There are a few topics in cognitive science that are like fingernails on a chalkboard to me. I find the very mention of them irritating, and the irritation can stick in my craw for days. At or near the top of that list are mirror neurons. These little cells have been made so sexy, either by their sheer coolness or massive, overblown publicity (it’s hard to tell which), that they seem to have become the solution to virtually every problem in the study of cognition and behavior in the minds of some researchers. Of course, the blogosphere has also jumped on the mirror neuron boat (I was particularly impressed with the one on mirror neurons and prayer).
What are mirror neurons? Well, they’re neurons that were first discovered by Gallese et al.1 in the premotor cortex of macaques (the red and yellow areas in figure A below2). They respond both to the movements that a monkey makes and when the monkey observes another monkey making the same movements. Hence the name. There is evidence, all be it more indirect than in monkeys, that humans have a mirror neuron system too. In us, it’s thought to be located around the right superior temporal sulcus (red and yellow areas in figure B below).
What functions, exactly, the mirror neuron system subserves in humans is not really known. Ignorance, however, is no barrier to speculation. Researchers have implicated mirror neurons in everything from autism to marketing. Given how little we know about the mirror neuron system in humans, the hype has been unbelievable. Consider some quotes (stolen from Hurford, 20043, emphasis mine):
USC’s Michael A. Arbib, Ph.D., says ”the neurons, located in the premotor cortex just in front of the motor cortex, are a mechanism for recognizing the meaning of actions made by others… For communication to succeed, both the individual sending a message and the individual receiving it must recognize the significance of the sender’s signal. Mirror neurons are thus the missing link in the evolution of language. They provide a mechanism for the sharing of meaning.” (ScienceDaily, 8/20/1998, repeated in USC Trojan Family alumni magazine, Spring 1999.)
Rizzolatti and Arbib think that mirror neurons may have provided the bridge from “doing” to “communicating”. The relationship between actor and observer may have developed into one involving the sending and receiving of a message. In all communication the sender and receiver have to have a common understanding about what’s passing between them. Could mirror neurons explain how this is achieved? Rizzolatti and Arbib think the answer is yes. (New Scientist, vol 169 issue 2275, 27/01/2001, p22)
“I predict that mirror neurons will do for psychology what DNA did for biology,” he [V.S.Ramachandran] says. “They will provide a unifying framework and help explain a host of mental abilities that have hitherto remained mysterious.” (New Scientist, vol 169 issue 2275, 27/01/2001, p22)
Are you kidding me? Mirror neurons are psychology’s DNA? OK, let me take a deep breath for a moment. I’m feeling better. Let’s move on. Not all of the proposed uses of mirror neurons are completely out in left field. Some (e.g., Gallese & Goldman, 19984) have suggested that it might be involved in mentalizing, or theory of mind, which doesn’t seem like too big of a stretch, particularly if you adopt a simulation theory of mentalizing. Some of the speculation, however, is just plain nuts. Since Seed Magazine has seen fit to publish a short article by Irene Pepperberg on one of those areas of speculation, I thought I’d say a little about it.
As Arbib and Rizzolatti suggest in the above quotes, some researchers (well, Arbib and Rizzolatti) have been suggesting for a few years now that mirror neurons may help to solve one of the toughest problems in cognitive science: how language evolved. The hypothesis works something like this (in Rizzolatti and Arbib, 1998, see footnote 2): spoken language developed from gestural communication (not a novel hypothesis, or a particularly batty one), and gestural communication was, in turn, made possible by the evolution of the mirror neuron system. They start by looking at “case grammar,” which they describe thusly:
Case grammar organizes sentences around action frames with slots for different roles. The key paper for case structure is ‘The Case for Case’ by Fillmore, although many of the ideas about case have now been absorbed in the thematic structure of the lexicon that is an integral component of the Chomskian approach to syntax known as ‘government and binding theory’. In a case analysis, the sentence ‘John hit Mary with his hand’ is viewed as the ‘surface structure’ for a case structure ‘hit’ (John, Mary, John’s hand), which is an instance of the case frame ‘hit’ (agent, recipient, instrument), which makes explicit the roles of ‘John’, ‘Mary’ and ‘John’s hand’. Clearly, many different sentences can express the underlying case structure.
