Response to Fred Adams' latest critique of "Embodied Cognition"

Fred Adams has a new article out online in Phenomenology and the Cognitive sciences entitled “Embodied Cognition”. Adams is renowned for being skeptical of the 4E movement in philosophy of mind (embodied, embedded, extended, enacted). He wrote a book with Ken Aizawa called “The Bounds of Cognition” that challenges the core claims of embodied cognition. However, given his familiarity with the literature, I am very puzzled by the paper. He starts off the paper talking about Varela and Gallagher as exemplars of the embodied cognition thesis, but then spends most of the paper talking about how to reduce sentential belief-symbols to literal simulations of motor output. He writes as if sentential comprehension is the main explanatory target of EC theorists when they say “cognition is embodied”.

Anyone who has read Varela and Maturana’s work on autopoiesis would be very confused about this formulation of the problems that embodied cognition sets out to study. Varela says, for instance, that “Living systems are cognitive systems, and living as a process is a process of cognition. This statement is valid for all organisms, with and without a nervous system.” Varela thinks that even the unicellular organism “cognizes” in virtue of its emergent self-organization of autopoiesis. This is the actual claim of embodied dynamic systems approaches to cognition, a far cry from the thesis that:

In the embodiment literature, we find the empirical step consisting of empirical correlations between certain kinds of cognitive processing and sentence comprehension and certain kinds of perceptual/motor performance.

Gibson was never concerned with “sentence comprehension”. While an admirable explanandum, Gibson thought we need to first better understand the more basic cognitive processes before we attempt to theorize about higher cognitive processes. He was almost always concerned with the cognition that we share with our animal cousins, not sentence comprehension or symbolic cognition. Many EC theorists actually propose a dual-level or dual-process model of reasoning wherein there exists a primordial, nonsymbolic level of cognitive processing shared by all animals (online processing) and a evolutionary recent and sententially grounded level of rational, serial processing (offline processing). I don’t know of any serious theorist proposing these two levels of distinctions  who makes the absurd claim that offline processing must be explained strictly in terms of online processing. Once external representations are taken up and integrated with the functioning of the cognitive system, there is no reason to suppose that the mechanism is only that of “simulation”. For example, Gibbs claims that representational (propositional) reasoning depends heavily upon analogical reason, which needs to be analyzed at the appropriate level of abstraction, not that of neurons firing. In all likelihood, it will require different explanatory tools and and terminology to explain both offline processing and online processing. Most EC theorists would simply emphasize the importance of recognizing that propositional reasoning comes after or “out of” online processing on both the phylogenetic and ontogenetic scales.

Accordingly, there seems to be a strange disconnect between Adams’ picture of EC and what the majority of serious theoreticians (that I know of) are proposing. The more I think about it, the more I think that this is a result of a widespread misunderstanding of what EC is, particularly in respect to the original formulations of Merleau-Ponty and Gibson. Some EC critics think that when we say “cognition is embodied” we are claiming that their conception of “cognition” is embodied. In actuality, we are trying to redefine what we mean by “cognition” and move away from definitions of cognition focused on sentential understanding. This is why Evan Thompson follows Varela in saying that all lifeforms exhibit cognition. Cognition is no longer manipulation of symbols, but regulation and coordination of emergent autonomous animacy/agency. This forces us to think about representations in terms of control and coordination of intrinsic movement rather than in terms of mirroring or “belief-formation”. Cognition is not sentence comprehension nor mastery of propositional concepts. We need to come up with a different concept to capture such higher-level processes.

I follow Julian Jaynes in making a distinction between what we can call cognition and narrative-consciousness. Narrative-consciousness enables the type of sentential mastery and understanding that Adams spent most of his time in the paper talking about. Giving the unique representational medium of sentential symbols, I see no reason why there cannot be an abstract analysis of such narrative mastery in terms that do not reduce to “sensori-motor simulation”. Which isn’t to say that we can make no progress on learning about the underlying functional circuitry which enables offline processing. Researching into resting state connectivity and anti-correlated functional networks is now opening up new vistas in understanding the neural distinction between online and offline processing.

