Posted by 
In case you haven’t heard about it, there’s a relatively new blog in the cognitive science section of the blog world called Cognitive Approaches to Literature. They don’t post very often over there, but if they ever start doing so, it promises to be an interesting read with lively discussions. The latest addition to the blog’s stable of contributors is Michael Kimmel, a “cognitive social scientist” in Vienna whose main interests are “metaphor, imagery and embodiment.” Now, if you speak the language of cognitive science, you can probably guess a few of Kimmel’s major inspirations from the three words, “metaphor, imagery and embodiment,” and if you’ve guessed them and you’ve hung around Mixing Memory long, you can probably guess what I’m about to start ranting on, too. But I was intrigued by the title “cognitive social scientist,” which is new to me (and Google), so I stopped over at Kimmel’s homepage (linked above) to see if I could find a link to a paper I could read just to get a feel for his work.
Unfortunately, there are no links to his papers there, but when I clicked on Resources and Links, and then “annotated list of my personal digest,” I got the information I was looking for. There’s some really good stuff on the list: Hutchins’ Cognition in the Wild, a book on my list of favorites in cog sci that I don’t know what to do with; Francisco Gil-White’s fascinating paper, “Are Ethnic Groups Biological “Species” to the Human Brain?;” Gilles Fauconnier’s as-of-yet unpaid promissory note, Mappings in Thought and Language; the classic, Metaphor and Thought, edited by Andrew Ortony; and Ronald Langacker’s Concept, Image, and Symbol (I’d have chosen Foundations of Cognitive Grammar, but that’s just me), to name a few. But then there are the books that you’re expecting, if you speak the language.
I have to admit it upsets me less now than it has in the past that people in departments other than, say, psychology, neuroscience, linguistics, computer science (AI), philosophy, and cognitive anthropology (not really a department, I know) who are interested in cognitive science are so often drawn to conceptual metaphor theory. But when I read this description of Lakoff and Johnson’s magnum opus, Philosophy in the Flesh, on Kimmel’s page, I felt some of the old frustration coming back:
A startling, but lengthy survey of near to 600 pages of recent research on conceptual metaphor. Most interestingly, it also deals with issues of methodology in cognitive research and introduces the notion of “convergent evidence” from varied methods as a core pillar. In the annex fascinating evidence is produced from a new Berkeley based interdisciplinary workgroup called the “Neural Theory of Language” that encompasses artificial intelligence, psychology, and linguistics. Some first, tentative steps are made in explaining a multi-level explanatory framework that connects the brain- sciences and the mind-science.
Let me quickly say that I don’t blame Kimmel for this description. It’s a description Lakoff himself might have written, as it’s based on the way in which Lakoff and Johnson describe not only their project, but their perceived originality and success. If you haven’t read any mainstream cognitive science, you might think that cognitive linguistics is one of the most successful paradigms in the field (they tell you so, in the new introduction to Metaphors We Live By). And of course, certain parts of Kimmel’s description are true. The book is lengthy, and it is about recent “research” on conceptual metaphor.
But the description is mostly misleading (in the ways that Lakoff and Johnson mislead their readers). You see, Lakoff and Johnson haven’t “introduced” the “notion of converging evidence as a core pillar.” That’s science! Sure, the academic world is such that people, even entire research areas, can get overly enamored with a single research paradigm. Hell, cognitive linguists have written at least a couple dozen books based almost entirely on analyzing metaphors in everyday speech. But “converging evidence” is a core pillar of all cognitive science, and has been since well before “cognitive linguistics” existed.
Then there’s this pair of sentences: “In the annex fascinating evidence is produced from a new Berkeley based interdisciplinary workgroup called the “Neural Theory of Language” that encompasses artificial intelligence, psychology, and linguistics. Some first, tentative steps are made in explaining a multi-level explanatory framework that connects the brain- sciences and the mind-science.” So cognitive linguistics is producing an innovative research program that “encompasses artificial intelligence, psychology, and linguistics,” sort of like the rest of cognitive science has been doing since, oh, the 1950s? A fact that belies the claim that within cognitive linguistics, the first steps, tentative or otherwise, are being taken “in explaining a multi-level explainatory framework that connects brain-sciences and mind-sciences.” Because, you know, that’s what cognitive scientists have been doing all along.
Again, I don’t blame Kimmel for this nonsense. It’s Lakoff and Johnson’s nonsense. They really do seem to believe that they’re doing something truly unique, as though embodiment wasn’t a central tenet of much of 20th century philosophy (at least on the continent — think of Bergson, Heidegger, or Merleau-Ponty, to give a few examples* [UPDATE: Commenter Cosma asks that we not forget James and Dewey.]), or thinking about thinking, and how the brain and language relate to thinking, is something they invented whole cloth in 1980 while everyone else in cognitive science was studying the ether.
Perhaps worst of all, they seem to think that they’ve been successful. As I’ve mentioned here before, outside of Lera Boroditsky’s work on temporal metaphors (some of which is now being seriously challenged, as researchers are finding it difficult to replicate her findings), there really is no experimental evidence for conceptual metaphor theory, or any of the other major theories in cognitive linguistics (as I said before, Fauconnier’s blending theory is just a promissory note, to date, despite what he and Mark Turner are now claiming, thought they at least have the guts to admit that they can’t really make predictions). Even more damning, there’s plenty of falsifying evidence (e.g., in this paper), and there are strong theoretical arguments against conceptual metaphor theory as well. Aside from Raymond Gibbs Jr. (in this paper, e.g., but see Murphy’s reply), few cognitive linguists have publicly addressed such evidence and objections. I’m not sure I’ve ever seen Lakoff even mention it. One gets the distinct impression that he and some of the other major figures in the field (again, with the notable exception of Gibbs) tend to just ignore recalcitrant data.
I’ve tossed around a dozen theories as to why Lakoff talks about conceptual metaphor theory as though it were so well confirmed as to be deserving of the same status as theories like relativity or evolution, and I’ve settle on two possibilities. The first is that he’s so attached to it that he’s incapable of recognizing a.) that mainstream cognitive science generally ignores it, particularly in the study of concepts (with the exception of the recent perceptual symbol fad), where it should be the most relevant, and b.) that after almost 30 years, you’d expect a great deal of experimental evidence in favor of the theory, and little evidence against it, while in both cases the reality is the opposite. The second is that he fully recognizes both a. and b., and like so many scientists whose pet theories have failed within their fields, has taken to selling it to people outside of that field (like politicians, English professors, etc.) rather than letting it die a much deserved death. And because conceptual metaphor is such an intuitively appealing theory — we think in metaphors! wow, that sounds cool — the ideas have sold well (literally and, dare I say it, metaphorically).
It’s because I’ve mostly accepted the second explanation that I’m no longer deeply troubled by Lakoff’s success outside of mainstream cognitive science, and especially outside of academia altogether. Lakoff’s turned into a salesman, and he’s a very, very good one. So people who haven’t been following the huge and often confusing literature on categories concepts that’s arisen in the wake of Elanor Rosch’s ground-breaking work in the early 70s (this page has a nice selection, and notice no Lakoff), or the productive though still inconclusive research on metaphor by people like Dedre Gentner or Sam Glucksberg, might not be aware that he’s selling them snake oil. I just wish there was some sort of filter that would allow me to avoid ever having to read about it again.
*While we’re at it, why don’t we note that even the idea that concepts are metaphors isn’t really a new either. There’s a wonderful little essay written by some 29-year old German guy in 1873 that takes a similar position.
June 17, 2007
June 18, 2007