The Analysis of Verbal Behavior 2009, 25, 87–98

Relational Frame Theory: An Overview of the Controversy

Amy C. Gross and Eric J. Fox, Western Michigan University

Although Skinner’s Verbal Behavior (1957) was published over 50 years ago, behavior-analytic research on human language and cognition has been slow to develop. In recent years, a new behavioral approach to language known as relational frame theory (RFT) has generated considerable attention, research, and debate. The controversy surrounding RFT can be difficult to fully appreciate, partly because of the complexity of the theory itself and partly because the debate has spanned several years and several journals. The current paper aims to provide a concise overview of RFT and a summary of key points of debate and controversy. Key words: relational frame theory

The first comprehensive paper introducing view of verbal behavior in some respects relational frame theory (RFT) as a behavior- (Barnes-Holmes, Barnes-Holmes, & Culli- analytic approach to human language and nan, 2000), its proponents are also directly cognition was presented at the Association critical of key components of Skinner’s for Behavior Analysis annual convention in analysis. Given Skinner’s prominence and 1985 (Hayes & Brownstein). Since that time, importance in the development of behavioral RFT has inspired a great deal of research, psychology, it is not surprising that a theory discussion, and debate. It is not uncommon to challenging his view on a topic, particularly witness heated discussions about RFT during one as important as verbal behavior, might be both symposia and social gatherings at be- met with apprehension, suspicion, and even havioral research conferences, and a number contempt. of criticisms of the approach have been Second, if the RFT analysis is accurate, it published, usually in the context of reviewing has drastic implications for how we conduct the first book-length treatment of the topic a science of human behavior (Hayes & (Hayes, Barnes-Holmes, & Roche, 2001). Berens, 2004). The transformation of stimu- Whereas behavior analysts have grown lus functions seen in the literature on derived accustomed to harsh criticism from those stimulus relations indicates that stimuli can outside their discipline or worldview (e.g., acquire behavioral functions based solely on Chomsky, 1959), it is less common for a their participation in verbal relations with behavioral theory to generate so much other events. The research on stimulus equiv- intense debate within the field itself. It is alence has revealed that stimulus functions probably accurate to say that RFT has commonly transfer through members of equiv- become one of the most controversial, hotly alence classes (e.g., Augustson & Dougher, contested topics in modern behavior analysis. 1997; Dougher, Augustson, Markham, Green- Much of the controversy surrounding RFT way, & Wulfert, 1994; Dymond & Barnes, seems to stem from two primary sources. 1994; Hayes, Kohlenberg, & Hayes, 1991), and research on other derived stimulus rela- First, it is a treatment of human language that tions has revealed that the behavioral func- differs substantially from that offered by our tions of a stimulus can also be changed or field’s founding father, B. F. Skinner (1957). transformed based on its derived relation to Although RFT is an extension of Skinner’s other stimuli (Dougher, Hamilton, Fink, & Harrington, 2007; Dymond & Barnes, 1995; We thank Candice Jostad for her feedback on an Roche & Barnes, 1997; Roche, Barnes- earlier version of this manuscript. Eric Fox is now Holmes, Smeets, Barnes-Holmes, & McGeady, at the Saybrook Graduate School and Research 2000). For example, if an individual derives Center in San Francisco. an arbitrary relation of ‘‘A is greater than Address correspondence to Amy Gross, Depart- ment of Psychology, Western Michigan Universi- B’’ and B is then established as a con- ty, 1903 West Michigan Avenue, Kalamazoo, ditioned reinforcer, with no further training, Michigan 49008. A may begin to be more reinforcing for the 87 88 AMY C. GROSS and ERIC J. FOX individual (e.g., the individual will choose mentally similar to Skinner’s account, and is A over B or work harder to obtain A than distinct from most cognitive and linguistic B). Such changes in stimulus functions approaches to language, in that ‘‘it approach- mean that our ability to predict and influ- es verbal events as activities not products’’ ence the behavior of humans with verbal (Hayes, Fox, et al., 2001, p. 22). It is abilities will be greatly impaired if we rely fundamentally different from Skinner’s ac- solely on analyses of direct-acting contin- count in how it defines and accounts for gencies; we must also take into account the those verbal events and activities. individual’s relational or verbal behavior. Such analyses often require new experimen- Challenges to Skinner’s (1957) Account of tal procedures and can lead to interventions Verbal Behavior that seem foreign and perhaps unnerving to many behavior analysts. For example, the Informal conversations with behavior an- most prominent applied extension of RFT, alysts sometimes reveal a sense of compla- acceptance and commitment therapy (ACT; cency with regard to the topic of verbal Hayes, Strosahl, & Wilson, 1999), often behavior. Many seem to believe the field has makes use of mindfulness meditation and adequately addressed verbal behavior, thanks experiential exercises that can seem out of to Skinner’s 1957 book. Unfortunately, the place in behavioral psychology. impact of Skinner’s analysis on research and If nothing else, the RFT debate has application has been limited. Although re- brought important philosophical, conceptual, search based on Skinner’s verbal operants is and empirical issues to the forefront in our increasing (Sautter & LeBlanc, 2006), both field. Despite being the focus of much atten- the volume and scope of this research remain tion, however, it seems that RFT in general, underwhelming at best. Many researchers and the controversy it has evoked in parti- focus on teaching Skinner’s basic verbal cular, remain