Home » Blogs » kazarena's blog

Traditional Chinese vs modern Japanese Kanji forms

Submitted by kazarena on Thu, 2009-01-08 22:19

I was reviewing smith's singatures and names in Kanji the other day in order to figure out few things. There is a terrible mix at the moment between traditional Chinese and Japanese characters. While I tried to use traditional Kanji for the names, signatures use all different forms (e.g. I changed all 国 (as for kuni) to 國 in the names, but plenty of mei still contain 国. Same applied to 'hiro': 広 and 廣).

There are different conventions in different sources:

  • Toko Taikan uses simplified Japanese forms for smiths' names.
  • Yamanaka uses Japanese forms.
  • Fujishiro uses Chinese forms.
  • Hawley uses Chinese forms.

Both ways make sense, but with sightly different emphasis. Names in Japanese written for a modern reader would use Japanese forms (native Japanese readers, please forgive my educated guess). Names with Chinese characters would be easier to use for identification as signatures would be in vast majority written in old Chinese script.

Moreover, depending on the choise of Japanese vs. Chinese forms appropriate fonts must be chosen. Currently I'm using MS Mincho as the main font for Kanji. However, MingLiU font would be more appropriate for Chinese forms, and SimSun (I guess) - for Japanese.

I think I'll keep using Chinese forms. I also tried different fonts and MS Mincho would do fine for the moment, however I'm seriously considering MingLiU.

Mixing different forms becomes a problem for my automated verification tools though. I need quality conversion tables as it would be impossible to fix everything manually. In search for those tables I found that CJK Dictionary Institute provides them, not for free though. They offer premium quality vocabularies of words, names, orthographic variants, synonyms, Kanji characters and many more. You may also find a very relevant article by Jack Halpern there.

Then I looked into my favourite Jim Breen's and Monash University resources. I also wanted to find a more detailed list of Kanji, as the one I'm using (1864 characters) isn't comprehensive enough to deal with smiths' signatures.
#1Submitted by saipan59 on Thu, 2009-11-12 03:59.

kazarena wrote:I also wanted

kazarena wrote:
I also wanted to find a more detailed list of Kanji, as the one I'm using (1864 characters) isn't comprehensive enough to deal with smiths' signatures.

You probably already know this, but the kanji database in ToShoW has about 3000 kanji in "JIS Level 1", and almost as many in "JIS Level 2". Level 1 includes data on ON/KUN readings, stroke count, Nelson number, etc. They are in order by reading. Most of Level 2 only has the JIS number.

For Level 2, the kanji are in order by radical. The file "LEVEL2.TXT" indexes the first kanji for each radical, to provide a way to find one that you want. Then you will know the JIS number, which can help you find it using other references.

At the very end of the Level 2 kanji are a few that I entered myself - they are variations that are too obscure to be in the JIS list, but are found on sword signatures. For example, kanji # 6931 is "SUKE" from Sukesada. The standard/modern version is # 3270.

The kanji fonts in ToShoW were imported from a font file that I got from a former employer over 20 years ago. So they were based on an 'official' JIS list. The raw data used in ToShoW (in the KANJIS.DAT file) is not in any standard format, but it is not compressed or encrypted, either.

Pete
#2Submitted by kazarena on Thu, 2009-11-12 10:31.

KANJIDIC

Hi Pete,

Thanks for the information. I have KANJIDIC/KANJD212 set now (you may find it through the last link in the original post). This is the most comprehensive database of Kanji in the internet as far as I know, with full list of references, radicals, examples etc.

The point about traditional vs. modern versions of Kanji is more interesting. I reviewed all the Kanji in smith names about a year ago using a set of automated tools about a year ago. One of the decisions I had to make is what version of Kanji to keep there (either modern or traditional). E.g.  国 vs. 國 for KUNI. Usually when I have questions like that, I follow a 'look how Japanese do it' principle. In this case though, there doesn't seem to be a common agreement with different sword related books following different conventions. I went for 國 because this is the form you would most likely see in the original signatures (which may be more helpful when reading them). I didn't strictly follow the same preference (towards traditional forms) in other cases, only trying to make sure that the usage of a particular form for particular character is consistent across the whole database.

Stan