SOURCE: "Mead and Skinner: Agency and Determinism," in Behaviorism: A Forum for Critical Discussion, Vol. 16, No. 2, Fall, 1988, pp. 109-62.

With some behaviorists heeding the "call to cognition" (Deitz & Arrington, 1984; Morris, 1985), behaviorists are raising increasing numbers of questions about the role of thought, deliberate action, agency and determinism in behavioral theories. Most methodological behaviorists and radical behaviorists equate agency with free will, which they reject (Zuriff, 1985), or with inner causes that can be explained with behavior principles (Zuriff, 1975, 1979, 1985). In contrast, social learning theorists and cognitive behaviorists emphasize the importance of cognitive processes in the construction of deliberate action, and criticize the more traditional behaviorists for assuming that self-determined behavior can be completely predicted from genetic and environmental data (Bandura, 1986; Pierce & Epling, 1984;).

Is this polarization within behaviorism necessary? George Herbert Mead's behavioral theory demonstrates how one can avoid these conflicts about agency. Although it is often recognized that behaviorism has its roots in pragmatism, little attention has been given to Mead (1863-1931), who was one of the original pragmatiste and social behaviorists. As a Chicago school philosopher, Mead developed a form of pragmatism and social behaviorism that clarifies philosophical questions related to determinism, agency, and other issues of importance to behaviorists; yet Mead is seldom cited in the contemporary behavioral literature. For example, Zuriff's (1985) excellent summary of the philosophical foundations of behaviorism (with over eleven hundred references) makes no mention of Mead. Boakes's (1984) historical analysis of the emergence of behaviorism from 1870 to 1930 contains only passing references to Mead (pp. 145f, 162). Although Staats (1975), Powell and Still (1979) and a few others discuss parts of Mead's work, most behaviorists show little awareness of the full breadth of Mead's contributions.

Mead's ideas have been lost due, in part, to his being misinterpreted by his most outspoken followers in sociology and symbolic interactionism. Herbert Blumer (1955, 1962, 1966) and other interactionists selectively drew upon one subset of Mead's work—on mind and the self—to create a mentalistic model of the human actor. Citing Mead's lectures, Blumer and other students of Mead created a secondary literature on Mead's theory; and for decades, knowledge of Mead's work was transmitted by the secondary literature or by an "oral tradition" (Kuhn, 1964), with little serious attention to Mead's written work. As Mead's name became identified with a mentalistic theory of the self and interpretive processes, it is quite understandable that many behaviorists did not turn to his work for ideas.

Recent studies of Mead's original writings—that carefully document his views with numerous quotations and references—reveal that he has been seriously misinterpreted (Baldwin, 1986; McPhail & Rexroat, 1979, 1980; Wood & Wardell, 1983). The goals of this paper are (1) to demonstrate the relevance and utility of Mead's pragmatic and behavioristic philosophy for contemporary behaviorists, (2) to do so by showing how Mead dealt with the problems of agency and determinism, and (3) to show the similarities and differences between Mead and Skinner. The contrast of Mead's and Skinner's positions reveals only one small, nonscientific assumption that impedes a rapproachement between cognitive and more traditional behaviorism.

GEORGE HERBERT MEAD

Mead and the other pragmatists argued that the scientific method was superior to all other methods of acquiring knowledge and should be extended to deal with every area of inquiry. All ideas and theories—even philosophical and theoretical ones—should be treated as hypotheses that can be tested by their ability to solve problems and provide useful information. Any idea can be evaluated in terms of the type of consequences that result from it (Mead, 1900; 1908; 1913; 1923; 1925-26; 1938, pp. 454ff). In philosophical debates, the scientific method was to replace the methods of contemplation and idealistic thinking that were common before the development of pragmatism (Mead, 1938, pp. 513ff).

One of Mead's (1922, 1924-25, 1927, 1929-30) central goals was to use the scientific method to create a unifying theory that would synthesize empirical findings on all aspects of human behavior—including its inner and outer aspects—in a nondualistic manner. "We must consider inside and outside together, and the world cannot be divided into inside and outside" (Mead, 1927, p. 107). "Mind and body are not to be separated on the basis of our present physical science" (p. 167). Whereas Watson attempted to avoid dualism by excluding mental events from his theory, Mead (1934, pp. 10f, 101f) faulted this strategy and argued that the mind can be studied behaviorally. Although Mead and Watson were friends, Mead (1934, p. 10) criticized Watson's approach to mind as "misguided and unsuccessful, for the existence as such of mind or consciousness, in some sense or other, must be admitted—the denial of it leads inevitably to obvious absurdities." Using an argument similar to Skinner's (1969, pp. 227f; 1971, pp. 1900, Mead stated that "it is not possible to deny the existence of mind or consciousness or mental phenomena, nor is it desirable to do so; but it is possible to account for them or deal with them in behavioristic terms which are precisely similar to those which Watson employs in dealing with non-mental psychological phenomena" (1934, p. 10).

Mead developed a thoroughly nondualistic approach to mind and cognitive processes, tracing the emergence of cognitive processes first during the evolutionary process and second, in humans, during socialization. Biological evolution helps explain the emergence of the brain structures that establish the potential for language; and language provides the mechanism for the symbolic processes that make human consciousness different from the consciousness of other species. "The mechanism of thinking [is] that of inward conversation" (1932, p. 84). "An objective psychology is not trying to get rid of consciousness, but trying to state the intelligence of the individual in terms which will enable us to see how that intelligence is exercised, and how it may be improved" (1934, p. 39). Describing mind in terms of neural mechanisms and "inward" verbal behavior clearly places mind in nature, available for empirical study. "But if mind is simply an emergent character of certain organisms in their so-called intelligent responses to their environments, mind can never transcend the environment within which it operates" (1932, p. 118). "Mind can no longer be put outside of nature" (p. 152).

Mind and self are not inborn, but arise in social interaction. "What I want particularly to emphasize is the temporal and logical pre-existence of the social process to the self-conscious individual that arises in it" (1934, p. 186). Mead advocated studying self-conscious behavior by tracing it's development, working from the outside social environment to the emergence of inside events. This strategy "works from the outside to the inside instead of from the inside to the outside, so to speak, in its endeavor to determine how such [inner] experience does arise within the [social] process" (p. 8). This approach allowed Mead (1927, p. 106) to "avoid introspection . . . We do not approach the organism from within. . . . The actual process begins at the periphery and goes to the center" (p. 156). "Just as there is a functional relationship between the organism and its environment, so there is one between what is 'in the mind' and what is 'outside'" (1936, p. 307). In numerous ways, Mead traced the structure of the mind and the self back to the structure of the environment. Nevertheless, this approach did not lead him to develop a deterministic model of the human actor, as can be understood by examining Mead's views on metaphysics and decision making.

BEYOND METAPHYSICS

Mead's approach to agency can be useful to contemporary behaviorists. There are several ways of conceptualizing agency. One is to equate agency with free will. However, free will is a metaphysical construct, and as such can never be evaluated empirically. Mead (1917, 1929, 1932) argued that we can never resolve metaphysical questions about the ultimate nature of "reality"; thus we can never know whether people "really" have free will or not.

