Lars Marius Garshol wrote:

>
> * Lars Marius Garshol
> |
> | Are all of 20 (Manchu), 21 (Mandaic), 22 (Meroitic), 24 (Mongolian
> | clear script), 25 (Mongolian), 29 (Old Persian Cuneiform), 30
> | (Orkhon), and 32 (Punic) abjads? (Just checking.)
>
> * Peter T. Daniels
> |
> | 22 can be interpreted variously; one claim is that it's an abugida.
>
> OK. I'll leave it as unclassified for now then. Is the problem that
> there isn't enough data, or just that it's a difficult typological
> case?

It's barely understood.

> | 29 is a mess
>
> Looking at it again I see what you mean. :-) I don't know what made me
> describe it as an alphabet, but I agree that it's neither an alphabet
> nor an abjad.
>
> | 30 is partly syllabic
>
> I've missed something, then. The table on page 537 of WWS, and the
> sample on 538 both seem entirely alphabetic to me. What are the
> syllabic elements in it?

I was remembering the superscripts in the table. They indicate that some
vowel information is in the consonant letters.

> | Why do you separate Punic from the rest of Phoenician? It supposedly
> | started using matres, Aramaic-like, in contrast to earlier
> | Phoenician (but I've never studied any Punic)
>
> Simply because WWS seemed to describe it that way. It is shown
> separately in the diagram on p 89, and the two later references to it
> (ps 112 and 141) do not in any way suggest that it's the same as
> Phoenician. Is it really the same script?

Punic is just late Phoenician

> * Lars Marius Garshol
> |
> | 12: Script invented by the guy who made the Omniglot website.
> | <URL: http://www.omniglot.com/writing/geyinzi.htm >
> |
>
> * Peter T. Daniels
> |
> | never heard of it ...
>
> Not so strange. I think he's probably the only person who's ever used
> it, or even looked closely at it.
>
> * Lars Marius Garshol
> |
> | 36 (Tai Lue): Page 282 of WWS.
> |
> | wrong ref.; that's 35
>
> You are right. Tai Lue is the script Peter Constable posted about at
> length. See <URL: http://groups.yahoo.com/group/qalam/message/528 >.
>
> * Lars Marius Garshol
> |
> | Is the reason you don't consider 23 and 37 alphabets that the vowels
> | are given "lower status" as characters in the script than consonant
> | characters? To me, a non-expert, it does not look like the scripts
> | are syllable-based, but would you say they are? If not, does it not
> | make sense to classify them as alphabets?
>
> * Peter T. Daniels
> |
> | I don't need to "classify" them, since how they work is clear:
> | clearly they're different from alphabets -- they're compulsorily
> | vocalized abjads!
>
> >From the definition on p 4 of WWS I concluded that a compulsorily
> vocalized abjad is an alphabet. There are also statements to that
> effect in WWS (in the article on Syriac). If the rules given in WWS
> are incomplete I've just learned something new.

They're not "rules."

> What's the point of separating alphabets from abjads with vowels,
> though? Aren't they in effect pretty much the same? Does it make much
> difference for the relation between language and writing whether the
> vowel characters have secondary status?

I don't think the psycholinguists have looked much into the matter ...

> * Lars Marius Garshol
> |
> | [mandaic]
> |
> | I believe it when you say so, but even now I am unable to understand
> | that by reading it. The description still sounds like the description
> | of an alphabet to me. Why do you consider it an abjad? Is it because
> | it only writes some of the vowels, and not all?
>
> * Peter T. Daniels
> |
> | Well, of course!!!!
>
> It's not that obvious to me. If an alphabet is writing with vowels and
> consonants, and abjads are writing with only consonant, a writing
> system with some vowels, but not all seems like a border-line case to
> me. But given the rule that abjads are abjads even if they write the
> vowels, but give the vowels lower status as characters, I see what you
> mean.
>
> * Lars Marius Garshol
> |
> | I have a web site about scripts. I believe it is useful to present to
> | the reader a set of script types, and a classification of the scripts
> | using those types. To be able to do so, however, I need a system that
> | works.
>
> * Peter T. Daniels
> |
> | You still don't answer Michael's and my question: WHY is it useful?
>
> There are many interpretations of this question, but I'll try my best.
>
> It's useful because then I can write "the Phoenician script is an
> abjad", or "'Phags pa is an abugida", and people who know nothing else
> about the scripts suddenly have at least some idea of how they work.
>
> It's also useful because I can present the different classes of
> writing systems and having read that presentation the reader will have
> gained a real understanding of the different kinds of scripts used
> around the world.
>
> Were these answers to the question you meant to ask?

Then the typology as-is serves its purpose, and you don't need something
mathematically precise and exhaustive.

> * Lars Marius Garshol
> |
> | Another reason is that I am trying to learn more about scripts and how
> | they work, and I am having a hard time understanding how to apply the
> | definitions of the different script types. It seems that the line
> | between alphabet and abjad, for example, is fairly subtle, and since
> | nobody, to my knowledge, has made an authoritative list of the
> | classifications of various scripts, I am reduced to finding the
> | correct answers by asking people who know.
>
> * Peter T. Daniels
> |
> | But there's exactly one example of each, and the transition is
> | perfectly clear:
> |
> | Phoenician >>> Greek.
>
> Perhaps.
>
> This the transitions my topic map tells me exists:
>
> Classical syriac    >>> Modern syriac   
> Arabic script    >>> Thaana   
> Phoenician script   >>> Greek   
> Pahlavi >>> Avestan   
> Mongolian clear script >>> Buryat   

Avestan letters come from a number of sources. Oktor Skjaervo had never
bothered to write out his suggestions but directed me to Hoffmann's
article in Encyclopedia Iranica; I sent him my codification of Hoffmann
and he replied with his own, and you'll find the result in my Handbook
of Linguistics (Aronoff & Rees-Miller) chapter.

> The requirement of equal status for vowels in alphabets you use
> removes the first two cases, and the third is the one we all know. You
> did not disagree that Avestan and Buryat are alphabets, however, so it
> seems like the two last cases should hold.

Avestan and Buryat, of course, are made with full knowledge of the Greek
and Cyrillic alphabets respectively, so they're not "unsophisticated
grammatogeny," so not interesting.

> (BTW, I used the following query:
>
> derived-from($A : predecessor, $B : successor),
> direct-instance-of($A, abjad),
> direct-instance-of($B, alphabet) ?
>
> to produce this list.)

--
Peter T. Daniels grammatim@...