The computational approach is the orthodoxy in psychological science. We try and understand the mind using the metaphors of information processing and the storage and retrieval of representations. These ideas are so common that it is easy to forget that there is any alternative. Andrew Wilson is on a mission to remind us that there is an alternative – a radical, non-representational, non-information processing take on what cognition is.
I sent him a few questions by email. After he answered these, and some follow up questions, we’ve both edited and agreed on the result, which you can read below.
Q1. Is it fair to say you are at odds with lots of psychology, theoretically? Can you outline why?
Psychology wants to understand what causes our behaviour. Cognitive psychology explanations are that behaviour is caused by internal states of the mind (or brain, if you like). These states are called mental representations, and they are models/simulations of the world that we use to figure out what to do and when to do it.
Cognitive psychology thinks we have representations because it assumes we have very poor sensory access to the world, e.g. vision supposedly begins with a patchy 2D image projected onto the back of the eye. We need these models to literally fill in the gaps by making an educated guess (‘inference’) about what caused those visual sensations.
My approach is called radical embodied cognitive psychology; ‘radical’ just means ‘no representations’. It is based on the work of James J Gibson. He was a perceptual psychologist who demonstrated that there is actually rich perceptual information about the world, and that we use this information. This is why perception and action are so amazingly successful most of the time, which is important because failures of perception have serious consequences for your health and wellbeing (e.g. falling on ice)
The most important consequence of this discovery is that when we have access to this information, we don’t need those internal models anymore. This then means that whatever the brain is doing, it’s not building models of the world in order to cause our behaviour. We are embedded in our environments and our behaviour is caused by the nature of that embedding (specifically, which information variables we are using for any given task).
So I ask very different questions than the typical psychologist: instead of ‘what mental model lets me solve this task?’ I ask ‘what information is there to support the observed behaviour and can I find evidence that we use it?’. When we get the right answer to the information question, we have great success in explaining and then predicting behaviour, which is actually the goal of psychology.
Q2. The idea that there are no mental representations is hard to get your head around. What about situations where behaviour seems to be based on things which aren’t there, like imagination, illusions or predictions?
First, saying that there are no mental representations is not saying that the brain is not up to something. This is a surprisingly common mistake, but I think it’s due to the fact cognitive psychologists have come to equate ‘brain activity’ with ‘representing’ and denying the latter means denying the former (see Is Embodied Cognition a No-Brainer?).
Illusions simply reveal how important it is to perception that we can move and explore. They are all based on a trick and they almost always require an Evil Psychologist™ lurking in the background. Specifically, illusions artificially restrict access to information so that the world looks like it’s doing one thing when it is really doing another. They only work if you don’t let people do anything to reveal the trick. Most visual illusions are revealed as such by exploring them, e.g by looking at them from a different perspective (e.g. the Ames Room).
Imagination and prediction are harder to talk about in this framework, but only because no one’s really tried. For what it’s worth, people are terrible at actively predicting things, and whatever imagination is it will be a side-effect of our ability to engage with the real world, not part of how we engage with the real world.
Q3. Is this radical approach really denying the reality of cognitive representations, or just using a different descriptive language in which they don’t figure? In other words, can you and the cognitivists both be right?
If the radical hypothesis is right, then a lot of cognitive theories will be wrong. Those theories all assume that information comes into the brain, is processed by representations and then output as behaviour. If we successfully replace representations with information, all those theories will be telling the wrong story. ‘Interacting with information’ is a completely different job description for the brain than ‘building models of the world’. This is another reason why it’s ‘radical’.
Q4. Even if I concede that you can think of the mind like this, can you convince me that I should? Why is it useful? What does this approach do for cognitive science that the conventional approach isn’t or cant’?
There are two reasons, I think. The first is empirical; this approach works very, very well. Whenever a researcher works through a problem using this approach, they find robust answers that stand up to extended scrutiny in the lab. These solutions then make novel predictions that also perform well – examples are topics like the outfielder problem and the A-not-B error [see below for references]. Cognitive psychology is filled with small, difficult to replicate effects; this is actually a hint that we aren’t asking the right questions. Radical embodied cognitive science tends to produce large, robust and interpretable effects which I take as a hint that our questions are closer to the mark.
The second is theoretical. The major problem with representations is that it’s not clear where they get their content from. Representations supposedly encode knowledge about the world that we use to make inferences to support perception, etc. But if we have such poor perceptual contact with the world that we need representations, how did we ever get access to the knowledge we needed to encode? This grounding problem is a disaster. Radical embodiment solves it by never creating it in the first place – we are in excellent perceptual contact with our environments, so there are no gaps for representations to fill, therefore no representations that need content.
Q5. Who should we be reading to get an idea of this approach?
‘Beyond the Brain’ by Louise Barrett. It’s accessible and full of great stuff.
‘Radical Embodied Cognitive Science’ by Tony Chemero. It’s clear and well written but it’s pitched at trained scientists more than the generally interested lay person.
