Experimental design: a cloak for intellectual sterility.

Lewin's (1931) warning against applying in psychology the Aristotelian notion that frequency of occurrence is equivalent to lawfulness fell on deaf ears. To-day Aristotelian modes of thought are as prevalent as ever, and are backed up by an elaborate structure of statistical methods and rules for experimental design. By reason of their logical tidiness these methods have the unfortunate effect of reducing (to a level below the threshold for effective action) the discontent that should be engendered by the rotten state of current psychological theory. What psychology needs is not statistical laws but scientific laws, that is, statements, expressed in general terms, that can generate specific predictions when specific values are assigned to the general terms. Such laws are unlikely to be found by statistical investigations, since they are concerned not with populations of systems but with single systems. Thus the law of gravitation is not a catalogue of the properties of solar systems, but it explains how any given solar system hangs together. It is argued that since mind occurs in an organic system its laws are likely to be laws of an ultrastable system (Ashby, 1952), and therefore not discoverable by statistical methods. Two kinds of investigation are required. The first is exemplified by Piaget's studies of the development of mind in individual children, and the second by Kohler's work with distorting spectacles. The first shows how the organism, considered as a multistable system, adapts itself by stages to its environment. The second involves a deliberate disorganization of both behaviour and perception, and enables us to observe simultaneously the processes of behavioural re-adaptation and the correction of perceptual distortions. It is claimed that the results are compatible with the author's theory of the behavioural basis of perception.