Re: Sin once more

From: dgkilday57
Message: 59627
Date: 2008-07-25

--- In, "tgpedersen" <tgpedersen@...> wrote:
> --- In, "dgkilday57" <dgkilday57@> wrote:
> >
> > --- In, "tgpedersen" <tgpedersen@> wrote:
> > >
> > > --- In, "dgkilday57" <dgkilday57@>
> > > >
> > > > --- In, "tgpedersen" <tgpedersen@>
> > > > >
> > > > > Pokorny: '
> > > > > pe:(i)-, pi:- "weh tun, beschädigen, schmähen";
> ...
> > > > > ai. pa:pá- = arm. hivand- "krank" (Ernst Lewy).
> > > > > WP. II 8 f., WH. II 234 f., 264, 283.'
> > > >
> > > > Everything but the kitchen sink. Sometimes Pokorny does
> > > > Another example is his *pel-(1).
> > > >
> > > > > How about *pe:-ik- "something that hurts", later
> > > > > 'de-sabellized' (cf. Osc scriftas, Umbrian screhto, but
> > > > > script-), pe:ik- > pekk-, thus Latin pecca: (old
> ...
> > > > Nor does "de-sabellization" work in your example, since no
> > > > P-Italic language makes -ik- from -kk-. Umbrian <pesetom> is
> > > > generally held to be written for *<peççetom>, as mentioned
> > > > earlier. Oscan retains -kk-; see e.g. <akkatus> nom. pl.
> > > > 'advocates', <Dekkieis> gen. sg. 'of Decius'. Thus, even if
> > > > "de-sabellization" were a valid way of producing Latin forms,
> > > > it could not yield *pecca from an earlier *peika: (and no
> > > > Italic language retains inherited long diphthongs as such).
> > >
> > > I hadn't read Buck's §143 closely enough; it seems the rules in
> > > Umbrian are *-kt- > -ht-, but *-kVt- > -it-, eg *re:kte: >
> > > *re:hte:, but *weghe-to:d > -veitu. But he also, confusingly to
> > > me, wants to posit an intermediate stage *wekto:d
with 'secondary
> > > -kt-'. Since the internal rules of Umbrian were probably as
> > > opaque to the average Roman as they are to me, I don't think
> > > 'de-Sabellizing' -it- as -kt-, and by generalization, -Vik- as
> > > -Vkk- is out of the question.
> >
> > Buck's intermediate stage is necessary to explain the vocalism of
> > <ar^veitu> and <kuveitu>. Without it, one would expect
> > *-weheto:d > *-we:tu, with a long monophthong instead of a
> > diphthong.
> But how would they distinguish 'primary' from 'secondary' -kt-?

They (speakers of Umbrian) would not. Buck's point is that an
intermediate -kt-, resulting from syncope and assimilation, is
necessary to explain the sequence written <-eit-> in these words,
rather than the *<-ehet-> or *<-et-> to be expected if the inherited
*/gh/ had changed to */h/ before syncope could take place.

> > My objections are, first, most average Romans knew little or no
> > Umbrian anyway, would be unlikely to know that <feitu> (for
> > example) corresponded systematically to <facito>, and would have
> > reason to "de-sabellize" a word anyhow.
> Latin dictionaries are full of explanations of this or that word as
> Oscan or Umbrian loans into Latin; I think it must have been closer
> home, some low sociolect in Rome with Sabellian features. We know
> there was such a sociolect, the shibboleth -au-/-o- (Petronius
> 'copo' for 'caupo'). If so, it would for the average Roman have
been a
> question of register, and they are easily muddled.

