The ambiguity is found in the way such behavioral terms as interpersonal reflexes, social interaction, and response are applied to what seem to be two different kinds of interpersonal events. This usage leads to confusion because it is not made clear whether the writers mean that the events are different and the terms are used broadly, or that the events are really alike and the terms are used strictly. On the one hand, the phrase interpersonal relation is often used with the clear assumption that what is designated is an interaction between organisms describable in the terms of a behavioristic social psychology.* On the other hand, the same term is extended to activities which are even recognized by the writers as being in some sense different from the directly observable behavior of organisms. The ambiguity appears in the description of the behavior of both psychiatrist and patient. Thus those studying the patient find it natural to speak of the objective study of his behavior and also of an “interpretive content analysis” of what he says.† And the behavior of the psychiatrist is described as “participant observation.” The psychiatrist not only enters into a conversation as other people do; he also preserves a posture of objectivity from which he takes note of the patient’s behavior, and his own, according to the principles of his science. One is free, of course, to designate all these activities by some such term as behavior or interaction. But if it is meant that these activities are really alike, it is not clear in what ways they are alike. Or if it is allowed that they are different, it is not clear wherein they differ or under what larger canon they may be brought into some kind of conceptual order.
The anomalous position of empirical scientists vis-à-vis intersubjective phenomena has been noticed before. Even Mead declared that an ideally refined behaviorism could explain the behavior of the observed subject but not that of the observing behaviorist. The social psychologist, it seems fair to say, sets out to understand social behavior as a species of interaction between organisms.‡ Yet by his own behavior he seems to allow for a kind of interpersonal activity which can be called “interaction” only by the most Pickwickian use of language. For the social psychologist observes, theorizes, and writes papers which he expects his colleagues not merely to respond to but to understand as well.* His behaviorism does not give an account of his own behavior. The awkward fact is that verstehen, that indispensable technique by which the social scientist discovers what another person “means,” is not provided for by neobehavioristic psychology. The anomaly is implicit in social psychology but explicit and acute in psychiatry because of the peculiar nature of the therapist-patient encounter. It is not possible to ignore the role of the scientist when he comprises one half of the social dyad under study. The social psychologist studies the interactions of persons and groups. But the psychiatrist is very largely concerned with the “interaction” between the patient and himself. And so the psychiatrist has come to be called the “participant observer.”
But the term participant observation expresses rather than clarifies a dilemma of the social sciences, and it should be accepted heuristically rather than as an explanation of what the psychiatrist is doing. The persistent ambiguity, however, is not occupationally peculiar to psychiatrists, and is not to be resolved by psychiatric theory. It comes about not as a result of some peculiar exigency of the therapist-patient relation but rather as a result of a fundamental incoherence in the attitude of empirical scientists toward that generic phenomenon of which the therapist-patient encounter is but a special instance: human communication. And it is to communication theory, considered both as the empirical science of symbolic behavior (psycholinguistics) and as a unified theory of signs (semiotic), that one must look for the source of the confusion and its resolution.
The Incoherence of a Behavioristic Theory of Meaning
About thirty-five years ago Edward Sapir called attention to a serious oversight in the then current psychology of language, writing, “…psychologists have perhaps too narrowly concerned themselves with the simple psychophysical bases of speech and have not penetrated very deeply into its symbolic nature.” He called for an empirical study of speech as a mode of symbolic behavior. Ten years later another great linguist, Benjamin Lee Whorf, took issue with his colleagues’ practice of “recording hairsplitting distinctions of sound, performing phonetic gymnastics, and writing complex grammars which only grammarians read.” “Linguistics,” he reminded them, “is essentially the quest of meaning.”
The warnings of Sapir and Whorf have not been heeded. On the contrary. The trend of theoretical linguistics in recent years has been in precisely the opposite direction. Linguists are quite frank about their aversion to meaning, to symbolic behavior, as a fit subject for empirical investigation. As Carroll has summed it up, the trend has been away from a psychology of verbal behavior—that is, the empirical investigation of the language event as a natural phenomenon; the trend instead has been toward “communication theory,” which abstracts from the event itself and concerns itself with a statistical analysis of the capacity of various systems of communication, and “discourse analysis,” which is a formal determination of the recurrence of morphemes in connected speech. The upshot has been an incoherent attitude toward symbolic behavior. Language is held to be a kind of sign response and so understandable in behavioristic terms as an interaction between an organism and its environment—which consists, in this case, of other organisms.* At the same time, the peculiar status of symbolic behavior is recognized by treating it formally—there are no formal sciences, as far as I know, devoted to the syntax or semantics of animal utterances. Thus, there is a natural science devoted to the study of reaction times and learning behavior; there are formal sciences which treat the logic and grammar of sentences. But where is the natural science which treats sentence events—not a sentence written on a blackboard, but the happening in which a father, replying to his son’s question, utters the following sounds: “That is a balloon”?
