dive (was Re: Sos-)

> > > > > My proposal is to start from a form with a nasalised vowel,
> > > > > thus
> > > > > *saN- -> *so:, and
> > > > > *saN-i- > *so:m-i
> > > > > so that the seed, so to speak, of the -m- Thomsen notices
> > > > > is absent in suo, is wrapped up in the basal vowel.
> > > >
> > > > If you allow for an -i suffix, you could as well allow for an
> > > > -mi suffix.
> > >
> > > Why would I add an -m- to the suffix, when I can derive it from
> > > the nasality in the root? Occam.
> >
> > Why would you assume nasality in the root rather than in the
> > suffix, if it only occurs in the suffixed form?
>
> But I have nasality in the root already; it's the *saN- of my
> *saN- "wet stuff" ->

Still just hypothesis. The actual fact is that _suo_ has no nasality.

> > > The claim that some language which was substrate to Saami was
> > > spoken all around the Baltic is not so unreasonable.
> >
> > No, it's not so unreasonable. But taken this one instance alone,
> > it is not reasonable either to think that this WAS the case.
>
> Let's take some more then:
> http://tech.groups.yahoo.com/group/cybalist/message/50246
>
> http://tech.groups.yahoo.com/group/cybalist/message/50252
>
> http://tech.groups.yahoo.com/group/cybalist/message/50254
>
> http://tech.groups.yahoo.com/group/cybalist/message/50264
>
> http://tech.groups.yahoo.com/group/cybalist/message/50268
> http://tech.groups.yahoo.com/group/cybalist/message/50271
>
> http://tech.groups.yahoo.com/group/cybalist/message/64056
> http://tech.groups.yahoo.com/group/cybalist/message/64057
>
> http://tech.groups.yahoo.com/group/cybalist/message/57180
>
> http://tech.groups.yahoo.com/group/cybalist/message/56270
>
> http://tech.groups.yahoo.com/group/cybalist/message/50267
>
> http://tech.groups.yahoo.com/group/cybalist/message/62590
>
>
> ...
>
> ...

This probably needs a different topic to digest... Some of these look fairly good, others don't appear to be required to posit for Pre-Samic, yet others seem to require your usual substratal variation hijinx. I did not see anything requiring a fully circum-Baltic distribution; being shared with pre-Scandinavian and pre-Samic, which would have gained its first substratal loans in southern Finland seems to suffice.

> > > > The words should be linked because they're all part of your
> > > > proposal, and because they should be linked, you are proposing
> > > > this specific proposal? Holy circularity, Batman.
> > >
> > > My proposal is that the words are linked, yes. Do you have a
> > > problem with that, Robin?
> >
> > If you try to use your proposal to argue for it, then I do. Which
> > is what you appear to be doing here:
> > "It includes 'sump' "swamp", thus it is not well-limited to
> > putative derivatives of the Uralic 'mouth' word"
> >
> > My argument of semantical well-limitedness is based on the data;
> > your "counterargument" appears to be based on just restating your
> > proposal.
>
> But the semantics of "swamp" also occurs in Finnic *so:- which means that if we posit that the semantic 'suck' group and the semantic 'swamp' group are from the same substrate language, they would have been indistinguishable in that language.

It is again a circular argument that "if we assume (or 'posit') they come from something identical, then they come from something identical". We can only reconstruct "suck" back to *suk/p, and "swamp" back to *swa/ump, which are not identical. (BTW Scand. *swu- would naturally simplify > *su-, so no need to assume alternation on that part, just the usual a/u.)

I do not see how the existence of *soo implies anything for the Germanic items.

> > > but then I got the idea that it could be handled by deriving
> > > the auslaut consonants from the diphthongs which were the
> > > result of the denasalisation of the nasal vowel I posited for
> > > another reason (that of accounting for the a/u alternation).
> >
> > This is where you go off the track of conclusions and into the
> > woods of wild speculation.
>
> Do you have any factual objection here?

Yes: you have provided no factual argument.

> > > > Actually, "sap" looks like it remains quite finely separate as
> > > > well.

> > This is what your runeberg.org link lists under "sump" as well,
> > but I don't see why that would be the same root.
>
> I think the connection is here
> http://en.wikipedia.org/wiki/Soma

Your creativity is mildly amusing. I'll note
1) Soma isn't a mushroom
2) Amanita muscaria isn't associated with swamps
3) A substrate behind Indo-Iranian cannot without further argument be assumed to also lie behind Scandinavian/Finnic

> The problem is, the semantic dividing lines you can set up for my proposed set of reflexes of *saN-, do not match similar semantic dividing lines in reflexes of *daN-.

Semantic dividing lines are not required to conform to any pattern if these are originally unrelated roots. But perhaps you can elaborate on what you mean by "similarities between semantic divisions"?

> > Try not proposing sound changes based on single forms if you
> > strive to be held in higher esteem.
>
> I told you already, I don't.

The *saiwa stuff sure seem'd to be such a case.

> > it seems to be quite possible to just speculate
> > without ever getting to the level of real arguments.
> > At that point the burden of proof (or, more correctly, burden of
> > argumentation) is still on you.
>
> What? When?

Perhaps you do not fully understand what the term means. It's not a legal or moral obligation; it means that no proposal can be _logically_ correct by default, just by force of having been proposed.

> > > How are Lithuanian and Latin relevant to ad hoc assumption in
> > > UEW of a change /a/ > /o:/?
> >
> > We do not need to assume any such change, as Finnic can have
> > gotten the word from a language where it had long *a:, which
> > would then have regularly changed into *o:.
>
> Okay, it's a loan then. I suspected UEW claimed this ad hoc change because it was actually a loan, and now you concur.

