The methods and principles of the natural sciences are not a diet we need to follow: A close look at the terms of the question

Abstract

In this article, I discuss each of the elements of this special issue’s question, that is, “should,” “psychology,” “follow,” “the methods and principles,” and “the natural sciences,” and first argue that the natural sciences are many and diverse, and the choice to emulate them would still leave plenty of room for variety. There are, moreover, good ontological reasons to resist the urge to restrict what we call “psychology” to the study of human life with the “methods and principles of the natural sciences.” Psychologists should feel free to adopt and adapt (rather than follow) what has been developed in other fields of research in terms of principles, methods, techniques, and instruments. That includes fields of research other than those in the natural sciences.

The question that is the topic of this special issue of Theory & Psychology (Eronen et al., 2024) is a familiar one. There is a long history of philosophers, psychologists, and other scholars making statements about the relation between psychology and the natural sciences. Canonical examples are Immanuel Kant (1786/2004), who did not think psychology could be a natural science, and John B. Watson (1913), who did think it could be and stressed that it had to be. One way to get further in this debate, I believe, is to look closely at the terms in which the issue is addressed. Different terms construe the issue in different ways. Kant (1786/2004), for example, was specifically discussing the question whether “the study of the phenomena of inner sense and their laws” could be a natural science (p. 7). Thus, Kant was discussing the possibility of a science of consciousness, but that is a particular and currently rather uncommon view of what psychology should be about. ¹ Before we make a decision about the weighty issue of whether psychology should follow the methods and principles of the natural sciences, we should take a good look at the terms in which the question is posed. First, we can avoid confusion if we realise that words are sometimes used in different ways. Not everyone means the same thing by “the natural sciences,” for example. Second, these different uses can relate to actual variety in what the terms refer to. What is commonly referred to as “psychology,” for instance, is in some ways a very diverse discipline. Third, a close look at the terms is necessary because words not only describe, but also make worlds. Classification matters. “Science,” for example, remains an extraordinarily powerful word, and wielding the word science in the right way and in the right circumstances can make things possible. We should investigate how these terms are used, by whom, and to what end, and if and how they could be used differently, by other people, for other, perhaps better purposes. I will discuss each of the elements of the question in reverse order: “the natural sciences,” “methods and principles,” “follow,” “psychology,” and “should.” It is not my intention to criticise the question. It is an excellent question, precisely because it raises so many questions itself.

The natural sciences

What is science? What is natural science? 20th-century philosophers of science have tended to separate the question of what natural science is into a normative issue (what should it be?) and an empirical issue (what is it in practice?). A distinction between the rational, intellectual, internal aspects of science—its methodology, its context of justification, on the one hand, and its irrational, external aspects, its social and psychological processes, its context of discovery, on the other hand, was a leading principle of the work of philosophers like Hans Reichenbach and Karl Popper. Imre Lakatos (1970), a late exponent of this tradition, noted that in his own “methodology of scientific research programmes” (p. 91), many aspects of science that had been external in previous methodologies had become internal and thus rational. His methodology painted a considerably more complicated picture of the history of science than, for instance, the falsificationism of Popper. Nevertheless, Lakatos too stressed the necessity of the distinction between the internal and external history of science.

Sociologists and historians of science have since abandoned the conceptualisation of the history of science in terms of internal and external factors, or, more productively, have critically analysed the role of this distinction in the historical study of science (Shapin, 1992) and have investigated the way the boundaries of science are drawn in practice (e.g., Gieryn, 1983; Shapin & Schaffer, 1985). Thomas Gieryn and other sociologists and historians of science have described the work required to demarcate science from nonscience, and the work that this boundary does. ² Defining what science is and what it is not is both a local, contingent achievement and one that has practical consequences in terms of who is included and who is excluded from science, who gets funding, who shares in the epistemic authority of science, and so forth.

The same empirical sensibility that has turned the boundaries of science from a normative starting point into a topic of historical and sociological inquiry has also made the diversity of and the divisions in science noticeable. A book title like The Disunity of Science: Boundaries, Contexts, and Power (Galison & Stump, 1996) is typical of the emphasis on diversity in science studies of the last four decades. Epistemic Cultures by Karin Knorr-Cetina (1999) is another example. Based on her ethnographic study of their respective laboratories, Knorr-Cetina describes high energy physics and microbiology, both undeniably natural sciences, as two cultures, in which knowledge is produced in very different ways.

