Hallo conlangers!

On Saturday 26 November 2011 14:47:10 R A Brown wrote:

> On 25/11/2011 22:24, Garth Wallace wrote:
> > On Fri, Nov 25, 2011 at 1:05 PM, Jörg
> > 
> > Rhiemeier<[log in to unmask]>  wrote:
> >> Hallo conlangers!
> >> 
> >> On Friday 25 November 2011 18:40:33 R A Brown wrote:
> >>> Syllabaries such as the Japanese hiragana&  katakana
> >>> 
> >>>  or the ancient Mycenaean Linear B are quite
> >>> 
> >>> certainly not featural, period.
> >> 
> >> Amen.
> > 
> > Not 100% though! In kana, the dakuten is a symbol for a
> > phonemic distinction (voicing of the consonant). ;)
> 
> I know about that - it's a diacritic!

Indeed.  This diacritic brings a whiff of featurality into the
syllabary, but the basic characters are entirely arbitrary and
not featural in any way.  The German orthography uses a diacritic
for the vowels /æ ø y/ which evolved from fronting of /a o u/
(and in the meantime, /æ/ has pretty much merged with /e/, thus
leading to two graphemes for one phoneme), but that doesn't make
the German orthography featural.  All this shows is that pure
types of writing systems are rarely if ever implemented.

> ================================================
> 
> On 25/11/2011 22:09, Garth Wallace wrote:
> > On Fri, Nov 25, 2011 at 12:37 AM, R A Brown wrote:
> >> On 25/11/2011 01:18, Garth Wallace wrote: [snip]
> 
> [snip]
> 
> >> But as the word "abugida" is of Ethiopian origin and
> >> referred to the so-called "Ge'ez alphabet", it seems to
> >> me rather odd to restrict it just to the writing
> >> systems of a an Asian subcontinent.
> > 
> > I never said it was restricted to the Asian subcontinent.
> > I just said I thought it referred to writing systems LIKE
> > those.
> 
> OK - sorry.
> 
> It's just that I have come across abugida confined to the
> Indic scripts with their own peculiarities; e.g. unlike the
> Ethiopian abugida, the Indic scripts show the different
> vowels in syllables that are just V by different symbols as
> in an alphabet; they are not shown by a 'zero consonant'
> with the vowel marking of other consonants.
> 
> Also AIUI the Indic scripts show word final C by using a
> diacritic with the -a series. AFAIK there's no counterpart
> to this in the Ethiopian scripts.

Yes.  The South Asian scripts do not work *exactly* like the
Ethiopian script - which means that a workable definition of
the word _abugida_ should abstract from these differences.
At any rate, it would be bizarre to restrict the word to the
South Asian scripts at the exclusion of Ethiopian - because
it is the latter what the word was originally coined for!
 
> > Granted, I can't read Amharic, but looking at a chart of
> > 
> >  the glyphs arranged with consonant rows and vowel
> > 
> > columns, the schwa-column syllabograms appear to be the
> > most basic, and all of the other vowels appear to be
> > marked by additions to the schwa-column glyphs, like
> > added tails. There doesn't seem to be a distinct schwa
> > element that can be separated from consonant elements;
> > the different schwa-column syllabograms don't have that
> > much in common with each other.
> 
> Yes, that's what the Wikipedia page seems to show.  But as
> set out on the Omniglot page, the [a] column is the basic
> one and the [ə] (shwa) column is a derived one.  This is also
> what Hans Jensen shows in "Sign, Symbol and Script" and what
> is shown in "The Blackwell Encyclopedia of Writing Systems"
> by Florian Coulmas.
> 
> I guess what you mean is that an the basic series in an
> abugida has an inherent vowel, namely /a/ (which may be
> pronounced [ə], and that _other_ vowels will be shown by
> adding marks to the basic symbols.

Right.  There is a default vowel encoded together with the
consonant in the basic glyph.  In South Asian, you add a
diacritic to suppress the default vowel; in Ethiopian, it
seems to be left ambiguous.

