The ``radical'' of a character is more or less its ``base component.'' For example, yama radical is a character and a radical, but appears as a component radical in characters such as:
<list of characters>

Sometimes the radical appears different depending on how it's used. The character hito radical is a radical, but can also appear like ninben radical . Some examples:

<list of characters>

Actually, those last two are different versions of the same character, and their radical really is hirabi radical (which is, be careful now, different from hi radical ). When the character was changed from the hirabi radical component was lost from the character, but it still remains as the traditional radical, but many dictionaries list them both under hito radical . Is it ``correct''? I dunno' -- it's not like science with an absolute truth.

Sometimes the differences among forms is larger. The radical mizu radical usually appears as the left part of (i.e. the part that's doesn't look like a cross). Some examples:

<list of characters>

More Illogicalness

Due to the historical (like, 2000 years) nature of kanji, there are lots of illogical things about radicals. For example, the two radicals kuchi and kunigamae radicals appear almost the same. Actually, there are two distinct radicals that do appear exactly the same (the tsuki or nikuzuki radical -- take your pick part of
<list of characters>
the first five have one radical, the second five have the other).

Some characters don't even physically contain the radical anymore, due to changes over the years. The from above is one example. Another is kuru , whose traditional radical is hito radical because the old form of kuru is . Some other kanji/radical/orig sets are

There are lots more.


The things I've mentioned here are just the tip of the iceberg. It all started when the book shuowen jiezi was published in China around the time of Christ. Characters were indexed via 540 radicals. More than 1,500 years later (when Benjamin Franklin was a wee lad of 10, still 16 years before George Washington was born), the influential kangxi zidian was published, classifying some 42,000 kanji via 214 radicals. These 214 ``traditional radicals'' are still commonly used today.

(the history blurb taken mostly from Jack Halpern's ``New Japanese-English Character Dictionary'')

``Fixing'' the radical system

Because of changes and errors and a lot of the silly things I mentioned above, it seems those that write dictionaries try to ``fix'' things to some extent or another. Andrew Nelson's ``Modern Reader's Japanese-English Character Dictionary'' uses the 214 traditional radicals, but reclassified some characters. Some say this is harder to use, although it might be the case that it is simply different (I don't use it myself so don't know).

Jack Halpern's ``New Japanese-English Character Dictionary'' also reports the traditional radicals, and at times they differ from Nelson. I don't know which are the ``authentic'' traditional radicals, or even if such a thing exists for modern characters (although Halpern does report the ``lost radicals,'' such as with from above). I do know that the examples reported above came from both sources, so there are probably some inconsistencies. Welcome to the wonderful world of kanji.

Spahn and Hadamitzky

When Spahn and Hadamitzky's ``Japanese Character Dictionary - With Compound Lookup via Any Kanji'' was published in 1989, they realized that merely tinkering with the joke that the traditional radical system had become wasn't enough -- they completely reworked the radical concept, compressing similarly appearing radicals together, identifying the ``logical'' radical based upon each character's present form. The result was 79 distinct radicals, with one extra grouping for ``no radical''.

This was both good and bad. Virtually all Japanese and Chinese dictionaries (for the native speaker or the student) available used something relatively similar to the traditional radicals. Even if the status-quo is silly, there is indeed some benefit to sticking with it. Not only did it offend the traditionalist's crowd as both blasphemous and rinky-dink, it could potentially harm the new student. If one learns on the S&H radicals, they might well have trouble using any other dictionary.

Personally, S&H was the first kanji dictionary I ever looked at, so had no ``traditional sensibilities'' to be hurt. I was just starting to study kanji, and wanted whatever would make it easier. My one data point with a native Japanese using it was during a conversation where he was in a rush and I wasn't understanding the word, he saw the S&H on my desk and found the word in about 10 seconds. At least with that one he had no problem (S&H indicate that their changes only influence about 15% of all characters).

Still, there are a lot of very common characters in the ``no radical'' group, so anyone learning via S&H would do well to at least check out the traditional radicals. Personally, in 7 years, I've never felt limited by my using only the S&H radicals. But then, only on occasion have I ever looked up something in a Japanese-Japanese kanji dictionary -- those that use them day-to-day would probably run into the differences more readily.

The New Nelson

The new Nelson, scheduled to be out in 1997, appears to have one very nice tidbit. Quoting from Tuttle's page (from when it still existed):
The chore of hunting for kanji by determining their primary radicals has been simplified with the new Universal Radical Index, and characters can be identified immediately using any one of their radical components.
Thus, it seems that if you can identify any component of a character as a radical (even if it's not the radical), you can find it via the new index.

Halpern's SKIP System

Although not really related to radicals, I'll mention Jack Halpern's `SKIP' system of kanji index. It is based not on radicals, but on the physical appearance of a character (which may be coincidentally related to radicals, but that's, well, coincidental). I have a small introduction to SKIP here.

I find SKIP to be extremely useful for quickly finding a character. In fact, I even use SKIP when searching kanjidic. Most traditionalists scoff at it as just a new gimic, but of those (few) I know that have given it a reasonable try, all have eventually decided it is faster and more reliable for them.

Comments appreciated
[Return to Index] [Return to Main Page]
(this page's master source last modified 17 years, 3 months ago)
This reply to request 119,564,802 made just for you Tue Feb 28th 2017 2:28pm JST [load currently averaging 4133 requests/day over a 1087-second sample]