Some ancient Greek philosophers and mathematically oriented thinkers assumed the presence of some preexistent ultimate truths beyond the range of our fallible perceptual organs (Mead, 1917); and for centuries various influential philosophers sought to reach those transcendental truths through contemplation and logic (1938, pp. 513f). However, the pragmatists noted that despite thousands of years of contemplation, philosophers had made no more net advance than the mythical figure Sisyphus had in repeatedly rolling his stone up the same hill. In contrast, science—by abandoning metaphysical puzzles and focusing on careful observation and experimentation—has made enormous advances in understanding the world in the past few centuries. The pragmatists advocated abandoning contemplation and metaphysics, to rely solely on the scientific method as our means of inquiry. When we do this, the "despairing sense of the philosophic Sisyphus vainly striving to roll the heavily weighted world of his reflection up into a preexistent reality" drops "away and the philosopher can face about toward the future and join in the scientist's adventure" (1964, pp. 389f). A belief in transcendental truths creates a problematic dualism between ideal knowledge and earthly knowledge; but a strong adherence to the scientific method and empirical data helps avoid the problem. "The solution of the problem carries with it the disappearance of the problem and the metaphysical system at the same time" (1964, p. 10).

Much of Mead's work was devoted to reconstructing philosophy and psychology using the methods of modern science, with no reference to contemplative philosophy or metaphysics. Relevant to philosophy, Mead (1964, p. 210) stated: "I have endeavored to present the world which is an implication of the scientific method of discovery with entire abstraction from any epistemological or metaphysical presuppositions or complications." Relevant to behavioristic psychology, Mead stated: "Psychology . . . has not been interested in these epistemological and metaphysical riddles, it has been simply irritated by them" (p. 269). "The behavioristic psychology has tried to get rid of the more or less metaphysical complications involved in the setting-up of the psychical over against the world, mind over against body, consciousness over against matter. That was felt to lead into a blind alley" (1934, p. 105). Among the metaphysical issues to be dropped is the debate over free will and determinism.

Nevertheless, some branches of science have deterministic qualities, and Mead dealt with these in his discussions of mechanism. Newtonian physics produced mechanistic, deterministic models of the world that left no room for purpose, freedom or choice (1936, pp. 250-281). The great success of the mechanistic sciences in developing deterministic laws reinforced the notion that everything might fit such models. "Seemingly, the whole world would be absolutely fixed and determined" (p. 250). However, Mead pointed out that mechanistic, deterministic assumptions are merely "postulates" or "working hypotheses," and science can never prove them to be absolute truths.

In the mechanistic disciplines, scientists often espouse methodological determinism as a perspective that focuses their efforts on uncovering regularities in nature (Kaplan, 1964, pp. 124f), and taking this perspective is often reinforced by the discovery of empirical generalizations, basic principles or laws. However, "Methodological determinism states only that laws are worth looking for here, not that they surely exist here, and surely not that they necessarily exist always and everywhere" (Kaplan, 1964, p. 124). Mead approved of methodologies that focus our efforts on constructing propositions and laws to explain the regularities in behavior. He stated that behavioral psychology has "its sympathies . . . with the presuppositions and method of the natural sciences" (1964, p. 269). "We are interested in finding the most general laws of correlation we can find" (1934, p. 39). This process builds toward a scientific psychology in which "All the distinctions must be explained by the same general laws as those which are appealed to to account for animal organs and functions" (1964, p. 82).

However, Mead recognized that no theory that we can craft—even in the physical sciences—is likely to ever become completely finalized, fully accurate, or totally deterministic: novel and unpredictable events continually emerge that force us to revise and reconstruct our theories (1917, 1929, 1932). As fallible organisms, humans are not likely to ever develop such perfect models that all future events could be predicted; hence all sciences repeatedly confront novel and emergent events that are not completely predictable. When scientists use methodological determinism, "the emergent has no sooner appeared than we set about rationalizing it ..." (1932, p. 14); however, these rationalizations and theories usually have to be modified and reconstructed as later unexpected events emerge. Thus, while science's ability to locate regularities and patterns has allowed the development of increasingly sophisticated empirical generalizations and laws, scientific laws are unlikely to ever become completely finalized and totally deterministic. In fact, "the scientist's procedure and method contemplate no such finality. On the contrary, they contemplate continued reconstruction in the face of events emerging in ceaseless novelty" (pp. 10lf). After constructing any theory or doctrine, "the scientist himself expects this doctrine to be reconstructed just as other scientific doctrines have been reconstructed" in the past (p. 105).

Instead of deterministic laws, Mead anticipated only probabilistic theories of our world of emergent events. "[W]hatever does happen, even the emergent ..." happens under conditions that "lie within probability only" and "these conditions never determine completely the 'what it is' that will happen" (p. 15). "[T]here is also the indeterminateness of what occurs" (p. 96). "And the indeterminate 'what' involves always a possibly new situation with a new complex of relationships" (p. 97). In an indeterminant universe, scientists construct probabilistic laws to explicate the regularities identified to date, without the illusion of attaining the perfect knowledge needed to construct totally deterministic laws.

Finally, Mead emphasized that mechanistic laws do not adequately deal "with the characters which belong to living organisms" (1936, p. 260). Namely, mechanistic theories fail to treat living things as adapting organisms, with qualities captured by the vocabulary of design and purpose (p. 269). For example, a purely mechanistic analysis of the digestive process omits the issue of the function—or purpose—of digestion in maintaining life. "The complete mechanical statement would not take account of the end, of the purpose ..." that is seen in living systems; "[a]nd that seems to be necessary to our comprehension of the world" (p. 272).

Mead used the term "teleology" to describe the purposeful, life-sustaining processes that were overlooked by purely mechanical analyses of living systems (1927, pp. 108ff; 1936, pp. 268ff). However, he clearly dissociated himself from the nonscientific form of teleology espoused by vitalists (1936, pp. 292-325). "Sometimes, like the vitalists, we abuse [mechanistic] science because it ignores life. But there is only a short distance we can go on the ideological program" (1927, p. 171). That short distance is to make the point that life is not fully described by mechanistic terms alone: Science needs to have concepts for explaining the self-organizing, purposive qualities of living systems, not just concepts about the mechanical aspects of the universe.

Both mechanism and teleology are postulates, not metaphysical dogma about the ultimate nature of "reality" (1936, pp. 264-291). They provide two different languages—mechanistic and action languages—that help us describe different aspects of living systems. Both of these languages can be useful in behavioral sciences; and they are not mutually exclusive: They can complement each other. "In biological science you bring in both these points of view" (p. 269). Consider the example of a physician and district attorney analyzing a murder (pp. 2680). The physician explains the mechanism of the murder, how the "bullet entered the body in a certain way and led to a given result" (p. 269). In contrast, the district attorney explains the motives and purposes of the murderer: "That is, he gives a teleological explanation, while the doctor who performs the autopsy gives a mechanical explanation" (p. 269). Since neither teleology nor mechanism is a dogma in science, it is easy to see that "[t]here is, then, no real conflict between a mechanical and a teleological account of the world or of the facts of life" (p. 271). "Thus we see that science has gotten away from metaphysical dogma as to what the nature of things is ..." (p. 275). Mechanistic, deterministic hypotheses are useful in describing some aspects of nature; and purposive, teleological models are useful in describing other aspects. Neither type of hypothesis warrants conclusions about ultimate metaphysical questions, including those concerning free will and determinism.

If we agree with Mead that we can never resolve metaphysical questions about the ultimate nature of "reality," we can put aside the metaphysical issue of whether or not humans have free will and ask if there is any other way to conceptualize and evaluate agency. Mead followed the behavioral strategy of analyzing the behaviors that are the referents of the abstract term agency. When concepts such as agency—which in the past have been treated as abstract, metaphysical concepts—are reformulated in purely behavioral terms, they are clearly removed from the arena of metaphysics and contemplative philosophy.

