Role of ideophones in Bantu and Niger-Congo in historical linguistics

[jess tauber]

Been oh so quiet here discussion-wise since that whole thing about He-who-shall-not-be-brought-up. Hope all noses have been successfully re-jointed.

Since summer I've been back working on ideophones, finally doing that mass crosslinguistic survey that I'd been meaning to get to for decades now, after the suggestion by Johanna Nichols that I do some actual statistical analysis back while I was at UCB a long, long time ago.

Currently Bantu and Niger-Congo are the focus. Despite claims by some that ideophones there are not reconstructable some in fact ARE- with the bases from which they are snipped going back to common Bantu. Because of this one cannot escape from historical changes- unusual in many cases for ideophones in other language families, but apparently not here.

Even so there is a pattern underlying the ideophonic systems, so far as I can tell. Vowels generally fit the common crosslinguistic pattern, with for instance /a/ meaning 'flat(ly)', /o/ having to do with rounded volumes, and so on. Consonants (once you've peeled away the historical changes) also fit nicely- for instance many, many labial initial forms dealing with collapse against some firmer (though not necessarily dry) surface.

Interestingly, the derivational extensions also appear to be in many cases iconically organized, a surprise. For instance 'locational' -am- has a far more specific reference than usually admitted- orientation at an angle less than normal/perpendicular to some reference plane segment. Other extensions also have more concrete spatial reference as well, some temporal. 

Indeed, the entire construction type, with ideophone/noun prefix, unidentifiable intervening segmental materials, and spatial/temporal suffix, looks a lot like the 'bipartite construction' so popular in Western North America, and Yahgan in South America. However, the senses are flipped.

In Yahgan, for instance, the entire set of instrument/bodypart/manner prefixes feeds into the grammaticalization of voice, while in Bantu the spatio-temporal suffixes do. Yahgan pathway/location suffixes grammaticalize into TAM markers- while it may be the ideophones which serve this function, at least for some elements of aspect (other inflections being on the auxiliary).

The American bipartite systems deal primarily with controlled activities, with a minority having to do with loss of control- yet so many ideophones in the Bantu languages appear to deal with loss or lack of control, in keeping with ideophones in many other families.

The questions I have about the Bantu (and larger Niger-Congo) situation are- even through multiple historical changes, how can you keep the system iconically organized? Does this require constraints on allowable changes to keep them relatively transparent? Or does it mean that replacement is the norm, with new, more iconic forms pushing out old, opaque ones?

Finally, are we dealing here not with 'grammaticalization' per se, but its evil twin, antigrammaticalization, which would explain the continuing use of these constructions in the creation of 'new' ideophones?

How many other languages families are amenable to such analysis?

Jess Tauber
phonosemantics at earthlink.net


_______________________________________________
Histling-l mailing list
Histling-l at mailman.rice.edu
https://mailman.rice.edu/mailman/listinfo/histling-l