[Spellyans] Spellyans, KS, and reconstruction

Michael Everson everson at evertype.com
Mon May 28 19:26:46 IST 2012

stOn 20 May 2012, at 12:40, Daniel Prohaska wrote:

>> One of the most aggravating things about "reconstruction" in our context is it practically asks for every bloody KK spelling that turns up in the SWF to be examined.
> I agree that it was not the best decision to use KK as a starting point rather than examining every word a new and starting fresh. May I remind you that you did the same with KS? You simply used UCR as a starting point. 

That isn't true. We used the SWF specification as our starting point. KS is not UCR with a set of derogations. KS1 wasn't even UCR with a set of derogations. 

KS took the SWF specification for rules about spelling, and derogated from that. 

We also did not start out by examining every word according to its etymology. 

> I see no problem, however, with allowing the study of historical phonology to inform the study and reconstruction of Cornish for the purpose of its Revival. 

I see a problem when it is used to shape the orthography, rather than to check an item here and there. 

>> Was there a *Meurth? The texts don't show evidence for it. 
> Well there's Welsh ‹Mawrth› which makes */œ/ all the more likely. I don't think this is really in dispute, only the formalism of saying the attested  form overrules the construction, even if it is very likely.

How do you know it was likely? There are plenty of things in Cornish which are not in Welsh and Breton. And there are things in Breton which are not in Cornish and Welsh. And there are things in Welsh and Cornish which are not in Breton.

> It is not so much the question of whether one is more 'correct' than the other, but rather how to deal with this issue in the case of Revived Cornish. Taking your view to the extreme and used only attested spellings,

That has NEVER been my view. 

> it would be very difficult to achieve a standardised orthography in the first place.

And yet as I have said, it is not 1904 (or even 30 years prior). We are not starting from scratch. 

> How far to take this issue of allowing non-attested spellings for the benefit of standardisation is a matter of discussion, consensus and agreement. 

If you base a Cornish orthography on exactly what you would expect if every regular sound change implied by modern Welsh and modern Breton were implemented, then what you are going to get is a nice tidy conlang, which is not the same thing as a normalization, regularization, and standardization of the corpus. I guess we don't have a *Murth attested, though. Must we have *Meurth? Isn't Merth good enough?

>> They do not publish etymologies justifying their tinkering; they expect everybody to take it on faith.
> No, anything can be doubted and discussed. there's no faith involved. Just some capacity of informed judgement. 

Where is your list of corrections you have applied to the SWF on the basis of your examination of the KK material? If there isn't a list, no one will know what corrections you have applied, or why?

>> "Ah, yes, those fellows over there, they know their Brythonic, they're doing all the checking of the reconstructions just fine." But there aren't any publications showing the work.
> That's not entirely true. Albert, Nicholas, Ben and Talat are publishing, tackling questions such as vocalic alternation, consonants, prosodic shifts etc., etc. . 

In the general broad view, yes, Dan. But not in examining individual etyma in the SWF. You're going through the KK wordlist, it seems, and judging George's reconstructions as against the SWF specification. All behind closed doors, with all of that getting in to the editor's books. But you're not documenting it for the public. 

>> Eric Hamp used to publish paragraphs about individual etymologies. Fascinating stuff. Made his case. Do we have this in Cornish linguistics? No. Are changes being made already to the SWF in light of such unpublished work? It appears so, at least in the background. 
> Fine, perhaps this is something Cornish Studies departments at e.g. Exeter can concentrate on… 

You're missing the point.

>> In Spellyans we have worked since 2008 in an endeavour to meet the brief quoted above. We didn't wait five years for... whomever to use the SWF and find out what is wrong with it. It was obvious from the publication of the SWF spec what was wrong with it.
> Michael, you had a very clear agenda then, some of your views on Cornish spelling and phonology are ideologically charged (as are mine!) have to do with a preference for phonological theories that can be interpreted to fit the evidence, but so can other theories.

I'm fairly sure that the phonological model we use for the revived language is robust and practical. We do not indulge in "aspirational" phonology that has no connection the revived language. And we manage not to have to deviate much from the look of traditional Cornish.

> There are certain features you wanted to see in the SWF and were so adamant about that you derogated from the SWF right after the AHG meetings. 