They then argue that in monkey brains, neurons in F5, including mirror neurons, organize actions (e.g., John hitting Mary with his hand) into frames that include the action (x hit y with z) and the specifics of the action (it was John who hit Mary, and he hit her with his hand), by representing the actions in the same cells that would represent hitting someone with your own hand. Now, how these neurons represent John or Mary is a mystery, but we’ll get to that in a moment. The important thing to recognize, for now, is that going from representing the meaning of non-communicative actions to representing communicative gestures is pretty simple: you attach a meaning to a gesture by representing it in the mirror neuron system in the same way that you represent the meanings of non-communicative actions.
Do you see the big leaps that Rizzolatti and Arbib have taken? They’ve gone from representing actions to representing actions to representing their “meaning,” and from there to communicative gestures. The next big leap is to move to speech. At first, this leap seems like the least problematic. There are theories of speech representation that are based in the motor system5. In these theories, speech patterns are represented as motor programs, rather than as sounds. Furthermore, these motor programs are thought to be represented in or around the areas that may house mirror neurons in humans. So it’s not too much of a stretch to argue that we represent, and perhaps even learn these motor programs by simulating observed speech in our mirror neuron system. But Rizzolatti and Arbib don’t stop there. They argue that this representation is also given meaning by the mirror neuron system, which somehow (god knows how – Arbib mentions affordances, but gives us no real theory of how mirror neurons represent affordances6) connects the speech patterns to concepts (like John and Mary in their example above). Thus, the evolution of mirror neurons is what made spoken language possible.
I don’t know about you, but at this point, my legs are tired from all of that leaping. So I’m going to wrap this up now. The problems with this theory are twofold:
- Almost all of what we know about mirror neurons comes from studies with monkeys. Sure, there’s evidence that we humans have neurons, but no real evidence about what they do in humans. Furthermore, we’re not completely sure that mirror neurons are involved in communication in monkeys, except, perhaps, in learning/imitating sounds and gestures. There’s no evidence that mirror neurons are involved in the processing of meaning in monkeys. So, it’s a huge leap to say that mirror neurons are not only involved in the motor representation of speech, but that they also give those representations meanings (through affordances, e.g.).
- Even if mirror neurons are associated with affordances, it’s a big stretch to say that they give speech patterns meaning. Speech patterns are arbitrary, and even if we adopt a theory in which affordances (which are just actions that particular objects in the world allow us to perform, by the way) are how we represent meaning (i.e., that affordances are concepts), Rizzolatti and Arbib give us no theory of how speech patterns represented in mirror neurons get connected to other motor patterns represented in mirror neurons and elsewhere in the brain.
In short, they’ve given us nothing but a bunch of wild speculation that seems to be based more on the belief that mirror neurons are really cool than on any empirical evidence about what mirror neurons actually do. You can’t solve long-standing problems like connecting arbitrary speech patterns to meanings, and the evolution of language, just by saying, “Look, mirror neurons!” But for some reason, taking just that approach has garnered Rizzolatti and Arbib, and mirror neurons themselves, a great deal of attention.
1Gallese, V., Fadiga, L., Fogassi, L., & Rizzolatti, G. (1996). Action recognition in the premotor cortex. Brain, 119(2), 593-609.
2Rizzolatti, G., & Arbib, M.A. (1998). Language within our grasp. Trends in Neurosciences, 21(5), 188-194.
3Hurford, J.R. (2004). Language beyond our grasp: what mirror neurons can, and cannot, do for language evolution. In D.K. Oller & U. Griebel (eds.), Evolution of Communication Systems: A Comparative Approach, pp.297-313. MIT Press: Cambridge, MA.
4Gallese, V., & Goldman, A. (1998). Mirror neurons and the simulation theory of mind-reading. Trends in Cognitive Sciences, 2(12), 493-501.
5E.g., Vihman, M.M. (1993). Variable paths to early word production. Journal of Phonetics, 21, 61-82.
6Arbib, M.A., (2002). The mirror system, imitation, and the evolution of language. In C. Nehaniv & K. Dautenhahn (Eds.), Imitation in Animals and Artifacts, MIT Press: Cambridge, MA.