Contextualism: Is the Act in Context the Adequate Metaphor for Scientific Psychology?

Psychonomic Bulletin & Review 1994, 1 (2), 239-249

Contextualism: Is the act in context the adequate metaphor for scientific psychology?

E. J. CAPALDI and ROBERT W. PROCTOR Purdue University, West Lafayette, Indiana

According to some, psychology as it has been practiced is based on a world view known as mech anism. Individuals from a number of different areas of psychology, most recently within the behavior-analytic community, have strongly argued that psychology should be based on a differ ent world view, contextualism. They emphasize a variety of characteristics that, in their view, differentiate a contextualistically based psychology from one based on mechanism. We examine these characteristics and find them to be of dubious value for differentiating a contextualistic approach to psychology from others. One proposal of some advocates of contextualism is that con textualistic approaches should develop independently from most ofthe remainder ofpsychology, which they regard as mechanistic. This proposal is said to be derived from the metaphilosophy ofPepper (1942). We evaluate this proposal and reject it. We go on to suggest that the mechanis tic/contextualistic dichotomy is too constraining to realistically describe various approaches to psychology.

In 1942, Stephen C. Pepper published a book entitled sequently, by critically focusing on the case advanced by World Hypotheses , in which he presented a framework the behavior analysts, we should be able as well to sup for classifying philosophic systems. This framework has ply a general critique of the more salient arguments met with considerable acceptance among a variety of psy offered by various adherents of Pepper 's metaphilosophy. chologists from different fields over the past two decades. Pepper's classifications have been used by some ofthese PEPPER'S METAPHILOSOPHY psychologists to reject what may be regarded as the dom inant or conventional approaches to theory and research According to Pepper (1942), most philosophic positions in psychology in favor of certain newer alternatives. If can be subsumed under one or another of four relatively these alternative approaches should come to be widely ac adequate world views. He goes on to say that each ofthese cepted, the result would be a revolution in the way psy world views derives from a distinctive root metaphor that chological research is practiced and theory is generated. provides a different, coherent, and defensible way of in Any arguments advocating radical change obviously de terpreting the world. These four world views arejormism, mand close scrutiny. However, although a few of the ar organicism, mechanism, and contextualism. guments stemming from Pepper's metaphilosophy have The root metaphor of the world view known as formism encountered some opposition (e.g., from Kendler, 1986; is similarity. The members of any dass of objects or ideas Marr, 1993; Staddon, 1993), so far they have escaped are similar to one another either because these members the extensive examination they deserve. We intend to rec are the reflection of some independently existing norm tify this shortcoming here. (Plato) or because we can see in the particulars that all The most recent proponents of radical change in psy members of the dass manifest the norm (Aristotle). For chology based on Pepper's metaphilosophy are individu Plato, oak trees are oak trees because they participate in als within the behavior-analytic framework stemming the norm for oak trees, whereas for Aristotle, all oak trees from the work of B. F. Skinner. Because the behavior have a common essence. The truth criterion of formism analyticarguments, more than others that have been made, is correspondence of the individual to the characteristics are addressed specifically to how experimental psychol of the dass. For example, this particular bird is a bird ogy is practiced, they are most relevant to members of because it has the characteristics peculiar to the dass of the Psychonomic Society and deserve particular attention. birds. Although some of the arguments advanced by certain For organicism, the root metaphor is the growing or members of the behavior-analytic community are unique, ganism. Things are seen as developing toward certain in many cases they correspond to those advanced by other ends, according to predetermined stages. Piaget's stages psychologists, such as developmental psychologists. Con- of cognitive development make up one example of organi cism within developmental psychology. The criterion of Correspondence should be addressed to E. J. Capaldi, Psychology truth for organicism is coherence. Systems-as, for ex Department, Psychology Building 1364, Purdue University, West ample, scientific theories-progress over time (i.e., de Lafayette, IN 47967-1364. velop) toward greater and greater inclusion or coherence.

Thus, Newton's theory is superior to that of Anaximenes. (1991), a behavior analyst, has argued that "for psychol Why? "Because, answers the organicist, it includes vastly ogy, mechanism is the wrong world-view for its subject more data, because these data are much more determinant, matter, and always was" (p. 124). He goes on to say, and because these determinant data are so closely in "Contextualism, though, is the bold new alternative-an tegrated so that in very large measure they are all mutu alternative that is continuing to evolve" (p. 143). ally implicative" (Pepper, 1942, p. 3(0). Contextualism has had its initial and major impact on Of almost exclusive concern here are the remaining two developmental psychology (see, e.g., Ford & Lerner, world views, mechanism and contextualism. According 1992; Lerner & Kaufman, 1985; Reese, 1991). It also to Pepper (1942), the root metaphor ofmechanism is the has been influential in environmental psychology (e.g., machine. The parts of the machine are assumed to inter Altman & Rogoff, 1987), social psychology (e.g., Ger act in a lawful manner to produce the functioning of the gen, 1989), and clinical psychology (e.g., Terwee, 1990). entire machine. For example, the earth revolves around Arecent book by Gillespie (1992) argues that cognitive the sun in a fixed pattern determined by the mutual gravita psychology should become contextualistic, and contex tional attraction ofthe two bodies. Mechanism is regarded tualism has been strongly advocated as the proper ap by many as the metaphor of science and of most contem proach for behavior analysis in particular and for psychol porary psychology (e.g., Gillespie, 1992; Morris, 1988). ogy in general (e.g., Hayes, Hayes, & Reese, 1988). Mechanism would be identified with Newton's approach, Despite the numerous differences among the advocates for example. Newton attempted to explain a variety of of contextualism, a theme that runs through the various phenomena on the basis of a small set of laws. The truth contextualistic accounts is that stimulus-response (S-R) criterion of mechanism is correspondence between hy psychology and cognitive psychology are highly similar, pothesis and experimental findings. This version of the in that both are exemplars of mechanistic systems and have correspondence criterion of truth is so well known and serious shortcomings as approaches to psychology. widely accepted as to require no elaboration. One may judge the fervor with which some have come The root metaphor ofcontextualism is the ongoing act to embrace contextualism from remarks by Sarbin (1993), in context. Contextualism sterns from pragmatism and is a personality-social psychologist. He states that coming associated with the philosophic systems ofWilliam James, to accept contextualism in place of mechanism is a very C. S. Peirce, and John Dewey. In contextualism, things difficult experience, requiring "something akin to reli are always seen as changing, but the change is not neces gious conversion" (p. 53). To begin to appreciate how sarily progressive, as in