poorly understood by many in operants (primarily mands and tacts) to behavior analysis. This is due, in part, to the children with developmental disabilities complexity of the theory itself. Although (Sautter & LeBlanc). Moreover, Dymond, the core claim of RFT is relatively simple O’Hora, Whelan, and O’Donovan (2006) (deriving stimulus relations is a learned found that during the period of 1984 to 2004 operant), fully understanding RFT research ‘‘the majority of citations of Verbal Behavior and analyses can be daunting and requires were from nonempirical articles’’ (p. 81). familiarity with the theory’s overall approach, Advocates of RFT have argued that the concepts, and terms. Even seasoned behavior limited impact of Skinner’s (1957) analysis analysts must commit to learning some new of verbal behavior may be due to the manner technical terms to grasp RFT. An additional in which he defined verbal behavior and problem is that the critiques of RFT and the verbal stimulation. Skinner defined verbal responses to those critiques have been scat- behavior as behavior that is reinforced tered across several journals over several through the mediation of another organism years. It can be difficult to fully appreciate who is trained by a verbal community to the nature of the controversy surrounding mediate such reinforcement. RFT, along with the excellent points made Hayes, Blackledge, and Barnes-Holmes on both sides of the debate, without conduct- (2001) claim that this definition is too broad ing an extensive literature review. To help because many behaviors are socially mediat- remedy this situation, the present paper aims ed in this manner, and the definition does to present an overview of RFT, summarize not provide any way to distinguish verbal the primary criticisms of the theory found in behavior from virtually any other social the literature, and present the responses to behavior. Skinner noted that his definition those criticisms by proponents of RFT. included the behavior of the nonhuman animal in an experimental chamber, with a WHAT IS RFT? nonhuman animal and experimenter com- prising ‘‘a small but genuine verbal commu- RFT is a behavior-analytic account of nity’’ (p. 108). Such a broad definition raises human language and cognition. It is funda- the question of why a separate definition or RFT CONTROVERSY 89 treatment of verbal behavior is even neces- that a girl in the woods hears a ‘‘cuckoo’’ sound sary, and does not lead to any obvious and says, ‘‘clock.’’ Is this response a tact or an advances in the methods we might use to intraverbal? Using Skinner’s definitions, it study verbal behavior (Hayes, Blackledge, & depends not on the function of the stimulus Barnes-Holmes, 2001). for the girl but on whether or not it is the Relational frame theorists have argued that product of verbal behavior. To classify the girl’s Skinner’s (1957) definition of verbal behav- response in this case, the key variable is the ior is not functional because it includes the source of the stimulation, not its function. If the behavioral history of another organism (i.e., ‘‘cuckoo’’ sound were the product of her the listener) as a defining feature (Hayes, brother’s verbal behavior, for example, the Blackledge, & Barnes-Holmes, 2001). In all girl’s response would be an intraverbal. How- other areas of behavior analysis, individuals ever, if the ‘‘cuckoo’’ sound were an instinctual define behaviors by their function for the noise made by a bird (i.e., not the product of organism of interest, not by the behavioral verbal behavior), the girl’s response would be a history of another organism. By incorporat- tact. Classification of the girl’s behavior differs ing the behavior of another organism into his depending on the source of the controlling definition of verbal behavior, RFT theorists stimulus, despite the fact that the girl is claim that Skinner placed the researcher in responding in the same way to perceptually the peculiar position of needing to study the identical stimulation in the two scenarios. This history of another organism (the listener) in method of classification is akin to classifying order to classify the behavior of the organism stimuli according to their topographical prop- of interest (the speaker). This requirement, erties, and is decidedly not functional. according to some, can lead to both concep- tual and methodological confusion when one The Promise of Derived Stimulus Relations attempts to conduct research on verbal behavior. Other behavior analysts do not find The research on stimulus equivalence and Skinner’s definition of verbal behavior to be derived stimulus relations offers behavior troublesome in this regard (e.g., Leigland, analysts new insights into understanding 1997; Palmer, 2008). verbal behavior and rule-governed behavior In addition, RFT researchers and others (Hayes, Barnes-Holmes, & Roche, 2001). We have criticized Skinner’s (1957) definition of now know that most humans are capable of a verbal stimulus (Hayes, Blackledge, & deriving arbitrary relations among stimulus Barnes-Holmes, 2001; Hayes & Hayes, events without direct training or instruction to 1992). Again, the core objection is that do so. For example, a person who is taught Skinner’s definition is not functional. For that A 5 B and A 5 C, where the letters Skinner, a verbal stimulus is simply the represent novel stimuli, will likely derive that product of verbal behavior. This is a peculiar B 5 C. This relation between B and C is way to classify a stimulus in behavior arbitrary because the stimuli are not physical- analysis, because it relies on the source of ly similar or equivalent. It is also derived the stimulus rather than its function for the because it has not been directly trained. Many individual organism of interest (Hayes, relations other than equivalence, such as Blackledge, & Barnes-Holmes, 2001). Be- relations of comparison and opposition, can havior analysts define other stimuli (includ- also be derived in this manner (Dymond & ing reinforcers, punishers, and discriminative Barnes, 1995; Green, Stromer, & Mackay, stimuli) by their functions, not what pro- 1993; Roche & Barnes, 1996). Moreover, duced them. Some might see this as a minor individuals can generate very complex rela- quibble, but the meaning and coherence of tional networks with just a few directly trained Skinner’s taxonomy of verbal operants suffer relations (Steele & Hayes, 1991). Describing without a functional definition of a verbal the origin of derived relational responding is stimulus. To properly classify any response beyond the scope of this paper; however, as being an instance of one of Skinner’s others have discussed this issue (e.g., Carr, verbal operants, with the exception of the Wilkinson, Blackman, & McIlvane, 2000; mand, one must first identify the controlling Devany, Hayes, & Nelson, 1986; Hayes, Fox, stimulus as verbal or not. For example, imagine et al., 2001, p. 28). 90 AMY C. GROSS and ERIC J. FOX

