Methodological Behaviorism: Historical Origins of a Problematic Concept (1923–1973)

Abstract

“Methodological behaviorism” is a term that frequently appears in the behavioristic literature, but one accompanied by considerable semantic confusion: the term is used to denote very different theoretical positions and the authors classified as methodological behaviorists are many and various. In order to understand the polysemic character of this term, we propose a historical analysis of its origins and development in the literature from the 50 years following its first appearance in 1923. The results reveal that it has been used by authors as diverse as Karl Lashley, B. F. Skinner, Herbert Feigl, and Gustav Bergmann. Moreover, it has been defined in terms of two central features (one a methodological assumption and the other a metaphysical one) and used to demarcate positive and negative forms of behaviorism, depending on how each author has understood those features and forms. We conclude that the term’s polysemic character and different uses can be traced back to its roots in the 1920s, which helps us to understand the semantic confusion in the contemporary literature.

A quick survey of the behavioristic literature reveals that the term “methodological behaviorism” has been widely used by different authors but with different purposes. Recent uses include attempts to criticize (e.g., Costall, 2018) or defend (e.g., Manning, 2014) specific conceptions of methodological behaviorism, as well as to evaluate its compatibility with certain psychological theories in terms of similarities (e.g., Cheng, 2018) or differences (e.g., Moore, 2017; Rachlin, 2013).

A close look at these diverse uses reveals that (1) different definitions appear in specific contexts and (2) the authors classified as methodological behaviorists defend distinct, sometimes incompatible, positions.

The polysemic character of the term is undeniable. For example, Segraves (1982) suggests that it denotes the application of the scientific method to psychology: “This viewpoint advocates controlled experimentation, independent replication of results, operationism of concepts, and testability of hypotheses. Methodological behaviorism is really a misnomer. It refers simply to the use of scientific methodology” (p. 110). By contrast, Wiebe (2015) defends the view that methodological behaviorism involves an ontological compromise, insofar as it accepts the existence of mind: “methodological behaviorists did not so much deny the existence of mental states as argue that such alleged states never needed mention, since intersubjectively observable stimuli and responses were adequate to describe and explain human behavior” (p. 22). Then again, Stegmüller (1969) claims that any psychological theory that restricts itself to observable behavior is a methodological behaviorism, regardless of its ontological compromises: “methodological behaviorism . . . either leaves open the question of the existence of psychical phenomena or, while expressly acknowledging that they exist, restricts itself to supplying behavioral criteria for all that is psychical” (p. 428). Meanwhile, Martin (1978) argues that it is indeed possible to identify different forms of methodological behaviorism, but that they all share a disposition to analyze psychological events in terms of observable data alone, and this for methodological reasons, independent of other philosophical justifications.

A second feature involved in the conceptualization of methodological behaviorism is the choice of its representatives. For the most part, conceptualizations include an indication of who is to be considered a methodological behaviorist, or at least what position should count as methodological behaviorism. In this sense, the diversity and incongruence in such indications is explicit. For example, B. F. Skinner has been labeled a methodological behaviorist (e.g., Martin, 1978; Woolfolk, 1983) but also as someone whose work stands in contrast to methodological behaviorism (e.g., Day, 1983; Moore, 2011/2012, 2013). The same applies to authors as distinct as John B. Watson (Moore, 2017; Morris & Todd, 1999; Smith, 1984) and Willard v. O. Quine (e.g., Cheng, 2018; Rappaport, 1979). Moreover, representatives of various and competing theoretical views have also been included among the methodological behaviorists. This is the case for Robert Yerkes (Roback, 1923), Edwin G. Boring (Skinner, 1945a), Elliott Sober (Trigg, 1987), Gustav Bergman (Flanagan, 1980; Moore, 2013), James R. Angell (Kessen & Cahan, 1986), Paul Oppenheim (Henle & Baltimore, 1967), Richard J. Herrnstein (Smillie, 1978), and Wilfrid S. Sellars (e.g. Manning, 2014; Tripodi, 2011). Not even Kant (Zimmer, 2001) and Plato (Shanon, 1984) have escaped such classification. This plurality and lack of consensus in the identification of the term’s referents reflect its polysemic nature as well as its different uses. In short, there is much confusion in the literature.

