[Spellyans] gawas 'to get'

[Michael Everson]

On 8 Sep 2011, at 15:00, Daniel Prohaska wrote:

> Now it has frequently been stated by many, including Nicholas and Michael, that Lhuyd’s spellings are owing to Welsh influence.

He often shows Welsh influence, whether in vocabulary, syntax, or phonology. There is nothing wrong with observing that he does this. Nor is it surprising if he does. Every linguist of that period did the same. (As do some today.)

> But I have my reservations about this assumption. It is just as possible that Lhuyd actually heard [ð] in the words and transcribed them with <dh>.

Even if he didn't hear it, he might think he did because that's the corresponding segment in Welsh. 

> He also writes them with <th> which he read in the texts that were available to him and may also have heard [θ].

I don't think it's likely that he heard both. As a Welsh speaker, he know what [b] and [p] were, and he knew what [ð] and [θ] were. If these sounds had been in free variation, one would have expected him to have made a note of it, because that would stand out as unusual to him. ("We'd never do that in Welsh...")

> This may mean that words final /ð/ may have been voiced in some environments (e.g. at the end of stressed syllables always, at the end of unstressed syllables in voiced environement), and unvoiced in others (e.g. absolute final position in unstressed syllables and in unvoiced environemts).

Keith Bailey plays the Breton game in his version of KK, writing voiced consonants in final position in unstressed syllables and evidently devoicing them in absolute auslaut and voicing them in sandhi. Your suggestion is more or less the same thing, isn't it? I'm not suggesting that you're a conlanger like Keith, but the argument is the same. And in this matter, he's consistent, writing mab and methewneb, neb and heveleb, hweg and karreg, mog and galloseg, dov and warnav, nev and enev, ov and esov, gradh and nowydh, badh and gelwydh. 

On the other hand we write mab and methewnep, neb and hevelep, wheg and carrek, mog and gallosek, dov and warnaf, nev and enef, ov and esof, gradh and nowyth, badh and gelwyth, according to the standard analysis: that final consonants in unstressed syllables are devoiced. Your suggestion, Dan, that ð/θ operates differently from the other consonants isn't, in my opinion, convincing, or tempting. The standard analysis is easier for learners, and with the pronunciation of Revived Cornish. Everyone I heard at Gorseth Kernow says [ˈɡɔrsəθ] -- everyone, including those who learned KK. Why should there be a special rule for -dh that people can devoice it? Why shouldn't they just write -th as they say it?

I think that the only reason we're discussing this is that Ken George thought to impose his interest in etymology on Cornish. That's the only reason -dh is in the SWF in unstressed syllables, and the only reason you bothered to look for a rationale for it. You say you think you've found one in Lhuyd; I say that the evidence really doesn't support that view. 

> Since it is secure that Cornish at some point in history had voiced /ð/ in these words and those with similar etymology (derived from British /d/ or /j/ in hiatus), it may be prudent to spell them with <dh> and, if desired, devoice them in pronunciation by rule in unstressed syllables.

It is more prudent to spell them with <th>, and, if desired, voice them in pronunciation by some rule, if you want to do that. We do not want to do that, and we do not teach that pronunciation in our books. 

There is no reason -th should differ from -k and -p and -t and -f in unstressed position. Having looked at Lhuyd myself and listened to your arguments, I am still convinced that etymological [ð] is devoiced to [θ] in final position in unstressed syllables. Since orthographically we write -k and -p and -t and -f in this position, it would make no sense to do otherwise for -th.

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