Reconstruction talk:Proto-Celtic/miskati

A-coloring

@CodeCat Is this correct? Nothing about the PIE derivation implies an *a-coloring? —JohnC5 14:28, 19 October 2016 (UTC)

But the Old Irish form does. —CodeCat 14:29, 19 October 2016 (UTC)

Is there evidence that his extends to PC besides Goidelic? —JohnC5 14:39, 19 October 2016 (UTC)

@Anglom —CodeCat 15:24, 19 October 2016 (UTC)

If we determine this a-coloring to be the case, then I could reconstruct this instead as a denominative *miskos + *-ati and not as a direct PIE reflex. —JohnC5 15:40, 19 October 2016 (UTC)

Right, that's what Matasovic was suggesting, though his entry was very incomplete, lacking most of the Brythonic forms I found. --Victar (talk) 15:47, 19 October 2016 (UTC)

A denominative would have a long vowel, -āti. —CodeCat 16:00, 19 October 2016 (UTC)

Exactly, but that would have changed the Brythonic forms, which is why not really seeing eye to eye with that theory. --Victar (talk) 16:07, 19 October 2016 (UTC)

But Old Irish has a class A I verb, which is the descendant of the -āti class. —CodeCat 16:14, 19 October 2016 (UTC)

I guess I'm confused. Can you only get a-coloring through a denominative suffix? Because if so, than the Brythonic form must be from a different form, no? --Victar (talk) 16:36, 19 October 2016 (UTC)

CodeCat, indeed I meant *-āti. Victar, there are several ways in which you can get a-coloring, but all of them involve the *h₂. PIE had an innovative denominative verbal suffix *-eh₂yéti which initially came from nouns in *-eh₂. This suffix accounts for the majority of conjugation I verbs in Latin and was highly productive in Proto-Italic (*-āō) and Latin. It operated in PC as well, and while *miskāti may go back to *miḱ-sḱ-eh₂yéti (a denominative verb from a deverbative noun), there aren't any other cognates as far as I can tell. PIE did have a deverbative noun *miḱ-sḱ-ó/éh₂-, so it's only one step in PC to use the productive suffix *-āti. —JohnC5 17:22, 19 October 2016 (UTC)

Thanks. My concern still is that Proto-Celtic *miskāti would have yielded Proto-Brythonic *mɨskọd, no? If so, than that doesn't work for its known descendants. --Victar (talk) 17:35, 19 October 2016 (UTC)

The only solution I can see is to move this entry to *misketi and derive Old Irish mescaid from PCelt *miskos +‎ *-āti. --Victar (talk) 20:00, 19 October 2016 (UTC)

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘ This also makes me notice that the following verbs should have *-ā-:

• *ɸirnati → *ɸirnāti < PIE *pr̥-né-h₃-ti ~ *pr̥-n-h₃-énti < *perh₃-
• *kʷrinati → *kʷrināti < PIE *kʷri-né-h₂-ti ~ *kʷri-n-h₂-énti < *kʷreyh₂-
• *linati → *lināti < PIE *h₂li-né-H-ti ~ *h₂li-n-H-énti. < *h₂leyH-
• *brusnati → *brusnāti < PIE *bʰrus-néh₂-ti ~ *bʰrus-nh₂-énti < *bʰrews-
• *barnati → *barnāti < PIE *gʷr̥-né-H-ti ~ *gʷr̥-n-H-énti < *gʷerH-
• *binati → *bināti < PIE *bʰi-né-H-ti ~ *bʰi-n-H-énti < *bʰeyH-

In all the cases that there are Brythonic descendants, they seem to have a short *a. This means that this syllable was regularly shortened in either PC or PB. This could perhaps be the generalization of the 3pl. *kʷrinanti (which must be short either because of the derivation from PIE oblique form *kʷri-n-h₂-énti or because a generalize *ā in *kʷrinānti would be shortened due to w:Osthoff's law). —JohnC5 00:07, 20 October 2016 (UTC)

The vowels would shorten in Old Irish anyway. However, there's one important point of distinction: the 3rd singular conjunct form. For a short -na- there'd be no final vowel there, while for long -nā- there'd be an -a still. That's exactly the difference between Old Irish classes A I and B IV. They shouldn't be confused. —CodeCat 01:35, 20 October 2016 (UTC)

Very interesting. So these were shortened in PC. But, you will admit that these short vowels is not the expected outcome. Should we say all *-néH- formations resulted in PC *-na-? —JohnC5 04:09, 20 October 2016 (UTC)

@JohnC5: The Celtic suffix probably wasn't shortened by phonological rule; rather it probably comes from a generalized zero grade -nh₂- and/or was re-formed on the basis of the plural (i.e. **kʷrināti become *kʷrinati under the influence of *kʷrinanti). —Aɴɢʀ (talk) 21:43, 7 December 2016 (UTC)

@Angr: So, *-āti can only apparently come from *-eh₂yéti? What's the situation with *marnāti (< *mr̥-né-h₂-ti from *merh₂- (“crumble, destroy”))? Is it incorrect? Matasovic has *mar-na-, which I believe means the vowel is short. —JohnC5 22:03, 7 December 2016 (UTC)