Re: [OT] ANN: Site about scripts

From: Kenneth Whistler (kenw@sybase.com)
Date: Fri Oct 12 2001 - 15:28:25 EDT


Lars,

> Actually, these definitions are both Daniels'. The one you attribute
> to me appears on page xxxix (first page of "Abbreviations,
> Conventions, and Definitions") of Bright and Daniels.

Hmm. I missed that. I guess Daniels and Bright decided not to
take a definitional approach to their discussion -- probably because
when you get down to the dirty details there is so much overlap
in the categories, anyway!

>
> Does the definition
>
> "A type of writing system whose characters denote syllables and
> where the characters are systematically composed from more fundamental
> symbols denoting phonetic features."
>
> sound sensible for a type called "Featural syllabary"?

I'd still avoid the "denoting". The pieces usually would not
*denote* phonetic features, but they do tend to be correlatable
with them. This is almost always the case because as soon as
a convention is in place, it is going to start undergoing
historic drift -- someone adds an extension here, someone changes
a form there, and so on. So even if an original intention was
for some formal distinction to be perfectly correlated with
a phonetic feature, it may drift away from that, and in any
case doesn't *denote* the feature, per se.

> | Also, regarding your current class of "alphasyllabary", I don't
> | think the critical issue here is linear order "that is congruent
> | with their temporal order in speech."
>
> I'm not sure I understood this. Is the issue that there is no such
> class as "alphasyllabary", or that I haven't been precise about the
> distinction between "abugida" and "alphasyllabary"?

The point was that in abugidas (which most of the scripts in
your alphasyllabary grouping are), it isn't necessary that any
vowels be represented out of logical (your "temporal") order.
It is an accidental feature of many abugidas primarily because
of those /e/ or /i/ vowel marks flopping off the top of consonants
around to the left of the consonant bases sometime in various
scripts' histories.

>
> If there is such a class, how does one tell instances of it apart from
> instances of the "abugida" class?

I don't think that "alphasyllabary" is a useful separate class
from syllabary.

Here's how I would approach the general typological problem.
First, take it as a given that while there may be some generally
clear types, almost all scripts are going to have mixed
characteristics, so the typology is always going to be blurred
around the edges. It is a matter of archetypes in a fuzzy space,
rather than really hard and fast distinctions.

Then I would ask the basic question: what is the nature of
the enumerated units generally proffered when one asks what
the "letters" or "characters" of the script are. Then with a
few simple criteria, the answer to that question can pigeonhole
scripts based on the archetypes. (Of course then you go on to
investigate all the funny edge cases which make the pigeonholing
less than exact.)

Criteria for the answers:

1. The units correspond to isolated consonants and
   isolated vowels (at least in principle) --> alphabet

2. The units correspond to isolated consonants only,
   with vowels left out or indicated by subsidiary marks --> abjad

3. The units correspond to consonants with an inherent
   vowel (usually /a/), with other vowels indicated by
   separate dependent units, and usually with a "killer"
   in the system to drop the inherent vowel --> abugida

4. The units correspond to syllables (usually CV), without
   separate units for isolated consonants. --> syllabary

5. The units correspond to syllables (usually CV), but in
   addition, formal aspects of the system can be correlated
   with segmental characteristics (i.e., there is more or
   less regular indication of consonant or vocalic rank,
   or of other phonetic information through form). --> featural syllabary

6. The units correspond to syllables, but not to the *sounds*
   of syllables, but rather to *morphemes* as conveyed by
   syllables. (At least as a predominant pattern.) --> logosyllabary

Of course many actual scripts don't fit neatly into any one
category, but as a practical approach to pigeonholing, this
is pretty good.

If you ask what are the units of Script X, and the answer
is like:

A B C D E F G H I... (with vowels in the list) ==> alphabet

ALEF BET GIMEL DALET HE VAV... (with no vowels in the list) ==> abjad

KA KHA GA GHA NGA CA CHA ... I II U UU E EE... (with consonants
    all cited with inherent vowel, and vowels pulled out separately)
    ==> abugida

KA KI KU KE KO SA SI SU SE SO TA TI TU TE TO... ==> syllabary

"The list of syllables is too long, but here is a table for it, and
 see how the dot on this side means X and the dot on the
 other side means Y...." ==> featural syllabary

"Look them up in the dictionary -- each one means a different
thing." ==> logosyllabary

Also, while abugidas clearly have a strong syllabic orientation
(they are a kind of half-syllabary), one notable difference
from the true syllabaries is that an abugida is almost never
presented as an enumeration of syllables, whereas syllabaries
*are* presented as an enumeration of syllables (even though
the enumeration may not be complete for the language -- as
is the case for the way Hiragana and Katakana are presented,
for example). Because of the way abugidas are structured,
it would be considered redundant to list out all the syllables --
the syllables are simply constructed by juxtaposition of the
consonant and vowel parts of the abugida. (Of course, for
pedagogical reasons, depending on the abugida, you might
have to actually do at least part of the enumeration, since
you can end up with complex rules for how the parts end up
shaped when they are juxtaposed, but that is secondary to
the *concept* of the abugida.)

>
> | (Of course nothing is black and white here -- the voicing marks
> | of Kana obviously are featural, and the tone marks for Yi are
> | also featural.)
>
> Would you say the diacritics, as used in French, are featural?

To the extent that application of a diacritic in French, as for
e vs. , is correlated with a regular distinction in vocalic
quality, it would be featural. However, the problem is that
as the language changes, some of the original distinctions are
lost, and the accented vowels just become arbitrary spelling
conventions using different letters. To that extent, it is
no longer featural.

Many alphabets contain such featural aspects in their diacritic
usage. For example, the hacek applied to s, c, z, is usually
there to indicate a systematic place of articulation difference.
Placement of hard or soft signs in Cyrillic could be considered
featural, and so on.

--Ken



This archive was generated by hypermail 2.1.2 : Fri Oct 12 2001 - 14:19:39 EDT