A blog about problems in the field of psychology and attempts to fix them.

Sunday, February 10, 2013

Could Affordances Structure Light?

This post should have a subtitle. It should read: Could Affordances Structure Light? Ken Aizawa gets a 10-month-delayed response

Last April I started blogging about my (now-submitted) attempt to update Cutting's paper distinguishing between Gibson's approach on ecological psychology and the emerging Connecticut approach. The first post generated many comments, and I promised to follow up on some of them soon. Well... in publishing time this is still "soon", even though in blogging time it is ages. In particular Ken Aizawa hit me with a few hard questions including the perennial stickler, "Can affordances structure light?"

I think it is important for ecological psychology that they do, and I think they do. However, my position (which I associate with Gibson's thinking) is less extreme that that promoted by the Connecticut approach (e.g., Turvey, Shaw, Mace, and the young Reed). I also think it is pretty darn simple to support:

As I argued in my Brief Introduction to Ecological Psychology, the term "affordances" came late in the development of the Ecological Approach, and I think it should stay late in the priorities of the approach. What came before was a broad argument for direct perception, a broad argument that the state of physical objects structures the various ambient energies (light, pressure waves, chemical gradients, magnetic fields, etc.) in ways specific to certain properties of the objects. For example, though a cylinder and a square may appear identical from a particular angle, they structure light in very different ways, such that moving around them will quickly disambiguate the situation.

This must be the first and highest priority claim for any ecological approach, and if this argument fails, there is no need to move further. This claim traces back, I believe, to the lineage of American Philosophy. James, paraphrasing Peirce, tells us that "There can be no difference anywhere that doesn't make a difference elsewhere." That is, there can be no difference between two things that does not have some consequence. The ecological psychologist adds, to be more specific, that the difference will, under at least some set of circumstances, structure ambient energy in a way characteristic of precisely that difference. 

Given that I am here trying to answer Ken's question, we may now fast forward to one of the last steps in setting up the armature of Ecological Psychology. We will simply assume, for now, that what an organism can do is limited by the properties of the organism and the properties of the world (a generally uncontroversial claim). The question then is whether or not those properties of the world are specified in ambient energy. That is, we must ask whether the difference between a world in which a cave entrance affords my passing-through and a world in which the cave entrance is too small has a corresponding difference in ambient energy arrays, and so on for other affordances.

I have never quite understood why this particular issue is controversial, and maybe Ken and others will, in the comment section below, smack me around for my ignorance.

1) If we have already established that all properties of objects are, under the right set of circumstances, specified in ambient energy, then why would it be controversial to claim that, under at least some circumstances, affordances --- opportunities for action --- are specified in ambient energy?

2) Maybe the question was whether or not all affordances structure ambient energy? To which I would say "Yes, they all do. ALL OF THEM. While some will structure energy in a way the organism cannot detect, that is a separate issue. Whatever the properties of the world are that create an opportunity for a particular organism to act, they create a unique configuration of ambient energies. How do I know? Because that is what it means for such situations to be different from other situations. QED. A difference that doesn't make a difference is no difference at all."

3) Maybe I'm cheating though. Maybe the question was specific to light? To which I would say, "Well, some affordances structure light and some don't. That's alright. Though Gibson talked the most about visual perception, it should be clear that his method was intended to apply to all perceptual systems (see his 1966 book). His system can also handle "multi-modal specification"; as when the property creates a unique pattern across modalities. Tom Stoffregen has argued that all specification must be multi-modal, and his argument could be stretched to potentially undermine the idea that light-on-its-own specifies affordances. However, it certainly doesn't undermine the broader notion of specification."

4) Maybe the question is whether ambient energy always specifies the presence of a certain affordance? To this I reply, "Well of course it doesn't! No one would claim that there are light-specified affordances in a dark room. The situation is even more complicated in organisms using perceptual systems where the pattern of ambient energy depends on their own activity. Good examples of this include the electric-field-sensing electric eel and sonar-using bat. In both of those cases, when the organism stops doing its part in generating the energy, the patterns-unique-to-the-affordances also disappear." It is worth noting that this is point on which, I believe, the Connecticut approach is in a much worse spot, philosophically, than is Gibson.