In as much as a general character is ascribed to the sciences in works like these, it refers to what scientists do in practice rather than what they should do in theory. Knorr (1979) claimed that “the experience of the laboratory proves that the model of hypothesis-testing spread by methodology textbooks and theoretics of the social sciences is inadequate” (p. 362). Instead, science is a process of tinkering, in which scientists employ whatever “assets” they have at hand—instruments, materials, people, ideas—in order to find a solution to a problem that will secure them a publication (p. 362). Such a claim rejects a normative methodology of science (Popper’s falsificationism in this case), and replaces it with a general characterisation of science in practice that leaves room for a wide variety of specific methods. One consequence of the diversity of science in practice is that if psychology were to decide to emulate “the natural sciences,” guided by some idea of what it is in general to be a natural science, it would still have to discover its own way of being a natural science.

Methods and principles

The focus on science in practice has put in doubt the usual ways of demarcating it from other parts of culture. A tenacious topos in this respect is “the scientific method,” and one thing to note about the phrase “methods and principles” is that it avoids this notion and instead combines two plurals. That is a good thing. There is no definition of the scientific method that is specific enough to be useful and at the same time general enough to cover the work of all or at least most of the people we would like to call scientists (Cartwright et al., 2022). Susan Haack (2012) puts it succinctly: “there is no ‘scientific method’ used by all and only scientists” (p. 88). When people speak of the scientific method, they usually mean some form of the hypothetico–deductive method: science proceeds by formulating hypotheses and testing them. But scientists do much more than testing theories, and vice versa, it is not uncommon for other people, nonscientists, to put their predictions to the test.

It is worth noting here that Popper (1976), who is often referred to by proponents of the scientific method, had such a broad notion of what science is that in essence the social sciences already have exactly the same method as the natural sciences: it “consists in trying out tentative solutions to certain problems: the problems from which our investigations start, and those which turn up during the investigation” (p. 89). Scientism, which he defined as urging the social sciences “to learn from the natural sciences what scientific method is,” is a mistake (p. 90). That is, for Popper, the question posed in this special issue is a distraction. In as much as it is a science, psychology already is a science in the same way as the natural sciences are. It tries to solve problems, both theoretical and social, by conjecture and refutation. And if it does not do that, it simply isn’t a science.

But Popper’s notion of science covers over the considerable diversity of the work of scientists, including the fact that, as Knorr-Cetina (1999), Lakatos (1970), and many others have noted, they are not always trying to refute their own or other researchers’ conjectures. This also complicates the demarcation of science from other domains. In his discussion of the internalism/externalism debate in the history of science, Shapin (1992) offered the suggestion that

it might plausibly be argued that what we are accustomed to call “science” . . . is a diverse set of cultural practices, which may not have common methods, conventions, or concepts, or at least common features distinguishing them from “non-science” or the common culture. (p. 346)

That is not to say that science is indistinguishable from, say, religion, but it means that there is not one method or principle that differentiates science from all other cultural domains, and that it will share features with many of them.

The plural “methods” suits the diversity of the work of scientists better than “the scientific method.” The diversity of the natural sciences and their research methods poses a problem for those who insist that psychology should follow them: which of those methods should we follow? The methods that researchers in one field employ are not necessarily of use in another. Electrophoresis and infinite descent are important methods in their respective fields (microbiology and number theory), but not in other disciplines, and neither of them is ever likely to play a role in the work of psychologists. They are not methods psychologists will want to “follow.”

Philosophers and historians of science have also looked for other ways to chart the wide array of ways of doing in science than as a collection of methods. Hasok Chang (2022) has proposed the concept of an epistemic activity, defined as “a programme of action designed for the achievement of a recognizable aim” in the context of the production, evaluation, and use of knowledge (p. 15). Chang emphasises that what we identify as an epistemic activity is to some extent a matter of interpretation, guided by what we see as the aim of the activity. Epistemic activities are, moreover, always analysable into simpler, constituent activities, which can be constituents of other activities as well. Thus, the work of scientists is not made up of atomic, unchangeable methods, but of much more fluid, changeable epistemic activities. Maybe there aren’t definite methods for psychology to follow.

