Today on the Perception in Action podcast, the second episode in a series looking at the incredible contributions of J.J. Gibson to the study of perception in action. What is affordance theory and how can it be applied to sports?
Where do effectivities or action capabilities fit in? So it's time for a call to action. Hi, everyone.
This is Rob Gray from Arizona State University and PerceptionAction.com. Welcome to the Perception in Action podcast, where I discuss how psychological research can be applied to improving performance, accelerating skill acquisition, and designing technologies. In today's episode, I want to look at what is perhaps Gibson's most famous idea, affordances.
To be honest, up front, this is a concept I've struggled with over the years, not so much with understanding, As is the case with most of what Gibson did, I think it's an elegantly simple and intuitive idea, but more so with how we can take the general idea and get into the details of how we control our motor actions. I will get into that in the final episode of the series, where I look at work that has extended Gibson's basic idea. Before diving in today, I would also like to thank everyone for the hugely positive responses I received about the first episode in this series. I'd like to take credit for it.
but I think the lion's share should go to the amazing source material. One final thing I'll note before diving into today's topic is that for those interested in helping to support the podcast, I recently posted a bonus episode on a topic that fits quite well with Gibson's ideas. Demonstration and imitation as rate enhancers in skill acquisition.
For more information, please check out patreon.com forward slash perception action. Okay, now on to today's topic, Gibson's theory of affordances. If you recall from my last episode, Gibson's concept of invariance showed how, that despite an ever-changing retinal image, we can directly pick up information from our environment that is relevant to the control of actions. That is, we can pick up information that directly specifies physical quantities like time to contact, speed, distance, etc.
The question we want to tackle today is, what exactly do we do with this information? In particular, how do we give it meaning relative to the goal action we're intending to perform? Imagine you're a baseball infielder, and you pick up invariant information from the optic array, which tells you that a ball that was just hit is going to pass by three steps to your left in two seconds. What exactly does this mean? What should you do?
How do we get from direction of motion and depth and time to passage, that is perceptual information, to running and catching, that is action? If our perceptual system's primary job was to pick up information about the physical properties of the world, like size, speed, time, and distance, we would be stuck here, because such properties have no meaning in and of themselves. To get unstuck, again, this would seem like a place where we need to bring in higher-order cognitive processing to the rescue.
An obvious answer to this meaning problem would seem to be that our brain takes in the perceptual information specifying the physical properties, and then does some sort of inference in order to interpret what it means and decide which action to perform. For my infielder example, this cognitive processing might involve using memory stores of previous plays as kind of template matching. So, for example, a player might recall from previous experience that a ball hit with this particular combination of direction and time I mentioned above is one they have successfully fielded before, so they decide to go and play the ball. Or... Maybe their coach has taught them an if-x-then-y type rule they've stored in memory.
For a very skilled fielder, this might also involve some complex mental representation and internal model that takes into account the relative position of the bases and the fielder and predictions of ball trajectory, etc. Gibson did not like this idea of bringing in cognitive processing to the rescue at all, of course. But before discussing his alternative, I want to talk briefly about why Gibson was so anti-cognition.
For Gibson, appealing to internal models or mental representations was a completely unsatisfactory way of addressing the questions he was interested in, because in his view, all it does is displace the problem of controlling actions from the environment and move it into the brain. What are these mental representations comprised of? Where do they come from?
How exactly does deciding something like, I will play the ball, lead to the specific pattern of movement and coordination which allows you to play it? One of the reasons I think the idea of indirect perception is so popular and attractive is it has what I would call illusory parsimony. Need the organism to be able to do something?
Well, that's easy. We just take in the perceptual information related to a memory representation, make a decision, then generate an action. Wow, what a nice clear explanation.
We could even put a nice flow diagram up on the board with boxes and arrows for each of these steps. But have we really understood the control of action anymore by saying all of that? No, not really.
