Re: Graphemes defined

cowan@... wrote:

>
> Peter T. Daniels scripsit:
>
> > > If you don't like my pirating of the term "grapheme" for this purpose,
> > > I invite you to provide a substitute term.
> >
> > What purpose, term for what?
>
> The purpose of establishing a term for "a class of written marks which the
> participants in a tradition of writing recognize, without hedging, as
> corresponding to one of the abstract basic units of the system."

IOW, it's not the slightest bit like a phoneme, a morpheme, a tagmeme,
etc.; so why should it be an -eme word?

> Thus a-z A-Z are among the graphemes of English, but neither cedilla

<a> is a grapheme, and <A> is a grapheme? How do you capture the
relationship between them?

Is <sh> a grapheme of English? <th>? <ng>? <ough>? (NB those four
examples are in a deliberate order.)

What are the graphemes of Chinese?

> by itself nor even c-cedilla is, despite its occasional appearance in the word
> "façade". Certainly alpha is not.

That's like saying there's no /x/ in English despite "Bach."

> > BTW if you know who you pirated it from, please let Gerhard Augst know
> > -- he wrote a rilly big book about it & couldn't decide.
>
> The Unicode Standard used this term between versions 3.0 and 4.0, when they
> abandoned it in favor of the unanalyzable (in this context) term "grapheme
> cluster".

The word has been around for many, many decades, but no one seems to
know who first made the mistake of thinking of it.
--
Peter T. Daniels grammatim@...