Principles are not the same as methods. Methods are ways of doing things, a means to an end, whereas principles are starting points. There are two kinds of principles in science. The first kind are described in handbooks like Principles of Organic Chemistry (Ouellette & Rawn, 2015). The book describes the basics of organic chemistry in a manner that is accessible to students, teaching them for example that “(t)he physical properties of haloalkanes depend on the lengths and strengths of the carbon-halogen bonds” (Ouellette & Rawn, 2015, p. 85). Ouellette and Rawn do not list the principles, but most of the chapters end with a one- or two-page summary of reactions. Principles in this sense are central to Kant’s (1786/2004) conception of science: a science is “a whole of cognition ordered according to principles” (p. 3). They are specific to each scientific field, and they are not followed but arrived at through research. ³ As a natural science, psychology would have to discover its own principles of this kind.

A second meaning of “principle” applies to science in general. These are the starting points of science as a whole. Kant’s definition of science, with its emphasis on the systematic ordering of rational principles, is itself a principle in this sense. Verificationism and falsificationism were proposed as scientific principles in the 20th century. This sort of principle can be used for demarcation and for defining an identity. The boundaries that are drawn with these principles can then be policed. If your hypotheses are not falsifiable and you do not put them to the test, you are not doing science. Astrology is not a science, astronomy is. Astronomers can apply for funding at the National Science Foundation, astrologists cannot. I will return to this policing in the final section of this article.

Whereas methods are plentiful, principles are few. Perhaps this is in the nature of principles. It would not make sense to organise one’s life or work on the basis of hundreds of principles. There were only 10 commandments. Robert Merton (1942) thought that there are only four norms that define science (p. 115). By defining science on the basis of just a few principles (or even just one) we can draw straight, unambiguous lines around (a) science. Principles function as starting points that are mostly not questioned and furnish an identity. As Roger Smith (2005) noted, it is the very fact of not questioning certain assumptions that allows a group of researchers to get on with producing knowledge. Perhaps, then, the principles of the natural sciences offer better guidance for psychology to follow than their multiple methods.

Follow

Following principles is more complicated than it sounds. Dominant principles or norms tend to be shadowed by alternatives. Merton (1942) described four norms of science, but later came to realise the existence of contradictory sets of norms in every social institution, including science. To deal with the contingencies of their work, scientists alternate between norms and counter-norms. Drawing on interviews with scientists involved in NASA’s lunar programme in the early 1970s, Ian Mitroff (1974) described four counter-norms that complemented Merton’s original four of communism, universalism, disinterestedness, and organised scepticism. For example, these scientists rejected the ideal of the disinterested scientist as a myth, but on the other hand also criticised colleagues when they seemed too committed to their pet theories. According to Mitroff (1974), “it is the interplay between personal and impersonal forces that makes for the rationale and ultimate rationality of science” (p. 580).

Thus, different norms can apply to a situation, depending on what meaning we give to that situation. Nor is it always clear what a principle, or norm, or value means in a specific context, particularly if that context is new. The Mertonian norm of communism (or “communalism” in later versions of Merton’s essay) has to be reinterpreted in the light of the emerging culture of open science (Hosseini et al., 2022). Merton meant that scientists have to share their findings, but does the same apply to research materials, as open science enthusiasts would like? And what does communism mean in a digital technological environment in which whatever we share quickly becomes data mined by the proprietary algorithms of large, decidedly noncommunist corporations? As Wittgenstein argued, following a rule is a practice, and when the world changes a new practice must develop in which following a particular rule such as scientific communism has meaning in the actions of scientists. Science, with its shifting research fronts, operates in a continually changing world. It creates new sociotechnical landscapes (Latour, 1993). In the resulting situations of uncertainty, methodological rules cannot not function as unambiguous prescriptions. Instead, their meaning in this new context has to be reinterpreted (Mulkay & Gilbert, 1981). Rather than following principles, scientists have to find them again and again in each new field of research, as Cartwright et al. (2022) argue in their book, The Tangle of Science. Objectivity is an important principle of scientific research, but what being objective means in the context of each new problem that scientists deal with depends on what the relevant aims are in those circumstances, how these can best be achieved, and how we should judge what is best. Likewise, precision is an important value or principle in science. But what does precision mean exactly, and what can count as precise enough? In Wundt’s reaction time experiments, his programme of mental chronometry, precision was deemed to be key, but how precision should be attained in this new science and what should count as “precise enough” had to be worked out in practice (Benschop & Draaisma, 2000).

