On 2008-10-20 22:08, raucousd wrote:
> New question: Why does everyone reconstruct the Germanic form for As-
> Aesir as *ansu, with an -n-. It appears in only one or two sources
> where it looks like an epenthetic -n- in Germanic, of which there are
> reportedly many (and they are often lost with compensating vowel
> lengthening). The rest of the forms (except the Latin/Germanic form
> anses- which I can't take too seriously) were reconstructed by Grimm
> to show what form the word *would* have taken in various Germanic
> languages, but they are not attested or they are not attested in that
> form in those languages. Are you trying to argue that Aesir is not
> cognate with I-I forms like asura/ahura? I think it is.
What you need is some background knowledge. Germanic *n was lost before
*s independently in North Sea Germanic (English, Frisian, Old Saxon) and
Scandinavian. A preceding *a was lengthened and nasalised as a result.
The nasalisation was soon lost, but left a trace in Anglo-Frisian, since
the nasalised *ã: was rounded there and ended up as _long_ /o:/:
PGmc. *Gans- > Ang.-Fr. *gã:s > OE go:s 'goose' (but cf. OHG gans)
In Old Norse, we have _long_ /a:/, spelt <á>, from the same source
(hence ON gás 'goose'). Thus, the very correspondence of OE o:s vs. ON
áss points to a nasalised vowel, and this is confirmed by those Germanic
languages that normally preserve the *n, i.e. German and Gothic. A
Gothic cognate is preserved as <ansis> (acc.pl.), recorded by Jordanes,
and we have the OHG onomastic element Ans- 'divine' (as in *Ansu-xelmaz
> Anselm, *Ansu-waldaz > Answald), corresponding to Old English O:s-
(O:shelm, O:swald, not to mention twenty-odd others with the same first
element).
One interesting consequence of this loss of nasals is the way it
affected the Anglo-Frisian version of the Runic script. The fourth Runic
letter was originally called *ansuz; its phonetic value was accordingly
/a/. In Anglo-Frisian, the _name_ of the same letter evolved into *ã:s >
o:s, so the Anglo-Saxon and Frisian Runic writers adjusted its phonetic
interpretation to /o/ or /o:/ (ignoring vowel length, as they always
did). This is why the original Runic alphabet is called "futhark" (its
first 6 letters were f, u, þ, a, r, k), whereas its insular variant is
known as "fuþorc" or "futhork".
If Indo-Iranian *asura- is related to Germanic *ansu-, its _short_ *a
must correspond to Germanic *an- (> attested an-, a:-, o:-). The
correspondence can't be straightforward, but if the PIE ancestor of the
Germanic word was something like *(h)ansu- or *(h)onsu-, the
Indo-Iranian term can reflect the _related_ zero grade *(h)n.su-,
extended with the adjectival suffix *-ro- (syllabic *n. regularly became
Indo-Iranian *a). We can't be certain that the relationship is real, but
at least it is not ruled out on formal grounds, and therefore remains
possible. By contrast, a simultaneous relationship with *a:tar- is
impossible for formal reasons. Even if both roots begin with *h2a-
(which is by no means sure), that makes them no more related than
English <hat> and <ham>.
Historical linguistics is not about finding similar-sounding words and
writing imaginative stories about their similarity. You can't propose a
correspondence if it doesn't match the known sound changes in the
languages being compared.
Piotr