Crossing the great divide: the sonorant-obstruent contrast in phonology
Crossing the great divide: the sonorant-obstruent contrast in phonology. NWO VENI (2008-2012); researcher: Bert Botma
Project data
| Full title | Crossing the great divide: the sonorant-obstruent contrast in phonology | |
| Duration | February 2008-January 2012 | |
| Nature | NWO VENI | |
| Researcher | Bert Botma |
Project description
A fundamental issue in cognitive science, including linguistics, concerns the way in which categories are represented in the mind. In phonology, the smallest units of representation are usually taken to be features, members of a set of discrete categories which combine to form speech sounds. One such feature is [sonorant], which divides the sounds of a language into two classes, sonorants and obstruents.
However, it has been observed that some languages have sounds with both sonorant and obstruent characteristics (see e.g. Rice 1993). The realisation of these sounds suggests that they are obstruents, but their behaviour suggests that they pattern as sonorants, and should therefore be identified as sonorants phonologically. It has also been shown that some sonorants, e.g. /m/, display behaviour which is in some respects more obstruent-like than that of other sonorants (see Botma 2004). The status of such ambivalent sounds has so far received only scant attention in the literature. Their existence raises important issues with respect to the nature of phonological feature systems and linguistic categorisation.
The aims of this project are:
- to investigate two cases of ambivalent sonorants, viz. the voiced stops of languages of western North America and of Benue-Kwa, and
- to ascertain by experimental means the potential 'ambivalence' of Germanic /v/, an apparently obstruent sound which in other languages, e.g. Hungarian, displays sonorant-like behaviour.
The project's empirical findings are interpreted in the model of Botma (2004), an innovative approach to segmental structure. The model's reliance on relative prominence, formalised in terms of head-dependency relations, allows it to distinguish between normal and ambivalent sonorants by structural rather than featural means. Thus, the model can account for segmental ambivalence in a discrete categorical fashion.
The project's combined empirical and theoretical orientation makes it of interest to a wide audience of linguists.