phonology of Plan B

Joerg & Ray concur that Plan B does not have 16 phonemes with consonantal and vocalic allophones:

Jörg Rhiemeier, On 02/07/2007 21:08:
> On Mon, 2 Jul 2007 17:01:25 +0100, R A Brown wrote:
>> Jörg Rhiemeier wrote:
>> [Plan B etc]
>>> On Mon, 2 Jul 2007 07:27:29 +0100, R A Brown wrote:
>> [snip]
>>>> The language *simply is not spelled phonemically*
>>>
>>> It can indeed be argued that Plan B has 16 consonant phonemes and
>>> 16 vowel phonemes with a rule that forbids both consonant clusters
>>> and vowel clusters;
>> It most certainly has 16 consonant phonemes. Whether it has 8 or 16
>> vowel phonemes is a matter of interpretation. If phonemic status is
>> given to /r/, then we have 8 vowel phonemes plus 8 combos of /r/ plus a
>> vowel. The rule is then that syllables must be of the form: C(r)V(C)
>
> I did not remember that the "16 vowels" included 8 /r/+vowel combos.
> I'd say that /r/ is a phoneme, so there are 17 consonants, one of them
> being /r/ that patterns fundamentally different than the other 16, and
> 8 vowels.
>
>>> this is probably a better analysis than saying
>>> it had 16 phonemes each with a consonantal and a vocalic allophone.
>> That IMHO is a simply a ridiculous analysis!
>
> Concurred.  It is very far-fetched, and if someone was to analyse
> a natlang that way, he should ask himself what he is doing.  It is
> like saying that English had a phoneme that is realized as [h] in
> syllable onsets and as [N] in codas - only worse (surely, the
> English word _hang_ is not a phonemic palindrome!).
To my eyes, the most sensible analysis of Plan B is the 16-phoneme dual-allophony
one. It's the analysis you get if you apply basic principles of
contrastiveness; and the alternatives that Joerg & Ray put forward fail to
account for the systematic equivalence of consonants and vowel phonemes.

What is unnatural is Plan B phonology itself, not the 16-phoneme analysis of it.

--And.

Reply