Abstract
Pragmatism is ever more popular amongst those who study international relations. Its emphasis on practice is generally acknowledged as a defining characteristic. There is, however, a general tension within pragmatist thought concerning practice, for pragmatism may emphasize the theorizing of practice. It is, then, distinguished from other theories in International Relations (IR) such as neo-realism or constructivism as a contender in their midst. We delineate a pragmatist theory of IR in the first part of this article, but insist on going beyond merely establishing the next paradigm, for pragmatism may also emphasize the practice of theorizing. Theories are, then, considered different tools useful for dealing with the social world. This will be corroborated in the second part by a close reading of William James. Finally, we submit that instead of a paradigm war, a metaphor such as that of the Papini hotel is needed in IR — a metaphor that accounts for theory competition without neglecting the limitations set by the practice of theorizing itself.
To treat abstract principles as finalities, before which our intellects may come to rest in a state of admiring contemplation, is the great rationalist failing. (William James, Pragmatism, 1987a[1907]: 527) But the moment philosophy supposes it can find a final and comprehensive solution, it ceases to be inquiry and becomes either apologetics or propaganda. (John Dewey, Logic, 1986[1938]: 42)
Introduction
The current situation in the study of international relations has been described by metaphors such as that of a ‘paradigm war’ or a ‘battle for the last variable standing’ (see Bauer and Brighi, 2009: 1; Buzan and Little, 2001: 32; Hellmann, 2000: 170; Herborth, 2004: 63; Isacoff, 2002: 604; Owen, 2002: 654; Smith, 2008: 726). Instead of a critical and open dialogue (see Hellmann, 2003), paradigm wars seem to chart the course of normal science in IR (see, for instance, Legro and Moravcsik, 1999; Vasquez, 1997).
If we follow this imagery, the increased interest in pragmatism (Hellmann, 2009: 638) might be interpreted as the emergence of a new contender on the battlefield. Earlier, there have been references to classical pragmatist thought in the writings of some of the most prominent IR theorists — explicit as in Wendt’s (1999: 170f.) recourse to Mead’s symbolic interactionism or implicit as in Waltz’s (1979: 11ff.) and Kratochwil’s (1989: ch. 7) critiques of deduction and induction. Yet, as a perceivable label, pragmatism has fully arrived in IR perhaps only with the Millennium special issue conference on ‘Pragmatism in International Relations Theory’ in 2002 (see, inter alia, Bohman, 2002; Cochran, 2002; Haas and Haas, 2002). 1 Pragmatism has since been prominently represented in IR by Hellmann (2009, 2010), Kratochwil (2007a, 2009) and Friedrichs and Kratochwil (2009), but also, as far as the ‘practice turn’ is concerned, by Neumann (2002), Büger and Gadinger (2007), Adler (2008) and Pouliot (2008).
The pragmatist tradition, however, has left us with several sometimes overlapping and sometimes mutually exclusive versions. Besides the pragmaticism of Charles Sanders Peirce (1839–1914) and those versions of other classical pragmatists such as William James (1842–1910), John Dewey (1859–1952) and George Herbert Mead (1863–1931), there are today a host of philosophers traded under labels such as logical pragmatism, 2 analytic pragmatism (see Brandom, 2008), process pragmatism, 3 prophetic pragmatism (see West, 1989), linguistic pragmatism, neo-pragmatism (Rorty, 1982), neo-classical pragmatism (see Haack, 2009) and lately even neo-neo-pragmatism or new pragmatism (see Misak, 2007). It has been observed that pragmatism ‘does not even come close to possessing a common method or style, or to constituting a single web of discussion’ (Glock, 2008: 85). 4 There is, however, a tension that runs through these different versions of pragmatism and that has to do with the pragmatist emphasis on practice, which as such is not disputed. Yet, pragmatists seem to draw different lessons from this emphasis. For some — and this is perhaps the majority in IR — practice is what needs to be theorized, that is, what a theory has to be about; for others, the lesson is rather that any theorizing is always a practice itself. This gives rise to two different, and we suggest perhaps complementary, understandings of pragmatism, each drawing on a different notion of theory. One is to understand pragmatism itself as a theory which seeks to outdo all other theories by theoretical virtue only or which is held to precede — logically or for some other reason — all other theories. The other is to view pragmatism as a philosophy in which theories — regardless of whether pragmatist or not — are regarded as so many tools and instruments as are useful for dealing with the social world.
