50 Informational environments
of some object provide only a set of boundary conditions that underdetermine
the possible ways of perception. At any rate, they are not represented in vision
(Purves et al. 2011, 15592).
In this circumscribed sense, the Empirical Strategists share Gibson’s view that
perception is not a process of representing the physical properties of some object
or scene but an interaction with some concrete object in its concrete context under
concrete conditions of perception. What one can do with or about that object in a
given situation or, in Gibsonian terms, what that object affords to the organism is
what fixes the way in which it is perceived.
As conceived under the Empirical Strategy, the relation between perceptual
qualities and their target appears at once rather simple and fairly complex and
abstract: some perceptual token will be reproduced if it is closely associated with
successful behaviours towards its source or towards some correlate of that source.
This condition is sufficient for a very specific sort of empirical adequacy. This
empirical adequacy is very specific because it does not build upon any direct relation – and perhaps not even a proper covariance – between a perceptual quality
and the physical properties of the perceived object. Divergence between perceptual qualities and the physical conditions at the source does not amount to misperception: “Since the measured properties of objects are not perceived, they cannot
be misperceived” (Purves et al. 2001, 296).
Instead, what is decisive for getting things right in perception are the frequency
distributions of different retinal patterns, which are mapped not onto variance in
physical variables but onto variance in behaviours that differentially respond to
certain world affairs, in accordance with the probability distributions of the occurrence of these affairs (Purves et al. 2011, 15594). For example the observable
mismatch between differences in lightness or brightness of an object as perceptual
qualities on the one hand and measured illumination and luminance of the physical objects on the other is attributed to the relation between two factors:
( f 1) Frequency distributions can be determined for variant luminance values of
some object as they obtain for the contexts of the various natural scenes in
which it, individually or as a member of a type, appears. Objects of some
kind will be more often encountered under certain lighting conditions than
under others.
( f 2) Frequency distributions are assumed for the rates of success of behaviours
of the perceiving organism or his ancestors towards that kind of object
under variant conditions. These frequency distributions will be affected
by processes of selection, on phylogenetic or ontogenetic levels, of variant
behaviours.
These two types of frequency distributions can be mapped onto each other, so
as to see how reliable behavioural success with respect to the object in question will be under the predominant conditions of appearance in the perceiving
organism’s environment, and what the cost of failure under less frequent conditions will be. The general strategy is to match perceived qualities of some object
of some object provide only a set of boundary conditions that underdetermine
the possible ways of perception. At any rate, they are not represented in vision
(Purves et al. 2011, 15592).
In this circumscribed sense, the Empirical Strategists share Gibson’s view that
perception is not a process of representing the physical properties of some object
or scene but an interaction with some concrete object in its concrete context under
concrete conditions of perception. What one can do with or about that object in a
given situation or, in Gibsonian terms, what that object affords to the organism is
what fixes the way in which it is perceived.
As conceived under the Empirical Strategy, the relation between perceptual
qualities and their target appears at once rather simple and fairly complex and
abstract: some perceptual token will be reproduced if it is closely associated with
successful behaviours towards its source or towards some correlate of that source.
This condition is sufficient for a very specific sort of empirical adequacy. This
empirical adequacy is very specific because it does not build upon any direct relation – and perhaps not even a proper covariance – between a perceptual quality
and the physical properties of the perceived object. Divergence between perceptual qualities and the physical conditions at the source does not amount to misperception: “Since the measured properties of objects are not perceived, they cannot
be misperceived” (Purves et al. 2001, 296).
Instead, what is decisive for getting things right in perception are the frequency
distributions of different retinal patterns, which are mapped not onto variance in
physical variables but onto variance in behaviours that differentially respond to
certain world affairs, in accordance with the probability distributions of the occurrence of these affairs (Purves et al. 2011, 15594). For example the observable
mismatch between differences in lightness or brightness of an object as perceptual
qualities on the one hand and measured illumination and luminance of the physical objects on the other is attributed to the relation between two factors:
( f 1) Frequency distributions can be determined for variant luminance values of
some object as they obtain for the contexts of the various natural scenes in
which it, individually or as a member of a type, appears. Objects of some
kind will be more often encountered under certain lighting conditions than
under others.
( f 2) Frequency distributions are assumed for the rates of success of behaviours
of the perceiving organism or his ancestors towards that kind of object
under variant conditions. These frequency distributions will be affected
by processes of selection, on phylogenetic or ontogenetic levels, of variant
behaviours.
These two types of frequency distributions can be mapped onto each other, so
as to see how reliable behavioural success with respect to the object in question will be under the predominant conditions of appearance in the perceiving
organism’s environment, and what the cost of failure under less frequent conditions will be. The general strategy is to match perceived qualities of some object