Beyond Bourdieu’s ‘body’ — giving too much credit?

Blogging on Peer-Reviewed ResearchA while back, Daniel recommended to me that I check out an article by Mauro Adenzato and Francesca Garbarini in Theory & Psychology. It’s a great piece, and there’s a lot of positive things I could say about it. For example, Adenzato and Garbarini’s principal point is that the model of ‘mind’ as the workings of organic brain systems is inconsistent with much social theory, built instead on a treatment of mind as a kind of dis-embodied amalgamation of logical and cognitive processes. This recognition that phenomenological and cognitive mind and biological brain are inextricably linked is part of what makes neuroanthropology possible, or rather, necessary. As they write:

The embodied cognition perspective views the mind no longer as a set of logical/abstract functions, but as a biological system, which is rooted in body experience and interwoven with action and interaction with other individuals. Action and representation are no longer interpreted in terms of the classic physical–mental dichotomy, but have proven to be closely interlinked. Specifically, embodied cognition means that acting in the world, interacting with the objects and individuals in it, representing the world, perceiving it, categorizing it and understanding its meaning are merely different levels of the same relationship that exists between an organism and its environment. (Adenzato and Garbarini 2006:748) 

Adenzato and Garbarini point out that cognitive science has become increasingly concerned with grounding theories about thought and the mind in observable workings of the organic brain. Although I agree, following Clark (1997), I would refer to this movement as ‘third generation,’ rather than ‘second generation’ (as the authors do), as the first two over-arching waves in cognitive science might be termed the ‘logic machine’ and ‘connectionist’ models.I have some niggling problems with their argument, such as the ‘merely different levels’ off-handed comment, which I would dispute, and the tired technique later in the article of dragging a folk term from an ethnographic case study and showing how it’s like one’s new theoretical term, as if the natives having a word for a phenomenon is proof you’re on to something. Nevertheless, this passage is a very coherent statement, as is the whole article, of the need to incorporate more sophisticated models of the brain’s working into cultural theory.

The pair discuss canonical and mirror neurons, two sets of sensory-motor neurons (I discussed mirror neurons in an earlier post and in a couple of articles that are in press or review). The existence of these neurons lends strong support to ‘simulation’ theories of motor perception, which assert that we experience ‘as if’ we were ourselves doing actions when we observe others acting, and to ‘affordance’-based theories of object perception, associated with ecological psychologist James Gibson. According to Gibson, we directly perceive ‘affordances’ in our environment; that is, we perceive opportunities for action based upon our abilities, interests, and potential for interaction with the environment (see Rizzolatti and Craighero 2004 for a review).

All of this is well and good, but, in my opinion, things start to go astray when the authors credit the late Pierre Bourdieu with proposing a theory of culture consistent with the findings of researchers like Rizzolatti, and then resort to a ‘representational’ theory of mind. According to ‘representational’ theories of mind, thought works primarily be generating ‘representations’ of the sense world. There are enormous slippages between the ways that neuroscientists use the term ‘representation’ and the way that anthropologists use the same term. The resulting slippage is typically disastrous for cultural theory. Anthropologists typically mean a surrogate symbol or sign when they say ‘representation’; neuroscientists mean the neural-physiological dimension to any mental phenomenon, such as a perception, thought, or experience.

To me, a simulation theory of motor perception – what mirror and canonical neurons suggest – is radically different from a representational theory. Rather than implying that we perceive movement by creating a mental ‘representation’ (in the anthropological sense), mirror and canonical neurons strongly support the idea that a neural ‘representation’ (in the neuroscientific sense) is instead very similar to actually doing the same action. In other words, perception of movement is mediated not by a symbol of movement, but rather by, well, an experience of oneself moving, even if that is an inchoate simulation. The cognitive and phenomenological distance implied by the symbolic notion of representation, its cognitive-ness, is instead kinetic, visceral, and directly sensed.

But my problems with arguing that mirror and canonical neurons support Bourdieu-ian-style post-structuralism (or ‘crypto-structuralism,’ if one is critical of him) also extends to other ways in which his theories are given an overly generous reading. When it comes to giving too much credit to Bourdieu, one of the most interesting passages is this:

With its constant reference to the ideas of ‘body’ and ‘embodied’, Bourdieu’s practice theory shares a common background with the paradigm of embodied cognition, that is, the analysis of body concept conducted by Merleau-Ponty in his Phenomenology of Perception (1945/1962). Indeed, Merleau-Ponty intended ‘body’ as both a physical structure (the biological body) and an experiential structure (the living, moving, suffering and exulting body). (Adenzato and Garbarini 2006:750) 

Equating Bourdieu’s theory of the habitus with Merleau-Ponty’s rich discussion of embodiment, including his references to the most sophisticated contemporary analyses of severe brain injury and psychological disorders, does an extreme disservice to both, muddying a distinction that Bourdieu was at pains to make clear.