The Italian writer Giovanni Papini (1881–1956), by way of James, has given us the metaphor that we believe most aptly captures this tension (Papini, 1961 [1905]: 405). In James’s words (1987a [1907]: 510):
[Pragmatism] lies in the midst of our theories, like a corridor in a hotel. Innumerable chambers open out of it. In one you may find a man writing an atheistic volume; in the next some one on his knees praying for faith and strength; in a third a chemist investigating a body’s properties. In a fourth a system of idealistic metaphysics is being excogitated; in a fifth the impossibility of metaphysics is being shown. But they all own the corridor, and all must pass through it if they want a practicable way of getting into or out of their respective rooms.
Applying this metaphor to pragmatism in IR, we suggest that some pragmatists understand their work in terms of establishing a primacy-seeking theory (pragmatism) as if they were in one of the hotel rooms; others understand pragmatism alongside Papini and James as a philosophy, a method or an ‘attitude’ (Hellmann, 2002), that is, as the corridor of the same hotel, turning theories into handy tools, instruments or vocabularies. Faced with a scientific puzzle or a problematic situation, the pragmatist theorist in the hotel room would always opt for a pragmatist theory, while the pragmatist in the corridor would do their utmost to find the optimal theory for the situation at hand, even if it means opting for a non-pragmatist theory. The point of the latter is that a neo-realist theory might of course be pragmatically successful to engage with, say, the stability of bipolar international systems. The pragmatist theorist would disagree on theoretical grounds, and there is nothing wrong with this disagreement as such. Yet, we submit that the study of IR should work towards a self-understanding where theorists make it clear that they are busy working inside one of Papini’s hotel rooms, while they must necessarily pass through the corridor should they wish to get out of the room and have their theories set to work. In short, rather than getting bogged down in the imagery of a paradigm war, we should follow the more comprehensive and colourful picture of the hotel metaphor.
Following up the first of the two different emphases on practice within the pragmatist tradition, the first part of this article deals with Prospects for Pragmatist Theory in IR. Inspired by Peirce, Dewey and Mead it pursues the project of pragmatism as a theory inside one of the rooms at the Papini hotel. The second part, which offers a Jamesian perspective on pragmatist theorizing of theories and thus focuses on Pragmatist Limitations of Theory in IR, is instead firmly located in the corridor of the Papini hotel. It not only addresses the status of primacy-seeking IR theories, but also that of the pragmatist version thereof. Some consequences of both pragmatist emphases on practice are drawn in the Conclusion.
Inside a hotel room: Prospects for pragmatist theory in International Relations
The approach to IR developed in this part seeks to highlight two strong points of pragmatist theory over against other theories or paradigms in IR. Both address peculiarities of the way ‘paradigm-driven research’ (Katzenstein and Sil, 2008: 110) is focused on single variables or core categories such as — in the broadest terms — the distribution of capabilities, power or security in the case of neo-realism, cooperation in the case of neoliberal institutionalism, or norms in the case of constructivist approaches. The first point holds that unlike other theories which are restricted to inductive and deductive methods (or modes of reasoning), the chance to gain new insights is maximized in pragmatist theory by means of Peirce-style abduction. The second point holds that conceiving the world in terms of variables presupposes a socio-theoretical understanding of the subject matter that discounts major characteristics of practice, that is, human action.
Enforcing abduction
Adherents of variable-centric research usually formulate favourite variables in advance and impose them on their subject matter. As a consequence, they merely find out whether or not a distinct set of variables helps to make sense of their research puzzle. Such a research process is organized like a binary wager: the scholar involved makes a bet on behalf of the favoured variables. Yet, even if the wager wins and a ‘positive’ role for the variables at stake can be established, the decisive question of whether or not something else could make still more sense remains unresolved. 5 This problem is inherent in variable-centric approaches as they are restricted to the use of inductive and deductive modes of reasoning. According to Dewey (1986 [1938]: 415), scientific procedures along simple lines of induction and/or deduction are more often than not ‘forced into the straitjacket of irrelevant preconceptions’ and unfit for social inquiry. To escape this dilemma, a third mode of reasoning is required, which, referencing Aristotle, Peirce calls abduction.
