Keywords: Observational Learning, interbehaviorism, interbehavioral psychology, stimulus substitution, rule-governed behavior, social learning, modeling, behavior analysis
Observational learning stands as a cornerstone in the fields of psychology and behavior science. Understanding how behavior change arises through observation is crucial, particularly for behavior analysts. This article provides an in-depth exploration of observational learning, beginning with a review of foundational research, followed by an examination of behavior analytic interpretations of these findings. We then introduce the interbehavioral perspective, highlighting its unique insights and addressing some limitations within existing behavior analytic approaches. Finally, we discuss the implications of adopting an interbehavioral framework for comprehending complex behaviors, especially those involved in observational learning.
EXPLORING OBSERVATIONAL LEARNING: KEY STUDIES AND CONCEPTS
The study of observational learning marked a significant advancement in psychology. The pioneering work of Albert Bandura and his colleagues laid the groundwork for the social cognitive theory of learning (Bandura, 1986). This perspective emerged, in part, as a challenge to the notion that respondent and operant conditioning could fully explain the breadth of human behavior. Social cognitive theory emphasized the critical roles of modeling and cognitive processes in understanding how we learn and behave.
While social cognitive theory gained prominence, behavior analysts have maintained that observational learning can be effectively explained through established behavioral principles, including generalized imitation, conditioned reinforcement, and rule-governed behavior (e.g., Catania, 2007; Pear, 2001; Pierce & Cheney, 2008). However, these explanations encounter challenges when we delve deeper into the psychological mechanisms at play in observational learning. Despite ongoing research in observational learning within behavior analysis, theoretical progress in this area has been somewhat limited.
This paper aims to re-examine the core findings of observational learning research from a naturalistic, behavioral viewpoint. Recognizing the significant role of verbal processes in observational learning, we will address these processes both generally and specifically throughout our discussion. To achieve this, we will draw upon J.R. Kantor’s interbehaviorism and the scientific system of interbehavioral psychology. We will explore the potential advantages of adopting an interbehavioral perspective for gaining a richer understanding of observational learning and complex behavior more broadly.
The Foundational Role of Modeling in Observational Learning
Albert Bandura and his colleagues’ groundbreaking research in the 1960s and 70s cemented observational learning as a crucial area of study within social psychology. Their early experiments, now considered landmark studies in psychology and behavior science (e.g., Bandura & McDonald, 1963; Bandura, Ross, & Ross, 1963), were motivated by several factors. These included challenging prevailing psychoanalytic (Bandura & Huston, 1961; Bandura, Ross, et al., 1963) and developmental theories (Bandura & McDonald, 1963) and, importantly, to investigate the fundamental role of observation as a mechanism for behavior change.
Initial studies focused on the impact of modeling1 on the acquisition of behaviors like aggression (Bandura, Ross, & Ross, 1963) and moral judgment (Bandura & McDonald, 1963). These studies provided the empirical foundation for social cognitive theory, which often positioned itself as an expansion beyond traditional behaviorist accounts, questioning whether behaviorism alone could provide a complete explanation of learning. Given the lasting influence of this research, we will now summarize some key findings from observational learning studies, while acknowledging that our review is selective and aims to highlight major themes within this body of literature.
Modeling as a Catalyst for Behavior Acquisition
A central and enduring theme in observational learning research is the power of modeling in shaping behavior (e.g., Bandura & Huston, 1961; Bandura & McDonald, 1963; Bandura, Ross, & Ross, 1961). For instance, a seminal early study explored how incidental behaviors demonstrated by an experimenter could be learned while participants were engaged in a different task (Bandura & Huston, 1961). The consistent finding across these studies is that behavior change can and does occur through observation, even when that observation is unintentional and occurs within the context of other activities. While seemingly straightforward, this conclusion has profound implications for how we conceptualize learning. As we will discuss further, this finding can pose specific conceptual challenges for traditional behavioral learning theories.