Furthermore, the behavioral functions of a related potential measures, and neural acti- stimulus may be transformed based on its vation patterns that resemble those in seman- derived relations to other stimuli. For exam- tic processing (Barnes-Holmes et al., 2005). ple, if A is a conditioned reinforcer for an individual, and the individual subsequently Relational Framing: Deriving Stimulus derives that B is the opposite of A, then B is Relations as Learned Operant Behavior likely to acquire conditioned punishing functions. This effect has been demonstrated Derived stimulus relations present a chal- with many stimulus functions, including lenge to behavior analysts because the results conditioned reinforcing functions (Dymond are not expected from a strict conditioning & Barnes, 1995; Hayes, Brownstein, Devany, paradigm; this is why the relations are often Kohlenberg, & Shelby, 1987; Hayes, Koh- called derived or emergent. According to lenberg, & Hayes, 1991; Roche & Barnes, RFT, deriving stimulus relations is a special 1997), discriminative functions (Hayes et al., generalized form of relational responding. 1987), elicited conditioned emotional re- Most organisms, from insects to birds to sponses (Dougher et al., 1994), extinction mammals, can learn to respond to the rela- functions (Dougher et al.), and self-discrim- tive physical properties of stimuli, such as ination functions (Dymond & Barnes, 1994). selecting the brightest or longest of two Derived stimulus relations may provide a stimuli (a phenomenon traditionally known useful model for analyzing language and as transposition; see Reese, 1968, for a other complex human behavior. The arbitrary review). Relational responding of this sort nature of derived stimulus relations parallels has some interesting characteristics. The first the symbolism or referential quality of is that stimulus relations are bidirectional, language, wherein words and their referents and any relational response reflects this. If A typically share few formal properties (e.g., is related to B in some way, then B is the word fox looks nothing like an actual necessarily related to A in some way. When fox), yet individuals often respond to them as an organism selects a stimulus because it is though they are equivalent and share many longer than another is, for example, it is psychological functions (Sidman & Tailby, simultaneously not selecting the other stim- 1982). The phenomenon of deriving com- ulus because it is shorter. A relation between plex networks of relations after direct train- two stimuli in one direction (A to B) entails a ing on just a few relations may explain the relation in the other direction (B to A). RFT remarkable generativity of language (Barnes- refers to this feature as mutual entailment (in Holmes, Hayes, Dymond, & O’Hora, 2001). equivalence relations, this feature is often In fact, there is increasing empirical evidence referred to as symmetry). Another feature of linking derived stimulus relations to lan- relational responding is that stimulus rela- guage development. Indeed, verbally able tions can mutually combine to reveal new humans easily demonstrate derived stimulus relations. When arranging three objects relations, but verbally deficient humans and according to size, for instance, responding nonhumans have not demonstrated such to just two stimulus relations can reveal a relations convincingly or unequivocally third stimulus relation. One might place a (Barnes, McCullagh, & Keenan, 1990; De- watermelon the left of an apple because it is vany et al., 1986; Dugdale & Lowe, 2000; bigger, and a cherry to the right of the apple Hayes, 1989; Sidman & Tailby, 1982). In because it is smaller. Responding to just addition, researchers have shown that the these two relations (the watermelon is bigger ability to derive stimulus relations correlates than the apple and the cherry is smaller with cognitive and verbal skills (Barnes et than the apple) reveals a third relation: The al., 1990; Devany et al., 1986; O’Hora, watermelon is bigger than the cherry. This Pela´ez, & Barnes-Holmes, 2005; O’Hora relation is entailed when the other two et al., 2008). It emerges in infancy, but relations are combined, and can be deter- develops gradually and at about the same mined without even directly comparing the time as language skills (Lipkens, Hayes, & watermelon to the cherry. The manner in Hayes, 1993). Finally, derived relations which bidirectional stimulus relations can produce priming effects, differential event- mutually combine to reveal new relations is RFT CONTROVERSY 91 called combinatorial entailment (in equiva- frames of perspective-taking (Hayes, Barnes- lence relations, this feature is often referred Holmes, & Roche, 2001). Although RFT to as transitivity). researchers use the noun form of ‘‘relational RFT researchers argue that the early frame’’ for convenience, it is important to language training received by most humans remember that relational frames (and relation- allows arbitrary contextual and social cues, al networks) describe behaviors or repertoires, rather than just the formal properties of the not hypothetical or inferred mental structures related stimuli, to control relational respond- or knowledge constructs. Specifically, rela- ing. For example, in