In formal terms, abduction stands for ‘the process of forming an explanatory hypothesis’ and is considered ‘the only logical operation which introduces any new idea’ (Peirce, 1998 [1903]: 216). In a broader sense, abduction describes any creative coping with problematic situations both in the academy and everyday life. The latter aspect is not uncontroversial, though (see Hellmann, 2009: 641). Since Peirce himself had struggled with the concept for nearly 40 years, this cannot be much of a surprise. Starting from ‘a reduction of a manifold to unity’ in 1868 (Peirce, 1992a [1868]: 34), Peirce in 1878 formally conceptualized abduction (or hypothesis as he initially termed it) as one of three kinds of reasoning based on a syllogism composed of rule (all men are mortal), case (Enoch and Elijah were men) and result (Enoch and Elijah were mortal). Whereas deduction ‘is merely the application of general rules to particular cases’, the two inversions of deduction are defined as ‘the inference of the rule from the case and result’ (induction) and — abduction — as ‘the inference of a case from a rule and result’ (Peirce, 1992c [1878]: 187f.). 6 Twenty-five years later, Peirce in his Harvard Lectures on Pragmatism (1998 [1903]) created what became a standard in qualifying abduction: the introduction of any new idea. Grasped as the suggestion ‘that something may be’ he distinguished the concept both from ‘necessary reasoning’ which ‘proves that something must be’ (deduction) and from ‘a course of experimental investigation’ which ‘shows that something actually is operative’ (induction; Peirce, 1998 [1903]: 212ff., italics in original). Ultimately, in a letter to the Italian pragmatist Mario Calderoni (1879–1914) Peirce (1966 [1905]) considered the three modes of reasoning as consecutive phases. After abduction (‘in examining a mass of facts’) has suggested a new theory, the consequences of this theory are deduced by ‘necessary reasoning’ (deduction). Since these consequences take the form of experiences that confront us when we act, they can be tried by means of ‘experimental research’ (i.e. inductive reasoning; see Peirce, 1966 [1905]: 167f.). Hence, a theory’s predicaments understood as consequences for our action can either prove their worth in practice (be it research or everyday practice) or fail so that abduction is required once more and the cycle of reasoning and research starts again.
To IR scholars the concept of abduction is in no way unfamiliar as Friedrichs and Kratochwil (2009: 709) point out, although they bemoan that ‘none has ever thoroughly specified its meaning’. Particularly accentuated, however, are the creative, associative or problem-solving dimensions of abduction as well as its engendering of new insights (see Friedrichs, 2009: 647; Hellmann, 2009: 641; Onuf, 1989: 100, 126; Rytövuori-Apunen, 2009: 644f.). When it comes to discriminating between abduction and induction, there are instructive differences among IR interpreters of Peirce. The bid that induction is ‘a much stronger kind of inference’, whereas abduction is ‘a bolder and more perilous step’, for instance (Peirce, 1992c [1878]: 198, 192), has been taken up by Friedrichs and Kratochwil (2009: 725) who aphoristically conclude that ‘abduction is a risky endeavor’, but that ‘the risk inherent in abduction is also the price for learning something genuinely new’. On the other hand, these authors localize abductive reasoning ‘at an intermediate level’ (2009: 715) — a metaphor that suggests security and the happy medium rather than risk and peril. Unlike Friedrichs and Kratochwil who conceive abduction as ‘a comparative case study method’ somewhere between imposing ‘an abstract theoretical template’ and ‘inferring propositions from facts’ (2009: 719, 715), Onuf focuses on a re-conceptualization of Peirce’s three modes of reasoning. Contrary to the founding father of pragmatism who back in 1869 grasped both induction and abduction as ‘inference from the parts to the whole’ (Peirce, 1992b [1869]: 78), Onuf (1989: 99) holds deductive reasoning to proceed ‘from a whole to its parts’, inductive reasoning ‘from parts to whole’ and abductive reasoning to be ‘a matter of leaping from one whole to another without having to proceed down to the parts and up to the next whole’. Regardless of the appropriateness of parts and wholes as basic categories, Onuf by referring to a leap congenially builds on the dimension of risk, boldness, creativity, innovation and spontaneity, which is not only central to abductive reasoning, but has also been associated by Mead with the ‘I’ — that part of the self through which ‘we surprise ourselves by our own actions’ and create the new (Mead, 1974 [1934]: 174; cf. Oevermann, 1991: 307ff.).
