dive (was Re: Sos-)

> > > > > > > *sá:-/*sák-/*sáp-/*sank´-/*samp´-
> > > > > > > *sú:-/*súk-/*súp-/*sunk´-/*sump´-
> > > > > > >
> > > > > > > which would solve the 'Suomi' mystery

> 'Several etymologies exist, we don't need another one'. What were you thinking?

I'm not saying it's pointless bringing new etymologies on the table at all, but that this word had no real NEED for a (new) etymology.

Mostly this comes from interpreting your "solving the mystery" to mean "I think have found the solution", and not "I think I have found yet another possibility" (after all, the latter is not quite "solving".)

This then would imply that you had misunderstood the situation to be that there is NO good etymology for _Suomi_ at all, rather than that there are several even without your proposal.

> > What are you exactly proposing as the answer, anyway? The long
> > vowel items here can only explain the _suo_ part.
>
> *saN-i- -> *so:mi

*aN > *o:m is not in your table of variants your sound changes can derive, so far.

Since you're basing this cluster on other "swamp" words, just deriving *so: "swamp" would actually look a bit better. (This word by contrast does not have even a proposed etymology.)

> > > > If we have X number of fairly plausible etymologies, adding
> > > > one that seems less plausible is not progress.
> > >
> > > That's not what you just said, you added a new and subjective
> > > premise. Why are you cheating on the scale?
> >
> > I would've thought that me considering anything based on your
> > substrate alternations not a plausible explanation was implicitly
> > quite well estabilish'd by now.
>
> Yes, you don't accept my proposals. That is an unreasoned opinion,

I've tried to explain my reasons several times.

> > > > > > (at least one involving all sorts of hypothetical forms).

> > > What exactly about it is it you object to?
> >
> > My prime issue is that the involved soundlaws do not appear to be
> > based on regular correspondences, but on a very limited set of
> > attested words and thus they're fairly ad hoc.
>
> The sound laws apply to a limited set of words, yes, as in
> Schrijver's article. For the whole set, you'd probably need to
> consult
> Boutkan, D. 1998.
> On the form of North European substratum words in Germanic,
> which Schrijver refers to.

Note taken.

> > > > > look at 'sump' here:
> > > > > http://runeberg.org/svetym/0995.html
> > > >
> > > > I don't see a need to assume any velar variants if that's all
> > > > we are explaining.
> > >
> > > There are all the "suck" words here:
> > > http://tech.groups.yahoo.com/group/cybalist/message/62677
> >
> > Semantically quite well-limited. I wouldn't consider this to be
> > from the same root. I see the resemblance to Uralic "mouth" tho.
>
> It includes 'sump' "swamp",

What "it"? The -k set is limited to "suck", and the -mp set is limited to "swamp". There's no overlap between these and I see no grounds to connect them.

"Swamp" > "mouth of river" > "mouth" would be another of the semantic chains you pull out of your sleeve. Or is this "mouth of river" attested somewhere?

> > > There is the "sink" stuff:
> > > http://tech.groups.yahoo.com/group/cybalist/message/43771
> > > http://tech.groups.yahoo.com/group/cybalist/message/43779
> >
> > *saiwa does not appear to be linkable - referring to clear, not
> > muddy, water and containing the difthong -ai-.
>
> *-aN- -> *-aNw- > *-aiw-, cf Portuguese.

You just keep making these up on the spot, do you?

Any phonetic form is "linkable" in that sense anyway. If you have more than just that one example of *-Vjw- corresponding to *-VNk- (or whichever you think of as the closest related form), shoot.

...Hey, actually it's also not *-jw-, it's *-wj- (Inari Sami _savja_, Skolt Sami _saujj-)! That suggests Finno-Samic had this word first, and it was loan'd by Germanic after the metathesis -wj- > -jw- (regular in Finnic, and in most Samic varieties).

> > I can propose _säng_ to come from Uralic *s´äNki- "to cut".
>
> Why would you cut a bed?