Firstly, Bourdieu explicitly rejected – repeatedly – phenomenology, although he did so in a way that I have always found less than generous: by equating all of phenomenology to a stripped-down version of ethnomethodology. Although he clearly owed much to Merleau-Ponty (which is probably true of most French thinkers of his generation, even those who rejected him), he certainly never made the positive contributions to his own thinking clear. Bourdieu seemed obdurately to misrepresent phenomenology and repeatedly beat on a straw version of the perspective, which is actually much more attuned to experience, including embodiment, than his own work.

Secondly, mentioning ‘the body’ does not necessarily mean that one is overcoming Cartesian dualism. Bourdieu’s use of the term ‘body’ included no references to biology, except in critical terms to reject it or in glib terms (like Pinker) positing vague organs or tissues or physiological processes that presumably went with the forms of ‘embodiment’ that he wanted to discuss (I’ll return to this in a moment). The habitus gets discussed with much reference to ‘the body’ and no clear explanation of any physiology or plausible discussion of how habits might affect perception or subjectivity. In fact, I think habits, skills, and bodily conditioning do affect the subject to an enormous degree, but my instinct is not a full-blown theory until I do the hard yards involved in articulating a theory, such as describing some causal mechanisms for habit to have these sorts of effects. Bourdieu does not do this at all; in contrast, Merleau-Ponty grappled much more with psychological levels of influence on individuals’ perceptions.

Third, Bourdieu’s ‘phenomenology’ is better than nothing, but it’s not great. That is, he does a decent job of talking about experience in ways that avoid some of the more obvious and egregious problems of other scholars – such as treating social action as ‘rational choice’ – but it’s far from a rich discussion of emotion, motivation, experience, memory, and other processes that make up experience. He uses some pretty persuasive metaphors (such as ‘feel for the game’), and these are much to be admired, but the reason that he kept having to fend off charges that his was a ‘rational actor theory (RAT)’ (such as in the volume he did with Wacquant, Invitation to Reflexive Sociology) is, at least in part, because his phenomenology, as well as his psychology, always tilted too much in that direction (‘strategies,’ ‘games,’ and the like).

One of the severe problems with Bourdieu, even though he’s a step in the direction of an integrated neuroanthropology (note, Bourdieu, to my knowledge, never inquired into the brain sciences in spite of his discussions of the body, habits, and incorporation), is what he thinks the body might hold. Adenzato and Garbarini give him credit for shifting the grounds of constructivism from ‘theoretical’ to ‘pragmatic’:

We argue that Bourdieu’s theory represents, in and of itself, a sort of constructivism, which is non-’theoretical’, but ‘pragmatic’. In fact, although practice theory holds that categories ‘construct’ the world in a Kantian way, it does not refer to theoretical categories (which are determined a priori independently of experience) but to pragmatic categories derived from mutual co-determination between organism and environment: our ability to apply categories to phenomena is made possible by the fact that these categories have evolved under the selective constraints of the environment itself (intended as the natural environment and social environment, or, better yet, the natural and social environments together). (Adenzato and Garbarini 2006:751) 

But here’s the problem; they’re still ‘categories.’ That is, even though we are constructed by pragmatic action, according to Bourdieu, what is constructed in us is ‘categories,’ just like with more abstract systems such as language or cognition, but they’re ‘pragmatic’ categories. The body, conveniently, works pretty much just like the brain, establishing the same sorts of entities, much like the ‘representational’ problem discussed above. So, conveniently, we can sort of treat ‘embodied knowledge’ just like a ‘bodied’ version of normal knowledge, but with ‘pragmatic’ categories.

This is where I find the Bourdieu-ian version of embodiment least convincing. There’s lots to like in Bourdieu – his attention to the way everyday life influences perceptions, his linking of sociological categories (such as class) to performative differences, his focus on generative processes rather than reified cultural structures (although he frequently wanders close to reifying ‘structuring structures’). But the ‘body’ in Bourdieu is a very thin, non-corporeal, non-biological entity. As I think I read somewhere in Tim Ingold’s work, to paraphrase, we can keep putting cognates of ‘body’ and ‘meaning’ together in different variations without getting one fly-hop closer to really overcoming Cartesianism if we don’t truly treat the body as an organism, with all that this implies.

