kapalama
Member
Registered: 2008-03-23
Posts: 183
穃 （wheat plus Contain) seems to be one of those characters that ended up in the JIS list even though it is not a Japanese character. (Google 'Ghost Character' for other examples. Or if you can read full on Japanese the Japanese Wikipedia 幽霊文字 article covers it).
Is 穃 even a Chinese character? It apparently has a reading of ヨウ to match the on- reading of 溶ける and 内容 I guess.
Also、and unrelated, 保栄茂 reads as びん? the exact opposite of a Ghost character maybe. Is that one just an artifact of Okinawan being represented by Kanji?
Jarvik7
Member
From: 名古屋
Registered: 2007-03-05
Posts: 3946
My paste is also from kanjigen but doesn't mention anything about being a miswriting. It may be the case that both characters are correct in that they were both used simultaneously for a time, with one eventually winning out.
I think most of the ghost characters remain unknown because no one (with access to the source documents) cares enough to look into it. If something is determined to be an error it's too late at this point to remove it anyways, since it would break compatibility.
Last edited by Jarvik7 (2010 February 27, 8:01 am)
I wrote a paper on ghost characters; there's a lot of detailed information on them in the JIS Kanji Dictionary. They actually did a pretty detailed investigation of the ghost characters and were able to confirm most of them. I don't seem to have the paper on this computer, but there is only one character that can be confirmed to be a mistake (妛; the line in the middle was actually a shadow), one character that is almost certainly a mistake (彁), and only 2 or 3 others that are highly likely to be mistakes.
Aside from the JIS dictionary's report of the investigation, some other people have done independent investigations and cast some doubt on other characters -- despite what the JIS investigation said, it's highly likely that 粫 is a mistake for 糯, for instance.
kapalama
Member
Registered: 2008-03-23
Posts: 183
Jarvik7 wrote:
I think most of the ghost characters remain unknown because no one (with access to the source documents) cares enough to look into it. If something is determined to be an error it's too late at this point to remove it anyways, since it would break compatibility.
If the only place they exist is on an isolated source document (which is subject to that document's writer's error), isn't that then exactly what a ghost character is?
It may not have been the JIS people's transcriber error, but rather the transcriber error of someone else. Th articles I read seemed to want to claim that the ghost charcters were a result of the JIS's effort; but isn't it just as likely (or even more likely) that an earlier person could have made the same trasncription error? The JIS errors just tend to get noticed because word processors make picking them easy. I imagine that there are a bunch of character pairs that came about because of errors、where the writer knew the character parts but transposed them.
If it was for instance a place name (or a person's name) , then it would have shown up somewhere else, wouldn't it?
