Re: Rotokas (was: California Cheeseburger)

Racsko Tamas said:
> On 17 Jun 2004 "Mark P. Line" <mark@PO...> wrote:
>
>> I believe that Rotokas has a modern, phonologically engineered
>> orthography and that such orthographies lack subphonemic distinctions
>> (at
>> least, I can't think of a single counterexample -- hence my question).
> [...]
>
>   If you are right when you said: "I'd say that "wiliwili" is a
> loan word from Tok Pisin 'wilwil'", then "vi wiliwili vorepieriva"
> is a counterexample, since Rotokas is said to have subphonemic
> distinction between "v" [v] and "w" ~ "b" [B] ~ [w].
What evidence do we have that [w] is used when pronouncing 'wiliwili'?

I think we have to assume that it's written with 'w' in Rotokas because
it's written with 'w' in Tok Pisin. We can make further assumptions about
the most likely ways it'd be pronounced, but we've already covered that.


>   However, if borrowings are not adapted to the phonemic rules of
> Rotokas, that is the above orthography is not subphonemic then this
> means that "v" and "w" are phonemic, though. Thus whe had more than
> six consonant phonemes.
Only if you consider /x/ a phoneme of American English because some
Americans use [x] when pronouncing forms like (Scottish) 'loch' or
(German) 'Bach' (the composer).

Languages in contact are just not as cut-and-dried as some would have us
suppose.

My guess is that monolingual Rotokas speakers adapt 'wiliwili' to native
Rotokas phonology, while speakers who are bilingual in Tok Pisin might
sometimes use [w]. The latter situation doesn not mean that /w/ is
suddenly a new phoneme of Rotokas. (It might become so eventually, of
course, if Rotokas doesn't go extinct first.)


>   (IMHO the term "orthography" excludes the fully phonological
> engineering by the defintion: this is the difference between
> orthography and transciption. Orthography is designed for non-
> professional humans: it has confortable shortcuts for an average
> speaker. An average speaker cannot distinguish even between letter
> and sound, therefore, it cannot be expected that he/she could think
> during writing in a "real-time way" in a more precise distinction
> between sound and phoneme...)
The literature on literacy, applied phonology, orthography development and
the psycholinguistics of reading is full of discussions about the pros and
cons of making newly-engineered orthographies as strictly phonemic as
possible.

We couldn't possibly rehash all of that here. Google is your friend. :)


> Jean-François's analysis reveals that (A) it is very likely that
> grapheme "s" is subphonemic, the only syllable beginning with "s"
> is "si" and it is moderately frequent. And there is no *"ti".
That dataset is way too small to be able to claim there's no "ti". Also,
we have to ask ourselves why the Firchows would never have mentioned such
an obvious allophonic alternation. I don't think we can assume they
"missed" it.


> (B)
> Phoneme /g/ seems to be very rare, therefore we cannot make valid
> statements of its phonotactic behaviour.
Again, that dataset is way too small to really know how rare /g/ is.

The Firchow 1969 phonology gave frequencies of all the phonemes in their
corpus, IIRC. Have you looked there?


-- Mark

Reply