[Spellyans] RLC <h> for <gh>

[Owen Cook]

2008/6/26 Jon Mills <j.mills at email.com> rug scrifa:
> You are correct in saying that Lhuyd's transcriptions are phonetic. However not every Cornish
> word written in Archaeologia Britannica is phonetically transcribed in Lhuyd's General Alphabet.
> Some items are quoted as they are spelled in the Middle Cornish texts. So care needs to be
> taken when citing Lhuyd as an authority for any proposed pronunciation.

True enough.

> Using minimal contrast, it would be possible, as you suggest Owen, to extrapolate a phonemic
> inventory from Lhuyd's Archaeologia; though this would not be entirely free of problems.

Indeed, I would expect it to have problems!

> The results might come as a surprise to many. For example, I have so far been unable to find a
> minimal contrast pair for Lhuyd's <dh> and <th>. This suggests that in 17th century Cornish, [D]
> and [T] are allophones, not separate phonemes.

That is an interesting circumstance. How would we account for [T] if
it is an allophone of /D/? If it is an allophone, how is it
conditioned? If it is not conditioned, we must imagine that there is
free variation. But if we find relative stability of [T] in certain
words, rather than free variation, then we have a puzzle on our hands.

~~Owen