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

Saturday, January 5, 2013

TSRM 1981 & The New Connecticut Approcah to Eco Psych

Warning: If you aren't interested in Ecological Psychology, this will probably be a boring post....

In past posts I talked about the two ecological psychologies article by Cutting, which asserted that Turvey and Shaw were doing something fundamentally different from Gibson. I've also argued that Gibson was doing something grounded in American Philosophy, which his latter standard bearers did not appreciate. Much of the impetus for the new direction comes from the debate in 1981 between Fodor and Pylyshyn on the one hand and Turvey, Shaw, Reed, and Mace on the other. I've already summarized F&P's arguments, and stated in two places the main points I think should have made in a reply. If you recall, F&P started all this by publishing a 58 page critique of Gibson's 1979 book. Apparently, TSRM felt the need to up the anti with a 68 page reply. Good lord. Luckily for F&P we were past the days of duels!

TSRM is a philosophical tour de force, bordering on an unfocused mess. They spend an incredible amount of time getting sidetracked into trying to undermine more or less everything F&P say.  I don't want to discuss too much of that here, and instead want to summarize how they defend the ecological approach, and where they see the field going. 

Summary and initial evaluation of:
Turvey, M. T., Shaw, R. E., Reed, E. S., and Mace, W. M. (1981). Ecological laws of perceiving and acting: In reply to Fordor and Pylyshyn (1981). Cognition, 9, 237-304. 

Following the convention from the prior posts: Summaries of TSRM's position (and direct quotes) will be in blue, my comments will be normal black.  Bold will indicate F&P’s section titles.
------

TSRM begin by pointing out that:
Because Fodor and Pylyshyn place such a great emphasis on Gibson’s claims about ecological laws, and because they evaluate Gibson’s claims in the terms of the philosophy of science---terms in which Gibson’s claims have neither been criticized nor defended previously---our focus is this aspect of the ecological approach. (p. 283)
They go on to explain Gibson's approach for 5 pages, then get distracted for 15 pages responding to nuances of F&P's paper, discuss natural laws for 15 pages (establishing therein the basic principles of the Connecticut approach), spend 8 pages flailing around with the problem of error, and 2 pages trying to nail down "direct realism". Then they get distracted for another 16 pages reminding psychologists not to attribute elaborate conceptual abilities to actors unnecessarily, 6 pages showing that "intentional behavior" can be explained without elaborate internal-representation systems, and finally they end with a nice, blessedly short and optimistic, postscript. As expected, they make only a few of the points I was inclined to make when responding to F&P.

2. Gibson’s ecological approach in overview (pages 238-245)
TSRM point out that F&P might not even be discussing the phenomena that Gibson was concerned with: Gibson made "a framework that would do justice to the practical success of an organisms 'everyday' behavior.... the perceiving required to support running, flying, building, grasping and so forth." (p. 238-239)

That much is definitely true, but then TSRM start to play really fast and loose:
an animal runs on the ground and an animal sees the ground. This much should be common sense. There is no thing between the animal and the ground in the relation. This is what Gibson has always meant by direct perception... (p. 239)
What are they talking about?!? There are things between the animal and the ground in the seeing-relation. Usually, when I am running, there are things between me and the ground, i.e., shoes and socks. This can't be what Gibson means by direct!

However, TSRM quickly recover, and give a list the traditional intermediaries that Gibson rejects - e.g., retinal image, mental representations. They also smartly point out how a dualism starts to fall apart when you bring activity back into the story - as you can't walk on a mental image. They then say the can summarize Gibson's approach in terms of "a major denial and a major assertion" (p. 240). Gibson denies the existence of (traditional) intermediaries in perception, and asserts that perception is intentional.

I think this is a somewhat shallow way to approach Gibson. Certainly, those are important aspects of his system (though it is always hard for me to parse the philosophical use of the term 'intention' in these sorts of contexts). However, I think there is a lot more to Gibson than just these two points, and I don't like leading with the negative assertion. We should be able to explain Gibson perfectly well without ever saying what he denies (I'm not saying TSRM should have done this, but they could have. And now, 30 years later, I have still never seen it done).

TSRM go on to emphasize that perception does not involve "inference". They also assert they are not concerned with F&P's primary concern: "fixation of belief." They deny an interest in the perception of particulars, and insist they can pre-demarcate their topics of interest based on what can be specified in the light (inherently admitting that particulars cannot be specified!). They further demarcate their interest as "the perceiving that goes with acting" (p. 240). They readily admit that we lack a good vocabulary to even begin talking about nested multi-level actions. 

