Re: Portuguese, Spanish bode "buck"

From: dgkilday57
Message: 71130
Date: 2013-03-29

--- In cybalist@yahoogroups.com, Bhrihskwobhloukstroy <bhrihstlobhrouzghdhroy@...> wrote:
>
> 2013/3/28, dgkilday57 <dgkilday57@...>:
> >
> > --- In cybalist@yahoogroups.com, Bhrihskwobhloukstroy
> > <bhrihstlobhrouzghdhroy@> wrote:
> >>
> >> 2013/3/22, dgkilday57 <dgkilday57@>:
> >> >
> >> > --- In cybalist@yahoogroups.com, Bhrihskwobhloukstroy
> >> > <bhrihstlobhrouzghdhroy@> wrote:
> >> >>
> >> >> A Celtic origin would be phonetically implied by a comparison with
> >> >> Bavarian butz, butzel 'person or animal charatcterized by a short and
> >> >> thick form' < Germanic *butti-z, *buttila-z < PIE *bhud-n'i-s,
> >> >> *bhud-n'i-lo-s: PIE *bhud-n'i-s > Celtic *buddi-s >
> >> >> Proto-Ibero-Romance *bodde
> >> >>
> >> > With the accent after the cluster I would expect Celt. *butti-s by
> >> > Stokes'
> >> > Law. If the accent fell before the cluster I would expect *buddi-s by
> >> > what
> >> > I provisionally (perhaps inappropriately) call MacBain's Law.
> >>
> >> *Bhr.: You probably remember that I follow Stokes' Law according to
> >> Zupitza's formulation, which predicts precisely the opposite outcome
> >> (*buddi-s if the accent fell after the cluster, otherwise Celtic
> >> *budnis with retained cluster).
> >
> > Yes, we have so far agreed to disagree on this matter. However, if you
> > accept Zupitza's formulation, you are left with a number of Celtic lexemes
> > with tenues geminatae which must be explained otherwise. For example,
> > Stokes (IF 2:169) refers Celt. *bukko-s 'he-goat' (OIr _boc_, MW _bwch_,
> > etc.) directly to *bHug[^]-nó-s, making it cognate with the Gmc. words.
> > Matasovic' on the other hand suggests "the Celtic forms might be loanwords
> > from Germanic", and we all know how much you hate borrowing, at least in
> > pre-Roman times.
> >
> > Stokes' next example helps clear up another problem with Matasovic'. S.
> > refers OIr _brecc_, MW _brych_, etc. 'speckled, variegated' to Celt.
> > *mrekko-s, from *mreg-nó-s. (We would now write *mrikko-s from *mr.g-nó-s.)
> > He connects this form with Lith. _márgas_ (neglecting the accent) and OIr
> > _mrecht_, OW _brith_, etc. 'variegated, painted', these latter from Celt.
> > *mrekto-s, from *mreg-tó-s. (Again we would write *mrixto-s < *mr.g-t-ós.)
> > M. refers the _brecc_ group instead to PIE *pr.k^-, but cannot explain Celt.
> > *b- for expected *p-. As cognates of _mrecht_ he cites not only Lith.
> > _márgas_ but Grk. _amorbós_ 'dark', the root being PIE *merh2gW-, with loss
> > of the laryngeal in this position in the /o/-grade (illustrated by
> > _amorbós_) being generalized to the zero-grade in *mrgW-tó-s, underlying the
> > Celt. forms.

Oops! Of course Celtic would not have expected *p- but zero, as in the Irish 'salmon' words referred to the same root (with p- retained in Porcobera, naturally).

> > In fact Grk. _amorbós_ means 'follower, attendant' and its placement here is
> > an etymological stretch; one must assume 'not speckled' > 'dark' >
> > 'enslaved' > 'servile' or the like. We may presume instead that *pr.k^-nó-s
> > led to OWIE *pr.kkó-s, and *mr.h2g(W)-nó-s to *mr.hkkó-s, both meaning
> > 'speckled', with a contaminated form *mr.kkó-s and analogical
> > quasi-participial *mr.któ-s (perhaps distinguishing 'made speckled, painted'
> > from 'naturally speckled, variegated'). These would lead to Celt. *mrikko-s
> > (OIr *mrecc, later _brecc_, etc.) and *mrixto-s (OIr _mrecht_,etc.).
> > Alternatively, the contamination could have occurred at an early stage of
> > Proto-Celtic, with *ma(:)rk(k)o-s replaced by *mrikko-s under the influence
> > of *frikko-s (thus avoiding confusion with the new word for 'horse',
> > whatever its source).
> >
> > Back when Torsten and I discussed Gallo-Latin _beccus_ 'beak', we both
> > agreed with your Master, Johannes Hubschmid (Sard. Stud. 106) that it should
> > be considered non-IE in origin. However, Tristano Bolelli (ID 17:151)
> > attributes derivation from PIE *bHeg- 'to break' to N. van Wijk (IF
> > 24:232-3). In fact, v.W. does not propose such an etymology for _beccus_ in
> > this paper, but argues that Gmc. words including OE _becca_ 'hoe', OHG
> > _bicchen_ 'to pierce', and MHG _becken_ 'to hew' are inherited from *bekk- <
> > *bHeg-n-´- by Kluge's Law, not borrowed from _beccus_ or its Romance
> > derivatives. Nevertheless we may compare this situation with 'buck' and
> > infer that G-L _beccus_ indeed continues PIE *bHeg-nó-s by Stokes' Law.
> > Moreover if Gaul. *bekko- 'beak' was borrowed into early PGmc and regularly
> > shifted to *pekko- (later *pekka-), this noun generating *pekko:(ja)naN (wk.
> > II) 'to peck' and *pekkjanaN (wk. I) 'to pick', we can easily understand ME
> > _pecken_ against MHG _becken_, NHG _Pickel_ against _Bickel_, etc.
> >
> (...)
> >
> > And yes, lambs are short compared to oxen, but short and stout? How could
> > they frolic with all those extra pounds?
> >
> > DGK
> >
> *Bhr.: Professor Hubschmid's views about non-Indo-European etyma
> indeed changed, though very late, between late Eighties and early
> Nineties; after some years of shocked astonishment I realized that the
> classical Pre-Indo-European schema was based upon a stratification of
> unverified assumptions and that Proto-Indo-European etyma were
> immensely more trustworthy, but I'm still open to change my mind again
> and again if I'll have time to be made persuaded.