The Impact of Consequences on Observational Learning
The role of consequences received significant attention in observational learning research (e.g., Bandura, 1965; Bandura, Grusec, & Menlove, 1966; [Bandura & McDonald, 1963](#anvb-27-01-06-Bandura7], Bandura, Ross, & Ross, 1963). Studies investigating consequences typically compared behavior change in children who observed a model being rewarded, a model being punished, or were in a control condition (e.g., observing non-aggressive play or no consequences). Generally, less behavior change is observed when a child witnesses a model being punished for a behavior (e.g., Bandura, Ross, & Ross, 1963).2
Interestingly, many studies found no significant difference in behavior change between conditions where a model was rewarded and conditions where there were no consequences for the model’s behavior. For example, Bandura and McDonald (1963) compared the effects of three variables on the acquisition of moral judgment responses. These variables were explored across three groups of adult-child pairs: in group one, both the model and child’s target judgments were reinforced; in group two, only the model’s behavior was reinforced; and in group three (control), there was no model and only the child’s behavior was reinforced. Importantly, in the model-child groups, trials alternated between the model and the child. Groups one and two showed greater behavior change than group three in assessments 1–3 weeks post-treatment. The researchers concluded that modeling, rather than direct reinforcement, was the primary factor in the acquisition of moral judgment. 3 Other studies also reported no significant difference between reward and no-consequence conditions, while the model-punished condition consistently yielded different outcomes (e.g., Bandura, 1965).
Further research raised questions about the potential negative effects of incentives on behavior acquisition. In one experiment (Bandura, Grusec, & Menlove, 1966), half of the participants were placed in an incentive condition where they were told they would receive candy for correctly demonstrating what they learned from a movie. After watching the film, children in both incentive and no-incentive conditions were asked to demonstrate the observed behaviors. The researchers found that children in the incentive condition performed slightly worse than those in the no-incentive condition, suggesting potential drawbacks to using incentives in learning (see Bandura, et al., 1966, p. 505).4
It is crucial to note that the terms reward, reinforcement, and operant conditioning are often used loosely in observational learning literature. From a behavior analytic perspective, a stimulus change is only classified as a reinforcer if it demonstrably increases the future frequency of the behavior it is contingent upon (e.g., Cooper, Heron, & Heward, 2007). Therefore, many stimulus changes labeled “rewards” or “reinforcers” in observational learning studies do not technically meet the criteria for reinforcers in a strict behavior analytic sense, nor are they necessarily indicative of operant conditioning processes. Nevertheless, consequences do appear to play some role in observational learning. However, the frequent finding of no difference between observation with reinforcement and observation with no consequence suggests that if consequences are influential, aversive consequences (punishment) seem to have a more pronounced effect. Given these methodological considerations, these findings should be interpreted with caution.
The Significant Role of Verbal Behavior in Observational Learning
As observational learning research evolved, increasing emphasis was placed on cognitive factors, often described using terms like coding and rehearsal. In this context, coding generally refers to mentally representing or describing observed actions, while rehearsal involves practicing or mentally reviewing what was observed. For instance, Bandura, Grusec, & Menlove (1966) examined the impact of verbally describing a model’s activity (“coding”) on the acquisition of observed behavior. This study was partly motivated by a desire to challenge behavior analysts who, it was argued, failed to account for “delayed reproduction of modeling behavior” (p. 499), which was assumed to require some form of cognitive processing.
In their study, three groups of children watched a video. One group was instructed to “verbalize every action of the model as it is being performed” (p. 501), the second group to repeatedly count “1 and a 2, and a 3, and a 4, and a 5” while watching (p. 501), and a control group observed without specific instructions. The researchers found that children who verbally described the model’s actions showed the greatest behavior change when tested later. This study underscores the early recognition of “cognitive” elements in observational learning.
Expanding on this line of inquiry, Bandura and Jeffrey (1973) investigated the roles of “coding and rehearsal” in observational learning. They found that participants who “symbolically coded” (i.e., developed number or letter coding systems) the model’s actions and immediately rehearsed (practiced) these codes demonstrated the best learning outcomes. Neither coding without symbolic rehearsal nor symbolic rehearsal without coding was sufficient for optimal learning. In other words, both creating a coded description of the model’s actions and practicing that description were found to be important for acquiring observed behavior. Interestingly, physically practicing (“motor rehearsal”) the observed behavior was deemed less critical. This suggested a growing distinction between different aspects of an individual’s behavioral repertoire and the various processes contributing to their development.