Against this background, our pragmatist alternative to variable-centric research starts from the socio-theoretical assumption that researchers cannot directly comprehend social reality or, put differently, the ephemeral here and now that is surrounding them. Ephemeral social reality, to be conceived, has to be fixed first (see Dewey, 1986 [1938]: 231). In the social sciences, this act of fixation implies constituting a subject matter by means of the relevant (in terms of research question) extracts of social reality. Following Mead (1974 [1934]: chs 9–11), any of these extracts is constituted by objective meaning 7 and, therefore, can be read as text — no matter whether it is present in written form, like a communiqué, a diary, an interview and so on, or in non-written form such as a picture, a film, a sculpture, a landscape or something else. The question of how these extracts of social reality are read brings the aspect of abduction back in. Instead of subsuming them under categories or favourite variables that were formulated in advance, such texts have to be taken seriously as autonomous expressions of social reality and thus need to be read patiently in their ‘own’ language, sequence by sequence (see Oevermann, 2000). In this way, pragmatist researchers, unlike those who bet on a certain variable or core category to play or not to play a role in a specific research context, refrain from merely reproducing what they already knew before and leave sufficient space for being genuinely surprised by their ‘data’. Or as Rytövuori-Apunen (2009: 645) and Deibert (1997: 176) put it, abduction can ‘save us from becoming prisoners of privileged professional bodies of knowledge’ and from ‘circular confirmations of a pre-existing set of lenses’.
Against these discriminations, at least three objections can be raised:
Objection 1: The presented picture of variable-centric research is a caricature. Variable-centric research processes are much more complex and rest upon a series of pre-tests in which favourite variables or core categories are cautiously developed, formulated and reformulated.
Even if variable-centric research makes possible new insights, this possibility is only exceptional, whereas it happens regularly, systematically and in a sense is enforced in pragmatist research processes. Besides, if formulating and reformulating the core categories or variables really were the creative (and thus in a sense abductive) part of variable-centric research, would it not be better to relieve it from the opaque pre-test part of the research process and put it to the centre of action — just as is the case in research processes that follow pragmatist approaches?
Objection 2: A methodology based on abduction opens the floodgates to immunization by blurring the boundary between a researcher’s presuppositions and her findings.
Any researcher, regardless of the preferred methodology, should explicate presuppositions at the beginning of the research process. The major problem, however, is not the presuppositions that researchers can formulate in advance, but those deep-rooted beliefs in the realm of the pre- or unconscious. These beliefs only come to the fore, if at all, during the research process — for example, in the process of interpreting texts. Therefore, the problem of presuppositions is best handled if the research process is as transparent as possible. This is warranted either by attaching to the publication the ‘memos’ written during the research process to the publication or by making one’s interpretations as such part of the whole publication. In so doing, any member of the community of scientists can intersubjectively check the intellectual moves that a researcher who proceeds in a pragmatist way has perceived to have made.
Objection 3: Like any other research, the research process based on abduction does not work without pre-established categories.
Pragmatist methodologies do not pretend to approach social reality without pre-established concepts since all reasoning rests on such categories. Whoever interprets a text (or, more generally, whoever copes with the world in some way) draws on experiences made before in a continued process of inquiry. In terms of pragmatist philosophy, these categories or experiences can be grasped as routines, that is, solutions to unprecedented and problematic situations (see Dewey, 1986 [1938], 1991 [1927]) that have proven their worth in (research and everyday) practice so far. As life relentlessly goes on, however, routines do not preclude the emergence of new problematic situations in which the categories or experiences taken advantage of before fail. Being creatively established by means of abduction in the process of coping with one kind of problematic situation, routines collapse in other kinds of problematic situations so that abduction is required once more. Hence, Rorty’s interpretation that ‘“abduction” names something that everybody does all the time’ (quoted in Hellmann, 2009: 641) hits the point insofar as everybody has at least to be prepared for doing abduction, for coping with the collapse of routines (be they pre-established categories or other kinds of experiences) all the time. The contribution of pragmatist methodologies in this context is twofold: they enforce abduction by intentionally simulating problematic situations under the promising conditions of the ‘ivory tower’ where researchers are relieved of the everyday-practical pressures to find new routines swiftly and without leisure (see Oevermann, 2000: 154ff.) and they, instead of hastily subsuming the subject matter under very specific and highly contested categories located in the core of ongoing debates, resort to concepts that — like practice — constitute social life in more general terms.
