On 2006-03-15 13:50, Miguel Carrasquer wrote:
> I want to explain e.g. Hitt. kir, kardiyas as being a single
> paradigm, not the amalgamation of two.
Yes, but *kr.d-ós is well attested as well, isn't it? and regularising
extensions of various kinds (cf. Germanic *xert-an-, Slavic *sIrd-Ice)
are only natural with morphophonologically "difficult" words. Extended
variants like *k^r.d-ejo-m may well have existed beside simplex *k^erd-
already in PIE, influencing its declension. Examples of very similar
processes could be cited form the history of many languages, cf. Pol.
lis'c' 'leaf' (instead of older <list>), back-derived from the mass noun
<lis'cie> (*listIje) misinterpreted as a plural.
> But I wonder whether there is sufficient evidence to warrant
> a reconstruction n. *h1wosu, mf. *h1wé:sus. In particular,
> *h1we:su- is attested very poorly (Pokorny's Gaul. Vi:suri:x
> besides Gmc. Wisuri:h seems arbitrary).
True. With an archaic paradigm we use whatever broken fragments can be
found and glued back together. I'll think of further examples. I haven't
got a copy of Jens Rasmussen's "Studien zur Morphophonemik..." (1989) to
hand, but I believe he shows some evidence of the pattern there (Jens,
could you help us if my recollection is correct?)
Piotr