Bourdieu’s ‘body’ is no such organic thing; it is a unified ‘structuring structure,’ a kind of generative principle shaped by class and culture. It more closely resembles a grammar, a semiotic structure, or a cultural pattern than an organic flesh-and-blood creature, with all of the complexity, diverse levels of synthesis and nested dynamics, and inconsistencies.

In contrast, Adenzato and Garbarini (2006:752-753) also discuss an ‘emergentist’ approach to ‘expert systems,’ that is, to skilled actors who may not be so ‘expert’ in folk terms (they use the example of child who learns to walk). In this part of the discussion, they make reference to Andy Clark and to Esther Thelen and Linda Smith’s (1994) discussion of how children develop the ability to walk. Thelen and Smith (and Clark) in no way describe the child as acquiring a unified, simple ‘structure’ to walk, nor do they argue that a structure immanent in the environment is internalized (images more in keeping with Bourdieu). Rather, they discuss walking as an ‘emergent’ ability, posing a whole set of challenges for the child, such as aspects of balance, bodily control, integrating composite motions, restricting extraneous momentum, and compensating for motion as one updates self perception and interaction with the environment. Rather than an internalized ‘structure,’ the account of walking that Thelen and Smith provide leads one to see walking as a softly assembled set of different techniques, compensating for each other, that are not identical in each child (see, for example, Thelen’s discussion of reaching for an object in Piaget’s experiments).

I find Thelen and Smith’s account, and Clark’s discussion, much more persuasive than Bourdieu’s idea that each novice internalizes a ‘structuring structure.’ My own research on capoeira, dance, sport, and other physical techniques leads me to believe that there is no single way to do any technique of the body, and that even very similar kinetic solutions can be reached by varying routes. This variation seems utterly inconsistent with the idea that everyone is internalizing a structure immanent in the environment. Adenzato and Garbarini (2006: 753-754) cite a number of theorists (Ingold and Bernstein, for example) who offer accounts of enculturation and apprenticeship that are entirely consistent with Thelen and Smith — and inconsistent with Bourdieu’s overly homogeneous, overly distilled, overly cognitivist model of the habitus. All of these accounts point to a model of bodily enculturation that is more heterogenous, irreducibly complex, physiological and perceptual (rather than seeing bodily action as a form of ‘categorization’).

Embodied simulation through a mirror system means that gestures do not need to be reduced to a generative structure in order to pass from one person to another; they are incorporated directly through resonance a piece at a time. The brain does not need to reduce them to something else. Motor enactment is its own ‘logic’ (although I hesitate to use the word), not a representation of some symbolic or semantic meaning. Although I agree strongly with Adenzato and Garbarini that insights from embodied cognition should shape cultural theory, I emphatically disagree that it is liberating because it makes it ‘possible to maintain a representational conception of the mind’ (2006:757). On the contrary, the interdisciplinary study of embodied cognition allows us to see the limits of ‘representation’ as a rubric to model all the brain’s capacities.


Adenzato, Mauro, and Francesca Garbarini.
2006. The As If in Cognitive Science, Neuroscience and Anthropology: A Journey among Robots, Blacksmiths and Neurons. Theory & Psychology 16(6): 747–759.

Clark, Andy.
1997. Being There: Putting Brain, Body and World Together Again. Boston, MA: MIT Press.

Rizzolatti, Giancomo, and Laura Craighero.
2004. The mirror-neuron system. Annual Review of Neuroscience 27: 169–192.

Thelen, Esther, and Linda Smith.
1994. A dynamic systems approach to the development of cognition and action. Cambridge, MA: MIT Press.

Published by


Trained as a cultural anthropologist at the University of Chicago, I have gone on to do fieldwork in Brazil and the United States, and look forward to a new project in New Zealand. I have written one book, Learning Capoeira: Lessons in Cunning from an Afro-Brazilian Art (Oxford, 2005). I have also co-edited several books, including, with Dr. Daniel Lende, The Encultured Brain: An Introduction to Neuroanthropology (MIT, 2012), and with Dr. Melissa Fisher, Frontiers of Capital: Ethnographic Reflections on the New Economy (Duke, 2006). My research interests include psychological anthropology, sport, dance, human rights, neuroscience, phenomenology, economic anthropology, and just about anything else that catches my attention.

9 thoughts on “Beyond Bourdieu’s ‘body’ — giving too much credit?