They next turn their focus to a lengthy treatment of awareness vs. awareness-of. Best I can tell, this seems to be a really bad way of contrasting awareness of something "in your mind" vs. awareness of something in the world. TSRM then go through the usual process of unfairly bashing prior theories of perception (accusing them not just of being wrong, but unscientific), and in the process introduce the perennially problematic term "information" without explaining it. Then, F&P are rightly taken to task because: 
one cannot disprove an empirical hypothesis by assuming that it is an incorrect logical claim. (p. 243)
The latter is one of my favorite points in the paper because I think TSRM make it very well, and I think it is a common problem with people misunderstanding Holt and James. As for other misunderstandings:
[F&P] ridicule the idea that more sampling of optical structure can clarify a situation on the grounds that this idea introduces an arbitrary, unconstrained move that opens the door to a trivial interpretation of Gibson. But in a natural environment the invariably present option to sample further is an absolutely essential aspect of the adaptive behavior of organisms.... Most assuredly, part of visual perceiving is the ability of an organism to change and select samples according to its current challenges.... It is difficult to understand why Fodor and Pylyshyn think that the opportunity to extend the sampling of optical structure so arbitrary when it is so integral a part of the daily living of mobile, seeing organisms. (p. 243)
Alas, this would be the time to bring in American Philosophy (something like this), but TSRM don't have the background. Further, they will later undercut this by failing to accept the obvious corollary that organisms are often wrong... then they explore more... then they are correct.

In fact, in the next paragraph TSRM assert that the point of Eco Psych to study "an organism's apprehension of the environment and how it controls its acts with respect to that environment." They also explicitly admit, in a footnote, that there are other varieties of "perception" which they plan on ignoring. 

This is all muddled. They talk in generals that wouldn't hold up to scrutiny; they clearly mean something more specific, because they bash the study of operant conditioning (which is blatantly about how an organism controls its acts with respect to the environment) and they seem to reject a phenomenological account of "apprehension." Also, they clearly do not intend to accommodate the things they claim to be leaving for later.

Oh, and then they spend several pages playing the "I'm rubber and you're glue" game. They claim all of F&P's criticisms of Eco Psych really apply better to the Establishment Position. In so doing, they slip in many odd little bits they would later reject (later in the paper and later in life). These include: Admitting that "inference exists in cognitive life" (p. 244), which entails admitting organisms have a "cognitive life"! This bleeds into the next section, and is joined by a bunch of "you said X, but a recent author said Y" rhetoric. 

The last game produces text that is necessarily dated and uninteresting almost as soon as it is published.... because there is always a recent author saying what you want and another recent author saying something else... and because the games in other literature (e.g., philosophy of science) change so often that 30 years later it is unclear what, exactly, they are talking about and why we should care. 

<Skip two sections>

5. The specification of intension: affordances and ecological laws (p. 260-267)

This section has some very bizarre parts, but it also lays out the basic logic of the Connecticut approach to Eco Psych. They begin with this:
Let us ecological realists put our major ontological cards on the table: (i) There are no bare particulars (individual) and there are no pure forms.... (ii) Some properties are intrinsic to a thing and some properties of a thing are mutual, invariant only in its relations to other specified things.... (iii) Things may be regarded as individuals or classes depending on the context and the interests of the observer.... (iv) There are no things that do not change and no changes that take place independently of things. (v) Categories of kind may be distinguished, such as event, substance, place and relation.... (vi) All things, therefore, have both persistent and transient properties, both of which are real.... (vii) Properties are not a separate category of individual, for there are only propertied things.... (viii) Succinctly, properties are inclusion relations among things; hence we can now state (ix) The ecologica1 principle of nesting: (Gibson, 1979, p. 9): all things are more or leass complexly nested. This nesting has no limit either in scale or grain --- those relations (v,d) among things nested within a thing, and those relations into which a thing is nested, constitute a thing’s properties, intrinsic and dispositional.  (p. 260)
Most of those "cards" seem to me unnecessary. I don't think they gain by going overly technical, and they bring in lots of issues that are sever distractions.

Moving on, they give the dispositional definition of "affordances" and "effectivities" (which they have since disavowed), as apart of a tripartite relationship in which an organism that will do something given the opportunity, when it encounters the opportunity, does the thing. Though, to be fair, there are probably at least three incompatible definitions of effectivities given, each of which would provide its own interesting interpretation of what is going on here.

