Thursday 15 April 2010

Reading Group - Gibson (1979) Chapter 3

Chapter 3

In which Gibson defines everything. See here and here for a refresher.

Here’s the crux of it:

“The world of physical reality does not consist of meaningful things. The world of ecological reality, as I have been trying to describe it, does. If what we perceived were the entities of physics and mathematics, meanings would have to be imposed on them. But if what we perceive are the entities of environmental science their meanings can be discovered” (p. 33).

Gibson goes on to define a number of useful terms for a theory of surface layout.

Ground – This refers to the surface of the earth. It is generally level and horizontal. It is a reference point for all other surfaces. It is perpendicular to the force of gravity.

Open Environment – This would be a layout of the ground only (e.g., in an open & level desert). Layouts tend be more wrinkled and cluttered than a perfect open environment.

Enclosure – A surface layout that surrounds the medium. The surfaces are inward-facing.

Detached Object – A surface layout that is completely surrounded by the medium. The surfaces are outward-facing. It must be possible to move the object without breaking any surface.

Attached Object – A surface layout that is partially surrounded by the medium. The substance of an attached object is continuous with another substance so that moving it would require breaking a surface. This could
just be a convexity.

Objects can be counted while substances cannot.

Partial Enclosure – A surface-layout that is partially enclosed by the medium. It might be only a concavity, but it could be a shelter.

Hollow Object – From the outside, it is an object, but from the inside it’s an enclosure. Thus, some of the surfaces face outward, and some face inward.

Place – A location in an environment with an extended surface layout. Places can be located in other places (the hallway by the kitchen) and they don’t need to have clear boundaries. They can be named.

Sheet – Two parallel surfaces enclosing a substance (e.g., biological membranes).

Fissure – Two parallel surfaces enclosing a medium, where those surfaces a close to one another.

Stick – Elongated object

Fiber – A very think elongated object (e.g., thread).

Dihedral – The junction of two flat surfaces that either enclose the medium (a corner) or enclose the substance (an edge).

Curved convexity – A curved surface that encloses (or partially encloses) a substance.

Curved concavity – A curved surface that encloses (or partially encloses) a medium.

Environments contain lots of things (e.g., surfaces, substances, enclosures, etc) but only some of these will be perceptible to a given animal. For humans, the terrain is usually cluttered with stuff (e.g., obstacles, barriers), which means that we can locomote only where there’s an opening in the terrain. We can say that the terrain affords locomotion at openings. Gibson provides precise definitions for all types of things that can get in the way of locomotion (p. 36-37). Two properties that we might not think of as obstacles are steps and slopes. Whether a step affords locomotion depends on the size of the step relative to the size of the animal. Whether a slope affords locomotion depends on the steepness of the angle with the ground and the texture of the terrain.

Thinking about the properties of objects tells us how we can interact with them. If an object is the right size, we can grasp it. Attached objects, such as tree branches, are things to hold on to that permit swinging and climbing. Detached objects, such as rocks, are things we can hold and potentially carry or throw. Tools are special types of detached objects. When we use tools, it’s as if they are a part of the body, rather than something in the environment. This isn’t meant trivially; for all intents and purposes, tools really do become part of the user’s body. Clothing is the same sort of thing. This means that the boundary between self and environment is flexible and that the duality between self and non-self is false. At this point I imagine Gibson saying “Philosophical crisis solved. You’re welcome” in the style of John Hodgman on the Daily Show.

Animals are also types of detached objects. Animals have the special property of moving spontaneously and these movements are highly specific to particular animals. According to Gibson, animals “are the most complex objects of perception that the environment presents to an observer” (p. 41). There are many things an animal might be in relation to us - predator, prey, and relative, – and so many states an animal might be in – asleep, hungry, and awake. So, animals can afford lots of different things – avoiding, eating, and nurturing.

Displays are detached objects that have special significance for humans. Generally, a display is any surface that has been modified to convey information about more than itself. Displays include solid images (e.g., sculptures), pictures, and anything you can write on.

Finally, Gibson clarifies an issue about the personal nature of our environment. Only one observer can occupy a given place at a given time. It has been argued that we each have a personal experience of the environment that we can never really share with anyone else. Gibson rejects this view. He emphasises the fact that animals move through environments and revisit the same places again and again. So, animals’ cumulative experiences are shared and highly similar.

This is the end of Part 1. Stay tuned for Part 2, which looks at the information we use for visual perception. Hint - it's not images.


  1. Rejecting dualisms is at the heart of Gibson (I swear I'll post about Heft on James tonight); but the tools stuff I like a lot because its well researched.

    The dynamic touch literature is all about how we detect properties such as length, etc of objects. The answer is not 'by detecting length' but 'via the object's inertia' which is much more closely aligned to the affordances of the objects - turns out we don't care about it's length, we care about how we can wield it.

    Even better, it turns out we perceive the same kinds of properties about our own bodies in the exact same way - so one (non-visual) way I know where my hand is in space is by swinging my arm and perceiving the inertial characteristics. Adding a tool just makes my arm longer, figuratively, literally, and perceptually!

    Dynamic touch is an incomplete answer because there's no good answer to how the inertia tensor (a dynamic entity, ie it entails mass) is specified in information (which is only about motion; kinematic), and it falls down in a few places (eg in the size weight illusion literature). But it's still a great example of how information is everything - if you pick up a tool, you literally act as if your arm changed length and act on the resulting, altered affordances. Too cool for school :)

  2. The size weight illusion is an interesting example. You can experience it through dynamic touch information (it can be induced through different torques when lifting). You can also experience it without dynamic touch information. But alone, these variants of the size weight illusion are not as strong at combining the dynamic touch and (lets for the sake of argument call it) cognitive versions of the illusions. In this illusion at least, it is clearly an additive combination of factors - some ecological and some representational.

    Is this 'ecological' vs 'representational' all or nothing approach not doomed to eventual failure. Seems there is enough oomph on both sides of the argument that day to day perception is a combination/clever interaction of/between these two styles.

  3. Explain 'you can also experience it without dynamic touch information'. I've also never heard any talk about things additively creating the size-weight illusion.

    I don't believe this all or nothing approach is doomed to failure. Ecological psychology may not entirely succeed, but you'll never get a viable 'representational ecological psychology' because the two ideas are simply too different. It would be like trying to actually make a 'colourless green idea sleep furiously'; it makes no sense.

    Now - we might be wrong. But that's a whole other issue.

  4. I'm not sure the additive aspect has been directly tested, but you experience weaker size weight illusions when they are placed into you hands to be perceived passively (accounting for pressure per square inch - I think Charpentier himself showed this), or by haptics alone without vision or vision alone without haptics (older stuff my Lederman), or the same object can be made to feel different weights as a function of what the lifter is seeing (Koseloff) and/or what they think are lifting (some of my stuff).