Chapter 3: A contemporary functional-analytic account of perspective-taking
‘You never really understand a person until you consider things from his point of view--until you climb inside of his skin and walk around in it’.
Excerpt from To Kill a Mockingbird by Harper Lee
Introduction [A]
The ability to take the perspective of others and thus anticipate their beliefs and desiresis central to our social lives (Epley, Morewedge & Keysar, 2004). Three decades of psychological research have shownthat this ability starts in early infancy anddevelops throughout childhood and beyond (Baron-Cohen, 1994; Gopnik, & Slaughter, 1991). In recent years, a novelfunctional analytic accountprovided by Relational Frame Theory (RFT) has started to provide new insights into this ability with applications in a number of important applied domains. The key aim of this chapter is to describe the empirical research underlying thisRFT-based-approach, and to link it with applied work described in other chapters in the current text. It will start by describing sometraditional accounts of perspective taking. It will then introduce the theoretical basis for the RFT account. Finally it will present a brief overview of key RFT-based research to date.
Mainstreamapproaches to perspective-taking [A]
Selman [B]
Aprominentearly approach to perspective-takingwas provided byRobert Selmanwho describeda five-stage model of the development of this ability. According to Selman (1980; Yeates & Selman, 1989), children gradually develop the ability to discriminate their own perspective from those of others and to see the relationships between these potentially discrepant points of view. His research involved presenting children with ‘interpersonal dilemmas’ and then askingthema range of related questions (e.g., ‘Does X know how Y feels about situation Z?’).Responses to the latter led him to construct a five stage model of the development of ‘role-taking skills’, thefirst stage of which begins to emerge at roughly three years andthe final stage of which emerges in adolescence.
Selman’s work in the area of interpersonal development led to increased interest in how perspective-taking could contribute to social development. However, this account was descriptive only; it did not attempt to explain how children could take the perspective of another. At the same time that Selman was describing his model, however, another approach to perspective taking that did purport to explain this ability was also beginning to emerge. This was Theory of Mind (ToM) which has now become the dominant paradigm to conceptualize perspective-taking within mainstream psychology.
Theory of Mind[B]
The ToM approach suggests that perspective-taking is based on an ability to mentally represent the mind (i.e., including beliefs, desires, intentions, emotions, etc.)of another. This concept initially appeared in Premack and Woodruff (1978) whoinvestigated whether chimpanzees’ had a ‘theory of mind’ that would allow them to understand human goals. Numerous related comparative studies have followed,most of which suggest that chimpanzees lack this understanding (Penn, & Povinelli, 2007; Tomasello & Hermann, 2010).Some studies suggest some social understanding in chimpanzees’ that begins with the observation of others’ behavior, similar to how this develops for humans (e.g.,Okamoto, Tanaka, & Tomonaga, 2004).); however, none have identified the development in chimpanzees or other non-humans of more complex ‘theory of mind’ abilities such as understanding of false beliefs (see e.g., Call & Tomasello, 2008, for a recent review).
Developmental Research [C]
After emerging as an issue in the comparative domain, the phenomenon of ToM subsequently became a topic of intense interest in the human developmental domain, where its appearance in early life was investigated. Baron-Cohen and colleagues have suggested that there are five levels in the development of the attribution of information or mental states to the self and others that range from simple visual perspective-taking to predicting actions on the basis of false belief (Howlin, Baron-Cohen, & Hadwin, 1999). At Level 1 (simple visual perspective-taking) the person acts consist with the principlethat different people can have different views of the same situation. AtLevel 2 (complex visual perspective-taking) they realise that people can see things differently. Level 3(seeing leads to knowing) is based on the principle that people only know things that they have seen (Taylor, 1988). At Level 4 the person knows that actions can be predicted on the basis of true belief (Howlin, et al., 1999). Finally, at Level 5, the person knows that actions can be predicted on the basis of false belief and that it is possible that previous beliefs (held by oneself or others) could have been false.
An extensive quantity of research has been conducted into ToM abilities in children, a review of some of which is provided in Chapter 5 of the current volume. For instance, one consistent finding is that successful false belief performance tends to emerge at around age four (e.g., Perner, Leekam, and Wimmer, 1987). In addition to investigation into levels of ToM ability, there has also been theoretical speculation on and empirical investigation into possible precursors to this ability.Baron-Cohen (1991) hasargued that understanding of attention, suggested to emerge between 7 and 9 months, is an important precursor to ToM while Meltzoff (2002) has focused on other suggested precursors including understanding of other’s intentions and imitation (see Chapter 4of the current volume for a reviewof behavior analytic theory and research on precursors to perspective-taking).