For now, I'll skip trying to explain why this formula isn't very interesting and/or has deep flaws. There is also a lot of totally unnecessary variables and symbolic-logic formulae.

There are long sections on "disopositionals" vs. "occurrent properties"; about the reality of possibility, i.e., a real possibility to perform a certain action; and about what a "natural kind" is. Somehow, through all this unnecessary muck, they come to a two-part plan for empirical research:
The first stage: is very much a matter of ecological physics: to isolate occurrent physical properties that are invariant over the extension and the perspectives and which are non-accidentally related to the extension of the affordance.... It seems to us that the task of ecological science is... the empirical delineation of affordances... 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... that stands in strict correspondence to the occurrent property.(p. 264-265)
They then explain that when you have laws relating the object-properties to the affordances and laws relating the object properties to the patterns of light, voila, you have laws relating affordances to patterns in light. It is clear that, to them, SCIENCE! = LAWS!.


Here it is, the Connecticut Approach to ecological psychology. It is, at least in my read, clearly a very restricted version of Gibson's Approach. Not wrong, per se, but a subpart of a bigger picture that thinks it is the bigger picture. It also clearly displays all the tendencies that Cutting attributed to it, but that TSRM refused to recognize when Cutting's article came out. (To be sure, R recognized it later.)

6. The intensional view of law (p. 267-273)
This section compares TSRM's system with the Establishment view criticized earlier. Much is made of certain philosophers of science's views about what should or should not count as a "natural law". The punchline is that Eco Psych produces LAWS! but the Establishment does not. There is also a short section talking about empirical evidence that one can manipulate organism's behavior by cleverly arranging things so that movements produce the opposite visual stimulation to what would be predicted by the law. Towards the end, TSRM attribute the "unrelenting insistence on a lawful correspondence" to Gibson, drawing on work from his 1950 book to his final book.

7. The scope of laws (p. 274-275)
Apparently, before a book published two years before this article, all philosophers of science were confused about laws. Luckily, F&P made the mistake of submitting their article after that book came out. Fools! Now TSRM have a book to cite in claiming that their laws really do count in SCIENCE!

See, I told you this game looked silly.

Incidentally, based on their discussion, I think the book is probably correct on the relevant points. That still doesn't justify the posturing and condescending generalities they use. All they needed was a single sentence saying "If, for some odd reason, you doubt that these types of relationships count as laws, see recent discussion by Dretske (1977)." Done. In fact, they might have been able to delete 15 pages from the article by including that sentence.

8. Misperception misconstrued (p. 275-282)
Here it gets really ugly. Ecological Psychology can handle misperception, but the approach being birthed here must contort awkwardly to do so. Several examples follow. In each case, they argue that, for a particular organism in a particular environment some sort of action is correct, that appears to us wrong. Typically it involves an organism acting in a way that TSRM can argue is "further investigation" rather than "a mistake", i.e., a shark trying to eat something that isn't food.

This is all not very convincing. If, for example, the shark is responding based on a LAW!, then why would it ever be wrong. If it is responding based on something that is sometimes right, and sometimes wrong, then isn't it not, at that instant, responding to specifying information?!?

They cite Woodbridge (1913) early on, which makes me smile because he was then engaging the New Realism of Holt and others (none of whom are cited). Alas, Woodbridge is doing phenomonology (i.e., asking how else should a stick immersed in water appear?), and TSRM have for all intents and purposes disavowed such issues, so it is unclear how that work is relevant.

9. Direct perception: the one and only gambit (p. 282-284)
Here TSRM apply a bunch of formal arguments and symbolic logic to their already fairly technical presentation in section two. "Perception", they tell us, refers to an actual relation between an actual organism and their actual environment. More specifically -
The perceiving of X-having-a by Z presupposes a law, L:
an ambient energy property e is nomically related to a in that it is unique and specific to a in Z’s niche.
Given L, ‘Z perceives X-having-a' designates an actual state of affairs if:
(i) X-having-a: is present,
(ii) the e resulting from (i) and L is available to Z,
(iii) Z detects the e defined in (ii).  (p. 283)
The term "perception", they assert, should only be used if all these conditions are met. Interestingly, I am pretty sure this is the first time they have said that an energy pattern only has to be specific to a given affordance in the organisms niche. They also point out that if we use this convention, then the term "direct perception" would be redundant, there would be no other kind to speak of.

Next, TSRM talk about all the types of situations an organism might be, wherein all these conditions do not hold. Each time they assert that the term "perception" wouldn't apply, i.e., you could not properly say 'Z perceives X-having-a'. After dumping on F&P for scores of pages, they suddenly admit that if you have (i), (ii), and (iii), but not "L", then you probably should say that 'Z infers X-having-a'! They even go so far as to allow conditions in which it would be correct to say ‘(To) Z appears X-having-a' when X is not present! But, of course, those are not things (by their definition) that an Ecological Psychologist would be interested in.

Oye! It couldn't be more obvious than it is here that some sort of break with Gibson is occurring. Gibson was interested in explaining the Visual World of the organism. Admitted, he was particularly interested in arguing that the cases of the above conditions existed, in finding examples, and in working through them. But this was a subset of a much wider set of interests, and he was clearly interested in the full range of situations in which the term "perception" was traditionally used. Except that they have demarcated a narrow world in which they know (by definition) they can find laws, I have no idea why TSRM would so limit the scope of their interests.

I'm also not sure why "in their niche" suddenly showed up to weaken the requirements for lawful relations between the properties of objects and the structure of ambient energy. It seems to me that the logic of Ecological Psychology is very different if such a caveat is allowed, and everything else in the article leads me to believe that TSRM would much rather do without it.

Also, the admission that situations very close to the ones they are interested in could rightfully be described is "inference" makes the whole paper seem like mere semantics. It sure makes it seem as if organisms regularly experience  situations of the type F&P describe. We also know that there are centuries old literature using the term "perception" to describe those situations. Could it really be just that TSRM want to study only a subset of the traditional situations, and they want to have a monopoly over the use the term "perception"?!? I don't think that it what is happening, I don't think it is mere semantics... but their admission comes out of nowhere and it sure makes the whole rest of the article seem like much ado about nothing.

10. Intensional description and conceptual ascription (p. 284-292)
This section elaborates at great length upon the claim that, properly speaking, the term 'perception' describes something about the organism-in-relation-to-its-environment, not something about the insides of the organism in isolation from the environment. Of particular note, it argues that an organism can act in a certain way with respect to objects that have property X, without needing to have a 'concept of X' 'in its head'. There is confusion over this, they claim, because we often use properties of organism (i.e., how they behave) to describe a set of objects. They show that we do the same sorts of things in other sciences with no confusion, so psychologists need to suck it up and stop being confused so easily.

Important point. Didn't need 8 pages. Didn't need technical apparatus.

11. Toward a natural basis for intentionality
How, one might ask, can animals perform actions that seem to be controlled future events - i.e., a diving bird pulls seeming to pull back its wings because it is about to hit the water. Gibson has solved this problem, they claim, by showing that "real possibilities are specified by current states of affairs." (p. 293) More specifically:
In the flowing optic array at the eyes of the diving gannet there currently exists information specific to a future encounter, viz., contact with the water. There is an optical property (the inverse of the rate of dilation of the optic array structured by fish-in-water) that is lawfully related to the property ‘time-to-contact’ (p. 293)
It seems to me that this exactly treats the optic array as an intermediary between the fish and the water, and it does so by using an example that lends itself to a linear (world-to-organism) model of perception rather than a cyclical model. TSRM have hammered home that the lawful relationships they describe mean that 'responding lawfully to the patterns' will result in 'responding lawfully to the object-properties' specified by the patterns. However, I have not seen anything to convince the reader that we should there by say 'the organism is responding to the property' itself. Something is missing.

There is another few pages of criticism of the Establishment approach, and then you get a crude prelude to Bob Shaw's "intentional dynamics" (as it will be in later decades later). The argument roughly is that to 'constrain your behavior in certain ways given certain optic variables' is to 'intend the resulting outcome'.
The intention ‘to dive.. .’ entails that the contact with the water be headfirst and vigorous, with the wings retracted st some time prior to contact. The intention ‘to alight...’ entails that the contact with the water be feet first and gentle, with the wings spread at some time prior to contact. Presumably, the times relative to contact, at which the behaviors of retracting and spreading the wings are initiated, respectively, are not identical. (p. 297)
Again, there is a bunch of unnecessary symbolic apparatus, for example, you find out that if:
c = plant stem is climb-upable,
b = plant stem is collide-withable,
and, on the occasion of the incoming tide,
I = marsh periwinkle perceives c.
I is a typical intentional statement. Its truth evaluation depends on the marsh periwinkle Z, the statement c and the occasion 0. In brief, I = P(Z,c,O) where P stands for perceives. (p. 298)
The conclusory paragraph for this section is strong enough to put here fully:
To conclude, the Establishment treatment of the intentionality problem under analysis conjures up the image of an organism on the occasion of being hungry (such as the hermit crab) moving about with a concept of food in mind and looking for something in the environment that will match this concept; or an organism on the occasion of impending danger from the approaching tide (such as the marsh periwinkle) moving about with a concept of a thing that can be climbed up in mind and looking for some thing in the environment that will match that concept. The ecological approach’s treatment of the problem conjures up a very different image, viz., of an organism, on a given occasion, moving in the context of one set of (nested) laws rather than another. The latter image expresses belief in a natural basis to intentionality whereas the former image, that of the Establishment, does not. (p. 299)
12. Postscript
Eco Psych makes sense, and is ready to rock. F&P have been taken to school!

