Friday, January 21, 2011

S does not visually perceive pick-up-ability in the normal box

Here's the argument.

Take two ring boxes that are physically identical on the outside (in their surfaces, if you will), but one is normal where the other is rigged with a touch-sensitive explosive mechanism.  If you touch the box, it will explode.  Give them identical conditions of lighting, temperature, etc.  That's just my stipulation.

Suppose we agree that if you can't touch the box, then you can't pick it up.  (Some room to argue there, perhaps.)

Suppose we further agree that we do not visually perceive pick-up-ability in the exploding box.  It doesn't have the property of pick-up-ability, so you cannot (factively) visually perceive it.
(*) Since they are physically identical on the outside (in their surfaces, if you will) and in their background lighting conditions, one will have identical visual perceptions when viewing them.  
(*) is not a stipulation, but an empirical hypothesis.  So there is perhaps some room to challenge it.

The stipulation plus (*), however, appear to suffice for a very simple and kind of radical objection to EP.

If subject S does not visually perceive pick-up-ability in the exploding box, then since what S visually perceives of the exploding one, S perceives of the normal one, S does not visually perceive pick-up-ability in the normal box.

So, S does not visually perceive pick-up-ability in the normal box.

32 comments:

  1. The more I think about it, the less inclined I am to agree 'pick-up-ability' is the right affordance. I don't mean that in a nit-picky way: I think you are on the right track in saying you can't visually perceive pick-up-ability in this case, and I'm wondering whether you can ever.

    Think what's involved in picking something up: you have to be able to reach it, you have to be able to grasp something useful, it's mass has to be within a range, as does it's size; etc. Charles and I have been discussing nested affordances and effecting sequences of them to achieve a higher order goal, and this has got me thinking that maybe you're right: there's no visual information about whether you can actually pick it up. There's plenty of visual information for things like size, location, the availability of handles, etc, and these are enough to guide you to the exploration of the other properties, like mass etc. There's no visual information for the latter unless the box is moving.

    I'm not trying to duck the issue: I just really think your concerns lead to the conclusion that we have the wrong affordance in our task analysis, rather than 'EP is wrong'.

    ReplyDelete
  2. You are ducking the issue. There are obvious variants of the problem for affordances in good standing.

    The one box affords alight-on-ability, but the other does not. (One of TSRM's affordances on p. 261).

    Change the explosive box so that it has a proximity fuse so that it blows up when one gets within two inches of it. This box does not afford grabable thing (TSRM, p. 261)

    Make the boxes bigger, but everything else the same. One affords sit-on-ability and the other does not.

    Make the objects two ladders that are physically the same on the outside, one normal, the other exploding. One affords climb-up-ability, the other does not.

    Make two chairs that are physically the same on the outside, one normal, the other exploding. One affords sit-on-ability, the other does not.

    ReplyDelete
  3. From the previous thread:

    "One box does afford pick-up-ability. You can pick it up. You just can't perceive it."

    But how do you know that it does and that you can?

    In each of your (essentially equivalent) scenarios you have hypothesized that some "apparent" affordance isn't really offered, which raises several challenging questions:

    - What sort of thing can be an affordance"

    - What does it "really" mean for an affordance to be "persistent"?

    - How do you identify an affordance when you "see" (ie, perceive) one?

    - Since immediate actualization is a requirement for being an affordance, how do you deal with an "apparent" affordance that isn't immediately actualizable, ie, can't be effected without actualizing intermediate affordances.

    The "nested affordances" concept to which Andrew alludes (altho I'd prefer something like "contingent affordances" - recall my Hume's Law comment) is an attempt to address that last question. Whether successful or not is TBD.

    And here's a first baby step in addressing some of the others:

    <philosophical babble>
    Something about the concept of affordances - the required/assumed persistence, I suspect - seems to be putting it into the realm of Platonic "forms", "objective truth", "reality", "moral imperatives", et al. - all of which have a role in certain language games, but that may be all. I take Sellars to have been arguing (convincingly, IMO) that perception, inference, and social practice are all we have to work with. The result of employing those may not correspond to "reality", but we'll never know.

