The Behavior Analyst 1986, 9, 175-190 No. 2 (Fall) Mentalism, Behavior-Behavior Relations, and a Behavior-Analytic View of the Purposes of Science Steven C. Hayes Aaron J. Brownstein University of Nevada-Reno University of North Carolina at Greensboro

In a behavioral view, the purposes of science are primarily prediction and control. To the extent that a scientist embraces both of these as a unified and generally applicable criterion for science, certain phil- osophical and theoretical practices are counterproductive, including mentalism in both its metaphysical and metatheoretical forms. It is possible and often worthwhile to recast some mentalistic talk into an issue ofbehavior-behavior relations. When behavior-behavior relations are approached non-mechanisti- cally, however, analysis cannot stop at the level of the relations themselves. Several analytic concepts common in the behavioral community share some ofthe dangers ofmentalism ifnot employed properly, including such concepts as self-reinforcement, response-produced stimulation, and self-rules.

Criticism of a behavioral approach to amine some of the philosophical under- human behavior has been frequent since pinnings of a behavior-analytic view of its inception. Recently, a type ofcriticism science. Many of the points we hope to has emerged from knowledgeable critics make have been made elsewhere, but the suggesting that there is a more rapid path discussions have often been directed to- to the kinds ofscientific knowledge sought ward other specific concerns and have by behaviorists (e.g., Keat, 1972; Wes- not always been interconnected or given sells, 1981, 1982). Even individuals who a comprehensive rationale. Keystones of formerly have been sympathetic to a be- a behavior-analytic position can then ap- havior-analytic position have embraced pear to be dogmatic or arbitrary, rather this line of criticism. For example, Kil- than required for the intellectual integrity leen has suggested that we need to "re- of the position. store the excitement" in our field by ad- Our starting point will be the primary mitting mentalism (Killeen, 1984). purposes of science from a behavior-an- Behavorists would enthusiastically alytic viewpoint-prediction and con- embrace this suggestion if the alterna- trol. We will attempt to show that an tives being proposed (e.g., Killeen's emphasis on prediction and control is not "emergent behaviorism," 1984) ad- arbitrary in behavior analysis because it vanced the goals of science as seen by is a necessary part of successful forms of behavior analysts. Sadly, that possibility the philosophy that underlies behavior- does not appear likely. Instead, the critics analytic theorizing. We will examine seem to be proposing a kind of science mentalism from several vantage points that is ill-suited to the scientific ends and show that regardless of its form, sought by behavior analysts. Although the mentalism is necessarily counterproduc- criticisms may seem to be about scientific tive to the purposes ofscience embraced strategies or tactics, in actuality they con- by behavior analysts. Mentalism can, cern the behavior-analytic view of sci- however, contribute to the purposes of entific explanation itself. science as seen from other perspectives. The nature ofthe recent criticism sug- We will examine a behavioral translation gests that it may be worthwhile to reex- of some types of mental phenomena in terms ofbehavior-behavior relations, but warn against uses of the translation that Requests for reprints may be sent to Steven C. also interfere with the accomplishment Hayes, Department of Psychology, University of Nevada-Reno, Reno, NV 89557-0062. This paper of prediction and control. We will then is published in loving memory ofAaron J. Brown- briefly examine several concepts within stein (1932-1986). We will miss you, old friend. a behavioral perspective that are on a 175 176 STEVEN C. HAYES & AARON J. BROWNSTEIN slippery slope to the same problem cre- a function" (Skinner, 1953, p. 35). This ated by mentalism. position is not based on an a priori dic- tum that only what are commonly called THE PURPOSES OF SCIENCE: "external variables" can possibly influ- PREDICTION AND CONTROL ence behavior. In a sense, the flow is in the opposite direction. Ifan event can in When differences in goals are made ev- principle directly allow both prediction ident, many arguments seem to dissolve. and control ofbehavior, then it deserves The goals of science from a behavior- the name "environmental variable" or analytic viewpoint have been quite ex- "external variable." This relation can be plicitly stated: "We undertake to predict shown by the fact that, in behavior anal- and control the behavior of the individ- ysis, the words "external" or "environ- ual organism. This is our 'dependent mental" do not always refer to the world variable'-the effect for which we are to outside the skin: rather they refer to the find the cause" (Skinner, 1953, p. 35). It world outside behavior (most of which, seems only fair to evaluate a position with of course, is outside the skin). For ex- respect to the goals it sets for itself, while ample, it is sometimes useful ("useful" recognizing ofcourse that other purposes in terms of prediction and control) to might be well served by different posi- think of an "internal environment" in- tions. fluencing behavior. An emphasis on "external variables" The Emphasis on Control comes from the goals ofscience as viewed Prediction and control are the primary by behavior analysis. Seeking both pre- goals of behavior analysis (the goals of diction and control puts certain con- interpretation and explanation are dis- straints on the kinds of statements of re- cussed later). The behavioral approach lations that are useful for the scientist. places emphasis on the words "and con- Only statements that point to events ex- trol" in the phrase "prediction and con- ternal to the behavior of the individual trol." Behavior analysis has sought an organism being studied can directly lead explanation ofbehavior in terms ofevents to prediction and control. The logic of that are ofa kind that at least potentially this claim is as follows. Scientific state- allows both prediction and control si- ments, as Skinner pointed out, are not multaneously. As we discuss below, some themselves the causes ofthe phenomena kinds of descriptions of events and re- they encompass. Bodies do not fall, for lations can in principle only directly pro- example, because of the law of gravity. duce successful prediction and not con- Rather humans can bring their behavior trol. Other kinds ofdescriptions ofevents under the control ofthis verbal statement and relations can allow both, in principle, of a relation (i.e., this rule) and achieve though of course for practical reasons certain ends. Thus, scientific rules help control may presently be impossible. Be- us accomplish particular ends by describ- havior analysis is committed to empha- ing contingencies. They are rules for sci- sizing this latter kind of analysis. entists, not rules for the world. Thus, sci- The importance ofprediction and con- entists who seek prediction and control trol as a guide to behavior-analytic theo- must rely on rules that start with the en- rizing cannot be overemphasized. It is vironment, in the sense of the "world the key to understanding many behav- outside of the behavior," because scien- ioral positions that might otherwise ap- tists are, and can only ever be, in other pear to be arbitrary. For example, be- organisms' environment in this sense. havior-analytic accounts ofbehavior are Scientists cannot directly use a rule to always ultimately to be cast in terms of control behavior unless it starts where "environmental variables" or "external the scientist is-in the potentially man- variables": "Our 'independent vari- ipulable world outside of the behavioral ables'-the causes of behavior-are the system. If a scientific statement is used external conditions of which behavior is by the scientist to control phenomena but MENTALISM AND SCIENCE 177 does not itself start from where the sci- radical behaviorism is a dynamic spatio- entist resides there must be some un- temporal contextual unit-none of the specified and unanalyzed link