Re: How about a typology for input methods

> From: Peter T. Daniels [mailto:grammatim@...]


> > Non-linearity is systemic to some scripts. The same isn't true for

any

> > alphabetic *script* that I know of, but that doesn't mean that
> > *orthographies* based on those scripts can have non-linear
> > representations.
>
> But wait -- you just said you didn't like calling Korean by something
> that didn't reflect its relation to its substance (i.e. "featural"),

so

> you must want to call Korean an alphabet, and Korean isn't linear!

Indeed, if Korean were to be considered an alphabet, it would make sense
to say it is systemically non-linear (apart from modern linearized
variants). But I think it's insightful -- I know you disagree -- to have
a system of classification in which one class encompasses scripts in
which there are units of graphical structure that correspond to phones,
and also units of graphical structure that correspond to syllables. This
makes a class that includes both Korean and SignWriting, and (employing
prototypes) also Indic scripts. And "alphasyllabary" seems as good a
name as any.

Again, I know you disagree. (I have no intention of rehearsing our lack
of consensus from 2001.)



Peter Constable