    Ie, we may need to tweak the concept of persistent affordances a bit.
    </philosophical babble>

    ReplyDelete

  4. "One box does afford pick-up-ability. You can pick it up. You just can't perceive it."

    But how do you know that it does and that you can?


    You may or may not know. That's different a different matter.

    I'm not going to do the metaphysics of affordances here. I'm just going to work with a list, e.g. pick-up-ability, climb-on-ability, sit-on-ability, alight-on-ability. I don't want to get bogged down in all that. Let the EP-ists do not.

    That's to wander off topic as far as I'm concerned. I'm bearing down hard on what are essentially two crucial sentences from TSRM and Gibson.

    ReplyDelete
  5. Not having the benefit of the two books/papers in question, I of course can't decide directly, but it's hardly a stretch to infer from the discussion among several people who do - and presumably understand them - that part of the numerous disconnects may be due to deficient definitions and poorly worded explanations. I'm just trying to address those possible shortcomings.

    Specifically, although I think I understand the point in the Gibson quote of distinguishing between a surface and what affordances it may offer, that distinction is not useful unless you know precisely what being an affordance entails. And although I think I understand the relation between surfaces and objects (eg, according to Sabrina's proposed definition, the latter is just a special case of the former), I don't get the point of substituting one thing for another in those statements, all of which include "afford". If the definition of "afford" is deficient, its presence corrupts them all equally.

    Likewise in the TSRM quote there is an assumed but unstated (in the quote at least) definition of affordance. If optical properties specify affordances in accordance with that definition and the definition is deficient, it is no surprise that problems may arise.

    In particular, I take the point of the penultimate sentence in that post to be that if affordance is defined as we have been assuming, ie, in such a way that a supposed visual perception may be misleading because of imperceptible features of the environment (eg, booby traps), then optical properties can't always specify affordances. You conclude that there must be something wrong with the whole concept of affordances. I am merely suggesting the alternative possibility that there is correctable deficiencies in our understanding of them.

    So, here's my shot at addressing two general types of problematic scenarios you have described:

    1. those in which an affordance can't be actualized because of an unforeseeable event that intervenes between perceiving and effecting it

    2. those in which actualizing the affordance has unexpected consequences for the effector

    It seems to me that for "affordance" to be a meaningful concept, anything so labeled must be actualizable by some perceiver in some environment. But in such scenarios the presumed affordance isn't actualizable. But the problem is the assumed relationship between affordances and the environment. The scenarios are constructed so as to make the perception misleading for any perceiver in any similar environment. But since the topic under discussion is perception, how is some feature of the environment that is imperceptible for any perceiver in the environment relevant? To repeat a previous scenario, in between perceiving and effecting any affordance offered by the normal box, the world ends. Does that make every perceived affordance retroactively invalid?

    As for (2), surely there is nothing in a reasonable definition of affordance requiring that actualization must be satisfying to the effector. The affordance "sit-on-able" presumably doesn't necessarily imply "comfortably-sit-on-able".

    My point about "truth" et al was that we should take care not to define affordances so as to make them metaphysical. As both an engineer and a Rortian, I of course consider that a potentially fatal danger.

    Hence, the relevance of "But how do you know that it does and that you can?" I think we may be assuming things about the perceptual environment that can't be known by mere mortals in the environment - which is metaphysical and thus possibly fatal to the project. I'm trying to flush out any such assumptions.

    ReplyDelete
  6. Charles,
    I don't think Andrew misunderstands me at this point. I think he now sees what I am arguing and sees how the argument goes. But, I could be wrong.

    1. those in which an affordance can't be actualized because of an unforeseeable event that intervenes between perceiving and effecting it

    Ok. This is not a correct description of the exploding box. The exploding box simply does not afford pick-up-ability. It is not that it has this affordance, but that something stops its being perceived, e.g. the world comes to an end.