DECISION MAKING

In order to clarify Mead's pragmatic approach to agency, the following sections briefly review Mead's analysis of those behaviors—such as choice and decision making—that are generally considered to be central to agency. Mead analyzed choice, planning, and decision making in terms of their neural and behavioral components—which are open to empirical study—and not in terms of "free will," as metaphysically conceived. Mead's analyses of human conduct never involved hidden assumptions about or attempts to prove any metaphysical position about free will or determinism.

Mead (1924-25, 1932, 1934) took an evolutionary approach to the emergence of consciousness and deliberate action. By comparing the behavior of animals at different phyletic levels, Mead outlined the evolutionary preconditions for the emergence of the higher forms of human cognition and planned behavior. His evolutionary approach also underscores the fact that human consciousness is part of nature, not transcendental. "The genesis of mind in human society . . . is a natural development within the world of living organisms and their environment" (1932, p. 84). "I have wished to present mind as an evolution in nature . . ." (p. 85). As such, Mead attempted to identify the biological and psychosocial precursors of human mental capacities, with special attention to neurophysiological and social factors.

NEUROPHYSIOLOGY

Mead (1924-25, 1932, 1934) related the development of deliberate choice and self-control to the evolution of the encephalon during the emergence of the advanced vertebrate species. He discussed several cortical mechanisms that allow delay and inhibition of response, imagery of possible future responses, and the capacity for choice between various possible alternative responses (1932, pp. 68-76, 89ff, 124-136; 1934, pp. 98-100, 109-118).

First, the encephalon "is primarily the nerve center of the important distance senses" (pp. 700, hence its development makes possible the detailed processing of distant stimuli before contact is made. With the evolution of the cortex and increased sophistication of the distance senses, "the contact experiences to which [the distance senses] respond are delayed, and possibilities of adjustment and of choice in response are thus increased" (p. 71). Sensing a stimulus at a distance, before contact is made, allows for a time period before a contact response is required; and this time allows for the selective processes involved in deliberate choice.

"The cerebrum ... is an organ which integrates a vast variety of responses including the lower reflexes . . ." (p. 126). "In a sense all responses are so interconnected by way of interrelated innervation and inhibition" (p. 125). Integration and inhibition provide a second mechanism for creating delay in response. "This introduces delay in response, and adjustment by way of selection of type of response, i.e., choice" (p. 126).

Third, during the delay period, the organism feels its own incipient responses to the distant stimuli. "In the innervations of the attitudes that distant objects call out the animal feels the invitation or threat they carry with them" (p. 71). Invitation or threat call up response tendencies to act. "His responses to his own tendencies to act provide the control that organizes all his responses into a cöordinated act, so that these inner feelings wax in importance in the development of the mechanism" of the encephalon (p. 71). "It is here that we first meet the stuff of ideation" (p. 71).

Fourth, as the cortex evolved, simple internal feelings of incipient responses became more sophisticated forms of "response imagery" (p. 74). Images of past actions and their results help in anticipating the future results—or consequences—of similar actions. "This imagery gives us the result of the act before we carry it out" (1914, p. 29). Because the cerebrum allows the organism to anticipate the results of the act, the cerebrum has "introduced the future into the mechanism of the act . . ." (1932, p. 132). Imagery of the future consequences of various alternative acts allows the organism to select among them and make "purposive responses" (p. 74). "It is the ability of later responses to play back into immediate responses that gives us our flexibility and power of choice" (1927, p. 158). Note that choice and purposive responses are defined behaviorally—as behaviors controlled by anticipated future consequences, which are established from past experience.

Although neural mechanisms provide part of Mead's explanation of choice, language and verbal behavior have allowed the emergence of advanced levels of deliberate action in humans.

SYMBOLIC THOUGHT AND CONSCIOUSNESS

The full emergence of mind and deliberative processes can occur only through social interaction, especially through the social use of language (Mead, 1924-25, 1934). As children interact socially and acquire language, they gain the tools for symbolic thought and increasing self-determination. Language provides "word images" (1932, p. 75) that enrich the more primitive response imagery used in making choices. "Ideas are closely related to images . . . Since the symbols with which we think are largely recognized as word images, ideas and images have a very close consanguinity" (p. 75). Words allow us to call up symbolic images of all phases of our acts, along with the consequences of those acts. They also bring images of the past and anticipated future into our present, and this "[i]deation extends spatially and temporally the field within which activity takes place," leading to an "extended present" that reaches far beyond our immediate perceptual present (p. 88).

The meanings that words carry are a central component of the extended present and the deliberate actions we take in it. By analyzing meaning from a behavioristic perspective, Mead was able to avoid the metaphysical problems associated with traditional views of "ideas" and "meanings." In discussing "the meaning of things," Mead (1934, p. 127) stated: "We are here avoiding logical and metaphysical problems, just as modern psychology does." First, objects and actions take on meanings based on the results or completions of the acts in which they are involved. "The completions that need to occur before the act is completed are behavioristic meanings" (1927, p. 143). For example, "the ultimate act of driving a nail is for us the meaning of the hammer" (p. 130). The hammer's capacity for driving the nail establishes its meaning. "When we indicate this pattern of final manipulation we indicate the meaning of the act" (p. 143). Second, words such as "hammer" are symbols that come to stand for the meanings that are based on people's actions, such as using hammers to drive nails.

Not only physical objects, but also social gestures—such as a wink or a smile—take on meanings, based on the results of the acts involving them. "If that gesture does so indicate to another organism the subsequent (or resultant) behavior of the given organism, then it has meaning. In other words, the relationship between a given stimulus—as a gesture—and the later phases of the social act of which it is an early . . . phase constitutes the field within which meaning originates and exists" (1934, p. 76). The meaning of a smile is established by the results of the acts following the smile, which of course can vary across different contexts. In those contexts where smiles have been followed by cheerful and kind behavior in the past, a person will tend to perceive smiles as indicating cheerfulness and kindness; but the meaning of smiles can be quite different when context cues are present that in the past indicated that smiles might be followed by insincere or manipulative behavior. Thus, the meaning of the social gesture is based on the phases of the social act that follow it. "Meaning is thus a development of something objectively there as a relation between certain phases of the social act; it is not a psychical addition to that act and it is not an 'idea' as traditionally conceived" (e.g., by metaphysical or idealist philosophers). Rather. "[t]he gesture stands for a certain resultant of the social act, ... so that meaning is given or stated in terms of response" (p. 76). "The basis of meaning is thus objectively there in social conduct, or in nature in its relation to such conduct" (p. 80). As such it is amenable to scientific study.

"Awareness or consciousness is not necessary to the presence of meaning in the process of social experience" (p. 77). Two snarling dogs are exchanging and responding to meaningful signals in their exchange of gestures, but are not consciously aware of the meanings. "The mechanism of meaning is thus present in the social act before the emergence of consciousness or awareness of meaning occurs" (p. 77). Meaning becomes aware only when the "gesture becomes a symbol, a significant symbol" (p. 78), namely a symbol that people use in a conventional manner, so that it calls up the same meaning for those users. Meaning "is not essentially or primarily a psychical content (a content of mind or consciousness), for it need not be conscious at all, and is not in fact until significant symbols are evolved in the process of human social experience. Only when it becomes identified with such symbols does meaning become conscious" (p. 80). Both a human and an animal can see and be perceptually aware of a hammer on the ground; but humans can gain the additional awareness of the meaning of the hammer by verbally describing it and its use. "Mentality on our approach simply comes in when the organism is able to point out meanings to others and to himself. This is the point at which mind appears, or if you like, emerges" (p. 132). This approach to mind introduces none of the mysterious elements present in idealist and metaphysical conceptions of meaning and mind.

