GrandPiano wrote: ↑12 Aug 2020 06:07
Salmoneus wrote: ↑18 Jun 2020 16:23However, we also have to bear in mind that Vasconic had been living alongside Indo-European languages, including undescribed Indo-European languages that may well have been closely related to Italo-Celtic, for hundreds if not thousands of years before the era of Latin-to-Proto-Basque borrowings - and that it may well have migrated alongside Indo-European from the steppe, indicating millennia more of loans, if not indeed a direct family connexion. And numerals are tricky in particular because they're often subject to irregular shifts (like dropping final consonants as a result of the onsets of the next number when counting, for instance).
Hm, that's an interesting point. I suppose it's possible that
sei could be a borrowing from an ancient non-Italic IE language, but I still doubt that that's the case for
bi and
urre, since they only resemble Latin
bis and
aurum as a result of sound changes that happened in Latin (namely dw > b and z > r / V_V). And even if it's plausible on phonological and semantic grounds,
sei being an IE borrowing still seems unlikely to me, since it would mean that the word for "six" was borrowed but not any of the other numbers. Though I suppose it could be plausible if the ancient Vasconic people had a superstition surrounding the number six?
Well, and speaking not to promote a particular theory but just, as it were, to use this as an example of difficulties one can encounter...
...can you really rule out cases of similarity only due to later sound changes, when those soundchanges are themselves common, or areally common?
In the case of /z/ > /r/, for instance, we know that a few centuries after this happened in Latin in Italy, it happened again in northern Europe, in Northwest Germanic. Indeed, it's assumed to have happened independently in North and West Germanic, and possibly even several times independently in West Germanic. It's theorised that /z/ in Germanic, or at least Northwest Germanic, must have been in some way proto-rhotic, prone to rhoticising.
Is it just a coincidence that it's the branches of Germanic that were in western Europe that had this pre-rhotic /z/, when Old Latin (pre-Old Latin?) must likewise have had pre-rhotic /z/ in western Europe? Is it, we might then wonder, also a coincidence that sibilants in Iberia, a part of the Latin world under heavy Germanic influence, have always been so weird?
Let's be more specific in our hypothesis, actually. We know that across Western Europe, in the middle ages, both Germanic and Romance languages tended to have a retracted apico-alveolar /s/, as the reflex of ancestral *s - which has survived in various scattered dialects of both families. Let's assume then that the parents of these languages, Proto-Germanic and (pre-)Latin ALSO has retracted apico-alveolar /s/. [indeed, apparently this is theorised for Latin, explaining why it uses /s/ in borrowings from Hebrew: shabat>sabbath, jeshua > jesus.]
Both these families, then, seem to have two unusual traits: retracted apico-alveolar /s/; and loss of corresponding [z]. Proto-Germanic had /z/, but it became /r/ in all descendents others than those that left the Western European area very early on; pre-Latin had [z] as an allophone of /s/, but it had become /r/ already by the time of Classical Latin. Maybe these two traits are not coincidental? Because if you want a /z/ that might turn into /r/, then a retracted apico-alveolar might be exactly what you'd look for!
And if this form of sibilant was found in Germanic, Romance, and indeed Basque, but is lost (or never gained) in the Germanic that left Western Europe, maybe it was a general areal trait of Western Europe? In which case it could indeed have been found, with the resulting /z/ > /r/ shift, in other IE languages of Western Europe? Indeed, interestingly, Ogham inscriptions do indeed have two different letters for later /s/ - which doesn't mean, of course, that there was necessarily the same contrast between normal /s/ from /ts/ and retracted apico-alveolar /s/ from *s that there was in Germanic and Romance... but it would kind of fit!
And we don't even have to assume a third case of /z/ > /r/, actually. Since Basque doesn't have /z/, it could easily have borrowed a retracted apico-alveolar /z/ as /r/ itself.