the presence of the tional frames refer to contextually controlled contextual cue of ‘‘X is bigger than Y,’’ one patterns of relational responding that individ- can learn to respond to X as though it uals learn through contingencies of reinforce- actually is bigger than Y (e.g., by pointing to ment established by their verbal and social the X when asked, ‘‘Which is bigger?’’). communities. Furthermore, because mutual entailment is a According to RFT, arbitrarily applicable feature of relational responding, reinforce- relational responding is the foundation of ment is likely in this context for responding human language and cognition; hence, the to Y as though it is smaller than X (e.g., by definition of verbal behavior is simply ‘‘the pointing to the Y when asked, ‘‘Which is action of framing events relationally’’ smaller?’’). Likewise, because combinatorial (Hayes, Fox, et al., 2001, p. 43). According- entailment is a feature of relational respond- ly, the definition of verbal stimuli is ‘‘stimuli ing, if one were also told, ‘‘Y is bigger than that have their effects because they partici- Z,’’ one could combine the two specified pate in relational frames’’ (Hayes, Fox, et al., relations and, thus, respond to X as though it 2001, p. 44). The history of the acting is bigger than Z. All of these relations among organism is the basis for bringing about X, Y, and Z are arbitrary because they are not verbal stimulus functions, not the history of based on their physical properties, and some another organism or listener. In the RFT of them are derived (rather than directly analysis, both the speaker and the listener are trained) due to the bidirectional and combi- engaging in verbal behavior. The speaker natorial features of the relational response does so by producing stimuli that are based applied to them. From the perspective of on relationally framed events, and the RFT, humans are capable of deriving stim- listener does so by responding based on ulus relations because we learn to arbitrarily these relationally framed events. apply relational responses to stimuli based on The RFT approach to studying verbal contextual cues to do so. behavior is leading to a growing body of Different types of relations have different empirical research, applications, and concep- patterns of responding associated with them, tual analyses. It serves as the theoretical basis and the term relational frame is used to for an increasingly popular form of psycho- describe a generic pattern of arbitrarily therapy known as acceptance and commit- applicable relational responding that has the ment therapy (Hayes et al., 1999). Further- features of mutual entailment, combinatorial more, its implications have been explored for entailment, and transformation of stimulus topics such as psychological development, functions. Contextual cues specify both the rule following, logical reasoning, persuasion relevant relations and the functions to be and rhetoric, problem solving, social behav- transformed in a relational frame. The meta- ior, prejudice and stigma, cognitive perspec- phor of a frame is used ‘‘to emphasize the tive taking, sexual attraction, and even reli- idea that this type of responding can involve gion and spirituality (see Hayes, Barnes- any stimulus event, even novel ones, just as a Holmes, & Roche, 2001). picture frame can contain any picture’’ (Hayes, Fox, et al., 2001, p. 34) and parallels CRITICISMS AND RESPONSES Skinner’s use of the term autoclitic frame (1957). A number of relational frames have Novelty of RFT been identified and examined, including frames of coordination, opposition, distinc- Post-Skinnerian. One criticism of RFT is tion, comparison, hierarchy, and deictic that it is not truly post-Skinnerian because of 92 AMY C. GROSS and ERIC J. FOX its reliance on many of the same fundamental applied to verbal behavior. The full explana- principles as Skinner’s (1957) analysis (Os- tion of each of these theories is beyond the borne, 2003). Osborne argued that RFT is scope of this paper; however, the following simply an extension of Skinner’s work on is a brief explanation of how RFT could be verbal behavior because RFT, like Skinner’s the combination of set theory and symbolic analysis, describes complex human behavior logic. RFT has a basis in and expands on using a small set of behavioral principles. equivalence relations, and Burgos noted that He noted that fundamental principles of be- this approach uses set theory language. In havior are unchanged with the RFT account regard to symbolic logic, Burgos stated that of human and nonhuman behavior, verbal or RFT authors point out that entailment is seen otherwise. Thus, ‘‘no paradigm shift lurks in in symbolic logic. Burgos declared that RFT’’ (p. 22). additional RFT concepts, beyond entailment, RFT researchers maintain the position that are equivalent to those of symbolic logic. RFT is a new theory. These researchers Burgos also explained that many mathemat- admittedly use the fundamental principles of ical and philosophical theories combine set behavior proposed by Skinner (1957) and theory and symbolic logic in order to have a credit him for them; however, they have powerful theory. Similar to these mathemat- moved beyond the analysis of verbal behav- ical and philosophical theories, Burgos con- ior that Skinner provided. Therefore, RFT is cluded that RFT is potentially powerful and post-Skinnerian because of its refined ap- of broad scope because it is a combination proach to human language and cognition of set theory and symbolic logic applied to (Hayes, Barnes-Holmes, & Roche, 2001, verbal behavior. 