Visual Imagery and Consciousness
Nigel J.T. Thomas
Page 6
http://www.imagery-imagination.com/viac.htm
If the symbolic tokens that comprise computational representational systems represent as words do, then it seems to follow that, by analogy with the way language represents visual scenes, the brain-computer represents visual scenes with descriptions. From this perspective, such mentalese descriptions are the end-product of perceptual information processing, and thus responsible for our perceptual experiences. Of course, we do not experience the descriptions as being descriptions, but, it is assumed, they nevertheless constitute our perceptual experience. If mentalese descriptions of visual scenes are retrieved from memory rather than being the result of present sensory input, or if they are constructed out of bits and pieces of various descriptions in memory, then we have the experience of mental imagery.
Pylyshyn claims that, since descriptions can be partial, incomplete, and can leave out all sorts of information (not only details, but sometimes even facts about the global structure of a scene) description theory can explain the frequently indefinite and ambiguous nature of our imagery better than picture theory can. However, the theory's main motivation is clearly the belief that it better respects fundamental facts about the nature of computational representation.
If anything, however, description theorists have paid even less heed than quasi-pictorial theorists to the question of how images (or, come to that, percepts) come to be consciously experienced. Clearly, we are not aware of mentalese descriptions as such, but the idea seems to be that we can, at least sometimes, be consciously aware of what they represent. Like the quasi-pictorialists, however, description theorists appear to be implicitly committed to the view that mental representations, including the elementary symbolic tokens of mentalese, and the symbol complexes that make up visual descriptions (and, thus, mental images), are physically instantiated as patterns of brain excitation (just as representations in a computer are instantiated as electrical charge patterns in RAM chips and CPU registers). This brings us to exactly the same "hard problem" as before: it seems impossible to conceive how patterns of brain excitation (which, themselves, are reducible to movements of ions, etc.) could be, or give rise to, conscious experience as we know it.
Enactive Theory
Quasi-pictorial and description theories are both attempts to explain how imagery, quasi-visual experience, might be explained within the context of an information processing theory of perception. They disagree over whether it should be identified with representations from an early or late stage of visual processing, and over the format of those representations. The enactive theory of imagery depends upon a quite different way of conceiving of perception, one that was pioneered in the twentieth century by James Gibson, and is more recently exemplified in "active vision" techniques used in robotics, and the "sensorimotor" or "enactive" theory of perception advocated by J. K. O'Regan and Alva Noë, amongst others. Instead of regarding vision as, at root, a matter of information flowing in through the eyes into the brain, and toward some internal center of consciousness, enactive theory regards it as a matter of the visual system actively seeking out and extracting (or "picking up") desired information from the environment. Seeing is not like taking a photograph (even a digital photograph whose file gets passed on to a computer for further processing); it is more akin to performing a series of (many, rapid) scientific tests and measurements on the information-bearing ambient light that surrounds us (what Gibson called the "optic array"). Rather than being the passive reception of information, vision (and perception in general) is a purposive process of, as it were, asking questions about our surroundings, and actively seeking out the answers. Visual consciousness is the experience of being engaged in this exploratory, questioning, information seeking activity, involving not only the making of the tests and measurements themselves, but also the continual adjustment of our expectations, and thus our exploratory behavior itself, in response to their findings.
Consider how we identify an object by touch. If something is simply pressed against our skin, we can tell very little about it. However, if we actively explore it, moving our fingers around to feel its shape and texture, seeking out corners and edges, squeezing it to assess its hardness, and so on, we can discover a great deal. What we learn derives not simply from the sensations we feel in our skin, but, crucially, from the way those sensations change in response to the purposeful, information-seeking movements of our fingers. Enactive visual theory holds that vision works in fundamentally the same sort of way, except that the exploratory, information-seeking behavior mostly happens so quickly and automatically that its details are unavailable to introspection.
Eye movements are perhaps the most obvious (and experimentally accessible) aspect of this visual information-seeking behavior – our eyes constantly, and purposefully, but largely unconsciously, flit rapidly around to take in information from different points of interest – and the enactive theory of imagery finds support in a number of recent experiments showing that the stimulus-specific eye movement patterns produced when a subject examines a complex visual stimulus are (quite unconsciously) re-enacted when the same subject later forms a memory image of that same stimulus.
However, eye movements cannot be the whole story. There clearly are bottom-up, inner representations involved in human vision. For example, the pattern of illumination on the retina does indeed produce a corresponding pattern of excitation in the visual cortex, and this carries information about the scene momentarily before our eyes. From the enactivist perspective, however, such representations are not conscious or even mental in any very meaningful sense. Quasi-pictorial theory seems to imply that, when we see, these representations are what we are conscious of. Enactive perceptual theory denies this, and holds instead that we are conscious of the actual things out in front of our eyes. Like the optical image formed inside a camera, the excitation pattern of visual cortex carries information about our surroundings, but this only gives rise to visual knowledge and awareness inasmuch as this data source is purposefully explored, searched, and queried. It is this active, exploratory searching that turns the mere passive reception of visual information (something that a camera can do) into true perception.
It is important to note, also, that the particular pattern of this exploratory questioning, and the particular tests and measurements brought into play (both external eye movements and internal data analysis processes), will be quite different when different sorts of things are being looked at. For example, in order to see a cat we would need to go through a set of exploratory actions appropriate to cat-seeing. These might include moving our eyes to focus on the likely location of particular characteristic cat features (pointed ears, gently curving tail, etc.), but will also undoubtedly include purely internal processes whereby further questions are answered by querying the data flowing bottom-up into the brain. The key idea is that there is a specific structured set of exploratory queries, what has been called a visual routine, appropriate to cat seeing (or even Tiddles-curled-up-asleep seeing), and a different set for each other type of thing that we are able to recognize.
So, if visual experience arises from the active seeking for, and finding, of information, imagery, quasi-visual experience, arises when we actively seek certain information, and persist in going through the motions of looking for it, even though it is not there to be found. To visually recognize a cat is to run through a specific, cat-recognizing visual routine, and to imagine a cat is to perform (or partially perform) this same visual routine when there is no cat present. The experience is different from actual seeing because there is no perceptual feedback from the cat itself (and we must, as it were, force ourselves to go on looking for cat features even though it is becoming ever more clear that we will not find them) but it is similar to real cat-seeing inasmuch as the pattern of perceptual exploration that we perform is similar. Arguably, this reflects the phenomenology of imagery, which is subjectively both like and unlike true seeing. In particular, it usually seems to take much more effort to sustain an image in consciousness for any length of time than it does to simply keep looking at something.
We Make it Easy to Succeed
Successwaves, Intl.
Brain Based Accelerated Success Audios
 |