What are three of the most important tools which political scientists use to develop answers to research questions in the study of global politics?

The distinction between different traditions makes it possible to identify the fundamental differences within political methodology. I have identified three methodological traditions (Table 1, column 2–4, Lauer 2017) with the help of the concepts of von Wright (1971) and Lakatos (1978), which I will briefly introduce below. The foundations of these traditions can be found in the Aristotelian Organon (Aristotle 2019, fourth century BC).

In a first step, I will establish not just the limits, but above all the possibilities of the descriptive and the explanative-prognostic traditions (“Descriptive tradition” and “Explanatory-prognostic tradition”). The applied methodologies of these two traditions are discussed separately in “Problem-driven methodology: applied phronesis” and "Applied methodology: social technology". The limitations of the descriptive and the explanative-prognostic traditions, the barriers that political scientists cannot overcome, are established by showing first the possibilities of the two traditions in “Descriptive tradition” and “Explanatory-prognostic tradition”, and then their limits in “Problem-driven methodology: applied phronesis” and "Applied methodology: social technology". The fundamental limits of these applied methodologies are established, the foundations of which go back primarily to pragmatism and, for the scientistic tradition, to Bacon, Weber and Popper.

In “Philosophical foundations of the practical tradition” I reconstruct the philosophical foundations of the practical tradition. I then explain why applied or better practical political questions require their own methodology ("Applied methodology: social technology"). Then I briefly outline the fundamentals of a practical methodology, which is based above all on the fundamentals developed by Aristotle and Kant (“Practical methodology (normative, pragmatic and technical)”).

It is certainly true that there is better and worse work by qualitative researchers, just as there is by quantitative researchers. But that is not my point. Above all, I examine the method books that actually define the criteria for good research. My aim is not to show good and bad scientific work using individual examples, but to illustrate the discussions using the example of the methodological manuals.

Descriptive tradition

The descriptive tradition can be explained very briefly because there is no great controversy about the limits and possibilities of this methodology. This methodological tradition stands on its own and cannot be reduced to any other methodological tradition.

Philosophical foundations of the descriptive tradition

The goal of the empirical-descriptive tradition is understanding (sense making, meaning making). It is firstly about describing (visible) phenomena and secondly about the interpretation of symbols (text, image, audio and video), above all through language. It uses an empirical-descriptive methodology. Sense making and meaning making can be found above all in the human and cultural sciences (humanities), which is where hermeneutic, phenomenological and structuralist explorations of meanings and contexts have commonly been based. Hermeneutics, phenomenologists, structuralists, interpretivists and perestroikans all belong to this very heterogeneous tradition.

Creswell includes the following interpretative frameworks of qualitative research: postpositivism, social constructivism, transformative frameworks, postmodern perspectives, pragmatism, feminist theory, critical theory and critical race theory (CRT), queer theory, and disability theory (Creswell 2013 [1998], 22 ff.). Creswell distinguishes five qualitative approaches: narrative research, phenomenology, grounded theory, ethnography, case study.

All the scientists working in these fields criticize the naturalists, (neo)positivists, or scientistic scientists who constitute a very homogenous tradition which will be presented in the next section. A so-called Methodenstreit (Lauer 2017) erupted in the nineteenth century on the philosophical (axiological, epistemic, methodological and ontological) level between the descriptive and the explanatory-prognostic tradition. In the second half of the twentieth century, the fight shifted primarily to the methodical level in the strict sense of the term, namely between quantitative and qualitative methods. Since the turn of this century, the Methodenstreit has yet shifted from the methods level back to the philosophical level. The phronetic perestroikans’ argumentation for a “real social science” (Flyvbjerg et al. 2012) is also made on the philosophical level. The phronetic perestroikans are the latest (counter-) revolution to the scientistic Establishment.

Empirical-descriptive methodology

Scientists who are primarily interested in hermeneutics and phenomenology, and thus not in the natural sciences but in the humanities, can in the meantime resort to a very differentiated and elaborate methodology, i.e., hermeneutic-interpretative ways of reasoning and qualitative-interpretive methods (Bevir and Rhodes 2016; Creswell 2013 [1998]; Yanow and Schwartz-Shea 2014a [2006]). For some years now, logical-mathematical methods have also found their way into the humanities. The digital humanities are increasingly using logical-mathematical methods for text interpretation due to the emergence of big data (Budrick et al. 2012).

Explanatory-prognostic tradition

The explanatory-prognostic tradition works with an empirical explanatory-prognostic methodology. The scientistic scientists in political science wish to set themselves apart from the humanist traditions through their orientation on the natural sciences and their introduction of a logical-mathematical research methodology. This requires its own axiological, epistemic, methodological and ontological foundations. For the most up-to-date and comprehensive overview of the explanatory-prognostic methodology, see The Oxford Handbook of Political Methodology (Box-Steffensmeier et al. 2010 [2008]).

According to Goodin (2011 [2009], p. 13), there have been three revolutions in American political science. The first was the establishment of politics as a science and its orientation on the natural sciences, above all through the introduction of causal and empirical thinking, deductive and inductive argumentations, the introduction of professional and systematic approaches and the establishment of political science as a discipline especially at American universities at the beginning of the twentieth century. The second revolution, the “behavioral revolution”, came in the 1950s, and the “rational choice revolution” took place in the 1970s.

The explanatory-prognostic tradition has by far the most developed methodology. Therefore, its explanation takes up the most space. The methodology of the explanative-prognostic traditions consists of several methodological scientistic research programs that complement each other and are constantly evolving (Lauer 2017, pp. 53 and 56):

1. A.

   Orientation on the natural sciences

2. B.

   Causal thinking or causal reductionism

3. C.

   Empiricism or empirical reductionism

4. D.

   Rationalism

5. E.

   Quantitative research program

6. F.

   Model thinking (especially rational choice approach)

7. G.

   Qualitative-mathematical research program

8. H.

   Experimental-simulative research program based on experiment and simulation.

Empiricism and rationality are the overriding principles to which scientific research must adhere, since scientific theories consist of logical formalism and an empirical foundation. Science requires both a rational formalization and an empirical anchoring. These principles are fundamental for both empirical and practical knowledge and therefore for empirical and practical methodology.

In the following, I shall deal with three core elements: causal reductionism, empirical reductionism, and rationalism. This will be preceded by a quantitative overview of the dissemination of this tradition.

Dissemination of the explanatory-prognostic tradition

The explanatory-prognostic tradition does not dominate the mainstream of political science, as is claimed by many scholars. Quantitative analyses show that since the 1950s only a minority of researchers have carried out such investigations. Only 5% of US scientists, especially behavioralists and rational choice theorists, use a quantitative methodology. Only one-third of authors publish causal analyses in scientific journals. Although that number has risen to about 35% in the APSR and has been constant since the 1970s, the number for all articles in JSTOR was under 20% in the 1990s (Goodin 2011 [2009], p. 13).

Furthermore, there is a gap between causal (roughly one-third) and linguistic-interpretive (6%) studies (narrative and interpretative), the latter increasing only in the 1980s within the most important American publications (Brady et al. 2011 [2009], p. 1036).

Considering furthermore that quantitative researchers prefer to publish their results as articles, while qualitative-interpretive researchers more often publish their results in book form, causal thinking and quantitative research are clearly far from forming the majority of political science publications.

However, quantitative researchers of the explanatory-prognostic tradition have a very strong influence within political science institutions, not least because, according to some scientists, they dominate the APSA and other scientific institutions (Monroe 2005). Based on these analyses, it is not possible to speak of “mainstream”, but since this is a very influential minority, I call it the “scientistic establishment” (Lauer 2017). Representatives of this tradition can be found in every political science department worldwide. In his critique of the establishment, Mr. Perestroika writes of “East Coast Brahmins” (Mr. Perestroika 2005 [2000], p. 9) who control the APSA and the editorial board of APSR. However, the phenomenon is in no way restricted to the East Coast of the United States.

Causal reductionism

Objectives of causal thinking

Causality is the ontological condition par excellence. Like Goethe’s Faust, causal reductionists search only for what holds the world together at its core. More pragmatically, causality is “The Cement of the Universe” (Mackie 1974). Those who identify causalities can recognize and change the world: “Causal statements explain events, allow predictions about the future, and make it possible to take action to affect the future” (Brady 2011 [2009], p. 1054).

The search for invisible causalities is the common thread that holds together the explanatory-prognostic tradition. While phenomena or appearances are visible (the Greek word phainómenon means visible, appearance), causality is invisible, so one can describe appearances, but only explain causalities: “To explain is to expose the internal workings, to lay bare the hidden mechanisms, to open the black boxes nature presents to us” (Salmon 1989, p. 134). This is stated by philosophers of science and is also taught in the corresponding political methods books.

Causality is primarily about inferences: “Obviously, we do not thereby mean that one directly observes causation. Rather, this involves inference, not direct observation” (Seawright and Collier 2010 [2004], p. 318, emphasis in the original). “Inference” is the magic word: “The goal is inference [emphasis in the original]” (King et al. 1994, p. 7). Secondarily, it is about observations (data-set or causal-process observations, Brady and Collier 2010 [2004]).

All methodological innovations (quantitative methods, model thinking, qualitative-mathematical methods, experiments, simulations) only serve the purpose of identifying invisible causalities and therefore causal regularities between different events on the macro level and concrete causal mechanisms at the micro level. Within this tradition, therefore, only causalities are sought; causality is the only relation that counts, and other relations or even contexts are of no interest (see also King et al. 1994, p. 75). I therefore speak of causal reductionism (Lauer 2017).

From correlation to causality

Causal thinking requires different, very complex methodological approaches to identify causalities. These methodological approaches are based on corresponding philosophical theories of causality, which were developed primarily by philosophers of science. In this case, the methodological approaches presented by Brady (2011 [2009]) ultimately form the core of four causal theories (Salmon 1989), which have been operationalized in such a way that the procedure for finding causalities can be specified. The main purpose of these theories is to establish the properties of causality and criteria that can be used to distinguish valid from invalid causal explanations.