2003). Advocates of RFT rejected Malott’s (2003) New behavioral principle or theory. Al- attempt to explain verbal behavior based on though Malott (2003) regarded relational a behavioral chain. Malott’s analysis requires responding as an important aspect of human the use of a vocal repertoire, and Hayes et al. language, he was hesitant to call this a new (2003) cited several studies in which rela- behavioral principle. Rather, he argued that tional responding developed in the absence fundamental behavioral processes might ex- of vocal repertoires. In addition, RFT re- plain relational responding. Malott suggested that, through appropriate stimulus control, searchers pointed out the proposal of several reinforcement, and generalization, a complex other molecular mediational accounts of lan- behavioral chain might account for relational guage and cognition, but noted that they framing. Although he did not allege that he lack empirical support. Therefore, it is un- had identified the correct fundamental princi- likely that a behavioral chain or other mo- ple, he did argue for a molecular explanation lecular mediational events are sufficient to of human language, rather than the molar explain the complex process that takes place description that RFT researchers provided. with language and cognition (Hayes et al., Similarly, Salzinger (2003) reported ap- 2003). prehension about moving away from for- Hayes et al. (2003) argued against both the merly established behavioral principles. In rule-governed account of Salzinger (2003) fact, he stated that rule-governed behavior and the symbolic logic account of Burgos might explain relational responding rather (2003) by contending that there are no ade- than RFT explaining rule-governed behavior. quate functional explanations of either of Salzinger maintained that individuals form these phenomena in behavior analysis. Thus, relations based on rules. Accordingly, he attempting to explain RFT via logic or rules concluded that relational frames might is not very helpful until we have a functional, simply be a form of rule-governed behav- technical account of logic and rules indepen- ior rather than an act of framing events dent of RFT. Rather than rules or logic relationally. explaining RFT, the philosophy of RFT uses Burgos (2003) stated that RFT is not a new basic operant principles to account for theory because of its basis in other theories. derived relational responding in a manner Specifically, he argued that RFT is a com- that offers a functional account of both rules bination of set theory and symbolic logic and logic. RFT CONTROVERSY 93

Beyond stimulus equivalence. One of the relations are more general than the terms most common criticisms about the novelty of adopted by Sidman and other equivalence RFT is that it is no different from Sidman’s researchers. RFT introduced these new terms and others’ accounts of stimulus equivalence. to provide a language to talk about all types In particular, some claim that the new techni- of stimulus relations (e.g., bigger than, cal terms introduced by RFT to describe the before–after, opposites, darker than). Sid- features of derived relational responding (i.e., man’s terms, taken from mathematical set mutual entailment, combinatorial entailment, theory and applied to equivalence relations, and transformation of stimulus function) are do not work for types of stimulus relations redundant with the existing terms used by other than equivalence. Because RFT re- researchers who study stimulus equivalence searchers examine many derived stimulus (i.e., reflexivity, symmetry, transitivity, and relations in addition to equivalence, terms transfer of stimulus function). Thus, critics with broader scope were warranted. claim that RFT does not add new information Fourth, as suggested above, RFT research to the existing literature on stimulus equiv- and application focus on a wider range of alence (Fox, 2006; McIlvane, 2003). stimulus relations than equivalence. Stimuli The RFT account is different from Sid- can be related to one another in multiple man’s equivalence account in four important ways and along multiple dimensions, and ways. First, although Sidman provided one of presumably each of these types of relations the earliest behavioral accounts of stimulus can be brought under arbitrary contextual equivalence, his approach was, and is, pri- control. Although much of the excitement marily a descriptive one. In fact, he noted, about stimulus equivalence in the behavioral ‘‘My own theorizing has been directed not community was due to its apparent connec- so much at an explanation of equivalence tion to word meaning and vocabulary build- relations but rather, at the formulation of a ing, it is difficult to account for more com- descriptive system—a consistent, coherent, plex verbal units, such as sentences, using and parsimonious way of defining and only equivalence relations. By incorporating talking about the observed phenomena’’ relations other than equivalence into their (Sidman, 1994, p. 536). A precise, coherent analyses, RFT researchers have been able to description of empirical phenomena is im- provide cogent accounts of complex lan- portant, but it does not satisfy the need for a guage-related phenomena such as sentences, functional, behavioral explanation. Sidman’s rules, spirituality, morality, and more (see account, then, is a description of the behav- Hayes, Barnes-Holmes, & Roche, 2001). In ioral phenomenon known as stimulus equiv- addition, the complex patterns of responding alence, whereas RFT is a behavioral expla- seen when training multiple types of stimulus nation for how that phenomenon (and other relations (e.g., Steele & Hayes, 1991) and the phenomena) might come about. accompanying transformation of stimulus Second, RFT is a scientifically more functions are readily predicted by RFT, but conservative way of accounting for equiva- are not easily accounted for by Sidman’s lence because it does not require any new behavioral principles at the level of process (1994) theory of stimulus equivalence. (RFT proponents do propose a new behav- ioral principle at the level of outcome, but Clarity of Concepts that is simply due to the unusual effects seen from the transformation of stimulus func- Several authors have criticized RFT for tions; Hayes, Fox, et al., 2001). Sidman, on unclear or incomplete explanations of con- the other hand, suggests that equivalence is cepts. Malott went as far as postulating that probably a basic stimulus function and ‘‘the complexity and difficulty of under- that his mathematically based account in- standing their analysis may be so great that volves ‘‘new—previously unknown—behav- their analysis will also fail to persuade the ioral variables or theoretical principles’’ faithless hordes’’ (2003, p. 17). Another (1994, p. 537). author stated that as the topics of study in Third, the terms used in RFT to describe RFT broaden, the analysis becomes less equivalence and other derived stimulus precise (Tonneau, 2004). 94 AMY C. GROSS and ERIC J. FOX