Yes. What I call the "Sabino-Latin dialect" (as opposed to the
Sabine language itself, which was likely still spoken around Reate)
is one of my hobbyhorses, and it showed precocious monophthongization
of /au/ to /o:/ and /ae/ to /e:/, as well as secondary aspiration of
word-internal unvoiced stops clustered with resonants (hence
<pulcher>, <mamphur>, <sulphur>, <lachrima>, <Orchus>, <Vulchanus>,
<sepulchrum>, SEPVLCRHVM inscc.) and devoicing of original voiced
stops in the same environment (<scintilla>, <te:ter>, <alpus> 'white'
cited by Festus as "Sabine" when true Sabine would be *alf(o)s). No
doubt hypercorrections of register did occur, but there is simply no
basis for extracting *pecca from one. As shown by <pulcher> (note
here the plebeian demagogue P. Clodius Pulcher; the aspiration was as
real as the monophthongization), /k/ was not assibilated before front
vowels. Possibly /sk/ did become /s^s^/ or simply /s^/ before front
vowels in Sabine, hence the doublet <scirpus>/<sirpus> (implying that
Sabino-Latin <scintilla> came from Latin <scindula>, not its Sabine
equivalent), but that does not help with *pecca. If Umbrian
<peççe:tom> had entered Latin as *pessa:tum, the most we could expect
from hypercorrection would be *pescia:tum, with *pescia not *pecca

> > Second, even if they did regularly practice "de-sabellization",
> > they would be unlikely to generalize from "/Vit/ really means
> > /Vkt/" to "/Vik/ really means /Vkk/". In Latin itself, apices on
> > inscriptions show that /akt/ became /a:kt/ (which is Somebody's
> > Law, I forget whose), but no such lengthening occurred
with /akk/.
> That might be Lachmann's law, root vowels are lengthened in the ppp
> stem if the root ends in a voiced stop. No lengthening occurred in
> other roots. If it is, we should expect lengthening -kk- if that is
> from -d-k- (*ped-ka:), not otherwise.

If it was <pe:cca:re>, the length of /e:/ should show up in Celtic
borrowings, as it does for <be:stia> (cf. Ernout-Meillet).

> > The cluster /kt/ and the geminate /kk/ had different effects on
> > their surroundings.
> Unless you were talking about Lachmann's law.

I probably should have kept my mouth shut just then.

> > And finally, if any Romans knew Umbrian well enough to
> > "de-sabellize" its words, they would know that /kk/ had
become /çç/
> > in Umbrian, in the very word under discussion.
> If pecca: is the result of hypercorrecting from a Sabellian-like
> sociolect in Latin, that particular rule might not be present in
> sociolect. Of course that is almost too convenient.
> > > > > It would correspond to Norw. (un-Grimm) peik "böser Streich"
> > > >
> > > > Distinct roots *peik- and *peig-, in Gmc. *fi:h-/*fi:g- and
> > > > *fi:k-, appear to have been confounded due to synonymy
> > > > 'wicked, hurtful, treacherous' and 'marked, branded'.
> > >
> > > Kuhn's usual explanation is that some words were only partially
> > > affected by Grimm, eg the anlaut stop would be affected, the
> > > auslaut one not.
> >
> > That seems unnecessary. One of the difficulties of NWB theory as
> > presented in "Anlautend P-" is that three of the NWB roots have
> > interdentals, /þ/ or /ð/. I do not believe that NWB had these
> > sounds, but that the Gmc. words involved (including 'path') are
> > derived from pre-Grimm Gaulish loanwords.
> That would be from Gaulish *b-, from PIE *bh-, or PIE *gW-, *gWh-.
> Apart from Kuhn's own suggestion of PIE *gWem-ti- > Gaulish
> *ba-ti- > PGerm *pa-þi-, I can't think of any derivation from PIE
> along the path you suggest of the NWB roots Kuhn presents, several
> which have sound Latin and Greek cognates as they stand. Meid
> supplements with Gaelic lar, NWB Plore (in placenames), German Flur,
> English floor.

I regard 'path' as borrowed from Gaulish *bato- 'beaten' (possibly in
all 3 genders, since all 3 are found in Gmc. forms of the word, with
different referenda implied). This in turn I relate to Latin
<fatuus> 'silly, insipid' (lit. 'suitable for striking', cf.
<caeduus> 'suitable for felling', etc.), also <refu:ta:re> (lit. 'to
beat back', <futuere>, and the true-Gmc. 'beat'-words. The problems
with vocalism arise in my opinion from the PIE root *bheH2w- having
metathetic forms as well as a participial stem *bh&2(w)to- in which
the */w/ was squeezed out, leading to reformations and a cacophony of
forms seemingly derived from *bha:u-, *bhau-, *bhu(:)-, *bhut-,
whatever. My notes are currently misplaced.

