Reading Notes: June 30th, 2022

“All [ecological theories of perception] are subject to a major liability. All of them, following Gibson, draw a distinction between direct and indirect perception, and hold that our perception of the ecological environment in normal cases is direct. This as we shall see is a distinction that rests upon a submerged and implausible theory. As that distinction and its coordinate theory sink so will the ecological views based upon it….At the beginning of Chapter 9 of The Ecological Approach Gibson draws a distinction of a highly theoretical sort, between perceiving directly and perceiving indirectly (or mediately). Here is how he characterizes the difference: “Direct perception is what one gets from seeing Niagara Falls, say, as distinguished from seeing a picture of it. The latter kind of perception is mediated. So, when I assert that perception of the environment is direct, I mean that it is not mediated by retinal pictures, neural pictures, or mental pictures. Direct perception is that activity of getting information from the ambient array of light. I call this a process of information pickup that involves the exploratory activity of looking around, getting around, and looking at things. This is quite different from the supposed activity of getting information from the inputs of the optic nerves, whatever they may prove to be.” What lies behind these remarks, surprisingly, is a classical philosophical theory, a variant of traditional Cartesianism.” (Stroll, Reflections on Surfaces, 203-205) 

“I say [that Gibson’s] remarks are surprising because Gibson presents his theory as anti-Cartesian, as a theory of information extraction or “pickup.” He argues that from different sorts of perceived objects we extract different amounts of information about the world. From Niagara Falls itself we extract more information than from a photograph of it. This is so because the difference in the amount of information picked up is partly a function of the exploratory activities a person normally engages in. One can move his or her location with respect to Niagara Falls, and thus see it from different perspectives; one can wait and watch the flow of water change, and so on. But we cannot see such a change in a photograph. There are still other differences involving color, size, smell, and so forth. To be sure, the photograph will give us some information about the Falls, but less than seeing the object itself. When the information picked up is restricted, Gibson says that we see the object indirectly. And in contrast, when the object itself is seen, and one is maximizing the information flow, in his parlance one is seeing the object “directly.” Yet despite the apparent innocuousness of this account, I will argue that the Cartesian model has its grip here. Descartes drew a distinction between “internal” mental events (so-called “ideas” or “sense-data”) that we apprehended by a person directly, and so-called “external” events (extended physical objects) that were apprehended indirectly though the intermediation of these mental events. The Cartesian thesis thus holds that our ideas (sense-data, impressions, etc.) are the direct sources of whatever information we have about the external world and that the nature of the world or its constituents must be inferred from those directly apprehended sources.” (Stroll, Reflections on Surfaces, 205-206) 

“In the twentieth century, Cartesianism takes a wide range of forms. Most cognitive scientists hold that our perception of the external environment is mediated by mental representations, which are the last events in a causal chain beginning at the external object and terminating in the brain. At this point the Cartesian story takes two different forms. On one version, it is having the representation that constitutes seeing; on the other, it is the mental representation that is seen. On both versions it is the external object that is seen indirectly.” (Stroll, Reflections on Surfaces, 206) 

“According to…Gibson’s information theory, a photography of Niagara Falls is a representation of the Falls, though of course, it is not a mental representation or mental image. Yet, paradoxically enough, his view is in some ways like Moore’s. If all one can see of an opaque, spherical object, say, from a given standpoint at a given moment, is its facing surface, then it follows that at that moment one has to be seeing the object indirectly, since some information about the object will always be lacking (about its reverse surface, interior, etc.). Directness and indirectness thus seem to be matters of degree (degrees of information extraction) on Gibson’s view. A drawing of the Virgin Mary will be less direct than a naturalistic portrait of George Bush, and that will be less direct than seeing George Bush in person. But presumably seeing the facing surface of George Bush will also be less direct than seeing his whole surface, and that will be less direct than seeing Bush himself (whatever this might mean). One can of course see the whole surface of certain objects (say a putting green or a tennis court) in a single act of perception, so the contrasting between seeing all and part of the surface of something is a sensible one. But for Gibson the latter would be less direct than the former since it conveys less information about the object having the surface. But if that is a consequence of Gibson’s theory it is not dissimilar to Moore’s, which is a specimen of Cartesianism.” (Stroll, Reflections on Surfaces, 207) 