Thus, if we look at science in practice we see that, as Knorr-Cetina put it, “the rules of official science are not exempt from local interpretation” (Knorr, 1979, p. 361). None of this means that principles are useless, but it does show that it is important to keep an eye on how principles work in practice if we consider which principles to adopt for our own field of study. To talk about “following the methods and principles of the natural sciences” obscures the fact that principles do not come before practices, but are always embedded in them, and only have meaning in practice. Even if it is possible to distinguish certain values that are generally typical of scientific practices, what these values (or norms, or principles, or rules, etc.) mean in practice has to be discovered anew when these practices change or have to be applied in the new contexts that science constantly creates.

Another drawback of the phrase “follow the methods and principles of the natural sciences” is that it sounds like being on a diet. It suggests that these methods and principles restrict us, and keep us from going astray. We should use only these methods and not others, and follow these principles to stay on the straight and narrow of natural science that will lead us to the truth. But we can also look at the methods and principles of the natural sciences as a buffet, just like Christina Bergmann urged us to do with the practices of open science (as cited in Whitaker & Guest, 2020). The fact that the natural sciences employ a multitude of diverse epistemic practices is a problem if we have to “follow” them, but it is a great benefit if we look at them as a buffet we can choose from. Not everything will be to our taste, but there will surely be methods in other sciences that we can put to good use in our own. We can look at the methods of the natural sciences as, potentially, “useful transferable tools” (Haack, 2012, p. 80).

We can push the buffet metaphor further and note that we don’t even have to eat the dishes exactly as they are served. We can take out the pickles, others can add ketchup. Methods can change as they travel from one research context to another. Chang’s (2022) notion of epistemic activities is useful in part because it recognises this adaptability: they are composites of constituent activities, and these components can be rearranged into new coherent composites that serve different purposes. And this is exactly what has been happening in the history of psychology. Wilhelm Wundt adopted the experimental method from physiology but incorporated inner perception into it. Gerard Heymans had a psychological laboratory constructed in the new University of Groningen academy building with multi-layered, soundproof dividing walls to shield the observers in his experiment from the chaos of life outside, an adaptation required by his specific view of psychology and its role in society (Derksen, 1997). Another example: in the process of constructing precision in his laboratory, Wundt designed instruments of his own, such as the “control hammer” for calibrating the Hipp chronoscope, and used a chronograph adjusted to his own specifications for final calibration (Benschop & Draaisma, 2000).

It is worth emphasising that methods are not only used, but also produced in science. They are among the many products of the work of scientists (Cartwright et al., 2022; Chang, 2022). Moreover, the development of methods, techniques, and apparatus for the production and measurement of phenomena and the analysis of data is intertwined with the theories that are produced. According to Ian Hacking, in successful laboratory sciences the mesh between theories on the one hand and methods and equipment on the other is so tight that they are self-vindicating:

Any test of theory is against apparatus that has evolved in conjunction with it – and in conjunction with modes of data analysis. Conversely, the criteria for the working of the apparatus and for the correctness of analyses is precisely the fit with the theory. (Hacking, 1992, p. 30)

Hacking does not think psychology is a laboratory science in his sense of the word (because there is not enough use of apparatus to produce the phenomena it studies), let alone a successful one, but in any science epistemic practices and theory evolve together, leading ideally to a coherent whole.

