Scientific realism,or the risks of importing research models into psychology

In “Scientific Realism and the Issue of Variability in Behavior,” Arocha (2021) develops an acute critique of “the standard model of current research practice in psychology” (p. 376), specifically aiming at its positivist roots. The author does so by dissecting five unwarranted assumptions behind this model: from portraying human behavior as a random variable to the gratuitous use of population statistics. These assumptions, Arocha argues, lead to issues like dismissing inner processes driving behavior and reaching conclusions that tell us nothing about individuals—the subject matter of psychology.

To address these issues, Arocha (2021) proposes that behavioral research should adopt a nonpositivist philosophical basis, namely scientific realism (see Bunge, 1974–1989); which—the argument goes—could help the discipline to become an “advanced” scientific field. Arocha (2021) develops two examples to portray how a metatheoretical stance like scientific realism translates into actual research in psychology: perceptual control theory (PCT) and observation-oriented modeling (OOM). The former proposes that: “the method for the study of behavior is based on the dynamics of negative feedback control” (p. 381). Therefore, when scientific realism is taken to practice, human purpose, for example, could be understood and studied “as a control system.” As such, “the reference signal (some desired state of a perception, a want, an intention, or a goal), coming from other control units higher in the nervous system, is compared to the input signal, which stands for some current perception” (p. 381).

Arocha (2021) makes a strong case against the use of positivist ideas in behavioral research, while also proposing a philosophical approach that addresses the issues generated by them. Behind his argumentation, however, it is implied that solving the negative consequences of positivism is enough to make scientific realism a fit metatheoretical framework for the whole of psychology. I argue, on the contrary, that scientific realism is not fit for becoming that (certainly much-needed) philosophical basis for a simple reason: because it is open to reducing the discipline’s subject matter—the human person—in order to make it fit with parsimonious but mechanical models that have been fruitful in other disciplines, such as information science and engineering (see Dijksterhuis, 1961). In doing so, Arocha’s proposal repeats a critical issue of positivism, one that has plagued the history—not to say the present—of psychology. Let us consider three historical examples of how the allure of adopting models from other sciences to tackle the complexity of psychology’s subject matter turned instead into an effort for dismantling the discipline and assimilating it into other fields.

For his history of the origins of cognitive science, Jean-Pierre Dupuy (2009) chose a blunt title: The Mechanization of Mind. Why so? Because “the aim of cognitive science always was—and still is today—the mechanization of mind, not the humanization of machine” (p. xi). According to him, the cognitivist project involved modeling the mind as a physical mechanism, capable of computing algorithms, that “appears to us to contain meaning, finality, directionality, and intentionality” (p. 4) but is ultimately bound and explained by the laws of physics. While Dupuy spares no words in criticizing the antihumanism behind cognitive science, he also provides a comprehensive account of how these ideas developed. To achieve the latter, the author puts special emphasis on the role played by scientific models. Models in science typically begin as an instrumental but conscious simplification of the phenomena described, aimed to open a path of systematic inquiry. Yet in the case of cognitive science, these models became the very image to which mental phenomena ought to comply:

Although the scientific model is a human imitation of nature, the scientist is inclined to regard it as a “model,” in the ordinary sense, of nature. Thus nature is taken to imitate the very model by which man tries to imitate it. (Dupuy, 2009, p. 30)

Therefore, once the mechanical (and later, computational) model of the mind was conflated with the phenomenon of mind by early cognitive scientists, they saw no point in pursuing any other form of scientific inquiry devoted to this subject. In other words, any other psychological understanding of the mind was deemed unnecessary—not to say unscientific. Thus, the thrilling possibility of bringing advances in information science and engineering into psychology became an ultimatum against any area of the discipline that did not want to comply with the mechanical model of the mind.

Far from an isolated misstep, the tendency of psychologists to naïvely import models from other disciplines could also be seen in the adoption of operational definitions from physics (see Koch, 1992), or modeling the mind after computers (see Hurtado, 2017). In these cases too, psychology seemed to have “solved” methodological and theoretical problems at the cost of fitting human behavior, cognition, emotion, language, and interaction to physical and mathematical models meant to explain other phenomena. Why naïvely? Because, as these examples also show, researchers in psychology ignored—willfully or not—well-known limitations of the models imported; be it Bridgman’s intent behind operational analysis (Koch, 1992) or Turing’s halting problem (Hurtado, 2017); limitations that rendered these models ultimately incapable of addressing the variability of individual persons regarding, for example, meaning-making processes.

Psychology’s troubled history should not only be a reminder about the risks involved in embracing oversimplifying models of persons. It should also be a warning about the metatheoretical approaches that allow this kind of model to take root in research practice. This is why the openness of scientific realism to conduct psychological research based on models depicting human behavior as information systems should be a worrying matter. As seen, we psychologists are prone to forget about meaningful actions in order to fit them within models simplifying them to the point where they are no longer recognizable. This is how the voluntary choice of avoiding an annoying foreman (see Runkel, 2003, p. 67) ends up being seen as the same as the negative feedback mechanism behind a thermostat. As a schematic comparison the former is harmless, but it becomes problematic when this model reduces human choice to nothing more than negative feedback. By doing so, we preclude any meaningful understanding of why those persons chose to avoid the foreman; ironically dismissing other inner (and also social, environmental) processes driving their behavior.

As Arocha (2021) himself notes:

psychology’s object of study is the individual, who has characteristics not possessed by corn plants or manufactured devices. In particular, they are moved by their own inner processes [emphasis added], which is a critical distinction between human and animal behavior and that of inanimate objects. (p. 384)

The way in which those inner processes are modeled, however, is crucial for our understanding of human experience and behavior. When, under scientific realism, those processes are equated to controlling information systems for all practical purposes, the individual is at risk of being reduced (again) to inanimate, manufactured devices. Even if Arocha’s proposal offers important contributions (embrace individual variability, avoid statistics misuse, among others), the former consequence should be worrying for anyone who is interested in conducting psychological research without sacrificing the richness of human living—even if doing so fails to meet the standards set by sciences devoted to understanding different phenomena.