What evidence is there for any such variation anywhere in Cornish? On 3 Feb 2009, at 15:32, Michael Everson wrote: > context conditioned variation of [ð] and [θ]