Jos Tellings will be presenting at BLS 39 at Berkeley this coming weekend. His talk is entitled: “Clitics and voicing in Dutch.” His abstract is as follows:
In Dutch, the process of clitic attachment to verbal stems interacts with voicing processes such as Final Devoicing and voice assimilation. In addition free variation is attested for certain classes of clitics.
Two previous accounts for these data have been given. Booij (1995, 1996, 1997) developed a rule-based analysis within the framework of Lexical Phonology. The
main problem with this approach is the stipulative assumption of various prosodic structures for how clitics attach. Grijzenhout and Krämer (2000) present a monostratal OT solution in which Align constraints for prosodic and lexical words are the main innovation. The main problem with this approach is that it fails to account for certain syllabification patterns, and that it does not account for the free variation.
I propose a two-level OT solution, combining the virtues of both accounts: prosodic structures that are derivable from constraints, and a better empirical coverage than Grijzenhout and Krämer. As for the observed variation, I follow Lahiri et al. (1990) in that the variation is caused by the clitic attaching in different ways to the stem. These prosodic structures arise in my analysis by the variable ranking of the Exhaustivity constraint and a voice identity constraint.