From: dgkilday57
Message: 69565
Date: 2012-05-10
>Celts developing in place defy archaeology. Celtic place-names with conserved inherited /p/ (as you claim for Porcobera) defy the concept of Celtic.
> 2012/5/9, dgkilday57 <dgkilday57@...>:
> >
> > --- In cybalist@yahoogroups.com, Bhrihskwobhloukstroy
> > <bhrihstlobhrouzghdhroy@> wrote:
> >> Your theory is a quite classical one. I used to adhere to it. It's
> >> far from complex: before the Celts, there were IE Ligurians. Main
> >> differences between Celtic and Ligurian are this latter's treatment of
> >> PIE */gwh/ as /b/ and of syllabic */r/ before stops as /ar/. Ligurian
> >> was spoke until 5th c. BC(E).
> >> My theory is: before 5th c. BC(E) Celtic invaders there were
> >> already Celts, not only from 13th c. BC(E), but since PIE times. There
> >> are no differences between both varieties of Celtic.
> >> Which one is simplier?
> > [DGK:]
> > The one which explains /b/ and /arC/ from *gWH and *r.C without further
> > ad-hoc assumptions, that is, Kretschmer's theory of Ligurian.
>
> Bhrihskwobhloukstroy:
>
> Could You please be so patient to enumerate the ad-hoc assumptions
> in my theory? Yours are two, aren't they? (Being ad-hoc isn't wrong;
> it simply means "not independently founded"))
> > [DGK:] A FEW place-names which cannot be plausibly explainedBut you regard all innovations as taking place in situ. Your doctrine cannot distinguish Celtic from non-Celtic at any time-depth.
> > as Celtic suffice to establish a pre-Celtic substrate.
>
> Bhrihskwobhloukstroy:
> I'm still waiting for "a few place-names which cannot be plausibly explained
> as Celtic", but this very expression is far too simple. It includes
> every non-Celtic place-name (Latin and Germanic ones, for instance)
> and even every place-name whose Celticity isn't certain (and since a
> Latin etymology is almost always possible, e.g. even for Milan <
> *Mediola:num < *mediolus 'middle channel' + -a:num [I don't believe in
> this etymology, but it's nevertheless regelrecht and with this I mean
> on phonological, morphological (word-formation), and lexical levels],
> this would lead to the spectacular conclusion that Milan suffices to
> establish a pre-Celtic substrate). Surely You didn't want to take such
> a strong position. Let's see a bit closer:
> Place-names that show only a part (even just the Late IE section)
> of Celtic sound laws are pre-Celtic only in the sense they are
> conservative. They don't prove the existence of any substrate; at most
> they can suggest it, but they can as well (of course don't necessarily
> need to) represent only marginal areas.
> Place-names that show non-Celtic innovations but no Celtic
> innovations are post-Celtic.
> Place-names that show non-Celtic innovations and only a part of
> Celtic innovations can (= cannot avoid the suspicion to be) later
> incomers.
> Only place-names that show non-Celtic innovations and all local
> Celtic innovations suffice to establish a pre-Celtic substrate.
> The crucial point is then to catch at least one sure terminus anteEvery one has /o:/-grade?
> quem for Celtic innovations in situ. I've proposed one. You reject it,
> so You are left without any possibility to establish a pre-Celtic
> substratum, unless You are able to find another terminus ante quem.
> You Kildays are lucky, so good luck!
>
> BTW: . Note that You mean that they must be explained other than Celtic
>
> >> DGK: "Celtic /ar/ < /o:r/ is uncertain because of the hapax status
> >> of bairt; Celtic etymologies for -ate- don't explain Reate"
> >
> > Furthermore there should be abundant parallel /o:/-grade formations NOT
> > involving Osthoff's shortening from roots without resonants in this
> > position, in Celtic and elsewhere.
>
> Bhrihskwobhloukstroy:
> Every vrddhi formation, isn't it?
> > Tea:num against Tea:te indicates that neuter *-ti is parallel to *-nom, andThen I do not understand your objection to Rea:te as a counter-example. If all your place-names end in -a:te, that is ABSOLUTELY no big deal.
> > Gaulish <litanos> 'broad' (*pr.th4-nos, Greek <platanos> 'broad tree,
> > sycamore', Sanskrit <pr.thu-> 'wide') shows that Celtic appends *-no- to
> > laryngeal-final stems. Therefore, no great shakes that place-names end in
> > (short) -ate. No reason to assume development in place. No historic first
> > in comparative linguistics.
>
> Bhrihskwobhloukstroy:
> I beg Your pardon, I can't understand Your argument. All
> -ate-place-names I've considered have accent on the penult (/a/); this
> proves that it is - as in open syllable - long.
> Jump to "no reason to assume development in place" is in any case
> completely oscure to me. If You mean these place-names had originally
> short /a/ (the only interpretation I can let Your sentence have a
> logical sense, but I strongly hope to be wrong in this), You'll agree
> that this would really be an enormously ad-hoc assumption...
> >> Instead of this non sequitur I propose: let's see how manyYes. First, dump the development-in-place (DIP) doctrine. Since you formerly agreed with Kretschmer's Ligurian and linguistic stratification, and even taught this material, it would be good to know what epiphany led to your conversion. Did Alfredo Trombetti's ghost appear to you in a dream? Or did you recognize that Ligurianists vacillated on the outcome of *kW, and realized that you could get parallel results by dumping Ligurian and proposing various forms of archaic Celtic?
> >> etymologies are possible, beginning from those according to the three
> >> languages that are historically attested in situ. After that, let's
> >> try to establish a hierarchy of probability; all hypotheses have to be
> >> correctly formulated, just one can be probably right (let alone the
> >> case of folk-etymology or the like).
>
> > [DGK:]
> > Such a procedure is obstructionistic.
> >
> Bhrihskwobhloukstroy:
> Do You have a better one? (A less long but equally 'measuring' one, I mean)