On Thu, 6 Mar 2003, John H. Jenkins wrote:

> On Thursday, March 6, 2003, at 09:10 AM, Nicholas Bodley wrote:
> > Shortly after I invested in the Unicode 3.0 book, I started to
> > scan the CJK "corpus", including the Unified Extensions. To my
> > considerable amazement, I discovered (P. 608) U+3AB3 and U+3AC8,
> > both of which (at least in the typeface used to set the book)
> > include ovals. I'm curious why these two contain ovals, when they
> > are so rare in CJK. There must be some reason.
>
> These are actually both Korean coinages, as a quick glance at the
> online Unihan database (<http://www.unicode.org/charts/unihan.html>)
> reveals. Unfortunately, we don't have any data on them beyond that
> (meaning, pronunciation, and so on).

I think they're one of those cases of those unusual, if not borderline,
Han characters, which are a combination of 1) a Han character, which
contributes the bulk of the syllable, and 2) a Han character or hangul,
which contributes the rest of the syllable. For example, if you look at
the the characters with radical #5 (such as those starting at U+4E59 in
the URO), they seem to all end in -l. U+4E72's kKorean field says "CAL"
(was this a typo for "CHAL"?) and the readily identifiable top portion is
U+6B21, which has a KKorean field of "CHA". The final -l might be
imparted by U+4E59, which has a kKorean field of "UL", although U+4E59
does look like it could also be a fancy-looking version of the rieul
hangul letter U+3139 (initial r-, or final -l).

At the "hanguo teyong hanzi biao"[1] section of Taiwan's Ministry of
Education's Dictionary of Chinese Character Variants[2], I found a page[3]
with what I believe is U+3AB3. It's the second entry on the page, with a
reading of "dung" (transcription system not specified, but it does not
appear to be the same as that used in unihan.txt), and says its seen in
old books as a slave/serf name. i.e., it looks like U+6597 (kKorean says
"TWU", which looks like Yale romanization to me) with -ng suffixed to it.
Likewise, the first entry on the page[3] looks like it has the kiyeok
hangul U+3131 imparting a -k to it (I'm not so sure about the nature of
the last one on that page, but it is some kind of -t imparter--there are a
number of this -t type in Unicode already).

If one looks carefully, they'll see that the graphic given at
charts.unicode.org for U+3AB3 surely isn't quite right. kRSUnicode
(radical-strokes) makes it clear that the residual strokes is 2, so the
bottom part cannot be a circle, but a circle with a small stroke on top of
it, which makes it U+3181 HANGUL LETTER YESIEUNG, which the Unicode book
annotates as "archaic velar nasal". However, in modern times, both the
zero-initial hangul letter and the final -ng hangul letter are written as
plain circles, so I think this might be on the order of a font difference,
a unifiable variant, or an anachronism.

I suppose kKorean for U+3AB3 could be "*TWUNG", and U+3AC8 perhaps "*ENG"
or "*ONG", but these should be verified independently.

[1] http://140.111.1.40/fulu/fu5/kor/index.htm
[2] http://140.111.1.40/
[3] http://140.111.1.40/fulu/fu5/kor/kor068.htm


Thomas Chan
tc31@...