December 1968
The Construction of Meaning vs. the Detection of Meaning
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 References to these essays must cite them explicitly as unpublished manuscripts.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 construction of meaning
If stimuli are meaningless in themselves, meaning has to be supplied by the animal receiving the stimuli. This has almost been taken for granted in the theory of sense-perception. The exciting of receptors as such cannot specify objects or events in the world, that is, the sources of the stimuli. What could specify them so that meaning is achieved in perception? There are many answers.
(1) the pattern of excited receptors. The trouble is that this pattern changes with every movement of the observer.
(2) the “field organization” in the brain to which the pattern of proximal stimulation gives rise. This is Koffka (Köhler) but the trouble is it cannot be literally true.
(3) the connections in the brain from the sensory inputs, or the memories that have been associated with them. This is Titchner’s context theory of meaning, the climax of classical empiricism, but Koffka refuted it.
(4) the motor responses to which the sensory inputs of the brain give rise. This is the motor theory of meaning. Meaning depends on habits of responses, on association learning, or on response-tendencies.
(5) the concepts of the mind that exist prior to all sensory impressions. This is nativism or rationalism. It has all sorts of unwelcome implications.
(6) a process of “inferring” the existence of objects and events from the sensory impressions of the patterns of receptor-input. This theory considers the inputs to be clues (cues). It is Helmholtz’s ingenious combination of rationalism and empiricism. The assumption is usually that the inputs are analogous to signals or messages in code, the meanings of which are learned from “past experience” in the same way that children learn the “speech code.” Despite the plausibility of this theory and its widespread acceptance, it has difficulties.
The detection of meaning
If we assume that, although stimuli are meaningless, the available stimulus information in the flowing sea of ambient energy is meaningful (stimulus ecology), then meaning does not have to be supplied by the animal but instead can be detected. Stimulus information can be defined as invariants in the flow of ambient energy. Objects and their classes are assumed to exist independently of any observer; events and their causal chains are assumed to occur whether or not they are observed. The invariants in the flow that specify persisting objects, and the unique departures from these invariants that specify events in the world, are assumed to be available for any observer.
The question now becomes: how can an observing animal detect meaning? Do any of the mechanisms by which he has been supposed to construct or supply meanings help to answer this question?
Invariants under change seem to be detected most readily when the pattern of the stimulus array is changed. This arises from exploratory movement of the observer (as in Held’s experiments) and even from the observer’s being passively moved about, and it can obviously arise also from the external motions of objects themselves. This requirement for change in order to detect unequivocal information about objects (if it is a requirement) is not elucidated by any of the old ideas about perceptual mechanisms. Or do I fail to see how the old ideas might be applied to the new problems?