Focusing on practice 8
Human action, practice, is the primary concern of a pragmatist approach to IR. A continuous stream of acts, practice has neither a definite beginning nor a definite end. To deal with it, though, we conceive practice as made up of sequences of acts (see Oevermann, 2000: 64ff.). Any such act is linked to a preceding one on which it has followed and, itself, opens a scope of possibilities for successive action. Hence, any sequence amidst the endless stream of practice is grasped as both closing (by performing an act) a scope of possibilities for action that has been opened before and opening a new one. This repeats whenever two sequences are linked, that is, with any successive act.
Examining practice in its course, two aspects of a sequence need to be distinguished analytically: the rules which constitute the scope of those possibilities for action that could meaningfully (precede and) follow on any given act as well as the (disposition-)factors based on which one of these possibilities is selected. The first aspect, meaning-generating rules, is held to generate meaning (in the form of meaningful ‘connections’ for action) similar to the way Wilhelm von Humboldt conceived language, that is, by infinitely making use of finite means. From this perspective, practice indefinitely continues as any act that opens a scope of possibilities for succeeding ones reproduces the conditions of action. The second aspect of a sequence, however, ensembles of disposition-factors, accounts for the specificity of human action. It encompasses the whole ensemble of biological, psychic and social ‘dispositions’ which are usually split along the lines of competing disciplines, reified as motives, attitudes, values and so on and operationalized as single variables. Put differently, it is due to conscious or unconscious selections along the ever specific ensembles of dispositions that human beings, individually or on behalf of distinct collectives, make their decisions.
Against this background, the concepts of practice (human action), being constituted by rules, sequentiality and sociality, can all be treated interchangeably. Practice, to begin with, is considered as constituted by rules. Even more: it is considered social action inasmuch as it is constituted by rules. This is made possible by the assumption that rules refer to sequentiality, that is, a sequence of acts. What appears to be a single act is thus not an isolated move, but one link in an infinite chain of action. As a consequence, practice is not grasped as constituted by single or individual acts, but held to bring about what then may be ascribed to individual human beings. In contrast to individualist approaches, single acts are thus considered a feature (or an abstraction) of social action and not vice versa. For this very reason, the term sociality has a considerable advantage over inter-action and intersubjectivity in terms of adequately signifying the sphere in which single acts or individuals (subjects) refer to each other as (always already) socially constituted. The terms interaction and intersubjectivity suggest that a common context or environment, somehow in addition, needs to be established to link formerly isolated acts or subjects to each other. The pragmatist notion of sociality holds on the contrary that individual acts and subjects are brought about by and thus follow on interaction and intersubjectivity. Subjects do not constitute themselves in isolation, neither biologically nor socially. Instead, they do so with reference to other members of the species (which, from the individual perspective, are a given). To wit: sociality (and hence practice) and not an individual act or subject is the starting point for any pragmatist analysis. 9
Among IR pragmatists, this primacy of practice is widely accepted (see Friedrichs and Kratochwil, 2009: 711ff; Hellmann, 2009: 639; Kratochwil, 2007a: 11ff.; Rytövuori-Apunen, 2009: 643f.). Moreover, despite his refusal to offer ‘some “working” definition’ (2009: 19), Kratochwil’s conjecture that ‘actions of agents are meaningfully oriented toward each other’ (2008: 447, italics in original) perfectly fits the understanding of practice presented above. Nevertheless, most IR pragmatists mainly focus on criticizing the discipline’s ‘nearly hypertrophic concern with epistemological issues’ (Kratochwil, 2007a: 1), ‘the application of standardized methods and techniques for the purpose of definitively verifying or falsifying truth claims’ (Katzenstein and Sil, 2008: 115) or related aspects concerning knowledge generation driven by (logical) positivism (see e.g. Cochran, 2002; Friedrichs, 2009; Hellmann, 2010; Puchala, 1990, 1995). As a consequence, systematically unfolding the implications of a concept of practice as the dynamic interrelation between meaning-generating rules and ensembles of disposition-factors markedly adds to pragmatism’s intellectual surplus as an approach to IR. In the remainder of this subsection, this added value is presented along three dimensions: (i) the nexus of crisis, creativity and autonomy; (ii) historicity, contingency and change; as well as (iii) the relation between rules and norms.
