April 1975
The Psychophysical Experiment and the Perception Experiment
J. J. Gibson, Cornell University
The World Wide Web distribution of James Gibson’s “Purple Perils” is for scholarly use with the understanding that Gibson did not intend them for publication. References to these essays must cite them explicitly as unpublished manuscripts. Copies may be circulated if this statement is included on each copy.
The difference between a traditional psychophysical experiment in which a variable of physical stimulation is applied to an observer and then the corresponding variable of his phenomenal (or “sensory”) experience is established, and the kind of experiment that is required by an information-based theory of perception can be illustrated as follows.
In traditional psychophysics one would present a systematically varied parameter of optical (or “retinal”) motion to an observer and try to establish from his reports what the variable of phenomenal motion is. The assumption is one of psychophysical correspondence. The Uppsala program of research on motion perception, or event perception, seems to be of this sort. On this tradition one would never think of trying to establish what gives rise to the perception of ego motion or locomotion of oneself through the environment (e.g., R. Warren). But in information psychophysics (if it can be called that) one tries to display information rather than arouse sensations. That is, one ties to stimulate natural perception. The assumption is not one of correspondence between physical and phenomenal variables but one of specifyingcapacity of invariant variables in an array. Hence an experimenter tries to isolate and systematically vary some parameter of an ambient array, not of a stimulus, leaving it up to the observer to pick it up if he can. He would try to display an optical flow-pattern to an observer which would specify for him both ego-locomotion in a layout of surfaces and the layout of surfaces. He constructs apparatus for the experimental study of perception-and-proprioception (memo of May 1967). Is this control of the optic array that he makes available to the observer a “psychophysical experiment”? Not in the traditional sense, surely, and not in the sense of epistemological dualism, but it is an experiment. How do we characterize this new kind of experimentation?
Psychologists have not faced up to the difficulty of systematically controlling stimulus information. They have used only ready-made information in verbal items (written or spoken language) and have left experimentation up to the unsystematic artists and musicians. The dodge this issue by calling such items “stimuli” and then confuse everybody by calling the inputs sensory “signals” which allows them to beg the question of what constitutes information for perception. But this will not do.