I've agreed all along that it's ultimately a loan.

> > > > > http://tech.groups.yahoo.com/group/cybalist/message/65616
> > > > > as. solian etc, cf. Danish søle "mud, slush"
> > > >
> > > > OK. I still do not see how this makes the salt/island semantic
> > > > gap any smaller.
> > >
> > > No, it's a problem, but greater men than myself have ignored it
> > > before. Who knows if the reinterpretation of *saN- as
> > > "immortalizing (?) will bring them any closer. Sacred island?
> >
> > And there's it again - a conviction that there is something to be
> > solved here.
>
> If you are convinced there's nothing to be solved, what are you doing in linguistics?

There's a distinction between "not convinced that X" (aka skepticism) and "convinced that not X" (aka disbelief).

> > I'll continue to simply reject the comparision if you cannot come
> > up with any argument better than "they all have /sal/" for why we
> > should attempt to relate these.
>
> They have to with "soul", "immortality", "truth" and "the otherworld". That's why it's interesting to find out how they are related.

"Salt", "saliva", "island", "slush" have nothing to do with those topics. Do you really not see how illusory all this is is?

> > > The alternation is also manifested as a/o:, BTW.
> > >
> > > Germanic: cake/cookie, hat/hood etc and the whole Class
> > > VI strong verbs

> > /o:/ is not /u/.
>
> That's true, but those alternations seem to appear in the same contexts.

Which are?

> > Whatever happens in the verbs is obviously some kind of inherent
> > development involving ablaut and possibly some generalizations
> > (esp. since not all of these are limited to Germanic).
>
> That's what I'm saying. There must have been a time between PIE and PGmc where the language had ablaut-based verb paradigms from PIE, but nothing else, ie the weak verb paradigm didn't yet exist.

Most probably.

> Verbs loaned at that time would have been too drastically changed to be recognized if patterned on the existing ablauting verbs,

How so? There's no reason the loan-givers would still have to recognize the word, and it would be sufficient if your pre-Germanics would have had some particular form to generalize from.

A nice parallel for this kind of a process is the "triliteralization" of loanwords in Hebrew I've seen mention'd sometimes.

> Thus we only have to explain the a/o: alternation, which think was from a/a: (and then of course there's the problem of why only verbs with root vowel -a- were loaned).
>
> > Or do you have any examples where a paradigmatic alternation was
> > retain'd when loaning?
>
> No, and as you can see, I don't need to assume that either.

That appears to be what you are doing, actually. If a/o: is from pre-Germanic a/a:, the only reason to assume a substrate alternation of similar sort behind these would seem to be if you think that the Germanic alternation was modelled on an alternation extant in the substrate. And that was exactly what I was talking about.

Or if you think it's Germanic-internal after all, obviously then it cannot be the same process as a/u alternation in substrate loans.

> > > > > Give an example of r > n.
> > > >
> > > > Proto-Algonquian > Arapaho, Atsina, Ojibwe.
> > >
> > > I don't find that in Wikipedia. Examples?
> >
> > For Ojibwe, the guy I got these from was referring to this:
> > http://tinyurl.com/yfyznx3
>
> There appear no /r/ for Proto-Algonquian in the table.

See next note.

> > (Note that Proto-Alg. *r was previously reconstructed as *l.)
>
> Maybe they should have left it at that.

Which reminds me of a yet another possibility for a different explanation for your assumed ar/an alternation: *l > n, or *l > r.

> > > > "A split happens for no specific reason" (ie. your 1st-stage
> > > > split into a:/aG/aw/u:/uG/uw) is not an explanation, just a
> > > > more complex re-statement of facts.
> > >
> > > You could say the same of Grimm's law. And?
> >
> > Grimm's law has no splits - and is not based on idle speculation
> > on what pre-Germanic might have look'd like, but on systematic
> > external cognates.
>
> From the point of view of all of IE, Grimm's law has a split, namely between Germanic (and Armenian) and the rest. And it's based on idle speculation on what PIE might have looked like, considering systematic non-Germanic cognates.

You may have problems understanding either the meaning of the word "speculation", or how the comparativ method works, or both.

The fundamental difference between Grimm's law etc. and the substrate case is that we KNOW the different representations in fact come from different languages.

> > > > Or varying reflexes of an *o.

> > > > One word: "Merger"
> > >
> > > /a/ merged with /u/?
> >
> > *o merged with *a or *u.
>
> That would be possible, but it wouldn't explain the forms with prenasalised auslaut.

It doesn't attempt to. Those I could explain from, say, nasalized vowels. Or just prenasal auslauts that were there to begin with, whenever we do not have any alternation.

> > > > Not every word has prenasalized alternants, or labial/velar
> > > > alternants. I conclude that nasalization alternation is
> > > > independant of labial/velar or a/u alternations.
> > >
> > > Your conclusion is unwarranted. There is also the possibility
> > > that the prenasalised variant have been discarded.
> >
> > If in most cases there's no nasal variant, Occam suggests it's
> > because there never was one.
>
> Occam suggests either a solution which loses nasalisation or one with one group having nasalisation and one without.

> Torsten

No, for words that lack variants with nasals, Occam very much does not suggest loss of nasality; Occam suggests no nasality.

Does your impression of Occam also suggest, say, loss of anlaut whenever there is none? Loss of rhotic whenever there is none? Loss of vowel length whenever there is none?

John Vertical