Clinical Research [C]
Closely related to research into the development of ToM abilities in typically developing children is investigation of ToM deficits in children with autistic spectrum disorder (ASD). For example, Baron-Cohen, Leslie and Frith (1985) compared children with ASD with mental age controlled groups of typically developing and non-ASD developmentally delayed children and found perspective-taking deficits in the ASD group alone. On the basis of accumulating evidence such as this it has been suggested that children with ASD do not have a ToM and thus have particular difficulties with tasks requiring them to understand others’ beliefs and that this is a key feature of ASD (see Chapter 6 for deficits in perspective taking in groups including children with ASD).
More recent research has indicated that a number of other clinical and sub-clinical conditions also appear to be characterised by deficits in perspective-taking. These include most prominently schizophrenia (Corcoran, Mercer, & Frith, 1995) but also conditions related to it such as schizotypy (Chapman, Chapman, & Kwapil, 1995) and social anhedonia (Villatte, Monestes, McHugh, Freixa i Baque & Loas, 2008) (see Chapter X for elaboration on the link between clinical disorders and perspective taking deficits).
Training Theory of Mind [C]
In addition to assessment of ToM abilities, there has also been a limited amount of research on trainingToM. For example, Clements, Rustin, and McCallum (2000) gavetypically developing children between 34 and 60 months old feedback on a particular ToM task(‘unexpected transfer’). One group got a detailed explanation regarding their responses while a secondwere simply told whether they were correct or not. Findings showed that only the explanation condition increased children’s understanding of false belief. Fisher & Happé (2005) tested and trained ToM in ten children with ASD. Individual training sessions were administered with corrective feedback on all training trials. The participants were tested before and after training and at a two-month follow-up and showed significant improvements on ToM tasks compared with controls.
Kloo and Perner (2008) reviewed the literature on training ToM and concluded that multiple training sessions are required in order for training to be effectiveand that elaborated as opposed to brief feedback is superior. However, Knoll and Charman (2000) have questioned whether, despite their apparent success,training studies allow generalised ToM understanding. They provided children with false belief training that involved discussing the ‘unexpected transfer’ task and reported that though participantsshowed improvement on this task itself, they showed none onan alternative false belief test. Begeer, Gevers, Clifford, Verhoeve, Kat, Hoddenback & Boer (2010) used a randomised controlled design to test the effectiveness of a 16 week ToM treatment in 8-13 year old children with ASD. Results showed that compared with controls, the treated children improved in conceptual ToM skills but not in elementary understanding, empathic skills or parent reported social behavior.
A Contextual Behavioral Approach to Perspective Taking [A]
Theoretical and empirical interest in perspective-taking, the substantial majority of which has been conducted under the rubric of ToM, has been considerable in mainstream psychology and thus an increasing body of knowledge as to sequential development of particular skills, manifestation of deficits in particular populations and training and remediation of skills under certain circumstances has begun to accumulate. However, from a contextual behavioral perspective, there are fundamental issues that this mainstream work does not address and that are of key importance with regard to the utility of the research generated.For instance, questions remain over the basic and functional processes that comprise perspective-taking and the most focused means of establishing these. Therelative success in ToM training studies in accelerating the acquisition of ToM understanding suggeststhat thedevelopment of perspective-taking is ‘experience-driven’, but the details of this critical experience, or learning history, remain unclear. As a result training may not be targeting core response patterns, meaning less than optimal efficiency as regards required training time, for instance, and lack of generalization to other tasks or everyday contexts,as indeed has been reported.
One of the fundamental assumptions of a contextual behavioral approach to psychology is the importance of achieving both prediction and influence in the explanation of behavior.The latter in turn requires the specification of manipulable processes as a sine qua non and this affords contextual behavioral psychologists an advantage at the level of intervention. This approach has had notable success in certain domains in which it has been able to identify key causal variables and thus enabled relevant change. Up until relatively recently, it has not achieved the same level of success in domains involving more complex human behavior, such as perspective–taking. However, as explained previously in this text, this has now begun to change, based on the emergence of a contextual behavioral conceptualization of human language and complex behavior,namely, Relational Frame Theory (Hayes, Barnes-Holmes & Roche, 2001).
Relational Frame Theory [B]
Relational Frame Theory (RFT) is a functional analytic approach that accounts for the development of language and higher cognition in terms of learned generalized patterns of relational responding referred to as relational frames. The prototypical example of such relational responding is the relation of co-ordination between words and their referents (e.g., [actual apple] = ‘apple’) that children begin to learn around eighteen months. Continued exposure to the socio-verbal environment produces increasingly complex patterns of generalized relational responding including more extensive examples of the frame of co-ordination as well as additional non-co-ordinate frames including distinction (e.g., ‘Boys are different from girls’), opposition (‘Day is opposite to night’), comparison (‘This car is bigger than that’) and others (Hayes et al., 2001). According to RFT, all examples of this phenomenon possess the following three characteristics: mutual entailment (the fundamental bi-directionality of relational responding; for example,if A is B then B is A) combinatorial entailment (the combination of already known relations to generate novel relations; if A > B and B > C then A > C and C < A) and transformation of stimulus functions (the transformation ofpsychologically relevant functions of a stimulus in accordance with the underlying relation in a given context; if A > C and C is desirable then A is more desirable) (see also Chapter 1, current volume).
