RE : [tied] Re: North of the Somme

--- In cybalist@yahoogroups.com, "caotope" <johnvertical@...> wrote:

>
> > > > No, it's /n,W/ in the donor ar-/ur-/ and geminate language. It
> > > > manifests itself in the receiving IE and FU languages as
> > > > /-mm-/-mb-/-mp-/-bb-/-pp-/
> > > > /-n,n,-/-n,g-/-n,k-/-gg-/-kk-/
> > > > Alternation is what we find in the receiving languages.
> > >
> > > I can accept that not having /N_w/ themselves, IE languages
> > > would substitute things like *m, *gg, *Nk in its stead. I
> > > however see no possible motivation for alternation as wide as
> > > *mm *mp *mb *bb *pp entirely upon loaning.
> >
> > You'll like this one: Russ dub, Polish da,b, Estonian tamm "oak"
> > (/a,/ nasal /a/).
>
> *amb ~ *amm basically (~ *om in Mari & Mordvinic _tumo_). Nasality
> found in all cases.

Or else this is a/u alternation before nasal and thus ar-/ur- / geminate language (I gotta come up with a good name for it).
UEW notes it in some roots as something to explained away.

> > > The acceptors contrast all of those; it is
> > > also unlikely that they all would be allophonic in the donor.
> >
> > No, but note that many are verbs, it might be paradigm
> > alternation.
>
> Yes, I'm fine with that.
>
> > I once proposed that the supposed n-infix appeared originally in
> those number/person where stress fell on the following syllable of
> a latently pre-nasalized voiced stop.
> > http://tech.groups.yahoo.com/group/cybalist/message/46161
>
> You're now linking this to IE nasal infixes? Slo down a bit.
>

On the chance that the substrate, or a medium language it passes through had that kind of pattering: 3sg *-´D-, 3pl *-nD´-, D meaning some voiced stop.

> > And is this coda position only, or do we have initial nasal/stop
> > alternation? If not, I would quite certainly take at least the
> > vacillating nasality as a sign of phonemic alternation.
> >
> > How about the "apple" and "water" words?
> > http://tech.groups.yahoo.com/group/cybalist/message/64461
> > cf UEW
> > omena ~ omæræ 'Apfel' FW and
> > un,a 'Strom; strömen' FU
> >
> > Since I've added the Schrijver's language of bird names to the
> > two others the initial a alternates with zero, letting the
> > following /n,W/ (supposedly) become initial and degenerate to ma-
> > (malum; madeo:) na- (;Grm. nass), wa- (;Sw. våt).
>
> *uNa is a misreconstruction, this points unambiguously to *-w-.
> (Relying on the UEW for reconstructions is a bad idea.)

The entry in Collinder was very short, and I don't know how comfortable people are with long quotes in German or I would have quoted the whole entry.

> It's a fun idea to relate this to the traditional "water" roots,
> but that would likely have to go back a lot farther than typical
> European substrate words.

Why? Because it appeared on the first day of creation? It turns out there are a lot of water-related stuff in that substrate, whatever it is. Whoever spoke it must have been specializing in river travel.

> That, and the lack of evidence nasality,
> then leaves _nass_ basically unconnectable.

Perhaps you forgot to add Schriver's language of bird names.
http://tech.groups.yahoo.com/group/cybalist/message/62677
With that included the a/u of *amlu, the reconstruction for apple, and akWa/apa/upe etc becomes a prefix; removing that cancels the 'zero grade' vowel after the now initial consonant; in Schrijver's examples


'*mesVl-, *a-m(V)sl- 'blackbird' —>

Welsh mwyalch, Latin merula;

Old High German amsla, amasla, amisla, amusla, Old English o:sle


*la&wað-, *a-lawð- 'lark' —>

Old Icelandic lævirki, Old English la:werce,
Old High German le:rahha, le:rihha, Middle Dutch le:werke,
Finnish leivo(nen);

Gaulish (in Latin) alauda



*raud-, *a-ru/id- 'ore' —>

Latin raudus 'lump of ore',
(also Finnish rauta, Northern Lappish ruow'de,
Old Icelandic rauði; or these directly from a descendant of
Proto-Indo-European *h1roudh- 'red')

Old High German aruz, ariz,
Old Saxon arut'

That's exactly the pattern we have in
*amlu-; *malum
*an,W-; *ma-deo:, Dutch nat, Sw. våt

(and BTW Eng. net, Da. våd "net",
cf. Russ. nevod "net" with otherwise inexplicable ne-)

> Of course, we can
> suppose anything, but allowing everything to alternate with
> everything is pseudoscience. We need patterns.

Yup. See above.

> I'm not convinced about _malum_ either, but anyway, even if it
> works, it's clearly secondarily initial here (we can imagine a path
> like a~bal- > amal- > mal-). I was getting at absolute initial
> stop/nasal alternation.

*n,Wal- > *mal-

> > > (Or perhaps vowel nasality, but that could only explain *mp ~
> > > *p, not *m ~ *p, nor POA vacillation.)
> >
> > What's POA?
>
> Place of Articulation. And before you ask, "coronal" is the cover
> term for dental/alveolar.

I've been wondering before whether IE (if, as said, it passed through some IE language) already had spirantization, since it occurs in Iranian, Sabellic and Germanic (and Grimm could have been a generalization). -xW-t- would easily fall apart into -ft- and -xt- forms, from which the whole paradigm could be back-generalized.

> > > A question I have not seen mention'd here is - is the nasal
> > > alternation limited to the labial and velar POAs, or does it
> > > also appear with coronals?

Yes.
Eg. *kant-/*kass-/*katt-/*kunt- etc

> > > > > > > and I'm not sure what you are getting at with "dup-".
> > > > > >
> > > > > > You will be, after you read Schrijver's article:
> > > > > > http://tech.groups.yahoo.com/group/cybalist/message/62677
> > > > > > Search 'dup'.
> > > > >
> > > > > That racks up the alternation count a lot actually. How do
> > > > > we figure they're ALL original alternation and not later
> > > > > reshaping?
>
> > > The idea is that some of the "alternation" we
> > > are seeing would be modifications specific to the individual
> > > languages;
> >
> > But the supposed donor language could be a number of related
> > dialects.
>
> This is compatible with later rehaping.

My best instincts tell me it's a case of picking forms in paradigms, which might already have been regularized in the donor dialect. But of there's also reshaping.

> > > > > _duppe_ works from *-mp- as well.
> > > >
> > > > That's a halfway solution.
> > > > You can always try the compromise solution between trying to
> > > > jump an obstacle and not trying. The results are rarely
> > > > satisfactory.
> > >
> > > I'm trying to gauge how many original forms we REALLY need.
> > > Obviously, the less, the better.
> >
> > I agree. Use /n,W/ as much as possible and be happy.
>
> "Original" on the IE side - economicity at every step.
>
> It's abuse of typology and methodology to use any single feature
> "as much as possible", as opposed to only when required.

It's my new toy you're talking about.

> > > It's ludicrous to posit one substrate root which spontaneously
> > > splits into some two dozen different roots upon being loan'd
> > > into Germanic.
> >
> > Not if it had a couple of centuries first to split into dialect
> > which were then all steamrollered by the likes of Germanic,
> > Baltic and Slavic.
>
> Yeah, again, I can accept that. But not that there is only ever a
> single substrate word, and all the alternation is because the IE
> 'lects interpret things differently (as you write in the quote at
> the top of the message).

'manifests' was meant to mean just 'appears'.

>
> > > > > There may also be secondary influence from the _deep_ root
> > > > > here and there.
> > > >
> > > > Except that the 'deep' root would be part of the loaned
> > > > complex, so there would be nothing for it to influence.
> > >
> > > No - the roots are distinct on the Germanic side. They would
> > > have a lot of time to influence one another during their
> > > evolution towards attestation.
> >
> > In which language?
>
> Potentially any that has multiple forms of the same complex.
>
>
> > BTW Kuhn links in the articles in the files section the place
> > name Dover to his ar-/ur- language. Since it's obviously a
> > cognate, that links ar-/ur- language to the *dub- root and thus
> > to the language of geminates.
>
> You still keep talking like there's only a single language that
> covers all of Europe. This can obviously have not been the case.

No, it's a quote, so to speak. Kuhn talks about the ar-/ur- languager (I think).

> We're more likely talking about a lost family, maybe more than one,
> with a time depth of several millennia.

You're right, of course.

>
> > > > I don't follow you. The a/u alternation is what we find in
> > > > the receiving languages, the original vowel would have been
> > > > something like /a,/, a nasal /a/ (Slavic /a,/ becomes
> > > > Russian /u/).
> > >
> > > OK. That may work for Russian, but not so much for Germanic.
> >
> > Slavic to Russian is a separate issue. I just wanted show with an
> > example that /a,/ -> /u/ has happened before.
>
> There are many, many possible pathways that can turn /a/ into /u/.
>
> > > Another thought: original *o, interpreted as *a/*u by
> > > Germanic/Balto-Slavic?
> >
> > Possible, but then we don't catch the nasalized variants.
>
> Both of *o *o~ are possible. Surely you're not saying that there
> exists nasal alternation in EVERY case?

I think (in some vague sense) that it had to do with voiced stops in IE originally being pre-nasalized. If nasalization changed alternated once in IE mobile paradigms, it might have set a pattern which was copied with the other stops in auslaut.


Torsten