In this paper, an attempt has been made to examine critically what is known concerning the mechanisms through which drugs modify personality function, in terms of subjective experience, overt performance, and neurophysiology. A study of available data and the techniques by which they have been acquired leads to several important general conclusions: (1) Not only is there a considerable degree of incommensurability between "subjective" and "objective" data, but also between data in either category that are acquired with different techniques. (2) The "organism" can never be separated from its " environment," and the 2 can be described only in terms of mutual interaction. (3) A "stimulus" cannot be defined in terms of its own properties alone, since its capacity to evoke responses is determined in part by antecedent events, and by particular experimental arrangements.These conclusions can be reconciled with a monistic theory of "mind" and "body" (43). However, it may be questioned whether concepts such as "psyche" and "soma," or their equivalents and derivatives, have not outlived their usefulness, and impede progress in psychiatry more than they foster it. An approach more consistent with the facts, which appears to be more useful in research, is one that may be called "instrumental relativity." Its salient features may be stated as follows. In psychiatry, we are concerned with the prediction and alteration of changes in the organism-environment complex at the symbolic level of functional integration. Such changes may be described in terms of various parameters, such as those of language, performance, physics, and chemistry. Each group of parameters constitutes a "frame of reference" for the measurements that are made. The data so acquired may be "explained" in terms of operational constructs ("properties," "functions," "mechanisms," "theories," " laws") that are peculiar to each frame of reference. However, as the data discussed in this paper indicate, any "mechanism" (or other operational construct) in a given frame of reference can be dissociated from all other "mechanisms" in any other frame of reference, and furthermore, the rates of change of operational constructs may vary considerably in different frames of reference. It follows therefore that perfect correlations can never be made between the data acquired with one technique and those with another, although they may be related. This is true even when, because of semantic confusion, we use the same word to describe different operational constructs—e.g., stimulation, depression, inhibition, facilitation, stress, homeostasis, energy, level of integration, etc. Also, cause-and-effect relationships between successive changes in the organism-environment complex can be inferred with confidence only with respect to such changes as are described in a given frame of reference. The use of terms such as psychosomatic in a cause-effect sense is semantically unjustified and is fraught with serious sources of error.Hence there is little justification for the despair of the neurophysiologist who felt that, as far as "mind" was concerned, the head might just as well be stuffed with cotton wool(44). Although complete equivalence of what are currently termed "mental" and "material" mechanisms can never be attained, the goal of psychiatric research must be the elucidation of mechanisms, or combinations of mechanisms, in multiple frames of reference at the symbolic level of functional integration, between which correlations can be demonstrated in increasingly high degrees of probability. We must be prepared to adopt new techniques for acquiring data, and to revise our operational constructs, including dynamic formulations and classifications of psychiatric disorders, if by so doing prediction and treatment are facilitated. Such flexibility in research at nonsymbolic levels has been the foundation for progress in other medical fields, and promises also to have great value for psychiatry.Abstract Teaser