Centaurs redux.

I'm reading Dumezil's Mitra-Varuna: An Essay on Two Indo-European Representations of Sovereignty  in the Derek Coltman translation (Zone Books, 1988). I'm somewhat cautious, in that this work dates from 1948. Anyway. He discusses the distinction between the Roman luperci and flamines. The latter were the magic priest-kings, the apotheosis of religious and civil respectablity (think of Queen Victoria and Prince Albert) . The former was a 'secret society' of men outside of but in parallel to the public state religion, and allowed to do their herterodox highjinks once a year at new years (running naked through the streets, whipping women with goatskin strops); the older term for the luperci was 'februus' (as in February, which was once near the new year). He finds exact parallels to this in the Greek and Vedic documents.

He relates the stems Gandharva, Februo and Kentauro as reflexes of the same PIE word. Gandharva relates to the Indic version of the luperci, while the Centaurs are Greek version. Now, this stuff is 50 years old, but here goes. Material in [brackets] is supplied by myself.
 

We are therefore justified in regarding the identity of the three names, Gandharva, Februo, Kentauro- -- give or take a few articulatory nuances -- as a probability. From the phonetic point of view alone, it is true they can be explained in several divergent ways, but a convergent explaination is also possible. Gandharva by Indo-European *Guhondh-erwo [superscript u], Februo by IE *Guhedh-rwo [again, superscript u] (for the ending cf. -ruus from *-rwo in patruus), Kentauro- by IE Kent-rwo  [filled underdot under the r]. The differences between the first two can be explained by quite normal shifts (depending on vocalic stages, presence and absence of "nasal infix"). As for the third, its unvoiced occlusives (k-t-), contrasting with the voiced aspirate occlusives (guh-dh [superscript u]) of the other two, insert it into a set of doublets collated by Vendryes (Memoires de la Societe de Linguistique, XVIII, 1913, p. 310; Revue Celtique, XL, 1932, p. 436), and this consonantal shift, appearing precisely in roots that indicate a swift or expressive movement of hand or foot ("seize", "run", "recoil"), as well as in names of animals ("he-goat") and parts of the body ("head"), would be appropriate on more than one count in the names of beast-men, Indo-European maskers, swift runners and great ravishers.

I don't know what he's talking about. I think he is talking in part about the 'secret society' of men responsible for initiating boys into manhood (and their education in warlike pursuits) and those aspects of masculinity expressed among men in modern days in the locker room and the barracks -- the stuff we men prefer to keep away from the women, or at least, unmentioned to them as a kind of open secret; it's the vessel of the double standard.

Anyway. Is it possible to relate these three words to a common PIE ancestor -- and what does it mean? Male sodality? Lodge-brothers?

Or is this to be rejected out of hand? Robert Graves seems to have been infected by this idea when he writes of centaurs.

Mark.