Psychology

To say that psychology is diverse is almost a cliché. As Smith (2020) writes, it is more appropriate to speak of “psychologies” than of “psychology” as a unitary discipline. As any historian of psychology will remind you, Wundt did not “found” one kind of psychology, but two: a physiologischen Psychologie and a Völkerpsychologie. The latter more or less died with Wundt, but psychology has remained a discipline in which very different approaches are loosely united. Moreover, psychology is not only the name of a discipline, it is also used to refer to “a host of beliefs and practices found in everyday life,” an everyday psychology, and to “a set of states someone has – a person’s psychology” (Smith, 2020, p. 2). Restricting ourselves for the moment to the discipline, it is clear that history has brought together under the label “psychology” very different topics, theories, and practices, and never has any one of them been so dominant as to unify the discipline (Green, 2015; Smith, 2020). It raises the question what should follow the methods and principles of the natural sciences? All that is called psychology? None of it? Some psychologies, but not others?

Psychology’s lack of “unity” has been reason to declare the discipline to be in crisis on several occasions (Goertzen, 2008; Sturm & Mülberger, 2012). ⁴ Jason Goertzen (2008), who analysed the crisis debates in psychology (a very extensive corpus of texts), noted that they are “substantially intertwined with the literature on psychology’s status as a science” (p. 835). Typically, those who believe that psychology should be unified also think this can only be done under an overarching natural scientific framework. Obviously, none of these attempts at unification have yet been successful. The recent history of evolutionary psychology is instructive in this respect. What John Tooby and Leda Cosmides pictured in their 1992 manifesto “The Psychological Foundations of Culture,” was a science that would integrate not only psychology but all of the social sciences and link them with biology (Tooby & Cosmides, 1992). Tooby, an anthropologist, and Cosmides, a psychologist, took aim at what they dubbed the Standard Social Science Model or SSSM, which, according to them, assumes that biological and environmental influences on human behaviour form mutually exclusive sets. As a result, the social sciences and much of psychology ignore human evolution and portray the human mind as almost infinitely malleable by changing its environment. Worse, the more extreme proponents of the SSSM, such as Clifford Geertz, abandon the “epistemic standards of science” and no longer look for the causes of behaviour but approach culture as a text to be interpreted (Tooby & Cosmides, 1992, p. 22). Such people “are no longer in a position to use their intellectual product to make any claims about what is true of the world or to dispute the others’ claims about what is true” (Tooby & Cosmides, 1992, p. 22). Evolutionary psychology offers a framework to overcome the split between the natural and social sciences, and to connect the latter with the “increasingly seamless system of interconnected knowledge” of the natural sciences (Tooby & Cosmides, 1992, p. 19). This bid at integration has so far not been successful: evolutionary psychology has not in fact become the foundation of psychology and the social sciences. It is now itself an approach to psychology, and a topic taught alongside other topics in the curriculum of psychology programmes. Moreover, other evolutionary perspectives on psychology exist (Laland & Brown, 2002), so that even in this case it is better to speak of “evolutionary psychologies” in the plural.

Given that the various psychologies came together in one discipline as the result of more or less arbitrary processes, Chris Green (2015) thinks it unlikely that they will ever be unified. It is more likely that “some portion of contemporary psychology may be unified according to some set of principles” (p. 211). There could be a new discipline that includes parts of psychology and parts of other disciplines and follows the methods and principles of the natural sciences. This is what seems to be happening with cognitive neuroscience, an interdisciplinary field comprising, among others, elements of psychology and neuroscience that is becoming more and more independent, including having its own research institutes and bachelor programmes.

Resistance against the unification of psychology as a natural science is often supported by ontological arguments about the uniquely reflexive character of human beings. As Smith (2005) noted, reflexivity is itself one of those “great words with multiple meanings” (p. 2). The sense of reflexivity that has been invoked in particular against psychology as a purely natural science regards the relation between psychological knowledge and the objects of that knowledge: psychological knowledge can change human beings. Psychology, the discipline, can change people’s psychology, both in the sense of their everyday psychological ideas and practices, and in the sense of their psychological states. The social psychologist Ken Gergen ⁵ caused a furore in 1973 when he published a paper arguing that social psychology cannot be a natural science, firstly because social behaviour is sensitive to changes in the social and cultural context, and secondly because the knowledge that psychology produces about social behaviour can affect that same behaviour (Gergen, 1973). When people learn about the bystander effect (Darley & Latane, 1968), they may decide to be more proactive in the future when they see someone in trouble. Knowing the results of Milgram’s (1963) obedience experiments can give people the resolve not to blindly follow orders. In general, social psychologists’ predictions about their behaviour may provoke reactance and lead people to change that same behaviour. As a result, the phenomena that social psychology studies do not have the stability of the objects of natural science. ⁶ They are, as Hacking (2007) would later put it, “moving targets.” There are no laws of social behaviour to be discovered. Social psychology should become a form of history, “a systematic account of contemporary affairs” that aims to sensitise people to the influences that shape their behaviour and to the historical and cross-cultural variability of social behaviour (Gergen, 1973, p. 316).