We did not. We began discussing the need for derogation and the shape that derogation might take after the SWF specification was published. Until then we did not know what would or would not be in it. There were things promised to be in it which were not, for example. 

>> It had a Traditional Form that used non-Traditional graphs (the ban on final -y in stressed monosyllables).
> I agree this is very unfortunate. 

So we derogated.

>> It shifted ambiguity from ‹u› to ‹o›.
> And ended up with both ambiguity in ‹u› as well as ambiguity of ‹o›. This needs to be tackled in the 2013Review. 

We decided to tackle it in 2008.

You see, the Cornish Language Partnership and MAGA do not own the Revival, or the language, any more than the Kesva did. 

They were bound to use the SWF unchanged for five years. We were not. We were free to use UCR or even KK if we wanted to. Instead we wanted to contribute *meaningfully*. That meant accepting the political SWF for what it was and improving it on the basis of linguistic scrutiny. 

>> It introduced ambiguity between the bÿs/bës words and the bys and res words.
> Yes. I hope this, too can be tackled in the Review. But your own view on the bÿs/bës words, even the phonetic transcriptions you use, are to some extent ideologically charged. Whether archived, achievable or not, there are those who aspire to make a three-way-front-vowel distinction in RC, something that Nicholas says was abandoned early after the prosodic shift and was no longer valid for Middle Cornish.

Unachievable phonemes are of no interest. They are not practical. And they are not necessary, as those who are aspiring to them but not achieving them are able to communicate with other Cornish speakers just fine. 

> But there are people who do not espouse Nicholas' views on the PS, rightly or wrongly.

If it is wrongly, what is the point? Do you believe that his dates on the Prosodic Shift are right, or don't you? Don't sit on the fence, Dan. Tell us. Let us know where you stand. I have looked at the arguments and I agree with Nicholas that the late date Ken gives for the shift is ludicrously late, and entirely implausible. 

That means an early date for PS and that leads us to a phonology which is pretty much just what Jenner recommended in 1904, and what we recommend now -- and what revivalists actually do. All the theory and even the implementation of KK in the 1980s and after hasn't changed that phonology. 

> This is my problem with many KS spellings. They're not neutral in the context of their spelling system, they are very much choices made on the base of Nicholas' views on historical Cornish phonology.

And the views of me, and of others, vis à vis the reality of the revived language. In order to write we could not sit on a fence. It is impossible to try to have an orthography that supports both Ken George's unachieved phonology and the real phonology which I hear at the Lostwithiel conferences, where everybody from all sorts of backgrounds is speaking. That's one of the reasons I go to the Lostwithiel conferences -- to listen to the Cornish spoken by people. 

> And some of the KS spellings "enshrine" this theory in spelling, whether it's the i/y/e distribution, final voiceless/voiced fricatives etc. etc. etc.  

Would you rather that the SWF "enshrines" your theories, then? You proposed in three lines an i/y/e distribution on this forum in July 2008. It wasn't accepted by the Spellyans group in part because it would have meant a major shift from the basis i/y distribution used in monosyllabeles in the SWF, and it was part of our brief to tread lightly on the SWF. 

You've never since tested your distribution by showing people texts written in it. If you produce some, perhaps on the basis of some KS text or other, we might put them on the website for scrutiny. But I don't believe still that your distribution is practical or simple enough to taught easily to learners. 

With regard to the distribution of fricatives, if you want, we can start a thread on that and discuss it publicly. As far as I can tell you have made claims about Lhuyd but have not backed those up with data, or with a discussion of the phonotactics of your theory in the whole context of the language. I would welcome such a discussion.

>> It omitted the useful graphs ‹ai› and ‹au› for a set of loanwords.
> Yes, I would like to seem them, or equivalent graphs introduced in the SWF, though I may have some minor issues with their distribution across lexemes...

Make your list and share it, if you want people to evaluate it. 

>> It imposed KK's etymological vowels, particularly difficult in unstressed final syllables (kegin v. kegin).
> ‹kegin› v. ‹kegin›, not sure what you mean here.

‹kegin› v. ‹colyn› etc. Sorry for the typo.

