Friday 9 April 2010

Reading Group - Gibson (1979) Chapter 2

In the previous chapter, Gibson established the importance of the environment to perception. In this chapter, he goes into more detail about the properties of the environment that are relevant to perceivers. Gibson argues that although classical physics talks about the world in terms of bodies in space, this doesn't mean that this is a good way to talk about the environment as it applies to perception. In studying perception, it’s better to think about our environment in terms of mediums, substances, and surfaces. Two things occurred to me as I read this chapter. The first was, “Wow, I’ve never before heard the environment described in this way.” The second was, “Wow, this is so unbelievably straightforward and comprehensive, why don’t all psychologists talk about the environment in this way?”

Chapter 2

The medium

A surface occurs wherever two states of matter (e.g. gas & solid) meet. The majority of relevant surfaces to humans occur when gaseous air meets solid ground. Animals can move through gases and liquids, so these afford locomotion. They are mediums for locomotion. Light can move through gases and liquids, but light bounces off solids. Mediums have to both transmit and reverberate light. Reverberation causes illumination, which fills a medium with ambient light. Sound, as well as light, can move through gases and liquids. This allows us to hear sounds (be sensitive to vibrations). Volatile or soluble molecules quickly diffuse through gases and liquids, allowing us to smell (to detect a source at a distance).

“Animal locomotion is not usually aimless but is guided or controlled – by light if the animal can see, by sound if the animal can hear, by odour if the animal can smell...The medium thus contains information about things that reflect light, vibrate, or are volatile. By detecting this information, the animal guides and controls locomotion.” p. 17

As an animal moves around, the sights, sounds, and smells change. Each point of observation is unique.
Mediums are relatively homogeneous. The composition of air has been pretty constant for millennia, and the proportion of oxygen in it is pretty constant around the world. Water has somewhat more variability (e.g., saltwater differs from fresh water), but the transitions between types of water is fairly gradual. This homogeneity means that animals can breathe. It also means that waves emanating from a source do so with regularity. Mediums have absolute vertical reference points. Gravity pulls down. Sunlight comes from above. These properties (permitting breathing, locomotion, illumination, detection of vibrations, and detection of diffusing molecules) are invariant. Gibson refers to these opportunities as affordances.


Substances do not transmit light or odour, and don’t afford locomotion. Substances are usually rigid. Compared to mediums, they are heterogeneous. Substances are usually aggregates of different chemical compounds. Animals have to be able to distinguish between substances. For instance, animals have to be able to detect whether something is edible or inedible. It’s preferable to be able to detect this at a distance (e.g., by using vision).

Substances differ in terms of physical properties (e.g., hardness, viscosity), but this is not necessarily the level at which animals distinguish between substances. Substances also differ in terms of the degree to which they reflect light. Many substances, for all intents and purposes, do not change. This means that substances in the environment are often dependable entities. For terrestrial animals, water is a substance. For aquatic animals, water is a medium.


If a substance persists, then the surface interfacing it with the medium will persist. Ambient light makes surfaces potentially visible. All the cool stuff happens on surfaces. Light is absorbed, touch occurs, diffusion occurs. P. 23 lays out the “ecological laws of surfaces”:

1) Persistent substances have surfaces, which have layouts

2) Surfaces resist deformation. Substances that deform easily might not permit locomotion to heavy animals like us. We can move some substances, but not others, out of the way.

3) Surfaces resist disintegration. We can smush clay, but we have to cut wood.

4) Surfaces have both layout textures and pigment textures. Surface texture influences optical texture.

5) Surfaces have shapes

6) Surfaces may be in light or in shade

7) Surfaces may absorb a little or a lot of light

8) Surfaces may reflect a little or a lot of light

9) Surfaces have particular reflectance ratios of different wavelengths of light (i.e., different colours)

The level at which Gibson characterises the environment is very different from the level at which it is described by physics. Instead of talking about bodies in space, Gibson talks about mediums, through which animals can move, subtances, which do not afford locomotion and which are relatively stable, and surfaces, which permit detection of structure in light, vibrations, or volatile molecules. This characterisation describes the world with respect to perceivers. Important aspects of the environment depend on whether we (or some other animal) can detect them. For instance, ultraviolet and infrared are possible colours, as defined by the world of physics, but these are not relevant properties of the environment for humans, who are unable to see light with wavelengths outside a narrow band. Like I said, it's straightforward and comprehensive.


