On the Representation of Objects in the Visual System
Files
Publication or External Link
Date
Authors
Advisor
Citation
DRUM DOI
Abstract
There is evidence in the psychological literature for representations of objects (Pylyshyn's visual indexes) that refer to and track, not properties, but what in our sort of world typically turn out to be individual physical objects. I am concerned with how such representations acquire their content.
Two strategies for accounting for the content of representations are a) representations of particulars refer to the entity that caused them; and b) representations of particulars refer to the entity whose properties are represented by the visual system. The first strategy faces the "which link" problem: since any one of the links in the causal chain leading to the token representation counts as a cause of the token representation, no particular link is individuated as the referent. I examine a recent proposed solution to this problem (Fodor's counterfactual triangulation) and conclude that it fails to determine whether the referent of a visual index is an object, as opposed to a state of affairs, or an event.
The problems with the first strategy are a reason to explore the second strategy: representations of objects refer to the entity whose properties are represented by the visual system. I adopt Fodor's asymmetric dependency account (ADA) of intentionality to account for how representations of properties get their content. Fodor's account is chosen not because it is free of problems, but because it has the structure of a theory that promises to deal with many of the classic problems that befall informational semantics (e.g. the disjunction problem).
Since ADA is designed to work for causal relations between properties and not for causal relations between particulars, it cannot, by itself, account for how representations of particulars get their content. So I suggest that ADA be supplemented with conceptual role semantics to account for the logico-syntactic roles of representations of particulars. In particular, I suggest that to represent objects the visual system requires the capacity to form and store in memory definite descriptions containing: a) predicates referring to spatio-temporal relations; and b) temporal indexicals.