(i) Crisis, creativity and autonomy
Understood as sequentially structured practice, any extract of social reality — be it a speech or a communiqué, intergovernmental negotiations, student protests or a military operation — shows a specific concatenation of choices. At every sequence, the practitioners (actors) involved have, consciously or not, realized one possibility for action among many. As any such choice stands for a decision, it makes sense to characterize practice, or (human) life in general, as making decisions. In many cases, human beings are not aware of the alternative possibilities for action since they follow routines, that is, they cope with the world in the way they (more or less successfully) have already done before. When no such routine is available, however, human actors face a crisis, a new or problematic situation in which they have to rely on their own creativity (Dewey, 1991 [1927]). Then, practitioners — members of government or ‘ordinary’ citizens, employers or employees — make genuine decisions on a course of action as it is unknown in advance whether these decisions will turn out to be right or wrong, good or bad, wise or unwise, successful or unsuccessful. In case such a decision, such an attempt to solve a problematic situation, turns out to be or is deemed to have been right, good, wise or successful, it may become a routine. Routines can therefore be grasped as solutions to problems of action (crises) which have proven their worth in practice. 10 At the same time, it is by making these kinds of genuine decisions — by departing from routines and performing creative acts into an open future — that practitioners gain autonomy. Autonomy is thus inextricably linked to the capacity of creative problem-solving. This implies that autonomy can also be lost by taking too much time to address a problematic situation, for instance. Even a non-decision yields consequences (and for that reason may be considered a decision, too). Put differently, there is a necessity to decide. On the other hand, everyone is, due to his or her social embeddedness, committed to justify a decision, to give reasons for it (even if only to oneself). Accordingly, Oevermann (1991: 297) suggestively speaks of a dialectics of decision-necessity and justification-commitment.
Regarding its implications for the study of international relations, the pragmatist concept of creative problem-solving transcends both rationalist and constructivist accounts. Rationality, to begin with, is revealed to be a socio-scientific core category of limited value only. In line with Kratochwil’s critique ‘of the clear yes or no scheme of logic that does not allow for a third category in between the “is” or “is not”’ (2007a: 10), decisions in crises, with no routines on hand, are neither rational nor irrational. Practitioners of course do hope that their making and doing in problematic situations may be successful and turn out to be rational, but it is only with hindsight that they are able to judge the worth of their decisions. In a similar way, this argument holds vis-a-vis constructivist approaches resorting to norms. Due to its focus on the performance of creative action into an open future, a pragmatist approach to IR can better cope with situations in which norms are repudiated. Moreover, pragmatism, due to its focus on genuine decisions in crises when no routine is available, can cope with situations in which a norm does not (yet) exist or does not exist anymore — a move totally unavailable to approaches starting with norms. In sum, pragmatism, by taking the primacy of practice seriously, offers a socio-theoretically more thorough alternative to the narrow conceptions of the rationalist homo oeconomicus and the constructivist homo sociologicus. In terms of the agent–structure debate (see Wendt, 1987), both approaches predominantly emphasize what constrains human action, but neglect what enables it. Practice, potentially creative social action, means much more than exclusively being either forced into rationality by one’s environment or trained by norms, though. Human actors ‘are not cultural dopes, who directly translate prescriptions into practice’ (Kratochwil, 2008: 453). On the contrary, they have the capacity to gain and to exert autonomy by creatively coping with crises of decision.
(ii) Historicity, contingency and change
Conceived as a succession of openings and closings of possibilities for action, practice refers to a dialectics of possible and realized acts. This implies that whoever examines a distinct extract of social reality as a sequence of acts (each selected from a scope of several alternatives) is not only confronted with the concrete course of action or history of that extract (or case), but also with what it could have become if other options had been realized. On the one hand, this dialectics of possible and realized action illuminates the meaning of historicity and contingency of practice. It could all have happened differently. That it happened the way it did — and not differently — on the other hand, reminds researchers to take the concrete history of their cases seriously as an expression of autonomous decisions. Drawing on Mead and Hawthorn, Hellmann (2010: 167ff.) turns this into an argument against the demand by King et al. (1994) to systematically simplify the mode of inference. The better an explanation, Hellmann writes instead, the more it would depend on plausible alternatives (counterfactuals). Consequently, comprehensiveness and complexity matter more than simplification and economy when it comes to explicating practice.