I have to admit that I was seduced by the appeal of such cognitive processes early in my career. If you look at some of my early papers, you will see complex diagrams showing all the complex processing that was done on the information coming in, then one tiny little box coming out. labeled something like generate action, where I didn't really understand what was going on at all. It was only when I started to try to understand how perceptual information was specifically used to control the action that I really understood what Gibson was getting at. Gibson was not denying that we don't ever use cognition and inference in life.
He and his ecological psychology followers just tried to explain as much as possible about the control of behavior without having to appeal to it. because it results in a more parsimonious and explanatory theory of perception and action. So, what exactly was Gibson's alternative account of adding meaning to the perception of our environment? It's quite simple, really. Gibson first noted that in our environment, surfaces afford, that is, they provide or supply, opportunities for action.
Flat surfaces provide opportunities to rest. Surfaces with gaps provide opportunities to pass through. Hanging surfaces provide opportunities to seek shelter, and so on.
Furthermore, as I discussed in the last episode, when looking at invariance, there is information in the structure of light, the optic array, which directly specifies the qualities of these surfaces. Given these two things, it is possible that what we perceive is what a surface affords, or in other words, what we can and cannot do with it, rather than just action-neutral properties like its distance, size, etc. In Gibson's words, quote, how do we go from surfaces to affordances? And if there's information in light for the perception of surfaces, is there information for the perception of what they afford?
Perhaps the composition and layout of surfaces constitute what they afford. If so, to perceive them is to perceive what they afford. This is a radical hypothesis, for it implies that the values and meanings of things in the environment can be directly perceived.
End quote. Gibson proposed that direct perception of affordances is possible because the patterns of stimulus energy, for example the optic array, are specific. That is, they lawfully relate to the environment.
Thus, the environment can be perceived without the process requiring cognitive mediation. Returning to my baseball infielder example, if the light reflecting off the hit ball returning to the fielder's eyes is lawfully related to action-relevant variables like time to passage and direction of motion, and these are lawfully related to opportunities for action afforded by the ball, for example, a hit ball with a very large direction relative to us is not catchable, why not just perceive what is afforded directly? Why bother designing a system that is only sensitive to action-neutral physical properties of the world through the pickup of invariant information, then add meaning later, when meaning can be directly derived from the information itself?
So in other words, Why perceive time to passage and direction first, then determine what it means, when the information used to get these values directly specifies what is afforded? If we scale or relate the information to our capacity for action, more on that in a moment, this tells us directly whether a ball that is going to pass by three steps to our left in two seconds is catchable. So what we perceive is the affordance of catchability, or not.
There is no need to first perceive the time to passage and direction. If the affordance of catchability is perceived in the situation, then we can use the very same sources of information to guide our movement to catch the ball. If it is not, then we change our goal.
For example, running to retrieve the ball in the outfield, or letting another fielder play the ball and covering a base, which then makes our perception action system attuned to different affordances and information sources. Let me give an example used by Michael Turvey to further illustrate the link between information and affordances. Imagine a bird flying towards a branch at velocity v. From physics, we can determine that the branch affords fracturing of the bird's legs when v is such that the rate of deceleration is too high.
Remember that surfaces in our environment afford both good and bad opportunities. How does the bird know whether the branch affords fracturing at its current velocity? Does it need to have an internal model of physics and do some complex calculation? No, the affordance of fracturability is optically specified. I talked about this back in episode 22 when looking at braking in driving.
Tau, as we know, is the angular size of an object divided by its rate of expansion. The first derivative of tau, its rate of change, is commonly called tau dot. If the animal is attuned to the information provided by tau dot, they can directly perceive the affordance.
If the value of tau dot is less than minus 0.5, there will be hard contact and the exchange in momentum is likely to be strong enough to lead to fracture. If it is above this value, contact will be safe. But the information is even more powerful than just this, because the same optical information that specifies the affordance of fracturability can also be used to control the action of landing itself. Specifically, if the bird couples the movements of its wings to the value of tau dot in a manner such that it keeps tau dot at exactly minus 0.5, it will land with zero velocity. What the environment affords is directly specified by the information and with no need for interpretation or cognitive processing.
