> Peter T. Daniels wrote:
> > What are the graphemes of Chinese?
>
> Well, it is obvious that a Chinese logogram such as "媽" ("ma¹" =
> "mother")
> is composed of a left-side element, "女" , and a right side element,
> "馬".
>
> It is well known that, in this case, the left-side element "女" bears a
> generic semantic indication, 'woman', while the right side element
> "馬" bears
> a phonetic information, [ma].
>
> These two elements are also found in other logograms, either with the
> same
> function described above or with swapped functions (i.e. "馬" can
> stand for
> meaning 'horse', and "女" can stand for sound [nu]).
>
> Moreover, they also constitute two single-element logograms on their
> own:
> "nü³" ('woman') and "ma³" ('horse').
>
> Finally, neither "女" nor "馬" can be further subdivided in graphic
> elements
> conveying either a certain meaning or a certain sound.
>
> To me, the role of elements such as "女" and "馬" in the Chinese writing
> system seems quite analogous to the role of a phoneme in a language's
> phonetic system, or to the role of a morpheme in a language's
> morphology:
> they all are the "atoms" which we can use to analyze larger
> structures.

I think that in chinese could have the role of "atoms" also the basic
strokes (bi hua), that form the elements such as "女" and "馬". The
strokes consist of dots and line (may be six basic types), that can be
connected and combined to form more than thirty basic strokes.
I think that it could be similar to Martinet's double articolation
(doppia articolazione??? Marco help me!! : ) ), but with more levels:
first level dots and line, second level strokes, third level
radicals... and so on.
I suppose the problem is that "grapheme" is a term invented for latin
alphabet...