A good example of this incoherence is to be found in the otherwise valuable discipline of semiotic, which seeks to unite the several disciplines of symbolic logic, psychological behaviorism, and semantics into a single organon.* Semiotic is divided into three levels or dimensions: syntactics, pragmatics, and semantics. Syntactics is, as one might expect, a formal science having to do with the logico-grammatical structure of signs and with the formation and transformation rules of language. Pragmatics is the natural science of organisms responding to signs in their environments—psychiatry would be considered a branch of pragmatics. Semantics, which has to do with the relation of signs and their designata, is not a natural science of symbolic behavior, as one might have hoped. It is a formal deductive discipline in which “semantic rules” are proposed, designating the conditions under which a sign is applied to its object or designatum† Thus, in semiotic, symbolic behavior is studied formally in syntactics and semantics, but is disqualified in the natural empirical science of pragmatics—or written off as a refinement of sign-response behavior.
The embarrassing fact is that there does not exist today, as far as I am aware, a natural empirical science of symbolic behavior as such.‡ Yet communication, the language event, is a real happening; it is as proper a subject for a natural science as nuclear fission or sexual reproduction.
Neobehavioristic social psychology is not able to take account of symbolic behavior, let alone provide a heuristically fruitful basis of investigation. To say so is in no wise to challenge the accomplishments of the behavioristic approach. Learning theory is still valid as far as it goes. Reaction times still stand. It is still quite true to say that when a conversation takes place between two people, a stimulus or energy exchange makes its well-known journey as a wave disturbance in the air, through the solids of the middle ear, as an afferent nerve impulse, as an electrocolloidal change in the central nervous system, as an efferent nerve impulse, as a muscle movement in the larynx of the second speaker, as a wave disturbance, and so on. One is still justified in calling the interpersonal process what Mead called it
fifty years ago: a conversation of gesture in which my speech stimulus “calls out a response” from you. It is not enough to say this, however. For, as Susanne Langer rather drily observed, to set forth language as a sequence of stimuli and responses overlooks the salient trait of symbolic behavior: Symbols, words, not only call forth responses; they also denote things, name things for both speakers. Furthermore, behavioristic psychology is not able to take account of another universal trait of connected speech: Words are not merely aggregates of sound, however significant; in sentences or in agglutinative forms they also assert a state of affairs (or deny it or question it or command it). No alternative remains to the behaviorist semanticist but to disqualify the phenomenon of symbolization—to call it “an unreal but imputed relation between word and thing” or simply “wrong.” Again one is free to call symbolic behavior wrong or unreal or anything one likes, but such epithets hardly settle its status for the empirical scientist. It remains the task of empirical science to investigate phenomena as they happen, and everyone would agree that symbolic behavior does happen: People talk together, name things, make assertions about states of affairs, and to a degree understand each other.
The real task is how to study symbolic behavior, not formally by the deductive sciences which specify rules for the use of symbols in logic and calculi, but empirically as a kind of event which takes place in the same public domain as learning behavior. Sapir’s gentle chiding about the lack of a science of symbolic behavior and the need of such a science is more conspicuously true today than it was thirty-five years ago.
I am well aware, of course, that the altogether praiseworthy objective of the behaviorist is to get beyond the old mentalist nightmare in which interpersonal process is set forth in terms of my having “ideas,” “thoughts,” and “feelings,” and giving them names and so conveying them to you. If the word meaning refers to such mental entities, researchers do well to have nothing to do with it, for nothing has so effectively stifled the empirical investigation of communication as this misbegotten offspring of Descartes, the word-thing, the sound which I speak and which somehow carries my idea over to you like a note in a bottle. Yet the question must arise as to whether the alternatives lie only between a behavioristic theory of meaning, the energy exchange bouncing back and forth between speaker and hearer like a tennis ball, and the old miraculous mind reading by means of words. The phenomenon oíverste-hen, my understanding of what another person “means,” has been often called “subjective” by positive scientists and hence beyond the competence of empirical science. But such a ruling places the social scientist in the uncomfortable position of disqualifying his own activity—in the psychiatrist’s case, the activity of understanding his patient, writing papers, teaching courses.
Some Molar Traits of the Communication Event
The fact is that the generic traits of symbolic behavior are not “mental” at all. They are empirically ascertainable and have indeed been observed often enough. Both Ruesch and Jaffe have noticed that interpersonal events are peculiarly dyadic in a sense not altogether applicable to the interaction of the organism with its environment. Ruesch speaks of the structure of the interpersonal relation as a two-person system; Jaffe calls it a dyad. I would lay even greater stress on this feature as a manifestation of a generic trait of symbolic behavior. One may say if one likes that the bee dance is a communication event occurring in a two-bee system, but one is multiplying entities and it is not particularly useful to say so anyhow. A bee responding to another bee can be considered quite adequately as an organism in transaction with an environment, quite as much so as a solitary polar bear responding to the sound of splitting ice. But it has proved anything but adequate to consider language in the same terms. A symbol is generically intersubjective. I can never discover that the object is called a chair unless you tell me so, and my inkling that it “is” a chair is qualitatively different from the bee’s response to the bee dance of going to look for nectar.