> I think this is definitely something we can sort out in the Review, but my approach would be to take the two dialects of RC into account and find a way to represent the differences and commonalities where they are useful to speakers of these respective dialects.

I'm afraid that your approach would increase the distance between RMC and RLC dialects. Needlessly.

>> It lacked a phonotactic rationale for its use of -dh/-v in unstressed monosyllables. Its distribution of ‹i› and ‹y› was entirely arbitrary (unless one had learnt KK).
> No, this is an opinion, a point of view. I have already stated, which you can take on faith ;-) or not, that this has to do with the phonological and positional contrasts that Cornish was capable of.

I don't know what your "it" refers to here. 

> Sure, native Cornish phonology was augmented by loans from English which didn't conform to the native system, but the native phonology contrasted the plosives only in initial position whereas in internal and final position there was no phonological contrast.

List a dozen examples, please. 

> The fricatives however retained this contrast, at least in internally and so it is disputable whether it was or wasn't given up in final position.

How can you tell? List your examples, please. 

> Yes, your phonotactic rational for an initial and internal stressed voiced v. unstressed voiceless contrast is neat, but it is not unambiguously borne out by the evidence in the texts.

First, I have asked you to explain your phonotactic model. Second, there is ambiguity in nearly EVERY feature of EVERY vowel and consonant in the texts, so why should this be particularly damning of our model? 

> An orthographic solution in RC for a distinction may be called for, and leave the ideological approach to a set of rules that can be applied. This is what a standard (maybe not for everybody but for the greatest number possible) should aspire to, and leave "being right all along" out of the discussion where the evidence is not absolute and unambiguous. 

The safest thing to do in this case is to spell -th as the texts do, and if you want to have a rule that people who have learnt KK can voice these if they wish, you may. But I tell you, at Lostwithiel I have listed with particular care to speakers who used KK orthography, and they do NOT voice -dh in final unstressed position. 

>> From the end of the first month after announcing Spellyans we had 27 members; we now maintain 51 observers. People like Albert Bock and Ben Bruch and Tony Hak, who are concerned with (i.e. in positions of authority concerning) the SWF, do not participate, and do not observe as members. They don't see the discussion: they don't engage with the Traditionalist community. Have they read any of the publications in KS? I wouldn't know. 
> And maybe they have their reasons for not participating or even leaving this discussion. Maybe they feel it's too one sided. maybe they feel you are coming on too strong on certain issues and fining you difficult to talk to at times. Not speaking for them, but saying what I feel occasionally.

We have never had meaningful dialogue and discussion with them. What we say seems to fall on deaf ears. 

> I can get over it, but maybe others don't find this so easy. This, is probably also the reason why several people didn't want you on the AHG meetings. I know nothing concrete, but it's the feeling I get from people who said they had no desire to sit at a conference table with you. I'm certainly not one of these people, but I am wondering, if an exercise is diplomacy wouldn't have got some of your points across more easily... 

One can hardly address the might-have-beens of 2007.

>> What is an orthography for? A Cornish orthography that has all this etymology built into it... is... to help people do what? To learn Cornish? Or to use the Cornish they've learnt to bolster their investigations into Brythonic? 
> Again the pro- and contras-etymology-thing is now becoming idealogical.

No. It is a matter of practicality. There is no reason to write ‹kegin› because of a medieval Latin ‹cucîna›. It serves no function but to confuse the writer as to when to write -in and when to write -yn, which is common in many other words. 

And here's something that speaks to your analyses of Ken George's etymological spellings. He has ‹kolen› 'puppy'. SWF now has ‹colyn›. Why the change from -‹en› to -‹yn› in that word? Ken George also has ‹kegin› 'kitchen', Why doesn't the SWF change this -‹in› to -‹yn›?

We have ‹colyn› pl ‹kelyn› and ‹kegyn› pl ‹kegydnow›, by the way. 

> The texts don't distinguish between ‹y› and ‹i› although certain differences must have been there to warrant LC forms that occur, deciding the vowel in the unstressed syllables has to do with this issue as well. You distinguish ‹mis› from ‹bÿs›, but the texts have ‹mys› and ‹bys›.

