by the five-place relation
(and presumably must be listed separately in a dictionary).  It is actually
a different relation, having certain features in common with the original
one, but not formally related to it.

>(Tell me again John why the loglan predicate "bluer-than" is
>impossible to use with negation.)

In Institute Loglan, omitted places were at one time considered to be
existentially quantified over.  (I don't know whether this is still the
case.)  The negation of "x bluer-than" would be "it is false that there
exists a y such that x is bluer than y", which is probably a stronger claim
that intended.  This is shown more clearly by e.g. "slower-than".  "It is
false that there exists a y such that x is slower than y" means not that x
is faster-than something (which one might have wanted it to mean), but that
x is at least as fast as everything (which one would rarely wish to
express).

--                                  ____
Richard Kennaway                  __\_ /    School of Information Systems
Internet:  jrk@sys.uea.ac.uk      \  X/     University of East Anglia
uucp:  ...mcsun!ukc!uea-sys!jrk    \/       Norwich NR4 7TJ, U.K.