JOURNAL OF THE EXPERIMENTAL ANALYSIS OF BEHAVIOR 1988, 501,97-111 NUMBER 1 (JULY) FINDING THE PHILOSOPHICAL CORE: A REVIEW OF STEPHEN C. PEPPER'S WORLD HYPOTHESES: A STUDY IN EVIDENCE' STEVEN C. HAYES, LINDA J. HAYES, AND HAYNE W. REESE UNIVERSITY OF NEVADA-RENO AND WEST VIRGINIA UNIVERSITY

Behavior analysis has always had significant the grand scheme of philosophy, abstracted conflicts with other psychological perspectives. from the details of particular positions. At their most fundamental level, these conflicts The book is at once simple and difficult. are often philosophical, concerning such issues In some areas, Pepper focused on the debates as the nature of the human and the purposes of his own time and particular history, and of science. Why are these the conflicts? What, thus emphasized distinctions that do not seem if anything, can we do about them? Can we to have withstood the test of nearly half a resolve them? Can we avoid them altogether century. In general, however, the book is by simply abandoning philosophy? amazingly contemporary. It is like a series To answer these questions, we must be of colored spotlights cast on a complicated clear about the philosophy underlying be- scene. Irrelevant details of various philo- havior analysis compared to those of other sophical positions disappear like so many perspectives. To be clear is difficult, however, shades of blue under a blue spotlight. Fun- because the assumptions and postulates of damental differences leap out, now from one the position are not deliberately and un- angle, now from another. Although the book ambiguously laid down. Fundamental as- is not about psychology, it exposes the philo- sumptions, specific theories, and historical sophical sources of current conflicts within accidents are too often discussed concurrently behavior analysis and the nature of its conflicts and without adequate differentiation in be- with other psychologies. havior-analytic expositions. Skinner's philo- To make Pepper's position understandable, sophical writings are especially prone to this we must summarize some parts of the book difficulty, perhaps because he is so extensively fairly extensively. To show its value for a involved with so many nonphilosophical as- field he never analyzed, we must interpret pects of the field. it and extend it. Our purpose is not to promote In 1942 Stephen C. Pepper, a philosopher the use of Pepper's conceptual categories per and aestheticist, published World Hypotheses: se; it is to use them to illuminate the conceptual A Study in Evidence. His central insight was categories of modern behavior analysis and that philosophical systems cluster around a of competing psychological systems. few core models, or "world hypotheses," drawn from common sense. His strategy was to ignore details and personalities, and instead to present THE NATURE OF WORLD the central tenets of each world view in a HYPOTHESES general way. He used his own terms to describe A world hypothesis is a model of the uni- most of these tenets, avoiding excess or pa- verse of observations and inferences. (We will rochial meanings. He quoted and cited very also use the term "world view" as a synonym little. His style permits an understanding of for "world hypotheses" to emphasize that Pepper's use of the term "hypothesis" in this Pepper, Stephen C. (1942). World hypotheses: A study context is not precisely what psychologists in evidence. Berkeley: University of California Press. vii mean by a "hypothesis" in a scientific context.) + 348 pp. World hypotheses differ in scope (i.e., the Linda J. Hayes is the former Linda J. Parrott. Reprint requests should be sent to Steven C. Hayes, range of events incorporated) and precision Department of Psychology, University of Nevada-Reno, (i.e., the scarcity of alternative interpretations Reno, Nevada 89557-0062. made of the events incorporated), and their 97 98 STEVEN C. HAYES et al. adequacy is a matter of the degree to which its underlying metaphor becomes empty and each obtains. Ideally, a world view has un- meaningless. limited scope and is so precise that it permits one and only one interpretation of every event. In practice, all reasonably adequate current THE RELATIVELY ADEQUATE world views fall short of this ideal. Scope WORLD VIEWS and precision tend to be inversely related: Pepper identified four relatively adequate A model that specifies precisely how to in- current world hypotheses: mechanism, form- terpret a given event will tend not to in- ism, organicism, and contextualism. Mech- corporate as many events as a model that anism and formism are analytic: The whole permits viewing a given event in a number is reducible to its parts. The parts are basic, of different ways. the whole derived. Organicism and contex- Pepper argued that each world hypothesis tualism are synthetic: The whole is basic, is autonomous. As an attempt to provide a the parts derived. Formism and contextualism complete view of the world, each world hy- are dispersive: Facts are related when they pothesis creates its own field of play. Within are found to be so, not by assumption. Chance, that field of play "competing" world views therefore, is not denied in these hypotheses. can be interpreted but cannot compete directly. Mechanism and organicism are integrative: This argument anticipates the similar views Facts are related by assumption and order of Kuhn (1962) among others (there are good is categorical. As such, chance is denied. Dis- reasons to believe that many of Kuhn's ideas persive world views tend to be higher in scope were derived from Pepper's work; Efron, than in precision; integrative world hypotheses 1980) and has several implications. First, tend to be higher in precision than in scope. using the categories of one world view to Describing psychological systems in Pep- analyze and criticize another is illegitimate per's terms organizes them into coordinated and inherently useless. Second, no world hy- sets of related propositions and concerns. The pothesis can be strengthened by revealing likely conflicts between particular psycho- shortcomings in another. The weakness of logical systems may then be derived, and one world view in no way implies strength solutions to these conflicts proposed. For this in another. Third, eclecticism, if it involves reason, examining each of Pepper's four rel- combinations of distinct world hypotheses, is atively adequate world views seems potentially inherently confusing. Each world hypothesis worthwhile to an analysis of psychological entails a different set of conceptual categories, systems. many of which are mutually contradictory across different world hypotheses. No coherent Mechanism combination of currently popular world views The root metaphor of mechanism is the seems likely and none has yet succeeded. This machine. Any common-sense machine is com- does not mean that forms of integration are posed of discrete parts related to other parts not possible, but if Pepper were correct, suc- in some systematic way. Relations among the cessful integration could come only by com- parts do not change the nature of the parts, binations based on yet another coherent world however, because the parts exist independently hypothesis. A single exception exists in stra- of those relations. Further, in any common- tegic integrations of world views subordinated sense machine, some sort of force or energy to a single world view. As discussed later, is exerted on or transmitted through the system such an integration is possible in only one to produce predictable outcomes. of the relatively adequate world views. A simple example is the lever. A lever is According to Pepper, world hypotheses are composed of two discrete and independent derived from "root metaphors." A root met- parts-a lever and a fulcrum. When they aphor is a common sense conceptualization are related in a particular way (e.g., the lever of a domain, in accordance with which cat- is placed on the fulcrum), a force exerted egorical concepts have been constructed. The on one end of the lever produces a predictable root metaphor structures an understanding effect at the other end according to the amount of the technical categories. A technical con- of force, length of the lever, placement of struct sustaining no conceptual contact with the fulcrum, and so on. In more complicated BOOK REVIEW 99 machines, the history of forces applied may or theory. Scientists cannot use the corre- also be relevant to the operation of the ma- spondence between a verbal construction of chine. In this context, history means that one the machine and the facts as a test of the part may not function until other parts have adequacy of the construction if the very same