From: tgpedersen
Message: 61934
Date: 2008-12-06
>But if ever-y is to be divided that way, 'ever' is a noun.
> On 2008-12-06 00:37, tgpedersen wrote:
>
> > You forgot 'ever' (of course of different origin)?
>
> Not entirely different. Practically all writers on the subject
> derive OE (n)æ:fre in one way or another from PGmc. *aiw- <
> *h2aiw-. If Anatoly Liberman's analysis (a late comp. of <a:, a:wa>
> 'always, ever') is correct, the -r- comes from the *z of the
> comparative suffix *-izan- (incidentally, it seems to me the
> variant <a:wa> contains two *aiwa-'s = 'ever-ever'). The etymology
> is surely imaginative and thought-provoking, like most of
> Liberman's ideas, and explains the strangely late attestation of
> <æ:fre>. Still, I wouldn't say that it is definitely superior to
> the older etymology, æ:fre < *æ:-feore (*aiwi- plus the dat.sg. of
> <feorh> 'life'), a compound corresponding to the well-attested
> phrase <a: to: feore> 'for evermore' (also <a:wa to: feore> and
> even <æ:fre to feore>). If the second etymology is correct, the *r
> is part of the noun feorh < *ferxWu- 'life', not a suffix at all.
>