Differentiating Learning and Performance in Observational Contexts
Related to the role of verbal behavior, Bandura and colleagues began to observe a distinction between an observer’s imitative performance at a later time and their ability to verbally describe what they had observed when asked. The ability to describe the observed behavior was considered a measure of learning, while actually engaging in the observed behavior later was considered performance. For example, Bandura, Ross, & Ross (1963) found that children in both the aggressive-reward (model rewarded for aggression) and aggressive-punished (model punished for aggression) groups could accurately describe the observed sequences of aggressive behavior, despite showing differences in their own imitative aggressive behavior. Similarly, Bandura (1965) found that differences in imitation between groups disappeared on an “acquisition index,” where children were told they would be rewarded for accurately describing what the model did. These findings further highlighted the crucial role of verbal behavior in learning through observation, and pointed to different ways of measuring this type of learning. One way to measure learning from observation is through subsequent imitation, while another is through verbal descriptions of the observed actions. As these two measures appeared to be influenced by different factors, Bandura and his colleagues increasingly differentiated between them.
Theoretical Frameworks for Observational Learning
Throughout these studies, Bandura and colleagues progressively developed a theoretical model of observational learning. Driven by findings that individuals could describe observed behavior later, even if they did not imitate it directly in a testing situation (e.g., Bandura, 1965; Bandura, Ross, & Ross, 1963), they began to distinguish between learning and performance (also see Greer, Singer-Dudek, & Gautreaux, 2006). Specifically, Bandura and colleagues proposed that verbal processes were more influential in learning,5 whereas consequences were more likely to determine the extent to which an individual’s performance changed based on observation – that is, whether they actually engaged in the observed behavior. Theoretical accounts of observational learning, therefore, emphasize this distinction (e.g., Bandura & Jeffrey, 1973; Greer, Singer-Dudek, & Gautreaux, 2006).
Bandura and his colleagues conceptualized observational learning through an input-output, cognitive model. Bandura and Jeffrey (1973) described four key processes involved in observational learning: attentional, retention, motor reproduction, and motivational. They stated, “Within this framework, acquisition of modeled patterns is primarily controlled by attention and retention processes. Whereas performance of observationally learned responses is regulated by motor reproduction and incentive processes” (p. 122).
- Attentional processes were defined as cognitive abilities that “regulate sensory registration of modeled actions.”
- Retention processes were those that “took transitory influences and converted [them] to enduring internal guides for memory representation” (Bandura & Jeffery, 1973, p. 122).
- Motor reproduction processes are those that translate stored memory representations of component actions into overt behaviors that resemble the modeled actions.
- Motivational processes determine whether these learned behaviors are actually expressed as overt actions.
According to this model, observational learning allows for both immediate imitation and delayed reproduction of observed behaviors across various situations. Bandura and Jeffrey (1973) concluded, “After modeled activities have been transformed into images and readily utilizable verbal symbols, these memory codes can function as guides for subsequent reproduction” (p. 123). They also found that individuals who actively transformed modeled actions into verbal descriptions or visual images achieved higher levels of observational learning compared to those who did not.
As a result of these and other experiments, Bandura theorized that observational learning was a fundamental aspect of human development, shaping personality (Bandura & Walters, 1963) and the development of social and antisocial behaviors in children (Bandura, 1973). Crucially, this research demonstrated that humans can learn without directly experiencing the consequences of their own actions. Therefore, any comprehensive account of learning, particularly within behavior analysis, must address these instances of observational learning. Specifically, behavior analysts need to explain how novel behaviors can be acquired in the absence of direct reinforcement for the learner and articulate the role of verbal behavior in this process.
In summary, Bandura and colleagues’ studies challenged the exclusive emphasis on rewards in shaping behavior, suggesting that observation of a model was a critical factor, potentially even more so than direct reinforcement history. Furthermore, they proposed that learning from observation was mediated by cognitive processes, including “verbal coding.” These findings appeared to challenge the completeness of the behaviorist perspective and paved the way for social cognitive theory. However, it’s important to reiterate that Bandura and colleagues’ use of terms like reinforcer and reinforcement often differed from strict behavior analytic definitions, making it difficult to draw definitive conclusions about the precise role of consequences from their research alone. Nevertheless, their work undeniably established observational learning as a critical area for behavior science to investigate.
