Lecture 23: Ecological Optics Primer (Turvey, 2019, Lectures on Perception)

The 'barrage of conceptual details that was Lecture 22' has provided us with the notion of affordances, and these are to be the objects of both perception and action. This means the need to be perceived, and this brings us to the second part of the ecological ontology - information. This chapter will focus specifically on optical information, identify the key characteristics of an optic array, and walk through some ways of analysing perception ecologically by dealing with some problems supposedly raised by illusions. 

The first section is about the retinal image. For a long time, this has been seen as a) the starting point for vision and b) a major block to the possibility of visual perception being direct. It is, of course neither; the notion of a retinal image is effectively a conceptual error based on being able to sort of see an image projected onto it in dead, dissected eyes. The retinal image is simply 'a careless use of words'. 

Turvey then identifies 4 different kinds of perceiving that a perceptual system needs to be able to do. These are

• exteroperceiving (perceiving the layout of surrounding surfaces)
• proprioperceiving (perceiving the layout of one's body)
• pro-exteroperceiving (perceiving the layout of surrounding surfaces relative to one's body)
• ex-proprioperceiving (perceiving the layout of one's body relative to the layout of surrounding surfaces)

He discusses how each of these is affected by using prisms, which are traditionally understood to shift or invert a retinal image but which clearly does much more as each of these modes is affected.

He now lays out some key structural features of the ambient light in an environment after it has reflected off the surfaces. Specifically, what each point of observation provides is a unique and specifying collection of projections of the world into optics. That projection is not simply into light, or photons; but into solid visual angles. Each of these has as it's base a 2D surface, with the edges of that surface demarcating the boundaries of the visual angles. There are very many of these, nested amongst each other at multiple scales (i.e. multifractal) and informing about surface layout. Vision is specific to this information; what we see is what the optics specify. This then leads us to reconsider some illusion cases.

One type of illusion is how when a straight stick is placed half in water it looks bent. But this is not a mistake; it is the only possible thing you could see, given how light behaves in different media (water vs air). What about other 'mistakes', such as when an shark mistakes the electric field of a machine placed their by a researcher for the electric field of it's prey? Hardly a mistake; the mistake would be to not investigate (Turvey et al, 1981). So errors such as these, made because the information isn't a copy of the world, are perfectly coherent ecologically; the information isn't a copy, and the behaviour tracks the information!

What about illusions such as the Muller-Lyer? People judge the lines to be different lengths in different contexts, even though the line is the same length. Well, the problem is that it is only the same length when measured by a ruler, and eyes are not the same kind of measurement device as a ruler (which is shown by the fact it produces different lengths under different contexts). Assuming that the ruler is the actual 'truth' is a mistake made by the scientist. The Ames Room is another example of this; when actually the varying perceptions of the room as one moves around make complete sense given how the room structures light at different places. Turvey points to one of my favourite papers where Runeson analyses the Room ecologically (which I blogged here). 

Turvey also discusses the Necker cube, and notes that the experience of flipping between versions of the cube is a function of ambiguity in information, and not a mental trick of flipping between interpretations. Again, the mistake comes from misunderstanding what the data is telling us about what kind of measurement system an eye is. 

These examples provide a useful map for how to apply the principles of ecological optics to the question of 'what was perceived and why'.