*pitt- and *piþ- in my opinion are distinct roots, the former NWB,
the latter Gaulish (*bet- from *gwet-?) borrowed before the Grimm
shift, like most Gaul. appellatives in Gmc.

*peþil- / *peðil- could represent borrowing from Gaul. *betilo:n vel
sim. 'birch-place' (the Gaul. 'birch'-stem has many variants
reflected in Gallo-Romance), with transfer of sense (though I admit
birches are not specific to lowlands, so this is somewhat difficult).

> > (I am aware that the Much-Kluge derivation of 'path' from
> > remains popular after a century, and a posting arguing for
> > instead is in preparation.)
> I look forward to it. Perhaps also for Kuhn remaining roots in
> p- ? ;-)

At some point I would like to undertake a thorough review of all the
roots in "Anlautend P-", with an eye toward developing NWB phonology
in sufficient detail that other NWB roots, not beginning with /p/,
could be identified. Getting anything postable here is likely a year
or more away.

It occurs to me that *patt- could be a NWB root similar in sense to
Kuhn's *pau(w)- 'treten', which he relates to L. <pavi:re> and
identifies as the source of Gmc./Rom. *paut- 'Pfote'. Spanish and
Portuguese <pata> 'animal's foot and foreleg' could come from a
Frankish *patta, with the same ultimate source as Middle English
<patten> 'to strike, tap, pat', <patte> 'mass formed by tapping',
namely NWB *patt- 'to strike, beat down, tread upon' vel sim.,
perhaps originally a frequentative, cf. L. <pactum> 'driven in, fixed
in place'. (Conventional wisdom regards the Sp./Pg. and Eng. words
as independent onomatopoeic formations.)

> > > > Kuhn's *peik- is a good NWB example, in my opinion.
> > >
> > > I want you consider what you are saying here, since it is pretty
> > > momentous. If one admits of non-Grimm-shifted forms in Germanic
> > > outside of the NWB area between Weser-Aller and Somme-Oise (and
> > > possibly in OE and descendants, since ex-NWB-speakers might have
> > > participated in the Saxon invasion), that is, in Lower Saxony,
> > > Schleswig-Holstein and Scandinavia, there can be only two
> > > explanations:
> > > 1) People from the NWB area emigrated to there, or
> > > 2) Germanic is intrusive in those areas, people used to speak
> > > non-Grimm-shifted (but closely related) languages.
> > > Now that is basically the content of my dispute with George.
> > > Kuhn never mentions this as a problem, for some reason, he just
> > > seems to ignore it.
> >
> > I regard (2) as partially correct. I take the Elbe-Oder
> > plus southern Scandinavia as the Germanic homeland, and the place
> > where the Grimm shift actually occurred.
> Southern Scandinavia is out of the question.
> This is for Jysk, but as you can assure yourself by any dictionary,
> standard Danish and Swedish are full of words in p-.

They are now. By "Scandinavia" I did not mean to include Denmark; I
was thinking etymologically, not politically.

