Tuesday, September 21, 2010

Gibson, Runeson, Physics and Affordances

Andrew sometimes comments that ecological psychology has moved along, in some respects, since Gibson.  And Chemero argues that for a diversity of interpretations of affordances.  (Chemero manuscript here.)  Actually, the other papers in that issue of Ecological Psychology probably attest to a diversity of opinion as well.

So, once we get to the point where we recognize that Ecological Psychology is not a single, monolithic body of theory, we can move on to spelling out the differences.  And, some of those differences might concern how physical and geometrical quantities are to be used, or not, in Ecological Psychology.


  1. You're still talking as if there is some sort of ban in place on these things. There isn't - you are wrong about this. The only point Gibson and Turvey were making was that the correct units of analysis for psychology are not necessarily those of Cartesian physics, hardly a controversial claim. This reading of these papers is very weird.

  2. I think that Turvey's point is that agents don't need to perceive points, distances between points, etc. So, that bit of text was probably not best.

    On the other hand, the thing about Gibson that sticks in my mind is his decision to avoid talking about the concepts of physics, which is not what Runeson does.

  3. Gibson is making a critical point. People assume that physics is describing an objective state of affairs, the 'truth' about the world that a perceiving organism will need to access in order to be functional. A scientific account of perception should, by this account, establish how we acquire knowledge about this state of affairs.

    But Gibson has a different ontology in mind. While there is indeed a physical reality out there, the description of it from physics is not the only way to carve that up. Organisms need information about affordances - not 'that object is distance x from me', but 'that object is reachable by me'. The concepts of physics can be computationally converted to information, but Gibson's hypothesis is about the direct perception of these properties expressed with reference to the organism.

    Physics and it's laws are at the heart of how information comes to specify these 'world' properties; but perception is about information, which is animal-scaled.

    Gibson's not avoiding anything. One of the things I love about his writing is that he attempts to engage with everything that's relevant; the 79 book is the culmination of a career arguing with people about this stuff. But he's simply denying that the ontology of physics is the right level of analysis for a theory of perception, and laying out why he thinks this.

  4. "he's simply denying that the ontology of physics is the right level of analysis for a theory of perception, and laying out why he thinks this. "

    Yes, but it seems to me that Runeson is violating this by using physics to explain certain features of perception, namely, why we see a distorted room as square.

  5. Physics is clearly part of the game; the ecological, informational level is structured by physical properties. But behaviour isn't necessarily organised with respect to (e.g.) size, or weight, but maybe 'heaviness', which would be a behaviourally relevant variable expressed with respect to some organism property,