Deictic Relational Framing [C]
According to RFT, perspective-taking skills are based on a particular pattern of relational framing referred to as deictic framing. Deictic relational frames specify a relation in terms of the perspective of the speaker (McHugh, et al., 2004). The three frames that appear to be most important in this regard are those of I and You, Here and There and Now and Then. Unlike other frames, these do not appear to have formal or non-arbitrary counterparts and cannot be traced to formal dimensions in the environment. Instead, it is the relationship between the individual and other events that serves as the constant variable upon which these frames are based.Responding to, and asking, many questions contained within our common verbal interactions with others (e.g. “What am I doing now?” or “What are you doing there?”) appear to be critical in establishing them. Each time questions such as these are asked or answered, the physical environment is different. What remains constant across these and similar questions, however, are the relational properties of I versus You, Hereversus Thereand Now versus Then. Furthermore, according to RFT, these properties themselves are abstracted through learning to talk about one's own perspective in relation to the perspective of others. I, for instance, is always from this perspective here, but not from the perspective of another person there. Abstraction of an individual'sperspective on the world, and that of others, requires a combination of a sufficiently well developed relational repertoire and an extensive history of exemplars that take advantage of that repertoire.
According to RFT, the three perspective-taking frames described above can generate a range of relational networks, including: I-Here-Now; You-Here-Now; I-Here-Then; You-Here-Then; I-There-Now; You-There-Now; I-There-Then; and You-There-Then. Many phrases common to our daily discourse are derived from these eight relational networks. Consider, for example, the phrases; “I am here now, but you were here then”; “You were there then, but I am here now”; and “You and I are both here now, but I was here then”. Of course, when used in actual dialogue, these phrases would often include or substitute words co-ordinated with particular individuals, places and times. For illustrative purposes, consider the following example: “It is four o’clock and I am at work [Here and Now], but Fiona [You] is still at home” [There and Now]. What appears to makes perspective-taking frames particularly complex and useful is that they cannot be defined in terms of particular words. That is, words such as, “I”, “you”, “here”, “there”, “now”, and “then” (used to describe the perspective of the self and others) are merely examples of the relational cues that control the perspective-taking frames, and a range of other words and contextual features may serve the same function. As is the case for all relational frames, what is important is the generalised relational activity, not the topography of particular cues.
According to RFT, the establishment of deictic framing will likely be critical to the development of perspective-taking as traditionally defined. Hence, theorists would argue that there should be some correlation between competence in deictic framing and more traditional perspective-taking tasks. Furthermore, itwould argue that targeting the deictic frames directly should lead to improved performances on the latter.
Preliminary Empirical Support [C]
Thus, in contrast with mainstream approaches to perspective-taking such as Theory of Mind, the contextual behavioral RFT approach specifies the processes that give rise to perspective-taking in terms of the effects of manipulable environmental variables (e.g., use of contextual cues such as I and You), thus constituting a novel and potentially powerful new approach to this ability. Empirical support for this account of perspective-taking has been accumulating over the last number of years. The earliest published study in this area (McHugh, Barnes-Holmes, O’Hora, & Barnes-Holmes, 2004) developed a protocol that targeted the three perspective-taking frames of I-You, Here-There, and Now-Then, in conjunction with three levels of relational complexity, referred to as: simple relations; reversed relations; and double reversed relations. To illustrate the distinction between these levels of complexity explicit examples are provided below.
Consider a simple I-YOU perspective-taking task in which participants were presented with the following question: “If I (Experimenter) have a pen and YOU (Participant) have a cup: What do I have? What do YOU have?” Responding correctly to this trial (i.e., “You/Experimenter have a pen and I/Participant have a cup”) requires that the participants respond in accordance with the I-YOU deictic relational frame, under the contextual control of the if-then relational frame. In effect, the if-then frame determines the functions (i.e., the pen and cup) that become attached to the I and YOU related events in the deictic frame. The perspective-taking protocol also included simple HERE-THERE perspective-taking trials. In a simple HERE-THERE trial the participants may have been instructed as follows: “I am standing here at the yellow door, and you are standing there at the brown door. Where are you standing? Where am I standing?” A correct response on this trial requires that the participants respond in accordance with the I-YOU and HERE-THERE frames, under the contextual control of If-Then. The third simple relational task involved simple NOW-THEN relations. These trial-types differed from the other simple relations in that they did not involve responding to I and YOU simultaneously. Consider the following example: “Yesterday I was shopping, today I am washing the car. What was I doing then? What am I doing now?” Once again, a correct response requires that the participants respond in accordance with the I-YOU and NOW-THEN frames, under the contextual control of If-Then.