between the terms can be defined independently of statement and the scientist's behavior. It any of the others. Radical behaviorism may be useful, but it is necessarily in- is so thoroughgoing in its attempt to ana- complete in the sense that the rule itself lyze context that even the behavior of did not specify whatever was done to scientists as they conduct contextual control the phenomena of interest. analyses is to be understood through more contextual analyses (Skinner, 1945). Why Prediction Must Be Included and The explanatory model of the "act in Control Emphasized context" is shared by many perspectives on behavior, from certain forms of evo- Much of the confusion over a behav- lutionary biology (e.g., Dawkins, 1982), ior-analytic perspective on the goals of to some types of cultural anthropology science would have vanished had control (e.g., Harris, 1979), to Marxism. Because alone been emphasized as the defining a basic explanatory model (or "root met- property of science. To the extent that aphor") is at the core of any well-inte- control is pursued, it forces us to em- grated and consistent world view, Pepper phasize external variables that are func- (1942) has suggested that all perspectives tionally related to the behavior. Indeed, that rely on the "act in context" as an ifthe critics ofa behavioral position were explanatory model be thought ofas types right and the emphasis on control rep- of the world view he calls "contextual- resented merely an emphasis on tech- ism." According to Pepper (1942), the nological advancement, there would have underlying "truth criterion" of contex- been no reason not to emphasize control tualism is "successful working" or prag- alone as the issue. The criticism is a fairly matism (Pepper, 1942). A term, concept, common one. For example, Wessells or statement of a relation is not true or (198 1) objects to an emphasis on control, false simply according to public agree- saying "the kinds ofpredictions one aims ment about the correspondence between to make need not be dictated by the prag- it and other events, but according to the matic desire to change behavior for the impact that the use ofthe term, concept, better" (p. 161). The concern over con- or statement has on dealing successfully trol, however, is not primarily with the with the phenomena of interest. Radical development of technology, though the behaviorism clearly encompasses such a technological outgrowth of behavior view (e.g., Skinner, 1945). analysis is impressive and a legitimate The nature of contextualism and of source of intellectual support for the po- other world views, in Pepper's fullest sition. At a theoretical level, the concern sense, is a complicated topic. It would is primarily with the completeness ofthe take us beyond the scope ofthis paper to account according to the scientific goals defend fully the claim that behavior anal- embraced by behavior analysis. A be- ysis fits all of Pepper's defining charac- havior-analytic position on the goals of teristics for contextualism (but see Hayes science in turn derives its dignity from & Reese, in press). For our purposes, we the necessary relation between these par- will refer to behavior analysis as a type ticular goals and the overall philosophi- ofcontextualism only in the limited and cal integrity of the world view repre- uncontroversial sense of a view of the sented by behavior analysis (cf. Reese, world consistently based on an analysis 1984). ofthe spatiotemporal context ofacts, and Behavior analysis is based on a per- driven by a pragmatic theory of truth. vasive use of a particular explanatory The problem with the "act in context" model: the "act in context." In behavior as an explanatory model is that it does analysis, any event is to be understood not and cannot specify the scope of the and even defined through a contextual act or the context. Context can proceed analysis. The three-term contingency of outward spatially to include all ofthe uni- 178 STEVEN C. HAYES & AARON J. BROWNSTEIN verse. Context can proceed backward in coincidence, or the wrong construction time infinitely to include the remotest an- could have been placed on the proper tecedent, or forward in time to include units of behavior and thus on its ante- the most delayed consequence. The "act" cedents, and so on. It is primarily the in question can vary from the finest mus- confirmation of the validity of the units cle twitch to the most elaborate and ex- of the analysis that requires that control tended behavioral sequence. Conse- be emphasized in successful forms of quently, in behavior analysis, an operant contextualism. can be of almost any size and, in prin- An additional problem in behavior ciple, can be influenced by contingencies analysis (or any contextualistic perspec- that are extremely remote or indirect. tive) is that even the behavior of a sci- Under such circumstances, one might ask entist analyzing behavior is itselfbehav- how we are to know that a particular con- ior that can be analyzed. Contextual textual analysis (in behavior analysis, a analysis can easily be paralyzed, how- particular "contingency analysis") is ad- ever, by the need to analyze context, and equate? then to analyze the context ofthe analysis "Successful working" provides an an- of context, and so on ad infinitum. Pre- swer to this question. An analysis need diction and control both confirm the val- proceed only to the point at which suc- ue of the analysis and provide an end- cessful action can be based on it. Suc- point to the need for analysis-it need cessful action confirms the value of the continue only to the point at which pre- analysis-it changes the indeterminent diction and control is possible in prin- units of contingency analysis in the ab- ciple. Without control as an endpoint, stract into the determined units ofa given behavior analysis could be caught up in instance of successful contingency anal- an intellectual whirlpool of infinite ysis. In behavior analysis, the name for regressions ofever more massive contex- such successful working is "prediction tual analyses and would lose contact with and control." We know that a given event "successful working" as its truth crite- should be called a discriminative stim- rion. The argument can be made that the ulus, for example, because the effects of relative lack ofempirical work springing doing so are that we can deal more suc- from J. R. Kantor's extremely contex- cessfully with the behavior of interest. tualistic view of psychology shows what While prediction alone provides some can happen ifcontextualists let go ofcon- confirmation of the value of an analysis, trol as a primary goal ofscience. It is the only control "proves" that the units we successful operation ofcontextualism that have selected (the divisions we have made pragmatically requires that control be in context and behavior) are valid. One emphasized-it is not a postulate of the could ask, for example, how do we know philosophy. that this is the relevant stimulus for this Ifwe have explained why control must behavior? The answer is of the general be emphasized as a goal ofscience in be- form that when we change this stimulus havior analysis, why should prediction (and not that stimulus), we get a change be included at all as a goal? There are two in this behavior (and not that behavior). reasons. First, prediction is pragmatical- Without manipulation, the units we se- ly useful and thus naturally part of the lect could be completely mistaken. Be- truth criterion ofcontextualism. It would cause the explanatory model in behavior be arbitrary to exclude it. Second, control analysis does not and cannot specify the cannot be the only proximate outcome precise nature ofthese units a priori, this of successful science, because technical would present a grave problem. Without limitations often exist on our ability to manipulation, we could guess, for ex- manipulate events. Predictive relation- ample, that event X is a discriminative ships are not by their nature incomplete stimulus, but instead it could be part of when prediction is based upon variables event Y and not a separate event at all, that are in the domain in which the po- or it could covary with behavior due to tential for control exists because in