A PROGRAM FOR FINDING METAPHOR CANDIDATES IN CORPORA

Authors

  • Tony Berber Sardinha Pontifical Catholic University of São Paulo – PUC-SP, São Paulo, Brazil

Keywords:

conceptual metaphor, linguistic metaphor, Corpus Linguistics, metaphor identification

Abstract

In this paper, I present a computer program (‘Identifi cador de Metáforas’) for finding metaphor candidates (i.e., words that are likely to have been used metaphorically) in corpora. It works by matching each word in the corpus to five databases that contain several kinds of information about lexis and its relationship to metaphor, all extracted from extensive hand-annotated corpora. These databases store the probability of words being metaphorical based on their previous use in metaphors, on lexical patterns occurring near and around words, and on their word class. This article explains how the program was created and how it works.

Author Biography

Tony Berber Sardinha, Pontifical Catholic University of São Paulo – PUC-SP, São Paulo, Brazil

Tony Berber Sardinha is an Associate Professor of Applied Linguistics with the Linguistics Department and the Applied Linguistics Graduate Program, Sao Paulo Catholic University (PUCSP), Brazil. He sits on the Executive Committee of RaAM, Researching and Applying Metaphor, and of ALSFAL, the Latin American Systemic Functional Linguistics Association. He runs a number of websites including the CEPRIL portal for online corpus analysis tools that hosts applications for automatic analysis of a range of linguistic features. He heads several projects, among which are the Brazilian Corpus, a one-billion word online resource, and Br-ICLE, the Brazilian subcorpus of the International Corpus of Learner English.

Downloads

Issue

Vol. 31 No. 1 (2010)

Section

Papers

License

The authors grant the journal all copyrights relating to the published works. The concepts issued in signed articles are the absolute and exclusive responsibility of their authors.