Berthold Frommann wrote:
> No, no! I'm absolutely supporting Stalph's definition, no doubt.

OK, then I do too, and probably everybody else in the world...

> ALL radicals are graphemes.
> But there are graphemes which are not part of the 214 Kanxi-radicals,
> because ... there are more of them.

Agreed. But I dare say that it is not a big discovery...

Even in the oldest analyses and with the oldest terminology, most Chinese
characters (90%?) are said to be composed of a "signific" (aka "radical")
and of a "phonetic". "Significs" are certainly more that 214 (but however
probably less that 400), while "phonetics" should be nearly 1000.

> And please have another look at my examples - 寺 (temple) is not a
> Kanxi-radical; but it certainly IS both a grapheme (i.e. It
> occurs in other characters as a unit) and an independent kanji/hanzi.

Sure. I didn't comment about your examples because there was nothing to say:
in "詩", "侍" and "時", "寺" it is no doubt a "phonetic" (not sure about "特").

_ Marco