prin- MENTALISM AND SCIENCE 179 ciple they can provide a complete ac- insufficient scope: a wide variety of in- count. They take us outside the behavioral stances are not encompassed by the "ex- system. Whether they in fact do provide planation." a complete account in a specific instance Behavior analysts also agree that at can be assessed as additional techniques times science can have explanation with- allow direct manipulation and as addi- out control. Skinner uses the term "in- tional knowledge creates networks of terpretation" to describe such cases. It mutually supportive observations. occurs when a system can be thoroughly Critics ofthe goals ofscience embraced described, but for technical reasons con- by behavior analysis have often misun- trol is not possible. Even then, behav- derstood the basis for these goals and the iorists insist that an adequate explana- integral role they play in behavior anal- tion or interpretation be based on events ysis. It is to such criticisms that we now that are of the same kind as those per- turn. mitting control. As we have emphasized above, this insistence is not an arbitrary THE BEHAVIOR-ANALYTIC VIEW component ofbehavior analysis. It is also OF EXPLANATION consistent with the strategy ofthe natural sciences under such circumstances. Mod- Wessells (198 1) has objected to the cri- em astronomy, for example, is based terion ofcontrol by pointing out that "be- fundamentally on the controlled obser- havioral control can be achieved in the vations ofthe physicist, and the observa- absence of explanation and vice versa" tions are of the same kind. Someday we (p. 161). Astronomy is a commonly used may even manipulate cosmic events. The example (Martin, 1978; Wessells, 1981) limitations that occur in astronomy are because we cannot test astronomical technical, not problems with the kinds of principles based on our ability to control events themselves. cosmic events. The literal meaning ofex- planation comes from the same root as The Mechanistic View ofExplanation the word "plane" -literally, explanation means to lay out flat before us. According This view of explanation can be con- to the behavior-analytic position, "ex- trasted with others that are popular in planation" ultimately refers to prediction psychology. To put the best face on it and control with adequate scope and pre- from the point of view ofbehavior anal- cision. Thus, seen from the standpoint of ysis, some forms of cognitivism are use- behavior analysis, Wessell's statement is fully considered as efforts to describe the at least partially incorrect. We cannot structure ofbehavior (Catania, 1973). A have control in a sophisticated manner description of the structure of behavior, without what behaviorists take to be "ex- however, is a description ofthe phenom- planation." With thorough and general ena to be explained. Knowing the phe- control must also come prediction, and nomena to be explained is vital, but it to a behaviorist, prediction and control must not be mistaken for a contextual of sufficient scope and precision is "ex- explanation of these very phenomena. planation" -it is how behaviorists "lay The distinction is critical because if be- out" behavioral systems. Control that is havioral events are orderly, they them- limited in scope and precision does not, selves can allow prediction ofbehavioral ofcourse, justify the term "explanation" phenomena, without ever going outside and in that sense Wessell's observation the particular behavioral system. This is seems obviously valid. The observation not adequate behavior analytic expla- that pointing a gun at people and saying nation even though in a more limited "Your money or your life" will often con- sense it clearly can be useful. No amount trol monetary exchange is hardly an ex- of description of behavioral events will planation ofcharitable giving. Such coer- directly provide for control ofbehavioral cion fails as an explanation for monetary events in the same individual. To think contributions in general because it has otherwise is to make the "structuralist 180 STEVEN C. HAYES & AARON J. BROWNSTEIN error" (Skinner, 1974). We cannot con- and the nature ofthe parts does not change trol the behavior of an individual with- when they are combined into systems. out considering events other than the be- The conditions that gave rise to this havior of that individual. structure or the ways we can manipulate Thus, the cognitivist agenda does not it are irrelevant to the description of the necessarily meet the goals of science ac- operation ofthe machine and predictions cording to behavior analysis. Cognitiv- based on this description. ists themselves have no quarrel with this Mechanism has a correspondence- point: "Most [cognitive] investigators based truth criterion (Pepper, 1942). A agree that their current theories are de- term, concept, or statement ofrelation is scriptive and that an explanatory theory, true to the extent that we can agree that which would accommodate the effects of it corresponds with events. This presents the environment, is a distal goal" (Wes- particular restrictions to mechanistic the- sells, 1982, p. 77). Just as it is unfair, ories. It would be improper to develop a however, to criticize behavior analysis description of a system based on close without considering its goals and under- contact with it and then to treat this as lying philosophy, so too it is unfair to an adequate theory ofthe system because criticize cognitive perspectives in the obviously any detailed and careful de- same manner. Cognitivists do not believe scription of events will be in close cor- that they are confusing description and respondence with these very same events. explanation. Nor do they believe that the Correspondence can be used to test the lack ofconcern for control is a weakness. adequacy oftheories or concepts only by They are correct in both instances, but applying them to new situations, that is, only in the context of their own under- by deductive prediction. The more ab- lying philosophical position. stract the relation between the original Most, though not all, forms of a cog- situation, the theory, and the new situ- nitive account of behavior seem to be ation, the better (e.g., see Ericsson & Si- based on the world view of mechanism mon, 1984, on this point). When we can (cf. Pepper, 1942). The fundamental ex- deductively predict with sufficient pre- planatory model ofmechanism is that of cision and scope, we have "explained" the machine. The type ofcognitive theo- the phenomena according to a mecha- rizing that is based on computer meta- nistic account. The emphasis on hypo- phors and uses computer simulations to thetico-deductive theorizing that exists test the adequacy and operating charac- in all mechanistic perspectives (e.g., most teristics of various theoretical models is forms of cognitivism and S-R theory) is obviously mechanistic in this sense. We not arbitrary; it is integral to the func- will limit our discussion of cognitivism tioning of the underlying world view. to this mechanistic variety. (Some forms Thus, for mechanists, description and ofbehaviorism are also mechanistic, but theoretical prediction form an adequate not radical behaviorism. A detailed dis- basis for explanation; in contrast, con- cussion ofthis is beyond the scope ofthis textualists emphasize prediction and paper, but see Hayes & Reese, in press.) control. These differences about ade- In accord with its explanatory model, a quate scientific explanation correspond mechanist shows no hesitation in ex- with differences about the goals ofscience plaining a behavioral system by speci- according to the underlying world view. fying the component parts ofits structure Neither can be said to be "right" or and the nature of its orderly operation; "wrong" because any evaluation we can just as a person examining a car would make about the adequacy of scientific readily explain its action by an appeal to goals must itself be in terms of some set its component parts (e.g., pistons, spark of underlying values or goals. From the plugs) and structural organization (e.g., point ofview ofcontextualism, all we can the spark plug wire is connected to the say is that these two views ofscience may spark plug). In