Wertheimer, M. (1988). Obstacles to the integration of competing theories in psychology. Philosophical Psychology, 1, 131-137.

Obstacles to the integration of competing theories in psychology

Michael Wertheimer

ABSTRACT: Among the obstacles to the integration of competing theories in psychology is that it is unclear why and how they differ. Some thoughtful speculations have been offered for why they differ, but they remain preliminary. A plethora of schemes has been proposed for analyzing how theories differ, but there is no convincing basis for choosing among these schemes. Furthermore, several kinds of relationships can be identified between apparently competing theories, each of which has implications for the potential integration of those theories. If theories X and Y seem to be competing, they may (a) turn out to be intertranslatable, in which case they are not really competing; they may (b) be truly contradictory, and therefore logically impossible to integrate; and may (c) be apparently but only superficially mutually translatable, but with the translation doing violence to at least one of the theories; or they may (d) be mutually irrelevant, so that there is little point in trying to integrate them, and any attempt at integration is apt to be little more than a minimally organized list.

William James wrote in his Pragmatism (1907) that philosophy is a quest for unity, but a quest that may forever be frustrated. We may have to learn to be content with a world that is imperfectly unified, and that perhaps always will remain so. The world we experience is so complex, so multifaceted, that there can be no single orderly scheme with which all experience can be reconciled. James recognized that though such schemes may satisfy our need for elegance they are inevitably thin and brittle, and incomplete; any unitary worldview will constantly be forced to rationalize to reinterpret, or even to disregard features that common sense takes for granted (James as interpreted by Viney, 1986). More recently, Wachtel ended a provocative article (1980, p. 408) with a plea for "opening the door to a greater diversity of method and mode of thought." And Brewster Smith (Wertheimer et al., 1978) a few years ago argued for the survival value of methodological and theoretical diversity: immutable species perish while natural selection can permit variants of a mutating species to survive and indeed to improve. So there are ontological, epistemological and evolutionary arguments for pluralism. Such arguments are certainly obstacles to any effort to unify any field, including the theories of psychology. [p. 132]

But James also did acknowledge that our need for elegance makes us seek unified schemes for the pluralistic world of human experience. Unification of a scientific field, or of all science, indeed of all knowledge has been a persistent venerable human hope. The world would be more satisfying and elegant if it really were not a multiverse but a universe. One exemplar of the urge for integration is the Society for Unification Psychology, founded by Arthur Staats, which is intended to address the growing issues of diversity and fragmentation in psychology (Staats, 1986b, p. 70). Salvator Maddi and Alan Boneau recently instituted an annual competition through the Division of General Psychology of the American Psychological Association for the best paper that unifies or integrates several disparate aspects of psychology; somewhat paradoxically, they named it the William James award even though James would not have supported the uninomic psychology movement (though with his pluralism he certainly wouldn't have discouraged it either).

Without addressing the prior question of whether theoretical unity is possible, or even desirable, one can still argue that if unity is desired, it would be useful to find out in the first place why and how theories in psychology differ. Before that is known, it would be futile to try to integrate apparently competing theories.

Why do psychological theories differ? As Staats (1986c) has pointed out, among the causes that have been identified for the fragmentation of psychology are an emphasis on empiricism (Bevan, in press), and on novelty (Bevan, in press; Maher, 1985; Staats, 1983), the pressure to establish a unique position (Maher, 1985; MacIntyre, 1985), and to publish experimental studies (Wachtel, 1980), as well as other methodological, organizational, and social factors (Bevan, in press; Staats, 1983; Wachtel, 1980). As Wachtel (1986) observed, "Psychological theories tend to be based upon a limited set of observations, deriving from a preferred investigative paradigm... [D]ifferent investigators' conceptions of what 'doing research' consists of tend to include certain lines of investigation and (though not usually explicitly) to exclude others." Fishman (1986b, p. 20) surmised more broadly that "psychology's fragmentation [may be] a natural and permanent state of the cultural, historical, and linguistic embeddedness of human experience and behavior." One could also elaborate on inherent incompatibilities among certain kinds of theories (e.g. systemic versus connectionistic), and on a host of less internal, and more external, factors such as who happens to be popular at a given time in a particular field (for most psychological theories do not outlive their formulators), where research funds happen to be more readily available (as national priorities waver among fads -'research applied to national needs,' reading, cognitive science, military projects) so that more graduate students can be supported, administrators have more research overhead funds to be used for discretionary purposes, and more publications can be generated to enhance the likelihood of promotion and extramural recognition, and still other quite external but very real pressures on the theorist. The current rift in APA between research/academic and applied/practitioner interests (which not accidentally parallels very nicely the 'two cultures' that Gregory Kimble (1984), identified a few years ago) is another instance of the external pressures impinging on the psychological theorist, as [p. 133] competition for scarce resources intensifies with the growing scarcity of the resources. All of these factors constitute obstacles to the integration of diverse theories; and anyone wishing to unify psychology must address them. It is likely, incidentally, that these factors are not additive, but that many of them are systemic and interactive; furthermore, after a thorough analysis of them has been completed, an awesomely massive social program is likely to be required if there is to be any chance of changing them in such a way as to produce unifying rather than the current fragmentary efforts.