One way to understand and possibly explain the problematic nature of the concept of methodological behaviorism in the contemporary literature is to identify its historical origins, a task that has yet to be completed, despite some tentative first steps. Moore (2001), for instance, suggests that methodological behaviorism “derived from logical behaviorism, analytic behaviorism, and commitments to ‘truth by agreement’ and the operationism of Boring and Stevens” (p. 239). In addition, Moore (2013), marks the beginnings of methodological behaviorism in the first quarter of the twentieth century: “radical behaviorists argue that methodological behaviorism began in the first quarter of the 20th century and has been the modal, orthodox orientation that underlies nearly all of psychology since the middle of the 20th century” (p. 146). Scott (2016) limits himself to associating methodological behaviorism with positivism and empiricism: “Methodological behaviourism has its origins in the sociological theory of positivism and the philosophical theory of empiricism” (p. 38). However, neither author attempted a systematic historical account, meaning that important and influential publications, such as Lashley (1923), were ignored altogether. The historical details are thus lacking, and it remains to be determined whether the polysemic nature of the term is a recent feature that has emerged as a result of deviations from its original meaning or use, or a characteristic that has dogged the term since its inception.

The goal of this article is to address this issue by means of a systematic analysis of the primary sources related to debates on the nature of behaviorism. We will present and discuss the first appearances of the term in the literature as well as its usages by the respective authors. In order to do this, we ask three questions: (1) How is the term defined? (2) Who are the methodological behaviorists? (3) How is the term used?

We will restrict our analysis to the 50 years following the term’s first appearance, that is, from 1923 to 1973. There are two reasons for this decision. First, the analysis we offer here is sufficient to address the issue raised above. Second, from 1974, new ways of understanding methodological behaviorism were proposed that make things still more complicated, demanding a much wider analysis than we can offer within the space of a single article. Examples of those complications are the modification in the concept of methodological behaviorism implemented by Day (e.g., 1983), which was followed by many behavior analysts (e.g., Leigland, 1997; Moore, 1975, 1980, 1981, 1989, 2001, 2011/2012, 2013), and the subsequent debate about the classification of contemporary versions of behaviorism (cf., Baum, 2011; Moore, 2011). It will be part of a separate, future investigation.

The Idea before the Concept

Before the term “methodological behaviorism” appeared in the literature, general proposals for demarcating behaviorism had already been made. Once behaviorism emerged as a new approach to psychology, psychologists and philosophers began to discuss whether it had to deny the existence of consciousness and the mind. Such debates would contribute to the formulation of the concept of methodological behaviorism, as we will see in the next section.

James Pratt (1875–1944), an American philosopher and former president of the American Theological Society, argued that behaviorist attempts to describe conscious states in terms of overt behavior would necessarily be unsuccessful. In order to be logical and consistent, he said, “it is necessary for Behaviorism to deny the existence of consciousness (in the old-fashioned subjective sense”; Pratt, 1922, p. 603). It meant, above all, the assumption of a metaphysical position, namely, materialism. Otherwise, he insisted, if the mind existed and could influence the body, Behaviorism cannot be even a science of behavior” (p. 604).

It is interesting to note that, in Pratt’s discussion, behaviorism has two aspects, namely, a metaphysical one and a methodological one. The first involves the denial of the existence of consciousness. The second involves the methodological prescription of “carefully observing and writing down―or quite as often, imagining―various forms of human behavior and presenting the results as an objective description of mind” (Pratt, 1922, p. 602).

In a direct response to Pratt, philosopher Stephen Pepper (1891–1972) claimed that Pratt had misunderstood behaviorism. According to Pepper, the behaviorist did not need to deny consciousness in order to be consistent. “All a behaviorist has to affirm to be thoroughgoing and self-consistent,” he said, “is that there is nothing mental which is not susceptible to objective description in terms of behavior. The behaviorist need be nothing more than a thoroughgoing parallelist” (Pepper, 1923, p. 243). However, because psychophysical parallelism is a kind of dualism, Pepper accepted the conclusion that behaviorism is compatible with metaphysical dualism, provided that there is no interaction between the mental and the physical. In this way, the behaviorist is free “to give an objective description of introspective facts” (p. 243).

One finds in Pepper’s account of behaviorism the same elements as in Pratt’s. There is a metaphysical and a methodological discussion. Instead of a materialistic monism, Pepper defended a kind of dualism for behaviorism. Moreover, he stressed the failures of the introspective method and welcomed the methodological efforts to give “a behavioristic account of mind” (Pepper, 1923, p. 244).

These two aspects of behaviorism―its metaphysical and methodological assumptions―constituted the core of more refined characterizations that were to appear in the subsequent literature. Thus, the concept of methodological behaviorism was based on the previous understandings of these two central concerns. Not, however, in a consensual way, as we shall see.1

Roback and Lashley: The Origins

The first use of “methodological behaviorism” to designate a particular form of behaviorism seems to have appeared with Abraham Roback (1890–1965), a psychologist who taught at Harvard. In his book Behaviorism and Psychology, in which he intended “to give a comprehensive survey of the whole behavioristic system” (Roback, 1923, p. 11), Roback proposed four classes (prebehaviorism, behaviorism proper, psychobehaviorism, and nominal behaviorism) and 21 types of behaviorism. “Methodological behaviorism” appeared in connection with the work of Robert Yerkes (1876–1956), one of the main figures in comparative psychology and a pioneer in intelligence testing. According to Roback, “there are behaviorists and behaviorists, running all the way from those who consider consciousness not only a myth but a profanity down to the mitigated form such as Yerkes seemingly represents” (p. 42). Because “[Yerkes] is primarily interested in the study of behavior for methodological reasons,” Roback called this approach “methodological behaviorism” (p. 41).

Two points deserve attention in Roback’s classification. First, the adjective “methodological” refers to the effective (experimental) methods of studying behavior, that is, how to approach psychological phenomena. Second, consciousness is the central problem that separates Yerkes’s behaviorism from proper behaviorism such as Watson’s. A methodological behaviorist accepts the existence of conscious phenomena, whereas a strict behaviorist regards it as a myth. As a consequence, Roback’s conceptualization has an ontological component, namely, the necessary compromise of behaviorism with materialism.

This brings us to Roback’s use of “methodological behaviorism.” Roback intended not only to offer a general classification of behaviorism, but also a general evaluation. In this sense, methodological behaviorism represents a mitigated form of behaviorism, a superficial and inconsistent position in comparison with, for example, Watson’s behaviorism. In Roback’s (1923) words, “if behaviorism is not to be mutually incompatible with the orthodox view of psychology, it must be taken in the sense of a practical rough approach to phenomena without reference to their further analysis” (p. 43). Thus, methodological behaviorism is not properly speaking behaviorism, or it is behaviorism in name only (nominal behaviorism).

Despite Roback’s efforts to classify all forms of behaviorism, his classification did not have much impact on subsequent literature. Further developments and theoretical discussions of behaviorism did not take into account his classification.

It was Karl Lashley (1890–1958)―Watson’s ex-student and friend, famous for his contributions to American neuropsychology―who first gave “methodological behaviorism” a definition that was debated and subsequently developed. In discussing the behavioristic interpretation of consciousness, Lashley provided another taxonomy of behaviorism that was much simpler than Roback’s. For him, “in various discussions of the scope of behaviorism, three distinct and incompatible formulations are discernible” (Lashley, 1923, p. 238).

First, there are behaviorists who accept the existence of consciousness and the possibility of approaching it introspectively. However, because facts about consciousness are irrelevant to behaviorism, they are of no interest. For Lashley (1923), “this is merely psychophysical parallelism with emphasis on the physical. It is the view of Bechterew and other early objectivists” (p. 238).

Behaviorists of the second stripe claim that “facts of consciousness exist but are unsuited to any form of scientific treatment” (Lashley, 1923, p. 238). Lashley attributed this position not only to Watson in his early writings (Watson, 1913) but also to Albert Weiss (1879–1931), a German-American psychologist who defended behaviorism in its beginnings (Weiss, 1917a, 1917b). For him, this approach is what “we may call a methodological behaviorism” (Lashley, 1923, p. 239).

According to the third type, “the supposedly unique facts of consciousness do not exist. An account of the behavior of the physiological organism leaves no residue of pure psychics. Mind is behavior and nothing else” (Lashley, 1923, p. 240). Lashley attributed this view to much of Watson’s work because of “his fundamental denial of the fact of consciousness” (p. 240). As a result, Lashley explained, “I shall speak of this doctrine as strict behaviorism, or for brevity simply as behaviorism” (p. 240).