functioned, or that various parts have worn facts serve both as the source of the con- down. For example, a clock does not strike struction and as the means of its verification. until the spring controlling the hourhand has Truth is best established by examining the been wound and has operated for a pre- correspondence between the verbal construc- determined period of time. To take a more tion and a variety of new facts implied by contemporary example, a computer does not the construction. In line with the integrative operate until a code has been placed in its quality of mechanism, the more derived and "memory" chips and software in its read-only indirect these predictions are, the better (Er- "memory." icsson & Simon, 1984). Hypothetico-deductive According to the world hypothesis of mech- research methodology exemplifies this logic; anism, the entire universe is like a machine. many mechanists gravitate toward it. Different machines yield different variants of mechanism. The hydraulic statuary of the Formism Middle Ages leads to one theory, the computer The root metaphor of formism is similarity. model leads to another. Nonetheless, the key The kind of similarity implied here is the elements of common-sense machines and their recurrence of recognizable forms: blades of means of operation are always present. grass, sheets of paper, rows of doughnuts, The machine metaphor extends to the or the like. The position of immanent formism knower as well as the known. The knower (one of two major variants) is as follows. The relates to the world by producing an internal perception of any event involves contact with copy of it, through mechanical transformation. two aspects of that event-character and par- This epistemological stance preserves both the ticularity-that are absolutely distinct but knower and the known intact and basically cannot be experienced independently. Char- unchanged by their relation-a requirement acter refers to the qualities and relations that of mechanism. In some versions of mechanism, are tied to a given object: This doughnut has the internal copy is "mental"; in others it the qualities of being soft, warm, and sticky; is "physiological." Regardless, the knower it has the relation of side-by-sidedness with knows a copy of the world, not the world the other doughnuts in this row. Theoretically, itself. Truth is a matter of how well the copy a given object may have an infinite number corresponds to the world, as evaluated by of characters. corroboration among independent knowers. A given character may occur in an infinite Corroboration is required because the cor- number of particulars. Still, this doughnut respondence between the copy and the world is this doughnut and no other. This doughnut cannot be observed directly. is the particularization of a character, as well Correspondence between what we say about as the characterization of a particular. The the world and what we see in it is relatively doughnut involves the participation of one trivial when the material described is im- with the other. mediately present. Correspondence of this A collection, or "class," of particulars par- limited kind is not adequate for the mechanist ticipating in one or more characters is another for several reasons. The mechanist's goal is categorical concept in formism. "Gorilla," for to discover the parts and the relations among example, is neither a particular event, nor parts of the existent machine. Because mech- a character, nor a participation. It is all three anism is integrative, all the parts are assumed together. Certain characters that participate to fit together. Order is categorical. Thus, in a particular object lead us to identify that mechanists do not simply describe parts in object as "gorilla." Each other object that the common-sense world; rather, they seek is characterized in the same way is also a to discover the true nature of a given event "gorilla," and all these objects considered by specifying what kind of part it really is together are the class "gorilla." and by placing it properly in the machine. Systematic organizations of facts are not Such a goal is aided by an a priori model assumed by formists, hence principles of op- 100, STEVEN C. HAYES et al. eration (such as "force") are not required the world-it is neither known directly nor to explain them. Indeed, if all facts were mechanically transformed. integrated by a set of formistic principles, then those facts would necessarily form a Contextualism system (the integration itself) and formism Contextualism is the most important world would begin to have the character of mech- view for our purposes, so we here describe anism. A causal law in formism is no more Pepper's analysis of contextualism in more than a bridge from one set of characterized detail. The root metaphor of contextualism particulars to another. It is a form. is the ongoing act in context. Another term The truth criterion of formism, like that might be the historical act, but not as a dead of mechanism, is correspondence. In formism, description of a thing done. It is doing as however, the simpler sense of this criterion it is being done, as in hunting, shopping, (correspondence however derived) is adequate or making love. given its dispersive quality. Two fundamental categories of contex- tualism are quality and texture. Quality is Organicism the experienced nature of an act; texture is The root metaphor of organicism is the the details and relations that make up its process of organic development, as in living, quality. In contextualism, even these categories growing, organic systems. In such systems, might change (if they do) because nothing change is given and stability is to be explained. is final or ultimate about our knowledge of Versions of developmentalism that rely on the world-not even that the world will stay stage models reflect this philosophy. For ex- the same. In our present epoch, however, all ample, a person is assumed to move from events have quality and texture. one stage of growth to another in an orderly Each category is defined in terms of other way. To explain the person's current stage categories. Quality, for example, is made up we must explain the orderliness of changes of spread and fusion. Spread refers to the from stage to stage. In other words, we must extended present of an act in context. The explain how the rules of change operate, past and future of an act exist in the ongoing assuming that change occurs according to rules act. The act spreads, as we say, both backward of change that are themselves unchanging and forward. Fusion refers to the integration (Reese & Overton, 1970). The organicist notes of the textural details of a given event. Lem- "the steps involved in the organic process and onade has the texture of water, lemons, and . . . the principal features in the organic sugar. The quality of lemonade is a fusion structure ultimately achieved" (Pepper, 1942, of these distinct ingredients-so thoroughly p. 281). so that the ingredients are difficult to analyze In organicism, the whole is not a synthesis separately. Cooking a meal is composed of of parts; the whole is basic, the parts mean- many textural elements (e.g., picking up a ingless except in the context of the whole. pot, selecting ingredients, mixing the ingre- The organicist embraces teleology: "The struc- dients), all of which may be fused in the ture achieved or realized is always the ideal overall quality of cooking the meal. If one aimed at by the progressive steps of the pro- becomes more interested in the act of picking cess" (Pepper, 1942, p. 281). up the pot than in the act of cooking a meal, The truth criterion of organicism is co- the fusion of the larger act dissolves. Picking herence. When a network of interrelated facts up the pot becomes the experienced quality. converges on a conclusion, the coherence of Texture is defined in terms of other cat- this network renders this conclusion "true." egories, namely, strands, context, and reference. All contradictions of understanding originate Strands are the interconnections among the in incomplete knowledge of the whole organic details of an act that directly contribute to process. When the whole is known, the con- its quality. Context is made up of the in- tradictions are removed and the "organic whole terconnections among strands, contributing . . . is found to have been implicit in the indirectly to the quality of a given act. The fragments" (Pepper, 1942, p. 283). two cannot be fully distinguished because each Epistemologically, organicists adopt con- contributes to the nature of the other. As an structivism. The knower actively construes example, a cook prepares a dessert. The details BOOK REVIEW 101