Besides, the presented concept of practice makes it possible immanently to contend with change. Practitioners involved in a distinct extract of social reality can either reproduce or transform the patterns by which they make their decisions (in accordance with their ensembles of disposition-factors). Whenever these patterns are transformed, this means change. Accordingly, change — potentially — happens at every sequence and is not grasped as something ‘separate’ which is thrown on life from an imaginary outside. Change emerges from the present, from what already exists (see Oevermann, 1991: 304ff.). It is brought about by human actors’ capacity for creative action which, represented by Mead’s ‘I’, does not float freely, but is dialectically interwoven with society (represented by Mead’s ‘me’). Like the self, creativity and change are thus constituted ‘in continuous dialogue with society’ (Mead, quoted by Katzenstein and Sil, 2008: 115).
(iii) Rules and norms
Finally, the proposed notion of practice adds to the research agenda of IR by addressing the relation between rules and norms — a topic that Kratochwil (1989) and Onuf (1989), the ‘avant garde corps’ (Puchala, 1990: 71) of our discipline, already dealt with in response to regime theory more than 20 years ago. As to social theory, however, the dialectical relations between the two central aspects of practice — meaning-generating rules and ensembles of disposition-factors — are modelled on a couple of more prominent examples. They resemble the relations between Searle’s constitutive rules (which bring about what they deal with) and regulative rules (the matter which they regulate is logically independent of them), Chomsky’s competence and performance, as well as de Saussure’s langue (language as a system of rules) and parole (language as communicative practice). Moreover, as Herborth (2004) has shown, the relation between meaning-generating rules and ensembles of disposition-factors corresponds to the very dialectics that Wendt (1987) introduced into IR as structure and agency. Against this background, two kinds of questions applicable to any practice can be distinguished. The first question asks whether a distinct act is well structured according to the (meaning-generating) rules that have brought it about. The second question, directed towards the acceptability of the very act, asks whether it is well structured in terms of certain kinds of norms. If the US Congress, to give a very short and thus somewhat simple example, adopted a joint resolution titled ‘Let’s conquers Canada’ it would violate both the rules of grammar that brought this phrase about and the norm to refrain from violence in international affairs. Pragmatism, therefore, allows addressing both analytic and normative questions.
In this context, it is crucial to bear in mind that the differentiation between rules and norms, constitutive and regulative rules, as well as meaning-generating rules and ensembles of disposition-factors is not an ontological and absolute but an analytical one. It is the different effects of rules that matter (see Kratochwil, 2008: 457). Even Onuf, who rejects the difference between constitutive and regulative rules (1989: 50ff. 85), re-approaches this position when he concedes that rules ‘perform constitutive and regulative tasks’ (1989: 116). In consequence, whether constitutive or regulative effects are taken into account and what is considered a rule or a norm depends on the research interest. A widely accepted norm such as nuclear non-proliferation without difficulties can be analysed as a meaning-generating rule that constitutes possibilities for action. On the other hand, ‘not all norms exhibit rule-like characteristics’ (Kratochwil, 1989: 10). A critical asymmetry between both concepts remains, namely: any appeal to norms for shaping a concrete practice takes for granted objective meaning which, in terms of Mead (1974 [1934]), has previously been brought about by (meaning-generating) rules. This implies that those who repudiate the norm of nuclear non-proliferation nevertheless act in accordance with the rules that generated the possibilities to ignore, repudiate, modify or comply with it. The repudiators’ action, of course, has consequences (meaning) which need(s) to be examined. In turn, a few rules cannot be analysed as norms since they are irreducible to a preliminary type of meaning-generating rule (which, as we have seen, would be constitutive for a norm). Candidates for such socio-theoretically central, universal rules are the rules of logical inference, of generating speech acts in terms of Austin and Searle, of universal grammar in terms of Chomsky, of moral competence in terms of Piaget, as well as the rules of social cooperation understood as reciprocity in terms of Lévi-Strauss (see Oevermann, 2000: 66).