In Lashley’s account of behaviorism, it is clear that the central criterion is the place of mind or consciousness in psychological science, bringing him close to Roback. Because methodological behaviorists accept the existence of mental facts, they cannot be behaviorists in the strict sense. They share with strict behaviorists only the methods of studying behavior, but this is not enough. In contrast to methodological behaviorism, strict behaviorism “makes no concessions to dualistic psychology and affirms the continuity in data and method of the physical, biological, and psychological sciences” (Lashley, 1923, p. 240). For this reason, methodological behaviorism receives here a pejorative connotation. Moreover, Lashley understood strict or genuine behaviorism in terms of an ontological compromise with materialism, as Roback had done.

Thus, it is clear that in Lashley’s classification the use of “methodological behaviorism” serves to indicate what behaviorism, properly speaking, is not or should not be. It represents a weak or inconsistent position to be abandoned in favor of strict behaviorism. Thus, both Roback and Lashley conceptualized methodological behaviorism in a similar way.

Skinner: Methodological versus Radical Behaviorism

The most famous of all behaviorists, B. F. Skinner (1904–1990), also used “methodological behaviorism” in contrast to the form of psychological science he was willing to defend. The context for such a defense was a symposium organized by Herbert Langfeld (1879–1958), then chief editor of Psychological Review, in which invited members were to discuss the application of operationism in psychology (Langfeld, 1945).2

Skinner contributed to the symposium at two different points (Skinner, 1945a, 1945b). In both, he welcomed the idea of an operational analysis of psychological terms but indicated nonetheless its shortcomings. In general, he saw it as “a good thing in any science, but especially in psychology because of the vast vocabulary of ancient and non-scientific origin” (Skinner, 1945a, p. 271). At the same time, he added, “the possibility of a genuine operationism in psychology has not yet been fully explored” (Skinner, 1945b, p. 294).3 Why not?

Skinner argued that although classical behaviorism had been mostly engaged in “a thoroughgoing operational analysis of traditional mentalistic concepts” (Skinner, 1945a, p. 271), behaviorists had failed to offer “an acceptable formulation of the ‘verbal report’” (p. 271). “What we want to know in the case of many traditional psychological terms,” he continued, “is, first, the specific stimulating conditions under which they are emitted . . . and, second . . . why each response is controlled by its corresponding condition” (p. 272). However, in the case of subjective terms, there is the problem of how to deal with private events.

For Skinner, this was the central problem for a genuine operational analysis in psychology and, as a result, for behaviorism. And because there were at least two different ways of dealing with operationism in psychology, he concluded that there were at least two corresponding forms of behaviorism. The first, he said, was “methodological behaviorism,” according to which “the world is divided into public and private events, and psychology, in order to meet the requirements of a science, must confine itself to the former” (Skinner, 1945b, pp. 292–293). Skinner attributed this position to psychologists such as Boring and Stanley Stevens (1906―1973) at Harvard, who developed a kind of dualistic approach to psychology, following the distinction between public and private events (e.g., Boring, 1945; Stevens, 1935).

The second form, in contrast to methodological behaviorism, was Skinner’s own behavioristic program, defined as radical behaviorism. According to him, “the distinction between public and private is by no means the same as that between physical and mental. That is why methodological behaviorism (which adopts the first) is very different from radical behaviorism (which lops off the latter term in the second”; Skinner, 1945b, p. 294). It is worth noting that Skinner’s (1945a) paper was the first time he used “radical behaviorism” in his writings (Schneider & Morris, 1987).

The first thing to note about Skinner’s classification is the pejorative connotation he attributed to methodological behaviorism. “This was never good behaviorism, but it was an easy position to expound and defend and was often resorted to by the behaviorists themselves” (Skinner, 1945b, p. 293). The task of radical behaviorism was not easy, however. The radical behaviorist faced the challenge of giving an account of private events in terms of an empirical analysis of verbal behavior, that is, an investigation of how our community has taught us to behave in relation to private events. Philosophical analysis per se would not be enough. “The solution must be psychological, rather than logical” (p. 294).

