Hallo conlangers!

On Friday 25 November 2011 18:40:33 R A Brown wrote:

> On 25/11/2011 09:19, Peter Cyrus wrote:
> > I find the Daniels typology less useful than I need : too
> > focused on how a script works and not enough on what it's
> > trying to represent.
> 
> What's the difference?

Well, writing systems always represent the same matter,
namely languages.  They do so at different levels, though;
and different languages suggest different levels of
representation being recommendable.  Japanese, for instance,
with its small phoneme inventory and simple syllable
structure, is almost ideally suited to a syllabary (or an
abugida), while a syllabary for English would need thousands
of characters.

But I am not Peter Cyrus, so I cannot definitely answer his
question ;)
 
> > If you have a CV language with only a few vowels, and are
> > looking for a writing system for it, the big decision is
> > whether each unit should represent a CV mora or something
> > smaller.
> 
> In the real world people don't normally sit around making
> decisions about these things (yes, I know there are
> exceptions); writing systems generally evolve.  I don't
> think anyone sat around in ancient Ethiopia (or Abyssinia)
> wondering whether to use an alphabet or an abugida.  The old
> Ethiopian abugida evolved from earlier Semitic abjad.

Right.  Few of the writing systems in common use today (or at
any time in the past) have been consciously designed the way
they are.  They evolved from earlier writing systems, or in
some instances, from pictographic systems (which are not true
writing).
 
> I don't know how you define only a few.  The Ethiopian
> abugida distinguishes 7 vowels.  As some the vowels are
> long, presumably some some CV symbols represent monomoraic
> syllables and others bimoraic syllables; the same also
> applies to the Indic scripts, which IIRC represent even more
> vowels.

They do.  Devanagari has 18 signs for vowels, diphthongs
and syllabic liquids.  Others have similar vowel sign
inventories.

> The point Daniels was trying to make - and I agree with him
> - is that in a pure syllabary the basic unite is the
> syllable, represented by a unique syllabogram which cannot
> be further analyzed.  In an abugida, the symbols for each
> syllable is analyzable into onset and nucleus.

Exactly.  And hence, an abugida is more like an alphabet
than a syllabary, at least what regards the number of symbols
to learn.  Devanagari, for instance, has 40 consonant symbols
and 18 vowel symbols; add to this some extra symbols such as
visarga, anusvara, candrabindu etc.: altogether about 70 or
80 elements.  A true syllabary expressing the same vowels and
consonants would have 40x18 = 720 different symbols for CV
syllables, a much higher tax on memory!

> [snip]
> 
> > The fact that we're arguing over whether true syllabaries
> > can be featural
> 
> Syllabaries such as the Japanese hiragana & katakana or the
> ancient Mycenaean Linear B are quite certainly not featural,
> period.

Amen.
 
> > and what exactly Canadian Syllabics and Hangul are, is an
> > indication we need a better typology.
> 
> They are featural.  Yes, the whole terminology alphabet,
> abugida, abjad, syllabary, alphasyllabary is not clear cut;
> one of the problems is that writing systems themselves do
> not always neatly fall into discrete slots   ;)

Indeed not.  I think we can boil it down to:

Logography: unitary glyphs for morphemes
Syllabary: unitary glyphs for syllables
Abugida: unitary glyphs for phonemes, grouped by syllables
Abjad: unitary glyphs for consonants, linearly arranged
Alphabet: unitary glyphs for phonemes, linearly arranged
Featural abugida: glyphs for phonemes composed of elements
  representing phonemic features, grouped by syllables
Featural alphabet: glyphs for phonemes composed of elements
  representing phonemic features, linearly arranged

Yet, many scripts combine elements of two or more of the
above.

> That is why I've taken the trouble to define the way I am
> using the terms.  I'm not saying these are definitive
> definitions; I merely define what I mean in order to avoid
> misunderstanding.
> ==========================================================
> 
> On 25/11/2011 09:58, Michael Everson wrote:
> > On 25 Nov 2011, at 08:37, R A Brown wrote:
> >> it contrasts with a syllabary, where letters with
> >> shared consonants or vowels show no particular
> >> resemblance to each another,
> > 
> > Glyph design is irrelevant to whether something is a
> > syllabary or not.
> 
> Clearly not all authorities go along with that.

Clearly, glyphs should not be more complex than necessary.
Of course, the more glyphs you have, the more complex they
must be in order to distinguish between them.  Hence, a good
alphabet ought to have simpler glyphs than a syllabary or
a logography.  Not all conscript designers take heed of thus.
I have seen alphabetic conscript with glyphs that were hardly
less intricate than Hanzi!
 
> I do not consider myself an authority; but I feel it helpful
> to make it clear what I understand and mean.  It may be
> better if all writing where the glyphs represented a
> syllable were unambiguously called syllabaries, and we had
> terms for subsets of syllabaries according to the way the
> glyphs are or are not related to one another etc.  But AFAIK
> such unambiguous terminology does not exist.

Indeed not.  Terms such as "alphabet", "syllabary", "abugida"
etc. are bandied about with whatever definition the relevant
author sees fit.  Some are used in misleading and contradictory
ways in the popular press, as such examples as "Japanese
syllabic alphabet" or "Chinese ideographs" show.

I have made a proposal above.

> ==========================================================
> 
> On 25/11/2011 16:07, Jörg Rhiemeier wrote:
> > Hallo conlangers!
> 
> > On Friday 25 November 2011 09:37:43 R A Brown wrote:
> [snip]
> 
> >> [Okamoto] did not even want the featural aspect of an
> >> abugida.
> > 
> > I see.  So he did not see a point in representing units
> > smaller than the syllable in script - hence a syllabary
> > and not an alphabet or abugida.
> 
> That is correct.  He obviously thought that in a language
> where the syllabic shape is either CV or V, it made things
> more complicated to have anything smaller than a syllabogram.
> 
> I don't agree with him, and my reason for adopting a
> Roman-letter syllabary are different.

Yes.  You use it because it allows for shorter morphemes,
not because of some preconceived idea about how one should
write a language.
 
> [snip]
> 
> >> That seems to me like calling both true alphabets,
> >> abjads (and often abugidas) "alphabets", as is
> >> popularly done.
> > 
> > I have even come across the term "syllabic alphabet"
> > several times, sometimes referring to abugidas,
> > sometimes to plain old syllabaries such as the Japanese
> > kana syllabaries!
> 
> I have seen the term applied to abugidas; there is some
> justification in that abugidas represent both the alphabet's
> having symbols for both consonants and vowels and, because
> the symbols are fused to make a single glyph, each glyph
> represents a syllable.

Yes.
 
> [snip]
> 
> >> I understand _syllabary_ to denote a writing system
> >> where the basic symbols is a syllabogram, i.e. the
> >> symbol itself, as e.g in hiragana, katakana and Linear
> >> B (and, presumably, Linear A) denotes a whole syllable
> >> and cannot be broken down into smaller phone*ic units.
> > 
> > Yes.  How you classify abugidas is a matter of what one
> > considers the "basic symbols", but arguably, the
> > syllabograms are *not* the basic symbols in an abugida -
> > because they can be broken up into consonant and vowel
> > symbols.  In that way, an abugida is closer to an
> > alphabet than to a "true" syllabary.
> 
> Exactly.

Just that.

--
... brought to you by the Weeping Elf
http://www.joerg-rhiemeier.de/Conlang/index.html
"Bêsel asa Êm, a Êm atha cvanthal a cvanth atha Êmel." - SiM 1:1