Reading Notes: July 12th, 2022

“As we have already noted, the semantics of representations cannot literally cause a system to behave the way it does; only the material form of the representation is causally efficacious….What actually do the causing are certain physical properties of the representation state—but in a way that reflects the representational state’s content.” (Pylyshyn, Computation and Cognition, 39) 

“The states of a computer can be given many different semantic interpretations; indeed, the same symbolic states are sometimes interpreted as words, sometimes as numbers, chess positions, or weather conditions….What determines what a given (syntactically articulated) state represents? [What] causes certain mental events to have certain contents? [According to some theorists], at least some mental contents represent certain things because they resemble them. An image of X represents X precisely because the conscious mental representations, or images, look like X. Such a view probably is not far from the common notion of visual imagery. If you were to ask a group of people how they know their image of a duck actually represents a duck, rather than, say, a rabbit, they might reply that the image looks like a duck. For several reasons, however, this answer does not explain why the image is a representation of a duck. For example, even in the introspectionist approach, the image need not closely resemble a duck for people to take it as a duck since it is their image, they can take it as virtually anything they wish; after all, the word duck refers to a duck without in any way, resembling a duck. [The] image of a man walking up a hill may look exactly like the image of a man walking backward down a hill; yet, if they were my images, there would be no question of their being indeterminate—I would know what they represented. The relation of resemblance is not well defined. Whether one thing resembles another is not a physically (or geometrically) definable property; resemblance depends on what the viewer knows or believes. To me, most birds closely resemble one another, but to a birdwatcher friend they are as different as ducks and rabbits. Resemblance cannot be specified except in relation to a viewer….Resemblance provides no basis for specifying the semantic content of mental representations.” (Pylyshyn, Computation and Cognition, 40-41) 

“Mediational behaviorists and certain speculative neurophysiologists take the position that a brain event can be said to represent something if that event is sufficiently like (possesses a subset of the properties of) the event that takes place when that something actually is perceived. Another, more radically behaviorist version requires that the mediational event evoke an internal “preparatory response” that is sufficiently like the response that would have been evoked by the corresponding stimulus. Neither position is satisfactory, because of the properties of representations we have already noted (for example, we can think about objects we have neither perceived nor have any disposition to behave toward, such as, perhaps, quarks). In any case, the only mechanism behaviorism provides for explicating the representing relation is that of association. Association, in turn, must be established by such principles as contiguity and evoked by the activation of other associated items (otherwise we would not have provided the naturalistic account of the semantics of the functional states sought by behaviorism). A chain of continuous events mediating between a brain state and an object, however, cannot form the basis of representation, for reasons discussed above, namely, that it is neither necessary nor sufficient that the state of an organism be linked by a series of contiguous events to the object that the state represents. Not only can I think of things to which I obviously am not in this sort of relation (for example, nonexistent things), but when I do think of X, I do not thereby think of associates of X; indeed, I need not think of properties that are necessarily coextensive with X, such as shape, size, weight, color, and so on. The basic problem is that representing is a semantic relation, that semantic relations, like logical relations, appear not to be causally definable…” (Pylyshyn, Computation and Cognition, 41-42) 

“Colour perception is often associated with feelings of pleasure or displeasure. Most people have preferences for certain colours rather than for others. Indeed, the order of preference seems fairly constant among Western peoples. It is as follows: blue, red, green, purple, orange, and yellow. Intermediate colours are usually felt to be less pleasant than pure colors. But there are considerable individual variations; in particular, some people seem to prefer bright colours, others softer and less saturated ones. Again, particular colours may give rise to particular emotional reactions: red to excitement or anger, blue to calm pleasure, black and grey to sadness or depression.” (Vernon, The Psychology of Perception, 71-72)