Notes On Dretske "Meaningful Perception"

Mike's Introductory Notes

There are two general things to keep in mind as this chapter is read. The first is to consider it in the context of basic dichotomies that have been emerging throughout the course: innate vs noninnate processes, impenetrable vs nonimpenetrable processes, modular vs isotropic processes, early vision vs visual cognition. The second, with respect to the specific material covered in the lecture material, is to relate Dretske's views to the problems of representation. In particular, Dretske spends a lot of time on misperception -- how does this relate to what we talked about with respect to representation? Dretske also makes a strong distinction between two different types of perception. How does this distinction relate to our discussion in class of implicit vs. explicit content of representations?

Margin Notes For The Chapter

"One can perceive an object without knowing or understanding what it is." Perceiving something, and recognizing what it is, are two different processes. Meaningful perception "embodies a judgement or belief, some degree of recognition or identification of what one is perceiving. Meaningfule perception requires more than good eyesight. It requires the kind of conceptual skills needed to classify and sort perceptual objects into distinct categories." Dretske contrasts this with what he calls sense perception; sense perception vs. meaningful perception is analogous to perception of objects vs. perception of facts about objects.

Perceiving Objects And Perceiving Facts

What does visual perception perceive -- objects? properties of objects? events? facts? This general question is important to answer, "for it is not at all clear that the same processes, mechanisms and results are, or need be, invovled in the perception of these different things."

Perhaps one approach to answering this question is to apply conventions -- for example, cognitive psychology usually equates seeing with visual cognition ("a coming-to-know-by-visual-means"). Problem with this view is that we have to account for misperception, which is difficult to do if perception is equated with perceiving facts about objects. Dretske's solution is to draw a principled (NB: not conventional!) distinction between sense perception and meaningful perception. "It seems preferable, therefore, to distinguish between seeing objects and seeing facts, not by artificially reserving the word perception for one way of seeing, the way of seeing that requires knowledge of the thing seen (i.e., seeing facts), but rather by distinguishing two forms of perception, two ways of seeing."

The study of early vision is the study of sense perception. The study of later vision is the study of meaningful perception. So..."debates about whether perceptual processes are top-down or bottom-up, about whether they are inferential or constructive in character, about whether they are massively parallel or sequential, and about their comparative modularity, are topics that can be given sharper focus by distinguishing the kind of perception -- sensory or meaningful -- the debate is a debate about."

What We Perceive: The Objects Of Perception

"Meaningful perception exhibits a hierarchical structure." Because of this structure, some visual facts depend upon others. Therefore, are some of these visual facts basic? In other words, does all meaninfgul perception depend upon a prior knowledge of how things seem?

"The answer to this question depends on the answer to a somewhat different question, a question about sense perception. What objects do we see? .... Insofar as we regard the image appearing on our television or movie screen as the primary, or real, object of perception, we regard facts about these images as cognitively primary. Facts about the people and events being rerpesented are secondary." So, we need to study the structure of sense perception before we can make sense of the structure of meaningful perception. "Discussions of these issues are often clouded by failure to appreciate the difference between meaningful perception and sense perception."

For instance, consider the debate about sense perception being direct or not. "One might be a direct realist on sense percpetion, but an indirect (representational) realist on meaningful perception." This makes the debate about direct vs. indirect perception extremely technical!

How We Perceive: The Perceptual Process

"The debate about the objects of perception is realted to a debate...about the kind of processes underlying perception." Move often is to use "semicognitive" language to describe perceptual processing or the processing of simple machines. Is this anything more than a metaphorical device? Many modern researchers, because of the problem of underdetermination, believe that perception is a form of problem solving. Gibsonians challenge this view. The notion of modularity plays a role in this debate -- modules are isolated from lots of information. "Modular systems are therefore most naturally thought of in the second of the two ways described above -- as good extractors of preexistent informatioin, informaiton that is already in the stimuli, not as good detectives or problem solvers about the best interpretation of informationally ambiguous stimuli." (NB: Here, I think, Dretske is missing some of the technical points of Ullman's stuff, for instance, which is both antiGibson and anticognitive!)

All of this isnt'e helped by the tendency to confuse sense perception with meaningful perception. Real issue is whether sense perception is inteligent!

Perceptual Change: The Relativity Of Meaningful Perception

Do we learn to perceive objects? First, you have to be clear about what kind of perception you are concerned about here. Meaningful perception is learned -- but what about sense perception? Dretske argues that sense perception is much less likely to be learned.

Is perception relative, depending upon our beliefs? Sense perception does not appear to be relative in this way. For Dretske, sense perception is cognitively impenetrable. "Sense perception is comparatively modular. It isnot sensitive to those cognitive influences (e.g., one's language, conceptual scheme, scientific worldview) that affect one's perception of facts."

Bottom line -- questions are only sensible int he context of distinguishing sense perception from meaningful perception. "One invites only confusion by ignoring the distinction."

Pearl Street | "An Invitation To Cognitive Science" Home Page | Dawson Home Page |