Comment: Lakoff on Metaphor – More Heat Than Light

As a linguist, what struck me most was Lakoff’s (2016) exclusive focus on English. There are sweeping (and questionable) claims about all or most languages, for example, “this is called Aspect, and the same structure appears in every language in the world”—yet not a single mention of any language other than English. Overall the impression is that one can access psychological or conceptual reality of the human species solely on the basis of contemporary English. Like most linguists, I believe we need to engage with and learn from language diversity. The other great lacuna in Lakoff’s text is a complete silence about cultural (and historical) influences on language and emotion. To rebalance the picture, key resources from linguists would be Wierzbicka (1999) and Goddard and Wierzbicka (2014, Chapters 5–6), and the studies of more than two dozen languages included in the edited collections Athanasiadou and Tabakowska (1998), Harkins and Wierzbicka (2001), and Goddard and Ye (2014). One can add the work of cross-linguistically aware psychologists, historians, and anthropologists (e.g., Dixon, 2006; Fontaine, Scherer, & Soriano, 2013; Russell, 1991; Shweder, 2004).

Coming to Lakoff’s (2016) main themes, I first want to make some comments about so-called “primary metaphors” such as “affection is warmth” and “anger is heat.” It would be foolish to deny any association between embodied experience and concepts like “affection” and “anger,” but as critics have often pointed out, the supposed mappings from bodily experience to concepts are not one-to-one, but many-to-many. The picture is complicated and open to multiple interpretations, and, as far as I know, there is no well-documented or widely agreed itemisation of primary metaphors across the world’s languages. Furthermore, it is questionable whether complex concepts (or complex metaphors) can be derived from combinations of primary metaphors in a systematic or verifiable fashion (Lakoff’s explanation for the supposed “love is a journey” metaphor, for example, stretches credibility).

Second, when looked at in detail, many metaphorical ways of speaking actually work differently in different languages. According to Wierzbicka (2009), the meaning of Russian gorjačij “hot” in expressions like “want hotly” and “speak hotly” is associated not with anger, but rather with “wanting something good” and “wanting it very much.” The high frequency of such collocations reflects a central cultural value of Russian. Goddard and Wierzbicka (2014, p. 78) develop similar arguments about the differing figurative uses of “hard” and “heavy” in English and Polish. In short, there are “cultural underpinnings of language-specific metaphorical transfers” (cf. Geeraerts, 2006).

Third, regarding Lakoff’s (2016) neural version of “embodiment theory,” few would deny that how we think and feel is profoundly influenced by our status as embodied beings. The idea that experiential schemas influence language and thinking is an important theme in cognitive linguistics. Equally, it is obvious that findings from neuroscience are important to a full understanding of human cognition and experience. None of this makes it valid, or even coherent, to recast the description of everything symbolic or “contentful” in language with talk about brain structures and neural processes (cf. Rohrer, 2006). To my mind, it is literally nonsensical to say that a metaphor is a neural circuit, or that a thought is a neural process. Not only that, such biophysical reductionism leaves consciousness and first-person experience out of the picture altogether (cf. Zlatev, 2007).

In reality, neural development is strongly influenced by interactional and sociocultural environments. What we need, then, are theory formulations that adopt a “plus-plus” approach. “Either-or” thinking is simply not helpful.

Finally, I want to state my disagreement with Lakoff’s (2016) implication (clearest in his final paragraph) that word meanings are unbelievably profuse, metaphorical, and polysemous. Polysemy is indeed a linguistic fact of life, but, contra Brugman (1988), no word has over 100 meanings. Numerous Natural Semantic Metalanguage (NSM) studies show that rigorous methods enable a surprising level of precision and predictiveness to be achieved in semantic analysis, and in particular that it is possible to pinpoint significant meaning differences in emotion expression in the languages of the world (see Goddard & Wierzbicka, 2014, for extensive references). This is key data for the affective sciences.