and relations of this act, that is, its texture, mula. The fact that his formula is much more may be arranged in strands of various sorts. useful to many more people doesn't make This act could comprise the strand we might it any more real" (Pepper, 1942, p. 251). call "entertaining a dinner guest," occurring A powerful implication of this truth cri- in a context of the guest. It could be the terion is that on contextualistic grounds one strand we call "the performance of the cook" can adopt the analytic strategy of an alternative occurring in a context of other meals prepared world view in a given situation if doing so by the cook. It could be the strand of "mixing is useful toward some end. For example, a in the eggs," occurring in a context of breaking philosophical contextualist might adopt a the eggs, holding the bowl, and so on. The mechanistic theory because it is useful in quality of the act in each case emerges in identifying ways of "controlling" behavior. the interaction of the strand and its context. Strategic integration of this sort does not The third category of texture, reference, violate Pepper's warning against the destruc- is simply strands more intimately considered. tive effects of eclecticism, because no inte- Reference concerns the temporal relations or gration of the underlying root metaphors is interconnections among the details of an act, implied. The machine metaphor is merely specifically their point of initiation, course, used in the service of a contextualistic agenda; and satisfaction. The concept of reference is the truth of the analysis based on that usage worthy of note because it pertains to issues is evaluated against a successful working cri- of similarity and novelty as contextualistically terion. interpreted. Similarity, for example, is not a feature of events from a contextualistic standpoint: No two events in the world are IMPLICATIONS FOR BEHAVIOR inherently similar. Rather, similarity is an ANALYSIS attribution made when different initiations Pepper did not write World Hypotheses for converge on one satisfaction. Planting a garden psychologists, or even for scientists in general. and going to a restaurant, though formally Thus, the correspondence between his world dissimilar, are regarded as similar to the extent views and philosophies underlying science is that they produce the same outcome, namely only approximate. Nevertheless, particular sustenance. approaches to psychology do seem often to The quality of an act is necessarily threat- have the predominant character of one or ened by examining its texture because any another of Pepper's four relatively adequate given strand of that texture might be ex- world views. perienced as a quality in its own right. This circumstance is a consequence of the dispersive Behavior Analysis as a Contextualistic System character of contextualism: The parts, being The predominant character of behavior derived, may be derived in any number of analysis, or at least what is central and dis- ways. As such, the texture of this new quality tinctive about behavior analysis, is contex- might be examined, one of its strands ex- tualistic. Among the particularly contextu- perienced as a quality, and so on. Were anal- alistic features of behavior analysis are the ysis made for its own sake nothing would concept of the operant, the criterion for truth prevent this process from continuing ad in- or adequacy, the role of the scientist in sci- finitum. For the contextualist, however, anal- entific analysis, and the possibility of novelty. ysis is always for some purpose. The concept of the operant. Several char- The truth criterion of contextualism is acteristics of the operant correspond closely successful working. Analyses are true only to the categorical concepts of contextualism. in terms of the accomplishment of particular An operant is defined as a relation among goals. No postulational provision is made for behavior and stimulus events. The events the evaluation of the goals themselves. Truth participating in an operant cannot usefully may thus exist with regard to relatively trivial be examined independently because their na- goals. This pragmatic view of truth is quite ture depends on their relations to the other radically applied: "The quality of blowing participants. Similarly, for a contextualist, your nose is just as cosmic and ultimate as an act out of context is not an act, categorically Newton's writing down his gravitational for- speaking: "It is not an act conceived as alone STEVEN C. HAYES et al. or cut off that we mean; it is an act in and contextualism from being paralyzed by its with its setting" (Pepper, 1942, p. 232). fluidity is its criterion for the adequacy of Because context must be included in the analysis, namely, successful working. Suc- analysis of an act, contextualists analyzing cessful working always implies success with an act quickly find themselves outside the regard to the accomplishment of some po- confines of the original event of interest (the tentially attainable goal. For this reason, the act) and in the domain of other events (the contextualist rather disparages analysis for context). "The quality of an event is the fused analysis' sake. "Serious analysis for [the con- qualities of its strands, and the qualities of textualist] is always either directly or indirectly its strands come partly out of its context, and practical . . .. If from one texture you wish there we are outside of the event" (Pepper, to get to another, then analysis has an end, 1942, p. 249). The parallels to behavior anal- and a direction, and some strands have re- ysis are obvious. For example, "We cannot levancy to this end and others do not, and account for the behavior of any system while . . . the enterprise becomes important in staying wholly inside it; eventually we must reference to the end" (Pepper, 1942, pp. 250- turn to forces operating upon the organism 251). Likewise, Skinner commented: "It is from without" (Skinner, 1953, p. 35). true that we could trace human behavior not Membership in an operant class in no way only to the physical conditions which shape depends on the formal characteristics of the and maintain it but also to the causes of those behavior involved. Responses share mem- conditions and the causes of those causes, bership in an operant to the extent that they almost ad infinitum" but we need take analysis produce common effects on the environment: only to the point at which "effective action "The consequences define the properties with can be taken" (Skinner, 1974, p. 210). That respect to which responses are called similar" point is the manipulable environment, because (Skinner, 1953, p. 65). This type of clas- only there may successful working toward sification corresponds precisely to the con- the behavior-analytic goals of prediction and textualistic conception of similarity as derived control be achieved and evaluated (Hayes & from "convergence of action on a single effect" Brownstein, 1986). (Pepper, 1942, p. 255). Several behavior-analytic positions are un- An operant has no fixed boundaries. For derstandable in these terms. For example, example, an operant may be anything from behavior analysts object to analyzing private a thumb twitch to cooking a four-course meal. events as causes because an event "is useless We might speak of the larger operant as a in the control of behavior unless we can ma- composition of smaller elements (in Pepper's nipulate it" (Skinner, 1953, p. 34). Searching terms, strands of texture), but speaking in for private causes is thus a "tiresome and this way does not exclude examining the exhausting digression" (Skinner, 1953, p. 35) elements as operants in their own right. Sim- when considered in terms of the purposes ilarly, any aspect of a strand of texture may of analysis. be examined and consequently become the The commitment of behavior analysis to quality of interest. successful working as a truth criterion is also Finally, the behavior-analytic view of be- demonstrated by repeated appeals to it in havior emphasizes the verb-like quality of criticizing other positions. For example, "The all behavioral interactions (e.g., Hineline, objection to the inner workings of the mind 1980). This emphasis parallels precisely the is not that they are not open to inspection root metaphor of contextualism. but that they have stood in the way of the The criterionfor truth or adequacy. As Pepper inspection of more important things" (Skinner, noted, "the analysis of an event consists in 1974, p. 165, emphasis added), and "men- the exhibition of its texture, and the exhibition talism has obscured the environmental ante- of its texture is the discrimination of its strands, cedents