Thursday, 12 January 2012

What is Embodied Cognition? - A Brief Introduction - Pt 1/2

The Brain says "A complex of important misapprehensions center around the question of the provenance of thoughts. John thinks of me as the point source of the intellectual products he identifies as his thoughts. But, to put it crudely, I do not have John's thoughts. John has John’s thoughts." – Clark 1997

Embodied Cognition (EC) is a relatively new movement, incorporating a variety of fields such as Phenomenology, Artificial Intelligence, Ecological Psychology and Robotics. The problem in defining EC is that it is a broad research programme, rather than a theory per se and as such, is often conceptualised in various ways leading to quite different implications. Perhaps to begin with, it is best defined in terms of what it is not.

Mind is a Verb not a Noun – The Rejection of Cartesian Materialism
Embodied Cognition, at its most basic level, rejects the classical computational approach to cognition. Much of the work in current Cognitive Psychology,involves the brain being viewed as ‘housing’ the mind, with the body simply serving as a sensory/motor input/output of little significance,other than providing superficial data to and from the outside world, that the brain then computes.

However, Embodied Cognitivists argue there is a conceptual error here that goes back to Descartes’ dualistic mind-body distinction. Whilst essentially all Cognitive Psychologists would reject Descartes’ Supernatural Dualism, they fall at the last hurdle in replacing the ‘soul’ with the ‘brain’. EC argues that the mind does not reside in the brain, viewing the ‘movie’ in the ‘Cartesian Theatre’. The mind is part of a dynamical system, which we, as humans, do; We think, feel, perceive, act and so on.

EC also places a heavy emphasis on the importance of the body and environment in understanding cognition. The role of the body and environment incognition varies depending on how strong a view one takes, but broadly EC highlights the necessity of understanding cognition in the context of brain, body and environment. Hence, researchers must not isolate the brain when studying cognition, they must account for the content of cognition bystudying the body containing the brain. A stronger account conceives of cognition as a system that contains brain, body and environment, such that mind is extended into all three [1].

Before we consider the weak and strong views, it is important to emphasise EC does not deny the brain is essential for cognition, only that itis part of a larger system – rather like an engine is essential for a car but does not explain the whole story.

Why form Representations and Simulations? – There’s a Whole World out There!
So what is the evidence for EC? A common source comes from research showing that what appears to be complex computation by the brain, is often a much simpler process which relies not on forming representations, but on the direct information from our environment at that particular time. Consider the Outfielder Problem. On a classic computational model, in order for an outfielder to be able to successfully catch the ball, they must (subconsciously) compute the trajectory of the ball in order to judge where it will land. However, it turns out this isn’t actually what we do [2]. A much simpler, explanation is prospective control. This involves coordinating one’s actions and responding appropriately to the information perceived. In other words catching the ball is achieved by aligning oneself such that the ball is perceived as moving in a straight line rather than a curve.

A further, perhaps simpler, example is judging depth perception. Rather than requiring symbolic representation, we can judge depth from our head movements which cause objects in the foreground to appear to move, relative to the background.

The point is that (in these examples at least) it is shown how we cognize without computation and that our cognition is strictly dependent on our direct perception, action and our response. There is no need to form an internal model which we then use to compute, when we can use the actual information directly perceived in real time.

The weak/strong divide
I would argue at this point there seems to be a divide between weak and strong versions. Weaker versions may argue this doesn’t necessarily challenge the classical approach entirely, rather it provides an extra tool or context for which cognition is to be understood. 

The problem is that whilst this kind of research provides good evidence based examples in favour of EC, there is also the equivalent evidence for a computational approach for other aspects of cognition, such as dead reckoning [3]. One challenge then, EC faces, is to give its own account of these sorts of cognition which seem to be adequately explained within a computational paradigm. Could it be possible that some aspects of cognition are computational whilst others are not?

Further, perhaps the biggest challenge is explaining ‘offline cognition’, such as imagination, day dreaming, planning and so on, which (arguably) seem to have little or no direct involvement with the outside environment. For example, how can sitting in a chair imagining catching a ball or planning your day, utilise your direct perceptions of the environment?