Yes, but the texts do not have ‹mes› alternating with ‹mis› as you know. We distinguish ‹mis› [miːz] from ‹bÿs› [biːz]~[beːz]. The texts also have ‹bes›, ‹beys›, and ‹bis›. 

> You also claim that the two vowels had fallen in with each other in some varieties of Cornish (or certain texts thereof), but the same can be said for the ‹melin› : ‹melyn› contrast.

I don't follow you. We have ‹melyn› 'mill' pl ‹melydnyow› (variant ‹belyn›!), and we have ‹melen› 'yellow' with a derivative ‹melender› 'yellowness', and a bird name ‹melenek›, pl ‹melenegas› 'Carduelis chloris, green linnet". The choice of written for the [ən] in the final unstressed syllable of the singular hinges upon the colour of the vowel when stressed. It's true that the derived words for 'yellow' are neologisms of Nance (or earlier), but nevertheless they are there. In this matter we followed the "colon"/"colodnow" rule we came up with in KS1, if you recall, when we were tending to default to -an in final unstressed syllables. 

There is no contrast in the singular. The words rhyme in the singular. Our claim about [ɛ] and [ɪ] being often in free variation is more relevant to stressed syllables. Was this not clear? It's in ‹lyver›/‹lever› where that seems to operate (from a very early period indeed). 

The point is that for this very large class of words it is difficult to see any sense in giving both forms or in writing them ‹lÿver›/‹lëver› because the class is so large. Much larger than the ‹bÿs›/‹bës› class, which is important to mark because without doing so the ‹bÿs› words regularly get mispronounced with a short vowel, which is a serious mistake. [lɪvəɹ]~[lɛvəɹ]: free variation [biːz]~[bɪz]: error.

A standard orthography should mark what *must* be distinguished. 

(Regarding the linnets, as far as I can tell Ken George re-analysed Nance's ‹lynek› pl ‹lynegas› to give his ‹linoges› pl ‹linogesow›, and he saw fit to change ‹melenek›, pl ‹melenegas› 'green linnet' to ‹melynek›, pl ‹melynoges› 'goldfinch'. I don't see what's wrong with ‹melenek›, pl ‹melenegas›.)

> If ‹i› and ‹y› are distributed according to given rules in RC, but were not in traditional Cornish, you are doing something different from what the Cornish scribes did.

In making *any* 20th/21st-century normalization we are all doing something different from what the Cornish scribes did. 

> So, the arbitrary ‹i› : ‹y› distribution in TC (traditional Cornish) is standardised according to certain rules in RC, big deal, so discuss how to standardize them.

We've done that. We've discussed them. Years ago, Dan! We had a "tread lightly on the SWF" policy in 2008, and that is one of the reasons we did not accept your proposed redistribution. We chose a different one. And we implemented it. How long do you think this ought to be in flux? 

> I don't see much pick up for the KS solution, so let's discuss how we can bring our respective views together.

As far as "pick-up" is concerned, Desky Kernowek has only now been published. But, seriously, what is it that you want to do? Do you want us to trade the re-distribution of i/y/e which we have implemented and published in, for your re-distribution, for which we have not actually seen any sample texts since 2008? If you produce sample texts, do you think they will convince us? (By "us" I mean "Spallyans".) Or are you open to altering your view on this and coming on board? 

See there are four options with regard to the SWF anyway.

1) Do nothing with i/y/e in the SWF and leave it as the KK mess it is. I suspect that Maga and former KK users would prefer that.

2) Implement the Spellyans re-distribution which has already been seen and read by many people and which is quite easy to explain, to teach, and to use.

3) Implement your re-distribution (perhaps after seeing examples) which would not only change KS but change the SWF and probably confuse everybody.

4) Enter into SWF discussions offering both 2) and 3) as incompatible options, which is likely to guarantee that 1) will be chosen. 

> That's what I meant when I asked you whether you were open to compromises. You don't seem to be, you seem to want to push the KS solution onto the SWF because you see them as the only possible (or at least very best way) of dealing with the issues. This again, is a matter of opinion, and as you know, opinions can differ. 

And I asked you, what changes or compromises do you think we ought to be open to? I mentioned being open to ‹ei› replacing ‹ÿ›~‹ë› (not thinking it very likely), and you said that it would be impossible because ‹ei› in hiatus is [əɪ] in RLC. (In terms of orthographic design there would be complementary distribution of the sounds, so it would in fact work.) 