  1. As I read your piece, Greg, I see us both struggling with two basic concepts—representation and categories. So I am just going to throw out an idea about pragmatic representation and see if it sparks anything. I won’t talk much about categories here, but I do think the embodied approach to categorization, such as promoted by George Lakoff or Lawrence Barsalou, is a much better candidate than the typical linguistic/abstract/logical approach to categorization. We need a more pragmatic approach to categorization, which can mediate between people in actual contexts, development, and cultural knowledge.

    And similarly with representation. The pragmatic approach put in mind the work of Charles Peirce. Jay Zeman has a discussion of Peirce’s Theory of Signs that I drew on in writing this comment. So here’s a couple pieces from Zeman’s essay:

    “About 1903, [Peirce] tells us that: “signs are divisible by three trichotomies; first, according as the sign in itself is a mere quality, is an actual existent, or is a general law; secondly, according as the relation of the sign to its object consists in the sign’s having some character in itself, or in some existential relation to that object, or in its relation to an interpretant; thirdly, according as its Interpretant represents it as a sign of possibility or as a sign of fact or as a sign of reason (2.243).”

    “Firstness is, among other things, the category of feeling… Firstness, the category of feeling in this sense, is preeminently the category of the prereflexive. The difficult thing about talking about firsts is that when we recognize that something is grasped as a first, its firstness as firstness effectively evanesces.”

    So, the firstness or the iconic level would be the canonical sensory neurons, doing the most immediate processing, which, to draw on Gibson, can include the affordances—a representation that mediates between both environment and behavior/intention, even at the most basic levels of processing. (Indeed, in writing this passage, I had to stop myself from using the word “representation” too much, as it becomes too internal, too cognitive, and thus misses the environment-and-interpretation that can happen even at the level of primary processing… Or at least that’s the idea I am throwing out. I am not sure how well it would stand up to recent research on the visual system.)

    In other words, the primary processing is iconic, shaped from raw sensory material yet made “representative” according to on-going demands and contexts.

    For the second level I might draw on what you wrote, “In other words, perception of movement is mediated not by a symbol of movement, but rather by, well, an experience of oneself moving, even if that is an inchoate simulation. The cognitive and phenomenological distance implied by the symbolic notion of representation, its cognitive-ness, is instead kinetic, visceral, and directly sensed.”

    For Peirce, the second level is commonly called the index, the classic example being that smoke indexes a fire burning. Mirror neurons might, and I emphasize might, be thought of this way, an indexing of what the canonical sensory neurons are processing.

    This sort of “indexing” would be much broader than normally considered in the use of Peirce. But I would say indexing could be what you describe with walking: “an ‘emergent’ ability, posing a whole set of challenges for the child, such as aspects of balance, bodily control, integrating composite motions, restricting extraneous momentum, and compensating for motion as one updates self perception and interaction with the environment. Rather than an internalized ‘structure,’ the account of walking that Thelen and Smith provide leads one to see walking as a softly assembled set of different techniques, compensating for each other, that are not identical in each child.”

    And the third level would be the classic symbol, arbitrary according to its physical substrate but historically contingent nonetheless. For brains, the symbolic level would be the use of mirror neuron processing that then gets projected (metaphorically?) or utilized by other parts of the brain. So, a stone gets placed into a wall or a “a rose by any other name would smell as sweet.”

    I don’t know how this works, really, but I was struck by the importance that in the firstness, there are nonetheless shades of the second and the third. I think the Monkey Makes Robot Walk blog shows a concrete example of how neuroscientists see it this way—the representation of the robot reaching right down to the motor neurons of the monkey. And that then gets indexed to the experimental conditions and the experiences and rewards of the monkey. And from there, the third level, I don’t know…

    But there might be one place to look. Terrence Deacon’s The Symbolic Species is one of the few books of integrative anthropology that draws on Charles Peirce’s theory of icon, index and symbol, in this case to discuss the evolution of human language.

    Zeman on Peirce’s Theory of Signs:

    Symbolic Species:

  2. Why would the multidimensional mechanism behind children learning how to walk be a criticism of habitus/structuring structures? I.e., how is it a *cultural* process? Don’t centipedes and puppies walk too, without the use of symbolic representation? Would a child left to fend for itself in the woods not start to scoot or crawl as soon as able? I guess I’ll have to look at your work on capoeira or dance for more relevant examples.

    I also wonder how the neurological-representational process behind processing images of motion works in the profoundly disabled. If one has no “walking experience” of one’s own, how is seeing other people walking represented? Does it get symbolized by being mapped onto another experience of one’s own body *then*?

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s