Specific misunderstandings have occurred which is a well-established process in behav- regarding the description of higher order ior analysis. They also claimed that other operants, the development of transfer of functionally oriented behavior analysts had stimulus function, relational frames as a no difficulty accepting relational operants class of behavior, and the new behavioral without the development of a new behavioral process or principle. Galizio (2003a) consid- principle and are, thus, uncertain as to why ered calling a relational frame a higher order Palmer had concerns with this approach. operant to be unclear. He contended that RFT provides no definition of a higher order Variety of Research Areas operant; therefore, it does not make sense to label a relational frame a higher order Palmer (2004a) criticized RFT for having operant. Galizio (2003b) also questioned limited scope. He stated that RFT researchers how the transformation of stimulus function typically study only teenagers or adults; begins. He stated that the analysis provided moreover, when children have been included by RFT leaves out the initial development of in studies, they show only the emergence of the transformation of stimulus functions, thus equivalence. McIlvane (2003) made a similar lending itself to an unclear description of the claim in that RFT researchers primarily study phenomenon. Palmer (2004b) criticized RFT university populations. In doing so, the re- for the uncertain description of whether searchers cannot answer the important ques- relational frames are a class of behavior or tion of how relational responding develops. if relational frames are part of the history that Palmer’s (2004a) criticism of limited brings about a class of behavior. When RFT scope also applies to RFT researchers not proponents describe relational frames as both accounting for mediating events. He stated an outcome and a process, the distinction is that the dependent variables measured in vague (Palmer, 2004b). Finally, Palmer RFT research tap into only a small amount (2004a) proposed that, if RFT researchers of the behavior in which the participants claim that relational behavior is a higher engage. Palmer alleged that RFT research order operant and that the transformation does not consider the importance of the of stimulus functions occurs without direct covert behavior involved in the emergence training, they ought to present a new of relational frames. Thus, RFT researchers principle. He asserted that a ‘‘statement of are ignoring an important line of study, the principle’’ (Palmer, 2004a, p. 231) is neces- control of relational responding by covert sary to fully explain relational operants. behavior. RFT advocates maintain that the concepts Although relational frame theorists have involved in RFT are clear, yet they acknowl- admitted that additional research is needed, edge that some concepts are difficult to they have defended the position that RFT has grasp due to their abstract nature (Hayes, a broad scope. RFT researchers have con- Barnes-Holmes, & Roche, 2001). Therefore, ducted several studies with various popula- they have responded to critics by providing tions. They have shown relational responding additional descriptions and examples. Al- in children, teenagers, and adults on several though the details are beyond the scope of occasions (e.g., Berens & Hayes, 2007; this paper, further information regarding the Stewart, Barnes-Holmes, Roche, & Smeets, concepts of higher order operants, transfer of 2002). Thus far, studies have revealed rela- stimulus functions, and relational frames as a tional phenomena including, but not limited class of behavior is available in other articles to, opposition and comparison in typically (see Barnes-Holmes & Hayes, 2003; Hayes developing preschoolers, mutual exclusion in & Barnes-Holmes, 2004). human infants, and analogical reasoning in In response to Palmer’s (2004a) criticism children (Hayes & Barnes-Holmes, 2004). of the lack of a new principle, proponents of RFT proponents agree with Palmer RFT stated that a new behavioral principle (2004a) that individuals likely engage in is not necessary to address the concept of covert relational framing behavior, but they relational operants. Hayes and Barnes- do not consider such behavior to be a Holmes (2004) noted that relational respond- mediating event. Thus, RFT researchers have ing is a result of differential reinforcement, not limited their research scope but rather RFT CONTROVERSY 95 have studied the correlation between covert research, particularly when verbal behavior and overt responding using several procedur- is an important independent or dependent al variations such as the relational evaluation variable. Moreover, the wide range of topics procedure, respondent-type pairing proce- being addressed and methods being used in dures, developmental studies, talk-aloud pro- RFT may make the field of behavior analy- cedures, response latencies, semantic prim- sis somewhat more appealing to those who ing and implicit association test procedures, long ago deemed behaviorism ‘‘dead’’ and and recording event-related potentials. irrelevant. Hopefully, the intense debate and Through this experimentation, RFT research- controversy inspired by RFT will serve to ers are developing an empirical account for move the field forward and contribute to an covert forms of relational framing, seeking a increased behavioral understanding of the more precise explanation than the nontech- complexities and importance of human nical account of mediating events proposed language. by Palmer (Hayes & Barnes-Holmes, 2004). Furthermore, in response to criticism related REFERENCES to mediating