‘Embodied Cognition’ by Lawrence Shapiro that clearly lays out all the various flavours of ‘embodied cognition’. My work is the ‘replacement’ hypothesis.
‘The Ecological Approach to Visual Perception’ by James J Gibson is an absolute masterpiece and the culmination of all his empirical and theoretical work.
I run a blog at http://psychsciencenotes.blogspot.co.uk/ with Sabrina Golonka where we discuss all this a lot, and we tweet @PsychScientists. We’ve also published a few papers on this, the most relevant of which is ‘Embodied Cognition is Not What You Think It Is‘
Q6. And finally, can you point us to a few blog posts you’re proudest of which illustrate this way of looking at the world
What Else Could It Be? (where Sabrina looks at the question, what if the brain is not a computer?)
Mirror neurons, or, What’s the matter with neuroscience? (how the traditional model can get you into trouble)
Prospective Control – The Outfielder problem (an example of the kind of research questions we ask)
Mind Hacks 
I’d be more excited by this interview if it gave any clue what “information” means to Andrew Wilson: surely not just something measured in bits or negentropy. What is it, and what does it mean for information to be ‘about’ the world?
I mean information as in James J Gibson ecological information. This is structure in ambient energy arrays that can specify dynamical properties of the environment. See http://psychsciencenotes.blogspot.co.uk/2011/11/specification-i-what-it-is-and-why-we.html and http://psychsciencenotes.blogspot.co.uk/2012/11/task-dynamics-and-information-they.html for more details. plus Turvey, Shaw, Reed and Mace (1981) for the as yet best formal description.
Defining information from an ecological perspective is indeed critical, and something radical embodied psychologists (or, ecological psychologists) have given a lot of thought to. You are correct that we are not talking about Shannon information.
An accessible introduction is given by Gibson in the 1979 book mentioned in the post. Since it is central, explaining it will touch upon may other aspects of the theory that the interview did not address, so it will only raise more questions I’m sure. But I will do my best.
In short, information is a property of an ambient energy array (for example, light) that comes into a 1:1 relationship with a meaningful property of the organism-environment system. The former are often called “invariants” because they are the aspects of the ambient energy that invariantly correspond to some meaningful property despite variation in other properties. This 1:1 relationship is also called specification.
There is some debate among ecological psychologists as to the details of this account. See this blog post for more details: http://psychsciencenotes.blogspot.co.uk/2012/06/how-information-gets-its-meaning.html
To give a concrete example, an upcoming collision is specified by a certain pattern in the visual flow field: specifically, an expanding region of the visual field that includes the focus of expansion (if it doesn’t include the focus of expansion, the organism is on course to pass the object without collision).
In this case, there is a pattern in an ambient energy array that is a product of the array being sampled in a certain way (due to a point of observation moving through it). This pattern specifies a behaviorally meaningful property (upcoming collision). The information is “out there” in the environment because it can only be defined across the organism-environment system. All the properties involved are relative properties. The information only exists with respect to an active organism moving through the world—one could spend a lifetime studying static retinal images and never encounter this pattern.
Why wouldn’t it be something measured in bits (negentropy is not a unit)? Whether the brain uses representations or not, uncertainty (entropy) reduction (= information) ought to be the whole point of learning.
I guess it’s because I haven’t been classically trained in the cognitivist tradition, but “radical” embodied cognition always strikes me me as nothing but behaviourism reborn without the questionable idea that it’s limited to explaining externally observable behaviours.
Reblogged this on Shane O'Mara's Blog.
As a neuroscientist I’ve always found anti-representationalism really confusing. The examples they work with don’t help either.
Are neural representations as problematic to them as mental representations? For instance, is a retinotopic or tonotopic map a representation? Is a place field a representation?
Is a memory a representation? It seems to be that it has to be. Some memories are environmentally constrained, but many are only minimally so. I can remember my name, phone number and address in a very wide variety of contexts. What do I gain by asserting that these pieces of information are not ‘represented’?
And would the non-represtatationalists be happy if we just replace the word ‘representation’ with ‘mapping’?
Is this why we can’t download ourselves into computer and still be us and human? I welcome anything that challenges the idea that what we essentially are is a mind in a vat that can be converted into a computer program. It seems sophomoric beyond all belief.
Are people working on radical embodied cognition, perhaps working on developing language that helps us avoid the mind/brain division? It’s so limiting and this mind/brain, nature/nurture dichotomy is a pernicious zombie.
Suspicion: Many active scientists use ‘representation’ differently; some uses are more theory-laden, or ‘THICK’, than others. Anti-representational theories are often incredibly thick: they pose themselves against (some!) positively thick representations. Hence, active scientists who do not have a thick sense of representation find radical embodiment insignificant.
“If we successfully replace representations with information, all those theories will be telling the wrong story. ‘Interacting with information’ is a completely different job description for the brain than ‘building models of the world’.”