Brady (2011 [2009]) has presented four methodological approaches, all of which are necessary to identify causalities or explain events causally:

1. I.

   Regularity or neo-Humean approach to determine regularities (correlations, not causalities!).

2. II.

   Counterfactual approach to eliminate random correlations.

3. III.

   Manipulative approach to determine the time arrow or the identification of cause and effect.

4. IV.

   Mechanism approach for the exact determination of the causal mechanism.

With the regularity approach, a conjunction between two events, technically speaking between two variables, is established. Thus, a regular and constant correlation between two events can be demonstrated with correlation and regression analysis. With the help of experiments (thought experiments, but also laboratory and field experiments) and/or simulations, one can imagine, but also artificially generate, possible worlds in which the cause does not appear, and then see what happens. The aim of the counterfactual approach is to sort out random correlations. With the regulative and counterfactual approaches, one can recognize correlations and probabilities and thus confirm the simultaneous existence of two variables, but identify neither the cause (independent, explanatory variable) nor the effect (dependent, explained variable). That is, one can recognize the symmetrical, but not the asymmetrical property of causality. The manipulative approach, which also draws on experiments and/or simulations, should first determine the direction of causality or the arrow of time and thus identify one variable as a cause and the other as an effect. Since, by definition, the cause of an action precedes the action, cause and effect can also be identified.

These approaches are macro-level studies that cannot answer the following questions: How can a specific correlation at the micro level be identified? How does the causal mechanism work?

The first question is about solving the pairing problem at the micro level, with the micro level simply forming each case. The second question aims at the ontic conception of causality (Salmon 1989). In other words, causality also has an ontological property as well as being nomological.

Regulative, counterfactual and manipulative theories of causation, methodically generated using correlation methods, experiments and simulations, can demonstrate the nomological aspect of causality and even partially answer the “why” question. However, the question of “how” is an ontological question and should above all explain how a cause produces an effect. Macro-analysis consisting of hypothesis-testing methods is therefore not enough to move from correlation to causality; microanalysis is required.

Here we come to the fundamental limitation of deduction: there is no deductive certainty, just as there can be no inductive certainty. Even if one has recognized the biological causal mechanism at the micro level, one cannot conclude that this relationship applies in all cases at the macro level. Smoking cannot explain lung cancer in all cases (limitation of induction), and smoking causes lung cancer only usually, but not in every single case (limitation of deduction). Both deduction and induction therefore have structural limitations. This is also an argument against fallibilism. There is no asymmetry between induction and deduction in causal analysis, as Popper (1968 [1934]) claimed. However, it makes no sense to demonize deduction as Popper demonized induction. The covering law model was the most popular scientific argumentation in the nineteenth and twentieth centuries. This model is now ignored for good reasons; it is not even mentioned in The Oxford Handbook of Political Methodology (Box-Steffensmeier et al. 2010 [2008]).

Today, quantitative-mathematical methods are used at the macro level, and qualitative-mathematical methods (not to be confused with the qualitative-interpretative methods) at the micro level, e.g., process analysis and qualitative comparative analysis (QCA). Qualitative-mathematical methods are used in case studies or small-N studies to identify specific causal mechanisms as well as to solve the pairing problem.

Interpretivists also speak of “qualitative-positivist” methods (Schwartz-Shea 2014 [2006], p. 143, footnote 6). I use the term qualitative-mathematical methods, since they are based on logic and mathematics, more precisely on modal logic, set theory and statistics (Lauer 2017).

The four methodical approaches for determining causality will now be described using an example. With the regularity approach, one can, e.g., discover that there is a correlation, regularity or probabilistic law between smoking and lung cancer. The counterfactual approach can be used to show that this is not a random correlation, and the manipulative approach allows us to identify smoking as the cause of lung cancer and lung cancer as the effect of smoking; more specifically, it shows what temporal occurrences between these two variables exist.

The complexity is increased by the existence of other causal relations. Further causal analysis shows that other environmental contaminants also lead to lung cancer (equifinality), and that some people do not develop lung cancer despite intensive smoking while others who do not smoke get lung cancer (asymmetric causality). In other words, different effects may have a common cause (equifinality), and conversely, a cause in combination with other conditions may produce different effects (multicollinearity, conjunctural causality).

This brings us to the pairing problem. Now, if someone who has smoked dies of lung cancer, the question of the cause of death remains unsolved: smoking, or other environmental factors? Only when all these questions have been clarified can one state that this specific individual died from smoking.

Science is characterized by specialization, so reducing complexity is generally the beginning of any scientific work. Every scientist has to apply Occam’s razor. Unfortunately, there is no safe way to separate important from unimportant factors. A central goal in methodological handbooks is to determine how to filter out unimportant factors (King et al. 1994; Brady and Collier 2010 [2004]). But in spite of all caution, there is typically the danger that one overlooks or ignores a third factor. This already describes the Galton problem (Moses and Knutsen 2012 [2007], p. 105).

With the causal approaches presented so far, only a correlation was established and the chronology was clarified. But this does not yet explain the how, or what mechanism is at work. Yet this is necessary for the causal explanation to be complete. The next major problem is therefore the Mill problem. Statistical and comparative methods cannot specify any necessity for the different variables studied (Moses and Knutsen 2012 [2007], p. 105).

