4.0707 Numeric Direction (2/61)

Elaine Brennan & Allen Renear (EDITORS@BROWNVM.BITNET)
Tue, 13 Nov 90 09:34:24 EST

Humanist Discussion Group, Vol. 4, No. 0707. Tuesday, 13 Nov 1990.

(1) Date: Wed, 07 Nov 90 12:03:47 EST (20 lines)
From: Peter Ian Kuniholm <MCG@CORNELLC>
Subject: Re: 4.0699 Bi-Directional Lines

(2) Date: 7 November 1990 18:01:40 CST (41 lines)
From: "Michael Sperberg-McQueen <U35395@UICVM>
Subject: numeric direction

(1) --------------------------------------------------------------------
Date: Wed, 07 Nov 90 12:03:47 EST
From: Peter Ian Kuniholm <MCG@CORNELLC>
Subject: Re: 4.0699 Bi-Directional Lines (2/71)

Re. comments on directionality of numbers and how we allegedly took the
system over from the Arabs:

They of course write retrograde to the way we do, but they express their
n umbers "backwards," which is to say that the number 321 in Arabic is
expressed as "one and twenty and three hundred" which makes the sequence
look like 321 to us although they think of it as 1+2+3.

Every time I have to deal with an Arabic inscription, I thank them for
thi s oddity which makes reading their numbers that much easier for me.
But it is mildly disconcerting to see left-to-right numbers in the
middle of a right-to-l eft text (which doesn't bother Egyptian
first-graders one bit).

Peter Ian Kuniholm
Cornell University
(2) --------------------------------------------------------------50----
Date: 7 November 1990 18:01:40 CST
From: "Michael Sperberg-McQueen 312 996-2477 -2981" <U35395@UICVM>
Subject: numeric direction

Alan Corre is absolutely right! Our numerals run the way they do
because we got them from the Arabs, and integers, at least, run right to

Some very strange things have occurred with numbers in the transition
from integer to real, and from integer to string of numerals, of which I
mention just two:

- decimal numbers read neither left to right nor right to left,
but inward-out
- telephone numbers (which are numeral strings, not numeric
values) run left-to-right, like the text, in Roman
scripts -- and in modern Hebrew and Arabic they also
run left-to-right, *against* the textual direction
(I assume the direction was taken over with the technology,
just as it was earlier when Arabic numerals moved west)

The latter fact leads to the very strange phenomenon that in Arabic and
Hebrew, integers run the same way as the text, but numeral strings run
the other way, while in right-to-left scripts the reverse is true.
Neither script has a wholly consistent treatment of numbers, I think.

For whatever reason, however, the native speakers of Hebrew I've worked
with prefer to write telephone numbers and integers in the same order,
and it's left-most digit first.

Whichever way you look at it, it's tough on the software.

Back to the main inquiry, though: There was a very nice DOS-based
Arabic word processor based on Palantir and sold out of Washington DC or
suburban Maryland, which I looked at carefully a few years ago, but I
can't remember the name. It was far far superior to the Multi-Lingual
Scholar of the same period, but I infer that Multi-Lingual Scholar has
got past the days when the backspace key didn't work (you had to press
left-arrow and delete).

Michael Sperberg-McQueen