I do hope that we have a discussion on dh/th and v/f, because I believe that you will eventually see that the weight of the evidence is against you. 

>> Have we done a good enough job? Is there a list of etyma that we still need to argue about? Are there serious shortcomings that render KS unworthy of academic support? 
> You've done a great job in your publications and the amount of Cornish being published.

Which I judge to be better for the Revival than doing nothing for five years. 

> You have not done a great job in supporting a consensual form of Cornish, because you don't agree with some issues that in your view render the orthography unusable.

If it is unusable, its use cannot be justified. It has mistakes in it, and things which are not Cornish. Why should we consent to use something like that? We were kept out of genuine participation in the discussions which formed the SWF. We owe that process little. 

The SWF compromise was remarkable and admirable from a political point of view. A number of KK underpinnings of the SWF are as wrong in the SWF as they were in KK. Having edited two editions of SWF/K and SWF/T grammar books, we are quite well aware of those problems. There are things in the SWF you can't learn, unless you know the rationale for them in KK. Those things tripped us up again and again. 

Get it? We (Ray, Nicholas, and I) failed to be able to write accurately in the SWF, not because we didn't know Cornish. But because we didn't know about the arbitrary choices Ken George has made. 

We *have* supported the compromise, Dan. We (Nicholas and I, and perhaps others in Spellyans who may wish to name themselves) have moved on from UC and UCR and do not see a road back to them, because the corrected form of the SWF is at once more accurate and more inclusive than either, if one sees at its core the period 1550-1650 as a *standard*, a magnet attracting both ends of 1500 and 1700 toward a centre. From an ideological perspective, that is what the Revival needs. Not more fragmentation. 

> There's no list of etyma that we need to deal with in KS,

I am not sure what you mean by "deal with". KS1 was a proposal for discussion. KS has been a solution to problems in the compromise orthography. The spellings in KS are facts, and I would recommend that you not ignore things like spellings in the Beybel Sans. By which I mean, you should not "deal with" UC and UCR but ignore KS. We have published well over a million words in Cornish. They should be taken into consideration. 

> but as I potter along with my dictionary and check and come across KS forms that I may find problematic, I'll bring them to your attention for further discussion, if you want.

You are welcome to do so. You have always been welcome here though you seem to have recused yourself for a long time a few years ago. 

> I do the same with the SWF and Albert's dictionary. We are in frequent communication about such issues. 

I see. 

> As far as shortcomings,

Where? In the SWF?

> I don't think there are many, but there are some issues that cannot be answered as there are no traditional speakers of Cornish and only those would be able to answer these issues.

In which case, where a choice must be made, a sound phonotactic theory should bolster the choice. I really would like to see the th/dh f/v issue put to rest. 

> So in the context of RC they become matters of opinion and later are somewhat ideologically charged.

Perhaps. Nevertheless some form of standardization is required. In the matter of th/dh for instance, our Standard Cornish is founded on what we hear amongst speakers of RC from all camps, and we have a phonotactic theory which fits well with the evidence of the texts and with the realities of RC. In Common Cornish on the other hand we have a distribution here which is based on cognates in Breton and Welsh, and I have yet to see the theory that explains how those fit in with the behaviour of the other consonants in similar contexts. 

> This is something where I think we can find some room for common ground to allow the SWF to somewhat open for interpretation, while giving users guidelines by rule to make them pronounce the language as correctly as we can determine at this date.

I believe what we believed when we developed KS1: That the orthography should be as accurate as possible while remaining true to the texts. And we have to take the phonology of the revived language into account. It's not 1904. People are already pronouncing the language. (Yes, you will remind me of their English [əɯ] diphthongs and so on. We know about those and of course we advise the use of "pure" vowels.) 

I hope that when you receive your copy of Desky Kernowek you will study the section on phonology and orthography very carefully. I believe that we (Spellyans) have devised a system which does exactly what you describe above: It has rules sufficient to enable a learner to accurately pronounce words which he or she sees, and which enable him or her to write Cornish accurately.

Michael Everson * http://www.evertype.com/

More information about the Spellyans mailing list