--- In
phoNet@yahoogroups.com, Piotr Gasiorowski <gpiotr@i...> wrote:
>
> Just between us -- IPA _is_ a bungled job, isn't it?
Are there no plans to ever revise and improve it?
For years I've worked on first one variation, and then
another, of my own phonetic alphabet, but have never
been quite satisfied with any of them.
For one thing, I've always wanted the symbols to
be systematically built up from smaller components,
each component representing a feature of the sound,
rather than by randomly coming up with say a hundred
distinct symbols for a hundred different phones.
This would allow, among other things, in a phonemic
representation, for instance, of a language where
nasals assimilate the position of the following stop,
for one to omit that part of the symbol encoding
position. In fact if the language has only one row
of nasals, it would be possible in such positions to
include nothing but the component encoding nasality.
I've always thought that this would allow one, in
phonemic representations, to approach the brevity
possible in a well-designed syllabary, but without
abandoning the advantages of an alphabetic system.
Secondly, I've always wanted a way to have symbols for
broader categories, such as coronal, which should
be simpler in shape than those for the more narrowly
defined dental, alveolar, palatal, etc., and to have
the symbols for the latter systematically derivable
from the former by additions. I would want even more
narrowly defined categories possible, and for the
symbols for which to be arrived at by reiteration of
the same process. So that one could be as broad or
as precise as necessary in representations. In other
words, nearly all symbols would contain the equivalent
of diacritics, and many would have what would amount
to multiple layers of diacritics, but hopefully could
be elegantly composed.
All of this is quite possible, however when you approach
it this way, you find that all your symbols end up being
much more complex. /p/ is a voiceless bilabial stop,
but the character <p> doesn't have quite enough detail
to it to encode these three features, at least not if
part of larger system that needs to encode as many types
of features as reality requires.
So even for phonemic representations, the hope of greater
brevity mentioned above never seems to pan out. You find
yourself representing such simple and common sequences as
/pit/, /pat/ or /put/ with tangled monstrosities resembling
Chinese characters, it is very discouraging, and also makes
it difficult to convince yourself you've created anything
that can be called an improvement.
The imperative to have a systematic relationship in shape
between the symbols and sounds, and that to keep the symbols
as simple as possible, are at odds with one another. :-(
David