Judging Explanations of Behavior: Issues in Establishing Taxonomy, Hierarchy, and AdmissibilityRoger D. Carlson
Paper presented at the First Annual Meeting of the American Psychology
Society, Arlington, Virginia, June, 1989.
AbstractExplanations of behavior are wide ranging (e.g., scientific, historical, introspective, contextual, experiential). The adequacy of various explanations is debatable. This paper examines the possibility of developing a taxonomy of explanation, a hierarchy of desirability (or "goodness") of explanations, and standards for the admissibility or acceptability of an explanation for human action.
Psychology is generally regarded as having as one of its goals the explanation of human action. What constitutes "explanation" is often debated. Psychologists having different theoretical or applied approaches debate as to what explanation is the "best." Skinnerians, physiological psychologists, and cognitive psychologists debate over the adequacy of their respective explanations. Psychoanalysts, behavior therapists, humanistic psychologists, and cognitive therapists debate over which explanatory approach to therapy is the "best." Lay people wonder why psychologists sometimes have a tendency to belabor the obvious, explain away a phenomenon, or "beat a dead horse." They also seem puzzled as to why there is a multiplicity of theories and explanations for behavior about which psychologists have a hard time agreeing. At the bottom of such debates is the question of what is considered an adequate explanation. Often what is considered adequate is probably guided by the needs and purposes of the explainer.
The purpose of this paper is to examine the issues involved in creating a taxonomy, hierarchy, and standards for admissibility of an explanation for human action.
Methodological Note. This paper is concerned with the fundamental assumptions which are inherent in the ways psychologists pursue knowledge and understanding. The approach used herein in analyzing psychological explanation is not an empirical approach, but rather conceptual analysis which is commonly used by philosophical psychologists. In doing such analysis the author is drawing upon the methods of the later work of the philosopher Ludwig Wittgenstein (1953), and those of psychologist/philosopher Michel Foucault (1970). Wittgenstein's approach, known as ordinary language philosophy, was fundamental in discovering critical flaws in the philosophical foundations of logical positivism (Stevens, 1963, Hartnack, 1962) upon which most modern psychological methodology and theorizing is based. Wittgenstein argues for consensuality of knowledge as opposed to knowledge universals (or "logical atoms" derived by either experimental or logical analysis [Russell, 1956]). Likewise, Foucault's work which he called archeology of knowledge, endeavored to describe the taxonomy of knowledge/power systems developed by cultures which contribute to the creation of discursive and political practices.
The author's working assumption is that before methods and theoretical practices of psychology are accepted at face value, the assumptions entailed in those methods and theoretical practices ought to be critically examined.
Varieties of Explanation. There are numerous kinds of explanations of human action. This paper is not meant to be encyclopedic, but rather to use illustrative types of explanations which should be considered in developing a taxonomy, a hierarchy, and standards for admissibility.
Scientific explanations in psychology come in many varieties. Radical behaviorists are satisfied that behavior is adequately explained when a functional relationship is described. Prediction and control of behavior are what constitute "explanation" (Skinner, 1953). Such prediction and control elucidates the causal variables of which behavior is a function. By way of contrast, cognitive psychologists often do not find such behavioral explanations adequate, and seek to clarify such functional relationships by making reference to cognitive hypothetical constructs (e.g., schema [Bartlett, 1932], scripts [Schank & Abelson, 1976], and the like).
Commonly it is suggested that debates about theoretical adequacy are resolved by appeal to such criteria as parsimony, generality, and heuristic value (Marx, 1963). The use of such criteria have a long tradition in the history of science (Hempel, 1966), and in particular, logical positivism (Stevens, 1963). The difficulty with the use of such criteria in psychology is that although their use is often espoused, resolution of controversies between theories are rarely resolved by either their tacit or explicit use. The debates between adherents to differing theoretical approaches often resemble debates between holders of ideologies rather than dispassion and impartiality usually ascribed to scientists and professional practitioners. The adherence to a particular theoretical viewpoint probably has a lot to do with the purposes that the explainer has for the explanation rather than in any intrinsic worth of the explanation itself.
Other kinds of explanation are the kinds of explanations given by astronomers which are not causal but descriptive. Such goodness of fit is approximated by use of the psychological model.
In practice, the clinical psychologist has different kinds of problems than the psychological theoretician or model builder with respect to explanation. Explanation within the clinical context is often said to be achieved when the clinician discovers a history, psychological story, or habit patterns which makes intelligible or understandable the current behavioral displays of a client. Because the explanatory approach is directed toward the individual, it is an idiographic (vs. nomothetic) approach to explanation (Allport, 1961). The clinician endeavors to find ways in order to minimize or redirect the deleterious effects of the antecedents to the condition which are specific to the client's life. Because the purpose for explanation is goal directed for the clinician, it is questionable whether or not direct applications from scientific theory to practice are possible. Scientific concepts such as "scripts," "beliefs," "cognitions," "reinforcers" may help guide the clinicians hypotheses, but the specific explanatory power of such concepts comes alive only when the particulars of the client's life become explicated.
