Word-final devoicing in German: effects of phonetic and sentential contexts

Abstract

The present study of word-final devoicing in German investigates possible contextual effects that might influence vowel duration preceding a word-final obstruent, voicing into closure of a final stop, closure duration of a final stop, and frication duration of a final sibilant as acoustic correlates of underlying voicing. In particular, the study focuses on the effects of the phonetic environment and sentential position on the target word. The results of the study replicate previous findings that the word-final voice contrast is not completely neutralized in German. However, the results show that this is true only in some phonetic and sentential contexts. The results also show that the word-final voice contrast appears to be neutralized in other phonetic and sentential contexts. The present study suggests that the phonetic rules necessary to implement words that have undergone word-final devoicing must be sensitive to surrounding phonetic and sentential contexts, and the underlying voice contrast. To account for the implications of the overall findings, an allophonic devoicing rule is proposed that applies only to underlying voiced obstruents and results in an obstruent that is unspecified for the value of its [voice] feature.