>I read up on Thomsen's (above) text on the occurrence of pairs of nouns with and without suffixed -s (only in nom. and part., otherwise identical) in Livonian; seems to have been widespread, and not limited to loans from Baltic and Germanic. It occurred to me that this corresponds to the IE situation for m vs. n. paradigms, with certain modifications.
> --- In firstname.lastname@example.org, johnvertical@ wrote:
> > > > > I proposed that the *IE* s-stems were based on a
> > > > > reinterpretation of the *IE* genitive -Vs as nominative
> > > > > (which BTW I think is the origin of the IE nom.sg. -(V)-s.
> > > >
> > > > Okay. Then what does Finnic have to do with it?
> > >
> > > It borrowed both from PIE thematic stems (Germanic a-stems)
> > > (like rengas) and PIE s-stems, and PIE (and Germanic) s-stems
> > > (like lammas), with the same result (the vieras declension).
> > > One explanation for that is that they borrowed the stem of the
> > > nom.sg., which would have the same *-os ending (<- *-as in PPIE
> > > and -> *-as in some dialects, including Germanic), but *not* in
> > > Germanic, which had lost the s-stem nom.sg *-s. The cognates of
> > > the borrowed items may be found in Germanic, but the donor
> > > language can't be PGmc, as you also remarked (below).
> > Aaah. So the lack or presence of -s in Finnic would be not so
> > erratic after all? It IS commonly left away entirely or
> > substituted (hanhi "goose" < B. *Zansis, karja "cattle" < Gmc
> > *xarjaz, mair-ea "smeering" < Gmc *smairjaz), but you're saying
> > that where it does appear, all have it if not from the nom.sg,
> > then from an old -s- stem. Right?
> No, that's what Vilhelm Thomsen said might be the case in 'Über den
> Einfluss der germanischen Sprachen auf die finnisch-lappischen.
> Eine sprachgeschichtliche Untersuchung (1870)'
> and what de Vries tacitly implied
> > The words of wider Finnic distribution where this -s appears
> > appended to an Uralic root seem, then, like an issue here, if we
> > are to make this suffix unproductiv.
> There's no way of avoiding it being an issue if we insist that its
> origin is in loans from Germanic.
> > Some of them could be
> > explained by being original -ks stems that were changed to -s due
> > to the example of IE loans (old Finnic words in -aks are rare),
> > as _oras_ "thorn" vs. Livonian _voraaks_ explicitly suggests. The
> > Uralic status of _uros_ "male" and _nauris_ "turnip" is uncertain -
> > Hungarian _úr_ "lord" "has been considered a separate Iranian
> > loan" according to Häkkinen, and Ob-Ugric _*nëëG@... means "cedar
> > nut" (this would also be the only Uralic root featuring *-kr-),
> > with no other internal cognates.
> > Also have you spotted any examples of -s- stem correlation other
> > than *lambaz?
> As you can see in the Thomsen locus
> there are
> lammas "sheep"
> mallas "malt"
> labbes "lamb"
> males "meal"
> more dubious are
> lannas "ground, beach"
> lunnas "ransom" (cf. German Lohn), with lunnastaa "to ransom"
> porras "stair, path"
> teuras "animal for slaughter"
> > A word lacking an IE etymology is not really a prime
> > candidate for loaning from a para-Germanic language, or
> > pre-Germanic.
> That's what I thought too.
> Food for thought:
> or not, as the case may be.