Hold your horses, this is just the summary.

> > The Baltic Finnic direct descendant *säNki primarily means
> > "stubble".
> > A later development of sense is "a patch of field or garden",
>
> From "stubble"? Why would you want to have your garden there, of
> all places?

"stubble in field" (attested, primary)
"a section of field which has stubble (has been harvested)" (attested as the compound _sänkipelto_)
"a section of field or garden" (attested)

Also, straws used to be used as a mattress filling. That possibly cuts some corners: in a bed (as opposed to just sleeping on a bench, or on the floor) one would in fact be sleeping on something stubbly.

> > > > > The alternative is that both loaned from an unrelated
> > > > > substrate, ie the ar-/ur- etc language.
> > > >
> > > > False dichotomy. I'm going with the default alternativ of
> > > > "unrelated".
> > >
> > > Which entails that PIE had
> > > 1. bagn- "swamp"
> > > 2. pan- "swamp"
> > > Are you sure that holds up?
> >
> > Why not? You've seen it's quite possible for a language to
> > simultaneously have words such as _deep_, _dive_, _dip_ and so on.
>
> Exactly. And they are possibly related, so bad example.

I gather'd you were objecting to the possibility that PIE had two words of similar shape. Being related somehow IS one possibility for such words, but such a relationship may well be older than PIE (and have no relation to the substrates involved behind Northern IE).

> > Oh and you're now going into a false trichotomy.
>
> I am?
>
> > If unrelated, these do not have to come from PIE;
> > one or both can be loans from unrelated substrates,
>
> These two substrate languages have *pan- "swamp" and *bagn- "swamp", respectively, and those two words are not a common loan in both?

It's a possibility.

I keep seeing this apparent principle "if they have some resemblance, it cannot be a coincidence" behind your (and some others') reasoning, but this is a false conviction.
http://en.wikipedia.org/wiki/False_cognate

> > or even from a common substrate but unrelated within that too.
>
> That just repeats the conundrum in another language.

So it does. There would still be 40,000 years to go backwards until the invention of language, so one step of regress is not a problem.

You're proposing that the words are related within a *specific* substrate (or family thereof). I don't think you can be very sure of that.

> > > > > > > 1 pin´: (Sal. pinli), pl. pi`n´n´&^D (neu: sùomli, pl.
> > > > > > > -st) finne (finnländer);
> > > > > > > s. pin´-mo:, pi`n´n´&^mìez.
> > > > > >
> > > > > > Transparently a loan from _Finn_. This provides no new
> > > > > > insight.
> >
> > > > Another thing is that this word only occurs in Livonian.
> > >
> > > What would that prove?
> >
> > Livonian was only incompletely documented, while Finnish and
> > Estonian dialects are documented in exceeding detail. If the word
> > still only occurs in Livonian (out of all Uralic languages), we
> > can be rather sure it's not inherited from Proto-Finnic, ie. it's
> > of later loan origin.
>
> Most likely, but not 100%.

In etymology, nothing is.

> > > > > Okay, so there is a/u alternation in Komi and Udmurt, under
> > > > > some conditions.
> > > >
> > > > Um no, I said there's a development u > ï in Udmurt in some
> > > > words in some dialects (no alternation, no /a/, and not in
> > > > Komi).
> > >
> > > Sorry, I was being imprecise: there is an alternation a/u within
> > > Uralic
> > > http://tech.groups.yahoo.com/group/cybalist/message/62618
> >
> > I see reference to no such thing in this link.
>
> But there is one of a necessary ad hoc assumption of a development /a/ > /o:/

No, a: > o: is a regular soundlaw of Finnic. Compare eg. "vein": F _suoni_, Mordvinic /san/ (from PU *sëxni > *së:ni > *sa:ni).

> and in Komi (Zyrian) further to /u/.

Komi is /sol/, and /o/ is the regular reflex of *a.
http://tinyurl.com/yzkaemo

Loaning appears to have taken place independantly in Permic and Finno-Volgaic, at a date such that Permic had no *a: and substituted *a.