The causal mechanism of how smoking causes lung cancer has not yet been explained, i.e., the ontic property of causality still needs to be explained. In other words, only the “why”, but not the “how” question has been answered. Therefore, the mechanism approach is needed to identify the underlying causal mechanism (Salmon 1989, p. 181).

The existence of a variety of qualitative-mathematical methods for determining causality at the micro level within case studies and small-N studies shows that even scientistic studies use not only nomothetic but also idiographic investigations. This is precisely what the perestroikans and other interpretivists deny. Windelband (1900 [1894]) was the first author to make a distinction between natural sciences, which aim at recognizing the general by means of nomothetic methods, and humanities, which aim at recognizing the individual by means of idiographic methods.

Actor-centered or teleological explanations

Causal reductionism is criticized for being insufficient to explain actions. Explaining only the effects of actions is not enough; one also wants to explain the intentions of the actors. von Wright (1971) has pointed out that in addition to causal explanations, an action theory also requires fundamentally different teleological explanations. To use his terminology, the Galilean tradition needs to be complemented by the Aristotelian tradition.

The rational choice approach aimed at providing a comprehensive explanation of social and political behavior. While systems theory and (neo-) Marxism offer explanations from a holistic viewpoint and with methodological holism, rational choice models attempt explanations through methodological individualism, generated from the point of view of actors, where the actors can be individuals acting individually but also collectives (organizations, classes or the state): “‘Bringing men back’ was an achievement of the rational choice approach” (von Beyme 2000 [1972], p. 145, my translation). The rational choice approach deserves the credit for bringing actors and thus people back into the field of research. Coleman explicitly uses the teleological action theory put forward by Weber (Coleman 1990, p. 13).

Depending on the point of view—individual or collective, i.e., holistic or aggregate level—from which explanations are made, causal or teleological ways of argumentation will be used. Both are necessary for explaining political reality.

Empirical foundation and empirical reductionism

Empiricism plays a fundamental role in science. No scientist calls into question the need for an empirical foundation as a general criterion. What has frequently been controversial, however, is how the empirical foundations can be realized in concrete terms and which fundamental limits must be considered.

Many political scientists advocate an empirical reductionism, which states that all scientific knowledge requires an empirical foundation. Therefore, only an empirical political science can claim the status of a science. Others consider that there are important issues that can be rationally discussed but cannot be determined empirically, and therefore require no empirical foundation.

Objectives of empiricism: verificationism and fallibilism

Locke (1975 [1690]) is considered the founder of empiricism. But Bacon (2000 [1620]) is the first to assert that theories can be clearly confirmed empirically by means of decision experiments. These mark a theoretical dividing line and should enable a clear identification of causes. Two alternatives are designed, and one of these is refuted by the experiment while the other is confirmed (Bacon 2000 [1620], p. 159).

Verificationism was then extended and differentiated especially in the twentieth century by logical positivism and logical empiricism (neopositivism). Popper (1968 [1934]) also represents a form of empirical reductionism. According to him, theories are not confirmed but are falsified as much as possible by the experiment. Popper’s fallibilism is mainly directed against holism.

Quine–Duhem or holism thesis

Duhem (1978 [1906]) first denied that such experimenta crucis (Bacon 2000 [1620]) exist and propagated a holistic view of science. Quine generalized the relation established by Duhem for physics to all of science: “The unit of empirical significance is the whole of science” (Quine 1964 [1953], p. 42), because the “dogma of reductionism survives in the supposition that each statement, taken in isolation from its fellows, can admit of confirmation or infirmation at all” (Quine 1964 [1953], p. 41). This is wrong because “our statements about the external world face the tribunal of sense experience not individually but only as a corporate body” (Quine 1964 [1953], p. 41), and secondly because “[t]aken collectively, science has its double dependence upon language and experience; but this duality is not significantly traceable into the statements of science taken one by one” (Quine 1964 [1953], p. 42).

Criticism of empirical reductionism

In the case of ontological questions, an empirical examination or decision is not possible in principle because it can neither be falsified nor reasonably made empirical at all: “the question of structure and agency, or any other ontological issue for that matter, cannot be falsified—for they make no necessary empirical claim” (Hay 2011 [2009], p. 469). Empiricism (sometimes referred to as “hyperfactualism”) anchored in logical positivism and critical rationalism is an ontological premise. Hyperfactualism and hypermethodologism without proper philosophical reflection are criticized in an article in the Oxford Handbook without any noticeable impact on the other articles (Bevir 2010 [2008], pp. 68–69).

Rationalism (more geometrico)

Rationality postulates

Knowledge criteria serve to evaluate the success of empiricism and rationality. With the help of rationality postulates, general criteria of scientific research are formulated so that methodological precision can be guaranteed.

The following general criteria are recognized by all scientists who prefer a logical-mathematical research methodology:

1. a.

   Intersubjectivity Science seeks ways to find justifications that any rational and knowledgeable person can understand.

2. b.

   Objectivity Subjective desires or prejudices must not be incorporated into the work, only intersubjective reasons.