5) Maybe the question is whether the light specifies both the presence of the affordance and somehow tells the organism that it is the type of organism that can take advantage of the affordance? I suspect this is the real issue. Several attempts to do epistemology care not about what people "know" in the broad sense (e.g., "I know its time for a beer, that's what I know."), and not about what people "know" that is correct (e.g., "I know the capital of Maryland is Annapolis."), but rather they care in particular about what people can know they are correct about knowing (e.g., "I can doubt everything except that I am doubting. Therefore I am doubting."). These particular philosophers care not about proving whether or not people can have correct beliefs, but about whether people can infallibly justify their correct belief. For these people it is not enough to show that an organism can know its behavioral opportunities if it detects the correct pattens of energy, we must further show that the organism could be certain that it has done that task correctly.

If this is the concern, then I say, "That is indeed a problem, because I don't think such specification exists. However, that does not undermines Gibson's explanation for perception, i.e., his explanation for how organisms generally act accurately with respect to the environment. It might or might not undermine Gibson's argument for direct perception, but Gibson's system is built on an old-school American-philosophy foundation, and so this will become a much longer conversation. (See also Cornell West's discussion of The American Evasion of Philosophy.)"

The inquisitor might press further: "How is it," they might ask, "that the organism only perceives its own affordances, if there is nothing in the light to tell it which affordances are for-it-in-particular?" That question can only be answered, I would argue, through extensive empirical study of evolution and development. That is, organisms become attuned to certain affordances and not others, and we know an awful lot about how that occurs, but not as much as we would like.This is deep and complicated question, but it is scientifically tractable.

--------

Alright... have I answered some of the question? all of the question? or did I completely miss the boat?




11 comments:

  1. In both of those cases, when the organism stops doing its part in generating the energy, the patterns-unique-to-the-affordances also disappear." It is worth noting that this is point on which, I believe, the Connecticut approach is in a much worse spot, philosophically, than is Gibson.
    What do you mean by this; why is the UConn approach troubled by this?

    That is, organisms become attuned to certain affordances and not others
    This seems like an uncomplicated answer to the question to me; you just have to remember about learning and development to see it :)

    ReplyDelete
  2. What do you mean by this; why is the UConn approach troubled by this?

    Well, recall that I think the UConn approach is unnecessarily dogmatic about certain issues, and that this dogmatism (potentially) leads to some serious conundrums. Among other things, UConn says that affordances must lawfully produce the ambient energy pattern that an organism must be lawfully attuned to, or there is no affordance. Let's compare this with Gibson in an example:

    Let us assume we have a bat (Bat 1) with sonar capabilities at one end of a dark cave with a few obstacles in it. The cave is such that, by anyone's standards the cave affords flying-through for Bat 1.

    Now let us assume there is another bat at the end of the cave. The second bat is identical to the first, except without sonar capabilities (Bat 2).

    In Gibson's system, it seems clear to me that the cave affords flying through for Bat 2. The big empty parts of the cave are big enough for Bat 2 to fit through them, the atmosphere is sufficient to allow flight, the bat is capable of flight... and that's it. The fact that its dark and so Bat 2 cannot, at this exact moment, take advantage of the affordance, doesn't change any of that. The presence of the affordance is fundamentally independent of the perceptual abilities of the organism.

    However, in TSM, it seems to me the cave does not afford flying through for Bat 2. This is especially obvious if you take the dispositional account seriously. By definition, if the organism is not doing the behavior, then the organism is not afforded the behavior (as dispositions necessarily obtain). And similarly, if we turned on a sound system that interfered with Bat 1's sonar, suddenly the UConn people would have to assert that the cave did not afford flying through for Bat 1. That seems awkward to me, but it is not too bad.

    Now, to make it a bit more awkward... Normally wethink of ambient energy patterns as simply present in the environment. The optic information I would use to navigate through my office is there where I am present or not, whether I close my eyes or not, etc. But the ambient acoustic energy used by Bat 1 to navigate through the cave is generated by the bat! This means that, in the UConn approach, the cave affords flying-through to Bat 1 when the bat is making clicking noises, but not when the bat is not-making clicking noises.

