Patrick Ryan wrote:
> Pokorny lists it under *ser-, 'fasten together', giving some of the same
> examples.
>
> This certainly looks like an ancient way of expressing 'coupling'.
>
> Do you disagree with Pokorny and seek another source?
Just thinking aloud. I don't believe Wel. serth 'offensive, obscene', is
cognate, first because the initial <s-> is a doubtful match and the
final <th> no match at all, and secondly because <serth> also means
'sloping, steep', and the meaning 'obscene' may be a mere case of
figurative extension. The other Celtic words don't really match Gmc.
*serð-e/a-, so we are left without any plausible extra-Germanic cognates.
However, I've just come across an article by Craig Melchert (Journal of
American Oriental Society 122, 2002) on Skt. sárdi-gr.di- 'the portio
vaginalis of the uterus' (used in the description of the sexual part of
the as'vamedHa- ritual), interpreted literally as 'vaginal penis' ("the
penis inside the vagina") -- hence the reconstruction of Skt. *sarda- or
*sardi- 'vagina', not attested on its own. Now Hittite has the verb root
sart- 'wipe, rub', and in addition to ON serða there's also <sarða>
meaning 'polish, touch', so the story begins to make sense if one
reconstructs *serd(H)- 'rub' with an obvious semantic potential (Skt.
lacks the aspiration for some reason) or even, as Melchert does, *ser-
'bring together with different degrees of violence' plus various "root
extensions" (also *ser-p-, for example).
Piotr