Zompist Bboard Again

By accident I found out that Polish noun łosoś (salmon) retains PIE -os ending (from Proto Slavic *lososь from PIE *loḱsos)

Any other nouns like this one?

What makes you think the reflex in PIE is *loḱsos and not *loḱsos-? The final jers would seem to point to the latter.

Interesting. I would not reconstruct an affix -os- just for this one word, so unless it's popped up elsewhere in proto-Slavic, I'd lean towards saying that it is in fact the original PIE -os. As to why it was retained, though, I wonder if it might be a loanword, supposing the early Slavs migrated away from their ancestral habitat and, at least briefly, lost their knowledge of what salmon were. Southeastern Europe is not a particularly good fishing area, so this makes perfect sense to me, although it's just a guess.

Poland and Belarus and Ukraine are all excellent fishing areas with lots of rivers and lakes. Wiktionary at least does not say it's a borrowing

and if what would it be borrowed from? The only other other reflexes are Germanic, Ossetian and Tocharian. Germanic is *lahs so Slavic reflex would be something like *losь or *lohsь

Otto Kretschmer wrote: ↑Mon Apr 12, 2021 10:06 am Poland and Belarus and Ukraine are all excellent fishing areas with lots of rivers and lakes. Wiktionary at least does not say it's a borrowing

and if what would it be borrowed from?

from Baltic, even if that root was later lost in Baltic. That said, it seems proto-Slavic was spoken much further north than I'd pictured it ... I'd always thought it was from the area of present-day Bulgaria and that all future movement was towards colder climates ... so I dont know if my idea makes sense anymore. Even so, if there's no known PIE affix -os- that makes sense there, I'm sticking with the idea that it's most likely a loan, even if I can;t place it on a map.

am I right that Slavic did a vowel shift of o > a > o, at least conditionally?

Pabappa wrote: ↑Mon Apr 12, 2021 10:45 amam I right that Slavic did a vowel shift of o > a > o, at least conditionally?

Yes.

Note the palatalisation of the final sibilant. Palatalisation in Proto-Slavic is only ever caused by a following *i or *j, and that has to come from somewhere. PSl *lososь likely derives from PIE **loḱs-os-is, with -os- being some kind of derivational suffix, of which there are a myriad in PIE.

Much less likely, it was borrowed from a more conservative dialect to a more innovative one just as the open syllable isogloss was spreading out. But I doubt it.

It has to be noted, though, that there are at least two cases of an epenthetic final vowel that blocks the wouldbe falloff a final consonant - PSl *otъ 'from' < PBSl *at, and one other preposition that I can't recall on top of my head.

Kroonen has also suggested that it may represent *loḱ-os-, i.e. an s-stem.

Zju wrote: ↑Mon Apr 12, 2021 12:30 pmPalatalisation in Proto-Slavic is only ever caused by a following *i or *j,

And also *e, although that's not relevant to this particular case.

It should be noted that animal vocabulary in particular is one of the parts of the PIE lexicon which is most prone to somewhat idiosyncratic extensions and reformations. For example there's a relatively frequent -k- suffix which turns up on some words in a number of branches, seen in for exampe *neh1tr- "snake" in Italo-Celtic (Latin natrix, Gaelic NOM~GEN nathair~nathrach, and this same extension also crops up in Celtic on *kapros "goat, sheep" > Gaelic NOM~GEN caora~caorach (Scottish Gaelic would add a bunch more -ag suffixes to a number of animal names later down the line as well e.g. feòrag "squirrel", ultimately from a PIE *wi(r)wer-, but that's not as immediately relevent here). The reformation of the athematic root noun *loḱs as an s-stem *loḱos- as opposed to a thematic *lok-os is a bit harder to find parallels for, but I'm sure those with a much more comprehensive working knowledge of the IE lexicon can come up with an example or two.

Zju wrote: ↑Mon Apr 12, 2021 12:30 pm Note the palatalisation of the final sibilant. Palatalisation in Proto-Slavic is only ever caused by a following *i or *j, and that has to come from somewhere. PSl *lososь likely derives from PIE **loḱs-os-is, with -os- being some kind of derivational suffix, of which there are a myriad in PIE.

KathTheDragon wrote: ↑Mon Apr 12, 2021 1:53 pm Kroonen has also suggested that it may represent *loḱ-os-, i.e. an s-stem.

Becoming i-stems would be the normal development for consonant stems in Slavic. The accusative *-ׅṃ of the consonant stems had become *-im in Balto-Slavic, and Slavic created new i-stem paradigms based on that. Sometimes the old consonant-stem forms are still attested, e.g. OCS svekry "mother-in-law" (< PIE swekruH2), gen. sg.svekrъve acc. sg. svekrъvь, vs. Russian i-stem nom/acc svekrovь, gen.sg. svekrovi, but often we can only assume that development based on comparison with IE and based on knowledge which suffixes were associated with consonatnat stems and which with i-stems in PIE.
In this case, though, we have -yo- and -o- stems in Slavic (i.e., projected back into PIE, *lok's-os-(y)o-; Baltic has: Lith. lašišà f.(-ya:- stem) 2; Latv. lasis m.(-yo- stem), OPr. lalasso [lasasso] m. - so this looks like an s-stem was re-interpreted / further derived with different results in both Baltic and Slavic, long before Slavic lost final /s/.