which would have led to a much more and the full discrimination of its strands is effective analysis" (Skinner, 1974, p. 165, the exhibition of other textures . . . and so emphasis added). Many such examples are on from event to event as long as we wish available in the writings of Skinner and other to go, which would be forever or until we behavior analysts. Skinner in particular has got tired" (Pepper, 1942, p. 249). What saves been explicit about this truth criterion: "[Sci- BOOK REVIEW entific knowledge] is a corpus of rules for the unavoidable stamp of those particulars effective action, and there is a special sense and cannot thereby be presumed to char- in which it could be 'true' if it yields the acterize the universe. Pepper summarized the most effective action possible.. . . [A] prop- contextualistic view of science this way: osition is 'true' to the extent that with its "[Scientific] schemes, such as maps, diagrams, help the listener responds effectively to the formulas, functional equations, and symbolic situation it describes" (Skinner, 1974, p. 235). systems . . . have been developed on the basis As previously mentioned, a contextualist of past social experience, and their status is can make use of the categorical structures a good deal like that of a social institution. of other world views without becoming philo- ... They constitute what is called 'the science' sophically eclectic. Behavior analysts have of a period, and change from period to period" done just that on occasion, although very likely (Pepper, 1942, p. 267). without full awareness, and much to the con- The possibility of novelty. A central tenet fusion of others. For example, Skinner has of contextualism is the possibility of novelty. talked of humans as complex machines: "We At first glance, this position seems to disqualify have discovered more about how the living contextualism as a scientific philosophy. The organism works and are better able to see position, however, is not that events will be its machine-like properties" (Skinner, 1953, novel, only that they may be. If order be found, p. 47). He did not become a mechanist thereby, so be it. The only evidence for true novelty, because his world view and the truth criterion that is, the utterly uncaused event, is "an it entails are contextualistic, not mechanistic. absence of any evidence to the contrary" He merely borrowed mechanistic models when (Pepper, 1942, p. 260); and by this criterion, doing so seemed useful. Of course, whether true novelty has never been demonstrated. doing so is in fact useful is arguable in any As far as successful working is concerned, given instance. then, the possibility of true novelty can be The role of the scientist in scientific analysis. ignored: It makes no difference (until and The importance of a scientific analysis of the unless it does). behavior of the scientist has long been rec- This aspect of contextualism does not seem ognized within behavior analysis (Kantor, to have obvious parallels in behavior analysis. 1939; Skinner, 1945). Behavior analysts be- Skinner, for example, has noted that although lieve that scientists cannot stand apart from we may not be able to prove that behavior the world under analysis; they are, rather, is "fully determined," the evidence points in a part of that world. Pepper made a similar this direction (Skinner, 1974, p. 189). The point in describing a contextualistic per- contextualistic concept of novelty does indeed spective on this issue. Because analysis is itself show up in behavior analysis, however. Op- an act in context, "the contextualist . . . does erant behavior is said to be "emitted," not not make any exceptions to his analysis of "elicited." Its emission is "occasioned" by analysis, not even for that analysis itself" certain antecedent events, not "caused" by (Pepper, 1942, p. 252). The value of any those events in the mechanistic sense of direct analysis, even of contextualism itself, is to production. From a behavior-analytic stand- be determined by its usefulness in the ac- point, prediction and control pertain not to complishment of some explicit purpose. instances but to classes of behavior. Con- Behavior analysts acknowledge the fact that sequently, no attempt need be made to predict science is, among other things, the action of specific instances of an operant at specific scientists-action meaningful only by ref- times (Skinner, 1938, pp. 10-12, although erence to its context. Accordingly, science is see Skinner, 1957, p. 28). not assumed by behavior analysts to be directed Moreover, no attempt need be made by toward the attainment of ultimate knowledge. behavior analysts to explain variability in Ultimate knowledge is, by definition, context- responding. Variability is assumed, and it may free. As Kantor (1953, pp. 9-25) pointed out, be ignored until not ignoring it makes a dif- science is the work of particular individuals ference for some purpose. Much the same working at particular times in particular places position is taken by evolutionary biologists. with particular materials for particular pur- Genetic mutation and variability allow se- poses. The products of scientists' actions bear lection to operate, but mutations and vari- 104 STEVEN C. HAYES et al.

ability themselves can be said to be "random." metaphor of contextualism is an act in context. In this way, scientific forms of contextualism Accordingly, the experienced quality of an can make room for both orderly relations and act is dependent upon context. Consequently, unexplained variation. a contextualist would be hard pressed to abide by the above definition of behavior without Mechanistic Elements of Behavior Analysis losing contact with the contextualistic root Even though behavior analysis can be metaphor. Such nonfunctional definitions oc- understood as a contextualistic system, it has casionally show up in behavior-analytic writ- aspects that appear mechanistic. Some of these ing. To the extent that they reflect mechanistic aspects may be understood as uses of mech- thinking, they threaten the philosophical con- anistic concepts subordinated to contextualistic sistency of behavior analysis. criteria, but others may represent an intrusion Reductionism. Reductionism applied to psy- of an alternative world view into behavior chology holds that psychological events are analysis. reducible to neural or other biological events. Nonfunctional definitions. To a contextualist, This position is incompatible with contex- acts and their context are inseverable. The tualism, not because biological events are of contextualist who loses sight of this inter- no concern but because in reductionism the dependence loses contact with the quality of parts are primary and the whole derived. In an act, which ultimately leads to the dis- The Behavior of Organisms, Skinner (1938, solution of the contextualistic perspective: "To pp. 418-432) argued against the reduction think of contexts as existing in addition to of psychology to neural events, asserting the or apart from practices is like imagining smiles legitimacy of a science of behavior in its own alongside or beside faces" (Bhaskar, 1983, right. This position is contextualistically sen- p. 87). Skinner has emphasized a similar point: sible. If an analysis works at the level of "Neither [stimulus nor response] may be de- the whole organism interacting in and with fined as to its essential properties without a context, then it works. Analyses at the level the other" (Skinner, 1938, p. 9). of parts of the organism interacting with other The mechanist does not deny the behavior parts are "truer" than other analyses only of the whole, but asserts that it is derivable in terms of particular goals. from the behavior of the parts. The meaning Nevertheless, many behaviorists embrace of a part is its role in the behavior of the biological reductionism. For example, some whole. This role can change as the whole argue a priori that thoughts or feelings are changes. For example, the role of a cogwheel "really" activities of the nervous system. Skin- depends on whether it is part of a clock or ner himself has said: "[The physiologist of being used as a fulcrum. A cogwheel is a the future] will be able to show us how an cogwheel in either case, but the meaning of organism is changed when exposed to con- the part is defined by its potential functions tingencies of reinforcement" (Skinner, 1974, in relation to the potential functions of the p. 215). If he meant that physiologists may other parts of the whole. some day be able to give psychologists the Functional definitions of behavior can thus "real" explanation of reinforcement, then the be incorporated into either world view. For statement is