Sunday 1 May 2011

Reading Group - Gibson (1979) Chapter 6 Part 2

So, it's been four months, but here's part 2 of Gibson '79 chapter six:

Gibson believes that events are specified in the optic array, just as stationary objects are. In this part of the chapter he describes three classes of events (changes of surface layout, changes of surface colour or texture, and changes in the existence of a surface). He goes on to discuss the types of optical disturbances that can specify these events.


An important thread in this chapter is the dissimilarity between motion in the world and changes in the optic array. Gibson wants to do away with the notion that we perceive 2D projections of a 3D world. His solution to this problem is to explain how events are specified by changes in the optic array. Thus, whereas the old view involving 2D projections assumes that we can infer the meaning of events by virtue of the similarity between what is perceived and what is happening in the world, Gibson argues that we can directly perceive the meaning of events by virtue of the specification relationship between perceptual information and events in the world.

The optical information for perceiving events

Events create disturbances in the ambient optic array. Different types of disturbances specify different types of events. It’s important to recognise that these disturbances aren’t any kind of literal copy of the motion of objects in the world. Changes in optical structure aren’t going to look anything like changes in physical motion in the environment.

Mechanical events

Imagine an object moving in a rigid translation with respect to an observer  - the object moves from one place to another, but its distance to the observer is constant. The easy way to think about this is a figure moving against ground. But, this isn’t what’s happening in the optics, so it can’t be the basis of event perception. There is, however, progressive deletion of the occluded surface at the leading edge of the moving object and progressive accretion of the occluded surface at the trailing edge (remember this? if not, see Chapter 5 Part 2). Thus, events are changes in structure, not simply motions. Most events are not rigid translations, they involve something either coming closer or moving further away from an observer. When an object approaches, it progressively deletes more of the background and when and object recedes it gradually reveals more of the background. When the visual solid angle of an object approaching an observer reaches 180 degrees, then the object has come in contact with the observer. When a spherical object rotates about a central axis there is no deletion or revelation of the background. However, the motion is still specified in the optic array by the shearing of the object’s texture created by the spin. The big message Gibson wants to convey is that the perception of motion does not involve a 2D projection of a 3D physical motion. The sticking point for this idea seems to be that it can’t cope with our perception that occluded things still exist. He says, “Whatever the perception of space may be, it is not simply the perception of the dimension of depth” (p. 105).

Chemical events

When substances are inert they will maintain the same colour and reflectance. But, chemically active surfaces change colour and texture. For example, we blush when we are embarrassed and blackberries turn dark when they are ready to eat. Persistence of colour is actually an interesting problem for Gibson. He knows that there must be some invariant in the optic array that specifies a persistent surface colour, but he doesn’t know what this might be. One obvious issue is distinguishing black from dark and white from light. A probable solution to this issue is that dark and light are connected with moving sources of illumination.

Destruction and creation of surfaces

Any visual solid angle in the array that is structured specifies a surface and any visual solid angle that is unstructured specifies a hole. In the upper hemisphere of the visual field, clouds form structured surfaces. When a cloud materialises, its structured surface gradually replaces the unstructured sky in that region. On earth, evaporating water gradually recedes to reveal the structured solid surface beneath it. Dissipation, evaporation, sublimation, dissolution, disintegration, and decay are all versions of one surface replacing another.

The kinds of disturbance of optical structure

Now that we’ve characterised the way that many events are specified in the optic array, it’s worth thinking about how to treat these mathematically. Some of the events described preserve a one-to-one mapping of units over time (e.g., the shearing of texture during rotation), but some of them do not (e.g., deletion of texture during occlusion). An important point to emphasise here is that changes in the optic array bear no resemblance to the events they specify. Optical motion isn’t at all like physical motion. This is because optical motion does not consist of moving physical bodies with mass and inertia. However, optical disturbances will correspond in time to an event in the environment. The beginning of the event will be the same time as the beginning of the optical disturbance, and so on.

The causation of events

A simple example of causation is a collision between objects where one object hits another, causing it to move. Whether or not we can perceive causation is an old philosophical debate. Gibson hints at evidence that suggests we can perceive causation, but will take up the discussion again in chapter 10.

Summary

Gibson believes that events are specified in the optic array, just as stationary objects are. He classified events into three types: changes of surface layout, changes of surface colour or texture, and changes in the existence of a surface. The types of optical disturbances that can specify events are things like deletion/accretion, shearing, magnification/minification, deformation, and substitution (among others).

The chapter ends with a charming sentence in which Gibson makes the point that although the mathematics behind these optical disturbances is just beginning to be investigated, “nevertheless, strange to say, they are what we are visually most sensitive to, all of us, animals, babies, men, women, and moviegoers” (p. 110).

1 comment: