Relational frame theory (RFT; Hayes, Barnes-Holmes, & Roche, 2001) is a contemporary behavior-analytic account of human language and cognition. It facilitates the analysis and understanding of complex behavioral phenomena that have traditionally been the purview of cognitive psychologists, such as empathy, self-concept, intelligence, or creative behavior (Cassidy, Roche, & Hayes, 2011; Dymond & Roche, 2013; McHugh, Barnes-Holmes, & Barnes-Holmes, 2004a; McHugh, & Stewart, 2012; O’Toole, Barnes-Holmes, Murphy, O’Connor, & Barnes-Holmes, 2009). According to RFT, humans are able to respond relationally to objects or events when the relation is not defined by the physical properties of the objects but rather by contextual controlled cues (for a more complete account, see Hayes et al., 2001). For example, if a child learns that “A occurs before B,” then he is able to say that “B occurs after A” without any further training. This response is controlled by the contextual cues before and after, but not by any physical relation. Particular kinds of relational responding are called relational frames.

From the RFT point of view there are several such frames, including coordination (identity, sameness, or similarity—i.e., A is equal to B); opposition (A is the opposite of B); distinction (responding to one event in terms of its differences with another event); comparison (e.g., A is worse/better than B); hierarchy (e.g., A is an attribute of B); temporality (e.g., A occurs before B); spatiality (e.g., A is below B); conditionality and causality (cause–effect relationships established between events, the form “if . . .then . . .”) and deictic relations (specifying a relationship in terms of the perspective of the speaker—e.g., “I–you”, “ here–there,” and “now–then”).

RFT provides a theoretical framework for the understanding of many facets of complex human behavior. Research shows that performance on relational responding tasks correlates with normal language and cognitive development and that deficits in the former have been found in developmentally delayed populations (Hayes et al, 2001; O’Hora, Pelaez, & Barnes-Holmes, 2005). In the 14 years since the first RFT book (Hayes et al., 2001), a growing body of research has indicated the importance of relational framing in development as well as the possibility of training it when deficient (see, e.g., Dunne, Foody, Barnes-Holmes, Barnes-Holmes, & Murphy, 2014; Luciano, Gómez-Becerra, & Rodríguez-Valverde, 2007; Rehfeldt, & Barnes-Holmes, 2009; Walsh, Horgan, Jay, Dymond, & Whelan, 2014). Therefore, empirical intervention guided by RFT may generate new procedures for establishing relational skills aimed at overcoming language deficits in applied populations. In line with this suggestion, Rehfeldt and Barnes-Holmes (2009) developed a book-length account of a number of RFT training protocols for learners with developmental disabilities that target basic and complex cognitive skills. Furthermore, empirical studies have demonstrated the utility of considering RFT when designing and developing therapeutic interventions, such as acceptance and commitment therapy (ACT; for full details, see Hayes, Strosahl, & Wilson, 1999; Luciano, Rodríguez-Valverde, & Gutiérrez, 2004). In short, ACT is a clinical behavioral approach to psychotherapy that coheres with the RFT analysis of human language and cognition (Hayes et al., 2001).

According to RFT, deictic relational frames have been able to explain cognitive concepts such as theory of mind (ToM; concept initially appeared in Premack & Woodruff, 1978) and perspective taking (for a cognitive and structural-developmental approach, see Selman, 1980). From an RFT point of view, perspective taking involves a complex pattern of relational responding in accordance with I–You, Here–There, and Now–Then “deictic” cues, which are manipulable environmental variables. Such deictic cues are learned from a history of multiple exemplars of asking and responding to questions such as, “What am I doing now?”, “What did you hear there?”, or “Where were you singing then?” Each time these questions are answered, the physical properties of the environment will be different. However, the relational properties of I–you, here–there, and now–then remains constant across all exemplars (Hayes et al., 2001). That is to say, these frames that specify a relation in terms of perspective of the speaker do not appear to have formal (i.e., nonarbitrary) counterparts and therefore cannot be traced to formal dimensions in the environment (McHugh, Stewart, & Hooper, 2012).

