Surface analysis of the verbal cluster in Dutch
Simon van Dreumel, P.A.J.M. Coppen · Linguistics · 2003
In this paper, a structuralist account of the verbal cluster of Dutch is given, as implemented in a structuralist surface parser.The description is based on a general principle of expectation, but with a limited number of deviations from the general scheme.It is shown that the implemented parser is able to analyze all clusters, slightly overgenerating to capture those cases where semantic restrictions may apply.In conclusion, it is claimed that the parser incorporates the most complete description of the Dutch verbal cluster so far. The Dutch verbal clusterOne of the major issues in Dutch syntax is the analysis of the verbal cluster, that is, the group of verbs that occurs at the end of the clause, generally resisting intervening nonverbal elements, and followed only by elements that can be argued to have been extraposed.In this article, we aim to provide a full description of the Dutch verbal cluster, with an eye to its implementation in a surface-structure parser called Amazon.1 In this section, we define the verbal cluster, taking various analyses found in previous research as a starting point.With this definition, we narrow down the constructions to be covered in our discussion.The distributional facts of the verbal cluster in Dutch look deceivingly simple: verbs taking nonfinite verbal complements in Dutch surface at the end of the clause,2 where the verbal complements in general3 follow the selecting verbs.As to the arguments of the verbal complements, different verbs allow for different distributions: either all arguments can (or must) be placed adjacent to their governing verb, or all (or some) arguments can (or must) be placed in front of all other verbs: