Reconstruction talk:Primitive Irish/ᚋᚑᚏᚔᚅ

-n?
Where does the final -n in the nominative singular come from? —Rua (mew) 15:32, 18 December 2018 (UTC)

I imagine it comes from a reanalysis of neuters in general as ending in /n/ in the nominative and accusative singular? It is evidenced by nasalisation triggered in the corresponding Old Irish forms, as in the phrase

- "muir ndubtemen" [nom.] (Saltair na Rann p. 909), and

- "tar muir ngríste" ["tar" takes the acc.] (Studies 1921, 75 § 16)

~ Rforqs (talk) 20:49, 18 December 2018 (UTC)
 * Ok, but are i-stem or u-stem neuters actually attested in Primitive Irish? —Rua (mew) 23:15, 18 December 2018 (UTC)
 * As far as I am aware, no. At least I have yet to find a compelling example in the database. Maybe it would be appropriate to put the -n in parentheses on the off-chance that the later nasalisation is an Old Irish innovation? ~Rforqs (talk) 08:39, 19 December 2018 (UTC)

RFD discussion: January–March 2020
Reconstructed form with only one (immediate) descendant. It isn't even clear at what point in the history of the word it acquired the final -n typical of neuter nouns (see Reconstruction talk:Primitive Irish/ᚋᚑᚏᚔᚅ). —Mahāgaja · talk 14:31, 6 January 2020 (UTC)
 * The lack of multiple descendants is surely not a problem, as the linked reconstruction also has a Proto-Celtic parent entry at . If Primitive Irish is the ancestor of Old Irish and the daughter language of Proto-Celtic, and the term is attested in the former language and securely reconstructed in the latter, it would seem to me that a Primitive Irish entry can be reconstructed with certainty (as long as the Old Irish term shows all the signs of inheritance and no indication of borrowing or some other provenance, which I am in no position to judge, being unfamiliar with Celtic languages). — Mnemosientje (t · c) 15:21, 6 January 2020 (UTC)
 * But the lack of multiple descendants makes it unnecessary; if we're about to eliminate Proto-Albanian because it has only one descendant, then it makes sense to delete reconstructed Primitive Irish words, since they also have only one descendant. Moreover, the exact form of the word is uncertain. The Old Irish descendant triggers eclipsis because it's a neuter noun, and most neuter nouns ended in -an in Primitive Irish (from -om in Proto-Celtic), but that doesn't prove that the Primitive Irish word was morin and not mori. There's simply no way to know at what point the analogical assimilation of  to other neuter nouns happened. Then there's the vowel of the first syllable: at some point, o was raised to u when the following syllable contained a high vowel, but when? If it had already happened in Primitive Irish, then the word should be reconstructed as *muri (or *murin), not as *mori(n). I just don't see any benefit in reconstructing Primitive Irish words at all, but at the moment I'm only nominating this one for deletion, and not the other three reconstructed forms in CAT:Primitive Irish lemmas, because the other three are less uncertain than this one. —Mahāgaja · talk 16:26, 6 January 2020 (UTC)


 * This reminds me of this debate about the utility of Reconstruction:Old English/nihtmare (also with only one direct descendant, AFAICT). - -sche (discuss) 22:06, 6 January 2020 (UTC)


 * RFD-deleted. —Μετάknowledge discuss/deeds 21:58, 22 March 2020 (UTC)