Friday, October 15, 2010

Runeson's objection to starting with physics

Now, if the theory of physics cannot be claimed to have monopoly on descriptions of "what is really there" , there is no longer any reason to assume that the perceptual systems must necessarily begin by registering what is basic to physics. On the contrary, we should expect perceptual mechanisms which directly register variables of high informational value to the perceiver. (Runeson, 1977, p. 173).
This seems to me a rash argument.  Grant that physics does not have a monopoly on reality.  That there are things, such as say trees, that are not a part of physics narrowly construed.  There still might be reason to believe (forget about "assuming") that perceptual systems begin by registering (set aside the "must necessarily") what is basic to physics (set aside quarks and think of photons).  I mean, Runeson seems to assume that something like reductionism to physics is the only possible reason to think that vision science should begin with entities of physics.

So, for example, why couldn't there be some experimental result is psychology or neuroscience that supports the view that perceptual systems begin by registering what is basic to physics?  By this, I take it that Runeson is not merely challenging the idea that vision science should begin with fundamental physical entities, such as quarks.  I'm assuming that he would not like a vision science that begins with, say, photons.  Why couldn't there be an experiment that shows that people respond to flashes of light or patterns of light.  That really is the mainstream view of vision.  EP folks think this; they just disagree with it.  

Runeson, S. (2008).  "On the possibility of "smart" perceptual mechanisms".  Scandinavian Journal of Psychology, Volume 18, Issue 1, pages 172–179, September 1977


  1. It's not trees that are the interesting counter-example; it's affordances. If affordances are what there is to be perceived, then you get license to start wondering how you might get to these. This then points back to Gibson's analysis of the information available about affordance properties, and you then ask whether organisms behave as if they are perceiving that information.

    The empirical disentangling of all this is an interesting problem Sabrina and I are thinking about just now. Part of the problem is that if you ask a retina to transduce a photon of light, you will get a response; but you can also get the same sort of response with mechanical pressure. The issue is, what sort of questions do you ask the visual system to answer? This is the role of theory in science, and Gibsonian theory suggests one set of questions. Runeson is then suggesting a way to think about what sort of device might respond to those questions.

  2. But, affordances are (apparently) what Runeson wants and what I am resisting.

    Runeson is offering a false dichotomy here. He suggests that either one perceives photons (let's say) or one perceives affordances. (This is essentially the same false dichotomy that Gibson, 1979, offers circa pp. 54-55). He then proposes that there is more than just photons (i.e., more than just physics), so let's go with affordances.

    I'm not asking you to abandon affordances; I am, however, asking you to abandon the argument Runeson gives here for affordances.

    Take out all of Runeson's hedging and look at what he is saying: "Once we reject reduction of everything to physics, there is no reason to believe that the perceptual systems registers photons" There's *nothing* in biology, neuroscience, or psychology that gives any reason to think perception begins with photon capture? Surely that is overstated.

  3. It's not a false dichotomy; it's laying out the two extant hypotheses. Sensation based theories assume that the job of perception is to bring your mental state into correspondence with the state of the world, as described by the language of physics. This is Fechner and traditional psychophysics. They also claim that the raw input to this process are sensations, i.e. meaningless photons/retinal images from which that correspondence must be recreated. Gibson's problem was that such a Cartesian framework is doomed to failure - you simply can't get to the end product you want from such impoverished input. So he reconsidered that assumption, identified a plausible alternative (affordances and ecological optics) and moved on from there.

    Runeson's smart mechanism is a counter to the initial complaint people had about Gibsonian information: in order to get to these higher order invariants, they claimed, you had to first detect the components of that higher order variable and then transform those into the other. That's not direct perception, and thus that idea doesn't work. The planimeter, however, is an example of a device that directly measures such a higher order property without any access to the 'simple' variable; simple is relative to the measurement system.

    So Runeson isn't arguing for affordances, per se. He's suggesting a way in which higher order information for affordances might get detected in the way Gibson advocated - it's a proof of concept piece. Smartness exists; it enables direct detection of 'higher order' variables; if vision is smart, then Gibson's theory is in principle implementable; so it's game on for a perceptual theory that doesn't need to recreate the world of physics inside the head from lower order variables.

    So he's not quite offering the argument you suggest: it's not 'physics or affordances', it's 'the descriptions of physics are deeply problematic for perception, but luckily that level isn't compulsory, which means you might expect a perceptual system to be able to operate without such deeply problematic descriptions, and with something more useful' .

  4. It is a false dichotomy. You can have a theory that one perceives light as such. You can have a theory one perceives affordances as such. And, you can have a theory that one perceives middle-level basic category objects as such. Runeson's hasty move is to go from a bad way to get "not photons" to affordances or "variables of high informational value to the perceiver", whatever they are. (I don't see that Runeson ever explicates that.)