mechanism, each part can be more or less useful for accomplishing be described independently ofthe others particular ends, or that they might have MENTALISM AND SCIENCE 181 this or that effect on the culture that sup- with the incompleteness of the account ports them. Most behaviorists probably mentalism engenders when measured in believe, of course, that a contextualistic terms ofthe goals of prediction and con- perspective has positive consequences for trol. Our main point is that it is impor- scientific subcultures and perhaps the tant not to confuse these two types of majority culture. Evaluating truth on the objections. Many of the behavior-ana- basis of consequences, however, is itself lytic concerns with mentalism are meta- a distinctly contextualistic line to take. theoretical objections driven by the goals Few behaviorists would argue that be- of science embraced by behavior analy- havior analysis is leading to better and sis, and these objections are by no means better hypothetico-deductive theories. unique to the topic of mentalism. We For a mechanist, this may show the in- show later that they apply to topics with- adequacy ofbehavior analysis. For a con- in behavior analysis itself. textualist, it shows no such thing. To summarize our main points so far: The Scientific Unacceptability ofLiteral Prediction and control are reasonable Dualism goals for science. Together, they allow behavior analysis to progress efficiently, Originally, psychology was the study providing an adequate basis for scientific ofthe soul. The Oxford English Diction- explanation according to a contextualis- ary (OED) defines "soul" as "the spiritual tic perspective. Unlike prediction, con- part of man." The OED defines "spirit" trol requires that we go outside of the as an "incorporeal or immaterial being" behavioral system itself. Sometimes even and as a "being or intelligence distinct when we do, control will be technically from anything physical." The word impossible. For the sake of a complete "physical" comes from a word for nature account, however, it is necessary that we (thus the science of physics) and is de- explain behavior in terms of events that fined as "ofor pertaining to the phenom- in principle allow both prediction and enal world of the senses; matter." Thus, control. This requirement is not due pri- if you take the words literally, "soul" or marily to a concern with technology. "spirit" are inherently dualistic terms be- Rather, it is necessary for the successful cause they oppose matter and nonmatter. working of the explanatory system used We might label this "literal dualism." in behavior analysis. Other explanatory Literal dualism is the beliefthat there are systems imply other kinds of scientific two different essences in the world-one goals, and alternative practices can be de- type exists in space and time, while the fended on those grounds. second type is nonspatiotemporal. This type ofliteral dualism was to some MENTALISM degree transferred to the concept of "mind" and thus to its study. The OED An arena in which to examine the im- defines "mind" as "the mental or psych- plications ofthe preceding analysis is the ical being or faculty." An elaborating def- study ofprivate events. Indeed, many of inition explains that "mind" is "the seat the objections to behavior analysis are of a person's consciousness, thoughts, usually cast in terms of private events volitions, and feelings; also, the incor- and their investigation. The dominant al- poreal subject of the psychical faculties, ternative to a behavioral analysis of pri- the spiritual side of a human being; the vate events is mentalism. Objections to soul as distinguished from the body .... mentalism can be placed in two basic cat- Mental being; opposed to matter." The egories: metaphysical objections and spiritual meaning of"mind" is also shown metatheoretical objections. The meta- by the fact that God has long been re- physical objections are directed at the ferred to as "mind," as for example in concept ofliteral dualism and are shared the quote "That eternal infinite mind, by virtually all scientific perspectives. The who made and governs all things" (Locke, metatheoretical objections have to do 1690). Thus, in lay terms, the essence of 182 STEVEN C. HAYES & AARON J. BROWNSTEIN

the mind is fundamentally different from placed literal dualism (Skinner, 1969, pp. the essence of the physical unvierse. 280-284). "Mind" explicitly makes contact with a It is very popular to use the word cultural tradition of literal dualism. "mind" to mean "brain." Television The objection to literal dualism is two- shows or magazine articles on the action fold. First, such a view is not scientifically of the human brain, for example, are al- tenable. This objection is shared by vir- most always said to be about "the human tually all scientists, including cognitiv- mind." Cognitive psychology has made ists. By definition, nonspatiotemporal no bones about the connection. A recent events cannot have form, mass, accel- authoritative text on cognitive psychol- eration, beginnings, or ends. How they ogy asserted that in cognitive psychology can be regarded as "events" in such a case "the contemporary view [is] that mental is problematic. Any event that can be processes are synonymous with brain discerned is being observed in the world processes" (Ellis & Hunt, 1983, p. 11). ofspace and time, and must in that sense Study of the brain and the nervous sys- be regarded as "physical." Literal dual- tem is, of course, worthwhile and rele- ism also raises the impossibly difficult vant to a behavioral analysis. But it is no issue of how nonspatiotemporal events less troublesome simply to substitute the can cause physical events to occur. word "brain" for "mind" and then to This argument is not about terms. If engage in precisely the same kinds of someone wanted to call all events "men- analyses as before. tal," there could be no objection, though Cognitivists use talk of the brain for it would distort our normal understand- two purposes. One is simply to claim that ing of the term. The concern is more di- "mental functioning is not a mysterious, rected at literal dualism: "What is lacking nonphysical event" (Ellis & Hunt, 1983, is the bold and exciting behavioristic hy- p. 7), that is, to emphasize the rejection pothesis that what one observes and talks of literal dualism. The second reason is about is always the 'real' and 'physical' more subtle. As we have described above, world (or at least the 'one' world)" (Skin- in the kind of hypothetico-deductive ner, 1945, p. 276, emphasis added). As theorizing naturally promoted by mech- the last phrase in this quote shows, rad- anism, there is no requirement that terms ical behaviorism is monistic, but not used in the analysis of events refer to physicalistic in the sense ofnaive realism other events outside the original domain or related perspectives. of interest. In order that there can be a The second major objection to literal proper division of labor among the sci- dualism is that it leaves a gap in the do- ences, however, scientists must take a main of science. Who is to predict and complete scientific account to the point control mental events and relate them to at which the determinants ofa given phe- behavioral events if they are not in the nomenon are themselves being analyzed purview of science? (Skinner, 1953, p. by other scientists. 