    2. those in which actualizing the affordance has unexpected consequences for the effector

    This too is not an accurate description of the exploding box. The box does not have the affordance of pick-up-ability, so there can't be unexpected consequences of actualizing this non-existent affordance.

    It seems to me that for "affordance" to be a meaningful concept, anything so labeled must be actualizable by some perceiver in some environment. But in such scenarios the presumed affordance isn't actualizable.

    Ok. But, I'm not giving an example in which there are affordances that are not actualizable. The normal box affords pick-up-ability (although, I argue, an observer cannot perceive it) and that affordance is actualizable; the exploding box does not afford pick-up-ability, so there's no pick-up-ability affordance to fail to be actualizable.

    I think you are getting off track.

    ReplyDelete
  7. Charles,

    This whole thing about actualizing affordances is a red herring. TSRM believe something so preposterous you might not have thought of it. It's something you wouldn't think anyone would think: They believe that you can perceive unactualized affordances. They believe you can see the sit-on-ability of a chair even if you don't sit on it. They believe you can see the pick-up-ability of a box even if you don't pick it up. SO, TSRM have this long song and dance about how you can perceive unactualized affordance (dispositions), because the anchoring properties always structure light in such a way that you can see it. But, this last part is just crazy. Anchoring properties support dispositions, but they do not (typically) structure light. Surfaces structure light.

    ReplyDelete
  8. When Andrew says, "I think you are on the right track in saying you can't visually perceive pick-up-ability in this case, and I'm wondering whether you can ever." He seems to me to be conceding what I have been arguing for, what, ten days. I think this is why we've ceased to have a disconnect.

    Now, he's trying to limit the damage somehow, but saying that pick-up-ability is the wrong affordance. But, it's easy to see that changing affordances does not solve the problem. The same problem arises mutatis mutandis for lots of putative affordances that the EPists have proposed.

    ReplyDelete
  9. I think this entire problem is being mis-conceived. There seem to be numerous confusions, some of which Charles has pointed out. I've been trying to figure out where the problem is; here are some thoughts.

    Let's lay some stuff out:

    There is the world. It has properties.

    Some of these properties are dispositions; these have anchoring properties (these are apparently just standard for dispositions?)

    Some of these dispositions are affordances - they can be used by a perceiver to control behaviour.

    Which ones are affordances? One necessary condition: they are those which can, at least in principle, be specified by perceptual information. They have to be at least in principle available to a perceiver.

    So, let's go back to the box:
    The exploding box has the disposition to explode when touched. This disposition creates no information (as the problem is laid out) so it's not an affordance.

    The exploding box also has the disposition to be picked up by me (by virtue of it's size, mass, etc, all expressed relative to my action capacities - these, I would hazard a guess, will be some of the anchoring properties). This disposition can create information (although I'm not convinced vision alone will do the trick in this particular case, but that's besides the point); it is thus plausibly an affordance, although not necessarily the one guiding behaviour in this task - it's an empirical question and we haven't run the study.

    Affordances are therefore a select subset of the properties in the world. They are dispositions which can create perceptual information, and who's anchoring properties are relevant to my action capabilities. So the box's mass must fall within the range 'liftable by me', and it is straightforward to compute this (as a pi number, for instance). The box's disposition to explode can't be defined this way; touch may trigger the explosion but it's not measured in terms of my capabilities. Frankly, it sounds as if the explosion disposition can be coupled to a perceiver but never constitute any part of a perception-action system.

    So this is where I come back to information. Gibson knew that for affordances to be of any use, there had to be information for them. Affordances and information are therefore the inseparable pillars of ecological psychology. Ken's analysis seems focused solely on dispositions; but not all dispositions are affordances and the difference, in part, is about information.