In addition to discovering why theories differ, it will also be necessary to specify how theories differ, if we are to try to integrate them. There is no easy answer to this question either. Many different schemes have been proposed for how to analyze the differences among theories Gregory Kimble (1984), as mentioned a moment ago, talked of psychology's 'two cultures', and Loyuse Silvern (1986) contrasted the 'received' (experimental) with the hermeneutic approach. Muenzinger (Wertheimer, 1987) suggested that psychological theories vary on such characteristics as the preferred unity of study and the preferred method, and whether the orientation should be idiographic or nomothetic. Another set of dimensions proposed by Muenzinger (Wertheimer, 1987) asks such questions as the extent to which theories are physiologically oriented, make an atomistic assumption, are quantitative, and adhere to the scientific materialist tradition. Watson (1967) proposed eighteen dimensions such as determinism versus indeterminism, empiricism versus rationalism, and molecularism versus molarism. Wertheimer (1972) proposed ten so-called 'fundamental issues in psychology', including richness versus precision, theory versus data, and andsum versus transsum. Fishman (1986a) generated a list of ten somewhat different questions which could be asked about any psychological theory; he called his dimensions 'paradigm characteristics' and used them to compare an experimental, a technological and hermeneutic paradigm.

We do not want for schemes of dimensions along which theories should be, or could be, compared; but one comes away from this literature with the impression that rather than a clean, useable single analytic tool, it yields little more than a profusion of confusion. Fragmentation and diversity characterize even this small subdomain of metatheory. Incidentally, while it is possible to translate at least some of the specific dimensions of one writer in this field into the dimensions of another writer it is not clear how much or what is gained by such an exercise: what criteria should be used for deciding which dimensions are apt to be useful? A democratic one, such as how often a dimension is mentioned by various writers? Not likely to be a viable criterion. Further aren't at least some of the schemes systemic, so that what is ostensibly the same dimension in two schemes may actually be used rather differently in the two? And there is the further difficulty that at a deeper level than that intended by the scheme-writers there are other, possibly more fundamental and pervasive, dimensions along which theories and theorists differ that are not covered by the extant schemes: aside from the scope, goals and terminology of theories, there are such basic matters as rock-bottom epistemology and methodology, values, and, perhaps most important, the implicit assumptions that are unexamined but taken for granted about metatheory, ontology -and everyday life (Wertheimer, 1986). [p. 134] As Gergen (1973) pointed out, such matters are profoundly tinged by the theorist's sociocultural context.

Nevertheless, perhaps we can, at least in specific cases, go about the task of integrating theories A and B without regard to the overall unification effort: just try to pull theories A and B together without taking on the theoretical synthesis of all psychological theories. One such instance is the classic work of Dollard & Miller (1950), who translated much of Freud's theory into Hullian terms. They succeeded in showing that the principles and concepts of one prominent psychological system could be translated into the terms and ideas of an ostensibly entirely different system.

But this example raises a significant issue: What does the phrase, 'the integration of competing theories', really mean? It could be argued that translation of one theory into terms of another that initially appears entirely different suggests that the theories probably weren't really conflicting in the first place, but only different sets of terminology for basically the same theoretical schema. So translation may not be integration in more than a rather trivial sense. Perhaps the extent to which mutual translation becomes impossible with diligent effort is an indicator of the extent to which the theories are actually competing (or mutually irrelevant, a category to which we shall return in a moment). Hull and Tolman, a generation ago, did disagree on the role of need reduction in learning, even though many other aspects of their theories could probably be translated without much violence. True competition such as one theory saying 'X is necessary', and the other saying 'No it isn't', can generate research, as was the case for the Hull-Tolman debate, with some fascinating studies of latent learning, place versus response learning, and so on. This is a case of competing theories where integration is explicitly impossible on logical grounds. The optimal arena for their competition is the realm of empirical data, even though, historically, the results of the empirical competition have rarely turned out to be definitive (even in the Hull-Tolman skirmish).