Bandura identified limitations in the operant interpretation of behavior, albeit based on a potentially incomplete understanding of operant conditioning. Observational learning appears to contradict traditional stimulus—response—reinforcer analyses, even when considering more contemporary concepts like motivating operations. Notably, observational learning often involves the emergence of novel responses, behaviors that have not been previously reinforced in the observer. Additionally, delayed imitation is common, presenting conceptual challenges to traditional behavioral concepts (e.g., Bandura, Grusec, & Menlove, 1966, p. 499). It is therefore understandable why Bandura’s work might be seen by some as an expansion or departure from the behaviorist viewpoint. Despite any limitations, Bandura and his colleagues raised crucial questions about the role of observation and verbal behavior in behavior change processes.
However, Bandura’s model relies on hypothetical, internal entities that are not directly observable in the natural world. His theoretical constructs are not derived from directly observable events and, therefore, cannot be empirically studied in a natural science framework (see Kantor, 1957; Smith, 2007). Instead, they are inferences stemming from a mentalistic, dualistic worldview. Behavior analysts have long argued that embracing such constructs can hinder a truly scientific analysis of behavior (e.g., Skinner, 1953). Consequently, behavior analysts have proposed alternative conceptualizations of observational learning. The following section outlines the behavior analytic perspective on observational learning.
THE BEHAVIOR ANALYTIC INTERPRETATION OF OBSERVATIONAL LEARNING
The behavior analytic explanation of observational learning is primarily rooted in the concept of generalized imitation (Baer, Peterson, & Sherman, 1967; Baer & Sherman, 1964; Pierce & Cheney, 2008). Generalized imitation is a process where an individual is trained to imitate a series of modeled responses (e.g., “do this” while the model touches their nose). After successful training across multiple examples, the individual is then asked to imitate a novel behavior that has never been modeled before. Generalized imitation is said to occur when the individual performs this new, unmodeled, and unreinforced behavior; in essence, imitation has “generalized” to new behaviors. Furthermore, it is proposed that social interactions shape delays in imitative responses. Thus, it is argued that “all instances of modeling and imitation involve the absence of the Sd” (Pierce & Cheney, 2008, p. 252), meaning the discriminative stimulus for imitation is not always immediately present. For example, a child might watch a TV show and, much later, repeat a phrase from it, perhaps while playing in the car. A parent might then say, “Yes, that’s what you heard on TV!” In this view, the individual learns to imitate observed behavior even in the absence of a specific immediate stimulus, and potentially at a later time. The individual is said to “emit” behaviors that fall under the category of generalized imitation.
Conditioned reinforcement is also a key element in the behavior analytic understanding of observational learning and imitation in general. Behaviors that closely resemble the observed behavior of models are assumed to have a history of reinforcement. Therefore, behaving in a manner similar to a model can itself become a conditioned reinforcer. This conceptualization is particularly relevant to the behavior analytic explanation of delayed imitation (see Gladstone & Cooley, 1975; Rosales-Ruiz & Baer, 1997).
Behavior analysts also offer an account of the “verbal coding” aspect of observational learning. They propose that individuals develop self-rules based on their observations of the environment (e.g., Hayes, Barnes-Holmes, & Roche, 2001; Hayes, Zettle, & Rosenfarb, 1989; Poppen, 1989). It is assumed that social learning teaches individuals to tact (Skinner, 1957) relationships in their environment, and these verbal descriptions exert significant control over behavior. It is suggested that a substantial amount of rule-following behavior is reinforced throughout life. Combined with a history of reinforced tact repertoires, individuals both create self-rules (verbalizing if-then relationships observed in their environment) and subsequently engage in rule-following behavior based on these self-rules.