organicism. According to Pep different his contextualistic approach is from the so-called per (1942), "The ineradicable contextualistic categories mechanistic approach, consider the following: may thus be said to be change and novelty" (p. 235). Pep We have crossed the threshold into a new era. We have per states, "To give instances of this root metaphor in joined ranks with scholars who make the case for social our language with the minimum risk of misunderstand constructivism, for hermeneutics, for ethnomethodology, ing, we should use only verbs. It is doing, and enduring, for psychobiography, for dramaturgy, for narrative psy and enjoying" (p. 232). Contextualism favors an opera chology, for ethogeny, all patently contextual, all ready tional theory of truth, which has three distinct specifica to advance knowledge through the study of exemplars or tions. One of these truth criteria is that of "successful specimens. Wehavelived through a conversion experience, working." In terms ofthis criterion, Pepper states, "Truth andcan nowseethe worldfromthe perspective of contex is utility or successful functioning, and that is the end of tualism, the root metaphor of which is the narrative. We it" (p. 270). According to Pepper, the successful work live in a story shaped world. Looking back, the dreamof a productive andcomplete science of psychology basedon ing criterion eschews hypothesis testing altogether. The mechanistic principles served as a powerful incentive to true is whatever works in a particular situation. For ex do psychology according totheconventional paradigm. To ample, a rat finding its way to the goalbox has solved a day, we acknowledge that it was only a dream. (p. 63) practical problem and, thus, has isolated the true path. Because it will be more appropriate to consider the other THE CASE FOR CONTEXTUALISM two truth criteria later in this paper, we defer considera ADVANCED BY BEHAVIOR ANALYSTS tion of them. Among the most ardent advocates of contextualism in THE EMERGENCE OF CONTEXTUALISM the past few years have been several behavior analysts. IN PSYCHOLOGY Behavior analysis, in common with much of the remainder of psychology, has in the past been considered by some Since about 1970, the notion that mechanism is the to be based on the world view known as mechanism (see, proper world view for science in general and psychology e.g., Gillespie, 1992; Morris, 1988, 1993a). Many be in particular has come to be increasingly rejected. It has havior analysts continue to see psychology in general, but been argued by a diverse array ofpsychologists that con not behavior analysis, as mechanistic (Hayes, 1988; Hayes textualism provides either an additional approach to psy et al., 1988; Morris, 1988). Hayes et al. state flatly, "Be chology or one that should replace mechanism entirely havior analysis is a contextualistic system" (p. 110). Simi (Ford & Lerner, 1992; Hayes, Hayes, Reese, & Sarbin, larly, Morris (1988) asserts that "contemporary behavior 1993: Rosnow & Georgoudi, 1986). For example, Morris analysis ... is contextualistic in world view, in contrast CONTEXTUAUSM MI to its traditional characterization as being mechanistic" TUE ARGUMENTS FOR CONTEXTUALISM: (p. 289), and Hineline (1990) states, "in its purest form, A DETAILED EXAMINATION behavior analytic theory is contextualistic to the extent of virtually excluding mechanistic metaphors" (p. 309). The previously mentioned arguments advanced by The initial and most influential arguments for behavior Hayes et al. (1988) and by Morris (1988, 1991, 1993a, analysis' being contextualistic rather than mechanistic 1993b) for classifying a given psychological approach as were advanced in a review of Pepper's (1942) book by contextualistic will be examined below. In considering Hayes et al. (1988), which appeared in the Journal ofthe whether an approach is contextualistic, neither Hayes Experimental Analysis 01 Behavior. Some of the argu et al. nor Morris, nor any other adherent, has suggested ments made by Hayes et al. are unique to behavior anal that a specific combination of characteristics must be taken ysis; others overlap with those made by psychologists in together. Rather, they have merely listed in serial fashion other areas. According to Hayes et al. (1988), the con the characteristics that, in their opinion, render a partic textualistic features of behavior analysis that distinguish ular approach contextualistic. Therefore, in this paper, it from mechanism are the concept of the operant, the cri we evaluate each of these purported characteristics in terion for truth or adequacy, the role of the scientist in serial fashion. scientific analysis, and the possibility of novelty. Addi tional arguments suggesting that behavior analysis is con Contextual Characteristics of the Operant textualistic have subsequently been supplied by Morris According to Hayes et al. (1988), contained in contex (1988,1991, 1993a, 1993b). The arguments advanced by tualism is a view of behavior that is compatible with the Morris, who sees considerable similarity between be concept of the operant, as it has been developed in be havior analysis and developmental psychology, provide havior analysis. Therefore, the operant has several major additional correspondence to those advanced by contex characteristics that Hayes et al. see as particularly com tualists in other areas of psychology. Morris suggests that patible with contextualism. However, as discussed below, contextualistic behavior analysis, unlike mechanistic ap each of the purported contextualistic characteristics of the proaches found in S-R psychology and cognitive psychol operant identified by Hayes et al. can be found in other ogy, is holistic rather then elementaristic, postulates an psychological approaches that have not been considered active rather than a passive organism, rejects linear causal contextualistic. ity, and is not reductionistic. Act in context. In contextualism, the context must be It is one thing to describe contextualism and mechanism included in the analysis of an act, because an act out of as abstract systems, as Pepper (1942) has done. But it is context is not an act, and this is true of the operant, ac another thing to suggest what the implications of the two cording to Hayes et al. (1988). They state, "The events positionsrnight be for concrete scientificpractice, as many participating in an operant cannot usefully be examined psychologists have done. It is of course the latter activity independently because their nature depends on their re that is of concern here. We will suggest that many of the lations to the other participants" (p. 101). We suggest that characteristics said to render behavior analysis, or indeed although it may be true that behavior analysts include con any other area of psychology, a contextualistic system fall text in the analysis of the act, the importance of includ into one oftwo categories: Either they are for one reason ing context is similarly recognized in many branches of or another of dubious value for that purpose, or, as oc psychology, if not by all psychologists. However, recog curs in the majority of cases, they can be found in equal nizing that context may be of critical importance in many abundance in most approaches to psychology. If we are instances does not make one a contextualist in the sense correct, then either all of psychology is contextualistic or, discussed in this article. as we think, the criteria are vacuous. The importance of context has been recognized in a re Notice that, according to the position advanced here, cent book on learning, entitled Context and Learning we are not arguing that psychology or some area of psy (Balsam & Tomie, 1985), of which the following are chology is or is not contextualistic, or mechanistic for that among a small sampIe of the opinions expressed by chap matter. Rather, our purpose consists in demonstrating that ter authors. Balsam states, "At a logical and procedural a convincing case has not been made that