258). This objection is a metatheoretical In psychology, contextualism naturally one, and it applies equally forcefully to does this because it leads to the identi- mentalistic theorizing that explicitly de- fication of events outside the behavioral nies dualism. system, as we have already discussed. These events are themselves to be ex- Mental Physiology plained by others. Behavior analysis, for For the above reasons, virtually all sci- example, does not attempt to understand entists avoid obvious forms of literal how operant chamber key lights work. It dualism in their scientific work. A variety is enough to known the conditions under oflinguistic practices, however, have the which key lights can function as stimuli. same metatheoretical problems as literal The study of electrical lights is turned dualism. One such practice is the creation over to another science. Conversely, of a pseudophysiological analysis to re- mechanism (though it need not do so) MENTALISM AND SCIENCE 183 can leave a hole in the fabric of science Mental Activity as Private Behavior because scientifically adequate analyses from this perspective can stay entirely Watson and "behavior." "Behavior" inside the original domain. Behavioral is commonly taken to refer to a certain events, or processes inferred from them, subset of organismic action. For exam- can be explained by other behavioral ple, it is quite typical to hear theorists events, or processes inferred from them. speak of "thoughts, feelings, and behav- Although philosophically permissible in ior" as if behavior can be easily distin- mechanism, this kind ofcircularity is in- guished from events called "thoughts" or tuitively unappealling, even to cognitiv- "feelings." ists. Confusion over this issue can in part By appearing to study brain processes, be attributed to Watson. Watson's (1925) however, cognitive psychologists can behaviorism had both methodological identify causes that appear to go outside and metaphysical components. His ofthe behavioral system itself. Cognitive methodological behaviorism essentially psychology can superficially present itself said that scientists must be behaviorists under the umbrella ofthe neurosciences. because science can only deal with the Ifwe actually understood how the struc- publicly observable. Thus, even though ture ofthe nervous system was influenced other kinds of human action may exist, by events, and exactly how any current we can only deal with behavior because structure produced behavior, we would only behavior is publicly observable. This indeed have one type of complete ac- position might be thought ofas implicitly count (in fact, one quite comfortable to dualistic because it recognizes that be- mechanists). But studying behavioral ac- havior is only a subset oforganismic ac- tivity called "mental" is no more inher- tivity, and encourages a study that is nec- ently a study of the brain than studying essarily incomplete since science can only any behavioral activity. In order to main- deal with that subset, rather than the en- tain a proper division of labor among tire set, due to rules of proper scientific sciences, the connection between brain methodology. and mental activity would have to be Watson also made a second, somewhat studied explicitly. It is clear that cogni- contradictory point. He seemed to say tivists do not actually intend to study the that even if we could solve the problem brain or its connection with mental ac- privacy presents to a scientific analysis, tivity: behavior is still all that could be studied because only behavior exists. Although Of course, brain activity can be studied physiolog- it is possible to read Watson to mean ically, but cognitive psychologists use a different simply that nonspatiotemporal events do approach. Since the brain activity ofinterest cannot not exist, his emphasis on the peripheral be directly observed (for example, we have no idea locus ofsuch phenomena as thinking (e.g., what happens in the brain when a person remem- bers a grandmother), we must infer the existence of Watson, 1920, 1925) can be and was tak- these processes and then describe the processes in en to mean that thoughts, feelings, and abstract language. (Ellis & Hunt, 1983, p. 7) other private events are not real in their own terms. This position can be termed In summary, mental activity can be "Watsonian metaphysical behavior- thought of as nonspatiotemporal activi- ism." ty, in which case it steps outside of sci- What is important to note in all ofthis ence altogether. It may be thought of as is that the use of the word "behavior" a brain activity, but then psychologists keeps changing. In Watsonian method- sometimes act as if we need not explain ological behaviorism, behavior is viewed how the structure of this activity itself as a subset of organismic action that is came to be. There is a third way to view publicly observable and is therefore sub- "mental activity," however-as behav- ject to a scientific analysis. In Watsonian ior.This is the view taken by radical be- metaphysical behaviorism, behavior is haviorism, to which we now turn. viewed as the totality of organismic ac- 184 STEVEN C. HAYES & AARON J. BROWNSTEIN

tion, but there is the implication that only of contextualism, the scientific value of events that are publicly observable (at an observation in behavior analysis is ul- least potentially) should be thought ofas timately determined by the degree to real. which it enables prediction and control. Radical behaviorism and "behavior." The essence of Skinner's (1945) criti- Radical behaviorism can be distin- cism of operationism was thus that pub- guished from these other types ofbehav- lic agreement provides no assurance of iorism in part by the view it takes in re- the proper sources of control over sci- gards to "behavior" and the nature of entific observation, nor of its pragmatic scientific observations (Skinner, 1945, value. Conversely, in principle, obser- 1963). As in Watsonian metaphysical be- vations of private events can be tightly haviorism, behavior is taken to be the set controlled by these events themselves, of all organismic action. The word "or- given the proper history, and can be high- ganismic" is important. Actions by sub- ly useful. In this sense, observations of organismic units (e.g., a single neuron fir- private events are no more or less sci- ing) are not usually considered to be the entific than public events based on their behavior oforganisms, but under certain privacy per se. In radical behaviorism, conditions they may if they can con- behavior can thus be defined as all ob- veniently be viewed as the integrated ac- servable organismic action, not all pub- tion of a whole organism. For example, licly observable organismic action (cf. the controlled heart rate of a person in a Heidbreder, 1933, and her discussion of biofeedback training program would the flaws ofclassical behaviorism). Skin- probably be thought of as the behavior ner (1974) has made the point quite clear- of an organism, while the beating of a ly: heart removed from the body clearly [Radical behaviorism] does not insist upon truth would not. by agreement and can therefore consider events tak- Unlike earlier forms of behaviorism, ing place in the private world within the skin. It however, radical behaviorism makes no does not call these events unobservable, and it does commitment to public observability per not dismiss them as subjective. (p. 16) se as the defining characteristic of sci- In this view, then, no objection can be entifically valid events (Skinner, 1945). made to talk of events such as thinking Rather, observations are scientifically or feeling. This talk is not trivialized by valid or invalid based on the contingen- insisting it is only the talk itself that is cies controlling these observations. Sci- scientifically legitimate (Skinner, 1945). ence is an enterprise that promotes the A specific instance ofthinking is viewed development of verbal statements of re- as a scientifically accessible event- a cov- lations between events based on verifi- ert behavior. We may eventually find able experience. Scientific verbal behav- ways of identifying specific covert be- ior thus should be under the control of haviors in others. For example, we cur- the subject matter of the science and the rently have ways of knowing when re- value of this verbal behavior is deter- ports of private speech are in fact mined through the impact it has on oth- occasioned by the specified private speech ers attempting to come under the control (Hayes, 1986). of that same subject matter. Science attempts to restrict sources of control over scientific observation Mental Causality as Behavior-Behavior through the scientific method. Its pur- Relation pose is to ensure that scientific obser- Thoughts as behavior. One might ask: vations are controlled primarily by events Why insist that thinking be regarded as in the relevant subject matter and not by covert behavior when the physical prop- states ofdeprivation, audience factors, or erties appear so different from overt be- similar sources ofcontrol over verbal