Typology and classification Re: everything else

Lars Marius Garshol wrote:

>
> * Lars Marius Garshol
> |
> | To some extent that is true, but birds and bees have fundamental
> | differences of a kind that I am not sure abugidas and syllabaries do.
> | What fundamental difference do you see that keep abugidas and
> | syllabaries so far apart that they are not even allowed a common
> | supertype below 'script'?
>
> * Peter T. Daniels
> |
> | Just as with the biological analogy, their ancestry is different.
>
> So what? Even the scripts within the "abugida" class have different
> ancestries.
>
> * Lars Marius Garshol
> |
> | Do you really think that there is only one alphabet? If so, what is
> | your response to the following statement:
> |
> | "Alphabet
> |
> | A type of writing system that denotes consonants and vowels with
> | separate characters.
> |
> | There are 36 instances: Armenian, Asomtavruli, Avestan, Buryat,
> | Carian, Cirth, Coptic script, Cyrillic, Deseret, Etruscan script,
> | Georgian script, Geyinzi, Glagolithic, Gothic script, Greek, Latin
> | script, Lycian, Lydian, Manchu, Mandaic script, Meroitic, Modern
> | syriac, Mongolian clear script, Mongolian script, N'ko,
> | Nusxa-xucuri, Ogham, Old Persian Cuneiform, Orkhon, Osmanya, Punic
> | script, Runic, Shavian, Sidetic, Thaana, Utopian."
>
> * Peter T. Daniels
> |
> | Who came up with the idiotic number 36?
>
> The topic map engine. It's telling us that the particular database
> it's looking at has 36 instances of this class, not that the class has
> precisely 36 instances. The class is of course of potentially infinite
> cardinality.

What's a "topic map engine"?

> | [...] and this list includes several abjads as well.
>
> Which ones? I've used the information in D&B to create this list, but
> the information given there is very incomplete when it comes to types.

Obviously that's not correct, since the list contains several items not
even mentioned in WWS.

> You still haven't answered the question, though. Do you consider all
> these to be the same script? If so, can you explain why to a novice
> like me who thinks that that's self-evidently wrong? Are runic and
> armenian the same script? Are thaana and utopian?
>
> If you think there's only one alphabet, do you think the other classes
> also have just one member? If not, what's the difference?

During the day, I figured out what the problem is that seems to have
brought about all this wrangling that suddenly appeared on qalam list.

Typology in linguistics is an impressionistic thing, whether isolating /
agglutinative / fusional / polysynthetic; or SOV / SVO / VSO. No
language belongs purely to any of those types (or the other typological
axes that have been investigated); they represent characteristics that
give you an idea of what the languages are like. In particular, Joseph
Greenberg (and he confirmed this to me personally) entered into his
typological studies in order to weed out what wasn't relevant to genetic
classification.

You guys, however, seem to want my *typology* to serve as a
*classification*; to divide the universe of writing systems, or scripts,
uniquely and exhaustively. Of course it doesn't do that and doesn't
claim to do that; in particular, it's not particularly relevant to
scripts invented in modern times by linguists (or others) who are well
informed about the structures of languages and of scripts.

In linguistics, classification can only be genetic, and the genetic
classification of scripts is largely trivial (we may not yet understand
which particular variety of Semitic abjad was adapted into Brahmi, but
the general relationship is clear).

My typology is a typology and you (whoever it was whow were arguing
about it on the Unicode list) can't shoehorn everything into it as if it
were a classification.

My *typological* observations of the differences between Ethiopic and
cuneiform, between Cherokee and Cree, made me recognize that a number of
seemingly unrelated phenomena could be newly accounted for (and isn't
that what science is about?) in a unified way. (I misremembered before:
the earliest appearance in print of the typology was in "Fundamentals of
Grammatology," in *Journal of the American Oriental Society* for 1990.
That's where I took on Gelb's Principle of Uniform Development.)

> * Lars Marius Garshol
> |
> | Also, you refer to your system. Where is that defined? Which classes
> | does it consist of? Is "alphasyllabary" included, for example?
>
> * Peter T. Daniels
> |
> | You claim you have access to WWS, and you don't know?
>
> What's strange about that? WWS is written by you and Bright, who seem
> to disagree on whether or not the terms "abugida" and "alphasyllabary"
> should be used. How can I then know whether your system includes
> alphasyllabary or whether WWS includes it as a compromise between you
> and Bright? Given that the classes have almost entirely overlapping
> membership it seems to me that any system that includes both is
> inconsistent.

Every article is signed. My typology is presented in Section 1, which is
written by me and no one else. Bill used the word "alphasyllabary" in
the articles about the part of the world he was responsible for, and I
explained the difference as I then understood it in the footnote on p.
4.

No typology includes both.

He subsequently went into the question in more detail in the
Urbana/Seoul paper, having identified hPags pa as a test case. It is an
abugida, it's not an alphasyllabary, and that follows directly from our
two definitions.

I like my category better because it's functional; his is formal.

(And the journal SLS isn't a "Working Papers"!)
--
Peter T. Daniels grammatim@...