behavior anal level, all learning occurs in context" (p. 1). Similarly, ysis in particular or scientific psychology in general, as Rescorla, Durlach, and Grau (1985) state, "Modern dis currently practiced, is or has been contextualistic. Per cussions of Pavlovian conditioning increasingly acknowl haps convincing arguments can be advanced, but so far, edge the importance of the context in which learning and we shall argue, this has not been the case. Having reached performance occur" (p. 23). Also, Medin and Reynolds this conclusion, we will go on to examine and reject a (1985) state, view advocated by Hayes et al. (1988) that contextually The present chapter describes work where context is af minded psychologists should not compete directly with forded a central status and where context effects are mani mechanistically minded psychologists. Finally, we will fest as strengths as weIl as limitations. We aim to show raise a neglected question: How appropriate is it to apply that analyses of context can provide insights into some key the terms contextualism and mechanism to science in gen problems associated with learning and memory research. eral and to psychology in particular? (p. 323) 242 CAPALDI AND PROCTOR

Clearly, the authors of these chapters, in common with the meaning of behavior is determined by the context in other authors in the book, acknowledge the importance which it occurs. For example,jire uttered in a theater has of context for understanding learning. An extended dis an entirely different meaning thanjire uttered on the ri cussion of how a variety of modern learning theories deal fle range. Moreover, the phonemic restoration effect de with context may be found in Hall (1991). scribed above illustrates that context can alter the act of The inclusion of context is hardly limited to learning hearing itself. theories. Context is also presumed to playa crucial role A good example of a theory in which context is fun in perception, cognition, and memory. Jenkins, a theorist damental to determining what is learned is supplied by widely recognized as emphasizing context in his approach Gibbon and Balsam's (1981) scalar expectancy theory to memory, goes so far as to say this of modern memory (SET). Gibbon and Balsam state, "The central idea es theory: "Everyone now knows that memory phenomena poused is that the association between the signal and food are much more complicated and contextually determined depends on an appreciation of the improvement in the than we used to think they were" (1979, p. 430, empha average delay to reinforcement in the signal compared sis ours). The importance attributed to context in mem with the average delay overall or in the background" ory research is illustrated by the wide variety of studies (p. 248). According to SET, to judge the effect of a par examining environmental, internal, and semantic context, ticular learning trial, we have to take into account not only many of which are reviewed in the book Memory in Con that learning trial but its relation to other events. More text: Context in Memory (Davies & Thomson, 1988). specifically, strength of expectancy in the signal and Consider now perception. One of the classic studies in strength of expectancy in the background are constantly this area is an experiment by Bruner and Minturn (1955), undergoing change, both in absolute terms and relative in which an ambiguous character was read as a letter when to each other, depending on (I) whether the reinforcing presented along with letters and as a number when pre event follows the signal, and at what delay or trace inter sented along with numbers. In other words, the percep val, (2) whether the signal occurs alone, (3) whether the tion of the stimulus was influencedby the context in which reinforcing event occurs alone, and (4) the temporal spac it was embedded. Another example is the phonernic resto ing of events in the background, including no event in ration effect discovered by Warren (1970). He presented the background. Just as Hayes et al. (1988) say that the subjects with sentences such as the following, in which events participating in an operant cannot usefully be ex the asterisk indicates a phoneme replaced by a nonspeech amined independently, Gibbon and Balsam are saying that sound: the effect of a Pavlovian trial cannot be judged indepen dently of its relationship to (I), (2), (3), and (4) above. It was found that the *eel was on the axle. Outside the act. Contextualists analyzing an act find It was found that the *eel was on the shoe. themselves outside of the original event of interest, and so too do behavior analysts, according to Hayes et al. It was found that the *eel was on the orange. (1988). We suggest that while this may be true, it is also It was found that the *eel was on the table. the case that mechanists analyzing events often find them selves outside of the original event of interest. For ex Even though the disambiguating context occurred after ample, in an attempt to understand the effects of context the missing phoneme, subjects reported actually hearing on, for example, learning or perception, so-called mech for each of the four sentences above wheel, heel, peel, anistically oriented psychologists attempt to construct the and meal, respectively. Many other examples from other ories. In this way, they find themselves outside of the areas of psychology could be cited. original event ofinterest. Thus, a theory developed in an Although indicating that mechanists deal with context, attempt to understand the effects ofcontext on a particu Hayes (1988) and Morris (1991) suggest that they do so lar behavior may be applied to a variety of other differ differently from contextualists. For example, Hayes ent behaviors (see, e.g., Hall, 1991). It is the attempt to (1988) states, "For a mechanist, context is relevant, but understand context, and not whether one is either a mech it does not alter the nature of the act itself' (p. 11). Simi anist or a contextualist, that results in one's interest's being larly, Morris suggests that in contextualism, unlike in directed outside of the original act. This may be illustrated mechanism, "context imbues behaviour with meaning; by the recently reported finding that preexposure of the the meaning of behaviour emerges from its context" conditioned stimulus (CS) affected sensitivity to context (p. 136). We think that the view expressed by Hayes and (Swartzentruber & Bouton, 1992). Three distinct explan Morris is only partially correct. In some instances, the atory mechanisms were suggested-that preexposure of possibility that context is simply another variable influ the CS might affect its associability , its processing, or encing behavior, albeit an important one, is justified. Bad retrieval ofthe CS-US association. Each ofthese distinct deley (1982) has used the term independent context to refer explanations suggests a different line of further investi to such instances. In other instances, however, context gation outside of the original event of interest. does have more fundamental effects; but this is generally Acts versus movements. Responses share membership realized and accepted by all researchers. In these in in an operant class to the extent that they produce com stances, which Baddeley has called interactive context, mon effects on the environment, according to Hayes et al. CONTEXTUALISM 243