incompatible with contextualism, the contextualist, they are incorporated as because it is based on the primacy of parts an extension of the basic root metaphor; for over wholes. the mechanist, as a reference to the functions This does not mean that reductionism is established among the parts of the machine. always mechanistic, nor that all mechanists Nonfunctional definitions of behavior can are reductionists. The assumption that the be incorporated readily into mechanism as whole is reducible to its parts is categorical a description of the parts themselves, con- in mechanism, but the assumption that psy- sidered separately from their roles. For con- chology is (in principle) reducible to biology, textualists, however, nonfunctional definitions for example, is not required. A mechanical present more of a problem. conception of thinking based on the metaphor Consider the following nonfunctional def- of the brain as computer does not obligate inition of behavior: behavior is muscle move- a concern with neurology. By analogy, if the ments and glandular secretions. The root lever is the root metaphor of a particular BOOK REVIEW 105 mechanistic model, then the laws of the lever "this caused that." Such a causal statement in mechanics provide the categories of the is mechanistic only if it is an extension from model and the question of whether these laws a system of parts, relations, and forces. Ef- can be reduced to the laws of subatomic phys- ficient causality per se is compatible with all ics, for example, does not arise. of the world views except formism. In contextualism, parts are abstractions, Causal talk may carry certain philosophical and therefore reduction of some parts to other assumptions with it, however, especially for parts is merely an analytic, conceptual tool. those who have not carefully articulated their Reductionism of this kind does not imply that philosophical assumptions. "This caused that" the whole is literally reducible to the parts, may lead to the view that this and that are because the parts do not exist independent discrete parts that exist independently of their of analysis. Reducing parts to other parts is relation, or of the scientist's act of analyzing a fiction that may be useful in a given instance. the whole into parts. The construction of Thus, a contextualist might suggest a bio- independent parts invites commentary as to logical explanation for a psychological event the "force" by which this caused that. Perhaps if it is useful in understanding the whole. some contextualists embrace only descriptive Mechanists may say similar things, but forms of science (e.g., Rosnow & Georgoudi, believe that the parts are real-the whole 1986b) in an effort to avoid the seductive is the fiction. For the mechanist, the different effects of causal analyses. If such seduction scientific domains are hierarchically related is likely, then behavior analysts must be es- and must fit together in an integrated and pecially aware of their philosophical as- complete system of knowledge. Given this sumptions if they are to retain their con- assumption, the reduction of one domain to textualistic stance. another is the reduction of the whole to its Reactive organism model. Many behavior fundamental parts (e.g., Bugelski, 1973, p. analysts have argued that the organism is 62) and is thus mechanistic. viewed as active in behavior analysis. How- Causality. Some psychologists (e.g., Howard ever, it is not active in the sense usually & Conway, 1987; Sarbin, 1986) believe that intended by the phrase "active organism concern for efficient causality is mechanistic: model," which is a mentalistic model in which "Efficient causality description is the goal for the organism has purposes and exercises free scientists working with one or another par- will in seeking stimulation, selecting among adigm within the mechanist world view. Be- stimuli, transforming (encoding) potential haviorism and radical empiricism exemplify stimuli into forms that may function differ- psychological and philosophical movements ently from the potential stimuli, and selecting committed to this world view" (Sarbin, 1986, which responses, if any, to make to the trans- p. 6). Based on this belief and the behavior- formed stimuli. The organism in behavior analytic adoption of efficient causality, some analysis does not have these capabilities; rather have concluded that behavior analysis is mech- it behaves consistently with a "reactive or- anistic (e.g., Howard & Conway, 1987). ganism model" (Baltes & Reese, 1977; Reese, Whether the behavior-analytic treatment 1976, 1986a; see also Overton & Reese, 1973; of causality is mechanistic depends on the Reese & Overton, 1970). purposes of this treatment and the criteria by which truth is determined. If one's goals are prediction and control, as are Skinner's, Conflicts Between Behavior Analysis and Other calling something a cause may be understood Psychological Systems to mean that the analytic purposes of the According to Pepper, world views are or- scientist have been accomplished: Prediction thogonal to each other and therefore cannot and control have been achieved. Such a prag- conflict. Apparent conflicts are really pseu- matic interest in efficient causality does not doconflicts, wherein criticisms of one world imply mechanism. view are made in terms of the categorical Mechanists are interested in the primary concepts of another. These kinds of conflicts parts of a system, their relations, and the are illegitimate and cannot be resolved; they forces that make them operate. An interest can only be recognized. Some of the most in efficient causality may lead one to say that persistent conflicts between behavior analysis 106 STEVEN C. HAYES et al. and other psychological systems seem to fit they are vague and imprecise. The flexibility this pattern. of possible units of analysis across various In what follows we consider each world situations might be viewed as undisciplined view in relation to contextualism, from the thinking. Making a place for randomness standpoint of adherents of these world views might be seen as justifying ignorance. A engaged in illegitimate conflicts with each mechanist might argue that contextualists are other. We also consider specific arguments confused about the purposes of science and between behavior analysts and other psy- have no theory, at least no testable one. The chologists to see if they conform to what we post hoc quality of some contextualistic anal- might expect if the arguments were between yses might be cited to illustrate this inad- contextualists and adherents of some other equacy. Most especially, mechanists might world view. If so, they may provide evidence be disturbed by the pragmatic truth criterion that behavior analysis is contextualistic. We contextualists invoke, asserting that contex- intend, further, to evaluate whether Pepper's tualists are merely technicians, more interested thesis makes sense of arguments among psy- in changing events than understanding them. chologists and the difficulties encountered in The arguments between behavior analysts achieving their satisfactory resolution. and other positions seem to fit the pattern Arguments between contextualists and mech- just described. For example, stimulus-re- anists. When contextualists argue against sponse learning theory is a classic example mechanism, they are likely to question the of mechanism within psychology. It has all "thingness" of the pieces of the mechanistic the components of a mechanistic system: stable machine. They might ask: Are these things pieces (e.g., stimulus objects), elaborate models really concrete things? Can they really be of the structural relations of these pieces (e.g., known in isolation and is their nature really Hull's BehaviorSystem), and forces (e.g., drives) not changed by virtue of the relations into to activate the operation of the machine. Truth which they enter? Along these lines, a con- is assessed by correspondence. textualist may attempt to show that the same Behavior analysts have criticized this stim- pieces of the machine function differently given ulus-response formulation, adopting, instead, other histories or other contexts. To this chal- the concept of the operant and an emphasis lenge, a mechanist is likely to argue that on selection by consequences (Skinner, 1974, historical and contextual arrangements of this pp. 222-225). The specific objections to stim- sort change only the operation of the system, ulus-response learning theory are exactly not the nature of its parts. For example, "The what might be expected were the argument machine is built to execute an instruction only made by contextualists against mechanism. when it