> This corresponds to the ad hoc assumption necessary for Germanic of an alternation a/u for the "salt" word.

What Germanic language calls it "sult" again?

> > It's all explainable from a root of a shape such as *sa:la.
>
> UEW obviously disagrees with that.

I've told you, UEW is badly outdated in what comes to reconstructions.

> > > > The semantic gap between "island" and "salt" does not seem
> > > > any smaller in other languages.

> As for the semantics gap, "slush" is a stepping-stone.

I don't really see that. They didn't use road salt in those days. ;) Nor are islands made of slush. Nor am I aware what actual form you're alluding to?

> > > The alternation is also manifested as a/o:, BTW.
> >
> > Which alternation, and where?
>
> Germanic: cake/cookie, hat/hood etc and the whole Class VI strong
> verbs
> http://en.wikipedia.org/wiki/Germanic_strong_verb#Class_6

What makes you think this is the same alternation and not a different one? That looks like it would be easiest to explain from a substratal *o (Gmc, lacking that, would substitute either *a to retain the quantity, or *o: to retain the quality).

And isn't "cook" from Latin anyway?

> > > > So you've now turn'd an apparent ar/ur alternation into a
> > > > hypothetized aN/uN alternation. What did this accomplish,
> > > > other than the addition of some extra assumptions?
> > >
> > > It gives me an alternative way to explain the prenasalized
> > > forms.
> >
> > The prenasalized forms that have -Nk/-mp and do not alternate
> > with a rhotic?

> For a word which does occur before a word boundary, look at a noun
> like *aN- "water"(-> *ur-, *var- and -> *akW-, *am-).

This one pseudo-root alone does not convince me. Do you have others?

> > I see no evidence that the vowel alternation has to be link'd to
> > nasality.
>
> There is no other evidence than that other options are worse.

I disagree, I think at least the *o option and the ablaut option are quite firmly in the game.

> > You could just as well assume let's say incomplete rhotic
> > coloring ur > ar, and an incomplete change r > n (> m / _p, etc).
>
> r > n? You can't be serious.

What? You yourself proposed n > r.

> > Or ablaut that's independant of consonantal context.
>
> Calling something ablaut is a statement of fact, not a explanation.

Indeed. So is calling something "an alternation" (cf. your model). The "laryngeal", "umlaut" and "vowel length" scenarios are a few examples of how "ablaut" could come about.

> > Or varying reflexes of an *o.
>
> /o/ is part of the vowel triangle, with intermediates. An /o/ which moves around like that would imply the whole vowel triangle etc did;

No, one vowel change does not need to imply others. One of the features of the ar/ur hydronymy is the *lack* of /o/, so it's possible that in some languages (or contexts) o > a, in others o > u.

> > Or umlaut of some sort.
>
> Umlaut is conditioned by a following vowel. I don't see what that
> would be.

Something that's decayed since then.

BTW your lack or presence of prenasalization was conditioned by position of stress. Have you evidenced that yet either?

> > Or reduction + lowering of a short counterpart of *u.
>
> That wouldn't explain the prenasalised forms.

It would if the vowel length alternation was independant of consonant quality.

> > Or loss of a "laryngeal" that sometimes leaves coloring.
>
> That wouldn't explain the prenasalised forms.

Prenasalized forms in this scenario too would come from something that has the nasal to begin with.

> > Or any of a zillion other mecanisms that are possible but not
> > really reflected in the data.
>
> Why would I propose a mechanism that is possible but not really reflected in the data?
>
> Torsten

Because that's pretty much all I've ever seen you propose. You take an alternation, decide it comes from a different alternation, invent a trigger for this new alternation, and finally invent some soundlaws that turn this alternation into the attested one.

For example *N (invented trigger) > *G/*w (invented alternation) > *g/b > *k/*p (attested alternation).

John Vertical