IS THE IMAGERY DEBATE OVER? IF SO, WHAT WAS IT ABOUT?
Zenon Pylyshyn
Source: http://ruccs.rutgers.edu/faculty/pylyshyn-mehler.htm
Neuropsychological Evidence and the “new stage” of the debate
The arguments I have sketched (when fleshed out in detail, as I have done elsewhere Pylyshyn, 1981; Pylyshyn, submitted) should make it clear that a picture theory that appeals to inherent properties of a “picture-like” 2D display is no better at explaining the results of mental imagery studies than is the “null hypothesis” that the content of our images is encoded in some symbolic form which serve as the basis for inferences and for simulating various aspects of what it would be like to see some situation unfold (including the relative times taken for different tasks). The basic problem is that the phenomena that have attracted people to the picture theory (phenomena such as mental scanning or the effect of image size on reaction times for detecting features) appear to be cognitively penetrable and thus cannot be attributed to the nature of the image itself – to how it is instantiated in brain tissue – as opposed to what people know or infer or assume would happen in the real referent situation. Any attempt to minimize this difficulty, say by postulating that images are only “functionally” like 2D pictures, is powerless to explain the phenomena at all since functional spaces are whatever we want them to be and are thus devoid of explanatory force. But what about the literal conception of images as real 2D displays in the brain? This is the view that is now popular in certain neuropsychology circles and has led to what (Kosslyn, 1994) has described as the “third phase” of the imagery debate – the view that the evidence of neuroscience can reveal the “display” or “depictive” nature of mental images. So where does such evidence place the current state of understanding of mental imagery?
Neuropsychological evidence has been cited in favor of a weak and a strong thesis with little care taken to distinguish them. The weak thesis is that mental imagery in some way involves the visual system. This claim is weak because nobody would be surprised if some parts of visual processing overlaps with virtually any cognitive activity – much depends on what one takes to be “the visual system.” (for more on this question see Pylyshyn, 1999). The strong claim is that not only is the visual system involved, but the input to this system is a spatially laid out as a “picture-like” pattern of activity. Yet the evidence cited in favor of the weak thesis that imagery involves the visual system is often also taken (sometimes implicitly and sometimes explicitly) to support the stronger thesis, that images are structurally different from other forms of thought because they are laid out spatially the way pictures are, and therefore that they are not descriptive but depictive (whatever exactly that means, though it clearly implies a picture-like display).
An argument along the following lines has been made in the recent neuropsychology literature (Kosslyn et al., 1999, see also the accompanying News of the Week article in Science, April 1999, Vol 284). Primary visual cortex (Area 17) is known to be organized retinotopically (at least in the monkey brain). So if the same early retinotopic visual area is activated when subjects generate visual images, it would tend to suggest that (1) the early visual system is involved in visual mental imagery, and (2) during imagery the cognitive system intercedes and provides the visual system with inputs in the form of a topographic display, like the one that is assumed to be normally provided by the eyes – in other words we generate a display that is laid out in a spatial or “depictive” form (i.e., like a two-dimensional picture). This interpretation was also supported by the finding (Kosslyn et al., 1995) that “smaller” images generated more activity in the posterior part of the medial occipital region and “larger” images generated more activity in the anterior parts of the region, a pattern that is similar to the activation produced by small and large retinal images, respectively.
There are plenty of both empirical and logical problems with this argument [5] which I will not address in this essay (but do address in Pylyshyn, submitted). For purposes of this essay, I will put aside these (often quite serious) concerns and assume that the conclusion reached by the authors of these recent studies are valid and that not only is the visual system involved in mental imagery, but also (1) a retinotopic picture-like display is generated on the surface of the visual cortex during imagery, and (2) it is by means of this spatial display that images are processed and patterns “perceived” in mental imagery. In other words I will assume that mental images literally correspond to two-dimensional displays projected onto primary visual cortex to be reperceived by the visual system in the course of reasoning about imaginary situations. We can then ask whether such a conclusion would help explain the large number of empirical findings concerning mental imagery (e.g., those described in Kosslyn, 1980) and thus help to clarify the nature of mental imagery. The purpose of this exercise is mainly to make the point that neuroscience evidence has no more claim to primacy in resolving disputes concerning mental processes than does behavior evidence, and indeed neuroscience evidence is of little help in clarifying conceptually ill-posed hypotheses, such as those being considered in the research on mental imagery.
What if we really found pictures in primary visual cortex?
Notice that what the neuropsychological evidence has been taken to support is the literal picture-in-the head story that people over the years have tried to avoid. It is no accident that the search for concrete biological evidence for the nature of mental imagery should have led us to this literal view. First of all, our search for neural evidence for the form of a representation can be no better than the psychological theory that motivates it. And the motivation all along has been the intuitive picture view. Even though many writers deny that the search is for a literal 2D display (e.g., Denis & Kosslyn, 1999), the questions being addressed in this research show that it is the literal view of images as 2-dimensional somatotopic displays that is driving this work. Secondly, if we were looking for support for a descriptivist view it is not clear what kind of neural evidence we would look for. We have no idea at all how codes for concepts or sentences in mentalese might be encoded. Even in concrete apparently well-understood systems like computers, searching the physical properties for signs of data structures would be hopeless. If our search was for a “functional space”, which some people have suggested as the basis for images, we would still have no idea what to look for in the brain to confirm such an hypothesis. It is because one is searching for a literal 2D display that the research has focused on showing imagery-related activity in cortical Area 17 – because this area is known to be, at least in part, topographically mapped. The kind of story being pursued is clearly illustrated by the importance that has been attached to the finding described in (Tootell, Silverman, Switkes, & de Valois, 1982). In this study, macaques were trained to stare at the center of a target-like pattern consisting of flashing lights, while the monkeys were injected with radioactively tagged 2-deoxydextroglucose (2-DG), whose absorption is known to be related to metabolic activity. Then the doomed animal was sacrificed and a map of metabolic activity in its cortex was developed. This 2-DG map showed an impressive retinotopic map of the pattern in V1, with only cortical magnification distorting the original pattern. In other words, it showed a picture in visual cortex of the pattern that the monkey had received on its retina, written in the ink of metabolic activity. This has led many people to believe that we now know that a picture in primary visual cortex appears during visual perception and constitutes the basis for vision. Although no such maps have been found for imagery, there can be no doubt that this is what the picture-theorists believe is there and is responsible for both the imagery experience and the empirical findings reported when mental images are being used. I have gone into these details because many people who cite the neuroscience results deny, when asked, that they believe in the literal picture view. But the lines of inference that are marshaled in the course discussing the evidence clearly belie this denial.
.
We Make it Easy to Succeed
Successwaves, Intl.
Brain Based Accelerated Success Audios
 |