stands in a particular sort of relation The stability of the elements in stimulus-re- to that instruction.. . . In human psychology, sponse theory was challenged by showing similarly, we assume that people are so con- contextual limits on their application. The stituted" (Stabler, 1984, p. 604). Contex- need for such elements was challenged by tualists, cognizant also of the quality of events providing explanations derived more directly and not just their textural details, might argue from the relations among contextual and be- further that mechanistic theoretical structures havioral events. The various forces were viewed prevent scientists from experiencing the world as mere reification. The theorizing was chal- as it truly exists. Mechanists might answer lenged on the grounds that it takes the scientist that what is real or true is decided on other away from the real world and amounts to criteria. Finally, contextualists might attack vacant construction (e.g., Skinner, 1950). whatever force is invoked to explain the op- More recent arguments of behaviorists eration of the machine, arguing that the force against cognitive psychology can be seen as is inferred from the events it "explains" and arguments of contextualists against mecha- is, thus, an instance of reification. From a nism. Not all cognitive theorizing is mech- mechanistic standpoint, however, a driving anistic (Piaget was an organicist, for example), force of some sort is necessary; it is not a but the kind built on computer models and matter that depends on observation. computer simulation (e.g., Ericsson & Simon, For their part, mechanists might attack 1984) is clearly so. (Jenkins (1974) tried to contextualistic approaches on the grounds that develop cognitivism as a contextualistic po- BOOK REVIEW 107 sition but without great success, because the Arguments between contextualists and or- categorical concepts were simply borrowed ganicists. Contextualists have different con- from Pepper without the development of an cerns with organicism. They might view the adequate psychological system per se.) As with organicists' reliance on coherence as confusing stimulus-response theory, mechanistic cog- the personal reactions of the scientist with nitivists are criticized by behavior analysts the actual value of the scientific work. Like- as building unnecessary theories: "We can wise, the organicists' admission of final causes avoid hypothetico-deductive methods . . . by might be challenged by contextualists and is formulating the data without reference to likely to yield sputterings about circular rea- cognitive processes, mental apparatuses, or soning. The embrace of constructivism might traits" (Skinner, 1984a, p. 523). The stability be criticized because it ignores the context of their categories is challenged. Explanations in which such construction takes place. Fur- in terms of context are offered. ther, the organicists' explanation of events For their part, many cognitivists do indeed by an appeal to their orderly change is likely see behavior analysis as vague and imprecise: to be regarded as confusion of description "The concepts of 'behavior,' 'response,' and with explanation. 'reinforcement' appear to be used so broadly The objections behavior analysts have to that . . . there is hardly anything they do "developmentalism" show the nature of the not seem to encompass. Skinner has stretched last argument quite clearly. Skinner wrote these notions" (Kochen, 1984, p. 600). Sim- that "developmental schedules are really ilarly: "Skinner's proposals have a funda- schedules of changing environments. . . stages mental limitation. They . . . simply are not are changes in the way in which behavior sufficiently precise" (Scandura, 1984, p. 603). acts upon and is reinforced by the (primarily Behavior analysts are also said to have mis- social) environment" (Skinner, 1984b, p. 719). understood the purpose of psychology as a From Skinner's perspective, "We need to go science: "The success of the behaviorist en- beyond mere observation to a study of func- terprise would leave quite untouched the sci- tional relationships" (Skinner, 1938, p. 8). entific problem of accounting for how we are Behavior analysts have attempted to reveal capable [of behaving.] . . . Only an account the shortcomings of organicism by manip- of the machinery within the skin can explain ulating contextual variables that accelerate behavior" (Marshall, 1984, p. 637). And, or delay transitions through developmental finally, many cognitivists see Skinner more stages (e.g., accelerating the acquisition of as a technician than as a scientist: "Skinner's Piaget's object constancy through direct train- emphasis on control is wrong.. . . The em- ing). phasis should be on understanding" (Mill- Organicists should be completely undaunted ward, 1984, p. 528). by such data, however, because they presume The nature of the cognitivists' reply to that organic elements are subject to inter- behavioral criticism is also illuminating. Be- ference and believe that rejecting final causes havior analysts usually criticize mentalism is failing to admit the obvious. The operation on pragmatic grounds. A typical reply to this of a final cause on an acorn inevitably produces criticism is to deny a reliance on immaterial an oak tree. If an acorn is eaten by a pig, entities and processes: "mental processes are it is no longer an acorn and the final cause synonymous with brain processes" (Ellis & of an acorn cannot operate upon it. Hunt, 1983, p. 11). These replies are at times Some organicists attribute to behavior an- explicitly mechanistic: "Mentalists are neither alysts a belief in a passive organism, pushed vitalists nor physiologists, but rather engineers and pulled by mechanical forces in the en- concerned with discovering the type of machine vironment. Behavior analysts might reply that that is man" (Marshall, 1984, p. 637), and the "organism is not . . . passive in the sense "[Behavioral criticisms are irrelevant to] a of submissive" (Skinner, 1984b, p. 719) but form of cognitivism that is not teleological such a reply would probably not alter the but mechanistic" because the "basic analogy- organicists' conclusion. The pragmatic truth the computer-does not involve purposes but criterion of behavior analysis might be seen only mechanistically determined programs" as contributing nothing to intellectual co- (Furedy & Riley, 1984, p. 625). herence and therefore as unimportant to sci- 108 STEVEN C. HAYES et al. entific understanding. Behavior analysts' in- as a "transcendent formist." Plato, one of terest in control might lead organicists to view the original transcendent formists, used the behavior analysis as mechanistic (e.g., Howard idea of norms to explain how persons know & Conway, 1987). so much even though their experiential basis Arguments between contextualists and form- for knowledge is small. Chomsky (1986) has ists. A contextualist might accuse formists of dealt explicitly with one form of what he engaging in analyses that are of no use. What called "Plato's problem": how persons know good is mere classification of events or cat- so much about language even though their aloging of relations among them? Does this experiential basis for this knowledge is small. kind of activity make a difference? If it did, Preexisting and nonexperiential norms pro- the contextualist might be reassured, but only vide the answer: Language emerges according in that specific instance. Formists do not have to a plan or norm housed in a language-ac- practical usefulness as a goal of analysis in quisition device in the "mind/brain," as he general. A contextualist might be concerned calls it (Chomsky, 1986). Underlying sim- that the very ground of similarity, upon which ilarities in grammar occur as approximations formistic laws are built, is shaky. For the to this norm. Behavior analysts question contextualist, similarity is a functional rather whether the regularities discerned by Chom- than formal affair. sky are true (i.e., functional) regularities. For their part, formists might challenge Chomsky's criticism of behavior analysis contextualists on the arbitrariness of their also fits with formistic concerns about con- decision as to what makes a difference. What textualism. He considered the denial of struc- right has a contextualist to proclaim that only tural similarity absurd; terms like "discrim- relations serving their purposes are worth- inative stimuli" are mere metaphors, with while? Formists might also be baffled by the no direct relation to their laboratory referents. contextualists' refusal to acknowledge im- "We can account for a wide class of responses mediately obvious similarities among things. in terms