Talk:eilfed

Etymology
This word doesn't seem to have been formed in Welsh; it is attested in Middle Welsh. Therefore the essentially true statement as to its formation has to be demoted to a synchronous derivation, and supplemented by its derivation from Middle Welsh (or earlier) elements. Grr! We currently have no Middle Welsh lemmas. I'd rather have expressed etymology as you have.--RichardW57 (talk) 18:33, 21 September 2022 (UTC)
 * , the GPC entry for eilfed doesn't list anything earlier than the 16th century, which is Early Modern Welsh, not Middle Welsh. And we do have some Middle Welsh lemmas, but not this one. —Mahāgaja · talk 18:35, 21 September 2022 (UTC)
 * , on the other hand, is attested in Middle Welsh as . —Mahāgaja · talk 18:37, 21 September 2022 (UTC)
 * I wonder how I got them mixed up. Now the other is a confusing word.  I'm not sure that 'deufed' is actually attested outside grammars.  Most of what I've found for it is 'ddeufed', with no obvious reason for a soft mutation, as in 1 Chronicles 24.
 * As to Middle Welsh lemmas, I must have been confused by 'Middle Welsh lemmas' coming before 'A' in Category:Middle Welsh language RichardW57 (talk) 19:12, 21 September 2022 (UTC)
 * I've reconstructed my reasoning. It's an issue with the morphology.  MiW to ModW has the sound change ei>ai in final syllables.  Using modern consonant symbols, ModW ail+fed leaves *ailfed.  MiW eil+fed yields *eilfed, > ModW eilfed.  So, do we have an early ModW morphological rule to give us ail+fed > eilfed?  We may have a morphological rule that says that ///eil/// is realised as /ail/ in final syllables.  The allomorph /eil/ generally occurs in the prefix eil-, so the first syllable has more in common with the prefix eil- than the word ail.  However, we also get cases where the morpheme now surfaces as ail-.  The tale is neater if words in eil- hark back to Middle Welsh.
 * My memory on attestations then crossed with deufed. --RichardW57 (talk) 21:14, 21 September 2022 (UTC)
 * I would say that the ai ~ ei alternation remains productive in Modern Welsh. There are so many examples along the lines of ~  etc. that speakers must be (subconsciously) aware of it, so even from a purely synchronic point of view it's unproblematic to say that  +  yields . That would be true even if the word were attested in Middle Welsh. It's the same with the alternation between y  in nonfinal syllables and y  in final syllables – diachronically that was caused by a sound change, but the alternation remains synchronically productive. —Mahāgaja · talk 05:29, 22 September 2022 (UTC)