How does Mead's behavioristic approach deal with highly abstract and generalized meanings? In metaphysical and idealistic world views, highly abstract meanings seem to refer to universals at a Platonic or transcendental level. An idealist might think that "this meaning or universal character [is something] with which a behavioristic psychology is supposed to have difficulty in dealing" (1934, p. 82). However Mead disagreed: "It is the possibility of such a behavioristic statement that I endeavor to sketch" (p. 83).

Mead explained "universals" in terms of the behavior seen when any of a variety of different stimuli can call up the same response. When we have to write a brief note, we recognize that many types of pens, pencils or markers will suffice: All members of a certain class of stimuli can call up the response of picking up the object and writing. "[R]ecognition can be stated in terms of a response that may answer to any one of a certain group of stimuli" (p. 83). If a person is attempting to drive a nail and cannot find a hammer, the individual may recognize that "a brick or a stone" will also serve the same function (p. 83). "Anything that he can get hold of that will serve the purpose will be a hammer. That sort of response which involves the grasping of a heavy object is a universal" (p. 83). The brick or stone can function as if it were a hammer; hence all three fall into the same class of objects because they—as hard, heavy objects—are functionally similar in carrying out the behavior of hammering the nail. The recognition of functional equivalence allows the universal "response that answers to a whole set of particulars" (p. 84). "It is this which has been supposed to be beyond the behavioristic explanation or statement. What behavioristic psychology does is to state that character of the experience in terms of response" (p. 84).

Our membership in a verbal community—a culture with symbolic knowledge—helps us acquire a broad range of knowledge about things experienced first by others. We do not have to be swept away by a raging river to learn about the strength and danger of a powerful torrent. "[I]n the community of those who communicate with each other, the force of the torrent has taken on a meaning insofar as each is wont to indicate this to others and so to himself (1964, p. 336). The consequences of stepping into a powerful torrent establish its meaning as "something to be avoided"; and experienced people can communicate those meanings to individuals who have never had first-hand experience with the torrent and its consequences. This cooperative social process allows individuals to expand their understanding of universals—e.g., adding "the torrent" to the class of stimuli that carry the universal meaning of "something to be avoided"—based on the experience of others.

REFLECTIVE INTELLIGENCE

Mead's (1934, pp. 90-100) analysis of reflective intelligence reveals his scientific approach to yet another important aspect of agency. When we have several alternative responses open to us in a given situation, we can use significant symbols to heighten our awareness of and reflect on our choices of possible actions before acting. "[R]eflective behavior arises only under the conditions of self-consciousness and makes possible the purposive control ... of its conduct . . ." (p. 91). Words make us aware of the details of each alternative action and the likely future consequences of each one. People are especially likely to use this type of intelligence in problematic situations (pp. 122ff). "This [reflective intelligence] is the most effective means of adjustment to the social environment, and indeed to the environment in general, that the individual has at his disposal" (p. 100).

Mead identified several components of reflective intelligence, all of which are amenable to scientific investigation. "Intelligence is essentially the ability to solve the problems of present behavior in terms of its possible future consequences as implicated on the basis of past experience . . ." (p. 100). When we confront a problem, we can use symbols to imagine several possible solutions to the problem and to evaluate the anticipated consequences of each alternative, based on past experience, before choosing a course of action. People do this when they—either out loud or in their inner conversation—talk themselves through several alternative solutions to a problem and use their memories of past experiences to anticipate the possible future consequences of each alternative.

Mead (p. 91) stated that "[i]t is essential that such reflective intelligence be dealt with from the point of view of social behaviorism." By conceptualizing the components of reflective intelligence in behavioral terms, Mead avoided the metaphysical notions of free will that arise from nonempirical approaches to choice and agency. "What the behaviorist is occupied with, what we have to come back to, is the actual reaction itself, and it is only in so far as we can translate the content of introspection over into response that we can get any satisfactory psychological doctrine" (p. 105).

Although reflective intelligence is mediated by the mechanisms of the central nervous system and inner conversation, it is not a totally mechanical activity that produces completely predictable outcomes. Rather it is an organic, creative process.

That which takes place in present organic behavior is always in some sense an emergent from the past, and never could have been precisely predicted in advance . . . and in the case of organic behavior which is intelligently controlled, this element of spontaneity is especially prominent by virtue of the present influence exercised over such behavior by the possible future results or consequences which it may have. (Mead, 1934, pp. 98f)

The spontaneous, creative qualities of the organic processes seen in reflective intelligence are important features of agency. Problem solving processes—such as those seen in reflective intelligence—are an especially important stimulant for creativity (Weisberg, 1986): As we reflect on the possible solutions of problems, we become aware of novel and unexpected alternatives, reevaluate and reconstruct our memories of past experiences, and reassess the possible future outcomes of the alternatives. From the complicated interaction of the perceptual present with images of the past and future, complex decisions and actions emerge. Like all emergents, these events are not completely predictable from data on preceding events (Mead, 1932, pp. 14ff, 35ff, 96f; 1964, pp. 346f).

Although Mead traced the contents of mind and the methods of reflective intelligence back to the central nervous system and society, he did not believe that the empirical data on physiology and social structure were adequate to make perfect predictions about the decisions that people would make. This is most obvious when we have to predict the exact outcome of a complex choice made by a person who has an extensive repertoire of decision making skills, ponders a large number of possible alternatives, and takes creative approaches to the problem solving process. A precise prediction would require an enormous amount of data on all aspects of the person's genetics, physiology, prior learning experience, and current stimuli. Since it is unlikely that humans can ever attain enough knowledge about all the controlling variables, it is more feasible to expect probabilistic forecasts than deterministic predictions. Mead's rejection of fully deterministic scientific theories is not to be confused with a position on the metaphysical question of free will and determinism. Mead merely stated that complex processes are not likely to be predicted with total precision, hence the behavior that emerges from reflective intelligence and other complex mediational processes is likely to have a spontaneous, creative quality that is difficult to predict completely.

Since Mead's time, psychologists have studied the various and complex ways in which people lay out alternative solutions to their problems, draw upon memory resources, apply rational, emotional and intuitive procedures to evaluate each alternative, allocate different amounts of time and effort to problems of differing levels of complexity and import, and so forth (Anderson, 1982; Chase & Simon, 1973; Dreyfus & Dreyfus, 1986). These and related studies have facilitated the development of guidelines and training programs for helping people to improve their skills at reflective intelligence (Baron & Steinberg, 1986; Meichenbaum, 1977). Rather than devaluing reflective intelligence and creativity by "explaining them away" with purely mechanistic and deterministic models, the scientific study of these activities can help us become more reflective and creative.

SKINNER

How does Skinner's position on cognition, choice and agency compare with Mead's? Skinner appears to have vacillated in his arguments about the "causal efficacy of events that occur inside the skin" (Killeen, 1984, p. 25). He has also taken positions on determinism that can easily appear to be inconsistent and confusing. After examining the ways in which Skinner's analysis of choice and indeterminism is compatible with Mead's work, we will turn to the central differences between these two behaviorists.