3. c.

   Reliability The results of scientific investigations should be reproducible under the same conditions.

4. d.

   Validity A scientific result must have an argumentative weight and meet methodological quality criteria. A distinction is made between internal validity (credibility) and external validity (transferability).

Rejecting criteria, as qualitative researchers and postmodernists often do, is not a convincing approach. This has now also been recognized, with restrictions, by qualitative researchers: “The tendency to increasingly view qualitative research as an art doctrine (see Denzin and Lincoln 1994) or ‘research style’ (Strauss 1987, 1985) and less as a formalizable approach does not relieve one of the application of evaluation criteria” (Steinke 2015 [2000], p. 322), my translation, see also Schwartz-Shea 2014 [2006] and Yanow 2014 [2006]). Even qualitative researchers no longer generally reject intersubjective or objective criteria, even if one speaks of “intersubjective comprehensibility” as a criterion for qualitative research and “intersubjective verifiability” (Steinke 2015 [2000], p. 323, 324) for quantitative research.

Subjective versus intersubjective language

Some interpretivists dispute that one can formulate all recognition objectively or intersubjectively. Instead, they claim to reproduce subjective experiences in a subjective way by means of language. They refer especially to Wittgenstein’s (1953) late work. But Wittgenstein himself expressed doubts about this in his famous arguments against a private language, illustrated with the example of one’s own pain (Schmerz). Even when one speaks about a subjective feeling like pain, one uses an intersubjective tool, namely an intersubjective, objective and public language. In other words, since there is no private language, we are also required, when formulating subjective experiences and subjective views, to reproduce them with the help of an intersubjective and public language.

The generation, evaluation or justification of scientific knowledge requires both general and specific criteria at ten methodological levels. The ten vertical and three horizontal levels are the systematic foundation upon which knowledge is methodologically generated and evaluated (Table 1, Lauer 2017).

Logical-mathematical methodology

Empirical causal analyses have been produced in political science since the 1950s with the help of quantitative tools as well as deductive and inductive arguments. In the 1970s, logical-mathematical model analyses were introduced; in political science these were mainly rational choice models. In the same decade, qualitative-mathematical methods were added. Experiments were included in the 1990s (Morton and Williams 2010 [2008]; Gerber and Green 2011 [2009]). In political science, in contrast to sociology, simulations are rare.

Practical tradition

Differentiation and specialization, the two salient features of modernity, have also been integrated into political science since the emergence of the discipline as a science. This is especially true for analytic-empirical theories. Methodology (form) and theory (content) are here usually treated separately; therefore, there is a variety of methodological manuals.

In political philosophy or political theory on the other hand, form and content are very rarely treated separately when it comes to the generation of practical theories. There simply are no methodological manuals dealing with the methodology of practical knowledge and theories. Political philosophy and theory thus remains stuck with one foot in the pre-modern era.

The second shortcoming in political science methodology is that scientistic scientists and interpretivists believe that the same methodology can be used to generate empirical and practical knowledge. A practical discourse that satisfies the current logical-analytical considerations is not possible neither with the scientistic nor with the interpretive methodology. A practical discourse requires a genuinely practical methodology and not a reductionist one.

In the following, I will first discuss the philosophical foundations of a practical tradition (“Philosophical foundations of the practical tradition”). Second, I will briefly explain the practical methodology of the constructivist using the example of the perestroikans (applied phronesis and pragmatism) and will show their limitations (“Problem-driven methodology: applied phronesis”). Third, I will do the same with the practical methodology of the explanatory-prognostic tradition, i.e., normative rational choice theory ("Applied methodology: social technology"). Fourth, I will provide an overview of a genuinely practical methodology (“Practical methodology (normative, pragmatic and technical)”).

Philosophical foundations of the practical tradition

The discussion of axiological questions or value problems is very controversial within the scientific community (Lauer 2017). Only two central problems can and must be addressed here: the is–ought problem and the foundations of an applied methodology.

Is–ought problem, fact–value gap

The separation into theoretical and practical philosophy made in antiquity is based above all on the fact that “is” and “ought” are different objects and therefore require fundamentally different methodologies. Following this tradition, affirmed by Kant, Wieland used the example of medicine to highlight the fundamental distinction between theoretical and practical science (Wieland 1986).

Much more influential are the methodological distinctions made by Weber, which continue to characterize the axiological arguments even today. Weber unconditionally adopts the distinction between is and ought, since he is guided by the spirit of neo-Kantianism (Weber 2011 [1904], p. 50). However, he does not use the Aristotelian or Kantian terminology, preferring the terms “empirical discipline” and “empirical science” (Weber 2011 [1904], p. 52, 54 and 55) on the one hand and “practical social science” (Weber 2011 [1904], p, 56) and “social policy” (Weber 2011 [1904], p. 60 and 67) on the other. In his methodological writings, Weber’s main concern are the limitations and possibilities of an empirical science. The limits of empirical sciences that Weber identified have generally been accepted by the scientistic scientists until today. Weber demands that scientists should make a clear distinction between logical-analytical discussions, empirical analyses and practical valuations.

