Towards a classification of metaphor use in text: Issues in conceptual discourse analysis of a domain-specific corpus
Towards a classification of metaphor use in text: Issues in conceptual discourse analysis of a domain-specific corpus. In: Third Workshop on Corpus-Based Approaches to Figurative Language. Birmingham; 2005.
Investigation of metaphor in running discourse has become popular, particularly among European scholars (Charteris-Black, 2004; Musolff, 2004; Koller, 2005; and others). This stands in contrast to an essentially encyclopedic approach traditionally adopted by conceptual metaphor analysts (Lakoff and Johnson, 1980; Lakoff, 1996/2001; Dirven, 1994; Kövecses, 2005, and others). While the discourse metaphor analysts are in a position to assess the presence of metaphorical expressions in a large and selectively representative body of textual data, they face the nearly impossible challenge of identifying not only metaphorical expressions but also metaphoric conceptual representations in such a corpus. As a result, they largely confine their analyses to the identification of metaphor keywords (as above, also Izwaini, 2003) and their distribution. This, however, presents problems for drawing conclusions about the role keyword-based metaphors play in the conceptual underpinning of discourse. In particular, analysts have often taken the presence of metaphor keywords to be evidence for an underlying conceptual structure, such as the use of “faith” in political discourse for politics is religion with no other evidence of inherited inferential structure such as politicians are priests or laws are biblical laws (Charteris-Black, 2004).
This paper proposes a classification for metaphor use in text that can avoid such unwarranted conclusions and exemplifies it on a partial analysis of a corpus of educational discourse. It seeks to do so by enriching the conceptual metaphor theory with insights from Hallidayan functional linguistics as well as taking into account and integrating conflicting psycholinguistic research results (e.g. Gibbs, 1994; Glucksberg, 2001). At the heart of this approach lies the recognition of the fact that identical keywords can represent different uses of metaphor, and thus the aim is not to classify metaphor in an essentialist manner but rather to provide a heuristic device for the interpretation of data obtained by other means, viz the evaluation of the conceptual import of metaphorical expressions occurring in running discourse. For this, Halliday’s distinction between the ideational (cognitive), interpersonal (social) and textual functions of linguistic units provides an ideal framework. Both the interpersonal and textual functions have been long assumed by cognitive linguists to be a part of their analyses (Langacker, 2005) but so far they have received very little explicit attention. The textual function has been highlighted by some scholars (e.g. Kittay, 1987 or Goatly, 1997) but has received little systematic research attention except in spoken language analysis (e.g. Holt and Drew, 2001). Other classifications, most notably that of Steen et al. (e.g. Steen, 1999; Semino and Steen, 2001; Crisp et al. 2002) have focused on the nature of metaphor-bearing expressions themselves to the neglect of the exclusion of their cognitive and interpersonal functions, the latter of which has been stressed, for instance, by Schön (1979, 1983) in his theory of generative metaphor.
The proposed classification makes it possible to incorporate these seemingly incompatible attempts at metaphor classification. For instance, in the sentence “Schools must decide what kind of business they want to be” it is possible to identify the cognitive function of metaphor in use as structural (following Lakoff and Johnson, 1980 and Kövecses, 2002) and at the same time recognize its generative interpersonal and anaphoric textual functions. On the other hand, a phrase such “I propose the introduction of a ‘pedocratic oath’” has the same cognitive and interpersonal function while being textually exophoric. It is only through the recognition of these and other different uses that metaphors are put to in a text, that we can fully assess the metaphorical aspects of the conceptual structure of discourse and thus accentuate the advantages of corpus-based research of figurative language.