Related to the sort of explanation that the clinician is trying to obtain is the type of explanation given by a detective as to how or why a crime was committed, and by whom. In such instances explanations are more in the form of hypothetical creations which are testable against the criterion of proof (which has as a goal, substantiated confession). Truth comes in the form of a theory (or hypothesis) that is shown to be the case. The kind of truth ultimately achieved is a good deal more substantial than the kind of truth achieved by the model-building psychologist. The kind of truth which is discovered by the clinician is a kind of informative truth. By way contrast, a theory of the origins of criminal behavior is an elucidating truth leading to understanding. Nevertheless it should be noted that truth for the detective, the clinical psychologist, and the psychological theoretician is a function of a particular ontological and metaphysical perspective--a perspective created by the discursive character of the culture of which the explainer is a part--which includes the constituent culture building blocks of truth, knowledge, and language with which the explainer creates an explanation (Szasz, 1960, Foucault, 1965, 1970).
The gulf between the kind of discoveries which lead to advances in scientific theory or models, and the kind of advance which is made by the clinical psychologist might be described as the distinction between advancing knowledge versus information. The kind of acknowledgement that is shown by the person who agrees that he now understands in a way that he hadn't before by virtue of the development of a new theory or model, has new insight--a new perspective. When something has been discovered (or uncovered) as a new "truth," then information has been advanced. For example, the development of cognition as an important mediator in the understanding of depression led to profoundly important understandings of the ways in which negative life events and thoughts can be thought of as interrelated but the specific combination of attributional factors predisposing one to depression in the face of negative life events is informative. In the first case, our understanding is advanced; in the second case information is advanced. While the former is useful in terms of enlightenment, the latter is useful in terms of practice.
Ordinary explanations for human actions given and accepted by non-psychologists every day. Many ordinary explanations given by people daily would not be considered adequate by psychologists. The average person does not need or require the kind of explanatory rigor that is required by most psychologists. Why is that the case? Frequently behavior is said to be "explained" when there is no puzzlement left in the mind of the person making inquiry. Indeed, many ordinary behaviors are understood as being so much a part of the culture that explanation of them seems unnecessary, trite, trivial, and superfluous. Flying a kite, standing with one's hands in one's pockets, wearing a shirt (or not wearing a shirt) are all instances in which further explanation would seem superfluous if not trivial. Of course understanding of these behaviors is always context dependent. For example, flying a kite in the middle of a hurricane, standing with one's hands in one's pockets when being asked to play a drum, and not wearing a shirt while one is on a ski slope are all situations which probably would not be understood in context and thus have the need for further explanation. It is in instances that elicit puzzlement in people that psychologists are queried about the origins of a particular behavior. Behavior exhibited in unusual contexts is one instance which elicits puzzlement in the observers and thus invites further explanation (Leifer, 1964).
Standards for Admissibility. Before considering how to prioritize and create a hierarchy for various kinds of explanations, one needs to decide what is minimally acceptable to qualify for admissibility as an explanation of human action. At a minimal level, the use of a consensual criterion (Wittgenstein, 1953) seems necessary. That is, if people generally agree that what is offered as an explanation is what we would ordinarily call an explanation in ordinary language (family resemblance), then it ought to be admissible as an explanation. If we would agree that what is offered as an explanation is what we generally would accept as an explanation, then we have a first step to comprehensibility. Of course such a consensual step for admission is necessarily based upon a culture's conceptual and discursive habits in a particular time and place. It is necessary that there be recognition of other cultural practices in other times and places making the behavior comprehensible in a way which would be satisfactory to that culture.
Ultimate definition of what constitutes "explanation" is avoided here out of recognition that definition itself is based upon use. (Not left unnoticed is the fact that the very words which I presently write are based upon the cultural fabric of relativism in which I live. Certainly such a relativistic stance is not a historical and cultural constant, and thus dismissing universalism is not possible. Only relativism can provide for such an inclusive explanatory system. But as such it is necessary to universally posit relativism to provide for such inclusion. Such a positing is justified by the generally accepted desirability of acknowledging different perspectives of psychological explanation rather than closing court on the issue whereby limitations in outlook become paramount. Universalists who become committed to a singular perspective or paradigm (e.g., radical behaviorism) do not allow for the kind of flexibility which is endemic to and necessary for the scientist's continual search for truth.)
Taxonomy. While this is not the forum for the development of a specific taxonomy of explanations, it must be remembered that such a taxonomy ought to be developed in a lateral manner. Rather than prematurely assume superiority of scientific over lay explanations, both the history of such taxonomical systems (e.g., Foucault, 1970) and the use of consensual criteria for their acceptability necessitates the admission of a variety of explanatory approaches that need to be equally considered before arbitrarily making decisions based upon history, tradition, and prevailing thought. There are language using constituencies who would consider admissible explanations of the following sort: scientific, historical, introspective, contextual, experimental, theological, astrological, characterological, mythological, etc. The attributes and the social/political, epistemological, and metaphysical functions of each ought to be given an honest hearing before developing a hierarchy. The development of a lateral taxonomy of such explanatory approaches would serve the function of explicating the functional features of these consensually acceptable explanatory approaches before establishing hierarchy.
