We consider the problem of extracting specified types of information from natural language text. To properly analyze the text, we wish to apply semantic (selectional) constraints whenever possible; however, we cannot expect to have semantic patterns for all the input we may encounter in real texts. We therefore use preference semantics: selecting the analysis which maximizes the number of semantic patterns matched. We describe a specific information extraction task, and report on the benefits of using preference semantics for this task. Task and Approach Information extraction is the task of extracting specified types of intonnation from a natural language text -- for example, information about specific classes of events. Typically, however, the text to he processed will contain many types of events besides the classes of interest. The system designer therefore faces a quandary in imposing semantic (selectional) constraints. Selectional constraints could be strictly enforced: a senten...