(1988). They go on to say that an operant may be any employed in classical conditioning appear to fall in this thing from a thumb twitch to cooking a four-course meal. category" (pp. 42-43). Moreover, Adams (1984), an ex Morris (1991) makes the same argument, in a somewhat pert in the area of motor learning, has suggested that al different manner. He states, "Mechanism adheres to though concern with the act is legitimate, so is concern elementarism, representing behavior and the environment with movement sequences. He has gone so far as to state as but collections of materially fundamental, atomic re that preoccupation with the act in both human and ani sponses and stimulus elements (or switches). Complex ac mallearning has had deleterious consequences for the un tion is then but an associative compounding of the basic derstanding of movement sequences (see especially pp. 3 elements and their interrelations" (p. 132). and 4 of his article). Although some psychologists may identify the animal Verbs. In contextualism, the verb-like quality of be as learning particular and specific behaviors, the idea that havioral interaction is emphasized. This is true of behavior responses share membership in an operant dass to the ex analysis, according to Hayes et al. (1988), as well as tent that they produce common effects on the environment Morris (1993a), who states, "Behavior analysis takes as has long been recognized by many psychologists. Indeed, its subject matter what we describe in the grammar of it was examined experimentally by Muenzinger (1928) verbs-behaving, constructing, and thinking" (p. 28). prior to the time it was emphasized by behavior analysts. Compare Morris's statement to that of Tolman (1932): The well-known Muenzinger study was prominently men tioned by Tolman (1932), a cognitive behaviorist, to il A rat running a maze; a cat getting out of a puzzle box; a man driving horne to dinner; a child hiding from a lustrate the "multiple trackness" of behavior. In that stranger; a woman doing her washing or gossiping over study, guinea pigs given over 1,000 acquisition trials used the telephone; a pupil marking a mental-test sheet; a psy a variety of responses over the course of the experiment. chologist reciting a list of nonsense syllabies; my friend Thus, the animal might use its left paw for a number of and I telling one another our thoughts and feelings-these trials and then switch over to using its right paw, and then are alt behaviors. (p. 8) switch again to using its left paw. As may be seen, even for a cognitive behaviorist presuma Tolman (1932) cites another study by Muenzinger, bly on the opposite pole from behavior analysis, the oper Koerner, and Irey (1929), in which the experimenters lim ant need not be restricted to any particular dass of be ited the animals to the use of the right paw. However, havior, and each of the behaviors described by Tolman even here, the animals showed considerable response vari emphasizes verb-like interactions, no less than those de ability. In Muenzinger et al.'s words: scribed by Morris above. Would it be possible for any The right paw might be used with a firm and deliberate variety of behaviorism to ignore the verb-like quality of pressing down of the lever, or a bare touch on the lever, behavioral interactions? or a quick tap, or a succession of quick taps, or brushing It seems clear to us that a point-by-point comparison or sliding over the lever, or circular movement partly of the contextualistic position adopted by Hayes et al. around the lever, or a gentle placingof the paw on the lever (1988) on the operant with that ofTolman (1932) and sub with pressure after some time, or a crossing of the right sequent writers on the nature of the response would re paw over the left. (p. 427) veal them to be highly sirnilar, to the point ofbeing almost All of these behaviors were regarded by To1manand Muen identical. Perhaps this is not accidental. In his later writ zinger as equivalent, despite their physical differences. ings, Pepper (1967) favored a new world view, seen by Tolman's approach emphasizing response equivalence him as highly sirnilar to contextualism, called selectivism. continues to be influential on the contemporary scene. For Selectivism was heavily influenced by the writings ofTol example, Mackintosh (1974) specifies several theories of man. The root metaphor of this new world view is the instrumentallearning that "rnay be regarded as modern goal-seeking, purposive act, which Pepper saw as an adap vers ions of the theory of learning advanced by Tolman tation of the position advocated by Tolman in his 1932 (1932)" (p. 199). Mackintosh continues, "One of the book, Purposive Behavior in Animals and Men. It may be most direct sources of support for such a theory is that that in focusing on acts in context, one is driven to an provided by evidence of response equivalence" (p. 199). analysis of behavior that stresses goals and purposes, as Mackintosh then goes on to cite many additional experi exemplified by Hayes and Hayes's (1992) treatment of the mental examples ofthe sort of findings originally reported goals of analysis. Thus, many implications that follow by Muenzinger and colleagues. from contextualism, such as stressing goals and purposes, The view that behavior may be characterized in terms for example, may follow equally well from other positions ofacts rather than specific movements has been so widely that would not typically be considered contextualistic. accepted that we find the contrary opinion expressed by Hayes et al. (1988) and Morris (1991) to be utterly with The Role of the Scientist in Scientific Analysis out foundation. For example, Spence (1956), a leading There are two features of the contextualistic approach S-R psychologist, said, "Certainly all of the responses to science that Hayes et al. (1988) suggest differ from involved in instrumental conditioning and selective learn those of a mechanistic approach. First, contextualists "be ing situations would be classified as acts rather than move lieve that scientists cannot stand apart from the world ments, and contrary to common belief, even the responses under analysis; they are, rather, apart of that world" 244 CAPALDI AND PROCTOR

(p. 103). Second, contextualists suggest that science is not velopment. On the contrary, in some important respects, directed "toward the attainment of ultimate knowledge" though by no means all, Einstein's general theory of rela (p. 103). Although these positions may characterize the tivity is closer to Aristotle's than either ofthem is to New views of contextualists, they also wouldbe widely accepted ton's. (pp. 206-207) by psychologistsand scientistsof various persuasions, and To summarize, there seems to be little doubt that psy we cannot understand how this could be doubted. chologists in general, in common with contextualists, be Hanson (1988), a philosopher of science who is quoted lieve that scientists do not stand apart from their subject favorably by a wide variety ofpsychologists, suggests that matter and that ultimateknowledge is not withinour grasp. our very observations are influenced by the theories that we hold. He states: The Possibility of Novelty To say that Tycho and Kepler, Simplicius and Galileo, Regarding the contextualistic category of novelty, Hooke and Newton, Priestley and Lavoisier, Soddy and Hayes et al. (1988) note that "this aspect of contextual Einstein, De Broglie and Born, Heisenberg and Bohm all ism does not have obvious parallels in behavior analy make the same observations but use them differently is too sis" (p. 103). However, they go on to assert that "the easy. It does not explain controversy in research science. Were there no sense in which they were different observa contextualistic concept of novelty does indeed show up tions, they could not be used differently. (p. 193) in behavior analysis" (p. 103). One problem with this view is that novelty is not defined in a clear, unambigu It is hard to imagine a more intimate relationship between ous manner by either Pepper (1942) or Hayes et al. That scientists and their subject matter than that suggested by is, no contextualist, to our knowledge, has indicated how Hanson, because it involves perception itself. Most books we would recognize novelty when and if it should occur. on psychological research methods state flatly that our ob For example, how would we distinguish between an event servations are influenced by our preconceptions. Indeed, unexpected on the basis of current scientific theory and in an undergraduate research methods book, Meyers and one that was truly novel? Failure