In the experimental research conducted so far, deictic relations have been divided into three levels of complexity: simple, reversed, and double reversed (McHugh, Barnes-Holmes, & Barnes-Holmes, 2004b). In a simple relational response, none of the elements are reversed—for example, “I (experimenter) am sitting here in a blue chair and you (subject) are sitting there in a black chair. Where are you sitting?” In a reversed relational response, one of the relations is reversed—for example, “If I were you and you were me, where would you be sitting?” And a correct response reflects this relational reversal (i.e., the experimenter is sitting in a black chair and the subject is sitting in a blue chair). In a double reversed relational response, two relations are reversed simultaneously (e.g., “If I were you and you were me and if here were there and there were here, where would you be sitting?”), and a correct response would appear to require more complex derived relational activity (i.e., the experimenter would be sitting on the blue chair and the subject would be sitting on the black chair). This experimental approach allows researchers to conceptualize cognitive phenomena and influence complex human behavior in populations with deficits in perspective taking, such as those diagnosed on the autistic spectrum. Most of the research in this area has traditionally been the domain of cognitive psychology, under the rubric of ToM (Baron-Cohen, 2001). Although the ToM model was a descriptive account in the first instance, it was subsequently adapted for use as a teaching tool to train ToM skills to children (for an intervention guide that takes into consideration the levels of ToM, see Howlin, Baron-Cohen, & Hadwin, 1999). However, from a contextual behavioral perspective, ToM has been characterized as a description rather than an explanation of the psychological processes involved in perspective-taking (Barnes-Holmes, McHugh, & Barnes-Holmes, 2004b).

A behavioral perspective on the issue of perspective taking was first discussed by Hayes (1984). Several years later, the first RFT experimental study on perspective taking in terms of deictic relational responding was conducted by Barnes-Holmes (2001). In this study, a testing and training protocol was developed for establishing the three deictic relations across the three levels of relational complexity in young children. Since then, further empirical support has been provided on the deictic framing approach. Despite its youth, the deictic framing approach has generated a large and growing contribution to the study of domains involving more complex human behavior, such as the self, perspective taking, and empathy (Barnes-Holmes, Barnes-Holmes, Roche, & Smeets, 2001; McHugh, & Stewart, 2012). Barnes-Holmes, Foody, Barnes-Holmes, and McHugh (2013) contributed a book chapter to a text on advances in relational frame theory (see Dymond & Barnes, 2013) that reviewed the research on deictic relations and perspective taking since the first RFT book was published. The current study involves the first bibliographical review of deictic relations. The review includes articles that cited search terms related to deictic relations from the initial publication by Barnes-Holmes in 2001 to 2015. The present bibliographical review, therefore, provides a summary of the progress and limitations in the deictic relations literature with a view to promoting the development of future research in the area.

Method

Database Searches

The search terms deictic relations, perspective taking, and relational frames were individually entered into the ISI Web of Knowledge (Web of Science), PsycARTICLES, ProQuest Psychology Journals, Scopus, Dialnet, and PsycINFO databases. Searches were conducted for articles that included at least one of these key words.

The quantity of literature on perspective taking necessitated the adoption of strict inclusion and exclusion criteria. We decided to apply the term relational frame theory as a mandatory requirement; that is, those articles that did not include a relational frame theory approach were excluded. In this way, several studies that included perspective taking from a cognitive point of view or deixis from the semantic and pragmatic approach were not categorized.

An upper date limit of 2015 was employed, and the default lower date limit was 2001. Only journal articles were included in the final data set; that is, dissertation, books, or book chapters were excluded.

Article Type and Classification

Once the articles were identified, they were independently categorized by the present authors. Consistent with the approach adopted by Dixon, Small, and Rosales (2007) and Dymond, May, Munnelly and Hoon (2010), the articles were categorized according to whether they were empirical or not (empirical/nonempirical studies) and the population recruited.

Empirical articles reported original data involving the direct manipulation of at least one independent variable or measurement of at least one dependent variable. To further identify the content addressed by empirical articles, we classified them as either manipulation designs or non-manipulation designs. Manipulation designs systematically manipulated variables to change a participant’s behavior. Non-manipulation articles reported original data of at least one dependent variable but did not systematically manipulate variables. Nonempirical articles did not involve manipulation of any independent variables or measurement of any dependent variables and reported no data.

The population was determined based on the demographic information provided in each study. We identified the following categories: (i) the type and (ii) age of the samples. Sample types were classified as either typically developing if the participant did not report a clinical diagnosis or atypically developing if the participant had a clinical diagnosis. Sample ages were classified as early childhood if the participants were below 8 years (young children), as late childhood if the participants were between 8 and 17 years (older children), and adulthood if the participants were 18 years or older. The sample types and age produced six mutually exclusive categories: typically developing adults, atypically developing adults, typically developing young children, atypically developing young children, typically developing older children, and atypically developing older children. If the article included a mixed sample (e.g., adults and children with typical development), then they were included in both categories.

Analyses of Interrater Agreement

Interrater agreement for article assignment to the different categories was calculated by the authors. Specifically, the number of articles assigned to each category was divided by the total number of articles and the result multiplied by 100. In all cases, interrater agreement was higher than 90%. In the case that there was a nonagreement, the authors discussed which category was appropriate and allocated accordingly.

