Summer’s wrapping up, a new academic year’s about to roll in, and if all goes well, I might be returning to more active blogging around here.
I have also returned, about a week ago, from the 12th International Congress for Finno-Ugric Studies. You can check out my presentation online, too: Semivowel losses and assimilations, in Finnic and beyond. Longtime readers may recall me having first explored the ideas within many blog posts (and one blog platform) ago. Evidence has continued to turn up, and I’m by now quite convinced that my newfound soundlaw *je- > Finnic *i- indeed exists.
The title is admittedly a bit more general than what might be warranted per the presentation’s contents. For space concerns, I was not able to treat the topic of initial semivowels in Uralic languages in more general.
I could mention one fairly simple addendum here, though — while *wo- has been already traditionally well-established, and I attempt to show that abundant evidence of *je- can be found as well, by contrast it seems to me that **wu- and **ji- were not possible sequences in Proto-Uralic (and they remain impossible in most descendant languages as well).
Only one widely accepted instance *wu- has been proposed: the word for ‘new’ (> Fi. uusi, Hu. új etc.), traditionally reconstructed (modulo notation) as *wud₂ə. I however suspect that this should be instead reconstructed with *o, and that the evidence suggesting *u is due to the shift *o-ə > *u-ə in open syllables, as first proposed by Janhunen (1981) for the Finno-Permic end of the family. (Seen also in e.g. *lomə > *lumə > Fi. lumi ‘snow’. Sammallahti has later suggested that the change also affected Ugric; I am skeptical, however. More on this at some point in the future.) — For *ji- there are a few more potential examples, but the best-looking cases (? *jikä ‘age’, ? *jitɜ ‘night’) fall among those where I believe *je- should be rather reconstructed.
If so, then it seems to me that we can likely apply for Proto-Uralic a phonological analysis also known from various other languages of the world: to unite *j *w on one hand and *i *u on the other as allophones of each other.
Nice! I quite like the idea that PIE *ǵ was loaned as PU *j.
On the slide about ien, you call *kn > *ŋ “an extraordinary sound shift”. Is it extraordinary at the ends of words? In southeastern German dialects like mine, -gen /gn̩/ regularly becomes [ŋ̩] (legen [ˈlẽŋ̩]); the resulting sequence of a vowel and a syllabic consonant loses the latter’s syllabicity in more common words, like sehen > Verner variant *segen > [sɛ̃ŋ]. Importantly, /g/ is voiceless, so it’s the same sound as generic Uralic /k/ (or Spanish /k/ for that matter).
Likewise -ben, -den > [m̩], [n̩].
It would be extraordinary in the overall picture of the consonant history of Permic; normally all medial single stops are simply lost, plus there does not seem to have been any syncope that would have produced that type of awkward consonant clusters in the first place.
There’s also a longer story that I did not have space to go into in there, in that the Permic comparison appears to have been motivated by Northern Mansi & Khanty, where there has indeed been a shift *-ɣ(ə)n > ŋ(ə)n, and which had misled some people to reconstruct something like original *äŋV…
(Also, for the record, *ǵ → *j has been known for long — the classic example is PIE *h₂aǵ- → PU *aja- ‘to drive’.)
I see!
Are there any examples of PIE *g being borrowed into PU?
The only comparison involving PIE *g (which was, after all, kind of on the rare side) that I can find offhand is PIE *mozg- ~ PU *mośkə- ‘to wash’, but the palatalized *ś makes this difficult to analyze as a loan.
I see. Are there any examples of *k or *gʰ (both also on the rare side)?
I don’t know many loan etymologies with those, either; but on further digging, one with *g or *gʰ is PIE *ǵʰalg⁽ʰ⁾- → PU *śëlka ‘pole’. It seems likely me that PIE *(s)kerH- ‘bark, skin’ has also been the origin of at least one of PU *kerə and *karə ‘id.’
A couple other comparisons that may or may not be onto anything:
Very convincing presentation! Great job. :)