In a subtle and very precise analysis of the idea that reflexivity is ground for a division between the human and the natural sciences, Smith (2005) notes that the knowledge that science produces can change the natural world just as much as it can affect human beings: “Technology and material practices generally demonstrate how the physical world, as well as the human world, changes as our knowledge of it changes” (p. 14). As we are seeing all too clearly in the case of climate change, our knowledge of the natural world can lead to radical changes in the natural world. What sets the human sciences apart is that self-knowledge changes people, and that the role that the human sciences perform is “to make the reflexive process self-conscious” (Smith, 2005, p. 17). ⁷ Their knowledge therefore has a moral character that the knowledge produced by the natural sciences does not have. Smith adds that this is perfectly compatible with a view of psychology as a heterogeneous discipline that contains both natural scientific and humanistic approaches.

In my own contribution to this debate (Derksen, 2005, 2007) I have used the concept of self-cultivation to describe the work of mediating between human nature and culture that people do, and the role that psychology can play in this process. People do not simply have a nature, but are always in the process of tending it, celebrating it, enhancing it, probing it, suppressing it, and changing it if possible. Our nature is reflexive. Nature and culture are the provisional outcomes of what is for human beings a natural process of self-cultivation. To support this process it is vital that psychology offers different kinds of insight. Self-cultivation draws not only on knowledge of what is, but also on ideas about how things came to be the way they are and how they could be different. In cultivating their own nature, people can benefit from historical and cross-cultural studies of this process of self-cultivation that show its contingencies and make the reflexive process self-conscious, as Smith (2005) puts it. Pluralism is key: psychology should not restrict itself to following the methods and principles of the natural sciences (diverse as these are), but also embrace the methods and principles of the human sciences. ⁸

One could of course choose to restrict the use of the label “psychology” to the kind of psychological research that aims to follow the methods and principles of the natural sciences. Despite the diversity of the psychologies, this is the dominant definition of the discipline. Judging by the way that it presents itself at least, most of what is called “psychology” already follows the methods and principles of natural science. Introductory textbooks, for example, typically describe psychology as a science in the Anglo-American sense of the term: a natural science. Holtz and Monnerjahn (2017) analysed 10 popular introductions to social psychology, and found that they all refer to “the scientific method” to explain what it is that makes scientific psychology different from common sense. But even for a psychology that aims to follow the methods and principles of the natural sciences, the fact that people have their own ideas about their mind and behaviour has methodological consequences. Long ago, Saul Rosenzweig (1933) noted that it is commonly recognised that the fact that “everyone is a psychologist” (pp. 338–339) presents a problem for the experimental psychologist. To study the “natural,” “spontaneous” behaviour of the participants experimentally requires the use of special provisions to keep them “naive.” With the work of Robert Rosenthal, Martin Orne, and others in the 1960s, this and similar issues came to the forefront in what was experienced as a crisis in psychology (Faye, 2012). The realisation that psychologists study people who are in a way psychologists themselves, that the experiment is itself a social situation, and that the scientific and the everyday psychology can interact in all sorts of ways, led to a period of “epistemological dizziness,” in Jill Morawski’s (2015) fortunate phrase. Morawski concludes that these problems were eventually solved pragmatically, with technical solutions such as experimental deception to keep participants naive. Even if this doesn’t work for all participants in an experiment, reflexive disturbances are kept in check enough to create a reasonably valid experiment, or so it was thought. Another way of solving, or at least minimising the problem of the reflexivity of the experimental participants lay in shifting the focus to psychological processes that these participants are not and cannot be aware of and therefore cannot control. As Morawski (2015) wrote, “Bogus pipelines and other technical illusions eventually were replaced by techniques, like the implicit attitude test, that are believed to measure cognitions beyond the subject’s awareness” (p. 596).

