Richard Wordingham wrote:
> It'd be better to have the letters and sounds in the same orders,
> e.g. 'ALEF, YEH and WAW' ... '/ā/, /ī/ or /ū/'.

Ooops!

> It can be argued that this is not quite true - that /i:/ is /ij/
> and /u:/ is /uw/, in which case there is nothing impure about these
> two cases! The same applies to the classical diphthongs /ai/
> and /au/.

This might be the correct *historical* explanation of how the spelling
evolved, but I don't think this is how Arabic spelling is perceived today.

Notice that, in fully vowelled text, there is a subtle difference between
yaa' and waaw used to spell /j/ and /w/ and the same two letters used to
spell long vowels and diphthongs: in the former case, yaa' and waaw carry a
sukuun, in the second they don't.

> ALEF for /ā/ remains impure.

I assume that alif used to represent a glottal stop, before it became a mute
carrier for the hamza. So, historically, you could perhaps analyze /a:/ as
/a/ + glottal stop.

> There are false plene vowels (e.g. WAW in /abu/ <ALEF, BEH, WAW>),
> and they are impure usages.

OTOH, there are unmarked long /a:/'s, which are only indicated in vowelled
text by means of superscript alifs.

However, I guess that such tiny details don't belong to the short definition
of "abjad" that Michael is trying to come up with.

--
Marco