First, Skinner's position on the nature of consciousness and inner experience is similar to Mead's. Although methodological behaviorists have refused to analyze events inside the skin, Skinner and other radical behaviorists have argued that this position is unwise and unnecessary. Speaking of methodological behaviorists, Skinner (1969, pp. 227f) stated that "the charge is justified that they have neglected the facts of consciousness. The strategy is, however, quite unwise. It is particularly important that a science of behavior face the problem of privacy. It may do so without abandoning the basic position of behaviorism," by treating inner events "as part of behavior itself." Skinner (1953, 1974) has done a skillful job of behaviorally conceptualizing and analyzing many types of private experiences—including perception, thinking, memory, emotions, problem solving, the self, and selfcontrol—explaining how they are shaped by the external contingencies.

Second, Skinner not only treats inner events as stimuli and responses that can be traced to environmental causes, he has recognized at least ten types of inner causes (Zuriff, 1979). Although Skinner has long been critical of hypothetical inner causes that are not amenable to scientific analyses, he has dealt with inner causes that can be stated in behavioral terms and analyzed in traditional behavioral manners. "In the ten examples . . . covert variables control overt behavior in all the ways that external environmental variables do. . . . Thus, if the external environment may be said to 'cause' behavior, then in the examples presented, inner events may also be said to 'cause' behavior" (Zuriff, 1979, p. 1). These inner causes are different from nonscientific inner causes because "they are acquired in overt form, obey the same laws and have the same dimensions as overt responses, and are ultimately controlled by environmental variables" (p. 8). However, Zuriff points out that, "as the present discussion clearly demonstrates, these inner activities often play an important role in the causation of overt behavior" (p. 8). The same could be said of Mead's treatment of inner causes.

Thus, Skinner's (1969, 1971) numerous attacks on mentalistic psychology, hypothetical mental way stations, and autonomous inner man must not be confused as a criticism of all inner causes. Mead (1910, 1913, 1927, 1934) also was critical of the theories based on introspective psychologies, and was careful to describe his inner causes of choice and reflective intelligence in terms of stimuli and responses amenable to behavioral analysis.

These first two points clearly indicate that Skinner is not adverse to analyzing inner experiences and causes, working from the outside to the inside, much as Mead did. The next four points demonstrate that Skinner's behavioral analysis of external and internal responses does not produce a completely deterministic theory, but rather is compatible with Mead's model of an indeterminate science of behavior.

Third, Skinner's (1938, 1953) analysis of operant behavior clearly reveals that operants are more unpredictable than reflexes. In both unconditioned and conditioned reflexes, an antecedent stimulus elicits a respondent in a rather automatic manner than can suggest mechanistic and deterministic models. Skinner (1938, pp. 19f, 178; 1953, p. 107) was careful to point out that operants are performed much less automatically and predictably than are respondents. In describing the "spontaneity" of operants, Skinner (1938) stated: "I do not mean that there are not originating forces in spontaneous behavior but simply that they are not located in the environment. We are not in a position to see them, and we have no need to. This kind of behavior might be said to be emitted by the organism . . ." (p. 20). "The prior stimulus does not elicit the response; it merely sets the occasion upon which the response will be reinforced" (1938, p. 178). Skinner contrasted the "static laws" of respondent behavior with the "dynamic laws" needed to describe operant behavior, which is "an event appearing spontaneously with a given frequency" (p. 21). Since it is difficult to predict exactly when an operant will occur in the presence of its controlling S^D's, operants do not fit deterministic models as well as they fit the probabilistic models that Mead advocated. "Both prediction and control are inherent in operant conditioning, but the notion is always probabilistic . . ." (Skinner, 1974, p. 226).

According to Skinner (1969, p. 2270: "The skin is not that important as a boundary. Private and public events have the same kind of physical dimensions." Thus, covert operants involved in thought and choice would be expected to fit probabilistic models as much as overt operants do.

Fourth, Skinner (1953, 1957, 1968, 1969) has analyzed various types of verbal problem-solving behavior, one of which—the self-probe—is quite similar to Mead's description of reflective intelligence. In the self-probe, we probe ourselves for data and arguments needed to solve a problem, much as others probe us for information when helping us decide what to do. "Tentative solutions, perhaps assembled for this purpose, are systematically reviewed" (1953, p. 250). "We facilitate choosing or making a decision in various ways—for example, by 'reviewing the facts.' If we are working with external materials, verbal or otherwise, we may indeed re-view them in the sense of looking at them again.... In reviewing an argument we simply argue again" (1974, p. 112). A culture can help people learn to verbalize about their behaviors and related consequences. "As a culture evolves, it encourages running comment of this sort and thus prepares its members to solve problems most effectively" (1969, p. 143). "A crude description may contribute to a more exact one, and a final characterization which supports a quite unambiguous response brings problem solving to an end" (p. 142). Cultures also help people learn rules for problem solving. "Many rules which help in solving the problem of solving problems are familiar. 'Ask yourself "What is the unknown?'"" (p. 145). In the Technology of Teaching, Skinner (1968) suggested ways to teach thinking, problem solving, and creativity.

Skinner provided more detail than Mead did on the ways in which people learn reflective intelligence and other methods of problem solving. However, Skinner's type of behavioral analysis does not allow us to predict exactly when and how an individual will solve most kinds of problems. In discussing problem solving behavior, Skinner (1968, pp. 1380 stated: "The behavior is not unlawful, but we lack the information needed to predict the moment of its occurrence with certainty." In fact, it is not only the timing of the response, but also the topography of the behavior and its effects on the environment that can be difficult to predict with complete certainty. We could never gain enough knowledge about people's genetics, prior conditioning history, and current stimulus inputs to predict in perfect detail the solutions to novel problems, such as a scientist's solution to a complex problem at the frontier of research. Novel problems confront people with numerous S^D's—many of which do not have strong stimulus control over specific operants—mixed in unfamiliar patterns, making it unlikely that a rote or easily predicted response will emerge. When problem solving involves the interaction of numerous complex behavioral skills, the interaction effects increase the probability of unexpected results. Since problem solving often involves lengthy chains of verbal and nonverbal operants (each link of which provides the S^D's for the next link), the probabilistic nature of each operant in the chain makes it difficult to predict which operant and which S^D's will emerge to set the occasion for the next operant. Thus, as a chain is performed, the uncertainties of predicting each link of the chain are amplified at each step; and such amplification mechanisms sometimes lead to highly unpredictable outcomes (cf, Crutchfield et al., 1986, p. 49). Clearly, probabilistic models are more appropriate than deterministic models.

Fifth, problems present S^D's that can set the occasion for creative thinking. In his early work, Skinner (1953, p. 255) traced novel and original behavior to novel environmental conditions: "Novel contigencies generate novel forms of behavior. . . . The question of originality can be disposed of ... by providing plausible accounts of the way in which a given idea might have occurred." Later, Skinner (1968, pp. 169-184) recognized that original, unpredictable behavior could be explained in other ways. For example, the reinforcement of novel behavior or punishment of commonplace behavior can produce original and creative behavior (pp. 181-184). Experimental studies have documented that creativity can, in fact, be enhanced by this type of reinforcement (Goetz & Baer, 1973; Maloney & Hopkins, 1973). Since the response class being reinforced is very broadly defined—as behavior that is novel and/or not commonplace—it is very hard to predict precisely what behavior will be emitted next (aside from the fact that it may be novel). In fact, Skinner (1974, p. 114) compared creative thought with mutations in biology, which he described as "random." Creative operants "are, if not random, at least not necessarily related to the contingencies under which they will be selected" (p. 114). Namely, they are not determined by the controlling variables. This is the stuff of indeterminacy.