Summary of Initial Reactions
Did I mention that the article was 68 pages long!

The only way you could justify this jumbled, bloated mess, is if you were invited to write a paper in a prestigious journal, and you feared it might be your only shot, and so you said everything you could possibly think you might want to say in that journal. Frankly, that was probably the situation. If I was in their shoes, I don't know if I could have kept a cooler head. There are a few real gems scattered throughout, but overall, it makes for something that doesn't hold together too well, and is hard to read.

One wishes that it was broken up into at least two articles (Part 1: F&P are wrong, and their system is lame; Part 2: This is Ecological Psychology). One also wishes that even for a second you could get some Dewey-esque plain language. Many of the points are so straightforward that the posturing and incorporation of technical devices really gets in the way, as do the elaborate references to now-dated debates in other fields.  (Some of Holt's works suffer from the same problems, but worse.)

All that aside, this paper successfully brought the field together, and guided 30 years of darn successful empirical work. It really is a powerhouse, and people are still wrestling with its implications. 

---------

I think this is probably the last post I will do in this series. My article following up on Cutting's paper has been submitted to The American Journal of Psychology. I added this post for completion's sake, and because I think there are deep issues here still worth dealing with.


7 comments:

  1. they give the dispositional definition of "affordances" and "effectivities" (which they have since disavowed)
    Only Reed did; TSM are still happy with the dispositional account.

    The only way you could justify this jumbled, bloated mess, is if you were invited to write a paper in a prestigious journal, and you feared it might be your only shot, and so you said everything you could possibly think you might want to say in that journal.
    I've always thought this was the case too.

    ReplyDelete
  2. Only Reed did

    I thought that too. But I have been assured at several conferences now (by Turvey, among others) that the "dispositional account" is now passe and that the entire field has moved well beyond it by now. The fact that the field clearly has not seems irrelevant.

    ReplyDelete
    Replies
    1. Weird. I know Turvey and Chemero have done some stuff trying to get show that the dispositional and relational stuff are basically equivalent (they have a painfully technical paper involving hypersets, iirc), but I had no idea Turvey had backed off the dispostions stuff.

      I actually think dispositions are the place to go; I also think that Turvey's stuff was based in a super brief discussion in Bunge and that since those papers, Stephan Mumford has gone to town on sorting dispositions out. Yet another 'will get to this one day' project of mine is to work with a philosopher friend of mine who knows about this work to update the dispositional account of affordances to reflect all this. My hunch is that most of the problems arise from the fact that the logic and rules of dispositions are a recent development that we haven't been taking advantage of.

      I don't get the dispositon hate anymore. The affordance work I'm doing with Geoff on throwing just screams dispositions over relations.

      Delete
    2. I don't hate the dispositional approach, I just think it leads places we might not want to go, and I hate that most proponents of the view are myopic to the potential problems and are therefore unwilling to discuss them. (You are certainly not on that list.)

      I have never quite figured out what it is that Turvey now claims to support, but I know he now thinks his older position had some minor flaws, and that his new thing gets around those flaws. Maybe Tony knows?

      Delete
    3. Where does it lead that's a problem?

      Delete
    4. Well, seems to me that it leads away from what would normally be thought of as "psychology", and it leads towards a most extreme behaviorism, with a risk of tautology in places. It might also risk requiring such extreme specification of conditions that it stops being useful in any practical sense.

      I will work on a post.

      Delete
  3. Then Ed was ahead of the curve. I think it's telling that Ed does not site the TSRM paper in ENCOUNTERING THE WORLD.

    -Jim Jackson

    ReplyDelete