Hume (2007 [1739/1740], pp. 469–470) was the first who speaks of a fact–value gap; and according to Moore (1965 [1903], § 12, 40), one makes a naturalistic fallacy when one concludes from an “is” to an “ought”.

Foundations of applied methodology

Applied means the inversion of causal sentences, i.e., empirically determined causal sentences in the form of if–then statements are transformed into prescriptive instructions or regulations in the form of technical rules.

The foundations of applied methodology are rarely addressed, at least in political science handbooks. Two implicit assumptions, inversion of causal statements and equivalence of causality and action, are discussed below. Furthermore, the criticism that is directed against applied methodology and comes especially from philosophers of technology is also addressed. I have also compiled structural differences between empirical and practical methodology at ten methodological levels (Table 1; structural differences between sciences and engineering have been worked out by Poser 2001, 2016; Lauer 2017).

(a) Is–ought bridge The causal reductionists claim that causal statements provide scientifically valid world knowledge. At the same time, almost as the other side of the coin, causal statements make it possible to take action to affect the future (see “Causal reductionism”). This would make the identification of causality sufficient by itself. An applied methodology is then simply the inversion of causal sentences. This is only possible because Weber and others have built a bridge from “is” to the technical (not normative or pragmatical) “ought”.

Through the simple “inversion of causal sentences”, technical rules necessary for a success-oriented (purpose-rational) action are generated. Only means discourses for this purpose are possible within empirical science (Weber 1973 [1917], p. 517 [479]).

Weber’s is–ought bridge was later adopted by critical rationalism: “The task of science is partly theoretical—explanations—and partly practical—prediction and technical application. I shall try to show that these two aims are, in a way, two different aspects of one and the same activity” (Popper 1981 [1972], p. 349, emphasis in the original). And a few pages later: “This makes it clear now, from a logical point of view, both the derivations of predictions and the technical application of scientific theories may be regarded as mere inversions [my emphasis] of the basic schema of scientific explanation” (Popper 1981 [1972], p. 353).

Albert, the most famous German critical rationalist, speaks of a tautological transformation that needs no additional premises (Albert 1967 [1965], p. 192). This reasoning is untenable, as will now be shown.

(b) Equivalence between causality and action The inversion of causal statements requires a premise. Causality can only form the basis of both world knowledge and world change if there is an equivalence between causality and action. Today this is assumed and rarely discussed. Bacon was aware of the connection between causality and action: “Human knowledge and human power come to the same thing, because ignorance of cause frustrates effect. For Nature is conquered only by obedience; and that which in thought is a cause, is like a rule in practice” (Bacon 2000 [1620], p. 33). In today’s casuistic terminology, one would say: theory and practice come together, statements that are true in theory are efficient in practice.

In pragmatism, there is an equivalence between usefulness and truth: anything that is useful is true: “You can say of it then either that ‘it is useful because it is true’ or that ‘it is true because it is useful’. Both these phrases mean exactly the same thing” (James 1975 [1907], p. 98).

Likewise, Bacon argues that what is useful is also true: “The two pronouncements, the active and the contemplative, are one and the same; and what is most useful in operating is truest in knowing” (Bacon 2000 [1620], p. 104).

Thus, pragmatists and scientistic scientists share the same premise when it comes to the foundations of an applied methodology. The equivalence between causality and action and an equivalence between statements and rules thus leads to the division into empirical (theoretical) and applied sciences. Causalities are determined within the former; the applied sciences only need to invert them. Thus, instructions or advice can be formulated almost incidentally in applied (not practical!) political science.

I will now point to some objections advanced by philosophers of technology to such a reductionist methodology.

(c) Criticism of applied methodology Today, the inversion of causal sentences or of the fundamental explanatory scheme is one of those hidden or tacit assumptions that rarely being mentioned or even discussed. The equivalence between causality and action is a “transmission rule from knowledge of nature to the rule of action in nature” (Kornwachs 2012, p. 42, my translation). Mario Bunge calls this a pragmatic syllogism (Bunge 1967, pp. 132–139).

Importantly, pragmatic syllogism is not formally valid. The first sentence is a statement of the form “if x, then y”. Inversion is a (technical) rule of the form: “if you want to reach y then do x”. There is only a pragmatic, but no logical, relationship between regulative propositions, e.g., “if A, then B”, and associated (technical) rules, e.g., “B by A”, “if you want to achieve B, then try A” (Kornwachs 2008, p. 139, 2012, p. 64. Kornwachs takes over this notation from Bunge 1967).

Problem-driven methodology: applied phronesis

The practical methodology of the constructivists, interpretivists and perestroikans will now be briefly explained and criticized using the example of the perestroikans (applied phronesis). The perestroikans as well as most interpretivists reject a separation between “is” and “ought”, referring to pragmatism and the Frankfurt School. They go a step further, claiming that the axiological values of researchers almost necessarily influence research in the shape of guiding interests (Habermas 1968; Yanow and Schwartz-Shea 2014b [2006]).