Behaviorists have good reason to believe that behavioral analysis is quite well suited for explaining how people acquire creative skills; but this knowledge is not sufficient to allow us to predict the precise timing and topography of the behavior generated from a creative repertoire. Creative actions are among the events "emerging in ceaseless novelty" that prevent sciences from ever reaching finality (Mead, 1932, p. 102). Mead's form of behaviorism and Killeen's (1984) "emergent behaviorism" provide a more appropriate treatment of emergent behavior than does a behaviorism that purports to be totally deterministic.

Sixth, concepts such as purpose and intentional design are also related to the topic of agency. Broadly defined, all operant behavior is future oriented, hence purposive. Skinner (1974, p. 55) states: "Possibly no charge is more often leveled against behaviorism or a science of behavior than that it cannot deal with purpose or intention. A stimulus-response formula has no answer, but operant behavior is the very field of purpose and intention. By its nature it is directed toward the future: a person acts in order that something will happen, and the order is temporal." Although all operants can be seen as purposive, this definition of purpose is not likely to coincide with the lay person's uses of the word. Skinner's treatment of intentional behavior and design comes closer to lay definitions of purposive action.

All through Beyond Freedom and Dignity, Skinner (1971) was critical of nonscientific conceptions of autonomous inner man; however, near the end, he concluded: "What is needed is more 'intentional' control, not less, and this is an important engineering problem" (p. 177). The intentional design of environmental controlling variables—hence of behavior—is possible; and Skinner (1953, 1968, 1971) frequently advocated intentional design at both the individual and cultural levels. He argued that our increasingly detailed knowledge of behavior principles can be used in designing and engineering the environmental conditions needed to produce intentionally designed behaviors. When intentionality is defined and analyzed behaviorally, it is not off limits to behavioral theory. Even the intentional behavior commonly seen in everyday life—whether effectively designed and implemented or not—can be analyzed behaviorally. "[T]he behavior from which we infer choice, intention, and originality is within reach of behavioral analysis . . . " (1974, p. 239).

THE DIFFERENCES

There are numerous other facets of Skinner's writings that parallel Mead's theories (Baldwin, 1981). In these areas of commonality, the most conspicuous difference between Mead's and Skinner's work arises from Skinner's having access to a much larger and more modern body of empirical data and theories than Mead had. For example, both Mead and Skinner analyzed reflective intelligence and problem solving in terms of verbally reviewing each possible solution and its consequences, but Skinner has a much more precise theory of the role of social and natural contingencies in the acquisition and modification of verbal skills and reflective processes.

Probably the most important basic difference between Mead's and Skinner's versions of behaviorism lies in their different treatment of determinism. Much as Mead had, Skinner has recognized the probabilistic nature of behavior and the impossibility of metaphysical certainty about issues of free will and determinism. Speaking of methodological determinism, Skinner (1968, p. 171) stated: "Determinism is a useful assumption because it encourages a search for causes." In a similar vein, Mead (1932, 1936) had described methodological determinism as a "postulate" that has been useful in the pursuit of mechanistic laws. Nevertheless, Skinner emphasized determinism more than Mead did. Whereas Mead counterbalanced his discussion of methodological determinism with discussions of the emergence of unpredictable events, Skinner emphasized methodological determinism much more than probability or emergence.

In addition, Skinner developed a technique never found in Mead's work. He extrapolated from methodological determinism to develop interpretive analyses in which he examined complex behaviors as if methodological determinism might someday produce a totally deterministic science of them. Skinner (1973, p. 261) explained his strategy: "My Verbal Behavior was an exercise in interpretation . . . Beyond Freedom and Dignity is also an exercise in interpretation. It is not science as such, but it is not metaphysics, either." Skinner is aware of the limits of deterministic interpretations. "[A]lthough a science of behavior permits a person to interpret what he sees more effectively, it will never tell him the whole story about the individual case" (1974, p. 242).

In spite of an awareness of the limitations of deterministic interpretations, Skinner has at times taken them to the extreme, making statements that sometimes sound like pronouncements about complete determinism. For example, from the observation that there is "new evidence of the predictability of human behavior," Skinner concluded: "Personal exemption from a complete determinism is revoked as a scientific analysis progresses . . ." (1971, p. 21, emphasis added). "All behavior is determined, directly or indirectly, by consequences . . ." (1974, p. 127, emphasis added). When such statements are not understood as interpretive analyses, they sound like statements about absolute metaphysical truths—or naive assertions about the power of a science of behavior.

Skinner's defense of deterministic interpretations reveals a penchant for deterministic thinking: "We cannot prove, of course, that human behavior as a whole is fully determined, but the proposition becomes more plausible as facts accumulate . . ." (p. 189). Advances in the science of behavior might suggest the plausibility of determinism to some, but a serious recognition of the probabilistic and unpredictable nature of many types of behavior has the opposite effect, undermining the assumption "that human behavior as a whole is fully determined." Although Skinner frequently focused on determinism and overlooked sources of indeterminacy, Mead (1932, p. 14), in contrast, saw that his task was "to bring into congruence with each other this universality of determination which is the text of modern science, and the emergence of the novel...." Mead advocated looking for "the most general laws" possible (1934, p. 39), but he expected the laws to be probabilistic, not deterministic (1932, pp. 14f, 320).

Is it wise for behaviorists to take the extreme position on determinism that Skinner does in his interpretive analyses? One answer lies in the consequences of the behavior. On the positive side, exercises in deterministic interpretations may encourage behaviorists to apply methodological determinism to all types of behavior—including thinking, problem solving, self-control, teaching, and culture—in search for causes. Skinner's deterministic analyses provide behaviorists with models that show how such analyses might be done, and reading them can provide reinforcement for expanding the domain of behavioral study by demonstrating that behavior principles can deal successfully with issues once thought to be beyond the range of the discipline.

In addition, deterministic interpretations provide a vivid contrast between accounts of behavior based on free will and those based on methodological determinism. This contrast helps behaviorists learn to identify prescientific intellectual baggage and problems that have their roots in the metaphysics of free will. For example, in Beyond Freedom and Dignity, this exercise revealed important points about the literature of freedom, including its lack of attention to the dangers of weak control and positive control (Skinner, 1973). Such exercises may hasten the demise of prescientific interpretations of behavior. "When I question the supposed residual freedom of autonomous man, I am not debating the issue of free will. I am simply describing the slow demise of a prescientific explanatory device . . . The argument is, I believe, quite similar to that against vital forces in biology" (1973, p. 261). Is it true that the demise of prescientific causal explanations will leave only determinism as an interpretive tool? Skinner appears to have underestimated the importance of emergence, probability, and indeterminacy in developing scientific theories that introduce no metaphysical assumptions of free will. Mead did not.

Not all the consequences of Skinner's deterministic interpretations have been positive. His extreme positions have caused many serious scientists and scholars to reject behaviorism, much as Watson's extremes did. Skinner's attempts to explain—or explain away—all mental causes as fully determined by environmental contingencies have generated numerous criticisms of behaviorism (Black, 1973; Chomsky, 1959; Lefcourt, 1973; Ritchie-Calder, 1973; Toynbee, 1973). Even though the actual scientific contributions of behavioral research are proving to be quite valuable in child rearing, education, various branches of therapy, and so forth, many people dislike or fear the brave new deterministic world they identify with behaviorism. Even behaviorists, such as Mahoney (1974), Schwartz and Lacy (1982), and Bandura (1986) have reacted against Skinner's extreme positions on determinism and the environmental determination of inner cognitions, introducing serious schisms into the discipline.

