PIE *kap- and *ghabh-

These two roots which are reconstructed in Proto-Indo-European (PIE) as *kap- and *gʰabʰ- in the traditional notation are invading my mindspace lately. However, so far, I don't know what to make of this vexing doublet. They both mean the same thing, "to take", and so it's not a big leap of imagination to wonder if they're originally from one and the same root. The "uvular proposal" for PIE phonology that I've previously mentioned a few times on this blog (which would give us *qep- and *ɢeb- respectively) doesn't diminish the curious similarity between these verb roots but in some ways enhances it.

My immediate hunch is that there was some dialectal mixing going on between early IE dialects. It gives me a fuzzy feeling of psychological validation when I find others online who have suggested the same thing before me. Douglas Kilday suggested something similar on February 19, 2006 under the topic Indo-European Typology and Sanskrit Phonology (in relation to the rarity or absence of *b in PIE) on the Languages Forum of Groupsrv (link here):

"A variant *gʰab- (not *gʰabʰ-) is required by the Italic forms, Lat. habēre, Umb. habe, habia, etc., Osc. hipid, hipust. It seems we have a "standard" PIE *gʰebʰ-, with the regular e/o-ablaut, and non-ablauting "dialectal" variants *gʰab- and *kap-. Several IE branches have reflexes of both *gʰebʰ- and *kap-, while Italic has *gʰab- and *kap- but not *gʰebʰ-. One explanation is that Pre-PIE gave rise to a chain of dialects, with e/o-ablaut being peculiar to the "standard" PIE (spoken, presumably, by the leaders of the PIE diaspora). Other dialects had /a/ for the "standard" /e/ and /o/ arising from Pre-PIE /a/ in different environments, and different treatment of the stops inherited from Pre-PIE (perhaps originally only two series). Semantic devaluation of the "standard" root would be one motivation for replacement by a "dialectal" form, and this certainly applies to a root meaning 'seize'. It might also have applied to the third homonym (in Pokorny's classification) of *bʰel-, 'blow, swell', if the original sense was 'strong, powerful', devalued by way of 'large' (cf. Eng. big) to 'swollen, blown'. That is, PIE *bel- may have been a dialectal variant of *bʰel-₍₃₎. Likewise the extended roots in *-b- (if such they are) could involve the borrowing of a dialectal variant of *-bʰ- or *-bʰo-. Of course, other explanations are possible."

Kilday's idea of ablautless paradialects of PIE being at work is interesting but most probably false since ablaut as a whole cannot likely be a recent feature in the development of PIE, which means that any ablautless para-IE languages would be so far removed from PIE proper as to be an altogether seperate language family. I also don't know of any direct descendents of PIE that quickly got rid of their ablaut as Kilday would require. Rather, my instinct is telling me that the "ablautless" forms are coloured by uvulars, hence *-a-, and that ablauting forms stem from a lengthened Narten present, *ɢēb- (or *gʰēbʰ- in traditional notation), since it's already been established by other IEists that long vowels are not affected by laryngeal colouring and thus, by extension, they wouldn't be affected by uvular colouring. Hence the preservation of *e once subjunctives with short e-grade replace most Narten presents as per Jasanoff's theory. An interesting thing is that the only difference between the two forms in the end is the voicing of the plosives. If the voiced variant contains the original ablauting Narten present, then would this mean that *qep- (trad. *kap-) is not the original root form and merely a dialectal variant of an original form *ɢēb-? Perhaps we can modify Kilday's idea and be on the look-out for para-IE dialects lacking voicing contrasts in stops, although it seems even more likely that post-IE dialect mixing could solve this conundrum without an appeal to exotic para-dialects.