It is clear that Skinner’s characterization of methodological behaviorism depended on an ontological interpretation of his opponents’ views. At stake here was a rejection of ontological dualism (mental vs. physical) and an adoption of physicalism (everything is physical). As Skinner (1945b) himself recognized, “I contend that my toothache is just as physical as my typewriter, though not public, and I see no reason why an objective and operational science cannot consider the processes through which a vocabulary descriptive of a toothache is acquired and maintained” (p. 294).4

Some 30 years later, “methodological behaviorism” appeared again in Skinner’s work. In particular, in Beyond Freedom and Dignity (1971) he claimed that scientific psychology could not ignore the subjective domain of private events.

What is called “methodological behaviorism” limits itself to what can be publicly observed; mental processes may exist, but they are ruled out of scientific consideration by their nature. The “behavioralists” in political science and many logical positivists in philosophy have followed a similar line. But self-observation can be studied, and it must be included in any reasonably complete account of human behavior. Rather than ignore consciousness, an experimental analysis of behavior has stressed certain crucial issues. The question is not whether a man can know himself but what he knows when he does so. (Skinner, 1971, p. 181)

In this later passage, nothing has substantially changed. In place of the previous opposition between methodological and radical behaviorism, Skinner rejected the former in favor of an experimental analysis of behavior. However, the pejorative connotation of methodological behaviorism and the ontological implications of genuine behaviorism remained the same.

One can conclude that Skinner’s use of “methodological behaviorism” had a twofold purpose. First, he intended to demarcate good and sound behaviorism—that is, radical behaviorism—from a spurious form of behavioral research. Second, he insisted that the demarcation had ontological implications: accepting radical behaviorism entailed the acceptance of a monistic worldview.

Logical Positivists: Methodological Behaviorism as a Linguistic Program

Logical positivism, or logical empiricism, the philosophical program associated with the Vienna Circle (Uebel, 2019), was profoundly influential in English-speaking academic circles after many of its members immigrated to the United Kingdom and the United States (Dahms, 1995). In the latter in particular, it became the dominant view in professional philosophy, at least until the late 1950s (Hardcastle & Richardson, 2003; Richardson, 2007).

One of the central tenets of logical empiricism was the unit of science thesis and the resulting search for a universal and unified scientific language in physical terms (physicalism). This helps to explain why the logical positivists showed an interest in scientific psychology, in particular in behaviorism (e.g., Hahn, Neurath, & Carnap, 1929; Hardcastle, 2007; Smith, 1986). Beginning in the 1930s, Rudolf Carnap (1891–1970), Otto Neurath (1882–1945), and Carl Hempel (1905–1997), among others, published articles on the philosophical problems of psychology. The solution for all psychological puzzles, according to them, was a linguistic therapy of psychological concepts, that is, a logical analysis of the language of psychology (Carnap, 1932; Hempel, 1935; Neurath, 1933).

In this context, Herbert Feigl (1902–1988), a former member of the Vienna Circle, suggested “a metaphysically neutral (or, more precisely, strictly ametaphysical) logical analysis of the fundamental concepts” (Feigl, 1934, p. 420). For him, such an analysis “is one of the most important achievements of the new ‘Logical Positivism’ (Vienna School”; p. 421, emphasis in original).

In Feigl’s suggested analysis, “methodological behaviorism” took on a new meaning. It was no longer something to be avoided but the very solution for scientific psychology. According to Feigl (1934), “if psychology is to be communicable (intersubjectively verifiable), its assertions must be expressible in physical language (i.e. methodological behaviorism and the justification of the unmetaphysical component of materialism”; p. 442; emphasis in original).

Feigl’s (1934) proposal had three main components. First, by “physical language” he meant “those still rather obscure physiological processes and conditions” (p. 438). Thus, the traditional mentalistic or introspective concepts of psychology (e.g., sensation, memory, feeling) were to be reinterpreted in physiological terms. Second, his form of physicalism did not carry any metaphysical or ontological implications. It was only a linguistic—that is, semantic—approach. Third, it is important to recognize here that, for Feigl, “methodological” qualified a philosophical method, not a psychological one. In other words, methodological behaviorism was a logical analysis of psychological language or a philosophical prescription for using and interpreting concepts in psychology, not a method for studying psychological phenomena.

