--- In qalam@yahoogroups.com, "suzmccarth" <suzmccarth@...> wrote:

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

> > B. Richard Sproat (
> http://compling.ai.uiuc.edu/rws/newindex/indic.pdf
> > )interprets this as, 'They are alphasyllabic scripts (Bright,
> 1996a)
> > (though Daniels (1996) prefers the term abugida), meaning that

they

> > are basically segmental in that almost all segments are
> represented in
> > the script, yet the fundamental organizing principle of the

script

> is
> > the (orthographic) syllable'.

> THis is an article I refered to last year but did not then take the
> time to quote properly. This is what I read about Tamil. I am not
> quite sure what Sproat means about eliminating asksaras but
> definitely there are no consonant conjuncts in Tamil and there is

no

> orthographic syllable in the sense of Devanagari.

Well, there is <ksha>, which is quite transparent as Indian ligatures
go. I presume that one cannot treat it as a single letter because
that would upset the collation sequence.

I presume that 'eliminating aksaras' means that there are no
consonant sequences behaving as single consonants. Tamil syllable
structure guarantees that the only possile ones are those for which
the phonetic and orthographic structures conflict. The almost
complete elimination of conjuncts almost completely removes this
conflict.

> Tamil does frequently start with a vowel as the first _written_
> phoneme although as Mr. Daniels pointed out these are phonetically
> preceded by a a y/j/ or w/v/. This phonetic phenomenon has written
> expression between vowels, so 'visual basic' is written in
> Tamil 'visuval basic'. Nontheless independant written vowels
> frequently occur word initial.

Even Brahmi has initial written vowels. Only a few scripts of that
family have totally eliminated them.

> TAMIL
> "From a formal point of view, Tamil (Steever, 1996; Radhakrishnan,
> 2002) is probably the simplest Indic script since it has eliminated
> the
> aksara as a fundamental unit of the script. Virtually all consonant
> conjuncts
> have been eliminated, and consonant sequences are written leftto-
> right with unligatured consonant symbols. Steever (1996, page 426)
> suggests that this trend may have been related to the introduction

of

> typography from the West
>
> … the SLU in Tamil is a maximally CV(V) unit, with the

possibilities

> being V(V) (syllable initial short or long vowel), C (consonant-
> sequence-medial consonant)
> and CV(V) (consonant with diacritic short or long vowels). Ignoring
> for the moment the decomposability of CV(V) elements and the fact
> that there are simplex vowel glyphs representing both long and short
> vowels, Tamil orthography is almost what I termed in (Sproat, 2000)

a

> core syllabary, like Japanese kana. Tamil is moving towards being an
> alphasyllabic version of kana.
>
> If the SLU is the CV(V) unit, not an orthographic syllable, the
> theory
> predicts that vowel diacritics in Tamil can only attach to the

final

> consonant
> in a consonant sequence. …
>
> Tamil seems to be evolving towards kana in another regard, namely
> that the
> CV(V) units are becoming unanalyzable." p.22
>
> http://compling.ai.uiuc.edu/rws/newindex/indic.pdf
>
> Now from my reading of this text I would expect that there might be
> some similarities in the cognitive process of reading Tamil and
> reading Japanese. This is my main interest but not the only one.
>
> Ancestry is not so important, not even the label or definition but
> that scripts with similar pertinent features be classed together.

So

> scripts would be classed as having syllabic organization because
> that is a useful feature to know.

I think decomposability is an important issue. Don't forget that
scripts can be mixed, e.g. the logograms commonly used for numbers.
One may have a mix of pure-syllabary-like items (your example would
be the u(:)ksharas of Tamil) and alphasyllabary items. Even Thai,
which lacks independent vowels and typewritten ligatures, has some
pure-syllabary-like items - the symbol for the Inid syllabic 'r'.
There may also be important distinctions depending on how
orthographic and phonetic syllables relate. Some of these
distinctions may be a bit fuzzy.

> > One might object that an alphasyllabary should preferentially
> > partition CVCCV as CV-CCV.
>
> I don't see how that applies to Tamil either. Like Cree, Tamil has
> the V or CV (long or short vowels) unit followed by an optional C.
> However this C is not attached to anything in either case. So I
> think the partition is CV-C-CV.

Formally, this lone C might be regarded as an _orthographic_
syllable. It occurs in Sanskrit for example. In Thai it may be
bound to the preceding syllable because certain vowel forms require a
following consonant.

Richard.