of Skinnerian functional analysis Contextualists might be criticized for con- by identifying the controlling stimuli. But the structing similarities on the basis of only one word stimulus has lost all objectivity in this small aspect of the situation. usage" (Chomsky, 1959, p. 38). The prag- As an example, some of the work on per- matic goals of behavior analysis are taken sonality types is formistic. A personality type to confuse technology with science. For ex- is identified, its characteristics are explored, ample, Skinner's Beyond Freedom and Dignity and its relation to other personality types is (1971) was said to show "a striking failure articulated. From a formistic standpoint, the to comprehend even the rudiments of scientific analysis need proceed no further, and often thinking" (Chomsky, 1972). does not. Behavior analysts respond to this Dealing with arguments across world views. research as contextualists might be expected The list of arguments mentioned above is to respond. They question whether the "sim- not exhaustive. Moreover, many of the ar- ilar" events are really similar, applying the guments are shared among world views. For contextualistic concept of similarity as a test. example, organicists are as likely as mech- This point is the core of Skinner's criticism anists to see contextualism as vague and im- of the "formalistic error" (Skinner, 1969, p. precise. All others should agree that con- 89). Behaviorists also challenge the usefulness textualists confuse technology with science. of this kind of research. To a behavior analyst, Most of this sort of overlap can be predicted relations among traits of personality cannot, from the dispersive and analytic qualities of in and of themselves, directly achieve behavior- the various world views. On issues relevant analytic purposes, because to achieve those to the primacy of parts versus wholes, for purposes one must have access to the con- instance, both formists and mechanists will textual variables of which the relations are disagree with contextualists. held to be a function (Hayes & Brownstein, The philosophical arguments between be- 1986). havior analysts and other psychologists are Chomsky (1986) provides another example. pseudoconflicts among world views. Given Although he may be viewed as an organicist that arguments of this sort are illegitimate (Overton, 1984), he may also be categorized and cannot be resolved, what can we do about BOOK REVIEW 109 them? Pepper suggested three legitimate investigators are unaware of them. When courses of action that behavior analysts can philosophical assumptions are not properly take. They can (a) increase the precision and articulated and organized into a postulational scope of behavior analysis and make these system, scientific formulations based on them improvements evident; (b) analyze the pre- tend to be inconsistent and confusing (Parrott, cision and scope of other systems, when taken 1986). on their own terms and with relevance to If behavior analysis rests on an unsystematic their own purposes (but recognizing that if mixture of mechanistic and contextualistic such an analysis reveals weaknesses, the postulates, it is in danger of being replaced strength of behavior analysis is in no way by more consistent and less confusing psy- increased); and (c) clearly articulate the as- chological systems. Contextualism allows the sumptions and purposes of behavior analysis strategic use of categorical concepts from other and note the differences in the assumptions world views subordinated to contextualistic made by others. For example, one can describe criteria. Skinner can be thought of as a con- quite legitimately a pragmatic truth criterion textualistic philosopher who at times uses and explain how to proceed if such a truth mechanistic theorizing. Although such stra- criterion is adopted; one cannot insist legit- tegic mixtures are consistent with contex- imately that it be adopted. tualism, they carry with them notable hazards. Pepper's advice is difficult to take. Weighing First, avoiding the philosophical impli- in against an intellectual opponent's cate- cations of foreign concepts is difficult. The gorical concepts in terms of one's own is far mechanistic elements of behavior analysis, in easier. After all, one is on well-examined particular, may be erroneously taken to rep- territory in doing so, and if the battle is resent the philosophical basis of behavior engaged on one's own turf, the outcome is analysis. This is extremely dangerous to be- certain. Unfortunately, intellectual adversaries havior analysis because, as a mechanistic sys- will take the same stance, and each combatant tem, it is not very interesting. The analogy will end up on a separate field of battle shout- of the computer is far richer and more elab- ing at a distant foe. Such arguments are no orate, for example. Inevitably, the more el- more than elaborate bouts of name-calling egant form of mechanism will prevail. and bravado. A much more difficult task is Second, borrowing concepts must be com- to develop the adequacy of one's own position, pared to the usefulness of developing concepts to analyze other positions from within, or from within a contextualistic world view. simply to illuminate the nature of the phil- Prematurely using foreign concepts may delay osophical disagreement. Difficult as it may more consistent and effective system building. be, in Pepper's view it is the only intellectually Third, strategic use of other world views honest alternative. can degenerate into undisciplined thinking. A mere reference to "successful working" is no Conflicts Within Behavior Analysis: The justification for intellectual chaos. If foreign Costs of Conceptual Confusion concepts are to be used, they must be used Scientific work always involves philosoph- deliberately and compared to the working of ical assumptions that influence the kinds of alternative domestic concepts. These hazards problems addressed, the methods employed may be minimized, if not altogether avoided, for their solution, and the manner in which if behavior analysts clarify their philosophical findings are interpreted. When properly ar- assumptions. Pepper's analysis may be helpful ticulated and organized into a postulational for this purpose. system, these assumptions serve as criteria against which the internal consistency and The Natural Allies of Behavior Analysis cross-disciplinary compatability of a given If behavior analysis is, in principle, a con- scientific enterprise may be evaluated. textualistic system, who are its natural allies? Philosophical assumptions have their sources Many other scientific systems share contex- in cultural circumstances, and like those cir- tualistic postulates, including some forms of cumstances, they are not all of one piece. evolutionary biology (e.g., Dawkins, 1982), Unfortunately, philosophical assumptions in- cultural anthropology (e.g., Harris, 1979), fluence scientific work even when individual Marxism, ethology, and psychobiology. Yet 110 STEVEN C. HAYES et al. most contextualists think that behavior anal- analysts have not become openly mechanistic. ysis is mechanistic; given the subtleties and Seeing behavior analysis as a contextualistic inconsistencies of behavior analysis this is not system may permit a greater sensitivity to difficult to understand. the crucial choice points between these distinct Sharing the same world view does not mean philosophies. that different theories will not conflict. Indeed, Other goals could also be completely con- according to Pepper, all legitimate intellectual sistent with a contextualistic world view. A conflicts, like all forms of effective collab- contextualistic system directed toward goals oration, occur among proponents of the same other than prediction and control might appear world view. Advocates of different world views alien or even hostile to behavior analysis (e.g., will always either agree to disagree or will Rosnow & Georgoudi, 1986a), but it would see each other as intellectual enemies. Thus, be so only in the sense of being a rival. It to the extent that behavior analysis is mistaken would be a rival member of the same "family for a representative of a world view foreign of theories," that is, theories derived from to it, truly productive collaboration with other a single world view and therefore comparable compatible scientific systems is impossible. with each other on empirical grounds (Reese An example is psychobiology. Psycho- & Overton, 1970). The antipathy between biologists are contextualistic in their thinking. psychobiology and behavior analysis may be They are sensitive to the subtle interactions partially understood in this manner because of