A similar use of “methodological behaviorism” appeared in 1956, in an attempt made by Gustav Bergmann (1906–1987), a former member of the Vienna Circle, to honor the scientific contribution of Watson to psychology.5 He argued that, “Watson is not only an experimental psychologist . . . he is also a systematic thinker, that is, a philosopher of psychology, or, as one says, a methodologist” (Bergmann, 1956, p. 265; emphasis in original). Watson had made, said Bergmann, a specific contribution to psychology, but “the contribution was not, as probably Watson thought, his materialism or metaphysical behaviorism―i.e., the thesis, which is merely silly, that there are no minds―but, rather, his methodological behaviorism” (p. 269; emphasis in original). Bergmann was making two claims. First, he was attributing to Watson two kinds of behaviorism. Then, he defined the first kind in terms of an ontological thesis, namely, that minds do not exist, in the sense of “that ancient notion which is traditionally called a substantial mind” (p. 266). However, because Bergmann was placing methodological behaviorism in opposition to metaphysical behaviorism, we need a definition of methodological behaviorism to make Bergmann’s claims understandable.

Bergmann understood methodological behaviorism in terms of a central thesis: “It must in principle be possible to predict future behavior, including verbal behavior, from a sufficiency of information about present (and past) behavioral, physiological, and environmental variables. This is the thesis” (Bergmann, 1956, p. 270). This formulation is too synthetic, however, and demands further clarifications.

Bergmann’s interpretation implied, in the first place, a distinction between common sense and scientific discourse. For Bergmann, methodological behaviorism did not entail the elimination of traditional terms such as “personality,” “volition,” “state of mind,” and the like. This is popular parlance. Scientific psychology, however, should not content itself with such vague terms as if they referred to mysterious entities or processes. The question, then, is to coordinate with any statement in which these terms occur another one that “mentions only behavioral, physiological, and environmental items” (Bergmann, 1956, p. 270). In this way, however useful mentalistic terms can be in the early stages of scientific psychology, psychological statements containing them would be gradually paired with corresponding statements referring to behavioral, physiological, or environmental variables.

The second implication concerned the programmatic aspect of methodological behaviorism. Bergmann (1956) was cautious enough to mention this: “We accept something as a program to be acted on and which, therefore, after a reasonable time may be judged by its fruits” (p. 270). For him, methodological behaviorism was not a psychological theory, but a philosophical program, an a priori model and a frame of reference for a posteriori psychological theorization. “Methodological” denoted a philosophical method, not a psychological one. It was an expression of optimism regarding the power of logical analysis and behaviorist language, not a prescription for studying human behavior. However, it might ultimately transpire that such optimism was illusory and psychological science might fail to advance.

The programmatic aspect of methodological behaviorism allows us to understand the third implication of the main thesis, namely, the refusal of what Bergmann (1956) called “physicalistic behaviorism” (p. 271). For him, physicalism in psychology―understood as the elimination of traditional psychological words and the reduction of psychological vocabulary to a limited set of mechanical and physiological terms―was too strict and narrow to succeed. In practical terms, Bergmann was rejecting the conflation of behavioral research and physiological research or, as he put it, “physiological reduction” (p. 272). This was why he insisted that behavioral variables were not the same as physiological or environmental ones, although they could be shown to be related.

Taken together, the main thesis and the three corollaries were in perfect agreement with Bergmann’s previous attempts to deal with the theoretical and conceptual problems of psychology in terms of a behavioristic reformulation (e.g., Bergmann, 1940a, 1940b, 1951). This was why we can say that, for him, methodological behaviorism is only a name for the application of logical positivism to psychology. Logical behaviorism and methodological behaviorism were, in Bergmann’s conception, identical. In other words, methodological behaviorism was a logical-linguistic prescription for using and interpreting psychological language. It was a refinement and further development of the first attempts to apply logical positivism to the situation of scientific psychology, following the path of Carnap, Feigl, Hempel, Neurath, and others.

This justifies Bergmann’s particular use of the concept. He not only intended to defend Watson against himself (to save him from metaphysical naivety, conceptual confusion, and physiological reduction), but also to defend behaviorism as a sound logical prescription for scientific psychology. Watson was, so to speak, Bergmann’s case study. However, behind his appreciation of Watson’s enduring contributions to psychological science lay a defense of a particular philosophical program, which he called methodological behaviorism.