—I suspect I am not alone when I say that representations can be thought of as information in the brain that is (roughly) isomorphic with some external information. I am persuaded by prediction error minimization (PEM) accounts of cognition, so these isomorphic structures (representations) form due to top down and bottom up processes. This top down aspect makes interaction incredibly important. In sum: we build ‘models’ through ‘interacting with information’. Evolutionary processes have been relatively kind, so we rarely have models that radically misrepresent the (useful parts of the) world.
I so agree with your pointer to terminology. Concept-related concepts are variously used, likely because it is a budding topic in itself. In language comprehension studies, we find ‘processing,’ and ‘representation’ used quite alike, even along with ‘integration’ less often.
I really enjoy reading both Andrew’s blog and Mindhacks. I was delighted to find my humble little area of the sciences (ecological psychology) getting some exposure today, on one of my favorite blogs, no less.
Great interview, Andrew did a wonderful job summarizing some of Gibson’s key insights, and he did it in plain language, which is not always easy to do (it becomes tempting to lapse into our own jargon).
For those with questions, I’m glad you are interested (even if it’s incredulity more so than curiosity). All of the readings Andrew suggested are fantastic. I personally read Chemero’s book and Gibson’s 1979 book at the same time a number of years ago, and I’ve never been the same since (in a good way, if you can’t tell).
Gibson’s writings are surprisingly approachable for a Psychologist (let’s face it, experimental psychologists are generally not good writers).
“Are neural representations as problematic to them as mental representations? For instance, is a retinotopic or tonotopic map a representation? Is a place field a representation?”
I think a better question to ask is, given the neural behavior we observe, how might it be obtained without postulating representations? Unfortunately, there are not many non-representational neuroscientists out there, and the ecological crowd sometimes displays an aversion to the brain. But since you asked, yes, yes, and yes.
The brain is still very important, it’s the means by which perceiving-acting is even possible (in tandem with the environment and the whole organism, of course). Brains clearly plays some kind of role in perceiving-acting. How much of a role, and the specifics of that role, are greatly debated, though.
However, I’m confident that the brain’s job description is not to generate representations, infer, calculate, or construct perception. It doesn’t need to do those things if the information is already in the environment. As Andrew pointed out, there are a plethora of studies which show that the information is in the environment, and organisms do detect and utilize it.
“And would the non-represtatationalists be happy if we just replace the word ‘representation’ with ‘mapping’?”
Nope. The map is not the territory, just as the representation is not the object. All those “maps” in the brain probably don’t do what we think they do, anyway: Graziano, et. al.’s 2002 article published in Neuron, The Cortical Control of Movement Revisited is one example of how our knowledge of some cortical maps may be quite mistaken.
The idea that information is in the environment is not mutually exclusive with the idea that there are mappings. After all, the information has to reach motor cortex (for example) somehow right?
Imagine a game involving pressing a button whenever you detect squares on a screen, irrespective of their location, size and color. (A very easy game!) What do we call the neural network or circuit that facilitates this, if not an invariant representation? It is a ‘something’ that allows the behavior to be invariant or symmetrical with respect to changes in location, size and color.
In other words, there needs to be some type of neural activity that can connect to the motor cortical neurons controlling the finger that presses the button.
What do non-representationalists call this sort of neural activity if not a mapping or a representation? (Framing things in terms of invariant representations also allows for some useful mathematical treatments.)
How does a non-representational model go about solving classification problems? I can build a very useful artificial neural network that starts with a stimulus representation, such as a frequency map for sound. The hidden layers create higher order representations, and the output layer gives you a category or a target function.
We know these networks do what is advertized — hence the success of things like Deep Learning. What do non-representationalists think of these techniques?
The best work in brain behavior science, period, is Paul Cisek. Cognitive models are just human exceptionalism and violate all sorts of biological tenets.
Because cognitive models support cultural, top-down, higher order concepts and the primacy of subjective experience they are the best funded – in the U.S.. They are just dum and don’t even make common sense.
O.K. this excites me. I’m going to spend a lot of time on this. I am in the middle of reading at
http://psychsciencenotes.blogspot.co.uk/2011/07/theres-more-to-us-than-our-brains-so.html
and two things have been brought up that I thought about with my first readings on this post— change blindness, and William James, in the context of seeing. I think the aeroplane test might show quite different results in non-Western cultures and that the blindness is not caused by some weakness in our brains, but the fact that in the U.S. and other Western cultures we have grown unaccustomed to LOOKING. We see “plane” in the same way children see people in their lives in the stick figures they draw. It’s a symbolic interpretation. Also, I and many others cue to change and tend to be blind to things that don’t change, which is what makes me and others excellent scope dopes.
It also makes me wonder about Karl Jasper’s concept of “direct observation”. Perhaps we do observe things directly all the time, but edit and interpret things through filters that are formed by a system of values that is determined not to see the thing in itself but whether or not it is useful to us according to our philosophy and fast-paced lifestyle. Many Americans do not perceive the fruit growing on a tree in their yard as food. I surmise that it takes a lot of programming in not seeing to miss something of such animal obviousness.