> > Lower Saxony is outside the interfluve, Schleswig-Holstein
> > NWB-speakers appear to have had no presence in Scandinavia, since
> > Old Norse has hardly any native words with initial /p/.
> This is Kuhn's objection to that argument, from 'Anlautend p-':
> 'Heimat in einer niederen Schicht, darin stimmen die
> Wörter mit anlautendem b- und die meisten unserer Stämme mit p- im
> Anlaut überein. Davon ausgenommen sind bei uns fast einzig solche
> Wörter, die mit sachlichen Neuerungen und dem Handel gekommen sein
> können, lateinische wie andere. Ihre Sonderstellung tritt schon in
> alten Quellen klar an den Tag. Die Lieder der Edda im Norden, der
> Beowulf in England, der altsächsische Heliand und Otfrids Werk im
> Süden, diese vier ältesten großen dichterischen Werke oder
> in vier Teilen Germaniens zusammen enthalten von den erörterten
> Wortstämmen (ohne die lateinischen) nicht mehr als fünf, davon vier
> typische Kultur- oder Wanderwörter: *paþ-„Pfad" (Beow. Otfr.),
> „Rock" (Beow. Hel.), *pan(n)ing- „Pfennig" (Edda Otfr.) und *plo:g-
> „Pflug" (Edda Otfr.). Der fünfte Stamm ist *pleg-(pflegen), dessen
> Bedeutungskreis auffallend weit und dessen Herkunft und älteste
> Verwendung unklar sind. Andere alte Kulturwörter haben wir zum
> Beispiel in Pfand, Pfarre, Pflicht (als Schiffsteil) und Pfriem. Die
> leicht aus dem Indogermanischen herleitbaren Stämme sind unter ihnen
> sehr in der Minderheit.'
> ie. that OE, OS and OHG are equally lacking in words in p- (OHG pf-
> which could be explained by assuming that what is recorded first has
> the least of the words of the language of the unwashed (NWB et al.)
> masses.

Yes. That is a very good point. The earliest surviving literature
is likely to represent an "echtgermanische Dichtersprache" having few
if any NWB loanwords. Therefore, it was rather rash on my part to
exclude southern Scandinavia (sensu stricto) from the NWB homeland.

> > Norwegian could have picked up a few NWB words later on, perhaps
> > from Middle Dutch.
> A large part of the vocabulary of the three Scandinavian languages
> assumed to be Low German, which played the same role here as OF in
> England. The usual procedure is to assume that if a word exists in
> three Scand. l. and in Low German, but not in ON, it is a loan from
> Low German, but that is of course no proof that the word didn't
> in a low register here before the time of the dominance of the Hanse
> (Low German was the language of the Hanse, the records of the Hanse
> meetings, now in Mal/bork, German Marienburg, are partly in Latin,
> partly in Low German)
> > Krahe recognized the absence of Alteuropäisch river-names in the
> > Elbe-Oder interfluve as a problem for his theory. It is not a
> > problem for my version of the theory, since I take the bulk of AE
> > river-names (exclusive of those like Regana/Regina which do not
> > belong there) as Indo-Iranian and intrusive in the West.
> When do you think Indo-Iranians intruded there?

In the early second, or possibly very late third, millennium BCE.

> I think they're Venetic. That would explain why almost all river
> in Europe are in one language: they were the ones who used them.

Venetic did not merge inherited */e/ and */o/ with */a/. Indo-
Iranian did, and according to my analysis so did Alteuropäisch. I
should probably shelve whatever else I am doing and concentrate on
cobbling together the evidence into a coherent posting.

> Check the maps in the folder
> Placenames untouched by Grimm, maps 08-12, 16
> in the folder
> Maps from Udolph
> in the files section.
> Be aware that Udolph, in order to save Germania as the home of the
> Grimm-shift posits an pre_Grimm un-Grimm rule p > b, t > d, k > g,
> which took place on a select few Germanic words, among them,
> conveniently enough all the river names in Germania, before Grimm's
> b > p, d > t, g > k set in, which all resulted in the present
> situation, which scandalously can be much better explained by
> that there was no Grimm-shift on the territory of the later

Yes. Udolph's attempt to pre-undo Grimm is ridiculously
Procrustean. Just as we have Germanized Celtic towns like Remagen,
and rivers like Neckar, without the Grimm shift, we can easily
explain that the shift took place in a smaller area, before the
Germans expanded into their historical territory.