    Part of the problem may also be imprecision in the literature on affordances; we do tend to just whack '-ability' on the end of things and go from there, but that's a problem and we need to tighten up. So worrying about whether 'pick-up-ability' is, in fact, the relevant affordance is my role as a scientist. My second role is to keep information in the picture. If pick-up-ability was the affordance but there wasn't visual information for it, we'd need to go looking for the non-visual information; Ken, you trying to restrict the discussion to vision is not a fair move, because you've constrained the space so that the actual solution can't fit, then claimed the solution is clearly wrong. This is the spherical cow problem.

    ReplyDelete
  10. So, this is precisely what worried me about Charles' comments, that it would invite this mass of irrelevant obfuscation that would take the discussion off topic.

    I have described the environment as the surfaces that separate substances from the medium in which the animals live. But I have also described what the environment affords animals, mentioning the terrain, shelters, water, fire, objects, tools, other animals, and human displays. How do we go from surfaces to affordances? And if there information in light for the perception of surfaces, is there information for the perception of what they afford? Perhaps the composition and layout of surfaces constitutes what they afford. (Gibson, 1979, p. 127).

    Look at what Gibson is proposing in the last sentence. He is proposing that the composition and layout of surfaces constitutes what things afford. Given the context of the preceding sentences, it is clear he is talking about vision. My argument is that his claim about surfaces, affordances, and vision is false. Gibson is claiming that P and so it's perfectly fair for me to argue against P.

    ReplyDelete
  11. Ken's analysis seems focused solely on dispositions; but not all dispositions are affordances and the difference, in part, is about information.

    This is a red herring. It just reformulates the problem. Let us again stick to vision for simplicity and because it is what Gibson and TSRM were talking about.

    So, suppose you say that affordances are dispositions about which subjects get information in light. So, then pick-up-ability won't be an affordance, because subjects don't get information about them in light. So, the case is not a counterexample. Looks good for Gibson, right?

    But, if following consistently, this would be disastrous for Gibson, since the same argument applies to other dispositions as well. In particular, all the dispositions for which the composition and layout of surfaces does not constitute an affordance. If you look at the other cases I've hinted at and you understand the moral of the exploding boxes case, you'll see this knocks out lots and lots of affordances. So, sit-on-ability, grabbabilty, climb-up-ability, throw-ability, stand-on-ability, etc., etc. will turn out not to be affordances. All that will be left standing as affordances will be such things as "see-ability", if that is even an affordance.

    ReplyDelete
  12. Ken -

    From the exploding this-or-that scenarios you seem to draw this moral:

    If something offers an "apparent" affordance and some imperceptible aspect of the environment precludes actualization of that "apparent" affordance, it isn't really an affordance.

    And as best I can tell, most - if not all - of your objections to Gibson and TSRM hinge on that moral.

    My discussion of those type-1 scenarios argues that the injection of imperceptible aspects of the environment into a discussion of perception is impermissible. Relative to "real" prospective perceivers in the environment, it puts the injector in the position of a super-perceiver - a metaphysical entity that perceives things no one else can perceive. I take Andrew's comment "it sounds as if the explosion disposition can be coupled to a perceiver but never constitute any part of a perception-action system." to be pointing in the same direction.

    Unless resolved, that seems to leave you and Andrew at an impasse.

    I admit to contributing noise by temporarily buying into the validity of the type-1 scenarios and trying to reconcile affordances with them. But I stand by much of the rest of what I've said. You claim that EP is arguing P and that you're arguing ~P. I see you to be arguing ~P' and have been trying to reconcile P and P'. You may see that as obfuscation, but I see it as basic logic.

    And along those lines, that you find the concept "that you can perceive unactualized affordances" to be "preposterous" suggests to me a major disconnect in use of the vocabulary. I understand that to be precisely the nature of affordances - no need to sit on a chair to perceive the sit-on-able affordance notwithstanding possible adverse consequences of doing so (the point of my type-2 scenarios). That's a pretty basic disconnect that one might want to resolve before drawing any conclusions supposedly fatal to the whole EP project.

