The central topic of this dissertation is the phenomenon Multiword
Expressions. In this dissertation, Multiword Expressions (MWEs) are defined
as combinations of words that have linguistic properties not predictable
from the individual components or the normal way they are combined. MWEs
are studied from multiple perspectives.
The first part of this study describes the design, implementation and
population of DuELME, a Dutch Electronic Lexicon of Multiword Expressions.
DuELME has been created out of a need for a large number of lexical
descriptions of Dutch MWEs organized in such a way that they can be used in
a wide variety of different grammatical frameworks, approaches to MWEs and
The result is a resource that contains lexical descriptions of over 5,000
MWEs, and that can be integrated into Dutch NLP systems with a minimal
amount of manual effort.
A corpus-based study of the variation potential of MWEs has been carried
out in the second part of this dissertation. Central in this part is the
Idiom Variation Potential Hypothesis, which postulates that the presence of
certain properties of the idiom parts are responsible for their in
principle unlimited variation potential. The hypothesis has been tested for
25 idioms taken from DuELME using both corpus data and panel judgments as
Untangling Multiword Expressions is of interest to those working in the
domain of linguistics, corpus linguistics, and to those who are concerned
with the creation of linguistic resources.