writing systems

Emmon Bach (bach@coins.cs.umass.edu)
Wed, 8 Jun 1994 05:45:00 EST


About Writing Systems

Here are some things to think about about writing systems:

I. Kinds of systems

Writing systems differ according to what language units are represented by
each sign:

1. one sign - one meaningful unit (morpheme): Chinese, Arabic numerals
like 1, 2, 3 borrowed into many languages. Classical Aztec?

Number of signs will be in the thousands.

Usually called logographic, ideographic, or the like.

2. one sign - one syllable: two of the systems used for Japanese (along
with Chinese characters), Cherokee, classical Mayan.

Numbers: depends on how closely the system fits the language, on the syllable
structure of the language and whether all distinctions are represented (as
they aren't in Cherokee),

Syllabary, syllabics.

3. one sign - one significant speech sound (phoneme): Finnish, many
alphabetic systems used for aboriginal languages around the world.

Numbers; a perfectly matching system would have just as many signs as there
are phonemes in the language, which ranges from about 20 up to maybe 80 or so.

Note: a sign might be a combination of letters (English "ch, th" etc.)

The Cree syllabary and the syllabary for various Inuit (Eskimo) type languages
are actually more phonemic than syllabic, since the systematic orientation of
each of the basic consonant signs signals which vowel forms the rest of the
syllable. (As Bill Poser argued at the last meeting of the Linguistic
Society of America for the Carrier system of interior British Columbia.)

There are various variants of the above. For example, English spelling today
reflects relatedness between words and is not strictly phonemic. Japanese and
classical Mayan (as I understand it) mix syllable signs and ideographs.

The above are all bound to particular languages. Various signs are used to
represent concepts independently of language: traffic signs, skull and cross-
bones for poison etc.

II. Practical considerations:

1. The number of signs makes a difference in how long you have to spend on
learning the system. Japanese kids spend a large part of their early
schooling memorizing Chinese characters. A good phoneme based system can be
learned in a matter of hours. Irregularities make for more time. Kids
learning English spelling have to spend a lot of effort on weird spelling and
spelling conventions, as far as I know there are no Finnish spelling bees.
English spelling and the knowledge of it is an important part of the system of
classifying people as "educated" and "not educated." The rigidity of the
system is fairly recent. Emily Dickinson wrote "it's" for standard "its" (as
in "its house" vs. "it's hot".)

2. You can leave out a lot of stuff and the system will still work well FOR
PEOPLE WHO KNOW THE LANGUAGE WELL. This is the point of the recent discussion
of whether it helps or not to represent tones. We linguists are tempted to
want to dseign "perfect" systems. Research like that of Russ Bernard and
colleagues is important in settling the question of whether marking tone is
helpful or just clutter. The alphabetic systems for Chinese all mark tone
somehow (either by extra marks or special spellings). Traditional Chinese
writing doesn't directly, but since different morphemes get different signs
the tone marking is implicit. There is usually more information in the
consonants than in the vowels, hence lvng t [?] th vwls wrks prtty wll u ea ou
e ooa oe! (I guess vowels are optionally marked in Hebrew.) English leaves
out stress marking (perFECT PERfect) without too much pain, often predictable.
It probably makes a big difference how the language works: frinstance, tone
works very differently in African tone langages and in Chinese and other
languages of East Asia and Southeast Asia.

3. To borrow or not to borrow: most writing systems around the world have been
based on borrowing (Latin alphabet and its ancestors and descendants, Slavic
Cyrillic (different strand of same history), Semitic (Hebrew, Arabic). There
are (I think) two opposing thrusts: to want to have something as different and
unique as possible for your own, to want to have something as easy as possible
for people who write or are learning to write another language (e.g. kids in
school).

4. Is the system to be used by people who know the language already or not?
This is crucial for the decision about how much can be left out? Note that in
many NA native communities, many kids grow up learning English (Spanish,
French) as their first language, or have done so in recent days.

5. Beauty is an important consideration, some of the ugliest systems have been
invented by linguists (Boas's system for Kwakw'ala:kUa'k!Uala where U is a
raised little u).

6. Technology: current computer and printer technology makes it possible to do
just about anything you want, including making it easy to use several
different systems (say one with one without tones marked, maybe the first in
texts and the second in dictionaries, vocaularies etc.). At present at a
price! Here's where the whole discussion of MAC vs PC etc. comes in.

Final thought: the people who are to use the system have to be the ones who
make the choices!

This is too long already.

Emmon Bach
Linguistics
UMass
Amherst MA 01002