The Impact on the Literature

The definitions presented and discussed here indicate that although the concept of methodological behaviorism appeared for the first time in 1923, it was changed substantially by subsequent users of the term. Now, it is necessary to show that those definitions and uses were adopted elsewhere in the literature, thus showing their historical impact.

Lashley’s (1923) proposal, for instance, was mentioned by other authors in the 1920s, 1930s, 1940s, and 1970s (e.g., Diserens, 1925; Ewing, 1944; Hanson, 1971; Leeper, 1931; Mace, 1948; Mischel, 1970).

As for Skinner (1945a, 1971), his demarcation was the basis upon which Ducasse (1951), Copeland (1964), and Ellis (1972) discussed their positions. Moreover, from the 1970s onward, a new generation of behavior analysts took off from Skinner’s idea to build their own proposals (e.g., Day, 1983; Leigland, 1997; Moore, 1975, 1980, 2001, 2011/2012, 2013).

Both critics of and sympathizers with logical positivism used the conception of methodological behaviorism developed by Feigl (1934) and Bergmann (1956), especially Bergmann’s. In this sense, methodological behaviorism was discussed in terms of the logical analysis of language. This was the case with Brodbeck (1955), Spence (1957), Miller (1959), Grossmann (1960), Brody and Oppenheim (1966), and Henle and Baltimore (1967).

Our list of citations is not, and does not intend to be, exhaustive. Its purpose is only to illustrate the presence of those three conceptions of methodological behaviorism in the subsequent literature up to 1973.

Concluding Remarks

Our analysis shows that the first conceptions of methodological behaviorism were based on a methodological component and an ontological component. However, because these two aspects were understood in different ways, each definition reflected a particular understanding of them. For example, “methodological” could refer to the concrete methods of psychological research (e.g., experimental, descriptive), but also to a philosophical method of linguistic analysis (e.g., physicalism, operationism). At the same time, methodological behaviorism was seen by many as related somehow to particular metaphysical views, such as ontological dualism (mind versus matter) or monism (everything is physical). Here, it is curious that the same behaviorist (e.g., Watson) could be labeled a dualist by some and a monist by others. Moreover, with regard to metaphysical assumptions, methodological behaviorism could also be seen as neutral (ametaphysical), as prescribed by Feigl and Bergmann.

In addition, those regarded as representatives of methodological behaviorism were many and various. For instance, Watson was sometimes considered the methodological behaviorist par excellence; for some, he was both a methodological and a metaphysical (dualist or monist) behaviorist. Others, however, claimed that he was not a methodological behaviorist at all. Yerkes was seen by Roback as the prototypical methodological behaviorist, whereas for Skinner this place belonged to Boring and Stevens. And the list goes on.

As for its uses, the concept of methodological behaviorism has primarily had a demarcating function. Whatever the context in which it appeared, it served to differentiate good from bad or genuine from spurious behaviorism. This function notwithstanding, methodological behaviorism appeared on both sides of the spectrum: it was either the ghost to be avoided or the remedy to be adopted. Whereas for Bergmann it was the remedy for metaphysical behaviorism, for Skinner it was an opportunistic and defective form of behaviorism, against which radical behaviorism had to be developed. Its value, it seems, was in the eye of the beholder.

To sum up, the concept of methodological behaviorism involved a metaphysical discussion (monism, dualism, or neutrality) and a methodological prescription (philosophical or psychological methods). Determining who qualifies as a methodological behaviorist depends on how one understands the metaphysical and methodological compromises of each author in question. Likewise, depending on one’s metaphysical and methodological assumptions, methodological behaviorism appears as either positive or negative.

This leads us to the central issue raised in the introduction. The polysemic nature of the concept is not a recent feature that emerged out of deviations from its original meaning or use. It was there from the beginning.

Thus, it is small wonder that debates on methodological behaviorism in the contemporary literature repeat the same patterns of definition and use of the concept, although some behavior analysts have added new aspects to it (e.g., Day, 1983; Leigland, 1997; Moore, 1975, 1980, 2001, 2011/2012, 2013). The concept still serves to demarcate what good behaviorism is or should be.

In this sense, our historical analysis can contribute to a better understanding of the conceptual confusion regarding methodological behaviorism in the literature. Knowing its historical origins and early developments can give the contemporary reader a broader frame of reference for evaluating the implications of the term’s use in each context.