From: Piotr Gasiorowski
Message: 10885
Date: 2001-11-01
It's hard to assume a late split of *a into *o/*a, since that would presuppose an unlikely phonological conspiracy between far-flung peripheral groups. A merger (even one carried out independently in several groups) is far more probable. It is therefore commonly assumed that (non-Anatolian IE) *a/*o > *a was an areal tendency affecting the "North IE" dialects irrespective of their genetic subgrouping, but leaving out (Proto-)Armenian, Greek, Italic, Celtic, etc. -- including even the ancestor of Indo-Iranian. Names with *a for expected *o in "Italo-Celtic" Europe may point to prehistoric linguistic intrusions (inasmuch as the etymologies in questions are correct -- *al- and *ol-/*el- may after all be different roots).
> By the way, in Olontigi I see an hybrid toponym, formed by a substratum IE (Celtiberian) name *Olontium or something similar (cfr. Alontium in Sicily), plus a basque ending -egi/-tegi 'top' that could explain several endings in -igi/-tigi (just with this characteristic alternation of the basque suffix after a dental or not) in the southern Spain region. This since I think that basque could be a superstratum language to IE languages in Hispania and Aquitania, rather than a substratum.Whatever happened in this particular case, it is in general common for superimposed languages to assimilate the local placenames morphologically, to reanalyse them, or to adorn them with new elements, creating hybrid toponyms or tautological combinations like the River Avon 'River River'.Piotr