> > In my view the Proto-Germans were successful in keeping the
> > AE-speakers out of the interfluve,
> That would make sense if AE was Venetic and the Veneti lived
> (survived) in the De,bczyn culture.
> > and the Proto-NWBers were nearly as successful in keeping them
> > of the lower Rhine basin, where only some smaller rivers have AE
> > names.
> > Celtic expansion later shrank the NWB area down to the Weser-
> > Aller/Somme-Oise borders, but as I mentioned Hercynia cannot be
> > regarded as an original Celtic name,
> We can't be sure the assimilation *p-kW- > *kW-kW- took place in the
> para-Celtic spoken in the Hercynian forest.

My point is that the assimilation did not take place in the Hercynian
forest because the "para-Celtic" speakers there were not Celtic.

> > and Kuhn argued (I forget where)
> Grenzen vor- und frühgeschichtlicher Ortsnamentypen
> > that the tribal name Parisii contains *par(a)- which appears in
> > Celtic names as Ar(e)-, so it is likely an NWB name.
> But this is preserved p-, and 'Hercynia' has *p- > h-

Hercynia was taken over by Celts while they still had original /p/,
and so it went to /h/ (when the Greeks first heard of them) and then
zero. The "Belgic" Parisii entered Celtic territory and settled
around Lutetia after original Celtic /p/ had already been lost, so
their name was Celticized with the new /p/.

The river Inn, likewise, was reached by Celts when they still had
original /p/, and its Venetic name *Pennom was Celticized, lost
the /p/, and became Gaulish *Ennon, whence the later Germanic forms.

> > > > > No more need to put a foot in it (although *pe:(i) etc might
> > > > > ultimately be from *ped- "lower(?)" v.).
> > > >
> > > > We need less lumping, not more. Or at any rate we should
> > > > do as much splitting as possible, then look for lumping
> > > > opportunities.
> > >
> > > That's a timing issue. Or: Been there, done that. Ultimately, I
> > > think the *ped- root is a loan, cf. the extreme lumpiness in
> > >
> >
> > That list needs to be whittled down. Roots like *bhedh- are not
> > equivalent to *ped-. Zero-grade is found in Latin <agrippa>
> > 'person born feet-first', which Schulze explained as *agri-pd-a:
> > 'first with the foot'. Incidentally, Nicholson explained West
> > Romance *petitto- 'small' (e.g. French <petit>) on the basis of
> > Latin <pede tectus> 'covered by the foot' (i.e. 'minute',
> > to 'small' by semantic devaluation).
> Ingenious, but not convincing.
> Here's a quote from Burrows: The Sanskrit language
> 'Ancient thematic neuters in IE are very rare. Skt. yugám is shown
> be ancient by the correspondence of Gk. zugón, Lat. iugum, etc.
> Another ancient word is Skt. padám 'step', Gk. pédon, Hitt. pedan,
> which may be presumed to have originated in the same way, though
> direct evidence is lacking in this case. Gk. érgon 'work' with the
> same rare and no doubt ancient apophony as pédon is to be classed
> it. It should be noted that such primitive thematic neuters, which
> according to this theory are transformed m-stems, are not only
> exceedingly rare, but they are the only class which provide certain
> word equations between different IE languages. Thus the thematic
> neuters of secondary origin, namely (1) extensions of neuter
> consonantal stems and (2) the still later though numerous taddhita
> formations, are of later origin. It would be difficult otherwise to
> explain the absence of detailed agreement among these formations
> between the various languages. In this connection also we must note
> the complete absence of the latter two formations in Hittite.'
> and

So the words *yugom and *pedom are not reconstructible to the
earliest PIE. That is not the same as excluding the roots *yug- and
*ped- from the earliest recoverable PIE.

> It seems pretty fantastic, I know, but I think both *pad- and *yug-
> are well travelled Wanderwörter. Under those circumstances it would
> wrong to force either word into a PIE Procrustean bed.

Well, what are we left with if we kick out all roots showing any
peculiarity? An _empty_ bed at the PIE Motel?

> And BTW, I just fell over
> Vannetais Breton paoter "little boy", Dutch peuter "small child"

If there are pre-Celtic NWB/Venetic words in Breton, you could get a
paper published. But if that Vannetais word is only a recent
borrowing ... oh well.