It is important to note, however, that since EC is in its early stages, this is by no means a reason for it to be rejected, merely a consideration for future exploration. This is certainly the view taken by the stronger EC account, which shall be considered in part 2. 

References and further readin

[1] Clark, A., & Chalmers, D. (1998). The Extended Mind Analysis, 58 (1), 7-19 DOI: 10.1111/1467-8284.0009
[2] McBeath MK, Shaffer DM, & Kaiser MK (1995). How baseball outfielders determine where to run to catch fly balls. Science (New York, N.Y.), 268 (5210), 569-73 PMID: 7725104   

Embodied cognition is not what youthink it is - Excellent blog by notes from two scientificpsychologists

Embodied Cognition -Another great blog by Fixing Psychology

Six views of embodied cognition - M Wilson - Psychonomic Bulletin & Review,2002  

Brain Science Podcast -evidence for the computational approach with Randy Gallistel

Brain Science Podcast -evidence for embodied cognition with Larry Shapiro


  1. I liked the article very much Chris! Especially, the point of EC serving as an extra tool to understand the current perspective of cognition is a point I would like to support. Also, albeit, it gives us the triad of brain-body-environment that plays and important role in formulating the concept of cognition, it also fails to define the meaning of brain and mind at the same time. Is mind the conscious and the brain only the physiological center carrying out the chemical processes that are the only logical explanation of the thinking or feeling that we do? Or does it have any other explanation?

    As far as the point made in the weak/strong divide, which is in support of the weaker divide, regarding the incapacitation of EC to explain the processes of imagination, day dreaming and planning under the umbrella of offline cognition, it also fails to explain the mystic phenomenon of Synesthesia and also the altered cognition, which I say are the delusions and hallucinations. I'm eagerly waiting for the strong point.

    I'm happy that EC puts forth the brain-body-environment interaction with the influence of mind seeping into all three as the very basis of cognition, but as it is a neophyte of a concept, we've yet to see how it further explains the very many processes covered under cognition.

    Thanks mate, for this interesting article! Cheers!

    -Kartik Angara

  2. Hi Kartik thanks for the support and comments.

    RE your point about defining what it actually means for the mind and brain...As I understand it, EC argues the brain is one aspect of the process which is the mind. As such, it doesn't make sense to speak of the mind arising from the brain in isolation. The mind is a dynamic system, of which the brain is just one part but other parts are necessary. Check out pt2, I cover it there i think.

    RE your second point of synesthesia and hallucinations, this is interesting and I'm not sure what EC would argue to be honest. There's a good discussion of this 'offline' issue in general on the 'two scientific scientists blog' i referenced - both the blog and and the comments are very insightful. Also check out the 'fixing psychology' blog on embodied cognition - both are excellent and explain it better than me!

    Like you note, and what is commonly emphasised is that the current challenges to EC simply havn't been adequately explored yet so it's a case of not knowing, rather than EC failing to explain...

    I hope this is helpful?

  3. Forgot to mention in the blog RE evidence for embodied cogntion, that the success and failures of AI and robotics have had a big impact. See Link

  4. @eric it's on pt2 thanks for the comment I'll hav a read and comment back tomorrow :-)

  5. Kartik,
    One aspect of the EC argument regards the direction in which psychology should proceed. My impression (from sparse readings of works from Plato through the middle ages) is that once upon a time you could start psychology at any point, by wondering about any interesting phenomenon, and proceeding from there however you wanted. Sometime around Descartes, I’m not sure if he started it or just popularized it, everyone decided that you had to start by figuring out so-called higher mental functions. The idea was that if you got a handle on how that stuff worked, you would automatically understanding how lower mental functions. In those days, hierarchies of being were all the rage, so the higher vs. lower thing worked. The “offline” vs. “online” distinction uses a computer analogy, but is basically the same thing. I would never say that this was a bad hypothesis, but it really hasn’t worked out. It is not clear that all of our studies of thinking, reasoning, planning, imagining, etc., over the past centuries has told us much about how behavior works, and it is unclear how much it has even told us about thinking, reasoning, planning, and imagining.