There is agreement between the perestroikans and the scientistic scientists on the idea that with a reductionist methodology one can both recognize and change the world. The difference is that the perestroikans do not begin by recognizing the world but by solving problems. Cognition and truth then emerge as a by-product, just as the scientistic scientists think that the application is a by-product. For the perestroikans, political science should operate with a problem-driven approach rather than with a method-driven or theory-driven approach that is the hallmark of the scientistic scientists (Shapiro 2005).

Applied phronesis

The phronetic perestroikans question the methodological approach of the scientistic scientists and want to use an applied phronesis to revolutionize political science and to achieve greater public relevance for the discipline (Flyvbjerg 2001; Schram and Caterino 2006; Flyvbjerg et al. 2012).

The phronetic perestroikans, using the Aristotelian phronesis and the inclusion of questions of power, want to establish a better and, above all, a relevant alternative discipline to contemporary social and political science, a phronetic or real science. Consequently, the social sciences that are based on the natural sciences are criticized as unreal (Flyvbjerg et al. 2012, p. 1).

The goal of the phronetic perestroikans is to create a problem-driven methodology with the help of applied phronesis. The phronetic scientists aspire to a philosophy of engagement that can change unjust states. The applied phronesis consists not only in the practical knowledge of how to transform unjust states by discovering the tension points, but also in the revolutionary ability to transform these tension points (Flyvbjerg et al. 2012). In doing so, the phronetic perestroikans act together with the affected groups, to whom they convey the knowledge and the practical skills to enforce their concerns and achieve empowerment. Thus, scientists become both scientists and revolutionaries at the same time (Flyvbjerg et al. 2012, p. 10).

Limits of applied phronesis

The general goal of the phronetic perestroikans is “to make the world a better place” (Flyvbjerg et al. 2012, p. 11). But who knows what is better, and above all, who justifies this and how? A practical science must not just answer technical questions, but also justify them. The answer of the perestroikans is: the justification lies in the moral attitudes of the researching scientists and those of the groups involved.

Unfortunately, they fail to explain how one can justify these goals with an applied phronesis and which other scientific tools are necessary for this. Applied phronesis is sufficient only for the generation of technical knowledge. Thus, only a technical means discourse is possible, and neither a normative value discourse nor a pragmatic target discourse can be conducted with it.

Applied methodology: social technology

Normative methodology: normative rational choice theory

I will now examine Hardin’s (2011 [2009]) contribution to how applied knowledge can be generated within the explanatory-prognostic tradition. According to Hardin, the rational choice approach suffices to recognize the world (positive rational choice theory), and a normative rational choice theory also makes possible the normative assessment and practical changing of political reality.

The scientistic establishment believes not only that normative rational choice theory offers an adequate practical methodology, but also that it is the best practical-normative methodology available today. The normative rational choice theory is considered the only valid normative methodology and theory (it is seen as both); all other normative theories are dismissed as irrelevant and even esoteric (Hardin 2011 [2009], p. 99).

Limits of the normative rational choice theory

This work is not interested in how to formulate social-technological regulations with the help of game theory. According to Hardin, this is relatively simple (Hardin 2011 [2009], p. 99). What is important is the character of a practical discourse that works with the help of this methodical approach. The normative rational choice approach is a technical means discourse. Thus, technical regulations can be formulated as part of an applied (not practical) political science. The ethical-normative and pragmatic dimensions are not addressed at all, as has been the case since antiquity in political philosophy. Rational choice theory, positive and normative, therefore operates with an instrumental reason. Normative discourses are only of interest to the extent that they are concerned with the discussion of means. This is also clearly recognized by the adepts of rational choice theory. The criticism of the Frankfurt School, that the explanatory-prognostic tradition would produce only “technically competent barbarians” (Rothstein 2005; Horkheimer 1947), is still supported. However, this criticism is only partially justified. Norms and values are properly considered, but while the perestroikans refer to the values of the researchers and the disadvantaged groups they are committed to and engaged with, the scientistic scientists refer primarily to liberal and utilitarian values.

However, neither perestroikans nor scientistic scientists can justify their values, norms and goals. Just like other ontological and epistemological ideas, these are simply assumed. In other words, rational choice theory, if used for practical purposes, is concerned only with technical discourse, and any legitimating intention is simply banished to its underlying ontological (individualism, self-interest) and ethical assumptions (utilitarianism). These prerequisites cannot now be justified by the rational choice approach.

Hidden assumptions also include the philosophical foundations of practical methodology discussed above. Thus, an equivalence between causality and action is assumed by not differentiating between normative and positive theory. If, as Hardin believes, method and theory coincide, one has the problem that it is difficult to distinguish between empirical or normative assumptions and the empirical or normative knowledge that this approach generates (Hardin 2011 [2009], pp. 93–94).

The terms “rational choice approach” and “rational choice theory” are often used interchangeably (Hardin 2011 [2009], p. 99). This leads to some misunderstandings and problems. I find it necessary that scientific knowledge should be differentiate between methodology (formal knowledge and tools) on the one hand and theories (content knowledge) on the other. In addition to the rational choice approach, it also makes a fundamental difference whether one wants to use the rational choice approach to produce truth-apt statements about the political reality or to generate technical rules that are not truth-apt (Table 1, the fourth row, level of ideals and properties).