Do the benefits of deterministic interpretations outweigh the costs of alienating serious scientists and scholars who otherwise might profit from knowledge of behavior principles? Perhaps only history can tell. Even though Watson has been severely criticized outside the field of behavioral psychology (Broadbent, 1961; Gould, 1982; Harrell & Harrison, 1938), he is still viewed with respect by behaviorists for his contributions to the growth of our discipline (Boakes, 1984; Bolles, 1979). Perhaps Skinner's effectiveness in inspiring behaviorists to pursue methodological determinism is more important than his alienating nonbehaviorists. Nevertheless, we may still wish to ask if behaviorism would be better served by indeterminant and probabilistic models such as those advocated by Mead.

INDETERMINISM

In his careful analysis and critical restructuring of the philosophical tenets of behaviorism, Zuriff (1985, p. 177) states: "Determinism . . . is a methodological working assumption for the behaviorist, rather than a metaphysical commitment." This working assumption derives in part from behaviorists' commitment to the natural sciences. "Because behaviorists generally subscribe to the world view of the natural sciences, they profess grave philosophical doubts about the possibility of a free agent operating in behavior" (p. 177). However, doubting free will does not necessarily lead the natural scientist to espouse determinism or take positions similar to Skinner's deterministic interpretations.

Although earlier, mechanistic physics portrayed a deterministic world, contemporary physicist have moved to a more probabilistic view. Research on radiation along with the development of theories of relativity and Heisenberg's uncertainty principle led physicists to withdraw from completely mechanistic and deterministic theories. Mead (1932, 1938) discussed the theory of relativity as part of his argument for recognizing the importance of emergence and probabilistic theories. By 1936, both physicists and philosophers of science were questioning the old "deterministic methods which have brought such great success to the physical sciences. . . . And now we are witnessing physical science itself not only raising doubt as to the adequacy of Newtonian mechanics and the laws of the conservation of energy, but also challenging the entire classical conception of physical causation" (Cohen, 1936, p. 327).

Skinner's response to these developments in physics was different from Mead's. Speaking of the principle of indeterminacy, Skinner (1953, p. 17) stated: "In our present state of knowledge, certain events therefore appear to be unpredictable. It does not follow that these events are free or capricious." It is true that the indeterminacy principle does not introduce a free agent, but it does not support Skinner's interpretations in terms of complete determinism. It reinforces arguments for probabilism and indeterminant models.

More recently, the theory of chaos has lent even stronger support to the position that numerous phenomena—including Brownian motion, fluid turbulence, weather, genetic variability, behavior, and society—have random, chaotic, and capricious elements (Crutchfield et al., 1986; Mayer-Kress, 1986). In systems that involve complex chains of events operating over time, minor variations in initial conditions can influence subsequent events such that small errors in original measurements are amplified as events proceed, sometimes leading to highly unpredictable outcomes, i.e., to chaos. For example, it would be impossible to make accurate predictions about the position of billiard balls on a table under frictionless conditions one minute after they have been hit by the cue ball. "The large growth in uncertainty comes about because the balls are curved, and small differences at the point of impact are amplified with each collision. The amplification is exponential: it is compounded at every collision. . . . Any effect, no matter how small, quickly reaches macroscopic proportions. . . . With chaos, predictions are rapidly doomed to gross inaccuracy" (Crutchfield et al., 1986, p. 49). As a result, "long-term predictions are intrinsically impossible" (p. 56). The same is true for complex human behavior and cognition. In a twenty link chain of cognitive operants, we can make much better predictions of the final outcome if we have knowledge about as many links of the chain as possible, rather than having data only about the original environmental conditions, as one would if using Skinner's deterministic interpretive approach to trace inner causes back to environmental contingencies rather than studying them in process. The less deterministic types of behaviorism advocated by Mead, Mahoney, Killeen, Bandura, and others recognize the utility of studying all links of cognitive chains as they function together for increasing the accuracy of prediction, control, and therapy.

Finally, there are questions about the wisdom of Skinner's (1953, 1971) applying his deterministic interpretations to society and social evolution. Random processes appear to be important features of social interaction and cultural evolution (Brenner, 1983; Duncan, 1986; Kaufman, 1985). Since social systems are influenced by so many controlling variables that are difficult to measure and model precisely, they often change in random and chaotic manners. Deterministic interpretations of social processes may lead us to overestimate our capacity for prediction, control, planned change, and the minimization of human error. A healthy recognition of the indeterminacy of social systems tempers ones confidence about long-term or extremely accurate predictions and sensitizes one to the need to monitor the countless unpredictable events that affect the next occurrences in social systems. Methodological determinism remains the best approach we have for studying social systems; but it can be expected to produce only probabilistic propositions, not completely deterministic laws.

CONCLUSIONS

G. H. Mead was a philosopher who developed sophisticated theories of pragmatism and behavioral psychology that avoid the deterministic and "empty organism" models that so many scholars find objectionable in some versions of behaviorism. Unfortunately, many contemporary behaviorists are not aware of Mead's contributions, which could be useful in remedying some of the problems in modern behaviorism and providing a strong philosophical position that could unite radical and cognitive behaviorists. This paper has attempted to show that Skinner the scientist has taken a position quite similar to Mead's. Skinner's awareness of the importance of inner causes—along with the spontaneous and probabilistic nature of operants, problem solving and creativity—makes it clear that Skinner, the scientist, did not anticipate a completely deterministic or empty-headed science of behavior. It is primarily his deterministic interpretations, which he explained were neither science nor metaphysics (1973, p. 261), that deviate from Mead's position. Since these deterministic interpretations are not scientific, they can be separated from the science of behavior, in which case, Skinner's position becomes even closer to Mead's and is compatible with cognitive and emergent behaviorism.

Mead attempted to create a unified world view that integrated data on biology, psychology and society in order to explain overt behavior and inner processes. His theory can be useful in helping contemporary behaviorists overcome some of the problems and schisms within our discipline. This paper has focused on agency because of its central importance to a science of behavior that has significant deterministic proclivities; but there is much more in Mead's work that warrants attention (Baldwin, 1986).

REFERENCES

Anderson, J. R. (1982). Acquisition of cognitive skill. Psychological Review, 89, 369-406.

Baldwin, J. D. (1981). George Herbert Mead and modern behaviorism. Pacific Sociological Review, 24, 411-440.

Baldwin, J. D. (1986). George Herbert Mead: A unifying theory for sociology. Beverly Hills, CA: Sage.

Bandura. A. (1986). Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ: Prentice-Hall.

Baron, J. B., & Steinberg. R. J. (1986). Teaching thinking skills: Theory and practice. New York: W.H. Freeman.

Black, M. (1973). Some aversive responses to a wouldbe reinforcer. In H. Wheeler (Ed.), Beyond the punitive society (pp. 125-134). San Francisco. CA: W.H. Freeman.

Blumer, H. (1955). Attitudes and the social act. Social Problems, 3, 59-65.

Blumer, H. (1962). Society as symbolic interaction. In A. Rose (Ed.), Human behavior and social processes (pp. 179-192). Boston, MA: Houghton Mifflin.

Blumer, H. (1966). Sociological implications of the thought of George Herbert Mead. American Journal of Sociology, 71, 535-544.

Boakes, R. (1984). From Darwin to behaviorism: Psychology and the minds of animals. New York: Cambridge University Press.

Bolles, R. C. (1979). Learning theory (2nd ed.). New York: Holt, Rinehart and Winston.

Brenner, R. (1983). History—the human gamble. Chicago, IL: University of Chicago Press.

Broadbent, D. E. (1961). Behaviour. New York: Basic Books.