Thus, psychologists have felt the need to introduce epistemic practices that are very uncommon among (other) natural science, precisely in order to fulfil their promise as a bona fide science. Deception, for example, is only used sometimes by scientists who study animals, such as when they use camouflage or a shelter to observe them unobtrusively. There is, moreover, one procedure in psychology that it does not share with any natural science: that of informed consent. In order to work with human participants in their studies, psychologists have to follow a formal procedure with legal consequences that implies that people have the right and the capacity to refuse co-operation. No other objects of scientific study have that legal right. Currently at least (informed consent has not always existed and may disappear again), this sets the sciences that work with people apart from other sciences.

Should 1

Recently, a crisis was once again declared in psychology, this time because it seemed that psychologists were not doing what they arguably should be doing according to the principles of the natural sciences. A central concern was the lack of so-called direct replications: studies that closely follow the procedure of an earlier study to see whether the same results can be obtained. Popper’s (2002) The Logic of Scientific Discovery was often quoted to support the argument that direct replications are indispensable in scientific research: “Only when certain events recur in accordance with rules or regularities, as is the case with repeatable experiments, can our observations be tested—in principle—by anyone” (p. 23). When reform-minded psychologists began doing such replication studies, they often could not reproduce the results of the original studies. Perhaps psychology is not following the methods and principles of the natural sciences very well. A reform movement emerged that aimed to redress psychology’s flaws. A sometimes-heated debate followed, that received new energy every time a failed replication was published, particularly if the original study was well-known.

Some social psychologists have responded to the crisis discourse and the calls for more replications in psychology by arguing that the unique characteristics of their object of study, social behaviour, make it unreasonable to expect direct replications always to “work.” Once again, the variability and context-sensitivity of social behaviour is put forward as its distinguishing feature, and even Gergen’s (1973) article is referred to (Iso-Ahola, 2017; Stroebe, 2016; Stroebe & Strack, 2014). Now, however, the conclusion is not that social psychology should abandon its ambition to be a natural science. Instead, social psychology must be a science in its own specific way. For these authors, it is a given that the phenomena that social psychology studies are historically changeable and culturally variable. “In matters of social psychology, one can never step in the same river twice” (Crandall & Sherman, 2016, p. 94). Repeating the exact same procedure in different historical, cultural, or social circumstances may therefore have an entirely different effect than in the original study—or no effect at all, and “it is even possible that the fact that . . . findings are reported in most social psychology textbooks and are therefore widely known among student participants could have affected the results” (Stroebe & Strack, 2014, p. 62). But, according to these authors, underlying these changeable phenomena are invariable and universal cognitive mechanisms, and they are what social psychology is primarily about. Social psychology is, they claim, first and foremost a study of the cognitive processes behind social behaviour, not of that behaviour as such. Theories about these cognitive processes are best tested and developed in conceptual replications, by testing different operationalisations of the same idea. “If the predictions of a theory are supported across a range of operationalizations of independent and dependent variables, then we gain confidence that we have learned something about a theory, rather than a single effect” (Crandall & Sherman, 2016, p. 95). The role of direct replications is relatively small. It is still a good idea to do a direct replication of one’s own experiment before publishing the results (Stroebe, 2016), and in applied research, direct replications are essential to test the reliability of an intervention (Crandall & Sherman, 2016; Stroebe, 2019; Stroebe & Strack, 2014). ⁹ In all other contexts, the failure of a direct replication is only informative if it is accompanied by a theory that explains the failure. It is the responsibility of the replicators to supply this theory; one cannot expect the original researcher to find out what made the replication fail to produce the effect (Strack & Stroebe, 2018, p. 40). ¹⁰