For example, a child might observe a teacher praising another child for correctly matching a Spanish flashcard to its English equivalent (“Good job matching ‘perro’ with ‘dog'”). Two days later, the observing child might be asked to “match same” when presented with the same Spanish flashcard and correctly place it with the corresponding English flashcard. From a behavior analytic perspective, it might be assumed that the child already possesses a generalized imitative repertoire, enabling them to imitate the observed child’s behavior at a later time (consistent with conditioned reinforcement hypotheses). Furthermore, the child may or may not have verbally labeled (tacted) the observed relationship when it occurred (rule-stating) and then engaged in rule-following behavior when interacting with the flashcard later. Both of these possibilities are consistent with the behavior analytic viewpoint. Importantly, the behavior analytic position does not require rule-stating and rule-following for observational learning to occur. Supporting this, recent studies by Greer and colleagues suggest that observational learning can occur without rule-following. For instance, individuals have learned new words through observation without observing consequences for another person, and stimuli have become conditioned reinforcers through observing others interact with them, neither of which necessarily involves rule-governed behavior (see Greer & Ross, 2008, Greer & Speckman, 2009).
It’s important to acknowledge that many of these issues are at the forefront of ongoing debate and development within behavior analysis. Perspectives such as joint control (e.g., Lowenkron, 1998), naming (e.g., Horne & Lowe, 1996), relational frame theory (e.g., Hayes, Barnes-Holmes, & Roche, 2001), and verbal behavior development (e.g., Greer & Ross, 2008; Greer & Speckman, 2009) all offer accounts of the phenomena discussed here. The active engagement with these issues indicates progress and growth within the field. However, missteps are possible in the pursuit of understanding these complex phenomena, which could have varying degrees of impact on behavior analysis as a scientific discipline. We propose that the interbehavioral perspective can provide a valuable foundation for researchers as they continue to explore these areas (see Morris, Higgins, & Bickel, 1982).
In general, the behavior analytic conceptualization of observational learning relies on generalized imitation, conditioned reinforcement, and various verbal processes, depending on the specific theoretical approach. These processes attempt to explain how imitative behaviors that have never been directly reinforced can occur later and account for the role of verbal behavior in observational learning. The diversity of perspectives within behavior analysis on these issues can be seen as a sign of progress and intellectual dynamism, but it also highlights the ongoing need for further theoretical development and system-building in this area. In the following sections, we will examine the behavior analytic position through the lens of interbehavioral psychology. Before that, we will briefly introduce the interbehavioral perspective, as it is less familiar to many behavior analysts.
THE INTERBEHAVIORAL PERSPECTIVE ON OBSERVATIONAL LEARNING
From the standpoint of interbehavioral psychology, the primary unit of analysis is always a thoroughly naturalistic, psychological event. This event is defined as the interaction between a stimulus function (sf) and a response function (rf) (Kantor, 1958). Crucially, this interaction always occurs within a complex, interconnected field of factors. This field is represented by the formula: PE = C (k, sf, rf, hi, st, md); where PE is the psychological event, C signifies the interrelationship of all participating factors, k represents the unique organization of these factors, sf is the stimulus function, rf is the response function, hi is the interbehavioral history, st denotes setting factors, and md is the medium of contact. It is essential to understand that this is a single event, a single interbehavioral field. Any change in one factor alters the entire field. None of these factors are seen as independent, dependent, or having causal priority. Instead, all factors are considered equal and integral components of the unified whole (see Smith, 2006).
Of particular relevance to understanding observational learning and complex behavior is interbehaviorism’s explicit distinction between stimulus objects and stimulus functions (e.g., Kantor, 1924, pp. 47–48; Parrott, 1983a, 1983b, 1986). Kantor differentiates the stimulating action of stimulus objects from their formal, physical properties. He suggested that the borrowing of “stimulus” and “response” from biology, where these functions are more structurally determined, may have contributed to a failure to distinguish between object and functional properties in psychology (Kantor, 1958, p. 68). For instance, in interbehaviorism, a picture (stimulus object) is distinct from its psychological functions. This distinction makes it easier to explain how seeing something in the absence of the thing itself (like a picture reminding you of a past event) is possible without resorting to mentalistic explanations (see Parrott, 1983a, 1983b, 1986; Skinner, 1974). The process by which stimulus objects acquire these psychological functions is central to understanding complex behavior, including observational learning, and we will now elaborate on this process.
Kantor emphasized that association conditions are fundamental psychological processes (1921, 1924). Here, “association” refers to spatiotemporal relationships – the co-occurrence of various factors in the environment in space and time. Crucially, these associations exist in the environment, not within the organism. Furthermore, it is not the organism that is “associating”; rather, the environment is the locus of association. Association conditions can involve relationships between stimuli and responses, stimuli and stimuli, settings and stimuli, settings and reactions, settings and settings, and reactions and reactions (including both overt and covert variations; Kantor, 1924, pp. 321–322).