    ReplyDelete
  13. You've correctly identified that vision doesn't provide you with access to every property you might need. This is only a problem if you only have a visual system. We don't have this problem.

    You've also correctly identified that not every disposition is specified. This is only a problem for eco-psych if people acted as if they could perceive that unspecified disposition. You think sit-on-ability isn't specified, because you've restricted yourself to vision. Even if it turns out sit-on-ability isn't a visually specified affordance but it needs non-visual information then I won't cry about it unless it turns out people are actually effecting this when they shouldn't be able to.

    It may very well be the that the dispositions you propose as affordances turn out to not be what people are using. That's ok: it's always been an empirical question what information and affordances people are actually using. But you've stopped your analysis too soon: if these aren't visually specified but people are behaving as if they are using it, then either there's non-visual information which people use to perceive the affordance or people are actually effecting a different affordance. Maybe people aren't effecting 'sit-on-ability' until they are actually trying to sit, in which case there's plenty of non-visual information for that affordance. Up until then they may be effecting a different affordance; that's fine, perception and action are time extended processes in which the search for further information is just an action controlled by information that is available.

    So you've identified some issues about the visual perception of some properties which may or may not be affordances. That's actually only stage 1 of an empirical, ecological exploration of a task space, and you've given up too soon to be causing any problems to the programme itself.

    ReplyDelete
  14. From the exploding this-or-that scenarios you seem to draw this moral:

    If something offers an "apparent" affordance


    I'm not drawing that moral. I don't know what an "apparent" affordance is. I know what an affordance is. My argument does not use "apparent" affordances; it does not conclude with "apparent" affordances. This is your introduction.

    I think it is preposterous to think we typically visually perceive non-actualized affordances; EPists think we typically visually perceive non-actualized affordances. We are essentially diametrically opposed on this. This is not what I mean by a disconnected. We are in direct opposition. By a disconnect, I understand something like talking past each other.

    ReplyDelete
  15. You've correctly identified that vision doesn't provide you with access to every property you might need. This is only a problem if you only have a visual system. We don't have this problem.

    You are being miserly in your concession. The argument shows more than this. It shows that many (most?) of the properties that EPists have proposed as affordances are not perceived by vision.

    There are obvious permutations on the exploding box case that work for:

    grab-ability,
    climb-upability,
    dig-into-ability,
    copulate-withability,
    crawl-into-ability,
    leap-over-ability,
    alight-on-ability.

    Will you concede that these putative affordances are not visually perceived? Or, do I have to write up the cases?

    ReplyDelete
  16. I'd have to actually run some studies before I concede that. A careful analysis might show these are fully specified. You should feel free to write them up; as I said, your current analysis has stopped too early.

    I will concede people are a bit lazy in their use of the word affordance. It's partly that the word got checked out by a lot of people who didn't also check out 'information'; of course, Turvey et al might also being optimistic about these specific cases. But people who actually do empirical work (Bill Warren and Geoff Bingham are the two best examples) tend to be much more careful: Geoff especially is a disciplined scientist who I was lucky to study with.

    Regarding Charles' point: the objection still seems to run if you replace 'apparent' with 'candidate' or even remove it. The issue is the imperceptibility of something which 'precludes actualisation'; that's the problem.

    I think it is preposterous to think we typically visually perceive non-actualized affordances;
    How do you think we then come to do things, if we can't perceive the possibilities for action somehow?

    ReplyDelete
  17. So, Charles, here is why I don't need "apparent" affordance. The two boxes are perceived to be the same; the two boxes differ in affordances, so evidently one is not perceiving the affordances.

    The two boxed can appear to be 85 monkeys eating bananas. (I don't think they do.) That doesn't matter for the problem. So, I'm setting it aside.

    ReplyDelete
  18. Andrew,

    So, take to boxes 3' x 3' x 3' in size. They are physically identical on the outside, but one is rigged with a touch sensitive device so that it will explode if you sit on it.