to deal with this issue Grossen (1978) cite Hanson favorably in making this is a very serious deficiency, for we lack the criteria for point. evaluating a fundamental category of contextualism Few, if any, psychologists would claim that science novelty. Another problem with novelty, as specified by deals with ultimate knowledge. Consider a point of view pure contextualism, is that it does not preclude total dis suggested by Evans (1985) in a textbook on research order. As Cicirelli (1994) has suggested, "In pure con methods: textualism, not only is there no end-point to development, Reflecting the skeptical attitude, scientific understanding but the direction of development is totally unpredictable" is tentative and ever-ehanging. Thus, theories and conclu (p. 34). sions are constantly undergoing revision as additional data These problems aside, one possible meaning of novelty are accumulated. The replacement of Newtonian physics is that laws or phenomena that hold for one period, how with relativity theory and quantum mechanics is a classic ever long or short, may not hold in some succeeding pe example ofthis. In Newton's day, his view ofthe universe riod. Clearly, being a contextualist is not a necessary re was the truth. But the truth changed rather quickly with quirement for acknowledging this logical possibility. It the appearance of new accounts that could explain more. is one that seems as open to mechanists, formists, and (p. 19) organicists as to contextualists. None of us is in a posi The examples above were obtained simply by examin tion either to affirm or deny that laws are or are not im ing a few books at random. Dozens of similar examples mutable. In any event, whether or not the possibility of could be found with relatively little effort. If this be novelty in this sense is accepted would seem to have no doubted, consider Kuhn (1970), certainly one of the most impact on how science is practiced, unless or until such influential philosophers of science on contemporary psy novelty appears and, moreover, is recognized. chology. As regards standing apart from one's subject matter, Kuhn (1970) suggests, "To an extent unparalleled Active Versus Passive Organism in most other fields, they [the members of a scientific com Morris (1991) states, "Within the mechanistic account, munity] have undergone similar educations and profes causation is characterized in terms of the effects of effi sional initiations; in the process they have absorbed the cient independent causes on dependent material-material same technicalliterature and drawn many of the same les that is otherwise at rest" (p. 134). Morris suggests that sons from it" (p. 177). As a result ofthese educational in contextualism, as opposed to mechanism, performance experiences, members of a scientific community are con must be understood in terms of an environment interact strained to see matters in a particular way and find it dif ing with the organism. The notion that a passive organ ficult to communicate with others educated in a different ism is postulated by all but contextualists is simply in "disciplinary matrix." As for ultimate truth, Kuhn states: correct. That the organism is an active information I do not doubt, for example, that Newton's mechanics im processor has been advocated in any number of sources, proved on Aristotle's and that Einstein's improved on New a few of which we will cite here. ton's as instruments for puzzle-solving. But I can see in ln a classic paper, which deals with the problem of serial their succession no coherent direction of ontological de- order behavior, whose influence in a variety of areas of CONTEXTUALISM 245 psychology would be difficult to overestimate (see, e.g., contextualists have attributed to scientists beliefs they do Bruce, 1994), Lashley (1951) indicated: not hold, rnistaking strategic features for ontologicalones. My principal thesis today will be that input is never into a quiescent or static system, but always into a system which Reductionism is already excited and organized. In the intact organism, Morris (1993a) argues that, unlikecontextualism, mech behavior is the result of interaction of this background of anism, and thus psychology in general, is necessarily excitation with input from any designated stimulus. Only reductionistic in the sense that the aim is to reduce psy when we can state the general characteristics of this back chology to a more fundamental discipline, ultimately ground of excitation, can we understand the effects of a physics. We find this viewpoint to be fallacious for two given input. (p. 112) reasons. First, as Bechtel(1988)has indicated, within psy Similarly, Postman (1972), in characterizing the organiza chology in general and cognitive psychology in particu tional approach to memory, said the following: lar, there is a considerable difference of opinion as to whether reductionism, in Morris's sense, is possible. One of the sources of appeal of the concept of organiza Fodor (1974), for example, is an outspoken opponent of tion is undoubtedly that it implies an active learner who reductionismwithin cognitivepsychology, contendingthat imposes structure on the material to which he is exposed. it is not possible to construct bridge laws equating terms The subject is credited with discovering and utilizing the systematic features introduced into the material by the ex in psychology with those of lower-level disciplines. perimenter, and also with inventing idiosyncratic bases of A second matter that Morris's view overlooks is that, organization. (p. 18) even within mechanism, a major alternative approach to reductionism, interfie1d theory, has been proposed. Inter Ina widely used textbook concerned with human mem field theory is regarded by some as more realistic than ory, Klatzky (1980) described her approach as follows: reductionism. Bechtelindicatesthat interfield theories "do "From the information-processingperspective adopted by not attempt to derive one theory from another but rather this book, '" human memory is depicted as a continu seek to identify relationships between phenomena studied ously active system that receives, modifies, stores, by the two different fields of inquiry" (p. 97). After de retrieves, and acts upon information" (p. ix). More re scribing a variety of mechanistic examples of interfield cently, in a textbook devoted to animal and human mem theory, Bechtel goes on to state the following: ory, Spear and Riccio (1994) state that there has been an "increasing realization that humans typically are active The mechanistic interfield theories sketched here open rather than passive processors of the information to be another possibility. This model can accommodate inter learned" (p. 13). Thus, from at least the early 1950s to action between cognitive and neural inquiries without re the present, it has continuously been recognized that quiring reduction. It allows both analyses to inform each other in the attempt to develop an interfield theory, but does learners are active rather than passive in the processing not require subsuming one explanation under another. of information. (p. 107) Linear Causality The Criterion for Truth Another characteristic that Morris (1993a) attributes to mechanism and that has been accepted by contextualists Both Hayes et al. (1988) and Morris (1988) argue that in general is unidirectional, linear causality. Morris says, the truth criterion for contextualism is different from that "In adhering to unidirectional and linear causality, the espoused by mechanism and corresponds to that already mechanistsexplanatory model is an asymmetric, one-way employed by behavior analysis. This criterion is that of relation between independently defined causes and ef successful working, which "implies success with regard to the accomplishmentof some potentiallyattainablegoal" fects" (p. 34). Morris means by this that organisms are (Hayes et al. , 1988, p. 102). In another paper, Hayes merely the recipients of the forces that