Patrick Chew wrote:

>
> At 02:58 PM 12/13/2003, Peter Daniels wrote:
>
> >Morphogram would be better, but then you have to explain what morphemes
> >are.
>
> Either that or call it a syllabo-graph... there're far too many
> bound-morph/cran-morph graphemes to really justify "morphogram," I think.

No, syllabograms are like Cherokee or kana: they denote syllables.
Morphograms denote morphemes. (Hockett in the review of WWS suggested
that a logosyllabary is a syllabary that distinguishes homophones.)

> >Yet somehow speakers of the language can usually figure out unfamiliar
> >characters from the components and the context. Remember, they _already
> >know_ the language perfectly.
>
> hrm.. actually when reading and coming across unfamiliar
> characters, *IF* it has a phonetic component, one can usually try to make a
> guess from context. However, since phonetic components are usually not
> directly phonetic, it's not that straightforward. Usually people
> ask someone else or they look it up in the dictionary.
>
> "Already know the language perfectly" is actually a spoken
> language context. There is still a large gap between spoken and written
> language.. if you've never run across a character (especially a base
> character, or one where none of the compositional graphemes contributes
> phonetically), you've no idea how to pronounce it....

In less than 10% of characters is there not a phonetic component. While
its reading often is not exactly the same as the reading of the
character, it's generally similar enough for figuring it out. (Take it
up with Lee Sau Dan.)
--
Peter T. Daniels grammatim@...