Re: [indic] Re: Tamil Anusvara (U+0B82) glyph shape [ Re: Dot position in Gurmukhi character U+0A33]

From: Philippe Verdy <>
Date: Fri, 10 Feb 2012 04:38:53 +0100

2012/2/9 Michael Everson <>:
> On 9 Feb 2012, at 13:49, srivas sinnathurai wrote:
>> Dear Michael,
>> It is better if you do do some research before commenting.
>> you say Tamil has 10 vowels.
>> No Tamil has 5 basic PoA for generating vowels.
> Please learn what a phoneme is.

And not just phonemes.

Because spoken languages alos use other phonetic features (even if
they are not always constrasted semantically, but the only occur due
to the paticular phonology of a language or one of its regional

For example, you'll want to learn about tones (and tonemes when they
are represented in a phonetic notation), stress, length...

As well as coarticulations (in IPA, some coarticulations were given
specific symbols, but now they are preferably represented by their
components, and a tie symbol above them). Coarticulation are typically
found in European language with the letter "x", even if this letter is
ambiguous and represents several phonetic realizations like "ks" or
"gz" (in many languages, they are considered allophones, because they
don't contrast, and they are chosen according to the phonology), plus
some others which are simpler phonemes with a single articulation like
"h" or "k".

Then you'll want to add breath extensions (something that transforms,
e.g., "p" into something like "p-h", still contrasting with the "pf"
coarticulation. Depending on linguists, those breath extensions are
not always considered as coarticulations (because the extra breath is
not considered as a real articulation).

Alphabets are built by trying to not just represent a particular
phonetic, but a range of different realizations that are considered
allophones and encompass various regional accents. They are a sort of
"unification" of the phonetics at one time, but they also play a role
to help mutual understanding with reading, using some common roots and
a limited set of variants.

For this reason, there's absolutely no language in the world that uses
a pure phonetic transcription (except possibly some minority languages
for which there's still no orthographic tradition or standard). This
would not work. All alphabets use letters to represent a more or less
wide set of allophones; then the languages continue to evolve
(including in their phonology and distribution of dominant accents),
so that the same letters are kept (to preserve semantic roots), even
though they no longer are representing true allophones.

At the same time, foreign sounds will be added to the set of
allophones, and approximated by existing letters, without necessarily
adding a letter to the alphabet. As long as this does not create
confusion with too many homonyms, such approximations will remain
valid and understandable.

It will take much longer time for a language to accept foreign
phonetic distinctions and include them as distinct phonemes as part of
their phonology, even if today it seems easier to extend the alphabet
with new borrowed letters or optional diacritics, used first by
"purists" and linguistic theoricians. One good example (found in
French) is the current development of the transcription of Japanese
names using macrons to transcribe long vowels (French has lost the
phonemic distinction of long and short vowels since long, but not
Japanese, as well as many Asian languages) : there's still no
agreement to use those macrons, because French words absolutely don't
need it (short and long vowels are considered allophones as the length
vary freely depending on each person or regioanl accents, or depending
on context of use for the same semantic words, and thus, it is
difficult for French natives to hear and pronounce a distinction
between those allophones that only vary in length, tone or stress).

But this does not mean that the number of letters in an alphabet shows
the extent of distinctive phonemes : the orthography also uses
distinctions using additional mute letters or diacritics (to break the
usual contextual rule of transcription), as well as digrams or
trigrams. Then the orthographic traditions to preserve the semantic
roots that associate different words in the same family play a role in
breaking the initial tie of the phonology (not the phonetics!) and the
orthography (whever a language tends to preserve these roots is highly
dependant of the language : for example, French preserves these roots
much more than European Portuguese, which itself preserves them much
more than Brazilian Portuguese and creole languages that have a very
relaxed orthography more tied to the current local phonology than to
the preservation of semantic roots).
Received on Thu Feb 09 2012 - 21:47:07 CST

This archive was generated by hypermail 2.2.0 : Thu Feb 09 2012 - 21:47:09 CST