From: dgkilday57
Message: 69805
Date: 2012-06-12
>What precisely is your version of "the Celtic Hypothesis"? How could one possibly falsify it, when you have licensed yourself /p/-retention as an archaic Celticism, and all other traditionally non-Celtic features as regional Celticisms? I am reminded of this "Harold" character on sci.lang who etymologized almost all European river-names as Celtic.
> 2012/6/6, dgkilday57 <dgkilday57@...>:
> >> Bhrihskwobhloukstroy:
> >> (...) Ligurian's diachronic phonology has been put
> >> together in three phases, the first one in a non-IE perspective, then
> >> with more or less happy guesses as if it were an IE language different
> >> form any other one, finally in a Celtological frame; G. Petracco
> >> Sicardi's book is primarily intended as a toponymic reference primer
> >> and therefore makes scarcely any attempt to a dialectologically
> >> coherent representation, but it seems to me (and to her too) that time
> >> is ripe for an improved discussion of that topic and this is wat we
> >> are trying to do.
>
> > DGK:
> > These "phases" overlapped. Alessio was still arguing for a non-IE Ligurian
> > long after Kretschmer's paper (which is hardly a set of "happy guesses"),
> > and Rhy^s was arguing for Celticism even before Kretschmer's paper. I would
> > say there are three schools of thought here. I belong to Kretschmer's, and
> > you belong to Rhy^s's.
> >
>
> Bhrihskwobhloukstroy:
>
> Yes, sure. With "guesses" I meant something neutral: one realizes
> (or proposes) that language X is IE, but has then to discover the
> detalis of its diachronic phonology; he has to start with equations
> where, for instance, vocalism can either represent the direct
> evolution of one and the same vowel or exhibit the direct evolution of
> a different ablaut grade: just like Aryan /a/ in closed syllable vs.
> Greek /o/ can be either PIE */e/ or */o/, Ligurian vowels can be
> directly equated with their matches in the corresponding words from
> other IE languages (this is the guess) but can also represent, as an
> alternative, a different ablaut grade (this is a guess as well).
> In the Celtic Hypothesis, the only - great - guess is the
> hypothesis itself; all the rest are corollaries (evaluation of vowels
> included).
> >> > DGK:Albanian is irrelevant to Illyrian, since Durante showed that A. cannot be descended from I. Krahe regarded */or/ and */ol/ as pre-cons. reflexes of syllabic *r. and *l., and we have <-ikkos> '-horse' in Tarentine Greek personal names, thus Messapic (and presumably Q-Ill.) *k^w > */kk/; since onomastic and other evidence shows that Ill. is a centum lg., there is no problem with *pr.kWah2 > Q-Ill. *porkka: > Rhaetic *porka(:).
> >> > On another matter, however, since Celtic and Italic share the
> >> > assimilation
> >> > *p...kW... > *kW...kW..., reflexes of the tree-name *perkWu-,
> >> > *pr.kWeh2-
> >> > with */p/-drop must have been borrowed rather than inherited by Celtic.
> >> > This applies to Hercy:nia, Orku:nia, Arku:nia, and Piemontese <olca>.
> >> > Trentine <porca> is presumably "Rhaetic" (in Hubschmied's sense),
> >> > "Illyrian"
> >> > (in Krahe's sense); we might compromise on "Rhaeto-Illyrian", a
> >> > Q-Illyrian
> >> > language spoken in Rhaetia.
> >>
> >> Bhrihskwobhloukstroy:
> >> This hardly applies to Hercynia (where do You get long /y:/
> >> from?), because it had since long no */kW/ more (before /u/).
> >> Piemontese olca 'Swiss Pine' is in my humble opinion the same
> >> formation as its Gaulish homophone olca 'brownfield land' < PIE
> >> *polk'ah2 'Gewendetes' (Pokorny 807); in any case it can't reflect
> >> *orka: because no */r/ > /l/ is known either in Gaulish or in
> >> Piemontese (nor Western Lombard and Ligurian intervocalic */l/ > /*r/,
> >> to which a */r/ > /l/ mutation could theoretically consitute a
> >> hypercorrect reaction)
>
> > DGK:
> > Latin <quercus> shows that the assimilation preceded delabialization of *kW
> > before *u. The length presumably comes from the same place as in Lat.
> > <portu:nus>, etc.: denominal adjectival formations in *-h2no- and *-h2ni-.
> >
> > I accept your reasoning about Piem. <olca>, and I will stop citing it in
> > this connection, but we still have Trent. <porca>.
> >
>
> Bhrihskwobhloukstroy:
>
> In Trentine <porca>, <c> = [k], as long as it's before
> non-stressed /a/, can continue both Latin (and Pre-Latin) /k/ and
> /kw/.
> If it were from */k/, it would be a very strange Illyrian word,
> since it would exhibit IE short /o/ (not /a/) but satem treatment
> (/k/) of the labio-velar */kw/, unless You maintain PIE syllabic /r/
> has become /or/ (open [É"]) as in Italic instead of /ri/ as in
> Albanian.
> If it were from */kw/, it would be a conservative IE form (*porkwa:)
> >> > DGK:I think "think" or something dropped out of the question. Yes, I think most Celtologists today take for granted the assignment of Lepontic to Celtic. To his credit, Matasovic' proceeds with caution, observing the flimsiness of basing linguistic assignments entirely upon personal names.
> >> > In my view, we do not yet have the totality of phonetic isoglosses
> >> > between
> >> > Ligurian and Gaulish (the ONLY securely Celtic language spoken in the
> >> > area
> >> > in pre-Roman times), we have only begun to discuss lexical isoglosses
> >> > (*dHeigW- in Lig. but not Celt., and if I am right about interpreting
> >> > Lepontic <teu>, *dHeh1- in Lig. but not Celt.) and morphological
> >> > isoglosses
> >> > (-asc- in Lig., -isc- in Celt.). Therefore, it would be foolish to
> >> > regard
> >> > Ligurian as "almost Celtic" or "para-Celtic" or whatever you are
> >> > driving
> >> > at.
> >> >
> >> Bhrihskwobhloukstroy:
> >>
> >> No, please, no. You are again taking for granted that Lepontic isn't
> >> Celtic and therefore anything Lepontic is automatically non-Celtic. As
> >> for *dHeigW-, You are on one side taking MatasoviÃâ¡ for the Golden Book
> >> of Pure Celtic (as if nothing not included in MatasoviÃâ¡ could be
> >> Celtic), on the other side You are taking everything You want - You
> >> even stretch Ligurian's boundaries according to Your pleasure - as
> >> definite proof of Ligurianness, clearly two exceedingly different
> >> criteria: everything can be labeled as Ligurian if only it occurs in
> >> Western Europe, but nothing can be labeled as Celtic if only it
> >> doesn't occur in MatasoviÃâ¡. Either You take a coherent position or
> >> it's useless to continue
>
> > DGK:
> > All right, I will try not to take the position of Lepontic for granted (as
> > most Celtologists do!) any more.
> >
>
> Bhrihskwobhloukstroy:
>
> Do You most Celtologist take the position of Lepontic as a Celtic
> language for granted, don't You? (Just to avoid misunderstandings,
> it's so difficult to avoid them)