be- havior? Why not call thinking "mental havior (Moore, 1981). Consistent with activity" or even a "brain process"? If the "successful working" truth criterion "behavior" is defined so broadly, doesn't MENTALISM AND SCIENCE 185 this make the concept of "behavior" but it is not a mere repetition of axioms. meaningless? Such questions have often It is an identification of possibly mis- been forcefully raised. For example: chievous contingencies over the behavior of psychologists, one that the history of The omission of [mental] states left [Skinner] with an inadequate vocabulary, which he then expanded psychology gives us every reason to take by moving some stimuli inside the organism ("pri- seriously. vate stimuli"), and by treating all other aspects of When a radical behaviorist is less than mental states as responses. Seeing became behavior, enthusiastic about an account of behav- and imagination became "seeing without the thing ior that predicts that someone will re- seen." But these are assertions, not demonstrated facts. They may serve as the axioms of a parsi- spond in a given way after thinking a monious behavioral system, and that is largely how particular thought, it should be an ab- Skinner used them. But they cannot then also be sence of enthusiasm resulting from the used as arguments against other systems, or against incompleteness ofthe account rather than behavioral systems with augmented axioms, such a The im- as the assumption that covert events are sufficiently from the reference to thought. different from overt ones to deserve separate treat- mediate question then becomes what are ment as a separate category ofevents. (Killeen, 1984, the determinants of that thought and p. 27) (even less obviously) what are the con- tingencies that lead to a relation between Skinner himself (1974) is quite clear a given instance of thinking and overt that the issue is not one ofparsimony per responding in this individual. se. Asking himself how we could decide In behavior analysis, the view that between behavior analysis and mental- thinking causes overt behavior distills ism, he replies: down to the view that one behavior can We cannot say that one is simpler than the other cause another. In these terms, when we ... [but] accessibility [for use in control] is another ask such questions as "What role does matter. No one has ever directly modified any of thinking play in the control of behav- the mental activities or traits .... for most practical purposes they are changed only through the envi- ior?", we are actually asking about the ronment .... A decision [between the two posi- nature of a behavior-behavior relation. tions] is perhaps more difficult if we simply want Behavior-behavior relations are very im- to predict behavior .... [Traits] are ... useless in portant in behavior analysis in a variety control but they permit us to predict one kind of ofareas, and they are as worthy of study behavior from another kind. (pp. 208-209) as is any behavior. No matter how dy- Behavior-behavior relations as incom- namically one behavioral event may be plete accounts. By referring to "mental intertwined with other behavioral events events" as behavior we do three things. within the same individual, however, for First, we eliminate consideration of a contextualist a behavior-behavior re- "mental events" that cannot be thought lation is a phenomenon to be explained ofas observable organismic activity, such by appealing to particular contextual ar- as purely hypothetical constructs. Sec- rangements (e.g., contingencies of rein- ond, we emphasize that it is the task of forcement) that might permit prediction psychology to predict and control these and control ofthe behavior-behavior re- events. And third, we focus on analyses lation itself. A behavior-behavior rela- that can accomplish these two goals. If tion cannot be a complete explanation of "mental events" are a separate category behavior, except to a mechanist, whose of events, then they can be used to ex- world view does not insist on control as plain behavioral events and perhaps need a necessary goal of science. not themselves be explained by behav- "We may object, first, to the predilec- ioral scientists. Mental events that can- tion for unfinished causal sequences" not be translated into behavioral obser- (Skinner, 1969, p. 240). Killeen (1984) vations are particularly prone to this has criticized this concern: problem because by definition they are Skinner notes that after we have explained a re- seemingly a separate category of events sponse in terms of mental states or activities of from behavior. The problem can be used feeling, we still need to explain the mental state. as an "argument against other systems," But there is nothing wrong with that. Experimental 186 STEVEN C. HAYES & AARON J. BROWNSTEIN analysis ofone ofthe links in a causal chain should involved are clearly in the behavioral do- not necessarily be faulted because it does not in- main. For example, a researcher may no- clude the previous ones; analysis must inevitably stop at some point short of the Ultimate Cause." tice that many good Monopoly players (pp. 27-28) are also good poker players. Few of us would be tempted, however, to claim that For a complete account, however, be- people play poker well because they play havior analysts must take analysis to the Monopoly well or vice versa. Both of point at which prediction and control are these actions are obviously in the behav- directly possible in principle. Behavior ioral domain. Events outside of this do- ofthe individual being studied can never main must be found to explain the first satisfy that criterion. behavior, and equally important, to ex- Thus, despite the fact that environ- plain the relation between the two be- ment and behavior are always involved haviors. That is, what actual events lead in a dynamic interrelation, it is not ar- to good Monopoly playing, good poker bitrary (Bandura, 1978, 1981) to insist playing, and their interrelation? This need that analysis proceed to the environmen- seems relatively clear when the two events tal level. An environmental cause can in are obviously from the same domain. Be- principle be used directly; given the tech- havior-behavior relations are seldom nical ability to manipulate it, effective mistaken for causal analyses adequate for action can be based on it. A rule pointing the purpose of prediction and control to a behavioral "cause" might help locate when they are clearly stated as behavior- causal environmental events and rela- behavior relations. tions, but it is also likely, if one is not Note, however, that when behavioral careful, to stop the search for causes that events are apparently in different do- could permit a complete account. Skin- mains, the mistake is readily made. Ifwe ner (1974) has said this explicitly: observed that good poker players feel self- It has been objected that we must stop somewhere confident (not "behave self-confident- in following a causal chain into the past and we ly"), have aggressive personalities (not may as well stop at a psychic level ... It is true that "behave aggressively"), and are intelli- we could trace human behavior not only to the physical conditions which [cause it] but also to the gent (not "show intelligent behavior"), causes of those conditions and the causes of those we might feel as though we have at least causes, almost ad infinitum, but there is no point a partial explanation of their skill even in going back beyond the point at which effective though this is in principle identical to the action can be taken. That point is not to be found obviously less satisfactory claim that good in the psyche. (p. 210) poker playing comes from good Monop- Thus, "the initiating action is taken by oly playing. To most people feelings, per- the environment" (Skinner, 1974, p. 73, sonalities, intelligence, and the like, ifnot emphasis added; see also Skinner, 1984). explicitly termed types or qualities ofbe- If we explain behavioral events in terms havior, often seem to be something else. ofevents that are in the domain ofother Such things are then easily mistaken for sciences, we move toward the compre- causes of behavior. hensive knowledge that we desire. But if Insisting that we call "mental events" we explain behavioral events in terms of by the name "behavior" should not be other behavioral events of the same in- done to diminish the interest in these so- dividual, then a significant gap is left to called mental events, any more than call- be filled. If we will not fill it, who will? ing a public behavior in a behavior-be- If not now, when? havior relation by the name "behavior" The effect of calling thinking "behav- should be done to diminish interest in ior." By calling all organismic activity public behavior. Behavioral scientists appropriate to psychological study "be- should call events "behavior" to keep havior," we naturally inoculate ourselves clear the fact that it is theirjob to explain to some degree against incomplete ac- such events and to avoid incomplete ac- counts. Acceptance of an incomplete counts based on these events. analysis is much less likely when all events A brief summary seems in order. The MENTALISM AND SCIENCE 187 purpose of this paper has been to place sons for them. Other goals or values make the behavior-analytic rejection ofmental alternative practices and beliefs justifia- causes into the larger context ofthe goals ble, and with these we can have no quar- of science embraced by behavior analy- rel, provided only that the researcher sis. Both critics and supporters ofbehav- honestly states what these goals are. For ior analysis seem at times to confuse the example, if a particular field of psychol- various reasons for the behavior-analytic ogy wishes to eschew control as the end concern with mentalism. Literal dualism product of science, it has the obligation is rejected because it is scientifically un- to make that clear to all, whether the per- tenable and because it leaves claimed son is a psychologist, a client seeking help, sources of control over behavior outside a congressman, or a taxpayer helping fund the scientific enterprise. This rejection of federal research. literal dualism, however, is common to We see no relative disadvantage for a virtually all scientific forms of psychol- radical behavioral approach when pre- ogy. The rejection of mentalism on the diction and control are being pursued. grounds of the requirements of scientific We do not mean to say, of course, that method, as done by early forms ofmeth- research generated by nonbehavioral odological behaviorism, is not recog- psychologists necessarily results in in- nized as legitimate by radical behavior- complete analyses as viewed from a con- ism. Neither is the attempt to define away textualistic perspective. See, for example, the independent existence of private Ornstein and Naus's (1978) demonstra- events, as in Watsonian metaphysical be- tion that recall was increased after the haviorism. What, then, is unique about manipulation of environmental events the radical behavioral rejection of men- controlling overt rehearsal. We must talism? guard against responding to the quality Our point has been that the primary of language rather than to the quality of radical behavioral objection to mental- the analysis. The issue is not theform of ism (other than literal dualism) is a meta- science but its function. It is the quality theoretical one. Nondualistic analyses ofthe analysis that is at issue rather than based on mental causality usually boil an attempt to urge that some simple down either to pseudophysiologizing, to physical correspondence exists between theorizing based on hypothetical con- private and public behaviors. structs, or to elevating disguised behav- ior-behavior relations to causal status. Examples ofthe Danger The concern with all ofthese maneuvers is that they interfere with the behavior- Behavioral analyses that implicate analytic agenda of predicting and con- "covert behavior," rather than "mental trolling behavior. As we have tried to events," are equally incomplete if they show, this agenda is not an arbitrary ele- fail to extend the analysis to environ- ment ofbehavior analysis. It is a required mental variables. Indeed, there are sev- element for the successful functioning of eral topics within behavior analysis that the perspective. Thus, mental causality share with mentalistic accounts some of is a form of theorizing rejected because the same potential metatheoretical prob- its pursuit threatens the successful op- lems of encouraging incomplete ac- eration of science as viewed from the counts. standpoint of behavior analysis. Self-rules. There can be little doubt that humans talk to themselves. Presumably, these verbal events can relate in reliable THE SLIPPERY SLOPE OF NON- ways to other behavior. To explain other MANIPULABLE CAUSES behavior simply in terms of "self-rules," however, is not adequate behavior-ana- We believe that much of the criticism lytic explanation. As in the cases dis- ofthe behavioral approach arises from a cussed earlier, the term "self-rules" does failure to recognize its goals and the rea- not obviously place these events in the 188 STEVEN C. HAYES & AARON J. BROWNSTEIN same category as "behavior." Thus, when image seen (in favor of the more "be- talking of a relation between a self-rule havioral" term "response-produced and behavior, it is not always obvious stimuli") it is troublesome. that it is a behavior-behavior relation that Responses obviously usually produce is at issue. Because it is, we must explain stimuli. For example, all operants, by both how external events gave rise to the definition, do so. The concept of re- private talk, and how the private talk sponse-produced stimuli, however, is came to control the behavior of interest. usually invoked when there is no possi- When this is done, "self-rules" may par- bility of discerning or manipulating the ticipate in an overall causal relation, but stimuli independently ofthe behavior it- they should not themselves be seen as self. The danger is that, because stimuli causes. are apparently in a different class than When behaviorists say they have lo- behavior, explanations based on "re- cated a "cause," it seems they should have sponse-produced stimuli" will be ac- identified a relation that can perform the cepted as complete even when we have functions required ofa scientific relation in principle no hope of manipulating or within a behavioral perspective, namely, even discerning these stimuli indepen- prediction and control. In a sense, causes dently. Explanations are not complete are not objective independent facts in the under these conditions. We move no world. All events in a thoroughgoing con- closer to prediction and control simply textualistic system are "events" only be- by replacing a behavior-behavior rela- cause construing them that way serves a tion with a "behavior-response-pro- purpose. A causal relation can only be duced stimuli-behavior" relation unless tested as a verbal construction, and only we have independent access to the stim- when this verbal statement of a relation uli. It is useless to "explain behavior by serves the purposes of science should it appealing to independent variables which be called a "cause." Perhaps when we have been inferred from the behavior thus have finished a complete analysis of a explained" (Skinner, 1969, p. 264). The behavior-behavior relation, we could radical behavioral objection to hypo- think of the first behavior as a kind of thetical constructs (when they are used "intermediate cause" (e.g., see Skinner's as other than merely a shorthand for be- reluctant agreement on this point, 1984), havior) has the same metatheoretical ba- but it is surely safer not to do so. Some sis as the primary objection to mental- other term is needed. "Controlling be- ism. In some cases, "response-produced havior" suggests itself because of Skin- stimuli" can have the status of a purely ner's use ofthe term to serve this function hypothetical construct. in his analysis of self-control. For ex- This objection does not necessarily ample, self-rules might be said to func- mean that we should stop using the term tion as controlling behaviors as parts of "response-produced stimuli." We should an overall causal relation. They are not distinguish, however, between three types themselves causes, at least not in the be- of situations. In the first, the "stimuli" havior-analytically acceptable sense ofthe referred to are clearly in the domain in word "cause." which control is possible in principle. An Response-produced stimuli. At times, ordinary operant is an example. A key- behaviorists have used the term "re- peck that produces food is producing sponse-produced stimuli" to explain be- food-related stimuli. That is presumably havior-behavior relations. The apparent why this behavior occurs in the first place. effects of private behaviors on other be- We could speak of the food as response- haviors may be explained on this basis. produced stimuli. In this case, we are us- For example, we may say that a person ing the term "response-produced stim- visualizes one's bedroom, and due to the uli" in a manner totally consistent with stimuli produced by this behavior, re- the goals of science as seen by behavior members where the car keys were left. analysts because we can manipulate these Even ifthe analysis avoids the idea ofan stimuli independently from the behavior MENTALISM AND SCIENCE 189