It is important to note that this proposal is based on three connected assumptions. First, that a distinction can be made between contextually variable social behaviour and the universal, underlying cognitive mechanisms that cause that behaviour in conjunction with contextual factors. Second, that there is a corresponding clear distinction between basic research that produces theories about the mechanisms, and applied research that develops behavioural interventions. And, finally, that social psychology should primarily be a basic science. Gergen (1973) rejected each of these assumptions. In his view, there are no basic social–cognitive processes. Social psychological phenomena are more or less durable depending on the extent to which they are tied to physiological processes, but there is no dichotomy between underlying social–cognitive mechanisms and the variable surface of actual behaviour. Thus, there is no basis for a division between pure and applied research. Social psychology should have an “intensive focus on contemporary social issues, based on the application of scientific methods and conceptual tools of broad generality” (Gergen, 1973, p. 317). Social psychologists might counter that the existence of basic social cognitive processes is a fact rather than an assumption, and that they are described and explained by the theories of social cognition that have been developed since the 1970s. But the evidence for those theories is a product of the same system of epistemic practices that is at issue in the replication debate. And even if one would grant the possibility of a pure science of basic social–cognitive processes, that would still leave a vast area of study focused on the very phenomena that make the basic science so difficult and direct replication “practically impossible” (Crandall & Sherman, 2016, p. 97). This psychology would study the historical and contextual variability of social behaviour and the interaction between the science of psychology, people’s own ideas about their mind and behaviour, and their mind and behaviour, in short the reflexive process. Direct replications would be very useful in such a science because, as the past decade has shown, they can provide evidence of the contextual variability of behaviour (Derksen & Morawski, 2022; Greenfield, 2017). Such a psychology would have a different politics as well as a different ontological and epistemological starting point. It would be a psychology of and for world-making, a science for which creating better, more equitable lives is not an application, but an inherent part of “basic” research (Derksen & Morawski, 2022; Gergen, 2015; Power et al., 2023). It would be a future-oriented science, that documents, explains, and feeds the processes in which people create their social worlds (Power et al., 2023).

Should 2

Rather than turn to the natural sciences for prescriptions—methods and principles that we should follow—we should look at their epistemic activities as a source of possibilities that we can adopt and adapt in our own work if that seems useful, just as we can draw on the work of philosophers, sociologists, literary scholars, psychiatrists, historians, novelists, poets, and artists, among others. Prescribing what legitimate science is and then policing these boundaries is counter-productive and itself unscientific. We should suppress “the demarcationist impulse” (Haack, 2012, p. 86). Science is not a fixed set of methods and principles that are applied in the study of the world “out there,” science develops methods and principles as it produces knowledge about the world and changes that world at the same time.

Let’s stop telling people what to do. Let them do their work in the way they see fit and we’ll engage with it when they have published it. The point of open science, which currently is a more important lodestar for psychology than the natural sciences, is precisely to make that possible. In as much as there are normative aspects to open science, they concern procedures and practices that allow others to see all aspects of the research process, all the epistemic activities that together led to the outcomes reported in the paper, and engage with them: with comments, criticism, or suggestions, or by drawing on them in their own work. Of course, open science practices are not epistemically neutral. A procedure like preregistration fits some forms of research better than others. But because it prioritises the principle of openness rather than specific methods, open science is potentially a broad church that can encompass both quantitative and qualitative approaches and accept members from all disciplines, enabling fruitful conversations between then. To some extent, this is already happening (Field & Derksen, 2021; Haven & Grootel, 2019; Steltenpohl et al., 2023).

Whether psychology should follow the methods and principles of the natural sciences is an old question, and in light of the diversity of both the natural sciences and psychology itself (or better: the psychologies themselves) it may seem to be a question that we should finally move beyond. Roger Smith (2020) has argued that

there is little profit in returning to argument about whether psychology is or is not a science, meaning in English a natural science . . . . Whatever the arguments, there are psychologies, in the plural, and discussion of the nature of scientific psychology belongs in discussion of that diversity. (p. 5)

But we should not underestimate the power of the word “science” understood as “natural science,” and the extent to which the curricula of psychology programmes in many parts of the world are still dominated by a view of psychology as at its core a quantitative, experimental discipline. That the psychologies are diverse and that this diversity should be fostered is a point that bears repeating.