A fraction too much friction

Abstract

Typological work shows that voiced fricatives like /β ð/ occur more often without their voiceless counterparts than with them, contrary to what would be expected on the basis of markedness relations between voicing and obstruents. This paper suggests that many of the offending fricatives are more appropriately viewed as sonorants, whose unmarked status is to be voiced. This view has an important consequence for the interpretation of intervocalic voicing (e.g. a f a > a v a ), which we suspect is the diachronic origin of most of the fricatives in our corpus. We propose that intervocalic voicing is sonorization, formalized in terms of the suppression of melodic material.