  1. It's funny how this sounds like it comes from a child's "first book of what's in the world". And depressing that it's not actually.

  2. Thanks so much for this clear and lucid commentary on the book. I am also amazed that what Gibson describes is not part of our very early conversation about the world. It is certainly a conversation that it implicit in the behavior of all of us - particularly infants and children as they (we) explore surfaces and mediums with such avid and responsive persistence. For me the exciting part of this body of work is the emerging thinking on how we as bodies with consciousness are also surfaces and substances to ourselves and that this reflexivity makes the exploration of the world that much more complex and interesting. I can identify with a substance and it then appears (according to Mead) as an 'object'. I am wondering if 'substance' and 'object' are the same for Gibson?

  3. Object and substance are divisions of no real interest for Gibson; they are categories of physics, not psychology. Gibson's claims are actually ontological - he's proposing that the way things really are, taken from the perspective of a perceiving organism, is that there are affordances (and events, depending on where you fall on that argument). Things are 'the same' to the extent that they contain the same information.

    So this distinction (identifying with a substance that then appears as an object) sounds like it entails exactly the kind of dualism Gibson was dodging by suggesting that the 'objects of perception' aren't 'objects' (things) but 'affordances and events' (information). Does that sound too vague? The specification constraints for information do nail this down more, I think, so there's more to come from us on that :)

  4. I'm not sure there's a necessary dualism here. I'm new to this material, but I would guess that one could treat 'substance' as the superordinate category and 'object' as one basic type of substance. For example, the ground is a substance that would defy what we typically mean by 'object', as are walls and ceilings. These substances have surfaces that constrain our movement through the environment, but they are not clearly bounded, or potentially movable, as objects tend to be. Not to say that objects must be manipulable, but an obvious boundary might be a prerequisite. In terms of what Andrew is saying, it seems like the intuitive differences between what we typically refer to as objects and other substances (like the ground), would map onto reliable differences in the affordances. So, objects might be substances that afford being picked up, or somehow moved (we can move them with respect to us), whereas non-object substances might afford our moving with respect to them (we walk over the ground, around walls, etc).

  5. This comment has been removed by the author.

  6. I'm just getting into chapter 3 and Gibson distinguishes between the ground, enclosures, and objects in way that maps onto what I was saying above (although much more clearly). So, I'll post about this soon.

  7. One obvious question pops out of that: why bother talking about objects and surfaces if what you actually have is all just affordances?

  8. This comment has been removed by the author.

  9. Talking about it this way just defines another class of affordances in the same way that Gibson already distinguishes between mediums and substances in terms of their affordances (mediums afford being moved through, substances do not). Now he's saying that within substances, some things are objects.

    Gibson finds 'object' to be a useful category in the environment and I agree. Instead of saying "this thing is a surface that completely surrounds its substance and is not continuous with any other surface" we can summarise that information and call it an object. This makes an object a very different type of thing than, say, the ground. The differences between objects and substances like the ground are defined by the affordances, but using a specific term usefully identifies a class of thing with common ecological properties.

  10. To once again play Devil's advocate, are affordances not 'in the eye of the beholder'? And while this is fine on an individual to individual level, how does this work on an evolutionary scale? How can things not afford the same action same action for everyone?

  11. Affordances are personal - a thing that affords sitting to me, probably will not afford sitting for a toddler. Or, a surface that a mouse perceives as sufficient to bear its weight, may not be sufficient to support at cat. But, importantly, an environment will afford similar things to groups of similar organisms. I can offer a guest in my home a seat and be reasonably sure that he or she will be able to sit on it. To the extent that there is variation, there is room to take advantage of an ecological niche, which may end up pushing the population to evolve. Mutations that cause changes in an organisms physical structure, may change what the environment affords that organism. And, these changes might be good or they might be bad. So, it seems to me that there is sufficient stability within a group so that affordances are relatively predictable within a species, but there is some variability to allow for adaptive change in behaviour.

  12. Interestingly the biological concept of 'niche' and the ecological concept of 'affordance' are, indeed, deeply connected just as Sabrina described, and this is exactly as powerful evolutionarily as she makes out :)

    Remember, Gibson is all about naturalising perception, ie making it consonant with how we know the natural world works (specifically biologically).