Re: Re: [tied] Osthoff's Law and T > M in Gaulish loanwords (was: Is Basque IE?)
---In email@example.com, <firstname.lastname@example.org> wrote:2013/10/8, dgkilday57@... <dgkilday57@...>:
> 2013/9/20, dgkilday57@... <dgkilday57@...>:Bhr.:You'll admit a spelling <Argio-> for /ar-gi-(j)o/ is better
>> >> [...]
>> >> [DGK]
>> >> But _landa_ 'campo, pieza de terreno' occurs widely in Basque (Bisc.,
>> >> Guip.,
>> >> Aezc., Lab., High & Low Nav., Ronc.) and appears to continue Gaul.
>> >> *landa:
>> >> directly. Moreover a Late Gaul. *lanna would have given Bq. *lana,
>> >> since
>> >> Latin _anno:na_ gives Bq. _anoa_. If _larra-_, _larre_ is borrowed from
>> >> Gaulish, it probably continues a collective *la:rja: (or *larja: by
>> >> Osthoff's shortening) 'flat area' from *la:ro- 'flat surface, floor',
>> >> PIE
>> >> *pl.h2-ró- or *pléh2-ro-. (Latin _pla:nus_ can represent *pl.h2-nó- and
>> >> provides no evidence for a heteroclite, pace Matasovic', only for
>> >> different
>> >> suffixes.)
>> > Bhr.:
>> > All right, I just wonder why Osthoff's shortening in *la:rja:
>> > (wouldn't it rather develop a Sievers' variant *la:rija:?)
>> > DGK:
>> > In his draft glossary Matasovic' has several examples of heavy syllables
>> > not
>> > developing a Sievers' variant, e.g. Proto-Celtic *argjo- 'white; snow' >
>> > Gaul. Argio-talus, etc.
> Bhr.: Are You sure that Argiotalus is /ar-gjo/ and not /ar-gi-(j)o/?
> DGK: No, of course not, just as I cannot disprove the existence of the
> Easter Bunny.
conceivable [than] Easter Bunny
Perhaps, but I get suspicious when Sievers' Law is invoked and it only involves *j and *w (or *j alone in Ringe's treatment of Proto-Germanic). Properly understood, Sievers' Law should affect all the resonants, *j, *w, *r, *l, *m, and *n, in post-heavy position. When most of the evidence is missing and explained as effaced by subsequent processes, it is like the Easter Bunny's missing tracks.
>>Bhr.: "Only marginally better" includes omitting the reference to an
>> >[DGK] I do not know whether Osthoff would be applicable
>> > in this time frame; hence my parenthetical guess. A Basque borrowing
>> > would
>> > not distinguish length.
>> > Your own
> Bhr.: van Windekens'
>> > [DGK] etymology of _glo:ria_ does not involve Sievers.
> Bhr.: nor Osthoff
>> > [DGK] My problem is
>> > that I never heard of being "green with glory",
> Bhr.: "bright, enlightened with glory"
> DGK: With a greenish tint? That might work for Popeye, whose glory came from
> spinach. This etymology is only marginally better than van Windekens'
> explanation of Greek _thálassa_ as 'flat water' (BzN 1:200, 1949-50), using
> IE *tel-, *akWa:- with his Pelasgian soundlaws.
unattested language like Pelasgian, that for You is evidently no
problem at all
Van Windekens' Pelasgian is one to which I do NOT subscribe.
>>Bhr.: You perhaps remember I completely agree
>> > [DGK] so I find the route Gaulish
>> > *klovesja: (vel sim.) > Ligurian > Old Latin more plausible, as with
>> > _gladius_. However, in order to investigate this theory that initial
>> > Gaulish tenues became mediae when borrowed into Ligurian (which I now
>> > attribute to differences in initial consonant strength, NOT aspiration),
>> > I
>> > need to get a copy of J.U. Hubschmied's long paper on Late Gaulish in
>> > Vox
>> > Romanica vol. 3 from the university library. Once I get this paper, it
>> > may
>> > turn out that counter-examples kill my theory.
> Bhr.: if You formulate Your theory in these very terms, I don't think
> You'll find counter-examples there, because Hubschmied doesn't mention
> DGK: No, but I had to peruse the paper to ensure that nothing deduced by H.
> for late or dialectal Gaulish could explain the anlaut-voicing without
> further assumptions. It appears that my theory is still viable. Two other
> issues in the paper are worth noting. First, H. places *-eu- > *-ou- within
> the history of Gaulish, against Matasovic' who considers it a Proto-Celtic
> shift. If H. is correct about Tissiniva < Gaul. *tegesa newia (pp. 49-50)
> and Leuca/Leu(g)k < *leuca: (86-7), it becomes difficult to accept M.'s
> contention that -eu- is merely a "spelling variant" of -ou- in old names.
Yes, but you also lump Ligurian together with Celtic, which amounts to extending the base of Celtic downward until it becomes Western IE itself. Since I distinguish Ligurian from Celtic, the question whether Gaulish had -eu- in historical times is not trivial, and separate from the attestation of Ligurian *nevio- in Nevius and Neviodu:num.
On a related matter, I see that Hubschmied earlier (Ausdrücke der Milchwirtschaft gall. Ursprungs, VRom 1:88-105, 1936) was forced to posit a bifurcate development of Late Gaulish intervocalic stops (notes 2 & 3, pp. 99-100). He has (Late) Gaul. *jutta corresponding to Ins. Celt. *juta:, but Late Gaul. *ridda for *rita:, by-form of _ritus_ 'ford', in the place-name Riddes where the ancient road crosses the Rhone. I would explain *jutta by Stokes' Law (which, pace tua et Zupitzas, Stokes intended as having the same outputs as Kluge's Law). But *ridda can be understood as the form in which the Ligurians borrowed (Late) Gaul. *rita. Likewise, H. did a good job of showing Late Gaul. *tw- < *dw- and *-pr- < *-br-, so *-bbr- from the same earlier *-br- is a bifurcation, until we recognize that the Romance forms requiring *-bbr- in the word for 'two-handled tub, firkin, Zober' (from Gaul. *dwibros, Late Gaul. *twipros elsewhere) are restricted to Liguria and adjacent Piedmont and Lombardy. If the remnant Ligurians borrowed *rita as *ridda, they could equally well have borrowed *twipros as *twibbros. Recognizing both Stokes' Law for Celtic and the voicing (with word-internal gemination) of Gaulish tenues when borrowed into Ligurian could clear up much of the difficulty with these footnotes.
>Bhr.: Sorry, I fear I don't understand Your last sentence; could You
> DGK: Second, H.'s treatment of the Interlaken/Inderlappen problem (pp. 52-8) is
> flawed by ad-hoc assumptions (Celt. *-kw-, not *-p-, maintained by
> Systemzwang, then Late Gaul. *-gw- > *-bb- > Alem. -pp-). Harald
> Bichlmeier's more recent explanation (HS 122:257-62, 2009) is not much
> better. I think an Illyrian place-name makes the most sense, but the
> prospects of convincing Celticists of this are dim.
reformulate it, please?
What I mean by a dim prospect is a low probability. My arguments about Hinderlappa (to be presented in a separate post) are not likely to convince most of today's Celticists, who have little or no interest in Pre-Celtic languages. Similarly I doubt they would attach any significance to *ridda or *twibbros.