Stimulus Substitution: The Core of Interbehavioral Learning
Stimulus substitution is the outcome of an organism’s history of interaction with various association conditions (Kantor, 1924, 1958; Parrott, 1983a, 1983b, 1986). When stimulus objects A (e.g., coffee shop) and B (e.g., Peter) have frequently occurred together in space and time, and an organism has interacted with this relationship, stimulus objects can acquire the stimulational properties of other objects, even when those objects are not physically present. This is how, for example, you might “see” Peter when you enter a coffee shop you used to frequent with him, even if he is not actually there. In this scenario, stimulus A (coffee shop) and B (Peter) were associated in space and time, and through interaction with this association, B becomes A (B[A]) and A becomes B (A[B]), psychologically speaking (see Hayes, 1992a). This process is vital for understanding complex behaviors. Furthermore, it allows interbehaviorists to conceptualize the past and present as interconnected, avoiding both mentalistic and reductionistic approaches that locate the past “within” the organism in some way (see Hayes, 1992b).
Moreover, through generalization, stimuli that share physical features with those involved in spatiotemporal associations can also develop substitute stimulus functions. A coffee shop that resembles the one you used to visit with Peter might also evoke a sense of Peter’s presence. Substitute stimulus functions can generalize to stimuli that have never directly participated in spatiotemporal associations but are physically similar to stimuli that have, and therefore share similar stimulus functions. This type of process can become quite subtle and likely plays a role in complex phenomena like imagining and dreaming.
At this point, it is important to address a potential misunderstanding regarding interbehaviorism, specifically concerning association conditions and the development of substitute stimulus functions.6 We are suggesting that all stimuli that co-occur in space and time and with which an organism interacts can develop substitute stimulus functions for one another, given the appropriate interbehavioral history. In theory, any stimulus could potentially develop substitute stimulus functions for any other stimulus, given a sufficiently complex interbehavioral history. As an individual’s interbehavioral history becomes increasingly elaborate, one can imagine how all stimuli could become interconnected through substitute stimulus functions, leading to a sense of psychological interconnectedness. However, it’s crucial to remember that the stimulus function←→response function interaction is always embedded within a uniquely complex, multifactored field. As Kantor stated, “Each interaction is always absolutely specific. What the reacting organism and the stimulus object do in each interaction constitutes a distinctly unique relational happening” (1977, p. 38). Thus, while a specific stimulus object may acquire a wide range of potential substitute stimulus functions, the actualization of specific substitute stimulus functions is always context-dependent and occurs within a unique interbehavioral field. For example, a glass of sangria might evoke associations with a particular friend in one context (remembering drinking sangria together), while in another context, the same glass of sangria might evoke associations with the music of a live band (recalling music played at a restaurant where you drank sangria). While a wide range of potential substitute stimulus functions may exist for any stimulus object, in each specific psychological event, particular substitute stimulus functions are actualized.
We have briefly introduced key aspects of interbehavioral psychology that are particularly relevant to understanding observational learning. From this perspective, individuals observe (interact with) spatiotemporal associations in their environment (e.g., a child observing another child put paper in a recycling bin and receive praise). Subsequently, the stimulus objects involved can substitute for the prior observation (e.g., the scrap paper may acquire the stimulus functions of praise experienced in the observed interaction). The scrap paper, in effect, develops the stimulational properties of the observed relationships; it substitutes for them. Psychologically, the scrap paper becomes those relationships (see Hayes, 1992a, 1992b).
The role of verbal behavior must also be considered within this interbehavioral framework. One outcome of interacting with an observed relationship is the ability to describe it verbally. Describing an observed relationship requires interaction with it, making verbal descriptions a strong indicator that the observed relationships have indeed been engaged with. However, from an interbehavioral perspective, verbal behavior, including rules, does not explain observational learning. Whether or not an individual verbally describes an observed relationship does not cause behavior change at a later time. However, verbal description is likely to be correlated with later behavior change, as it indicates that the individual has interacted with the observed relationship. Furthermore, to the extent that rule-statements can substitute for a history of direct reinforcement, they may enhance learning through observation. Importantly, verbal behavior is not seen as “mediating” responding in this view. While verbal behavior participates in observational learning, it should not be given causal or special status. Observational learning can and does occur in the absence of verbal behavior, as demonstrated by animal research in this area (e.g., Biederman, Robertson, & Vanayan, 1986; Meyers, 1970; Reiss, 1972).