    That seems highly problematic to me. To use the UConn terminology: The "effectivity" and the ambient energy and the affordance seem problematically entangled with each other.

    ReplyDelete
    Replies
    1. Effecting an affordance requires you to be able to perceive it. You aren't the right kind of dispositional partner if you can't perceive it, as evidenced by the fact that you aren't effecting the affordance. TSM have no trouble saying the cave affords flying through for the bat, and that at time A the bat cannot perceive the affordance but can at time B.

      TSM aren't conflating affordances and the perception of affordances, I think you are, here.

      Delete
    2. Alright... but who is afforded?

      For Gibson the Bat is the right kind of organism, but circumstances are off.

      By your account, for TSM, the bat is not the right kind of organism. And if you are not the right kind of organism, you are not afforded the opportunity, by definition.

      I am not accusing them of conflating anything. I am accusing them of getting into a bind where certain concepts are connected (entangled) in undesirable ways.

      Note, in either system you can not specify the affordance without specifying who is afforded the opportunity for action. The issue is how much detail goes is required in the descriptions of the organism and the circumstances.

      For Gibson the Bat is afforded flying, but can't at the moment.

      For TSM the Bat is NOT afforded flying, because it can't see. This is because, for TSM, the affordance and effectivity are "lawfully related" (with a very narrow interpretation of the term "law") and so one cannot be present (in any real situation) without the other.

      Note, in TSM you could say, "If the lights were on, or if the bat had sonar capabilities, then there would be an affordance." But you would also have to admit there was no such affordance, for that bat, at this moment.

      Delete
    3. P.S. Note, such an entanglement doesn't necessarily mean TSM is "wrong". I am making the lesser assertion that we should better understand the consequences of heading that direction, and that should choose the direction we want to go with such consequences in mind.

      Delete
    4. For Gibson the Bat is the right kind of organism, but circumstances are off.

      By your account, for TSM, the bat is not the right kind of organism. And if you are not the right kind of organism, you are not afforded the opportunity, by definition.

      I'm saying the bat who can't see the cave isn't currently in a position to effect the affordance because there is no information for the affordance. The affordance persists (it will be defined in terms of things like the size of the bat and the space it needs to maneuver, and won't include 'presence of light').

      So everyone, Gibson and TSM, can be happy that there is an affordance but there is no information for that affordance that the bat can detect just now. Affordances and information for affordances are two things for both these 'camps'.

      Delete
    5. Ok... backing up.

      Aren't (in TSM) Affordances and Effectivities complementary, such that one cannot exist without the other?

      And also such that if an affordance meets the proper effectivity it MUST obtain?!?

      The same way a "fragile" glass MUST break when it hits a hard surface?

      I think you are wussing out on the rigidity of what TSM are asserting.

      Delete
    6. Dispositions are physically formed by anchoring properties; the glass is fragile because of what it's made of. The disposition to break exists and can be defined, even if there are no hard surfaces (as far as I know). This is one of the things I like about the dispositional account; it makes affordances something that might serve as a target that can shape evolutionary and developmental selection pressures. The cave was always flyable through to something like a bat, and therefore bats are the way they are because that opportunity existed and they evolved for various reasons to take advantage of that opportunity (rather than others).

      This argument about 'well you must effect dispositions and we don't always therefore no' doesn't hold much water for me. The affordance half of the disposition exists; whether or not I am the other half varies over time and given the multidimensional nature of the dispositions we're talking about, it's not hard for me to be mostly but not exactly the right thing. Only when I fit properly is the effecting compulsory; maybe this is part of automaticity in expertise.

      One useful thing I got from arguing with Ken, though, was getting pointed to the bit in TSRM on anchoring properties. This aspect of dispositions is never talked about but it seems to me to solve many problems.

      Delete
  3. First, Gibson made claims about structuring light, not ambient energy. The former is a stronger principle than the latter. Structure light entails structuring ambient energy, but not vice versa. So, you are not defending Gibson.