Perceiving affordances is to carve up the world in meaningful units of action, rather than using meaningless units of physics. In Gibson's words, quote, the perceiving of an affordance is not a process of perceiving a value free physical object to which meaning is somehow added in a way that no one has been able to agree upon. It is a process of perceiving a value rich ecological object. Any substance, any surface, any layout has some affordance for benefit or injury to someone. Physics may be value-free, but ecology is not.
End quote. Trying to express this idea in one last way, in the context of sports, to a performer, gaps don't look wide, opponents don't look close, and pitches don't look fast. Instead, performers see passability in a gap, tackleability of an opponent, and hitability of a pitch. perception of opportunities for action, not just physical properties. At the end of today's episode, I'll talk about the implications of this idea for sports training, but next I want to look at the somewhat tricky issue of how affordances can be specific to an individual.
So far, we've talked about how information in the optic array directly specifies opportunity for action, affordances to a perceiver. However, it is obviously the case that affordances are not the same for everyone. We need to somehow take into account the action capabilities of the perceiver. For example, a baseball pitch with a particular tau value that affords hitting for a player with a high bat speed may not have the same affordance for someone with a lower bat speed.
Or, a regulation-sized soccer ball may afford kicking and dribbling for an adult, but not for a young child. As Gibson put it, quote, affordances have to be measured relative to the animal. They are unique for that animal.
They are not just abstract physical properties. They have unity relative to the posture and behavior of the animal being considered. End quote. In further describing this relationship, Gibson has a passage in one of his books that I think is very profound, but I have to admit I'm still trying to figure out what it all means. Here it goes.
Quote, an affordance is neither an objective property or a subjective property, or it is both if you like. An affordance cuts across the dichotomy of subjective-objective and helps us to understand its inadequacy. It is equally a fact of the environment and a fact of behavior.
It is both physical and cyclical, yet neither. An affordance points both ways, to the environment and the observer." Taking into account the action capabilities of the perceiver has actually been a bit of a sticking point in the theories that have sought to advance these ideas since Gibson. In particular, there's a heated debate between people that argue affordances are dispositional versus relational. Briefly, consistent with Gibson's thinking, Michael Turvey has argued that affordances are dispositional properties of the environment. That is, they exist independently of the animal. These are then complemented by dispositional properties of the animals themselves, turned effectivities. Quote, Whereas an affordance is a disposition of a particular surface layout, An effectivity is the complementing disposition of a particular animal. An effectivity, as the term suggests, is the causal propensity for an animal to affect or bring about a particular action. End quote. I will dive into effectivities a bit more next time when talking about some research that has examined these ideas. Alternatively, Thomas Stofgren has argued that affordances are emergent properties of an animal environment system and are not properties of the environment alone. Instead, according to Stofgen's perspective, affordances are undefined without jointly considering properties of both the animal and the environment. I don't want to go into too much under this because it gets pretty philosophical. If you're interested in learning more about this issue, I would highly recommend Andrew Wilson's recent post on his blog, Notes from Two Scientific Psychologists, where he describes how his thinking has evolved with regards to this issue. Link in the show notes. But the main point to take away is that affordances in some way need to be related to our action capabilities, both in terms of anatomical and skill. More on this in the next episode. Before wrapping up today's episode with some comments on what all this means for sports coaches, I want to highlight some key aspects of affordances I haven't touched on yet. First, perceiving affordances supports creativity in behavior, because it involves perceiving opportunities for action that are not necessarily the same as what a surface or object was designed for. Think about the affordance of sitability. When we really want to sit down, we do so on a multitude of objects that are not designed for that purpose. Tables, desks, counters, threesomes, etc. To perceive an affordance is not to classify an object into a predefined category. It is to detect opportunities for action based on information not necessarily dictated by the design. Of course, good design should take into account affordances, but that's a topic for another time. Second, affordances are dynamic. Opportunities for action come and go on a moment-to-moment basis. At one moment in a match, a teammate may be open, and a pass to the teammate is afforded. A fraction of a second later, a defender may step into the passing