Chase, W. G., & Simon, H. A. (1973). Perception in chess. Cognitive Psychology, 4, 55-81.

Chomsky, N. (1959). A review of B.F. Skinner's Verbal behavior. Language, 35, 26-58.

Cohen, M. R. (1936). The statistical view of nature. Journal of the American Statistical Association, 31, 327-347.

Crutchfield, J. P., Farmer, J. D., & Packard N. H. (1986). Chaos. Scientific American, 255(6), 46-57.

Deitz, S. M. & Arrington, R. L. (1984). Wittgenstein's language-games and the call to cognition. Behaviorism, 12, 1-14.

Dreyfus, H. L., & Dreyfus, S. E. (1986). Mind over machine: The power of human intuition and experience in the era of the computer. New York: Macmillan/Free Press.

Duncan, O. D. (1986). Probability, disposition, and the inconsistency of attitudes and behavior. Synthese, 68, 65-98.

Goetz, E. M., & Baer, D. M. (1973). Social control of form diversity and the emergence of new forms in children's block building. Journal of Applied Behavior Analysis, 6, 209-217.

Gould, J. L. (1982). Ethology: The mechanisms and evolution of behavior. New York: W.W. Norton.

Harrell, W., & Harrison, R. (1938). The rise and fall of behaviorism. The Journal of General Psychology, 18, 367-421.

Kaplan, A. (1964). The conduct of inquiry: Methodology for behavioral science. New York: Harper & Row.

Kaufman, H. (1985). Time, chance and organizations. Chatham, NJ: Chatham House.

Killeen, P. R. (1984). Emergent behaviorism. Behaviorism, 12, 25-39.

Kuhn, M. H. (1964). Major trends in symbolic interaction theory in the past twenty-five years. Sociological Quarterly, 5, 61-84.

Lefcourt, H. M. (1973). The function of the illusions of control and freedom. American Psychologist, 28, 417-425.

Mahoney, M. J. (1974). Cognition and behavior modification. Cambridge, MA: Ballinger.

Maloney, K. B. & Hopkins, B. L. (1973). The modification of sentence structure and its relationship to subjective judgements of creative writing. Journal of Applied Behavior Analysis, 6, 425-433.

Mayer-Kress, G. (Ed.). (1986). Dimensions and entropies in chaotic systems. New York: Springer-Verlag.

McPhail, C, & Rexroat, C. (1979). Mead vs. Blumer: The divergent methodological perspectives of social behaviorism and symbolic interactionism. American Sociological Review, 44, 449-467.

McPhail, C, & Rexroat, C. (1980). Ex cathedra Blumer or ex libris Mead. American Sociological Review, 45, 420-430.

Mead, G. H. (1900). "Suggestions toward a theory of philosophical disciplines." The Philosophical Review, 9, 1-17.

Mead, G. H. (1908). "The philosophical basis of ethics." International Journal of Ethics, 18, 311-323.

Mead, G. H. (1913). "The social self." Journal of Philosophy, Psychology, and Scientific Methods, 10, 374-380.

Mead, G. H. (1914). "1914 class lectures in social psychology." In D. L. Miller (Ed., 1982). The individual and the social self (pp. 27-105). Chicago, IL: University of Chicago Press.

Mead, G. H. (1917). "Scientific method and individual thinker." In Creative intelligence: Essays in the pragmatic attitude (pp. 176-227). New York: Henry Holt.

Mead, G. H. (1922). "A behavioristic account of the significant symbol." Journal of Philosophy, 19, 157-163.

Mead, G. H. (1923). "Scientific method and the moral sciences." International Journal of Ethics, 33, 229-247.

Mead, G. H. (1924-1925). "The genesis of the self and social control." International Journal of Ethics, 35, 251-277.

Mead, G. H. (1925-1926). "The nature of aesthetic experience." International Journal of Ethics, 36, 382-393.

Mead, G. H. (1927). "1927 class lectures in social psychology." In D. L. Miller (Ed.), The individual and the social self (pp. 106-175). Chicago, IL: University of Chicago Press.

Mead, G. H. (1929). "A pragmatic theory of truth." In Studies in the nature of truth, vol. 11 (pp. 65-88). University of California, Publications in Philosophy.

Mead, G. H. (1929-1930). "The philosophies of Royce, James, and Dewey in their American setting." International Journal of Ethics, 40, 211-231.

Mead, G. H. (1932). The philosophy of the present. A. E. Murphy (Ed.). Chicago, IL: Open Court.

Mead, G. H. (1934). Mind, self and society, C. W. Morris (Ed.). Chicago, IL: University of Chicago Press.

Mead, G. H. (1936). Movements of thought in the nineteenth century, M. H. Moore (Ed.). Chicago, IL: University of Chicago Press.

Mead, G. H. (1938). The philosophy of the act, C. W. Morris (Ed.). Chicago, IL: University of Chicago Press.

Mead, G. H. (1964). Selected writings, A. J. Reck (Ed.). New York: Bobbs-Merrill.

Meichenbaum, D. (1977). Cognitive behavior modification: An integrative approach. New York: Plenum.

Morris, E. K. (1985). "Wittgenstein's language-games and the call to cognition:" Comments on Deitz and Arrington. Behaviorism, 13, 137-146.

Pierce, W. D., & Epling, W. F. (1984). On the persistence of cognitive explanation: Implications for behavior analysis. Behaviorism, 12, 15-27.

Powell, R. P., & Still, A. W. (1979). Behaviorism and the psychology of language: An historical reassessment. Behaviorism, 1, 71-89.

Ritchie-Calder, L. (1973). Beyond B.F. Skinner. In H. Wheeler (Ed.), Beyond the punitive society (pp. 212-216). San Francisco, CA: W.H. Freeman.

Schwartz, B., & Lacey, H. (1982). Behaviorism, science, and human nature. New York: W.W. Norton.

Skinner, B. F. (1938). The behavior of organisms. New York: Appleton-Century-Crofts.

Skinner, B. F. (1953). Science and human behavior. New York: Macmillan.

Skinner, B. F. (1957). Verbal behavior. New York: Appleton-Century-Crofts.

Skinner, B. F. (1968). The technology of teaching. Englewood Cliffs, NJ: Prentice-Hall.

Skinner, B. F. (1969). Contingencies of reinforcement. New York: Appleton-Century-Crofts.

Skinner, B. F. (1971). Beyond freedom and dignity. New York: Knopf.

Skinner, B. F. (1973). Answers for my critics. In H. Wheeler (Ed.), Beyond the punitive society (pp. 256-266). San Francisco, CA: W.H. Freeman.

Skinner, B.F. (1974). About behaviorism. New York: Knopf.

Stoats, A.W. (1975). Social behaviorism. Homewood, IL: Dorsey.

Toynbee, A. (1973). Great expectations. In H. Wheeler (Ed.), Beyond the punitive society (pp. .113-120). San Francisco, CA: W.H. Freeman.

Weisberg, R. W. (1986). Creativity: Genius and other myths. New York: W.H. Freeman.

Wood, M., & Wardell, M. L. (1983). G.H. Mead's social behaviorism vs. the a structural bias of symbolic interactionism. Symbolic Interaction, 6, 85-96.

Zuriff, G. E. (1975). Where is the agent in behavior? Behaviorism, 3, 1-21.

Zuriff, G. E. (1979). Ten inner causes. Behaviorism, 7, 1-8.

Zuriff, G. E. (1985). Behaviorism: A conceptual reconstruction. New York: Columbia University Press.