“Instead of challenging the whole [Cartesian] model which discriminates between seeing directly and seeing indirectly, Gibson is playing on the Cartesian turf by adopting the direct-indirect distinction, only he has turned it on its head. For him it is the external thing, like Niagara Falls, that is apprehended directly and the representation, the picture, that is apprehended mediately. But the direct-indirect model is clearly playing a fundamental role in his theory….So then what is wrong with any form of Cartesianism? Its defect can be identified when we note that according to the theory we cannot see the same thing directly under some conditions and indirectly under others. We can only see our sense-data or representations directly; we can only see external objects indirectly. But these are different objects. Suppose we try to see a mental representation indirectly or a physical object directly: how are we to proceed? Shall we squint when looking at the mental representation? Shall we look harder when we look at the physical object? But neither activity will allow us to contravene the theory. Clearly, the theory makes it impossible to do so, no matter what we do. In order to see something indirectly, the object will have to be a different sort of object from that which we see directly: not a mental representation or a sense-datum, but a physical object, And, conversely, to see something directly will entail that what is seen cannot be a physical object.” (Stroll, Reflections on Surfaces, 207-208) 

“That Gibson has bought the Cartesian model is evidenced by his saying that we must look at a picture or an image, namely an object that is different from Niagara Falls if we are to see Niagara Falls indirectly. If we are actually looking at Niagara Falls (i.e., standing at a locus on the American-Canadian border) we cannot be seeing Niagara Falls indirectly. But if we look at a photograph of Niagara Falls then we are not (no matter what we do) seeing Niagara Falls directly. But still, one might ask, so what? Isn’t Gibson just correct? Aren’t we seeing Niagara Falls via the picture and hence indirectly? Here again the answer is “no,” and for this reason. In the normal use of contrasting terms, the terms apply to the same object. Thus, I can scratch this table accidentally or intentionally; I can pick up that book willingly or unwillingly. I can see that tree distinctly or indistinctly. In each case it is the same object to which the contrasting predicates apply. But somehow, I cannot do this with “directly” and “indirectly” as Gibson uses those terms. But why can’t I? Why must the terms apply to different objects: to pictures versus the real thing? The answer is that Gibson is committing a logical fallacy—a subtle one to be sure, but this is where his error lies.” (Stroll, Reflections on Surfaces, 208) 

“The fallacy consists in his presupposing that in the cases of looking at a picture of Niagara Falls and looking at Niagara Falls itself we are in both cases seeing Niagara Falls itself and it is only the mode of seeing that differs. That is what is entailed by his remark: “Direct perception is what one gets from seeing Niagara Falls, say, as distinguished from seeing a picture of it.” He thinks these cases are analogous to seeing the same tree distinctly at one time, indistinctly at another. That is, Gibson is presupposing that when we look at a picture of Niagara Falls we are perceiving Niagara Falls all right, but in a special way which he calls “mediately.” But he is wrong. For when we see a picture of Niagara Falls we are not seeing Niagara Falls at all. There is an equivocation in his account that makes it seem as if in both cases we are really seeing Niagara Falls—only, as it were, in different ways: indirectly or directly.” (Stroll, Reflections on Surfaces, 209) 

“This equivocation is papered over by his appeal to the degree of information extracted from the objects we are perceiving. But one is not literally seeing Niagara Falls when one sees a picture of Niagara Falls. For Niagara Falls is composed of water whereas the perceived object in the photograph is composed of paper and ink and hence the two objects are not identical. There is, of course, a sense in which it is true to say when one is looking at a picture of Niagara Falls that one is seeing Niagara Falls. But this use of “see” is elliptical for “seeing a picture of.” So one seeing a television film on the Roosevelt Administration might say in that elliptical sense, “I saw President Roosevelt today.’ Bit that would be short for “I saw President Roosevelt on TV today,” or some such idiom. If one had actually seen President Roosevelt today, i.e. on some date in 1991, that would entail that Roosevelt was alive in 1991, which, of course, is not so. The object on the TV screen is composed of black and white dots, but Roosevelt was not; therefore the object on the screen cannot be identical with Roosevelt. So one who says that I saw Niagara Falls today is implying that he or she was actually standing at a site on the Canadian-American border today. One can look at a picture of Niagara Falls without being at a particular place on the Canadian-American border, and is therefore on that occasion not seeing Niagara Falls in some special way, such as indirectly. Rather, one is not seeing Niagara Falls at all in that case. What one is seeing is a picture of Niagara Falls. One is thus seeing two different objects in the scenarios that Gibson gives us…The upshot is that no form of Cartesianism will accommodate the facts of perception….Hence, any ecological theory [of perception] that incorporates the Cartesian model must be rejected. So far as I know, all of them do.” (Stroll, Reflections on Surfaces, 209-210)