    I say they are visually perceived to be the same, but that they afford different things (one affords sit-on-ability; the other does not). So, one does not perceive sit-on-ability.

    ReplyDelete
  19. Maybe one does, but incorrectly in one case. This would explain why minefields work.

    ReplyDelete
  20. In fact, that might be it; you're in error because of the evil philosopher. You can always be made to be in error because of the evil person running the experiment - but we've had this discussion before, I think, and it's not that relevant to laying out the mechanisms of perception except that the error hints at a limitation.

    Information has scope; you've broken the scope by imperceptibly rigging the box. So the fact that information no longer helps you is hardly the fault of the person perceiving the specified affordance and not the unspecified disposition.

    ReplyDelete
  21. When I initially entered this discussion, Gibson's approach was being described in system engineering terms. And once I got the general idea, it appeared that the approach could be viewed as a conceptually straightforward feedback control system (though presumably quite complex analytically) with light as input and motion (AKA action) as output.

    Within that view, the light input is considered to be structured and the structure is considered to be a vehicle for information about the environment. The system's primary function is to extract that information for use in moving purposefully within the environment. And there is a concept called affordance that is a characteristic of the information extracted that is useful in relating the information to appropriate responsive movements.

    If that summary is more-or-less accurate, affordances should be some mathematically specifiable entity in the mathematical model that I understand to be the real core of the Gibson approach. And as such, it isn't clear to me why they need to be described in seemingly metaphysical terms like persistence, actualizability, et al. I think Ken and I are on similar paths here, only they diverge in that Ken seems to want to KO the EP project, I just want to understand how it's supposed to work. I would take purely prose passages in the books as merely motivational whereas Ken wants to find logical inconsistencies. I don't really care about verbal imprecision. To paraphrase the famous movie line, "Show me the math!" And I'm counting on Sabrina to do that in due time.

    In short, I realize that I no longer even know what the discussion is about. But I am now clear that I have nothing to contribute on the topic of exploding boxes.

    ReplyDelete
  22. In fairness to Gibson, affordances do have metaphysical work to do, in the sense that they need to be real properties that can be detected by a perceptual system and guide action. These aren't trivial concerns; Gibson had to reinvent perceptual information to get this idea to be of any use to a theory of perception.

    But they do need to persist (so that they are stable enough to support learning). They do need to be actualizable, ie support an action that someone might actually want to do. Turvey's language gets a little carried away sometimes, but none of this is by accident. They are attempts to formalise a solution to a problem that all theories of perception face but that Gibson thought all except his failed to solve.

    ReplyDelete
  23. Andrew,
    Scope is another red herring. The basic point you are not getting is (in the TSRM framework) that anchoring properties do not structure light. In the Gibson picture, it is not the surfaces that make a thing explosive or no (hence pick-up-able or not).

    What makes one of these boxes explosive or not (hence pick-up-able or not) has to do with the inner workings of the thing that are not exposed to light. That's why these cannot structure light. All you do with the (unprincipled) appeal to scope is try to dismiss the cases that make the obvious obvious.

    You are letting all thing EP baggage obscure some very elementary physical facts.

    ReplyDelete
  24. And there is a concept called affordance that is a characteristic of the information extracted that is useful in relating the information to appropriate responsive movements.

    Charles, I don't think that this is right. To get in the ballpark, let me appeal to the TSRM dispositional approach. So, a glass has the dispositional property of being fragile. A piece of food has the dispositional property of being edible. So, the TSRM idea is that what makes glass fragile or food edible (an anchoring property) also structures light. This structure in light is the information. The affordance is not a feature or characteristic of the light.

    ReplyDelete
  25. Here is yet another way to see what TSRM are saying. They claim that anchoring properties do two things. They are what make affordances affordances and they structure light. My objection has been trying to make it clear that anchoring properties typically only do one of these jobs, namely, make affordances affordances.