Schachtel set forth another trait of symbolic behavior when he observed the genesis of an attitude among children which he called “autonomous object interest,” an attitude which he was careful to distinguish from need-satisfactions and wish fulfillment. It is not difficult, I think, to demonstrate that this autonomous object interest is intimately associated with the genesis of object language in the second year of life and is in fact an enduring trait of all symbolic behavior.
Two observations by Martin Buber are also of the utmost relevance to the basic structure of symbolic behavior. One of the main theses of Buber’s thought is his concept of relation, or the interhuman, which he holds to be beyond the reach of a behavioristic psychology. The other is the concept of distance. In contrast to the organism which exists wholly within its environment, man sets things at a distance. He is the creature through whose being (Sein), a phenomenon, “what is” (das Seiende), becomes detached from him and recognized for itself. Buber’s observations are developed within the framework of a philosophical anthropology; the traits of distance and relation are expressed as modes peculiar to human existence rather than as directly observable features of human relations. Expressed thus, Buber’s insights are perhaps somewhat uncongenial to many American social scientists with their strict empirical methodology—although it would be quite possible to defend the thesis that Buber’s analysis of human existence and human relations is also empirical in the broad sense of the word. It may be true that these existential traits of distance and relation are not “mental,” but they must strike the empirical scientist as vague in meaning and difficult to define operationally. Man is after all an organism, whatever else he is, and he does live in an environment. If he exists in uniquely human modes of being, such as distance and relation, it is not clear how these modes are grounded in or otherwise related to the present empirical knowledge of man. Precisely what does it mean to say that the human organism enters into the interhuman relation and sets things at a distance? Such theoretical grounding is, I believe, forthcoming from an empirical analysis of symbolic behavior. Indeed, it seems clear that Ruesch and Jaffe’s more-than-one-person system, Schachtel’s autonomous object interest, and Buber’s distance and relation are neither random nor reducible characteristics of human behavior. They are rather among the prime and generic traits of the highly structured meaning-situation found in symbolic behavior. What is more important, these traits are ascertainable not by a philosophical anthropology—which source is itself enough to render them suspect in the eyes of the behavioral scientists—but by an empirical analysis of language events as they are found to occur in the genetic appearance of language in the encultured child, in blind deaf-mutes, and in the structure of everyday language exchanges.
The greatest danger of the narrow behavioristic framework within which American behavioral scientists almost instinctively conceive the interpersonal process is that peculiarly human phenomena, such as language, are held either reducible to response sequences which leave out symbols altogether, or else describable by analogy, which does not so much shed light on the subject as close the door. Thus it may be unexceptionable to compare genes and symbols as the permanent characters of their respective systems and to speak of “levels of organization,” but such semantic shifts shed little or no light on intersubjective processes.* In an article about Buber, Leslie Farber wrote not long ago, “Having used only the single mode of scientific knowledge for the past hundred years or so, we are uneasily aware that this was the wrong mode—the wrong viewpoint, the wrong terminology, and the wrong kind of knowledge—ever to explain the human being.” This is true enough, I believe. There is a danger, however, in setting philosophical anthropology over against empirical science in such a sharp dichotomy. It is apt to confirm the positive scientist in his determination to have nothing to do with the existentialist-phenomenological movement—and so further impoverishes his social behaviorism. At the same time it encourages from the opposite quarter all manner of irrational and antiscientific prejudice—in particul
ar the ill-assorted crew of post-Cartesian mentalists who want to rescue “man” from “science” and restore him to the angelic order of mind and subjectivity. No, the present crisis of the social sciences need not polarize itself into an ideological issue between American positivists and European existentialists. Surely the better course is an allegiance to the empirical method—but not, let me carefully note, an allegiance to a theoretical commitment. The watchword of the empirical social scientist who confronts interpersonal phenomena should be, Let us see what is going on, and not, Let us see how we can fit it into a stimulus-response transaction.
The Structure of Symbolic Behavior
It would not, perhaps, be inaccurate to say that American psychology, as well as other behavioral sciences, has settled on an eclectic behaviorism in which the cruder features of Watsonian psychology have been refined by the work of Tolman, Skinner, Hull, Mowrer, Dollard and Miller, Sears, and Angyall. In this view, also put forward at the pragmatic level of semiotic, the organism, whether human or subhuman, is regarded as an open system living in an environment and adapting to that environment through its response to elements which are called signs. A sign is defined as an element in the environment which, through congenital or acquired patterns of behavior, directs the organism to something else, this something else being understood either as some other element or simply as biologically relevant behavior. Thus, the scent of deer directs the tiger to the deer; the scent of the tiger directs the deer to flight. A good representation of this relation is the semiotic triangle, shown in Figure 5.*