Our contention that verbal behavior should not be assigned causal status in observational learning might seem to conflict with some prominent perspectives in behavior analysis. Research on naming (e.g., Miguel, Petursdottir, Carr, & Michael, 2008), joint control (e.g., Lowenkron, 1998), and generalized imitation (e.g., Horne & Erjavec, 2007) often suggests a mediational role for verbal behavior. Again, we acknowledge that verbal behavior can be helpful in many contexts, but we caution against giving it a privileged status. Verbal behavior may, but importantly also may not, participate in learning from observation. It is not necessarily “mediational.” This interbehavioral perspective is both parsimonious and comprehensive. It avoids unnecessary assumptions or constructs and accounts for observational learning both with and without verbal behavior.7
From an interbehavioral viewpoint, observational learning is not conceptually puzzling. Stimulus substitution provides a straightforward, naturalistic, and parsimonious way to understand complex processes, including those involved in observational learning. Furthermore, the interbehavioral perspective avoids certain shortcomings found in behavior analytic interpretations of observational learning, which we will now address specifically.
Critique of the Behavior Analytic Perspective on Observational Learning
As discussed earlier, the behavior analytic conceptualization of observational learning relies on generalized imitation, conditioned reinforcement, rule-governed behavior, and verbal processes. From an interbehavioral perspective, these analyses do not fully articulate the nature of stimulation within the psychological event. Interbehaviorism emphasizes that the psychological event is always the stimulus function←→response function interaction. The generalized imitation analysis raises questions about the nature of the stimulus being interacted with. It is unclear what precisely the stimulus is. This issue is further complicated by the suggestion that generalized imitation involves responding in the absence of a discriminative stimulus (Pierce & Cheney, 2008, p. 252). Given the interbehavioral premise that psychological events always involve sf←→rf interactions within multifactored fields, this account becomes problematic. Similarly, the concept of deriving and following self-rules leaves us questioning the nature of the stimulus. It is unclear what an individual is interacting with when deriving a self-rule and when following such a rule. Again, based on interbehavioral principles, both of these analyses require a more thorough consideration of the stimulus involved.
In addition to concerns about the nature of the stimulus, behavior analytic conceptualizations often fail to explicitly specify the location of the stimulus. It is unclear where the stimulus being interacted with is situated. This lack of clarity regarding the stimulus’s nature and location can inadvertently create openings for mentalistic explanations to re-emerge. In the case of generalized imitation, we may find ourselves saying that the response is “in the repertoire” of the individual, implying that the stimulus is private, covert, or biological (also see Hayes & Fryling, 2009). Alternatively, it might be suggested that the organism “derives” or “relates” in the context of verbal processes. In these cases, we either avoid specifying the stimulus altogether, locate it within the organism, or relegate it to the domain of other scientific disciplines, such as biology.8 In each scenario, we fall short of providing a fully psychological account of the event of interest, leaving our task incomplete. Historically, when scientific explanations are incomplete, dualistic and reductionistic perspectives readily step in to fill the gaps. While it could be argued that much of contemporary work in complex behavior avoids some of these concerns, a lack of explicit attention to these crucial issues can lead to long-term confusion and a potential resurgence of mentalistic thinking.
CONCLUSION: INTERBEHAVIORISM AND THE FUTURE OF OBSERVATIONAL LEARNING RESEARCH
The behavior analytic community maintains a strong interest in the important processes involved in observational learning (e.g., Alvero & Austin, 2004; Bruzek & Thompson, 2007; Greer & Singer-Dudek, 2008; Greer, Singer-Dudek, Longano, & Zrino, 2008; Moore & Fisher, 2007; Ramirez & Rehfeldt, 2009; Rehfeldt, Latimore, & Stromer, 2003). Furthermore, there is compelling evidence suggesting that observational learning procedures hold significant clinical value compared to other intervention methods (see Hayes, Kohlenberg, & Melancohn, 1989). What is needed is a thoroughly naturalistic conceptualization of observational learning, one that rigorously avoids mentalism (i.e., internal, unobservable mediating steps within the organism). As we have argued, the interbehavioral perspective offers precisely this: a clear, consistent, and entirely naturalistic framework for understanding observational learning. Moreover, it achieves this comprehensiveness without requiring any additional constructs to explain complex processes.
