Systematic analysis of the Japanese lexicon (specifically, Kanji)

Question

I'm wondering whether anyone has carried out an analysis of the Kanji (or Hanzi), the traditional Chinese characters used in Japanese and some dialects of Chinese.

For example, I wonder whether anyone has put in the effort to annotate the various typographical and semantic links between elements of the lexicon. For example, a 'typographical' link would involve e.g. some shared radicals in the ideogram; a 'semantic' link would e.g. link a 'composite element' (that is, one composed of several ideograms) to each of its constituent ideograms, which would themselves already be considered elements of the lexicon even if they never appear alone.

Basically, I am looking for a WordNet-like construction that places the 'typographical' level at the forefront of the investigation.

Answer 1

I assume you're looking for an electronic database (as opposed to the rich tradition of character analysis in paper dictionaries and such). KanjiVG is based on character appearance and includes data on the components of each kanji; it's easy to parse the XML to do things like finding all kanji with a given component, etc. Sorry for the self-plug, but I once used it to estimate the predictive power of Japanese phonetics components; perhaps the data files in that post can prove useful to you in some way.

There's also Wenlin's Character Description Language, but it's non-free.

The individual role of components – which may be semantic, phonetic, or both – is much fuzzier and harder to analyze with certainty, and many competing proposals exist. For character-level information, you can use Kanjidic and Unihan. There are some online resources which deal in decomposing characters; e.g. in English, Sear's or Zhongwen.com (I have a simple tool to compare such sites; see the documentation for details). I've tried to contact these authors to ask if we're allowed to mine their sites and use the data, but got no answer.