Metaphor and Corpus Linguistics

Growing interest in metaphor over recent decades has led to an increasing need for guidance on how to identify and analyze this linguistic phenomenon in corpus data. While Cognitive Linguistics offers various perspectives on interpreting metaphorical meanings, it lacks the means to document how, where, and with what frequency these meanings are realised textually. Corpus Linguistics can address these issues, but only once one crucial problem has been addressed: how to find metaphorical meanings in data that is typically searched via word forms. This contribution starts with an overview of conceptual and linguistic metaphor, and how metaphors are identified in language. It then dedicates space to issues of specific interest to corpus-based studies of metaphor, particularly regarding how to find metaphors in corpus data. The first method described is the lexicographical approach, which starts with a word form and profiles its lexical environments in order to separate out distinct meanings, metaphorical or otherwise, and to schematize the regular lexicogrammatical patternings associated with each. The other methods focus on how to identify/extract candidate metaphors from corpus data. After identification/extraction, the data has to be examined to verify that each instantiation of a word is indeed metaphorical, following the same principles as in the lexicographical approach. Corpus studies of metaphor can enhance our understanding of what metaphor achieves in text by analyzing the lexicogrammatical patterns, the frequencies, and the distribution of metaphor in natural language data.