Wiktionary talk:About Primitive Irish

o-stem declension

These are the known declensions of o-stem nouns. Template:pgl-noun-m Compare them to Proto-Celtic:

Masculine o-stem
	singular	dual	plural
nominative	*maglos	*maglou	*magloi
vocative	*magle	*maglou	*maglūs
accusative	*maglom	*maglou	*magloms
genitive	*maglī	*maglous	*maglom
dative	*maglūi	*maglobom	*maglobos
locative	*maglei	*?	*?
instrumental	*maglū	*maglobim	*maglūis

We can see that os > as and i > i. From that, should we reconstruct the accusative as am and the dative as ui? User:ÞunoresWrǣþþe/Template:pgl-noun-m-reconstructed

Accusative was probably -an by Prim.Ir., and dative maybe -u. Are they never attested at all? —Aɴɢʀ (talk) 14:43, 28 December 2016 (UTC)

No, none of them. Every single inscription is in the "X son of Y" style, sometimes with "of the tribe of X", meaning all we get is genitive and nominative (one is vocative, I believe, since it directly addresses Lug, but it appears to identical to the nominative form). ÞunoresWrǣþþe (talk) 17:27, 28 December 2016 (UTC)

Well that's annoying. Continental Celtic inscriptions at least occasionally have datives in "X erected this stone for Y" constructions. —Aɴɢʀ (talk) 19:02, 28 December 2016 (UTC)

There's an oroit do degen, but that has no declension at all. All four archeologists who commented on it said it was "A prayer for Degen}}, which would make Degen dative, but it doesn't seem to be declensed. ÞunoresWrǣþþe (talk) 20:34, 28 December 2016 (UTC)

That looks like a very late form anyway, with final syllable loss. Degen could be an n-stem rather than an o-stem. —Aɴɢʀ (talk) 21:03, 28 December 2016 (UTC)

My first thought was that it might be Old Irish. For now, I'll add *-an and *-ui to the masculine-o-stem template. ÞunoresWrǣþþe (talk) 22:11, 28 December 2016 (UTC)

At the very least, if other words show the final m > n change, we can assume it for the o-stem inflection too. —CodeCat 21:49, 28 December 2016 (UTC)

I won't lie, I don't at all get why Proto-Celtic -os became -as. Is it possible some dialects of Proto-Celtic had "as" like declensions? And those Celts went to Ireland? Or does that development (os to as) make sense to both of you? ÞunoresWrǣþþe (talk) 22:11, 28 December 2016 (UTC)

It's just a sound change. Sound changes don't really happen with a reason, they just happen. —CodeCat 22:30, 28 December 2016 (UTC)

It's no weirder than PIE -os becoming -us in Latin. —Aɴɢʀ (talk) 06:42, 29 December 2016 (UTC)

gn > n

@Angr Why did you remove that part? While your other additions were correct, gn > n is perfectly attested. brocagni (nominative brocagnas) later turned into Broccán. rodagni later turned into Ródán. Gn was never preserved, it always became n or nn. Also, why did you remove the spirantisation? Isn't cattu > cath spirantisation? ÞunoresWrǣþþe (talk) 17:29, 28 December 2016 (UTC)

I removed it because it's too specific. Yes, word-internal gn > n happened (but it stayed gn word-initially, e.g. gníid), but there was a general loss of stops before sonorants, with compensatory lengthening, e.g. *skʷetlom > scél. So it's part of a bigger sound change; it's not just gn > n. As for *kattos > cath, that's Welsh. The Old Irish word for "cat" is catt /kat/. Or are you talking about the "battle" word? In that case it's *katus > cath with ordinary lenition. Goidelic doesn't have spirantization as a separate sound change from lenition. —Aɴɢʀ (talk) 19:00, 28 December 2016 (UTC)

Primitive Irish Cattus is Old Irish Cath. ÞunoresWrǣþþe (talk) 20:29, 28 December 2016 (UTC)

Doubled letters in Ogham do not necessarily stand for phonologically geminate consonants, though. Ogham inscriptions write double consonants seemingly at random, so an inscription that says CATTUS can still be representing *katus with a single /t/. —Aɴɢʀ (talk) 21:00, 28 December 2016 (UTC)

Like Pictish Nehhton or British Cattigern? ÞunoresWrǣþþe (talk) 21:27, 28 December 2016 (UTC)

No one knows anything about Pictish, I wouldn't use it to draw any conclusions. Cattigern is a compound of *katus and *tigernos, i.e. the two t's belong to different roots. —Aɴɢʀ (talk) 22:01, 28 December 2016 (UTC)

You're right, thanks. ÞunoresWrǣþþe (talk) 22:15, 28 December 2016 (UTC)

Ruamanduach

@Angr, CodeCat How could the name Ruamanduach be reconstructed? The root is ruaman, Old Irish for dye (and derived for "ruam" for red), but I can't figure out what duach means. I can't find the original Proto-Celtic of either word. ÞunoresWrǣþþe (talk) 18:14, 1 January 2017 (UTC)

You'll need to clarify what language these words are in. —CodeCat 18:23, 1 January 2017 (UTC)

Old Irish. ÞunoresWrǣþþe (talk) 18:28, 1 January 2017 (UTC)

So then both "ua"s are disyllabic? —CodeCat 18:30, 1 January 2017 (UTC)

Probably, it usually seems to be descended from Primitive Irish ova, ota or oda. ÞunoresWrǣþþe (talk) 18:32, 1 January 2017 (UTC)

If the name isn't attested outside Goidelic, there's no reason to assume it's Proto-Celtic at all. It may have been coined only in Old Irish. DIL lists four words spelled rúam, all of which appear to be monosyllabic; one of them means "fame, glory", which of course is also a common meaning for elements of personal names. It also lists a rúaman (“incursion, raid”), which I suppose is also plausible as an early Celtic given name. This site gives the spelling Rumaind Duach, and DIL says that Duach (disyllabic) is the genitive of the name Daui, so it's probably a patronym. —Aɴɢʀ (talk) 22:59, 1 January 2017 (UTC)