Consonant gradation in Fula suffixes the ugly truth

Abstract

In this paper, I will argue that Skousen's account of the differences between stems and suffixes requires substantial revision and elaboration. I will argue further that the complexities of the behavior of suffixes preclude the possibility of providing an insightful analysis in terms of auto segmental phonology/morphology (contra Lieber [1983, 1984], Marantz [n.d.]). Instead, the alternations found in suffix-initial consonants, like those in stem-initial consonants [Churma 1986b], require a fairly extensive use of morphophonological diacritics. Indeed, it would appear that an optimal account of gradation phenomena (or at least those that have been heavily grammaticized, such as in Fu1a and Celtic), will never crucially make use of autosegmental tier-separation and association (cf. Churma [1986b], Willis [1986]), although I will not be able to address this issue here.