Conscious Realism and the Mind-Body Problem - UCI Cognitive ...

106 HoffmanHere there is good news. We have substantial progress on the mindbodyproblem under conscious realism, and there are real scientific theories.We now have mathematically precise theories about how one typeof conscious agent, namely human observers, might construct the visualshapes, colors, textures, and motions of objects (see, e.g., Hoffman 1998;Knill and Richards 1996, Palmer 1999).One example is Ullman’s (1979) theory of the construction of threedimensionalobjects from image motion. This theory is mathematicallyprecise and allows one to build computer-vision systems that simulate theconstruction of such objects. There are many other mathematically precisetheories and algorithms for how human observers could, in principle,construct three-dimensional objects from various types of image motions(e.g., Faugeras and Maybank 1990, Hoffman and Bennett 1986, Hoffmanand Flinchbaugh 1982, Huang and Lee, 1989, Koenderink and van Doorn1991, Longuet-Higgins and Prazdny 1980). We also have precise theoriesfor constructing three-dimensional objects from stereo (Geiger et al. 1995,Grimson 1981, Marr and Poggio 1979), shading (Horn and Brooks 1989),and texture (Aloimonos and Swain 1988, Witkin 1981). Researchers debatethe empirical adequacy of each such theory as a model of humanperception, but this is just normal science.Almost without exception the authors of these perceptual theories arephysicalists who accept HFD and conceive of their theories as specifyingmethods by which human observers can reconstruct or approximate thetrue properties of physical objects that, they assume, exist objectively,i.e., independently of the observer (a claim about physical objects thatis explicitly denied by conscious realism). But each of these perceptualtheories can equally well be reinterpreted simply as specifying a methodof object construction, not reconstruction. The mathematics is indifferentbetween the two interpretations. It does not require the hypothesisof independently existing physical objects. It is perfectly compatible withthe hypothesis of conscious realism, and the mind-dependence of all objects.So interpreted, the large and growing literature in computationalvision, and computational perception more generally, is concrete scientificprogress on the mind-body problem, as this problem is posed by consciousrealism. It gives mathematically precise theories about how certain consciousagents construct their physical worlds. The relationship betweenthe conscious and the physical is thus not a mystery, but the subject ofsystematic scientific investigation and genuine scientific theories.What one gives up in this framework of thinking is the belief thatphysical objects and their properties exist independently of the consciousagents that perceive them. Piaget claimed that children, at about ninemonths of age, acquire object permanence, the belief that physical objectsexist even when they are not observed (Piaget 1954; but see Baillargeon1987). Conscious realism claims that object permanence is an illusion. It