> What is interesting here is the apparently widespread perception
> that the word is properly "Belltaine". The headword in DIL is
> in fact "Bel(l)taine", and this spelling with double 'll' goes
> back at least as far as Sanas Cormaic, where entry #122 begins
> "Belltaine .i. beil-tine .i. tine beil..." LEIA acknowledges
> that the word is "souvent écrit avec -ll- double", but does
> not address this in the discussion of possible etymologies.
> Hey David! Is there some OI phonological process that would
> naturally account for a meaningless -lt- / -llt- variation?
I think so, yes. It lies in the fact that lenited consonants are delenited next to homorganic consonants after syncope had taken place. That's the case with "l" and "t(h)",
which are both dental sounds. Thus all combinations of /l/ or /L/ + /d dh t th/ would end up as /Ld/ or /Lt/.