I agree that ceteris paribus, we should prefer an IE etymology for a word attested in an IE language, but it is possible to go overboard with this principle. (And also overboard in the other direction, ascribing almost every difficult word to non-IE, like Tavi and the early Bertoldi.)

> Of course, "a number of Celtic lexemes with tenues geminatae (...)
> must be explained otherwise" than through Stokes-Zupitza's Law,
> especially since plosive + /n/ + (accented) vowel isn't the only
> source for geminates (neither in Celtic nor elsewhere). This
> particularly applies to *bukko-s < *bhug'-ko-s, if we want to preserve
> its root connection with *bhug'no-s (*bhuk-n'o-s from *bheuk- 'big',
> Pokorny 1959: 100, would be reasonable as well, but less attractive).
> For beccus I had already proposed, following the suggestions by
> Delamarre 2003: 70 and 80, a root *bek- (or maybe *gWek-) 'sting',
> unless *bekko-s < *gWet-ko-s (cf. *gWet- 'bulge', Pokorny 1959: 481).

The obvious problem with *gWet-ko-s is that *-tk- should have undergone metathesis outside Anatolian and Tocharian, as in Celtic for 'bear', unless we presume that *-ko- remained productively in use with bare roots. Connecting 'beak' with 'bee' seems rather fanciful, even if a beak is pointy like a sting.

> *Mrikkos is, at least since Pokorny 1959: 733-734 (not in the
> otherwise corresponding entry in Walde - Pokorny 1927: 274), from
> *mrk- (most diagnostic Russian mórok 'darkness, fog, clouds, sunset'),
> of course for Thurneysen's pupils not from *mrk-n'os (but whence
> then?), which on the contrary I find completely regular.

Hmm. Pokorny likes 'speckled' ~ 'dark', while Matasovic' likes 'not speckled' ~ 'dark'. For now I will stick with my contamination scheme outlined above. At least it avoids these quantum leaps of semantics.

> I anyway follow Pedersen in positing *bhrk'- (in his opinion a
> doublet of *prek'-, unfortunately an unacceptable connection) as a
> second, complementary source of *brikto-s and *brikko-s.

I agree that Pedersen's postulation of doublets of this sort is unacceptable. Such pseudo-connections as *kap- ~ *gHabH- only foster laziness in IE research. False friends are inevitable. On the other hand if my contamination scheme holds, an additional root for the words in question is superfluous.

> That's all; every similar instance can be treated in the same way(s).

Yes, but this leads to robotic reconstruction. I have no immediate counter to your mechanism of explaining geminates with *-ko- (and presumably *-to-, still productive since old *-t/d-to- > *-sso-, and *-po-), but I wonder what Kluge would have done. The Law rightly bears his name, even though Bezzenberger and Osthoff made earlier use of it, because Kluge presented enough organized evidence to show that the words in question with geminates started life with /n/-suffixes, beyond reasonable doubt. There may be a way to apply his reasoning to the Celtic problem as well, and at least make it highly plausible that we are dealing with /n/-suffixation and not *-ko-, *-to-, *-po-, or whatever. But this cannot happen overnight.

> As to lambs' stoutness (does this word exist?), I perfectly agree
> with You, but You should rather ask our beloved Old (High) German (or
> Germanic or regional Indo-European) ancestors, because Butze is really
> 'Lämmchen' and butz, butzel (sorry I had mistakenly omitted -l,
> although my reconstruction implied it) does mean 'Person oder Tier von
> kurzer, dicker Gestalt'

Yes, and lambs could be fattened up for feasts, so I have no serious objection to this connection.

DGK