    Second,
    "The ecological psychologist adds, to be more specific, that the difference will, under at least some set of circumstances, structure ambient energy in a way characteristic of precisely that difference."

    The problem is that this is too weak a principle for what you need. It is not enough that, under at least some set of circumstances C, differences will structure ambient energy in a way characteristic of those differences. One would have to have it happen that those conditions C obtain in practice.

    So, in principle, the difference between a sit-onable chair and one that is not sit-onable (because it has a materials defect) might be "perceived" if I viewed them through some sort of X-ray machine. But, so what? If I don't have the x-ray machine, then I don't perceive the difference.

    Here is the point more simply. No one wants to say that one can't tell the difference between a sitonable chair and a non-sitonable chair. The problem is whether one can tell the difference between a sitonable chair and a non-sitonable chair just by looking, which is what Gibson said.

    ReplyDelete
  4. First, Gibson made claims about structuring light, not ambient energy. The former is a stronger principle than the latter. Structure light entails structuring ambient energy, but not vice versa. So, you are not defending Gibson.

    Well... while it is true that Gibson's last book focused almost entirely on vision, the 1966 book clearly dealt with the full range of "perceptual systems". Also, while the vast majority of Gibson's experimental work involved vision, he did perform several experiments looking at other modalities, and always using the same basic principles. So, while many people are interested primarily in Gibson's theory of vision, and most of Gibson's empirical work was on vision, it is still the case that he was offering a much broader theory of perception.

    Note, for example, the generality of the discussion in the working-paper here: http://www.trincoll.edu/depts/ecopsyc/perils/folder6/modality.html

    ReplyDelete
  5. Second,
    .... No one wants to say that one can't tell the difference between a sitonable chair and a non-sitonable chair. The problem is whether one can tell the difference between a sitonable chair and a non-sitonable chair just by looking, which is what Gibson said.


    Well, now we get to a more interesting problem: "What was Gibson up to, and did he succeed?" vs. "What do some people wish Gibson was up to, and did he succeed at that?"

    Frankly, many people have historically asserted that you cannot tell the difference between the two types of chairs.

    As I see Gibson, he is working within the tradition of American Philosophy, and trying to push back against fundamental assumptions of the mainstream Continental Traditions. There are a few major issues here... which people might not care about much any more, but which I think were once quite cared about. For example:

    It has long been asserted that there was not an environmental basis that could support veridical perception. Period. (Just as Chomsky asserted there is not sufficient environmental support for language acquisition.) If this is true, then there must, somewhere, be magic in the system to explain veridical perception - it cannot be just the body and the world. Gibson's analysis of ambient energy patterns suggests that there is sufficient environmental support, so a theory of perception without magic is possible. That alone is a pretty major victory... or at least it was at one time.

    Also, it does not seem to me that Gibson ever suggests all perception is veridical. (You know, cause when I read his books they appear to have elaborate discussions of situations in which non-veridical perception occurs.) As I see it, once Gibson has shown that there is a potential basis for veridical perception, he then tries to show that the necessary confluence of organism and environment obtain at least some of the time. Gibson clearly thinks it is the majority of the time, but it would be enough of a victory to show that it occurs, in practice, even some of the time.

    Back to your chair: Are there some things about the chair the organism cannot perceive due to its physiological limitations? Yeah, sure. Are there other things about the chair the organism can see? Yeah, sure. Can the organism see that the chair is sit-on-able? Well, it depends on what you mean. The organism can distinguish chairs that are too small, chairs that are too pointy, chairs at the wrong angle, etc., and in many cases the organism can visually distinguish chairs that are too frail. Now we just have to agree on the properties that make a chair sit-on-able and see if they match up.

    Is the organism sometimes wrong? Yeah again... but we already admitted that organisms were sometimes wrong, so I don't see why that is relevant.

    -----------

    I know this won't make you totally happy, but I'm not sure where to go next. What am I missing?

    -----------

    P.S. Thank you for the responses! I am glad you had a chance to look at this.

    ReplyDelete