Re: PIE for "eel"

From: tgpedersen@...
Message: 7838
Date: 2001-07-11

--- In cybalist@..., "Piotr Gasiorowski" <gpiotr@...> wrote:
> On second thoughts, Skt. ahi- is *h2ngWHi- rather than *h1eg^Hi- --
cf. Av. az^i (and the the phonetically repetitive poetic formula
<ahann ahim> 'he killed the serpent', which may well go back to pre-
Indic times). This reduces the evidence for *h1eg^Hi- 'viper'. The
main witness is Greek, where <ekhis> may be idiosyncratic
("stinger"? -- cf. the "sea-urchin" and "hedgehog" words). I wonder
if there is a real basis for reconstructing *h1eg^Hi- as a
PIE "snake" word. Baltic, Slavic and Latin require only *h2(o)ngWH-i-
for "snake" and its derivatives for "eel". Gk. ophis and Arm. iz^ are
a little irregular but at any rate require a root in *-gWH-, not *-
g^H-; <ophis> may be a compromise between the ablaut variants *aphi-
and *omphi- (as you seem to suggest) or between *omphi- and ekhi-.
> Sorry for these changes of opinion, but I'm just analysing the
material as the discussion goes on; please treat the above as
tentative thoughts to be verified later :)
> Piotr
Being the superstitious Platonist that I am (or rather, being one who
suspects these ancient peoples of being just that) I wondered whether
there was an underlying verb "to wind, to be bent, to be crooked"
which verb then might have an n-infix (or be perceived to have, which
by back-formation, removing the -n-, would add up to the same thing).
Were snakes then snakes (in an Linnéan sense) or were they emanations
of the principle of snakeness or crookednes? This discussion might
look fatuous, but I think it determines in the end what we will
accept as semantically "contiguous".
One of the reasons I wonder is because in the course of collecting
material for my clever Austronesian theory, in that material there
seemed to condense three collections having to do with "creation",
maintenance" and destruction, respectively, and *H-n-g- (or similar)
was exactly that (principle of) "destruction".