Practical methodology (normative, pragmatic and technical)

Now the possibility and necessity of a genuine practical political science, and certainly also a practical social science, will be outlined, building on the fundamental distinction between empirical and practical social sciences made by Weber (2011 [1904]) and between theoretical and practical sciences made by Wieland (1986). This practical methodology establishes a practical political science that is also in the Aristotelian and Kantian tradition. Neither the normative rational choice approach of the naturalists nor the applied phronesis of the perestroikans are rejected, but these are only very small parts of a much broader practical methodology.

The practical tradition works with a practical methodology, with the aim of formulating practical norms and regulations with which the political system can be changed. In the following, the basics of a practical methodology will be explained, before the normative, pragmatic and technical methodology are further developed on this basis. Practical norms and regulations are generated in three different discourses (normative, pragmatic and technical).

Practical discourses

Höffe (2009 [2007]) has presented a complex ethical rational design that draws on the entire Western ethical tradition. Höffe succeeds in proving that two of the most important types of Western ethics, the eudaemonistic/Aristotelian and the deontological/Kantian ethics, are not necessarily mutually exclusive but can be viewed as complementary. The basis for this is a practical methodology, which consists of three different methodologies and goes back to Aristotle and Kant.

At the lowest level, one evaluates ways and means for their suitability for any purpose or goal. By means of technical rationality, technical imperatives are generated within these discourses, be it technical individual rules or social-technological regulations (Höffe 2009 [2007], p. 23).

At the second level of evaluation, the objective that is only assumed at the lowest level is now assessed. Pragmatic rationality is used to generate pragmatic imperatives in the pragmatic objective discourse, be it individual-pragmatic rules or socio-pragmatic regulations (Höffe 2009 [2007], p. 24).

The third and highest stage in practical philosophy is the normative discourse of value. Here ethical-moral rationality is used. Höffe distinguishes between virtuous morality and ethical-moral norms on the one hand, and legal morality, (political) justice and legal norms on the other (Höffe 2009 [2007], p. 26).

Normative or value discourses

In a practical political science, one can first of all justify maxims of action (constitutional norms) and political values within normative value discourses. Here the question in the foreground is: Why should something be done? In this case, the political maxims of action which are decisive for the standardization or regulation of the political system as a whole or of a policy area should be discussed.

Pragmatic objective discourses

Pragmatic reasons, goals and purposes are to be justified within pragmatic objective discourses. The aim is to pragmatically tackle the goals set in normative discourses.

Technical discourses

How should something be done? Technical justifications are generated within technical means discourses. The objective is concrete technical means for the regulation of political problems.

However, it is not possible to derive technical regulations or individual instructions for action from pragmatic objectives or from maxims of action. Subsumption under action strategies or action maxims is not possible, although this is claimed and demanded by normative-ontological political scientists (details Lauer 2017).

Complementarity between three methodological traditions

Empirical explanative-prognostic methodology as a political methodology is not enough. Before one can determine a causal relation between events, one must describe these events (von Wright 1971). Political phenomena must be described and politically important symbols interpreted. This is not possible without an empirical-descriptive methodology. To recognize the political reality, therefore, a logical-mathematical as well as a linguistic-interpretive research methodology is necessary; only then is an adequate empirical (descriptive, explanatory and prognostic) knowledge (descriptions, explanations and forecasts) generated.

The methodological science war (Methodenstreit, Lauer 2017) is conducted in revolutionary mode. If a complementarity between the methodological traditions discussed above is recognized, then this “science war” (Flyvbjerg 2001, p. 1) can be overcome. There is complementarity between the descriptive and explanatory-prognostic traditions as well as between the empirical (descriptive, explanatory and prognostic) and practical (normative, pragmatic and technical) traditions.

The political methodology must first be explained, explicated, clarified and reconstructed. Of course, existing methodologies must be further developed and complemented by innovations. Progress is possible and meaningful only on the basis of tradition. Tradition and progress must be connected.

A practical methodology differs in principle from an empirical methodology. With an empirical methodology adequate descriptions, explanations and predictions about political reality can be generated, but not practical ones can, not even technical-instrumental regulatory proposals. Genuinely practical discourses (value, target and means discourses) require a practical methodology that is complementary to an empirical methodology. The arguments for this have been explained in detail elsewhere (Table 1, Lauer 2017).

A practical political science that supplements an empirical (descriptive, explanatory and prognostic) political science with a practical (normative, pragmatic and technical) methodology is the appropriate place to discuss political-practical issues. A practical methodology can justify a practical social science and thus also a practical political science.

I am only suggesting that within political science, all three traditions should be used complementarily. Because of their specialization, political scientists can only bake small rolls. As a rule, one cannot even use all methodological approaches to determine causalities in one article. Completely different methodical approaches and methods are used for the identification of correlations, for causal regularities and then for the determination of causal mechanisms (see “Causal reductionism”). These methods are also used complementarily in political science, but due to the complexity of the questions, this is done in different articles and even by different scientists.

What tools do political scientists use?

What are 3 major aspects of political science?

What are the 5 methods of political science?

What are 3 disciplines that political science draws from?