act upon them. The arch-mechanist, Newton, would hardly agree with (1988) states, "all events and all analyses are evaluated this. Newton's third law of motion states as follows: To in terms of successful working" (p. 12). Hayes also says, every action there is always opposed an equal reaction; "More than anything else it is this truth criterion that or, the mutual actions of two bodies upon each other are shows the contextualistic qualities of behavior analysis" (p. 11). There is a major problem with this argument, in always equal and directed toward contrary parts (Halli our view. day & Resnick, 1966). According to Pepper (1942), successful working is not How can such a serious misunderstanding of mecha the best truth criterion of contextualism (p. 272). How nism be explained? The likely reason has to do with the the contextualistically rnindedbehavior analysts have man way that experimental research is normally conducted. Inexperimentalarrangements, as a practical matter, seien aged to ignore Pepper's position on this issue is difficult to understand, because Pepper states it so clearly and tists seek to examine the effects of independent variables explicitly: on dependent variables. They are well aware, however, that effects occur in the reverse direction, but ignore them The "successful working" theory excludes hypotheses from not for epistemological or ontological reasons but for truth, yet hypotheses are prominent textures in contextu purely practical ones. If our reasoning here is correct, alism, and they are the very textures to which usage im- 246 CAPALDI AND PROCTOR

plies truth and falsity. A contextualistic theory of truth that should develop in isolation from each other. Must, be leaves hypotheses out ofthe theory is not fitting common cause effective communication between different world sense truth as closely as possible into the contextualistic views is impossible. And should, because conceptualcon categories. In fact, the "successful working" theory is only fusion arises from mixing of metaphors. a haljhearted contextualistic theory. (p. 272, emphasis ours) Hayes et al. (1988) state, "using the categories of one Pepper goes so far as to say that successful working is world view to analyze and criticize another is illegitimate not only the narrowest of the truth criteria of pragmatism, and inherently useless" (p. 98). In another place, Hayes but also "the one the enemies of pragmatism try to as (1988) says, "Pepper claimedthat it is illegitimate to criti sociate with it" (p. 269). On the basis of our reading of cize one world view in terms of the categorical concepts the philosophicalliterature on pragmatism, we would cer of another" (p. 12). That, in Hayes's opinion, there is tainly agree with Pepper's evaluation of successful work little to be gained from either competing with or criticiz ing and think that most pragmatists would too. It is be ing mechanistic positions, or attending to their criticisms, yond our understanding why modern proponents of is clear from the following statement: "The only true in contextualism in psychology would associate themselves tellectual allies and adversaries of a given specific posi with a position that in the past has been used to under tion will always be found within a given world view" mine their cause. (Hayes, 1988, p. 13). Hayes and Hayes (1992) state in Pepper (1942) describes two other more adequate con connection with mechanistic and contextualistic versions textualistic truth criteria. The first ofthese is that ofverifi of radical behaviorism, "The two positions are incom cation of hypotheses. "The slogan of this type of opera patible, and the differences cannot be resolved by way tionalism is that truth is verification. According to this of a compromise" (p. 231). The astounding implication formulation, it is not the successful act that is true, but of this position, it will be recognized, is that contextu the hypothesis that leads to the successful act" (p. 272). alistic behavior analysis should be as independent of The trial-and-error behavior of a rat in a maze would pro mechanistic behavior analysis as of most of the remainder duce true and false judgments according to the success of psychology. ful working criterion, but not according to the verified The first weakness ofthe preceding argument is, as we hypothesis criterion. However, ifthe animal were enter have shown previously, that the criteria which are said taining some hypothesis as to the correct path, its judg to distinguish between contextualistic and mechanisticap ments would be true or false. The verification criterion proaches to psychology fail to do so. A second weakness, seems more in character with the rest of psychology than in our opinion, is that the contextualistically minded be with behavior analysis. havior analysts have misinterpreted Pepper's (1942) po A third truth criterion is qualitative confirrnation, sition on the competition between world views, although which, like the second one, emphasizes confirrnation of we must admit that Pepper is to some extent ambiguous hypotheses. However, it adds to this the view that "the on this score. Consider the following statements from body of hypotheses possessed by science and philosophy Pepper: gives us a considerable amount of insight into the struc It is iIlegitimate to disparage the factual interpretations of ture ofnature" (Pepper, 1942, p. 278). Pepper doesnot one world hypothesis in terms of the categories of elaborate on what he means by the structure of nature. another-if both hypotheses are equally adequate. (p. 98) In any event, the two more adequate truth criteria of con textualism, as described by Pepper, are quite close to the It is the cognitive obligation of a world theory to interpret general position adopted by most psychologists. Conse the danda [a type of corroborationJ and categories of other quently, the truth criterion of successful working cannot world theories in terms of its own categories. (p. 100) be used to suggest that some area of psychology is more We need all world hypotheses, so far as they are adequate, contextualistic than the remainder of psychology. for mutual comparison and correction of interpretive bias. (p. 101) HAYES ET AL.'S INTERPRETATION OF When we say that world theories are mutually exclusive, THE RELATIONSHIP BETWEEN we do not mean that they stand apart from one another like MECHANISTIC AND so many isolated posts. Each theory is weil aware of the CONTEXTUALISTIC SYSTEMS others, criticizes and interprets them, and entirely includes them within its scope. (p. 104) According to Hayes et al. (1988), who claim to be ex Pepper's (1942) entire treatment ofthis issue seems to pressing Pepper's belief, with which they agree, differ suggest that the world views are in a constant state of ent world views "cannot compete directly" (p. 98). growth and healthy competition. How can our interpre Moreover, as regards conflicts between world views, tation of Pepper on the relation between world views be Hayes et al. (1988) state, "According to Pepper , world reconciled with that of Hayes et al. (1988)? In our view, views are orthogonal to each other and therefore cannot Pepper discouraged criticizing the fundamental assump conflict. Apparent conflicts are really pseudoconflicts.... tions of one world view in terms of the fundamental as These kinds of conflicts are illegitimate and cannot be sumptionsof another world view, much as one might criti resolved; they can only be recognized" (Hayes et al. , cize an ethical system from a scientificviewpoint. Clearly, 1988, p. 105). World views, they maintain, must and criticism ofthis sort is not particularly useful. Pepper did CONTEXTUALISM 247

