At 19:09 -0500 2003-12-12, Peter T. Daniels wrote:
>John Jenkins wrote:
>
> > What you're saying is that there is a single Japanese "script," which
> > consists of three (or more) components: kana, kanji, romaji, etc.,
> > nicht wahr? And that the overall set of characters used in the overall
>> Japanese script is not interchangeable with the overall set of Korean
>> characters, and so on?
>
>They don't look alike; it's probably more inappropriate to use a Chinese
>font to set Japanese (never mind that there are some Jpn. characters
>that don't exist in China) than it is to set English in Fraktur, or
>German in romain du roi.

It is true that the specific way Han characters are drawn is very
important to the users of those characters. Setting English in
blackletter (not to say German Fraktur) can done in wedding
invitations, but no Japanese would use Chinese fonts if he could
avoid it. The shape preferences for users of CJK are quite strong.

> > Or are you saying that the set of kanji used in Japanese, hanzi used in
> > Chinese, and hanja used in Korean are not interchangeable?
>
>They're simply not the same. They share a perhaps sizable core group of
>characters, but they don't look alike and they don't sound the same.

There are some 70,000 CJK characters encoded in Unicode; no language
uses all of them. There are characters unique to Viet Nam, unique to
Singapore, and so on. Of course most of the 70,000 come from old
Chinese dictionaries and are recognized by only a very few
specialists. All of the client languages use a subset of the unified
set as has been encoded. Historically, elaboration and preservation
of word/character lists has been important to all of the users of
this script. The Unicode Standard is a superset of all previous work.
This is significant.

>I trust we no longer have the typewriter problem of using the same
>character for one and ell, for zero and oh; for Russian <n> and sm.cap.
><h>; etc.

In Unicode, we no longer have this problem.

> > If the latter, then I must confess I find the conclusion rather
>> remarkable, as it's rather the opposite of the general impression of
>> people who live in East Asia, barring anti-Unicode rhetoric. While
>> there is some difference in the precise set of kanji/hanzi/hanja used,
>> and some difference is the way they're written, the fundamental
>> identity is rarely questioned.
>
>Perhaps that's because they're ordinary people, with
>ordinary-people-intuitions. Recall how ordinary-people-intuitions about
>language very often bear little relation to the fruits of linguistic
>investigation.

That's also been recognized. People on the ground want to have their
glyphs the shapes that they want. They don't care whether the
encoding unifies them with other glyphs or not. This has been a
challenge for Unicode, but font technology is meeting the challenge
and the requirements of the ordinary people.

>As with the typology, I'm saying that maybe _every_ term doesn't have to
>be tried to be applied to _every_ case. Maybe "script" isn't a useful
>term in discussing Japanese at all, since Japanese is so sui generis.

I think we're back to misunderstanding terminology. The Japanese
language has a particular orthography, which makes use of the
Katakana syllabary, the Hiragana syllabary, a subset of the huge
number of CJK characters which exist, and occasionally of the Latin
alphabet and other symbols.

>It's the computer engineers who insist on utterly dividing up the
>universe into watertight compartments such that every entity has its
>very own assignment, and there are no empty areas and no overlaps.
>That ain't the way human minds work.

Sure it is. Hiragana and Katakana and Kanji and Romaji have different
names because they act and interact differently in Japanese
orthography.
--
Michael Everson * * Everson Typography * * http://www.evertype.com