--- In qalam@yahoogroups.com, "Peter T. Daniels" <grammatim@...>
wrote:

> Michael Everson wrote:
> >
> > At 20:26 -0400 2004-08-01, Peter T. Daniels wrote:
> >
> > > >So you think that a Hebrew Genesis when unpointed is an

abjad, but

> > > >that a Hebrew Genesis pointed is no longer an abjad, but has

become

> > > >an alphabet?
> > >
> > >No, a text isn't any sort of writing system.
> >
> > You're being obtuse; please don't be.
> >
> > "So you think that a Hebrew Genesis when unpointed is written

with an

> > abjad, but that a Hebrew Genesis pointed is no longer written

with an

> > abjad, but has become an alphabet?"

> > >The former is written with an abjad, the latter with an alphabet.
> >
> > That's one way of putting it. I tend to prefer saying that the

Hebrew

> > script is an abjad, though it is used as an alphabet when used for
> > Yiddish. I am not sure the Hebrew points count as "letters",

though

> > -- in which case it may not be appropriate to say that the Hebrew
> > script is used as an alphabet when pointed.

> Scripts are no longer abjads when used with "pointing" systems that
> denote all the vowels, as in the cases of Syriac, Arabic, and

Hebrew.

Perhaps we should consider some cases, and then think how useful the
divisions alphabet v. abjad v. abugida are.

A pure abjad is clearly distinguished.

I think we can see several different categories:

a) 'Pure' abjad - no vowels present.
b) Impure abjad - most abjads have some consonants which can indicate
vowels. The vowel indicators _follow_ the consonants - I think this
may be an important point.
c) Almost all vowels are marked, and these marks, at least in
principle, follow the consonants.

At this point we have an alphabet, e.g. Greek. (My pronunciation
of English has an unmarked vowel in "isn't" - the <n> is not syllabic
in my speech, which is why I say 'almost all'.)

(a) to (c) is, I believe, the historical progression leading to most
undisputed alphabets.

d) Almost pure abjad, but word-initial vowels are distinguished, e.g.
Ugaritic. (What was the origin of the independent vowels in the
Brahmi family?)

e) 'Pointed abjads', e.g. Modern Syriac and the optional pointing of
Hebrew and Arabic. Do we need to distinguish pointing according to
what not pointing a consonant means. For example, in Classical
Arabic, the assimilation of the lam of the article before a word
starting with a solar consonant is indicated by writing no mark on
the lam and writing a tashdid on the initial consonant, thus
inidcating that the lam is silent and the initial consonant is
geminate. In Hebrew, silent consonants are indicated by having no
pointing. Some SE Asian scripts have an explicit mark to indicate a
silent syllable.

One could therefore argue that we have the following
inherent 'vowels' in the following systems:

Pointed Classical Arabic: Consonant is silent.
Pointed Hebrew: Non-final consonant is silent.
Final consonant has no vowel.
Devanagari for Sanskrit: a
Devanagari for modern
North Indian languages: Non final consonant: a
Final consonant: no vowel.
Thai: a/o/no vowel
Lao: no vowel

The overwhelming majority of the silent consonants in pointed Hebrew
are matres lectionis in unpointed Hebrew, but there are a very few
others, such as the first 's' in the Hebrew form of the proper name
Issachar.

Devanagari and Thai are classed as abugidas, and are both part of the
family descended from the Brahmi script. Most members of this family
group sequences of consonants into clusters, so the concept of
conjunct consonants arises. To my mind, conjunct consonants, rather
than the implicit vowel, are the most striking feature of this family.

The Thai script has dispensed with conjunct consonants. It retains
a 'no vowel' mark (phinthu), but I have never seen it used for
writing Thai, useful though it would be. It seems to be reserved for
foreign languages, such as Sanskrit and Pali. It also has a mark
signalling the presence of a conjunct consonant, but that seems to
have dropped out of use even for Sanskrit and Pali. Not all Thai
vowels are subordinate to consonants. The subscript and superscript
vowels are, but the preposed and postposed vowels are treated like
consonants when words are stretched out. Perhaps Thai is ceasing to
be an abugida. Another point is that (sesqui)syllables are not split
up (or does anyone know some examples?). For example, the
name /montho:/ (wife of Rama's enemy Ravanna) is spelt
<m><.n><o:><t'h>, not <m><o:><.n><t'h> as it would be in Devanagari.
By sesquisyllable, I mean a sequence /CaCV(C)/ treated as though
the /a/ were not there - the first example I met
was /khame:n/ 'Cambodian', spelt <e><~kh><m><r>. As far as I am
aware, sesquisyllables only occur word initially or as part of an
alliterating near reduplication.

The Lao script is very similar to the Thai script (I was surprised
that they are not unified in Unicode - perhaps the problem is that
there isn't a Thai nukta like the Devanagari nukta to derive the
extra letters - stretched <r> and <y>.), but Lao does not permit
inherent vowels. This development required one additional
superscript vowel symbol (for /o/). What does this make Lao - a
compulsorily pointed abjad or an alphabet?

(f) I am not certain if this class has any members. I suspect from
Michael Everson's notes that some varieties of Lanna script may be
examples. As far as I can interpret his notes, it seems that a
syllable or sesquisyllable /(C(a))CV(C)/ may be written in a vertical
stack. As they were only addressing the impact on encoding, they are
not complete, and I may have completely misunderstood what is going
on in the writing system. It seems like the Korean hangul system!
I'd like to find out more about the Northern Thai writing system -
it's use for Tai Lue is less important to me. Unfortunately, I read
Thai very slowly.

I would say pointed abjad (or do I need to 'abjad plus points' to
preserve the peace) is different to both an impure abjad and to an
alphabet. However, it should not be forgotten that many normally
unpointed abjads *do* have an associated set of points (or even two
sets in the case of Hebrew!). I'm not so sure that the difference
between an abugida and a compulsorily pointed abjad is so significant.

From a practical point of view, it is worth noting that the Unicode
ideal is effectively to represent a pointed abjad or an abugida as an
alphabet. Thai and Lao are rare exceptions to that rule. There does
not appear to be a consistent way of reporting a spelling in Thai!
One textbook gives most of a system that actually approaches the
Unicode ideal (ducking the difficult cases), but inquiry reveals that
many Thais do not use it. The difficult case was /ngau_5/ 'lonely',
written <e><h><ng><a:>, for the system described says that words such
as /au_1/ 'take', written <e><'><a>, should be spelt out <'><au>.

There are also arguably minor elements to the classification. How do
we handle cases where abugidas or pointed abjads seem to have become
syllabaries (Ethiopic and, so we are hearing, Tamil)? Should we do
anything about cases where part of the rime is encoded in the choice
of the initial consonant? This curious condition applies to many
S.E. Asian scripts - Tai Lue, Thai, Lao and Khmer to my certain
knowledge - as a result of similar sound changes in all these
languages.

Richard.