|Title:||Crossing morpheme boundaries in Dutch|
|Authors:||Marc van Oostendorp|
|Abstract:||On the basis of Dutch data, this article argues that many differences between types of affixes which are usually described by arbitrary morphological diacritics in the literature, can be made to follow from the phonological shape of affixes. Two cases are studied in some detail: the difference between prefixes and suffixes, and differences between 'cohering' and 'noncohering' suffixes.
Prefixes in many languages of the world behave as prosodically more independent than suffixes with respect to syllabification: the former usually do not integrate with the stem, whereas the latter do. Dutch is an example of a language to which this applies. It is argued that this asymmetry does not have to be stipulated, but can be made to follow from the fact that the phonotactic reason for crossing a morpheme boundary usually involves the creation of an onset, which is on the left-hand side of a vowel and not on its right-hand side, plus some general (symmetric) conditions on the interface between phonology and morphology.
The differences between 'cohering' and 'noncohering' suffixes with respect to syllable structure and stress in Dutch is argued to similarly follow from the phonological shape of these affixes. Once we have set up an appropriately sophisticated structure for every affix, there is no need any more to stipulate arbitrary morphological markings: differences in phonological behaviour follow from differences in phonological shape alone, given a sufficiently precise theory of universal constraint and their interactions.