ABSTRACT
Thus as a rule perception of form in everyday life does not involve the accurate discrimination of minute detail which was described in the previous chapter. Nevertheless, the capacity is available, and may be utilized for instance in certain scientific activities or in the estimation of precision measurements. In general, however, there is a redundancy of sensory information available, much of it corroborative, and the task of the observer is to select what is relevant to his identification of the objects and events visible, and to react to them appropriately. As we have noted, he makes use of schematized knowledge as to the type of situation and the relevant pattern of response. However, there may be occasions on which information is restricted or conflicting; and perhaps more commonly when it is so abundant that he must select what is relevant and discard what is irrelevant. In such situations he may make inferences as to the real nature of the situation or of the objects presented
to him, going far beyond what is given by the immediate sensory data. In fact, as Brunswik (1956) pointed out, what is immediately perceived seldom provides a completely veridical impression of the stimulus situation. Thus in making inferences, the observer must extract such information as gives him the truest impression, not of the immediately perceptible aspects of the stimulus but of the fundamental nature of the situation within which the stimulus occurs. This process of inference is clearly demonstrated in the 'constancies' (see Chapter VIII). It was also shown by Craik and Vernon (1942) to occur when silhouette pictures were presented in very dim light. The perception and identification of these depended only in part upon the absolute threshold for vision. If observers could make out the general nature of the pictures, they were able to infer their identity at an intensity of illumination below the cone threshold. Later, Vernon (1947) found that the identification of silhouette pictures, blurred by dazzling light, was related to factors such as intelligence. Again, inference may over-ride the erroneous impressions given by the visual illusions. Thus even if the Poggendorfillusion persists in pictures such as that in Fig. 21g (p. 55), we should be unlikely in real life to suppose that the rope was broken or discontinuous. Identification through inference may remain unchanged even when different perceptual cues are utilized. Thus Brand (1954) exposed some of the Rorschach ink-blots on two successive occasions, a week apart. Though different details of the blots were discriminated on successive occasions, the interpretation in terms of resemblance to real objects varied little if at all.