that produced them and discern their ef- behavior-environmental restriction- fects on behavior. behavior relation. In the second situation, the stimuli The examples of self-rules, response- cannot currently be controlled due to produced stimuli, and self-reinforcement technical limitations, but control is pos- show that behavior analysts are also sus- sible in principle. Analyzing back- ceptible to the tendency to dress up be- scratching in terms of sensory reinforce- havior-behavior relations in the cloak of ment maybe an example. Here, when we nonbehavioral events and then to forget claim that the back scratch is due to re- that they have done so. The cost of this sponse-produced stimuli we are engaging action is the same as the cost of mental- in interpretation, and in the future we istic talk or pseudophysiological talk- may or may not find the interpretation incomplete analyses are inappropriately to have provided a complete account. We accepted as complete and a resultant gap may ultimately find ways to block the in knowledge is produced. suggested sensory stimulation and to dis- cern the effects of this manipulation. CONCLUSION Finally, there are times when direct control is impossible in principle. A claim An embrace ofmentalism is not a sure that we have unconscious thoughts and road to an appreciation of the richness that these produce stimuli might be an ofprivate phenomena, and trivialization example. Here, we are using the term is not the necessary result of behavioral "response-produced stimuli" solely to translation. Behavior analysts should re- provide a consistent account, but at a ject mentalistic terms precisely in order considerable cost. We have disguised an to study the actual phenomena associ- analysis that cannot in principle meet all ated with them in a more thorough way the goals of science from a behavior-an- and in a way more satisfying to the goals alytic viewpoint in the cloak of termi- of science as viewed by behavior ana- nology that suggests these goals can be lysts. Allowing behavioral causes made met. seemingly less incomplete by calling them Behaviors as self-reinforcers. Even rad- "mental" ultimately tends to stop causal ical behaviorists sometimes claim that analysis before the point at which effec- one behavior can be maintained by the tive action is possible. We need to un- person involved providing other behav- derstand the actual phenomena pointed iors as "self-reinforcers" (e.g., Malott, in to by mentalistic terms, or terms such as press). The task ofbehavior analysis must self-rules or self-reinforcement for that be to explain both behaviors and their matter. The analytic discipline supplied relation. When the contexts giving rise to by the assumptions inherent in radical such behavior-behavior relations have behaviorism is needed most in exactly been manipulated experimentally, how- such difficult endeavors, not in order to ever (e.g., Hayes, Rosenfarb, Wulfert, pursue analytic discipline for its own sake, Korn, & Zettle, 1985), the apparent in- but in order to develop a more thor- fluence ofself-reinforcing behaviors have oughly adequate explanation of human always resolved into the effects of envi- behavior. ronmental events (Sohn & Lamal, 1982). REFERENCES An example of an apparently causal behavioral consequence is the Premack Bandura, A. (1978). The self system in reciprocal determinism. American Psychologist, 33, 344- principle. Superficially, this principle 358. seems to suggest that one behavior can Bandura, A. (1981). In search of pure unidirec- reinforce another in the same organism. tional determinants. Behavior Therapy, 12, 30- But if the external environment manip- 40. ulates the opportunity to engage in a par- Catania, A. C. (1973). The psychologies of struc- ture, function, and development. American Psy- ticular behavior following another, then chologist, 28, 434-443. this is not a simple behavior-behavior Dawkins, R. (1982). Theextendedphenotype. San relation. It is really best thought of as a Francisco: Freeman. 190 STEVEN C. HAYES & AARON J. BROWNSTEIN

Ellis, H. C., & Hunt, R. R. (1983). Fundamentals behaviorism, and radical behaviorism. Behav- ofhuman memory and cognition (3rd ed.). Du- iorism, 9, 55-77. buque, IO: W. C. Brown. Ornstein, P. A., & Naus, M. J. (1978). Rehearsal Ericsson, K. A., & Simon, H. A. (1984). Protocol processes in children's memory. In P. A. Ornstein analysis: Verbal reports as data. Cambridge, MA: (Ed.), Memory development in children (pp. 69- MIT Press. 99). Hillsdale, NJ: Erlbaum. Harris, M. (1979). Cultural materialism: The Pepper, S. C. (1942). World hypotheses. Berkeley: strugglefor a science ofculture. New York: Ran- University of Califomia Press. dom House. Reese, H. W. (May, 1984). Historical and philo- Hayes, S. C. (1986). The case of the silent dog- sophical analysis ofcausality. Paper presented at Verbal reports and the analysis ofrules: A review the meeting ofthe Association for Behavior Anal- ofEricsson and Simon's "Protocol analysis: Ver- ysis, Nashville. bal reports as data." Journal ofthe Experimental Skinner, B. F. (1945). The operational analysis of Analysis ofBehavior, 45, 351-363. psychological terms. Psychological Review, 52, Hayes, S. C., & Reese, H. W. (in press). A review 270-276. of S. C. Pepper's "World Hypotheses: A Study Skinner, B. F. (1953). Science and human behav- in Evidence." Journal ofthe Experimental Anal- ior. New York: Free Press. ysis ofBehavior. Skinner, B. F. (1963). Behaviorism at fifty. Sci- Hayes, S. C., Rosenfarb, S. C., Wulfert, E., Korn, ence, 134, 566-602. Z., & Zettle, R. D. (1985). Self-reinforcement Skinner, B. F. (1969). Contingencies ofreinforce- procedures: Another way to set social standards? ment: A theoretical analysis. New York: Apple- Journal ofApplied Behavior Analysis, 18, 201- ton-Century-Crofts. 214. Skinner, B. F. (1974). About behaviorism. New Heidbreder, E. (1933). Seven psychologies. New York: Knopf. York: Appleton-Century-Crofts. Skinner, B. F. (1984). Reply to Catania. The Be- Keat, R. (1972). A critical examination of B. F. havioral and Brain Sciences, 7, 718-719. Skinner's objections to mentalism. Behaviorism, Sohn, D., & Lamal, P. A. (1982). Self-reinforce- 1, 53-70. ment: Its reinforcing capability and its clinical Killeen, P. R. (1984). Emergent behaviorism. Be- utility. The Psychological Record, 32, 179-203. haviorism, 12, 25-39. Watson, J. B. (1920). Is thinking merely the action Locke, J. (1690). An essay concerning human un- of language mechanisms? British Journal ofPsy- derstanding. London: Awnsham, Churchill, & chology, 11, 87-104. Manship. Watson, J. B. (1925). Behaviorism. New York: Malott, R. W. (in press) The achievement ofeva- Norton. sive goals: Control by rules describing indirect- Wessells, M. G. (1981). A critique of Skinner's acting contingencies. In S. C. Hayes (Ed.), Rule- views on the explanatory inadequacy or cognitive governed behavior: Cognition, contingencies, and theories. Behaviorism, 9, 153-170. instructional control. New York: Plenum. Wessells, M. G. (1982). A critique of Skinner's Martin, M. (1978). Interpreting Skinner. Behav- views on the obstructive character of cognitive iorism, 6, 129-138. theories. Behaviorism, 10, 65-84. Moore, J. (1981). On mentalism, methodological