Grzegorz Jagodzinski wrote:
> 4) And finally *magh- is also not-so-clear as it could look. It is a very
> complex problem as there are some IE words with the meaning 'big, powerful'
> (if not convinced about relation between 'able' and 'powerful', just compare
> Slavic *mogti 'be able' and *mogQtjI 'powerful'). Just take some more roots
> into consideration:
> a) *me:-/*mo:-/*mE-/*mEi- like in Slavic -me^rU in proper names, Goth. mais,
> maiza 'more', Osk. mais 'more' etc. (see Pokorny, me:-(4))
> b) *meg^h- like in Sanskr. maha-
> c) *meg^- like in Greek mega- (strange -a- here, isn't it?, < *meg^H- ? ),
> arm. mec 'big', Lat. magnus (why -a- in the root?)
>
> How do you explain g^ : g^h : NULL here? And why shouldn't we make a link
> between *magh- and *meg^-/*meg^h-/*me:-? We could explain the strange
> variation assuming another instance of "stop + laryngeal" here (cf. esp.
> Greek mega-). If this laryngeal was H2, then g^H2 > gH2 (like in *magh-)
> would be a simple assimilation. And finally, the a-colour should be
> explained with the influence of H2 (see Latin magnus!).
I believe we have three originally different roots here. Let us first
put the *me:-/*mo:- word-family aside. There are some special problems
with that them (including their putative relatedness to *meh1- ~ *met-
'measure') but at least it's clear that they constitute a separate group.
There remain *meg^h2- 'large' and *magH- 'be able, powerful, capable'.
They are kept neatly separate in Germanic (*mek- etc. vs. *mag-an and
its derivatives) and apparently in Indo-Iranian and Greek. *magH- has
_only_ its original meaning in Slavic, where *meg^h2- doesn't occur as
all (*mog-ti 'be able', *moz^-InU 'mighty', *mog-tI 'power', etc.).
However, since size and strength often go together, folk-etymological
contamination between the two roots, or even their complete coalescence,
was easy, especially where favoured by phonetic convergence, and
happened at least in Italic and Celtic, and possibly in Tocharian
(PToch. *ma:ka: is ambiguous but may well contain *magH-; as for the
meaning, cf. Pol. moc 'power' --> 'many, much'). I'm not sure about Alb.
madh, since I don't know of any independent examples of *-gHj- in
Albanian, but if there are no counterexamples to rule out *-gHj- > dh,
one could analyse <madh> as fully parallel to Lat. ma:ior (cf. madhështi
: ma:iesta:s < *magH-jes-t..-)
Thus my best guess is that Lat. magnus and PCelt. *magino- are formally
cognates of Slavic *moz^InU < *magH-ino-, despite the semantic
contamination with *meg^h2-, and that Lat. ma:ior reflects the expected
comparative (*mágH-jo:s) of *magHu- 'able, powerful' -- cf. Iranian
*magu- 'magician' and perhaps *magHu- 'young person' as reflected in
Celtic and Germanic.
The question remains whether in the case of *magH- we have an underlying
e-grade coloured by *gH or a "fundamental" *a. Perhaps it's the latter,
if Gk. me:kHane: 'contrivance' (with Doric a:, as if < *ma:gH-m.n-a-h2
or the like) belongs here.
Gk. mega, Skt. maha:-, Gmc. mek- and Arm. mec all reflect the originally
athematic PIE adjective *meg^h2- (for details see Jens Rasmussen's
recent postings), often extended with productive suffixes in the
branches (Goth. mikils, Gk. megalo-, Hitt. mekki- < *meg^h2-i-). The
corresponding zero grade *m.g^h2- can be seen in some Greek compounds
with <aga-> 'greatly, very'. There's nothing mysterious about the
aspiration in Skt. (caused by the *h2 and generalised from the weak
cases, exactly as in <pántHa:>) or the /a/ in Gk. mega < *meg^&2, cf.
Skt. n. mahi.
Piotr