Message-Id: <9206180906.AA09012@relay1.UU.NET>
Date:         Thu, 18 Jun 1992 10:02:06 BST
Reply-To: CJ FINE <cbmvax!uunet!bradford.ac.uk!C.J.Fine>
Sender: Lojban list <cbmvax!uunet!pucc.princeton.edu!LOJBAN>
From: CJ FINE <cbmvax!uunet!bradford.ac.uk!C.J.Fine>
Subject:      Re: Metrology and exact/approximate numbers
To: John Cowan <cowan@snark.thyrsus.com>
In-Reply-To:  <no.id>; from "John Cowan" at Jun 16, 92 1:18 pm
Status: RO

>
>
> Earlier this month there was much talk about whether 0.9 is equal to 9/10
> or not, and what the significance of Lojban numbers is when measurements
> are involved.  Here's my current position on the matter, on which I solicit
> comments:
>
> 9/10 (sofi'upano) has the same value as .9 (piso), because both are exact
> numbers.  We do have the cmavo "ji'i", which has the same grammar as a digit,
> and is used to indicate approximation.  Current examples are
>
> 	ji'i vo no
> 	approximately four zero
> 	about forty
>
> 	re pi ze re ji'u ma'u
> 	two point seven two approximately positive-sign
> 	2.72 (rounded up)
>
> 	re pi ze pa ji'i ni'u
> 	two point seven one approximately negative-sign
> 	2.71 (rounded down)
>
> Based on these examples, it seems to me that if we want to say "3.1418" and
> indicate that the last digit is an approximation (in other words, our
> measurements are only accurate to .001 precision), we can say:
>
> 	ci pi pa vo pa ji'i bi
> 	three point one four one approximately eight
>
> Comments?
>
> --
> cowan@snark.thyrsus.com		...!uunet!cbmvax!snark!cowan
> 		e'osai ko sarji la lojban
>

Your exposition of explicit approximation in Lojban is clear and
admirable. My problem is that, except in rather special circumstances, a
precise number like 2.72 has no validity. (The exceptions are when we
are talking in quantised units, such as rupnu, and when we are talking
about the highly abstract concept of pure numbers). Every other use of
2.72 *in any language* has a precision, and I wish this to be recognised
in the definition of Lojban.
 I am not saying that we have to go about expressing the precision
everywhere, any more than we do in English; and we have several ways of
doing so if we wish to. All I am looking for is an admission in the
definition of the semantics of numbers that there is an implied
precision - and I suggest that it be the scheme I have already outlined.

			kolin