Results

At first, a total of 301 articles were identified with the initial search terms. However, 267 articles were excluded following the search strategy, thus remaining a total of 34 articles in the final data set. Figure 1 shows the increasing trend of published articles from empirical and nonempirical categories between 2001 and 2015. Although the trend of nonempirical articles has remained stable, empirical studies have been increasing. The greatest increase in empirical articles occurred between 2004 and 2007 with the early research of Louise McHugh, Yvonne Barnes-Holmes, and Dermot Barnes-Holmes.

Fig. 1
figure 1

Cumulative number of empirical and nonempirical articles per year between 2001 and 2015 that reported at least one of the search terms

Full size image

Of the included articles, 26 articles (76.47%) were assigned to the empirical category and eight (23.53%) to the nonempirical category. Of the 26 empirical articles, 14 (53.85%) of studies published between 2001 and 2015 were empirical articles with non-manipulative designs and 14 (53.85%) with manipulative designs. It should be noted that Table 1 shows more empirical articles with non-manipulative designs than with manipulative designs because some articles contributed to more than one population subcategory.

Table 1 The number of empirical articles categorized as typical or atypical adults and children and overall percentages
Full size table

Analysis of the populations studied in the 26 empirical articles, Table 1 shows that the majority of research has been conducted with typically developing populations (92.31%). In particular, 14 (53.85%) involved typically developing children and 10 (38.46%) involved typically developing adults. In terms of the type of design, the more frequent empirical articles were typically developing adults in non-manipulation designs (8/21) and typically developing children in manipulation designs (9/15).

Tables 2 and 3 present a brief summary of each of the empirical and non-empirical articles, respectively. In Table 2, some articles were included in more than one category because they have a mixed sample (e.g., adults, young children, and older children with typical development). In this case, they were divided depending on the population used. For example, if one article had two experiments carried out with typically developing adults and typically developing children, respectively, then each experiment was placed into a different category.

Table 2 Summary of empirical articles distributed by category and population
Full size table
Table 3 Summary of nonempirical articles
Full size table

An analysis of the aims of the articles (see Tables 2 and 3) shows that the RFT approach to the conceptualization of mental-state attribution through deictic relations has been the main thematic focus. Of the 34 articles included in the final data set, 25 (75.76%) mentioned the term theory of mind (ToM) in the text and connected their aims with the mainstream approach (Barnes-Holmes et al., 2001; Barnes-Holmes et al. 2004a; Barnes-Holmes et al. 2004b; DeBernardis et al., 2014; Gilroy et al., 2015; Heagle & Rehfeldt, 2006; Jackson et al., 2014; Janssen et al., 2014; Lovett & Rehfeldt, 2014; McHugh et al., 2004a, 2004b, 2006; McHugh et al. 2007a; McHugh et al. 2007b; Montoya & Molina, 2015; Mori, & Cigala, 2015; O’Neill & Weil, 2014; Rehfeldt et al., 2007; Rendón et al., 2012; Vilardaga et al., 2012; Villatte et al., 2008, 2010a, 2010b; Tibbetts & Rehfeldt, 2005; Weil et al., 2011). The remainder of the studies aimed to examine deictic relations in specific ACT techniques (8.82%; Foody et al., 2013; Luciano et al., 2011; Ruiz & Perete, 2015), to examine correlations between deictic framing and other complex cognitive behaviors (8.82%; Gore et al., 2010; Hooper et al., 2015; Vilardaga, 2009), to analyze performance on deictic relations protocol but without relating it to other cognitive variables (5.88%; Meléndez, 2010; Rendón, 2013), or to use deictic relational responding to explore a flexible self (2.94%; McHugh, 2015).

Looking at the journals in which the articles were published, Fig. 2 shows the journal The Psychological Record (TPR) has been the one which has published the largest amount of these studies, with 11 empirical articles and one non-empirical article. The second position was by far for International Journal of Psychology & Psychological Therapy (IJP&PT), with four empirical articles and one non-empirical article.

Fig. 2
figure 2

Summary of the journals deictic relations articles were published in. TPR = The Psychological Record; IJP&PT = International Journal of Psychology & Psychological Therapy; JEIBI = Journal of Early and Intensive Behavior Intervention; BDB = Behavioral Development Bulletin; SP = Suma Psicológica; RASD = Research in Autism Spectrum Disorders; JBT&EP = Journal of Behavior Therapy and Experimental Psychiatry; IJBCT = International Journal of Behavioral Consultations and Therapy; JCBS = Journal of Contextual Behavioral Science; BAT = Behavior Analyst Today; COP = Current Opinion in Psychology; EPR = Educational Psychology Review

Full size image

Discussion

The current bibliographical review provides growing evidence in support of the application of the RFT-based account of perspective taking to typically and atypically developing children and adults. In line with other studies (Dymond et al., 2010; Ruiz-Sánchez & Montoya-Rodríguez, 2014), the RFT approach to perspective taking has generated a series of intervention programs for establishing derived relational responding in different populations. For example, studies such as Weil et al. (2011) and O’Neill and Weil (2014) have trained deictic framing to establish theory-of-mind skills in populations who were preexperimentally deficient. Others have shown that deictic responding can help to understand clinical issues such as social anhedonia or schizophrenia (e.g., Janssen et al., 2014; Villatte et al. 2008, 2010a, b). A deictic relational frame account has also been used to explore ACT at a more basic level (Foody et al., 2013; Luciano et al., 2011; Ruiz, & Perete, 2015). The findings from these empirical articles indicated the importance of particular patterns of relational responding (including deictic relational responding) for core act processes and provide a basis for future applied research investigating deictic relational responding and ACT.