    Several of the approaches that come together in the more extreme forms of EC assert that the direction of inquiry has been backwards. We can make rapid progress on the problems of behavior – success at this has been demonstrated in several areas. It is also quite likely that a better understanding of behavior will help us understand how the “offline” activities work. Thus, you are correct that EC is not very good, at present, in dealing with phenomenon such as delusions, hallucinations, mystic experiences, synesthesia, etc., but that is to be expected. It is quite likely that a better understanding of behavior will lead us to redescribe those phenomenon, and provide at least the start of a good explanation for them.

    Does that seem reasonable (or at least coherent)?

  6. (from an email exchange with Greg Nirshberg who runs

    "I think the representation question is important though. And most of my academic work in this area is focused on explicating the positive aspects of EC while pointing out the need for representation. Representation gets a bad rap because it tends to immediately be associated with classical naive accounts of representation as "re-presentation" of pregiven features of the external world through some sort of pattern of neuronal firing. That there is some internal image that is created inside the brain. I agree that these old accounts fail. But regardless of what word you put on it, our conscious experience is constructed, and while embodied cognition provides a more reliable mechanistic account of perception and learning and memory and skill acquisition, there is nothing about these mechanistic accounts that provides a real explanation for why certain causal interactive processes in the universe allow consciousness to emerge or arise, whether those processes happen to be distributed through a body or not, and whether we take into account the interactive relationship the system has with its environment or not. Even if embodied cognition would want to deny any causality to mental states, and say that it's all epiphenomenal, there's still no explanation for why or how the epiphenomena. There's all sorts of causal interactive relationships between systems in our universe, why are certain ones conscious?

    EC seems to me to provide the empirical data needed to flesh out better notions of how consciousness arises. Representation can't be some simple correlation with things in the world if content is anticipatory, if it's constructed out affordances and satisfaction conditions. But the world also can't be "it's own best representation" as there are all sorts of perceptual errors and hallucinations that we're prone to. There's a discrepancy between our experience and the world, and something about the nature of that discrepancy is important for figuring out how to talk about the normativity, aboutness, intentionality of our mental content towards the world."

  7. (email continued...)
    Chris -

    are you essentially refering to the hard problem? I.e. that whilst EC gives us a more parsimonious account of the easy problems of perception, memory etc (what Chalmers calls awareness), it does little to explain how subjective experience arises. If so, I would agree, but would you argue representation does provide a better account?

    With regards to 'world cannot be it's own representation', How would you interpret the outfielder problem? This seems to be a good example of how the world can be used as it's own model, without the need for representation and computation?
    I find the issue of offline cognition more problematic for EC. It is hard to understand how EC could explain cognizing such as planning, day dreaming and imagery, as there is no perception of environmentally relevent stimuli to aid cognition.

  8. (email continued)

    "I am in a sense referring to the hard problem, but I'm also not so sure we should make such a sharp divide between the easy and hard problems. I think that representation *can* be a better account, but I'm not very impressed with most accounts of representation. I just hesitate to give up the word. I think we need to give up old computational/information processing accounts, and I've mostly embraced approaches within embodied cognition and dynamic systems, but I still find these to be correlative explanations. It's like cognitive scientists who are wrapped up in finding the neural correlates of consciousness. Even if that could be done, it doesn't actually tell us anything about consciousness (well, it tells us some things). Same with finding the sensorimotor full body dynamic correlates of consciousness. I think there's a big problem in philosophy of mind that confuses contact with the world with content about the world. So in your outfielder example, I wouldn't say the world serves as its own best model, it serves as the necessary contact with the environment for the biological system to engage in certain types of interactions. But that the subjective experience of the world, and the intentionality with which that interaction takes place, doesn't just fall out of embodied cognition. It needs to be explained. Whether the field calls it "representation" or not, I'm not terribly concerned with. But I'm always a bit miffed at the denial from EC people that there is anything more to explain.

    Like you point out, dreaming and imagining are problematic. As well as perceptual error, illusions, hallucinations, etc...there is something about our experience that does not correlate directly to the outside world, and how that is constructed, how that emerges from any amount of interaction, is what I find interesting. "