ontogenetic and phylogenetic contingencies. the behavior-analytic goal of control is not They appreciate functional definitions and embraced by most psychobiologists. the role of context. Unfortunately, psycho- Behavior analysis is a contextualistic sys- biologists view behavior analysis as an enemy tem. This conclusion significantly alters our (e.g., Gottlieb, 1984). The confusion in be- view of what is central to behavior analysis havior analysis about its own approach has and what is mere historical accident or the- made the identification and construction of oretical preference. For example, when viewed alliances very difficult. as a contextualistic system, behavior-analytic Because contextualism entails no stand on emphases on functional definitions and a prag- what the goal of analysis ought to be, any matic truth criterion stand out in high relief. potentially attainable goal may be entirely Pepper also provides guidance in difficult areas consistent with contextualism. For example, of system development. For instance, it be- many contextualistic findings are "specific, comes obvious that causal laws are ways of often taking the form of accounts justified speaking, not representations of nature. by their coherence" (Fiske, 1988). None- Behavior analysis is a position with a history theless, successful working may be evaluated of success in achieving its goals, but a position with respect to some scientific goals more that fails to develop is a position that fades readily than others. Success with respect to away (Reese, 1986b). How can behavior anal- the goals articulated by behavior analysts, ysis find its way, while retaining the useful namely prediction and control, is readily eval- qualities that made it what it is? Pepper shows uated, at least insofar as these goals apply behavior analysts what is truly unusual and to concrete instances. Thus, these goals and important about their position. His book helps contextualism seem especially harmonious put behavior analysts' hands on the tiller of (Hayes & Brownstein, 1986). In another their philosophical vessel. It may steer better sense, however, they are not. A constant threat from there. is that analyses that help achieve these goals will be taken literally. When causes and effects are viewed as preexisting things in the world, REFERENCES the holistic quality of contextualism is threat- Baltes, M. M., & Reese, H. W. (1977). Operant re- ened, and mechanism is but a short step away. search in violation of the operant paradigm? In B. The behavior analysts who have converted C. Etzel, J. M. LeBlanc, & D. M. Baer (Eds.), New to cognitivism may present examples of the developments in behavioral research (pp. 11-30). Hills- dale, NJ: Erlbaum. process. Given the lack of careful articulation Bhaskar, R. (1983). Beef, structure and place: Notes of the philosophical base of behavior analysis, from a critical naturalist perspective. Journal for the it is perhaps surprising that more behavior Theory of Social Behaviour, 13, 81-95. BOOK REVIEW 111

Bugelski, B. R. (1973). An introduction to the principles Nesselroade & H. W. Reese (Eds.), Life-span de- of psychology (2nd ed.). Indianapolis: Bobbs-Merrill. velopmental psychology: Methodological issues (pp. 65- Chomsky, N. (1959). Review of Skinner's Verbal Be- 86). New York: Academic Press. havior. Language, 35, 26-58. Parrott, L. J. (1986). The role of postulation in the Chomsky, N. (1972). Psychology and ideology. Cog- analysis of inapparent events. In H. W. Reese & nition, 1, 11-46. L. J. Parrott (Eds.), Behavior science: Philosophical, Chomsky, N. (1986). Knowledgeoflanguage. New York: methodological, and empirical advances (pp. 35-60). Praeger. Hillsdale, NJ: Erlbaum. Dawkins, R. (1982). Theextendedphenotype. San Fran- Pepper, S. C. (1942). World hypotheses: A study in cisco: Freeman. evidence. Berkeley: University of California Press. Efron, A. (1980). Pepper's continuing value. In A. Reese, H. W. (1976). Conceptions of the active or- Efron & J. Herold (Eds.), Root metaphor: The live ganism: Discussion. Human Development, 19, 108-119. thought of Stephen C. Pepper. Paunch, No. 53-54, Reese, H. W. (1986a). Behavioral and dialectical psy- 5-53. chologies. In L. P. Lipsitt & J. H. Cantor (Eds.), Ellis, H. C., & Hunt, R. R. (1983). Fundamentals Experimental child psychologist: Essays and experiments of human memory and cognition (3rd ed.). Dubuque, in honor of Charles C. Spiker (pp. 157-195). Hillsdale, IA: Brown. NJ: Erlbaum. Ericsson, A. K., & Simon, H. A. (1984). Protocolanalysis: Reese, H. W. (1986b). On the theory and practice Verbal reports as data. Cambridge, MA: MIT Press. of behavior analysis. In H. W. Reese & L. J. Parrott Fiske, D. W. (1988). Review of Contextualism and (Eds.), Behavior science: Philosophical, methodological, Understanding in Behavioral Science: Implications for and empirical advances (pp. 1-33). Hillsdale, NJ: Research and Theory, edited by Ralph L. Rosnow Erlbaum. and Marianthi Georgoudi. American Scientist, 76, 93. Reese, H. W., & Overton, W. F. (1970). Models of Furedy, J. J., & Riley, D. M. (1984). Undifferentiated development and theories of development. In L. R. and "mote-beam" percepts in Watsonian-Skinnerian Goulet & P. B. Baltes (Eds.), Life-span developmental behaviorism. Behavioral and Brain Sciences, 7, 625-626. psychology: Research and theory (pp. 115-145). New Gottlieb, G. (1984). Lingering Haeckelian influences York: Academic Press. and certain other inadequacies of the operant viewpoint Rosnow, R. L., & Georgoudi, M. (Eds.). (1986a). of phylogeny and ontogeny. Behavioral and Brain Sci- Contextualism and understanding in behavioral science. ences, 7, 688-689. New York: Praeger. Harris, M. (1979). Cultural materialism: The struggle Rosnow, R. L., & Georgoudi, M. (1986b). The spirit for a science of culture. New York: Random House. of contextualism. In R. L. Rosnow & M. Georgoudi Hayes, S. C., & Brownstein, A. J. (1986). Mentalism, (Eds.), Contextualism and understanding in behavioral behavior-behavior relations, and a behavior-analytic science (pp. 3-22). New York: Praeger. view of the purposes of science. Behavior Analyst, 9, Sarbin, T. R. (1986). The narrative as a root metaphor 175-190. for psychology. In T. R. Sarbin (Ed.), Narrative psy- Hineline, P. N. (1980). The language of behavior chology: The storied nature of human conduct (pp. 3- analysis: Its community, its functions, and its lim- 22). New York: Praeger. itations. Behaviorism, 8, 67-86. Scandura, J. M. (1984). New wine in old glasses? Howard, G. S., & Conway, C. G. (1987). The next Behavioral and Brain Sciences, 7, 602-603. steps toward a science of agency. American Psychologist, Skinner, B. F. (1938). The behavior of organisms. New 42, 1034-1036. York: Appleton-Century. Jenkins, J. J. (1974). Remember that old theory of Skinner, B. F. (1945). The operational analysis of memory? Well, forget it! Amertcan Psychologist, 29, psychological terms. Psychological Review, 52, 270-277. 785-795. Skinner, B. F. (1950). Are theories of learning nec- Kantor, J. R. (1939). The nature of psychology as essary? Psychological Review, 57, 193-216. a natural science. Acta Psychologica, 4, 1-61. Skinner, B. F. (1953). Science and human behavior. New Kantor, J. R. (1953). The logic ofmodern science. Bloom- York: Macmillan. ington, IN: Principia Press. Skinner, B. F. (1957). Verbal behavior. New York: Kochen, M. (1984). Problem solving as a cognitive Appleton-Century-Crofts. process. Behavioral and Brain Sciences, 7, 599-600. Skinner, B. F. (1969). Contingencies of reinforcement: Kuhn, T. S. (1962). The structure ofscientific revolutions. A theoretical analysis. New York: Appleton-Century- Chicago: University of Chicago Press. Crofts. Marshall, J. C. (1984). Mechanism at two thousand. Skinner, B. F. (1971). Beyondfreedom and dignity. New Behavioral and Brain Sciences, 7, 637. York: Knopf. Millward, R. (1984). Cognitive science: A different Skinner, B. F. (1974). About behaviorism. New York: approach to scientific psychology. Behavioral and Brain Knopf. Sciences, 7, 527-529. Skinner, B. F. (1984a). Methods and theories in the Overton, W. F. (1984). World views and their influence experimental analysis of behavior. Behavioral and Brain on psychological theory and research: Kuhn-Lakatos- Sciences, 7, 511-523. Laudan. In H. W. Reese (Ed.), Advances in child Skinner, B. F. (1984b). Reply to Catania. Behavioral development and behavior (Vol. 18, pp. 191-226). New and Brain Sciences, 7, 718-719. York: Academic Press. Stabler, E. P., Jr. (1984). Rule-governed behavior in Overton, W. F., & Reese, H. W. (1973). Models of computational psychology. Behavioral and Brain Sci- development: Methodological implications. In J. R. ences, 7, 604-605.