[Spellyans] RLC <h> for <gh>

[nicholas williams]

Besides, Jon, in your hatched job on KK in Cornish Studies, you do not  
demand a
one to one relationship of graph to sound. You actually say (and I am  
quoting from memory)
that all we need is a fairly close fit between sounds and spelling and  
Unified gave us that.
You then go on to say that KK was an expensive waste of time and effort.
Well what is needed is this: a good (if not perfect) fit of spelling  
with sounds
even if there are a few marginal inconsistencies.

The fit of SWF with the sounds is very poor. The use of <y> is  
chaotic. Moreover final -dh after unstressed vowels is
de rigueur, except in nowyth, even though most KK speakers have a  
voiceless segment passim.

We are not devising a phonetic transcription, either broad or narrow.  
Just an orthography
for a group of speakers, whose whole outlook has been severely warped  
by over
20 years of orthographical and phonological nonsense.

Nicholas

On 26 Jun 2008, at 17:26, Michael Everson wrote:

> At 15:42 +0000 2008-06-26, Jon Mills wrote:
>
>> Ultimately, it needs to be decided what the relationship is between
>> KS orthography and the pronunciation of Cornish.
>
> Which Cornish? :-)
>
>> If KS is intended to closely represent the pronunciation system of
>> Cornish, it must be phonemic in principle.
>
> It also has to be more, because there are dialect differences it must
> account for. <u> for instance represents /y/ and /i/, since one
> dialect has no /y/.
>
>> The phonemic system is the principal element of the phonological
>> system where orthography is concerned.  An orthography that is not
>> theoretically sound is just waiting to be shot down.
>
> I'm not much of a linguistic abstractionist. I don't believe i've
> defined any "theory" for KS. I've looked at the data and worked with
> Neil and Nicholas and the rest of UdnFormScrefys to put something
> together that "does the job". Perhaps that matches a theory?
>
>> The reason that we do not like KK is that it is theoretically  
>> unsound.
>
> On several grounds. First, George reconstructs Cornish phonology as
> he thinks it ought to be, introducing phonemic geminates and three
> conditioned vowel lengths as the core of a system which all the
> evidence suggests had short or neutral consonants  and two phonemic
> vowel lengths. Second, he insists on a theory that a one-to-one
> relationship between graphs is "good" and anything else is "bad".
> Third, he uses untraditional graphs which piss people off. (Thank
> goodness he did. It really let people dig their heels in against it.)
>
>> Umbrella graphs are being used to bridge dialectal variation not
>> allophonic variation. Nothing wrong with that per se.
>>
>>> At what level of abstraction are you talking?
>>
>> Phonemes are abstractions. Allophones are realisations.
>
> OK, but I'm not sure how this affects my task. Though KS 16 describes
> the phonemes and also uses IPA to show phonetic realizations. So
> maybe I took it into account.
>
>>> We have Revived Cornish and some pretty good
>>> recommended pronunciations that are actually
>>> feasible. What are you getting at? Starting from
>>> scratch and trying to reconstruct? Or?
>>
>> Well, I hope that we do not have to start from scratch. We have some
>> recommended pronunciations, didactolects. Just how good, however
>> .... We do not seem to think much of KK's accompanying orthoepy.
>> Theories concerning the pronunciation of Cornish are in a state of
>> flux and are likely to remain so for the forseeable future. An
>> orthography is not going to be robust if it is founded on shifting
>> didactolects. A more stable foundation is needed.
>
> I think the phonology of KS is based on an analysis of what people
> are doing, with some corrections where we would like to encourage
> better pronunciation. For instance, we recommend a flapped /r/
> intervocalically for people who can manage it and the approximant
> elsewhere. Neil Kennedy has taught this to his students and it sounds
> good and is not unfounded in Cornish dialect. I don't know what
> *other* recommendations about /r/ we would like to make instead.
>
> As I see it UC recommendations are improved by UCR recommendations
> and those were improved by working with RLC speakers, studying Lhuyd,
> and making the effort we began in September 2006.
> -- 
> Michael Everson * http://www.evertype.com
>
> _______________________________________________
> Spellyans mailing list
> Spellyans at kernowek.net
> http://kernowek.net/mailman/listinfo/spellyans_kernowek.net