We believe that the interbehavioral perspective presented in this paper can be effectively integrated with contemporary research and scholarship in behavior analysis. This integration is particularly fruitful when we maintain clear distinctions between investigative constructs and the events themselves, as advocated by interbehaviorists (see Fryling & Hayes, 2009; Kantor, 1957; Smith, 2007). Kantor (1958) suggested that investigative constructs are permissible within the investigative subsystem of science, but they should not be confused with the subject matter itself or the broader philosophical foundations of the science. The constructs we use to understand relationships among factors in psychological events should not be mistaken for representations of the subject matter as a whole, nor should they be seen as explaining or causing one another. For example, both operant and respondent processes can be understood within the broader framework of association and the resulting phenomenon of stimulus substitution. Contemporary behavior analytic research necessitates emphasizing specific aspects of the interbehavioral perspective, particularly the role of context (unique multifactored fields) and the actualization of specific substitute stimulus functions. In this regard, research on relational responding is particularly insightful. This line of research manipulates multiple historical association conditions in unique ways within varying contexts and then examines the “emergence” of a wide range of behavioral outcomes. When these outcomes are conceptualized as unique forms of substitute stimulation operating within historical, multifactored fields, their explanations remain consistently naturalistic. We contend that much of contemporary research and scholarship in behavior analysis can and should be integrated with the interbehavioral perspective. Such integration could serve to unify the efforts of researchers in the field and ultimately enhance our productivity as a scientific enterprise.
Despite the limitations of Bandura’s work, the process of learning through observation remains a compelling and crucial area for a comprehensive analysis of behavior. If we value comprehensiveness in our behavioral science, our core concepts and principles must be relevant to and capable of explaining observational learning. Furthermore, this comprehensiveness is valuable only when achieved within the context of validity (internal consistency) and significance (external consistency within the broader field of sciences; see Clayton, Hayes, & Swain, 2005; Kantor, 1958). The interbehavioral perspective is particularly valuable in this regard. Kantor’s conceptualization of the psychological event, in its full complexity, offers a pathway to fully integrate even the most complex behaviors, including those involved in observational learning, into a natural science approach to behavior analysis.
Footnotes
Cristin Johnston is affiliated with Spectrum Center, Oakland, CA.
1 The term modeling is used interchangeably with observation and demonstration in this context. Modeling refers to observing a demonstration of a response and the surrounding circumstances.
2 See Greer et al., 2004 for related studies on peer tutoring, where it was the observation of corrections, rather than simply reinforcement, that led to observational learning.
3 The researchers acknowledged that their positive statements might not have been optimal reinforcers, suggesting that the modeling plus reinforcement condition might have shown stronger effects with more potent reinforcers (Bandura & McDonald, 1963, p. 281).
4 The idea that rewards can sometimes hinder learning aligns with concerns raised by Alfie Kohn (1999).
5 In this literature, learning is defined as the individual’s ability to verbally describe observed behavior at a later time.
6 Some critics have described interbehaviorism as exhibiting a “loose form of associationism” (e.g., Hayes, Barnes-Holmes, & Roche, 2001, p. 8).
7 We acknowledge the importance of language in many socially significant behaviors and the value of understanding how to promote these behaviors (e.g., categorization). However, we argue against assigning special status to language in the conceptualization of observational learning itself.
8 It is crucial to note that even when biological factors are observed, they are never seen to be engaging in the psychological event of interest. We cannot observe the brain or any biological component of the organism directly engaging in the behavior we are analyzing (see Kantor, 1947). Confusion between what is measured and what is claimed to be measured is a common issue in science (see Kantor, 1957; Smith, 2007), and is particularly likely when the boundaries between scientific disciplines are not clearly defined.