events, Hayes and Barnes- Holmes noted that mediating events are not Augustson, E. M., & Dougher, M. J. (1997). the central issue of study in human language The transfer of avoidance evoking func- and cognition. The differing opinions on the tions through stimulus equivalence class- importance of mediating events is likely due es. Journal of Behavior Therapy and to irresolvable philosophical differences be- Experimental Psychiatry, 28, 181–191. tween individuals concerned with mediating Barnes, D., McCullagh, P. D., & Keenan, M. events and those less concerned with these (1990). Equivalence class formation in events. non-hearing impaired children and hear- ing impaired children. The Analysis of CONCLUSION Verbal Behavior, 8, 19–30. Barnes-Holmes, D., Barnes-Holmes, Y., & RFT researchers are studying an area that Cullinan, V. (2000). Relational frame most investigators in behavior analysis have theory and Skinner’s Verbal Behavior:A not previously studied. Since the beginning possible synthesis. The Behavior Analyst, of this endeavor, other researchers and the- 23, 69–84. orists have been quick to denounce the work. Barnes-Holmes, D., & Hayes, S. C. (2003). A They have criticized the novelty of RFT, the reply to Galizio’s ‘‘The Abstracted Oper- clarity of the concepts, and the scope of the ant: A review of Relational Frame research. Proponents of RFT have responded Theory: A post-Skinnerian Account of to each of these criticisms, although one’s Human Language and Cognition.’’ The view of these debates may be guided more by Behavior Analyst, 26, 305–310. underlying philosophical assumptions than Barnes-Holmes, D., Hayes, S. C., Dymond, logic or empirical data. S., & O’Hora, D. (2001). Multiple stim- Despite the criticisms, even critics agree ulus relations and the transformation of that RFT researchers are tackling an impor- stimulus functions. In S. C. Hayes, D. tant subject area (Malott, 2003: McIlvane, Barnes-Holmes, & B. Roche (Eds.), Re- 2003; Spradlin, 2003). Behavior analysts lational frame theory: A post-Skinnerian have long overlooked the area of human account of human language and cognition language and cognition, but proponents of (pp. 51–72). New York: Kluwer Academ- RFT are ambitiously addressing these diffi- ic/Plenum. cult and important topics (McIlvane). Al- Barnes-Holmes, D., Staunton, C., Whelan, though critics may not agree with all of the R., Barnes-Holmes, Y., Commins, S., concepts presented by RFT, they have Walsh, D., et al. (2005). Derived stimulus acknowledged that RFT is valuable in the relations, semantic priming, and event- study of language and cognition (Galizio, related potentials: Testing a behavioral 2003a; Malott; McIlvane). Applied and theory of semantic networks. Journal of experimental behavior analysts are encour- the Experimental Analysis of Behavior, aged to consider RFT when conducting 84, 417–430. 96 AMY C. GROSS and ERIC J. FOX

Berens, N. M., & Hayes, S. C. (2007). Dymond, S., O’Hora, D., Whelan, & O’Do- Arbitrary applicable comparative rela- novan, A. (2006). Citation analysis of tions: Experimental evidence for a rela- Skinner’s Verbal Behavior: 1984–2004. tional operant. Journal of Applied Behav- The Behavior Analyst, 29, 75–88. ior Analysis, 40, 45–71. Fox, E. J. (2006, June 16). How is RFT dif- Burgos, J. E. (2003). Laudable goals, inter- ferent from stimulus equivalence? Message esting experiments, unintelligible theoriz- posted to http://www.contextualpsychology. ing: A critical review of Relational Frame org/how_is_rft_different_from_stimulus_ Theory, edited by Hayes, S. C., Barnes- equivalence Holmes, D., and Roche, B. Behavior and Galizio, M. (2003a). The abstracted operant: Philosophy, 31, 19–45. A review of Relational Frame Theory: A Carr, D., Wilkinson, K. M., Blackman, D., & Post-Skinnerian Account of Human Lan- McIlvane, M. J. (2000). Equivalence guage and Cognition, edited by S. C. classes in individuals with minimal verbal Hayes, D. Barnes-Holmes, and B. Roche. repertoires. Journal of the Experimental The Behavior Analyst, 26, 159–169. Analaysis of Behavior, 74, 101–114. Galizio, M. (2003b). Relational frames: Chomsky, N. (1959). A review of B. F. Where do they come from? A comment Skinner’s Verbal Behavior. Language, 35, on Barnes-Holmes and Hayes. The Be- 26–58. havior Analyst, 27, 107–112. Devany, J. M., Hayes, S. C., & Nelson, R. O. Green, G., Stromer, R., & Mackay, H. A. (1986). Equivalence class formation in (1993). Relational learning in stimulus language-able and language-disabled chil- sequences. The Psychological Record, 43, dren. Journal of the Experimental Analy- 599–615. sis of Behavior, 46, 243–257. Hayes, S. C. (1989). Nonhumans have not yet Dougher, M. J., Augustson, E. M., Markham, shown stimulus equivalence. Journal of M. R., Greenway, D. E., & Wulfert, E. the Experimental Analysis of Behavior, (1994). The transfer of respondent elicit- 51, 385–392. ing and extinction functions through Hayes, S. C., & Barnes-Holmes, D. (2004). stimulus equivalence classes. Journal of Relational operants: Processes and impli- the Experimental Analysis of Behavior, cations: A response to Palmer’s review of 62, 331–351. Relational Frame Theory. Journal of the Dougher, M. J., Hamilton, D. A., Fink, B. C., Experimental Analysis of Behavior, 82, & Harrington, J. (2007). Transformation 213–224. of the discriminative and eliciting func- Hayes, S., C, Barnes-Holmes, D, & Roche, tions of generalized relational stimuli. B. (Eds.). (2001). Relational frame theo- Journal of the Experimental Analysis of ry: A post-Skinnerian account of human Behavior, 88, 179–197. language and cognition. New York: Klu- Dugdale, N., & Lowe, C. F. (2000). Testing wer Academic/Plenum. for symmetry in the conditional discrim- Hayes, S. C., Barnes-Holmes, D., & Roche, inations of language-trained chimpanzees. B. (2003). Behavior analysis, relational Journal of the Experimental Analysis of frame theory, and the challenge of human Behavior, 73, 5–22. language and cognition: A reply to the Dymond, S., & Barnes, D. (1994). A transfer commentaries on Relational Frame Theo- of self-discrimination response functions ry: A Post-Skinnerian Account of Human through equivalence relations. Journal of Language and Cognition. The Analysis of the Experimental Analysis of Behavior, Verbal Behavior, 19, 39–54. 62, 251–267. Hayes, S. C., & Berens, N. M. (2004). Why Dymond, S., & Barnes, D. (1995). A relational frame theory alters the relation- transformation of self-discrimination re- ship between basic and applied behavioral sponse functions in accordance with the psychology. International Journal of Psy- arbitrarily applicable relations of same- chology and Psychological Therapy, 4, ness, more than, and less than. Journal of 341–353. the Experimental Analysis of Behavior, Hayes, S. C., Blackledge, J. T., & Barnes- 64, 163–184. Holmes, D. (2001). Language and cogni- RFT CONTROVERSY 97