    ReplyDelete
  26. Here is a straightforward kind of experimental paradigm. An experimentalists wants to know what properties a subject responds to, property A or property B. So, the way to test this is to hold A constant and vary B, then hold B constant and vary A. So, you can think about the boxes case as having A properties of surface features and B properties of affordances. Subjects respond to surface properties, not affordances.

    What Andrew is would have to do to get his appeal to scope to work would be for it to rule out these kinds of experimental protocols. But, if it does rule out these sorts of protocols, then we should really reject the appeal to such strictures on scope.

    ReplyDelete
  27. Ken -

    As I said, our objectives have diverged. You are (I think) trying to find flaws in Gibson's and/or TSRM's presentation of their concepts, whatever those may be. I'm trying to make sense of the underlying problem and proposed solution that I assume EP to be addressing, and it appears that some tweaking of the underlying concepts - or at least the vocabulary in which they are presented - may be required. My current thoughts along those lines are in this comment at the Two Psychologists blog:

    http://psychsciencenotes.blogspot.com/2011/01/reading-group-gibson-1979-chapter-6.html?showComment=1296068820201#c702088374569105587

    In short, I'm suggesting that "affordance" as we have been assuming it to be defined doesn't appear to work. I'm trying to understand if that's really true, and if so what changes might be needed. And since I'm questioning the details of the definition of "affordances" that we have been assuming, whether that definition is consistent with some other concept (possibly equally vague) is irrelevant to my objective.

    ReplyDelete
  28. Charles,
    So, I am back home and thinking a bit more about these issues.

    Ken seems to want to KO the EP project, I just want to understand how it's supposed to work. I would take purely prose passages in the books as merely motivational whereas Ken wants to find logical inconsistencies. I don't really care about verbal imprecision.

    I indeed think that my objections spell trouble for the EP theory of how we perceive affordances, but there is more to EP than affordances.

    I am not, however, trying to find logical inconsistencies in Gibson or TSRM, although they may be there. Instead, I am trying to draw attention to some empirical claims that they make that seem to me to be false.

    Moreover, I don't think that the EP folks can get their chestnuts out of the fire by clarifying their terminology. Gibson and TSRM seem to me to spend an inordinate amount of time trying to set things up, but a lot turns on just a few very simple and implausible assumptions, e.g. that affordances are constituted by surfaces and that anchoring properties structure light. These are crucial assumptions/theses that I have been arguing are false.

    ReplyDelete
  29. SO, TSRM have this long song and dance about how you can perceive unactualized affordance (dispositions), because the anchoring properties always structure light in such a way that you can see it.
    Where in TSRM is this song and dance? I was looking for it and couldn't find it.

    ReplyDelete
  30. One spans pp. 265-266. These are the "e-ness-> o-ness" "laws".

    ReplyDelete
  31. Here it's relatively clear (for TSRM):
    The second stage in analyzing an affordance for an animal is very much a matter (for visually detected affordances) of mathematical optics: to describe the light patterned by an affordance-specific occurrent property and to characterize that patterning in terms of an optical variable, an optical property, that stands in strict correspondence to the occurrent property.

    What I have been driving at (formulated for TSRM) is that light patterns do not typically stand in strict correspondence to the affordance-specific occurrent properties. The reason is obvious: The affordance-specific occurrent properties are not properties of the surfaces; they are properties of things beneath the surface. So, what makes something rigid is not the surface so much as the whole of its structure.

    ReplyDelete
  32. There are plenty of weird turns in this section. So, for example, TSRM start discussing Quine and nominalism and (as I understand him) Quine was a nominalist, but Fodor and Pylyshyn are not. And, I don't think Quine would ever be considered an Establishment cognitivist. He was, as far as I can, tell some sort of Skinnerian. Fodor and Quine are pretty far apart on the cognitive science landscape.

    So, TSRM is very dense and sometimes awkward, but I think the idea is relatively simple: anchoring properties of affordances (which are occurrent, rather than dispositional) structure light. This, however, is typically not so.

    ReplyDelete