Hierarchy. Once one agrees that E is an explanation one needs to decide how it fares as an explanation. One that is "incorrect" or does not account for the facts will probably be rejected in favor of one which does account for the data.
Some explanations serve as critiques to popular theories of human action. To the extent that the scientist discovers facts which are contrary to prevailing thought that scientist is acting as a social critic. For example, the demonstration that scores on intelligence tests are highly modifiable by environmental circumstances serves as a corrective a commonly held belief that such scores reflect a constant in hereditary endowment. Although such information based upon empirical investigation changes our view of the malleability of intelligence to a limited extent, it does not provide a general understanding of intelligence.
Beyond that, criteria for a hierarchy become difficult to determine. What makes, for example, "scientific" explanation "better" than explanations given by a non-psychologist?
Often scientific explanations are judged to be superior to ordinary explanations of human action. When a psychologist explains a child's whining by speaking of the history of parents' reinforcing the child for whining, what makes such an explanation superior to that of a non-psychologist who says that the child is whining because she wants the candy? From a perspective given by ordinary language philosophy, one would say that "explanation" has been achieved when most language users in the language using collective would agree that was is given is an "explanation" of a particular human action. If one accepts such a standard for the admissibility of an explanation, then one needs to decide how to determine what explanations are "better" than others. Consensual criteria seem important to invoke at the outset. If the language using community agrees that explanation E' is "better" than explanation E, then E' should be accepted over E. In the context of the example above, it seems that the most appropriate explanation must be judged by those who are using the word "explanation." If the group who is listening to the explanation are non-psychologists, no doubt the child's wanting the candy is explanatory. If the group who is listening to the explanation are psychologists, the reinforcement explanation will no doubt be more satisfactory. If when the child whines and is given the candy, the child throws the candy on the floor, then the non-psychologist who accepted the "wanting the candy" explanation may turn to the psychologist for further explanation (e.g., other operative reinforcers). So on an everyday basis what constitutes explanation is probably consensual. As a general rule then, one might posit that ordinary explanations are usually expected for ordinary behavior, and the psychologist's more esoteric explanations for ordinary behavior are acceptable only when the behavior no longer appears to be ordinary.
In developing a hierarchy of explanation, three "levels" seem obvious. At the lowest level is the explanation which is agreed by the language using community to an what it considers an "explanation," but for one reason or another is so far removed from the data-base that it is considered to be unsatisfactory in its explanatory power by most people is the language using community.
At the next level, social criticism done by psychologists serve as correctives for erroneous beliefs which are not supported by factual content. Such criticism is important and powerful yet is limited in terms of its explanatory specificity of content, narrowness of scope, and purposive, goal-directed character.
Also at this level is the sort of explanation done by clinical psychologists who endeavor to "solve a mystery" concerning the behavioral displays of a client. This sort of explanation is also of limited value in giving a general understanding of human action. This limitation was probably one of the oversights of Freud in extending the explanations given to the clients that he saw to people outside of the psychoanalytic context. Probably it was this oversight which has led to a good deal of skepticism by the general public of psychoanalytic explanations of everyday behavior. Like social criticism, the clinical psychologists gives explanations which are limited by specificity of content, narrowness of scope, and purposive, goal-directed character.
The adequacy of general accounts are what are most often debated and are often used to counter ordinary explanations. If what is offered as explanatory is accepted as "explanatory" by the language using community to which it is offered, its adequacy must be contingent upon serving the needs of enlightenment to that constituency.
Since explanatory elitism or scientific imperialism can only serve to further social/political purposes and results in alienating people who might otherwise by served by the methodological skills and conceptual insights possessed by psychologists, an attitude of egalitarianism seems to be called for.
Allport, G. W. (1961). Pattern and Growth in Personality. New York: Holt.
Bartlett, F. C. (1932). Remembering: A study in Experimental and Social Psychology. Cambridge: Cambridge University Press.
Foucault, M. (1965). Madness and Civilization: A History of Insanity in the Age of Reason. New York: Pantheon.
Foucault, M. (1970). The Order of Things: An Archeology of the Human Sciences. New York: Pantheon.
Hartnack, J. (1962). Wittgenstein and Modern Philosophy. Garden City, N. Y.: Anchor.
Hempel, C. G. (1966). Philosophy of a Natural Science. Englewood Cliffs: Prentice-Hall.
Leifer, R. (1964). The psychiatrist and tests of criminal responsibility. American Psychologist, 19, 825-830.
Marx, M. H. (1963). The general nature of theory construction. In M. H. Marx, Theories in Contemporary Psychology. New York: Macmillan.
Russell, B. (1956). Our Knowledge of the External World. New York: Mentor.
Schank, R. C. & Abelson, R. (1976). Scripts, Plans, Goals, and Understanding. Hillsdale, N. J.: Erlbaum.
Skinner, B. F. (1953). Science and Human Behavior. New York: Macmillan.
Stevens, S. S. (1963). Operationism and logical positivism. In M. H. Marx, Theories in Contemporary Psychology. New York: Macmillan.
Szasz, T. S. (1960). The myth of mental illness. American Psychologist, 15, 113-118.
Wittgenstein, L. (1953). Philosophical Investigations. New York: Macmillan.