> But even with this stricter definition above, the  Canadian
> Aboriginal syllabics are an abugida in that one can consider
> the -a series to be basic, and the -e, -i and -o series to
> be formed transformations of the basic series.

Yes.  It is similar, though a rotation is not the same as an
extra twiddle on the glyph.

> But to me, the important difference is between those
> syllable scripts where:
> (a) all the glyphs are unique and not analyzable into C and V;
> and:
> (a) the glyphs are analyzable into a part which denotes the
> consonant onset (or lack of onset) and a part denoting the
> vocalic nucleus.
> 
> But if it is wished to restrict 'abugida' to a subset of (b)
> in which the vowel nucleus is shown by a zero sign, then we
> are left with what to call the superset (b).

"Alphasyllabary" could work.  Thus, "abugida" refers to scripts
of the Ethiopic and South Asian types, and "alphasyllabary" to
a broader category which includes Cree (and its descendants)
and Hangul.
 
> [snip]
> 
> >> The Indic scripts do differ from the Ethiopian abugida
> >> 
> >>  in that the basic consonant symbols denote consonant
> >> 
> >> followed by inherent /a/, and the symbols for V symbols
> >> are, in effect, alphabetic.  The term _alphasyllabary_
> >> is often used to denote the Indic scripts as they share
> >> alphabetic features with that of a syllabary, in the
> >> broadest sense of the word.
> > 
> > And the basic Ethiopian symbols don't denote consonant
> > followed by inherent schwa?
> 
> What I meant is that the Indic vowels without any consonant
> onset are shown by unrelated symbols as they are in an alphabet.

Yes.

> [snip]
> 
> > I thought a writing system that has distinct consonant
> > and vowel symbols, with both obligatory, was an alphabet.
> > Is that not true?
> 
> Not IMO if the consonant & vowel bits form a composite
> symbol.  Even in the Indic & Ethiopian abugidas, the short
> can be interpreted as a null symbol.
> 
> > It  sounds  like  if  you  set  syl la bles  a part from
> > each  o ther,  say  by  put ting  spac es  be tween them,
> > then  by  your  de fi ni tion  the  La tin  al pha bet
> > would  be come  an  a bu gi da.
> 
> That, with respect, is just silly.  All you have done is
> write a string of _discrete_ alphabetic symbols with
> variable white space: narrow within syllables and wider at
> the end of syllables or after punctuation marks.

Right.  I wouldn't call syllable-spaced Latin alphabet an
"abugida".

> BTW some Roman inscriptions did something similar 2000 years
> ago: white spaces between letters, but dots at the end of
> syllables.  But no AFAIK has ever suggested that made the
> Latin script an abugida    ;)

I have never heard of such a proposal, either.

> By that sort of arguing one might just as well say that
> English writing is polylogographic simply because we put
> wider whiter space at the end of words (not even after
> morphemes)!
> ==================================================
> 
> On 25/11/2011 21:05, Jörg Rhiemeier wrote:
> [snip]
> 
> > 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]
> 
> > 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 eatures, linearly arranged

Peter Cyrus asked for examples, thus:

Logography: Hanzi, Eyptian hieroglyphic, Mesopotamian cuneiform
Syllabary: Hiragana & Katakana, Old Persian cuneiform
Abugida: Ethiopic, Devanagari & other South Asian scripts
Abjad: Arabic, Hebrew
Alphabet: Latin, Greek, Cyrillic, Armenian, Georgian
Featural abugida: Hangul (sort of), Shwa syllabary & abugida gaits,
  Tengwar with Tehtar
Featural alphabet: Tengwar with full vowel letters (as on the
  Moria Gate), Cirth, Shwa alphabet gait

> I go along with most of that, except to observe that
> designers of featural systems seem IME more often to be
> concerned with phonetic features rather than phonemes.

Some, yes.  Some use only those features which are phonemic
in the target language.
 
--
... 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