This brings me to my next point: the misunderstanding of “meaning” and “affordances”. Adams follows Glenberg and Kaschak in defining affordances as “a set of actions available to the animal.” In this view, Adams seem to suggest that affordances are those cognitive systems which enable and support interaction between animal and environment. But this is exactly wrong. Affordances are not within the animal and they do not “arise” or “emerge” out of the interaction or “relation” between the animal and the environment. Affordances are real and objective. Meaning is external to the animal. For example, the ground affords support to all animals whether or not any particular one of them utilizes it for support. The affordance-property of support is embedded into the actual nature of the ground. What it really is determines what it means for the animal.

Accordingly, meaning is not generated by the interaction by the animal and environment, it is sought out and utilized. I get the feeling many EC supporters make this mistake as well. Meaning is external to the animal and needs to be found and used. For animals with the appropriate bodily capacities then, the process of finding the affordances can be decoupled from the process of using the resource. I therefore have problems with Zwaan and Madden, who Adams quote as saying “…there are no clear demarcations between perception, action, and cognition.”

I think this is stated poorly. For many higher animals, there is a clear distinction between the processing of detecting affordance-information (what Gibson calls “stimulus” or “ecological” information) and the utilization of that information for means of adaptive behavior. The is the distinction that Gibson makes between exploratory behavior and performatory behavior. However, it would be a mistake to conclude from this that the input-output model of perception is therefore right. The fact that the physical stimulus does not equate with the informational stimulus supports that idea that perception is but a perturbation upon an intrinsic dynamic network not a specific input which is mechanically read-off and used to send specific commands. As the frame-problem indicates, any concept of the cognitive system which understands the input to be “raw” or “meaningless” is bound to fail to produce functional specificity across widely changing environment demands. For embodied cognition, the given is already valenced in terms of what kind of information the animal is seeking in accordance with its internal dynamics and regulatory demands. The is the only way to avoid the input-out model. Doing so also allows us to escape from the Myth of the Raw Input, otherwise known as the Myth of the Given.

Advertisements

9 Comments

Filed under Philosophy, Psychology

9 responses to “Response to Fred Adams' latest critique of "Embodied Cognition"

  1. Tim

    Interesting piece, Gary. Just wanted to make a comment on this bit:

    “This brings me to my next point: the misunderstanding of “meaning” and “affordances”. Adams follows Glenberg and Kaschak in defining affordances as “a set of actions available to the animal.” In this view, Adams seem to suggest that affordances are those cognitive systems which enable and support interaction between animal and environment. But this is exactly wrong. Affordances are not within the animal and they do not “arise” or “emerge” out of the interaction or “relation” between the animal and the environment. Affordances are real and objective. Meaning is external to the animal. For example, the ground affords support to all animals whether or not any particular one of them utilizes it for support. The affordance-property of support is embedded into the actual nature of the ground. What it really is determines what it means for the animal.”

    I think you have to be careful here. If affordances are ‘real’ and ‘objective’, what are we to make of, say, the affordances of a pen. A pen might be perceived as a writing tool for one animal, nest-making material for another, a stabbing weapon for another, and something for another animal to ignore. So in what sense is the affordance of the pen ‘objective’? You could make the case that the affordance is ‘in the materiality of the pen’, but the pen’s material presence has no objective status shorn of a nervous system’s bringing-it-forth. In this sense, affordances are as much ‘in’ the eye of the beholder as they are ‘of’ the pen. This seems intuitive in so far as a perceiving organism has the sort of body to make use of said affordance.

    Look at it this way: what the organism senses is a function of how the organism moves. How the organism moves is a function of what the organism senses. If an organism thereby ‘structurally couples’ with its environment (in Varela and Maturana’s terms), doesn’t that mean the nervous system constitutes perceived affordances? If brain, body and environment are mutually embedded systems, then they cannot be internally and externally located with respect to one another. Surely this goes for affordances too?

    Here’s a quote from Gibson that might help: J.J.Gibson. “To perceive the world is to coperceive oneself … The optical information to specify the self… accompanies the optical
    information to specify the environment… The one could not exist without the other… The supposedly separate realms of the subjective and the objective are actually only poles of attention. The dualism of observer and environment is unnecessary. The information for the perception of ‘here’ is of the same kind as the information for the perception of ‘there,’ and a continuous layout of surfaces extends from the one to the other.” (Gibson, The Ecolological Approach to Visual Perception 1979, p.116).

    Awesome blog, Gary. Keep it up!

