CJ as pidgin; phonemes

[Henry Kammler]

> I also tend to see CJ as having quite a lot of phonemes.  As the language
> is one which absorbed only partially the features of any particular donor
> language, and as it only recently was certainly a pidgin, CJ is very
> likely to have a great many various sounds in it -- Sounds which have not
> become regularized into a neat, traditional structure of oppositions and
> symmetries such as one is taught in linguistics school to look for.

This is exactly why one *could* argue for a simple writing system. I.e., if
the range of sound variants would be more or less a question of where you come
from. Take English. It doesn't have a simple orthography but it serves L1 and
L2 speakers around the globe although pronunciation varies to an incredible
extent.

> Furthermore the possibility of strictly normalizing even Grand Ronde CJ
> may be a slightly remote one.  By this I mean to point out that there are
> varying pronunciations for quite a lot of words.  I'm unsure what
> principle would lead us to choose either /mILEt/ or /mILayt/ as the
> 'standard' form of <mitlite>, when we can simply record that both are much
> used and understood, and perceived to be the same word.

Different strokes for different folks, of course. I don't think the distorted
writing systems used by early recorders of CJ is any good for teaching the
language (the same with English: if you don't hear it spoken the orthography
is useless) so one should argue for a writing system close to the phonetics
employed by fluent speakers.
For practical reasons, e.g. in building a lexical database, it would be
advisable to have a header in a normalized writing for each entry. Under the
header you can still record all occurrences of the given lexeme in a phonetic
writing. The more material you have, the easier it will be to sort out to what
extent the variations are predictable, are conditioned by dialect variation or
actually have sth. to do with grammatical processes. Maybe it is rather
unlikely but variations *of the type* /mILEt/ vs. /mILayt/ could actually
convey differences in meaning, like intensity or duration, if you find the
alternating forms in the speech of one and the same speaker. For a database
entry /miLayt/ would be a good candidate, then you can reserve data fields
under this header, one for /mILayt/, one for /mILEt/ etc. and provide as much
additional info as you need with each alternating form. Modern software can
handle all these aliases neatly and it is not the paper work anymore that it
used to be. :-)
I feel "mitlite" should only have relevance for philology or for teaching the
students how to use historical sources ("you know there was a time when them
Boston people didn't know how to write properly ..."). BTW this writing
variant implies that /mILayt/ was the pronunciation at the place and time when
it was taken down as "mitlite".

> By the way, I *strongly* feel that it's important to do what Tony does in
> writing exactly how each CJ word is pronounced in his community.  However
> that is a separate question from this one.

Agreed. What if Boas hadn't used what he considered phonetic writing in his
"Chinook Texts" and "Kathlamet Texts"?

Henry