tion: Constructing an alternative approach McIlvane, W. J. (2003). A stimulus in need within the behavioral tradition. In S. C. of a response: A review of Relational Hayes, D. Barnes-Holmes, & B. Roche Frame Theory: A Post-Skinnerian Ac- (Eds.), Relational frame theory: A post- count of Human Language and Cognition. Skinnerian account of human language The Analysis of Verbal Behavior, 19, 29– and cognition (pp. 3–20). New York: 37. Kluwer Academic/Plenum. O’Hora, D., Pela´ez, M., & Barnes-Holmes, Hayes, S. C., & Brownstein, A. J. (1985, D. (2005). Derived relational responding May). Verbal behavior, equivalence class- and performance on verbal subtests of the es, and rules: New definitions, data, and WAIS-III. The Psychological Record, 55, directions. Invited address presented at 155–175. the annual meeting of the Association for O’Hora, D., Pela´ez, M., Barnes-Holmes, D., Behavior Analysis, Columbus, OH. Rae, G., Robinson, K., & Chaudhary, T. Hayes, S. C., Brownstein, A. J., Devany, J. (2008). Temporal relations and intelli- M., Kohlenberg, B. S., & Shelby, J. gence: Correlating relational performance (1987). Stimulus equivalence and the with performance on the WAIS-III. The symbolic control of behavior. Mexican Psychological Record, 58, 569–584. Journal of Behavior Analysis, 13, 361– Osborne, J. G. (2003). A review of Relational 374. Frame Theory: A Post-Skinnerian Account Hayes, S. C., Fox, E., Gifford, E. V., Wilson, of Human Language and Cognition. The K. G., Barnes-Holmes, D., & Healy, O. Analysis of Verbal Behavior, 19, 19–27. (2001). Derived relational responding as Palmer, D. C. (2004a). Data in search of a learned behavior. In S. C. Hayes, D. principle: A review of Relational Frame Barnes-Holmes, & B. Roche (Eds.), Re- Theory: A Post-Skinnerian Account of lational frame theory: A post-Skinnerian Human Language and Cognition. Journal account of language and cognition (pp. of the Experimental Analysis of Behavior, 21–50). New York: Kluwer Academic/ 81, 189–204. Plenum. Palmer, D. C. (2004b). Generic response Hayes, S. C., & Hayes, L. J. (1992). Verbal classes and relational frame theory: Re- relations and the evolution of behavior sponse to Hayes and Barnes-Holmes. analysis. American Psychologist, 47, 1383– Journal of the Experimental Analysis of 1395. Behavior, 82, 225–234. Hayes, S. C., Kohlenberg, B. K., & Hayes, L. Palmer, D. C. (2008). On Skinner’s definition J. (1991). The transfer of specific and of verbal behavior. International Journal general consequential functions through of Psychology and Psychological Thera- simple and conditional equivalence class- py, 8, 295–307. es. Journal of the Experimental Analysis Reese, H. W. (1968). The perception of of Behavior, 56, 119–137. stimulus relations: Discrimination learn- Hayes, S. C., Strosahl, K. D., & Wilson, K. ing and transposition. New York: Aca- G. (1999). Acceptance and commitment demic. therapy: An experiential approach to Roche, B., & Barnes, D. (1996). Arbitrarily behavior change. New York: Guilford. applicable relational responding and sex- Leigland, S. (1997). Is a new definition of ual categorization: A critical test of the verbal behavior necessary in light of derived difference relation. The Psycho- derived relational responding? The Behav- logical Record, 46, 451–475. ior Analyst, 20, 3–9. Roche, B., & Barnes, D. (1997). A transfor- Lipkens, G., Hayes, S. C., & Hayes, L. J. mation of respondently conditioned func- (1993). Longitudinal study of derived tions in accordance with arbitrarily appli- stimulus relations in an infant. Journal cable relations. Journal of the Experi- of Experimental Child Psychology, 56, mental Analysis of Behavior, 67, 275–301. 201–239. Roche, B., Barnes-Holmes, D., Smeets, P. Malott, R. W. (2003). Behavior analysis and M., Barnes-Holmes, Y., & McGeady, S. linguistic productivity. The Analysis of (2000). Contextual control over the de- Verbal Behavior, 19, 11–18. rived transformation of discriminative and 98 AMY C. GROSS and ERIC J. FOX

sexual arousal functions. The Psycholog- Skinner, B. F. (1957). Verbal behavior. New ical Record, 50, 267–291. York: Appleton-Century-Crofts. Salzinger, K. (2003). On the verbal behavior Spradlin, J. E. (2003). Alternative theories of of Relational Frame Theory: A Post- the origin of derived stimulus relations. Skinnerian Account of Human Language The Analysis of Verbal Behavior, 19, 3–6. and Cognition. The Analysis of Verbal Steele, D. L., & Hayes, S. C. (1991). Behavior, 19, 7–9. Stimulus equivalence and arbitrarily ap- Sautter, R. A., & LeBlanc, L. A. (2006). plicable relational responding. Journal of Empirical applications of Skinner’s anal- the Experimental Analysis of Behavior, ysis of verbal behavior with humans. The 56, 519–555. Analysis of Verbal Behavior, 22, 35–48. Stewart, I., Barnes-Holmes, D., Roche, B., & Sidman, M. (1994). Stimulus equivalence: A Smeets, P. M. (2002). Stimulus equiva- research story. Boston: Authors Cooper- lence and non-arbitrary relations. The ative. Psychological Record, 52, 77–88. Sidman, M., & Tailby, W. (1982). Condi- Tonneau, F. (2004). [Review of the book tional discrimination versus matching to Relational frame theory: A post-Skinner- sample: An expansion of the testing ian account of human language and paradigm. Journal of the Experimental cognition]. British Journal of Psychology, Analysis of Behavior, 37, 5–22. 95, 265–268.