    • Tim, you should read Turvey’s paper laying out the ‘affordances as dispositions’ ontology; affordances are not relations, but real properties of the world expressed with respect to an organism. A crude analogy to keep in mind is that a table has a real length, which might be 30cm to a cm ruler but 12in to an inch ruler. Same property, measured with a different device.

      I’ve written about the affordance as relations vs dispositions things here: http://psychsciencenotes.blogspot.com/search/label/affordances

      Gary, nice summary. This issue of the embodiment hypothesis being a redefinition of what it means to be cognitive is, I think, an important distinction that’s not getting enough play.

  2. Gary Williams

    Tim, you said:

    I think you have to be careful here. If affordances are ‘real’ and ‘objective’, what are we to make of, say, the affordances of a pen. A pen might be perceived as a writing tool for one animal, nest-making material for another, a stabbing weapon for another, and something for another animal to ignore. So in what sense is the affordance of the pen ‘objective’? You could make the case that the affordance is ‘in the materiality of the pen’, but the pen’s material presence has no objective status shorn of a nervous system’s bringing-it-forth. In this sense, affordances are as much ‘in’ the eye of the beholder as they are ‘of’ the pen. This seems intuitive in so far as a perceiving organism has the sort of body to make use of said affordance.

    Affordances are objective in the sense that they exist for a species population rather than for an individual. Think about it this way: if there is a population of animals with roughly the same body type (as there is for any species), all the affordances of the environment are going to persist overtime regardless of whether any particular individual utilizes those resources. It is precisely because the affordance property persists through time independently of its utilization that allows for the evolutionary selection pressure upon distributed function units since animals use the invariant affordance information to regulate their behavior.

    Moreover, the affordance is “objective” because as Gibson said, what a thing objectively is determines what it means. In your example, it is the objective properties of the pen which determines its function in respect to each particular organism that wants to use it for something. And, of course, you are perfectly right to point out that the function is entirely relative to each individual. But it is the objective properties of the pen that make is useful for making nests, just as it is the objective properties of the pen that make it a good writing tool. In this respect, affordance properties exist independently of us. Our task is to find, detect, and utilize that information. This is why Gibson always emphasizes the active or exploratory nature of perception.

  3. Ken Aizawa

    Gary, I’ve not read this version of Fred’s paper, but I think a charitable interpretation of what Fred is up to is that he is *not* going to talk about *all* the things associated with embodied cognition, but one segment or strand. This is how the paper begins:

    The view that cognition is embodied (Valera et al 1991; Gibbs 2006; Gallagher
    2005) is rapidly gaining prominence in the world of cognitive science, and is aiming
    for dominance. “Embodiment” in embodied cognition covers different things tor
    different scientists, so I will limit my remarks here to a particular strand that is being
    investigated largely in the psychological literature (Barsalou 1999, 2010; Gallagher
    2005; Glenberg and Kaschak 2002; Glenberg et al. 2004, 201Oa, b; Pecher and
    Zwaan 2005).

    • Gary Williams

      But I think he picked what is arguably the weakest argument for the thesis that “cognition is embodied in the whole brain-body system”, a thesis that Adams is skeptical of. I don’t think the cognition being discussed in the paper is what most EC theorists refer to when they say “cognition is embodied”.

      Why pick on the weakest strand in order to show that the thesis of EC has problems? Why not address Gibson himself? Or Varela and Maturana for that matter.

      • Ken Aizawa

        I don’t think that Fred shares your sense of what is the weakest argument for embodied cognition or perhaps what most EC theorists refer to when they say cognition is embedded.

  4. Pingback: Embodied cognition/Cognition incarnée : le débat fait rage | REALISTA Approches réalistes en linguistique

  5. “Cognition is no longer manipulation of symbols, but regulation and coordination of emergent autonomous animacy/agency……Cognition is not sentence comprehension nor mastery of propositional concepts. We need to come up with a different concept to capture such higher-level processes.”

    I think this is key and neglecting this idea is likely to obscure the value of the embodied approach to cognitive science. I’ve become interested because of how this line of thought contributes to my perspective in general and my thoughts about mathematics in particular.

  6. Pingback: Gilles en vrac… » cognition incarnée

Leave a Reply

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

WordPress.com Logo

You are commenting using your WordPress.com 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 )

Google+ photo

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

Connecting to %s