not intend, however, to discourage criticizing what we division of all philosophie systems into formism, organi shall call the products of a particular world view from cism, mechanism, and contextualism. Such seemingly un some other standpoint. Examples of products include such critical acceptance does not seem entirely healthy to us. things as some particular theoretical formulation that is Consider, for example, the proposition that the commonly developed from within a world view (e.g., HuIl's mech accepted metaphor for science in general and psychology anistic S-R theory) or some particular methodology that in particular is the machine. We have two observations sterns from a world view (e.g., factor analysis or statis about this proposition. First, even if one accepts the ma tics, generally). Our interpretation of Pepper is that such chine metaphor, and certainly it has been used extensively products of world views are fair game for competition, within some areas of psychology, it does not imply that conflict, and criticism. (See Gillespie, 1992, for a con one literally thinks that all aspects of psychology are textualist who agrees with our view on competition be machine-like. Take the metaphorical utterance, The Presi tween contextualistic and mechanistic systems.) In our dent is the captain ofthe ship ofstate. We do not literally opinion, then, Pepper encouraged, rather than discour think that the President is steering a ship. Rather , what aged, meaningful dialogue between individuals who sub we attempt to illuminate by this metaphor is that in form scribe to different world views concerning the products ing government policy the President must avoid making of those views. mistakes (i.e., avoid the reefs), that he must generate the Regardless of whether our interpretation of Pepper or best possible plan (steer a good course), that the final de that of Hayes et al. (1988) is the correct one, we would cision is his (as a captain is in charge of the ship), and suggest a third reason for rejecting their position that con so on. In other words, when using metaphors, we take textualistic systems should develop in isolation from other them not literally but metaphorically. As du Preez (1991) formulations, such as mechanistic behavior analysis and has indicated, cognitive psychology: It is a course of action that seems counterproductive. By not engaging in dialogue with What we ought to do is to attempt to explore the range and power of each metaphor as a heuristic device. Each opens others over specific interpretive and methodological dif up a set ofproblems, questions and possible answers. Each ferences, one invites the charge of being unreasonable, can be no more than a slice of the total. (p. 26) if not anti-intellectual, No scientific approach that is seen as being either unreasonable or anti-intellectual will be Our second point is that, as the preceding quote im able to exercise influence beyond its own boundaries. This plies, the machine metaphor is not the only popular surely is not an end to be desired, even if further, more metaphor within psychology. For example, du Preez adequate arguments than have been supplied so far should (1991) suggests that the foundation metaphor for social ultimately reveal some portions of psychology to be con psychology is theater (roles, actors, role distance, role textualistic. relations, scripts). Another popular metaphor in many We recognize that it is always difficult for people of areas of psychology is games. As other examples, the different persuasions to communicate with each other, but mind has been considered to be a blank slate to be writ we do not see this problem as an insurmountable one. On ten on by experience (Locke) or an active organizing space the problem inherent in such communication, we agree with apriori ideas and theories (Chomsky). Many other with the prominent philosopher of science, Karl Popper metaphors that are commonly used in psychology could (1970), who has commented as follows: be added to this list (see du Preez, 1991). Thus, characterizing science exclusively in terms of the I do admit that at any moment we are prisoners caught in machine metaphor has serious limitations. There is a better the framework of our theories; our expectations; our past way to characterize science, in our opinion. We prefer to experiences; our language. But we are prisoners in a Pick think of science as an empirieal, analytic endeavor that wickian sense; if we try, we can break out of OUT frame deals with closed as weIl as open systems. By a closed work at any time. Admittedly, we shall find ourselves again system, we mean one in which the ideal is to exclude from in a framework, but it will be a better and roomier one; and we can at any moment break out of it again. The cen consideration al1 variables except those that we wish to ex tral point is that a critical discussion and a comparison of amine. A prime example of a closed system is what seien the various frameworks is always possible .... The diffi tists attempt to achieve in conducting experiments. Note culty ofdiscussion between people brought up in different that the empirical/analytic approach is useful no matter frameworks is to be admitted. But nothing is more fruitful what the metaphor employedto characterize psychology-a than such a discussion. (pp. 56-57) machine, astage, agame, etc. By an open system, we mean one in which significant HOW APPROPRIATE IS IT Ta APPLY variables not under the control of the investigator intrude THE TERMS CONTEXTUALISM to influence the phenomena of interest. Open systems are AND MECHANISM TO to be found when one examines behavior in real world PSYCHOLOGY? settings. One example of an open system is what contex tualists mean by the term multilevel context. In a multilevel Having examined a number of articles and books writ context, behavior would be understood in terms of ten by contextualistically minded psychologists, we find "atomic, molecular, chemical, biological, physical, psy no contextualist who has quarreled with Pepper's (1942) chological, community, social, societal, cultural, and 248 CAPALDI AND PROCTOR

historical variables, and so on" (Cicirelli, 1994, p. 32). BALSAM, P. D., & TOMIE, A. (Eds.) (1985). Context and learning. Hills For many contextualists, exarnining phenomena in closed dale, NJ: Erlbaum. systems distorts the phenomena or limits their generality BECHTEL, W. (1988). Philosophy 0/science: An overview for cognitive science. Hillsdale, NJ: Erlbaum. to restricted if not artificial contexts. BRUCE, D. (1994). Lashley and the problem of serial order. American Gillespie (1992), an advocate of a contextualistic ap Psychologist, 49, 93-103. proach to cognitive psychology, utilizes a multilevel anal BRUNER, J. S., & MINTURN, A. L. (1955). Perceptual identification ysis in her attempt to decide whether or not to send her and perceptual organization. JournalofGeneral Psychology, 53, 21-28. CICIRELLI, V. G. (1994). The individual in the family life cycle. In daughter, Gemma, to kindergarten. Gillespie concludes L. L' Abate (Ed.), Handbook 0/developmental family psychology and that a better decision could be made by employing mul psychopathology (pp. 27-43). New York: Wiley. tilevel contextualisticcategories rather than those normally DAVIES, G. M., & THOMSON, D. M. (Eds.) (1988). Memory in con associated with closed systems. She may be correct in this text: Context in memory. Chichester, U.K.: Wiley. judgment. However, the problem of whether or not to DU PREEZ, P. (1991). A science 0/ mind. San Diego, CA: Academic Press. send Gemma to kindergarten is not one that admits of a EVANS, J. D. (1985).lnvitation to psychological research. New York: scientific answer. It is doubtful that we have generallaws Holt, Rinehart & Winston. that apply to this particular case. Even if we had such FODOR, J. A. (1974). Special sciences (or: Disunity ofscience as a work laws, we could not apply them to Gemma because we ing hypothesis). Synthese, 28, 97-115. FORD, D. H., & LERNER, R. M. (1992). Developmental systems the know little ofher social, cultural, and genetic background, ory: An integrative approach. Newburry Park, CA: Sage. to which these laws would presumably apply. We also GERGEN, K. J. (1989). Social psychology and the wrong revolution. lack knowledge of the