--- In qalam@yahoogroups.com, Michael Everson <everson@...> wrote:

> I intend to forward the following definitions to the Unicode Book
> Committee as improvements on the current definitions:
>
> Abjad. A writing system in which only consonants are indicated.

<Snip>

> Alphabet. A writing system in which both both consonants and

vowels

> are indicated.

<Snip>

> I am satisfied with both of these as being suitable with regard to
> the aims of the Unicode glossary

Should you be? You have changed Syriac from being an abjad (like
Arabic) to being an alphabet! Unfortunately, I'm not sure what the
aims are.

For the practical purposes of recording text, there is very little
difference between an abjad (as defined here) and an alphabet.
There is a simple progression of symbols - typically linear, but
zigzag in Korean. The biggest practical difference is probably
between systems that are ligated (such as Arabic) and those that are
not.

For practical purposes, I think there is little difference between
pointed abjads (now classed with true alphabets) and abugidas. As
an optimality issue, some meaning may be attributed to the lack of a
pointing. In abugidas as Peter Daniels defines them, it implies a
vowel. In Hindi, the meaning depends on the position of a word - at
the end it means no vowel sound, in other positions it means a
particular vowel. In the Thai script as used for Thai (as opposed
to Pali or Sanskrit), absence is ambiguous - it means either no, or
the inherent vowel sound, which may further depend on the phonetic
context. There are two major issues in most scripts derived from
the Brahmi script, most of which are clearly abugidas:

1. The conjunction of consonants in the onset of the quasisyllables
that Brahmi-derived scripts usually divide words into.

2. The progression of symbols, compared against the spoken sounds,
is not simple. Most particularly, there tend to be preposed vowels.

Issue 1 defeats simple font technologies, such as the use of a Type
3 Adobe font to render strings. Text *must* be preprocessed first.

For Issue 2, either 'logical' and 'phonetic' order will conflict, as
in Thai and Lao, or again, simple font technologies will be
defeated. I would add that being able to use a simple font
technology was once of great personal importance to me - it reduced
the time to write correctly spelt love letters from days to hours.

An example of the effects of Issues 1 and 2 lies in spelling. I
have seen Thai /ngau/ 'lonely', correct spelling visually
<e><h><ng><aa>, spelt <e><h><aa><ng> (an impossible Thai spelling)
because the correct (and logical) spelling feels too implausible.
(The <h> serves to indicate the tone, and is the only way it can be
indicated for the tone and phonetic initial of this word.)

On the basis of this example, and the paper on segmental awarenesss
that Suzanne McCarthy drew our attention to, I think Issue 2 is not
merely a computing issue.

The existence of an inherent vowel has nothing to do with either of
these issues. Diller describes the Thai and Lao scripts together.
By the addition of one vowel (for inherent vowel in closed
syllables) and the consistent use of another vowel symbol, Lao has
become an 'alphabet', while Thai remains an abugida. Issue 2
applies to both. Issue 1 has already been reduced in Thai because:

a) Conjuncts are not written as such in Thai.

b) As far as I am aware, C..CV pseudosyllables do not cross
sesquisyllable boundaries.

Issue 1 has been further reduced in Lao, partly by treating
sesuisyllables as two syllables, and partly by elimination of almost
all initial phonetic clusters.

While many divisions do have to be arbitrary, it does seem class
Tamil and Thai as abugidas and Lao as an alphabet, especially as the
practical consequences of the orthographic features are much closer
in Tamil and Lao than in Tamil and Thai!

To summarise: For practical purposes, the diagnostic feature of
abugidas is not what sets most of them apart from alphabets in the
broader sense. It's like making a hairy coat a diagnostic feature
for mammals.

I also think that the treatment of vowels is fundamentally different
in descendants of the Greek alphabet and in most 'pointed abjads'.

Richard.