    And, while Runeson does entitle his paper in such a way as to make it seem like he's only arguing for a "possibility" or "proof of concept", he does get beyond that to say that 'smart' mechanisms are likely. But, I'll come to that ...

  5. Ken,

    Perhaps Heidegger's notion of the "as-structure" is relevant here. I would say that animals perceive photons *as* meaningful, *as* affording certain possibilities. Why? Because the ambient array of light-beams in the environment is informationally rich. Learning how to differentially respond to that array is highly adaptive because the discriminated information is functionally relevant to the goals of the animals. It is important to note that the perception of possibility isn't restricted to just the possibilities of our own movement. We can perceive the possibilities of objects in the environment, e.g., that a boulder is about to fall or that someone is about to become angry.

    I think we could plausibly take a "one world" approach where both physicists and Gibsonsians acknowledge a real world that exists independently from us but also acknowledge that we can take any number of perspectives on it depending on our explanatory target. Physicists are going to investigate reality differently than biosemioticians. If you really want to investigate the theoretical background of affordance theory, biosemiotics is the place to go.

    The ecological psychologist looks at reality in terms of how it is vitally significant to an animal. Depending on the animal, certain features or properties of the environment are valenced as relevant, useful, valuable, threatening, etc. We see the photons *as* significant, or *as* meaningful, as a function of our intrinsic goals and needs. If I am hungry the photons reflecting off a cheeseburger are valenced as attractive; if I am satiated, the cheeseburger photons are valenced differently. This can be cashed out in terms of what's now call "attention-salience" models of behavior. They have done nice psychological work with drug addicts show how these salience-networks operate in terms of how we attend to reality. The sign-structure of reality is dependent on the specific functional fit between animal and environment. Recognizing this fact doesn't entail a rejection of realism about scientific entities.

  6. It seems to me that you are challenging Runeson if you say that animals perceive photons *as* meaningful. It seems to me that he is attempting to undermine the idea that perception begins with photon capture, but you are saying it is ok to have perception beginning with photon capture, since photons afford things.

    But, there are possibilities that have to be sussed out here. There is the matter of whether we perceive photons, affordances, or objects, then whether we perceive these directly or not.

    The Gibson/Runeson view is that we perceive affordances and that we perceive them directly. What I take to be, say, the Pylyshyn view is that we perceive middle-sized objects, like dogs and trees, and that we perceive them indirectly by a computational mechanism.

    All I'm doing in this post is trying to map out the space of possibilities so that one has a better sense of what needs additional argumentation.

  7. Indeed, it seems to me to run counter to Gibson as well to suggest that we see photons. That seems to be what he is rejecting at about p. 54 of Gibson, 1979.

  8. I didn't say Gibson thought we see photons. We see photons *as* meaningful. This means that we perceive the meaning directly, not the photons. We sense the photons, but we perceive the affordance of the collection of photons that make up the ambient array. In other words, we perceive the stimulus information, or affordance, not the physical information. This is why perception is "direct" for Gibson. Put another way, we perceive the molar level of reality, not the molecular. We pay attention to the abstract invariant structure, not the molecular flux.

  9. but we perceive the affordance of the collection of photons that make up the ambient array
    You need to be a little careful with terminology here. The affordance is a property of the world for which there is specifying information in the optic array. It's important to keep world and optical variables in the right places.

  10. "I didn't say Gibson thought we see photons. We see photons *as* meaningful"

    Your first sentence is true, but I take it that if you say we see photons *as* meaningful, this entails that we see photons.

    I'm thinking of how the seeing versus seeing-as distinction works in the recent Analytic tradition. Now, I could be wrong about this entailment in this tradition. Or, I could be wrong about what you mean by seeing *as*. You go link this to Heidegger and I know essentially nothing about Heidegger.

    And, I take it that the preceding story holds as well for perceiving versus perceiving-as.

  11. "but we perceive the affordance of the collection of photons that make up the ambient array
    You need to be a little careful with terminology here. The affordance is a property of the world for which there is specifying information in the optic array. It's important to keep world and optical variables in the right places."

    Andrew, I was a bit surprised by Gary's claim here too. I read him, however, as expressing his own view, rather than presenting Gibson's view. So, that would mean that Gary is not using the terminology differently, he just has a different view of photons. Photons, like chairs, afford things.

    *My* worry about Gary's comment, however, was to return to the concern about constraining what can count as an affordance. I mean is there really anything in the world that does not afford something, hence that a Gibsonian could say we directly perceive?

  12. One day you'll have time to read the ecological laws paper (Turvey et al reply to Fodor & Pylyshyn) and then that question will get an answer :)

  13. Yeah, I'm half way through the F&P, but I'm also working on Gibson, 1979, Chap. 14. I think I'm getting behinder and behinder.