Re: *ka/unt- etc, new conquests

> > > "This is above all its distribution area. It is bigger than
> > > that of Krahe's name groups and seems by far to go beyond the
> > > borders of Europe, which I included in my works.
> >
> > I notice he doesn't actually demonstrate this claim here.
>
> True. But I'll trust him on this one. He's usually reliable.

I don't tho, and I don't plan on starting to back on trust in matters of science. Or hey, I could just tell that I trust Pokorny or Redei or Sammallahti or whomever on the matter and we'd have nothing to discuss. (Which might actually be the smart choice if the experts were still cracking at the topic too, but Krahe won't be around to defend anything anymore.) I get the impression your view here goes beyond Krahe's, too.

> > > > But it occurs to me that Proto-Samic has the sound change *u
> > > > > *o (while Proto-Finnic doesn't), and you kno what Germanic
> > > > speakers would make of that.
> > >
> > > Yes.
> > >
> > > > Not extensible to central Europe either, however.
> > >
> > > Now there's your problem!
> >
> > Not really. It just means there is some language or family that
> > also includes an a/u alternation.
>
> No, more than that. It would mean that some substrate in Europe had a root *ka/unt- "hunt etc" which was unrelated to Uralic *kunta "group, to hunt", *kan-ta "to carry" and *kënta "stump, base".

This was Germanic, no? Perhaps I should look deeper into past discussions.

---

Okay, so there's *kan-tho- "edge" which looks like it could be related to Uralic *känta (up to minimal quadruplet now). The distribution and semantics here are sufficiently bad that I can believe it might be two separate words, with Samic "shore" related to the IE complex.

I continue not to see the semantic link to "hunting group" or the other Uralic words. The idea of a link in the form of "edge" ~ "cavalry wing of 100 horsemen" ~ "group" is unconvincing even for Germanic alone. That makes about as much sense as "edge" ~ "a sharp-toothed animal" ~ "hound", or "edge" ~ "knife" ~ "handheld tool" ~ "hand", and those kind of links can be created between any two sufficiently basic words (OTOH hound ~ hunt might have something to it, but it's not directly relevant now).

Finnic *kansa is a kno'n loan, so not relevant.

"Hat" ~ "hose", "hidden", "hazel", "cunt" ~ "kusi" etc. only put the semantical bredth further out of hand.

If you are saying what I think you are saying (ie. that these are all "related somehow") the criteria for related-somehow-ness seem to come down to:
1) Forget all about MOA
2) Forget all about semantics
3) Look for vowels that adhere to a pre-decided set
4) List any words that have consonants of the same POA and vowels

Heck, why not change #3 to
3b) Forget all about vowels
and add in "hit", "hint", "hind", "hiss", "kit", "kid", "Kind", "kind", "kiss" etc. as well?

> > > > Also the distinction between Uralic *kunta "group, to hunt",
> > > > "to hunt", *kan-ta "to carry" and *kënta "stump, base" is by
> > > > all evidence one inherited from PU; I see no links between
> > > > the three, other than that they have the same consonants.
> > >
> > > If you give up your assumption that they are native Uralic
> > > words, you will.
> >
> > It is not an assumption, it is a conclusion.
>
> No, a conclusion needs premises. It's an assumption.

The premises are that the words have a good distribution and adhere to regular Uralic sound laws. If you use a method other than the presense of phonetical irregularities, or spotty distribution combined with parallels elsewhere (or the lack of a deeper etymology - but that's not applicable here) for identifying latter loanwords in a protolang's reconstructed lexicon, I'd like to hear about it.

> > It is a disservice to etymology to turn perfectly well-behaving
> > words into "substrate lons" just because.
>
> The English words 'democrat' and 'democracy' are obviously not related since the English languages does not have the extensions -t and -cy, and since they are perfectly well-behaving there is no reason to assume they are loans either.

As long as we're doing satire, did you hear about this new substrate, the i-language (iish for short)? It has the curious feature of containing the vowel /i/. See, there are several words in the world's languages that both lack an etymology and contain the vowel /i/. This is an obvious pattern that needs an explanation. ;)

More seriously tho, there is an obvious semantic connection between "democrat" and "democracy" that does not exist between "stump" and "to hunt". Likewise, this alternation of non-suffixes **-t and **-cy is exactly what makes a word not "well-behaving".

> > Down that road, we could as well decide that all words are
> > substrate loans and call it a day.
>
> And then you'd have to read up on new foreign languages. We wouldn't want that.

The correct objection to that is that taken to its logical conclusion, it would abolish the concept of regular descent of lexicon from a reconstructable proto-language. An English word resembling a German word? Must be because they're cross-loans, or both loaned from the same substrate... hm, looks like this substrate contains alternations such as -k ~ -x... and -t- ~ -ts-...

> Of course we wouldn't to call a word a loan which had no cognates outside and didn't otherwise stick out.

And I repeat that these words in fact do not "stick out". Also external comparisions can just as well point to common inheritance (if not straight out coincidence). All I'm saying is that these go back to proto-Uralic.

> > > > And I have no idea what you are getting at with the other
> > > > roots with *ka- you list in msg #62525.
> >
> > It does not keep getting any clearer what are we to make of it.
>
> As I said, you will, if you're willing to give up your assumption that they're Uralic. But you're not.
>
> Torsten

I'd appreciate if you for once just told us what your thesis is on them, insted of expecting others to read your thoughts on the matter. We won't arrive to the same conclusions just because we have the same group of data available, if only because we come from different backgrounds (our "full sets of data" are different).

For the record, I'm quite open to the idea that the UEW contains fair amounts of substrate loans (take for instance the prevalence of Finnic-Samic-Mordvinic isoglosses, and their different phonotactical properties compared to common Uralic).

Or are you trying to say that *ka is an un-Uralic combination and therefore sufficient grounds for a word being a loan? ;) Also remember, the paucity of *k and *a in PIE is an exception, not a rule...

John Vertical