A Cognitive Semantics Approach to Darwin’s Theory of Evolution

21 1. Theoretical Background: Cognitive Semantics The quest for increased reliability of metaphor identification, speed and scope of data analysis, as well as the reduction of the role of intuition and retrospection is most vividly manifested in attempts to design computational programs that can screen large corpora for metaphorical expressions (e.g. the CorMet metaphor extraction engine described in Mason 2004). Although developments of research methods within CMT along these lines are both interesting and desirable, a word in defence of researcher’s intuition as part of the CMT methodology needs to be given. When skepticism against CMT is voiced, intuition is devalued, as if it were a researcher’s purely subjective opinion, or even hopeful guessing. However, although CMT scholars may differ in metaphor identification or interpretation, their judgement always springs from experience, knowledge, and a careful analysis of a discourse sample, often in its broad context: textual, social, cultural, etc. Computerized algorithms and formal protocols can accompany subtle powers of intuition, enhancing them and testing for accuracy, but should not replace them. Heavy reliance on computer search engines and metaphor identification protocols that mainly operate on the word level may, in the long run, lead to putting emphasis on metaphor in language at the expense of identifying metaphor in thought, thereby drifting away from the most valuable postulates of Cognitive Semantics. The next charge discussed by Gibbs (2009) concerns the treatment of conventional expressions as metaphorical by practitioners of Cognitive Semantics. Perhaps such criticism is best represented in the paper by Keysar et al. (2000). With the support of some experimental data, the authors reject Lakoff and Johnson’s claim for the metaphorical status of everyday expressions and argue that “conventional expressions can be understood directly, without recourse to underlying conceptual mappings” (Keysar et al. 2000: 578). They also allow for only a limited use of metaphorical mappings in understanding of novel expressions. We can partially agree with the general idea that large portions of language, though historically motivated by metaphorical mappings, when used on repetitive, everyday basis to communicate about everyday topics, do not require the activation of cross-domain mappings and can be processed as automated routines. We can also agree that many words are polysemous and their use does not necessarily involve systematic metaphorical mappings. However, there seems to be a flaw in the interpretation of the results of the experiments, which are intended to provide empirical evidence against conceptual metaphor. All the experiments the authors report are focused on the understanding of metaphorical expressions,