Stress as a morphological edge

In this paper, I argue that stress is not a phonological primitive. In the CVCV framework, both stress and boundaries have a skeletal exponent that conditions the realization of the segments. I show that in Old Norse, there is an equivalence between the skeletal exponent of stress and the skeletal exponent of the word boundary. By arguing that stress results from the word boundary, the fixed initial stress of Old Norse can be accounted for. Then, I support this hypothesis by showing that Proto-Germanic post-coda voiceless fricatives cannot be accounted for if we do not assume that stress is an internal boundary.