So there are apparently competing theories which at first sight are quite different, but which turn out to be mutually translatable, and therefore were not really competing after all. And there are apparently competing theories which turn out on analysis to be actually competing, that is, irreconcilable because they contain mutually contradictory components or assumptions. Before considering two other kinds of apparently competing theories, it should be mentioned that the Hull-Tolman conflict is by no means the only instance of this genuinely contradictory category. Any time a theory is introduced that defines itself explicitly as disagreeing in certain concrete respects with a prior theory, integration becomes logically impossible. If theory X says 'A', and theory Y says 'not A' they can't be integrated. Freud and Jung on the role of sexual factors in the etiology of psychopathology is another instance; an atomistic, connectionistic theory as contrasted with a systemic or field approach is another.

A third category of relationship between two apparently competing theories that impacts on their potential 'integrateability' could be characterized as superficial but disastrous. As a graduate student coming out of a Gestalt undergraduate orientation, the present writer encountered Skinner's idea that the probability of a [p. 135] response in a particular situation is enhanced if that response is followed by a reinforcer. It reeked of a blind mechanism that he had been taught is abhorrent and just plain wrong. So he tried to translate the Skinnerian formulation into the cognitive terms with which he was more familiar -and that were more compatible with his Psychologieanschauung. The animal learns what leads to what; it discovers that what it did made the reinforcer occur, so if it once again wishes to have that reinforcer occur in the same situation later on, it is reasonable for the animal to perform that same act again, with the anticipation that it might once again make the reinforcer appear. He had recreated a conception similar to Tolman's 'sign-Gestalt-expectation', and was content with it. But of course it was laughed down by his fellow graduate students (although Skinner was too polite to laugh openly at the naive effort), since among the central feature of Skinner's approach is the explicit denial of such subjective fictions as 'expectations', 'conscious anticipation', or 'cognition'. The attempted integration had done theoretical violence.

There is a fourth category, which is by far the largest: theories that are unrelated, mutually irrelevant, orthogonal, or tangential, theories about entirely different sets of phenomena, which come form entirely different sets of assumptions. Of course that makes them not genuinely 'competing' theories. Yet almost every introductory psychology text does try in a very loose sense to 'integrate' them, at least by mentioning them successively when the topics to which they are relevant come up. With the exception, perhaps of D. O. Hebb's textbook (1958) and more recently of Henry Gleitman's (1981), these efforts must be considered superficial, though perhaps harmless. John Gosling (1986), though, has proposed a disciple-wide integration in a serious and responsible manner.

Indeed, there have been many proposals for how to integrate competing theories in psychology, and this discussion will end with very brief comments upon some of them. Gosling (1986) proposes an analysis of the types of knowledge currently represented in existing psychological theory, followed in turn by an analysis of the nature of the theoretical devices by which knowledge is represented in current psychological theory and a consideration of their limitations. Gosling's vision may be inspiring, but it is likely to founder at the outset, because there will be irreconcilable differences among psychologists on what the specific knowledge requirements of psychology are. Staats (e.g. 1983, 1986a, b, c) has pleaded repeatedly for efforts to unify psychology's chaotic diversity, but the specific details of how concretely to go about such an endeavor are not yet fully clear. Fishman (1986c) wants us all to agree on a 'coherent epistemology', but does not tell us how to go about achieving this consensus. Minke (1986) proposes that systematization could be achieved by an historical analysis of definitions, but gives us little advice about how to achieve definitional consensus among psychologists. Wachtel (1986) recommends greater emphasis in training programs on theoretical inquiry and more support from granting agencies for theoretical work, recommendations which warrant the enthusiastic endorsement of all psychologists, but which [p. 136] advocate an external support system for more responsible theoretical work without specifying how this system will in fact enhance psychology's theoretical unity.

One can favor general psychology, and careful theoretical and philosophical work, without endorsing the possibly unrealistic and perhaps even undesirable goal of unifying all of psychological theory. But it is also conceivable that genuine integration of competing theories may be inherently impossible. In summary, several kinds of relationships can be identified between apparently competing theories in psychology, each of which has implications for the potential integration of those theories. If theories X and Y seem to be competing, they may (a) turn out to be intertranslatable, in which case they are not really competing; they may (b) be truly contradictory, and therefore logically impossible to integrate; they may (c) be apparently but only superficially mutually translatable but with the translation doing violence to at least on of the theories; or they may (d) be mutually irrelevant, so that there is little point in trying to integrate them, and any attempt at integration is apt to be little more than a minimally organised list.

Acknowledgment

This paper is based on a contribution to a symposium on 'Uninomic Psychology: Areas of development in creating a unified science' at the 1986 annual convention of the American Psychological Association.

References