kinds of situations that Gemma European Journal 0/ Social Psychology, 19, 463-484. might encounter in kindergarten. Is the teacher there to GIBBON, J., & BALSAM, P. (1981). Spreading association in time. In her liking? Will her classmates be friendly or unfriendly? C. M. Locurto, H. S. Terrace, & J. Gibbon (Eds.), Autoshaping and conditioning theory (pp. 219-253). New York: Academic Press. The failure to have answers to these and many other ques GILLESPIE, D. (1992). The mind's we: Contextualism in cognitive psy tions precludes a scientific analysis of this and like ex chology. Carbondale: Southern Illinois University Press. amples of open systems. HALL, G. (1991). Perceptual and associative learning. Oxford: Ox It may be that in some open, multilevel systems, an in ford University Press, Clarendon Press. HALLIDAY, D., & RESNICK, R. (1966). Physics (Parts land II). New terpretive (i.e., hermeneutical) approach is more useful York: Wiley. than one based on information from closed systems. As HANSON, N. R. (1988). Observation. In E. D. Klemke, R. Hollinger, one example, a contextualistic-interpretive analysis rnight & A. D. Kline (Eds.), Introductory readings in the philosophy 0/sei provide a better basis for action than an empirical/ana ence (rev. ed., pp. 184-195). Buffalo, NY: Prometheus. lytic one in the case of Gemma. As another example, a HAYES, S. C. (1988). Contextualism and the next wave of behavioral psychology. Behavior Analysis, 23, 7-22. hermeneutical approach to an open system might provide HAYES, S. C., & HAYES, L. J. (1992). Some clinical implications of interesting hypotheses (context of discovery) that could contextualistic behaviorism: Theexample ofcognition. Behavior Ther then be subjected to empirical analysis in a closed sys apy, 25, 223-249. tem (context of justification). No doubt much useful sei HAYES, S. C., HAYES, L. J., & REESE, H. W. (1988). Finding the philosophical core: A review of Stephen C. Pepper's World hypoth ence has been produced in just this way. Be this as it may, eses: A study in evidence. Journal 0/ the Experimental Analysis 0/ . we see the distinction between open systems (in which Behavior, 50, 97-111. interpretation may be practiced) and closed systems (in HAYES, S. C., HAYEs, L. J., REESE, H. W., & SARBIN, T. R. (Eds.) which variables may be isolated)as much more fundamen (1993). Varieties ofscientific contextualism. Reno, NY: Context Press. tal than that between mechanism and contextualism, which HINELINE, P. N. (1990). The origins of environment-based psycholog ical theory. Journal 0/ the Experimental Analysis 0/ Behavior, 53, are just two of the many metaphors that have been em 305-320. ployed to facilitate understanding. Regardless, it has been JENKINS, J. J. (1979). Four points to remember: A tetrahedral model the major purpose of this paper to demonstrate that the of memory experiments. In L. S. Cermak & F. I. M. Craik (Eds.), case has not been made that contextualism, as defined by Levels ofprocessing in human memory (pp. 429-446). Hillsdale, NJ: Erlbaum. its current adherents, provides an approach to scientific KENDLER, T. S. (1986). World views and the concept of development: psychology fundamentally different from that stemming A reply to Lerner and Kauffman. Developmental Review, 6, 80-95. from other metaphors. KLATZKY, R. L. (1980). Human memory: Structure and processes (2nd ed.). San Francisco: W. H. Freeman. KUHN, T. S. (1970). The structure ofscientific revolutions (rev. ed.). REFERENCES Chicago: University of Chicago Press. LASHLEY, K. S. (1951). The problem of serial order in behavior. In ADAMS, J. A. (1984). Learning ofmovement sequences. Psychologi L. A. Jeffress (Ed.), Cerebral mechanisms in behavior: The Hixon cal Bulletin, 96, 3-28. symposium (pp. 112-136). New York: Wiley. ALTMAN, 1., & ROOOFF, B. (1987). World views in psychology: Trait, LERNER, R. M., & KAUFMAN, M. B. (1985). The concept ofdevelop interactional, organismic, and transactional. In B. Stokols & I. Alt ment in contextualism. Developmental Review, 5, 309-333. man (Eds.), HaruIbook0/environmental psychology (pp. 7-40). New MACKINTOSH, N. J. (1974). The psychology 0/ animallearning. New York: Wiley. York: Academic Press. BADDELEY, A. D. (1982). Domains of recollection. Psychological MARR, M. J. (1993). Contextualistic mechanism or mechanistic con Review, 89, 708-729. textualism?: The straw machine as tar baby. Behavior Analyst, 16, BALSAM, P. D. (1985). The functions of context in learning and per 59-65. formance. In P. D. Balsam & A. Tomie (Eds.), Context and learn MEDIN, D. L., & REYNOLDS, T. J. (1985). Cue-context interactions in ing (pp. 1-22). Hillsdale, NJ: Erlbaum. discrimination, categorization, and memory. In P. D. Balsam & CONTEXTUALISM 249

A. Tornie (Eds.), Context and leaming (pp. 323-356). Hillsdale, NJ: REESE, H. W. (1991). Contextualism in developmental psychology. In Erlbaum. H. W. Reese (Ed.), Advances in child development and behavior MEYERS, L. S., & GROSSEN, N. E. (1978). Behavioral research: The (pp. 187-230). Orlando, FL: Academic Press. ory, procedure, and design (2nd ed.). San Francisco: W. H. Freeman. RESCORLA, R. A., DURLACH, P. J., & GRAU, J. W. (1985). Contextual MORRlS, E. K. (1988). Contextualism: The world view ofbehavior anal learning in Pavlovian conditioning. In P. D. Balsam & A. Tomie ysis. Journal of Experimental Child Psychology, 46, 289-323. (Eds.), Context and learning (pp. 23-56). Hillsdale, NJ: Erlbaum. MORRlS, E. K. (1991). The contextualism that is behavior analysis: An ROSNOW, R. L., & GEORGOUDI, M. (Eds.) (1986). Contextualism and alternative to cognitive psychology. In A. Still & A. Costall (Eds.), understanding in behavioral science. New York: Praeger. Against cognitivism: Alternative foundations for cognitive psychology SARBIN, T. R. (1993). The narrative as the root metaphor for contextu (pp. 123-149). Hempstead, U.K.: Harvester-Wheatsheaf. alism. In S. C. Hayes, L. J. Hayes, H. W. Reese, & T. R. Sarbin MORRlS, E. K. (l993a). Behavior analysis and mechanism: One is not (Eds.), Varieties ofscientific contextualism (pp. 51-65). Reno, NV: the other. Behavior Analyst, 16, 25-43. Context Press. MORRlS, E. K. (I993b). Contextualism, historiography, and the his SPEAR, N. E., & RICCIO, D. C. (1994). Memory: Phenomena and prin tory ofbehavior analysis. In S. C. Hayes, L. J. Hayes, H. W. Reese, ciples. Boston: Allyn & Bacon. & T. R. Sarbin (Eds.), Varieties ofscientific contextualism (pp. 137 SPENCE, K. W. (1956). Behavior theory and conditioning . New Haven, 165). Reno, NV: Context Press. CT: Yale University Press. MUENZINGER, K. F. (1928). Plasticity and mechanization ofthe prob STADDON, J. E. R. (1993). Pepper with a pinch of psalt. Behavior lem box habit in guinea pigs. Journal of Comparative Psychology, Analyst, 16, 245-250. 8, 45-69. SWARTZENTRUBER, D., & BOUTON, M. E. (1992). Context sensitivity MUENZINGER, K. F., KOERNER, L., & IREY, E. (1929). Variability of of conditioned suppression following preexposure to the conditioned an habitual movement in guinea pigs. Journal of Comparative Psy stimulus. Animal Learning & Behavior, 20, 97-103. chology, 9, 425-436. TERWEE, J. S. (1990). Hermeneutics in psychology and psychoanaly PEPPER, S. C. (1942). World hypotheses: A study in evidence. Berkeley: sis, New York: Springer-Verlag. University of California Press. TOLMAN, E. C. (1932). Purposive behavior in animals and men. New PEPPER, S. C. (1967). Concept and quality, La Salle, IL: Open Court. York: Appleton-Century-Crofts, POPPER, K. R. (1970). Normal science and its dangers. In I. Lakatos WARREN, R. M. (1970). Perceptual restorations of missing speech & A. Musgrave (Eds.), Criticism andthe growth ofknowledge (pp. 51 sounds. Science, 167, 392-393. 58). New York: Cambridge University Press. POSTMAN, L. (1972). A pragmatic view of organization theory. In E. Tulving & W. Donaldson (Eds.), Organization ofmemory (pp. 3 (Manuscript received December 21, 1993; 48). New York: Academic Press. revision accepted for publication February 14, 1994.)