However, to date most of research has focused on populations with typical development (92.31%). The least amount of empirical work has been conducted with atypically developing adults and young children. The articles reviewed for the current study support the postulate that a deictic frames protocol can be a developmental and educational tool. However, there are very few empirical attempts to establish deictic relations in populations who are deficient in these repertoires. The present review shows an equal number of non-manipulative rather than manipulative designs. While non-manipulative designs are informative, they do not allow the prediction and influence of behavior through the manipulation of environmental variables (Skinner, 1953). Although RFT research has indicated that deictic framing runs parallel to the developmental levels of perspective taking as assessed in the cognitive approach referred to as theory of mind (see, e.g., Barnes-Holmes, Gilroy, et al., 2015; Janssen et al., 2014; McHugh et al., 2004a, 2004b; Rehfeldt et al., 2007; Villatte et al., 2008), very few studies have evaluated the impact of deictic relational training on traditional ToM tasks. Specifically, only five studies of the 26 empirical articles have provided empirical evidence examining the relationship between these variables (Jackson et al., 2014; Lovett & Rehfeldt, 2014; O’Neill & Weil, 2014; Rendón et al., 2012; Weil et al., 2011). Of these, only three articles were conducted with participants with a known diagnosis (i.e., 3/26 empirical articles were conducted with atypically developing populations). While the results from these three studies were promising, much more work is needed. Future research should expand on these early investigations of deictic relation training in populations characterized by perspective taking deficits.

In addition to this, the training protocols derived so far for training deictic relational responding need further investigation and development. The number of outcome variables in the training studies thus far has been limited. Future work should test whether improvements in deictic relational responding (such as those demonstrated by O’Neill & Weil, 2014) impact on variables such as diagnosis, decreased disruptions in functioning, or enhancing quality of life. In addition to this some empirical demonstrations of basic theoretical assumptions need to be conducted. For example, the RFT literature suggests that empathy is the transformation of emotional functions across deictic relations (see Stewart & McHugh, 2013; Vilardaga, 2009). However, there is no published work extending the research on perspective taking to empathy by testing the effectiveness of training the transformation of emotional functions across deictic relations.

A number of sources have suggested that RFT can inform the development of acceptance and commitment therapy (ACT). The current review points to three studies that have linked the RFT deictic relations literature and the development of more effective ACT techniques (Foody et al., 2013; Luciano et al., 2011; Ruiz & Perete, 2015). This type of research is in its infancy. Future research needs to expand the literature on clinical RFT and to determine how RFT deictic relations work can inform and development ACT and psychotherapeutic interventions more generally (see Villatte, Villatte, & Hayes, 2016).

Finally, in reviewing the journals in which the work is published, the two most widely used outlets are TPR and IJP&PT. Both of these journals publish empirical and theoretical works from a behavior analysis approach; however, their relative impact is low. Specifically, IJP&PT does not have an impact factor and TPR has low impact (0.879 in the year 2014). Nonetheless, the last studies have been published in higher impact journals. For example, Gilroy et al. (2015) was published in Research in Autism Spectrum Disorders (RASD) with a 2.12 impact factor, and Mori and Cigala (2015) was published in Educational Psychology Review (EPR) with a 2.56 impact. Furthermore, RASD consists of an interdisciplinary genetic, neurobiological, cognitive, and behavioral approach. This demonstrates that across time the deictic relations work is extending its scope of publication with publications in higher impact outlets.

In conclusion, the present study indicates that RFT research into deictic relations has made an important contribution to the literature in the last two decades. For example, establishing deictic relations in those diagnosed with autistic spectrum conditions or analyzing the deficits in perspective taking from a behavioral point of view in clinical populations such as schizophrenia. However, there are still important future directions to an empirically supported comprehensive understanding of the development of deictic relations in typical and atypical development, as well how a bottom-up understanding of deictic relations can promote the design of effective tools to use in contemporary therapies such as ACT. Reviews such as this one are imperative for the consolidation of the dissemination of ideas and practices to the larger community of researchers and practitioners to help advance our understanding of how to remediate deficits in core skill sets.