Yet another application of inference in computational linguistics

Texts in natural languages consist of words that are syntactically linked and semantically combinable-like political party, pay attention, or brick wall. Such semantically plausible combinations of two content words, which we hereafter refer to as collocations, are important knowledge in many areas of computational linguistics. We consider a lexical resource that provides such knowledge-a collocation database (CBD). Since such databases cannot be complete under any reasonable compilation procedure, we consider heuristic-based inference mechanisms that predict new plausible collocations based on the ones present in the CDB, with the help of a WordNet-like thesaurus. If an available collocation combines the entries A and B, and B is 'similar' to C, then A and C supposedly constitute a collocation of the same category. Also, we touch upon semantically induced morphological categories suiting for such inferences. Several heuristics for filtering out wrong hypotheses are also given and the experience in inferences obtained with CrossLexica CDB is briefly discussed.