Configural Processing
Consortium (CPC) 2006: Program
Kyle Morrow Room, Fondren Library
General talks open to the
public:
Michael Kubovy,
Gestalt phenomena---such
as grouping and other forms of perceptual organization---and configural
processing---such as the perception of faces---are examples of emergent
properties (EPs). They are called EPs because one cannot predict the properties
of the complex object (a face, a melody) from the properties of the elements
(the list <eye-shape, nose-shape, …>, or the list of musical notes and
their associated durations). The classic example is our inability to predict the
state of matter of water at room temperature (liquid) from the state of matter
of its constituents (hydrogen and oxygen) at room temperature (gas). Some treat
emergence as if it were just a relation between properties of elements and
properties of what they constitute when they combine. This conceptualization is
observer-neutral; it makes claims about nature; it is ontological. I will review
the history of the notion of emergent property, and propose that claims of
emergence are not purely ontological; they must be stated with respect to a
epistemic framework. In other words one cannot assert that something is an
emergent property in a context in which the relation between elements and a
whole is fully understood. So within the framework of physical chemistry, the
relation between H, O, and H_{2}O is not an instance of emergence. I apply these
ideas to perception, and conclude with a classification of emergent phenomena,
based on two distinctions: (1) cognitive vs. perceptual emergence; (2)
eliminative vs. preservative emergence.
Daniel
Algom,
How does something like
beauty emerge? Why do some stimuli like faces or certain emblems form compelling
visual Gestalts, whereas others leave the casual observer unaffected? Why do a
brother and sister look alike? These impressions of beauty, holism, and
similarity are as immediate and compelling as are subsequent attempts at
justifying them look forced and unsatisfying. Is there a way to quantify the
processing of these stimuli, thereby distinguishing them from other
non-configural stimuli? The answer remains tentative, as attempts at deriving
objective measures of configurailty have been notoriously unsuccessful. There
has long been an informal, intuitive linkage between configurality and
singularly efficient processing produced through interaction among the various
stimulus features. Another influential idea stressed the strong perceptual glue
binding together parts of configural stimuli, compromising full selective
attention to parts. However, available measures of efficiency and/or capacity
failed to support the notion of supercapacity, or even dependency in processing
configural stimuli. In a similar vein, routine measures of selective attention
fail to distinguish configural from non-configural stimuli. A possible solution
to the conundrum implicates the tasks used. They require decomposition of the
stimulus or attention to a single dimension, thereby relinquishing possible
effects of stimulus holism. Tasks that preserve the stimulus as a whole do
reveal configural superiority.
Full program (invitation only)