'Ghost kanji' lurk in the Japanese lexicon

There are plenty of dead words in languages all over the world, but can the same be said about ghost words? Floating around the murky regions of digitized Unicode values are anywhere between 60 and 100 yūrei-moji — literally, “ghost characters” — haunting the Japanese kanji lexicon.

Though academically obscure, these characters’ origins can be understood readily enough by anyone who has mistaken their doctor’s handwriting for chicken scratch (or vice versa, for that matter). All it took was a few unintentional splatters of ink and some poorly rendered photocopies to bewitch people into seeing kanji that, for all we know, should never have existed.

The origins of yūrei-moji can be traced back to the 1970s, when the Japanese Industrial Standards Committee was tasked with generating computer fonts in preparation for the coming digital age. The committee took about seven years to accumulate thousands of kanji representing every corner of Japan, using data sourced from 日本生命 (Nihon Seimei, Nippon Life Insurance Co.), 情報処理学会 (Jōhō Shori Gakkai, the Information Processing Society of Japan), 国土地理協会 (Kokudo Chiri Kyōkai, the Japan Geographic Data Center), and 旧行政管理庁 (kyū Gyōsei Kanrichō, the former Administrative Management Bureau).

What resulted in 1978 was JIS C 6226, later renamed JIS X 0208, or — if you don’t speak computer — simply the JIS Kanji Code. This code contained almost 7,000 digital characters, most of which were kanji but which also included Roman, Greek and Cyrillic symbols.

Celebrations aside, the code was inevitably revised and expanded on three separate occasions over the next two decades. It was during this period of revisions that scholars began noticing some rather peculiar kanji in the code — peculiar in that they seemed to have come from nowhere at all.

Perhaps the most famous of these mysterious kanji is 妛 (whose pronunciation is listed in some dictionaries as akebi). After a formal investigation of the JIS Kanji Code in 1997, academics found no record of 妛 in the Kangxi Dictionary (the 47,000-odd-character Holy Grail of Chinese pictographic knowledge) and JIS technicians could locate no region or family name in Japan that used it — that is, until they came across 𡚴原 (Akenbara), the vernacular name of a small district in Shiga Prefecture. Look closely and you’ll notice that 妛 consists of radicals 山, 一 and 女 while the kanji from Shiga is exactly one brush stroke less complicated.

With literally no other lead to go on, the JISC concluded that it had mistakenly registered what was supposed to be 𡚴 as 妛, in effect creating an entirely new digital character with no historical or cultural substance — a ghost of a character, if you will. It was later discovered that the local district’s method of printing 𡚴 at the time involved cutting out the 山 and 女 radicals from other preprinted kanji and pasting them together before photocopying the result. The line-like shadow produced by the overlay is what led to the committee’s incorrect reading. Mystery solved.

Interestingly, 妛 is one of the few yūrei-moji to have a confirmed albeit Scooby-Doo-esque explanation for its appearance (if it weren’t for those meddling photocopiers!). The fact is, most ghost characters have yet to be fully understood.

One such example is 彁 (pronunciation unknown). Unlike 妛, which found its corporeal counterpart in the 国土行政区画総覧 (Kokudo Gyōsei Kukaku Sōran, National Conspectus of Administrative Districts), 彁 was documented as “completely unique among the graphic characters encoded in these standards, with absolutely no means of identification.”

Common theories propose that 彁 was mistakenly de-stylized from calligraphic depictions of 謌 (uta/ka, sing or recite) or 彊 (kyō/tsuyo(i), strong), or that it was a variant of 歌 (uta/ka, sing), or even that it was supposed to be 哥 (an older iteration of 歌) but was misread together with a splattering of ink on its left side. In any case, this character simply did not exist before 1976; yet now, inexplicably, here it is, puzzling historians like a mischievous poltergeist.

Despite another two decades having passed since the discovery of yūrei-moji, they can all still be generated using standard fonts in most word processors via Unicode input. The reason for this harkens back to the first JIS Kanji Code revision in 1983. These revisions added some new characters, modified roughly 300 kanji to depict simplified forms, and swapped some kanji around as per the request of the education ministry.

Unfortunately, due to the limited functionality of early-1980s computers, these revisions caused massive technical problems with international Unicode compatibility between both Chinese and Korean computers. So, when the 1997 investigation shed some figurative light on these spectral symbols, the JISC decided that it would be far less of a headache to just register the correct kanji in new code and allow the yūrei-moji to remain with their original Unicode values, awaiting the day when an overly enthusiastic translator writes an article about them.

And what practical use is all this to the modern Japanese learner, you ask? Well, not much, to be honest. If you’re taking the Japanese Language Proficiency Test anytime soon, I’d say you’ll be just fine sticking with the un-haunted study guides. But for those like myself who love diving down the rabbit hole, these kanji are certainly interesting in their own right.

Looking at the bigger picture, yūrei-moji provide us with unique tidbits of industrial and academic history, potential stories behind lesser-known regions or household names, or even just a brief look into the evolution of Japanese typography. And given the fact that somebody, somewhere, decided to call these characters yūrei-moji, it really does drive home just how much Japan loves its ghost stories.