[Spellyans] th/dh and Lhuyd
everson at evertype.com
Mon Nov 22 09:43:30 GMT 2010
Owen, the only reason -dh is in the SWF is that George admired Welsh and he put it in KK in the mid-1980s. He did not do an analysis of internal or external sandhi environments, or base his arguments on Lhuyd's transcription of JCH. Instead, he assumed that it would be nice if Cornish were like Welsh, and he innovated -dh. This is the same sort of shenanigans he engaged in when changing bêwnans/bôwnans to *bywnans, when putting *iw into the system, and so on and on.
I suspect Dan's interest in finding reasons for -dh was to give some justification for it. I don't think he's made a strong case. The stronger case is that which is analogous to what we find in other natural languages: where there is a devoicing of final consonants, it happens regularly and consistently.
In such a light, Lhuyd's inconsistency is easily explained: L1 interference. Dan was happy to stipulate that Lhuyd's -ev was L1 interference from Welsh, since he could only have heard "gene" in isolation, he simply mis-parsed "gene vy" as "genev vy" (W genef fi). It's no damnation of Lhuyd to make such observations -- but they make his inconsistencies make sense.
I would like to point out that Nicholas listed yesterday examples of numbers ending in -ek and in -eg in Lhuyd. This is the same inconsistency that Lhuyd shows when he writes nowyth and nowydh.
In for a penny, in for a pound. If Lhuyd is inconsistent about -dh/-th and about -g/-k, then either
1) he was in both cases correctly writing what he heard, in which case we should be writing (like Breton) final voiced consonants in all positions and devoicing them by rule (carreg, caregy; nowydh, nowodhow), or
2) his Welsh was leading him astray, in which case we should settle on the simple and logical stressed monosyllable/voiced v. unstressed final/voiceless model
The problem with 1) is that it is not borne out by the rest of the corpus.
Michael Everson * http://www.evertype.com/
More information about the Spellyans