Chapter 4. Locating metaphor candidates in specialized corpora using raw frequency and keyword lists

This chapter explains one method that can be used to extract linguistic metaphors from a specialized corpus of Italian political speeches, using statistically-based measures incorporated into most standard corpus query software – in this case, WordSmith Tools (Scott 2004). This method can be used alone or in combination with existing manual or semi-manual analyses. While software has been developed for the automatic extraction of metaphors in English, minority languages, including Italian, lack tools for semantic annotation and probability measures that underlie such applications. The method presented in this chapter is intended for users who have no access to lemmatizers, semantic taggers, etc., and/or are working with under-resourced languages, for which no such tools are generally available.