lane, and the pass is no longer possible. Action possibilities can evolve and devolve rapidly. They can also change over longer timescales. For example, when a fatigued player late in a match cannot accelerate quickly enough to catch a pass they would have caught earlier in the game. Excelling in this type of dynamic environment requires a performer to have good attunement between affordances and action capabilities through the process of calibration. I talked about this back in episode 120, and we'll discuss it a little more next time. Third, affordances allow us to prospectively control behavior. As illustrated in the bird landing example I talked about a few moments ago, one of the really attractive qualities of affordance theory for me is the fact that the same information on which decisions are made also supports execution of the behavior to achieve the goal. This is not the case for more indirect internal model approaches. Once we take in information to create an internal representation of the world, how then do we move to realize our goal? We need to invoke a completely separate motor control system. This is not the case with affordances, because like with all of ecological psychology, they capture the reciprocity of perception and action. That is the fact that perception and action are prepared. perpetually feeding each other. I discussed these issues back in episode 113 when looking at online versus internal model-based control. Finally, affordance theory does allow for the possibility of misperception or perceptual errors. As Gibson put it, if information is picked up, perception results. If misinformation is picked up, misperception results. As I mentioned in the first episode in the series, the fact that perception can be tricked, for example with illusions. is often used as evidence in support of indirect inference theories of perception. But if we examine the illusions, they can often equally be well understood in terms of the information present. The vast majority of illusions, for example, are deliberately constructed to present very unnatural and limited information. For example, they're typically two-dimensional, so don't provide the invariant information that comes from being allowed to explore one's environment. Okay, to sum up, in today's episode we looked at affordances. which are a way of perceiving meaning in the environment in terms of how an actor can behave in that environment. This allows an actor to control their actions with respect to behavioral possibilities in the environment, not merely using the action-neutral possibilities of physics. Stated another way, affordances provide functional semantics for understanding the control of action. So what does all this mean for coaching? Well, I think using affordance theory to understand decision-making in sports. presents a very attractive option for coaches. Think about a soccer player dribbling the ball up the field. The player could shoot, pass to an open teammate, pass to a less open teammate but one who is closer to the goal, continue to dribble upfield, etc. If we choose to understand such decisions in terms of affordances, that is using the information in the environment, and effectivities, the athlete's abilities, these are both things we can directly influence through practice design and training. Information can be manipulated by changing constraints, for example, for example changing the number of players, while action capabilities can be altered by fatiguing the athlete or strength and conditioning in a multitude of other ways. For me, this presents a much more attractive option to a coach than trying to infer the mental processes going on inside a player's head and add to these by giving indirect rules for making decisions. Also, as I've said a few times today, if you look at decision making in sports in terms of affordances, you get control for free. That is, if you design practice so that players learn whether or not to pass or hold on to the ball based on tau-like information. you have educated their attention to the very same information they need to control the action itself. For example, the rate of closure of a gap between players both specifies the affordance of passability and is the control variable you need to regulate the speed at which you pass the ball. Finally, thinking about practice design in terms of what each activity affords can be an effective way to ensure that the task difficulty is set at an appropriate level for practice. If you think about the distance between players or the height of a barrier in a practice activity in terms of passability and shootability rather than in feet and meters, then it helps to take into account the skill level and or age of the athletes you're working with. I'll get into a few more specific ideas for coaching next time when I look at research investigating Gibson's ideas and extension of it, particularly the information-based regulation of movement. Okay, that's it for today's episode. Remember, you can contact me at robgray at asu.edu or follow me on Twitter at shakyweights. To find out more about the podcast, please check out perceptionaction.com. Finally, to support the podcast and receive bonus materials, including an extra monthly episode and written transcripts, please head over to patreon.com forward slash perceptionaction. This is Rob Gray from ASU. Cheers for now and keep them coupled. We'll cut you quick.