Metaphor Annotation

Abstract

The goal of studies in corpus linguistics is often to characterize metaphors used in particular corpora of interest, such as metaphors used by a certain political figure, metaphors used in public discourse about a certain topic in a certain period of time, or metaphors in texts in a certain professional field [Cameron, 2003, Charteris-Black, 2005, Koller, 2004, Musolff, 2000, Partington, 1995]. The operating definition of what constitutes a metaphor, as well as the granularity of analysis (e.g., word or phrase level) may differ across studies. This might not be a problem for any given study, as studies usually operate with aggregate figures, such as the totals or proportions of certain kinds of metaphors in the given corpus; thus, even if some of the specific instances of metaphors identified in the corpus are somewhat subjective, significant differences between corpora are still likely to hold up (given the particular definition of metaphor that is used in the study). Corpus-linguistic studies often address the interpretive and strategic aspects of metaphor use, and might therefore focus on certain culturally, politically, or pedagogically interesting metaphors, leaving open the possibility that other, less salient, metaphors in the same texts might be ignored. However, if one seeks an exhaustive characterization of metaphor in a given corpus, or a comparative characterization across different studies, then the lack of methodological consistency in identifying metaphors becomes problematic: “Variability in intuitions, and lack of precision about what counts as a metaphor, makes it quite difficult to compare different empirical analyses” [Pragglejaz Group, 2007, p. 2]. Thus, out of the multiplicity of approaches and techniques, a recognition has grown that the field needs “principled solutions if an annotated corpus of metaphors in texts is to be produced for research purposes” [Heywood et al., 2002, p. 51].