An Electronic Dictionary of Collocations for European Portuguese: Methodology, Results and Applications

This paper discusses a lexicographic approach to collocations, presenting the methodology, options, results andapplications of an electronic Dictionary of Portuguese Collocations (DCP). The methodology underlying thedictionary involves the extraction from a corpus of contemporary Portuguese of lexical associations of pairs ofword forms, contiguous or not. The significance of the pairs is statistically measured by the Mutual Information(MI) calculus, as well as by the MI weighted by the frequency of the pair (MIF). Other issues are discussed:frequency of the word forms vs. frequency of the lemmas, the organization of the collocations in the dictionary,grammatical patterns as source of lexical information, as well as the splitting of collocations into sense-groups.