We describe the use of a clustering algorithm based on fuzzy set theory for the detection of common and distinctive topics within a document set. This approach generates a highly flexible data structure that has been applied to multidocument summarization.