[Author Login]
Title:Final Devoicing and Voicing Assimilation in Dutch Derivation and Cliticization
Authors:Janet Grijzenhout, Martin Krämer
Abstract:Final Devoicing and Voicing Assimilation

in Dutch Derivation and Cliticization

Janet Grijzenhout & Martin Kraemer

University of Duesseldorf

Final-devoicing and voicing assimilation have always been regarded as

evidence for level-ordering in Dutch (e.g., Booij 1995). In recent con-

straint-based accounts, some crucial data have been ignored. We will here

propose a constraint-based analysis for all voicing alternations in Dutch.

In obstruent clusters, we mostly find regressive voicing assimilation.

Progressive voicing assimilation occurs when the right member of a clus-

ter is a fricative or a plosive which is part of an inflectional affix

and this is unaccounted for in recent Optimaltiy Theory-approaches (e.g.,

Lombardi 1996). Stem-final obstruents in Dutch behave differently in dif-

ferent types of affixation and in cliticization. We provide evidence for

the assumption that one class of affixes form a prosodic word together

with the stem (i.e., they are so-called 'internal affixes' in the sense

of Selkirk 1995) and that the other class of affixes form an independent

prosodic word (i.e. they are 'external affixes'). Clitics, on the other

hand, are not part of a prosodic word, but rather of a higher prosodic

unit. In our view, final devoicing results from a markedness constraint

on the right edge of Prosodic Words (on positional faithfulness and mar-

kedness, see e.g. Zoll 1998).

Voicing assimilation is explained by us by a surface correspondence rela-

tion between adjacent obstruents. The different behavior of different

types of segments with regard to the feature [voice] will be explained by

the interaction of different Identity constraints on plosives and other

segments. Furthermore, we propose that Identity-[voice] constraints are

sensitive to prosodic domains. Derivational approaches to these phenomena

had to employ rule ordering as well as level ordering (see Berendsen 1983,

Booij 1995, Trommelen & Zonneveld 1989). In the proposal made in this

paper, neither intermediate lexical levels (Kiparsky 1998), nor so-called

'sympathy' (McCarthy 1998) need to be assumed to account for voicing al-

ternations in Dutch.
Type:Paper/tech report
Article:Version 1