01-04-2004 19:16, Brian M. Scott wrote:
> At 4:47:11 AM on Thursday, April 1, 2004, tgpedersen wrote:
>
>
>>BTW the names 'Sciri' and 'Bastarnae' are good evidence
>>that these peoples spoke some type of Germanic. And the
>>word 'saddle' (PGmc *saðula, cf Slavic sedlo, Latin sella
>>< *sed-la) which because it is different from what we'd
>>expect in an inherited Germanic word (cf. Geraman Sattel
>>"saddle", Sessel "seat") must be a loan.
>
>
> Not the only possibility: Watkins suggests the possibility
> of a suffixed o-grade *sod-dhlo-.
Unlikely. The instrumental suffix is not added to the o-grade; and
*-dHlo- is questionable -- Olsen's arguments against it (presented here
by Jens some time ago) are strong. But the most fatal objection to
Watkins's etymology is that the lexical slot "sit + instrumental suffix"
is already preoccupied in Germanic: *sed-tlo- > *set.tlo- > *set.lo- >
*seþla- 'seat' (> OE setl ~ seðel ~ sedl ~ seld). This simplification
looks very old; I wonder if it isn't PIE. There is no reason why Lat.
sella should not be derived from *set.(t)lah2 (a collective).
If Slavic instrumental *-dlo- reflects *-tlo-, we could analyse Slavic
*sed(U)lo in the same way (assuming for the sake of the argument that
the yer is secondary). The solution might be attractive, since it would
allow us to explain why *sedlo shows no Winterian lengthening: it was
*setla- at the time Winter's law operated! (cf. sedmI '7' < *se(p)tm-).
Gmc. *sad(u)laz is mysterious. I've seen suggestions that it might be a
loan from Indo-Iranian, but where in Indo-Iranian do we find a plausible
prototype attested with the meaning 'saddle'? If the word comes from the
same source as Slavic sed(U)lo (including the possibility of a loan from
Slavic), why on earth do they have different vocalism? Any ideas, anybody?
Piotr