What makes universal rules so particularly important, however, is not their presumed unchangeability, as their ‘valid’ interpretation indeed varies from time to time and from place to place. But it is impossible to criticize meaningfully universal rules as such. Whoever wanted to bring down the rules of generating speech acts would need to generate speech acts; and whoever wanted to bring down the rules of social cooperation would need to cooperate. From this perspective, there is a kind of demarcation line between pragmatist thinking and poststructuralist, postmodern or radically constructivist approaches. Recognizing universal rules and the difference between constitutive and regulative effects, pragmatists believe that meaning can temporarily be fixed and thus bears routines, but they also stress human beings’ potential for creatively transforming meaning. Dismissing universal rules and the difference between constitutive and regulative effects, postmodernists, on the other hand, hold that meaning can never be fixed. Human beings thus face a life in perpetual crisis. Consistently (and, in theoretical terms, following from the rejection to differentiate between a potential and its actualization), everybody engages in abduction all of the time and permanently (re-)constructs social reality. Pragmatists, however, are content with human beings’ potential for abduction, creative problem-solving and constructing social reality — a potential which is activated in moments of crisis, when human beings get the chance to gain autonomy by making decisions. 11
How does this pragmatist reliance on universal rules relate to Onuf’s (1989: 113) critique that structuralists such as Piaget, Lévi-Strauss and Chomsky consider human beings ‘not just competent, but genetically mandated, to’ act morally, to cooperate and to form grammatically well-structured phrases? We share these reservations, but see no need to discard the notion of competence as this would make it difficult to distinguish a potential (autonomy, creativity, etc.) from its actualization. Without aspiring to categorically substitute nurture for nature (which would only mean the resurgence of an overcome antagonism), we believe that human beings make competences their own ‘in dialogue’ with their respective social environment, their ‘significant others’, in early childhood. In this way, structuralism’s biological determinism can easily be replaced by the pragmatist dialectics of self and society.
In sum, the proposed pragmatist concept of practice allows for analysing human action independently of the tight borders between paradigms and disciplines. Instead of forcing the diversity of the world into the straitjacket of pre-established, isolated and reified variables such as power, interest, security, military capabilities, law, identities, norms or motives, whatever happens (in ‘international relations’ or elsewhere) is, quite simply, understood as the actualization of options out of scopes of possibilities for action. This concept of practice as a concatenation of selections realized by the actors involved in line with their autonomy is perfectly suited as the starting point for a sequential examination of social reality since the historicity of any case, of any concrete practice, emerges from alternatives that have not been actualized.
Inside the corridor: Pragmatist limitations of theory in International Relations
The pragmatist emphasis on practice may also be understood as pointing out some important limitations of any theorizing in IR, pragmatist or otherwise. Both emphases mentioned in the introduction find themselves acknowledged in IR literature, although primacy is clearly given to the theorizing of practice over against reflection on the practice of theorizing. For instance, a recent contribution claims (Bauer and Brighi, 2009: 2):
Pragmatism invites IR to abandon the search for an epistemological purity which the field simply cannot afford and should not pursue, and to orient itself firmly towards the study of practices and problematic situations.
Yet, giving up ‘the search for epistemological purity’ should not imply that one may dispense with epistemological problems; it simply means not to search for any neat and final answers. In what follows, we will illustrate the identified tension between the two emphases on practice and offer a reading of James that is supportive of understanding pragmatism as emphasizing the implications of practice on theorizing itself. We hark back to classical pragmatism in its Jamesian version for the following reason: James, although frequently acknowledged by pragmatist theorists in IR, has been markedly less often and less intensely discussed than Dewey, Peirce or Mead. Yet, we believe that James has something interesting to contribute to IR, particularly if it comes to pragmatist limitations of (pragmatist and other) theory in the discipline. Quite obviously, James was writing in a time different from ours, and that brings some issues with it which we should address. For instance, James wrote when the linguistic turn was about to take shape (depending on whether one sees the turn happening with e.g. Frege, Moore, Wittgenstein or Heidegger), but he himself can hardly be considered to have taken the turn himself. As a commentator has recently put it: ‘Pragmatism, for James, is a philosophy of experience — and not a philosophy of language, narrative, or vocabularies’ (Stuhr, 2010: 199). James, however, ‘is not averse to the linguistic turn’ and quite readily may be updated in this regard as he has indeed been updated by, for example, Rorty, as another commentator remarks (Lamberth, 1999: 213). Obviously, what kind of update is required for what part of James is itself a question that depends on one’s philosophical and theoretical standpoint and thus a question that might find radically different answers.
A Jamesian reading of the practice of theorizing
In his The Principles of Psychology (1950 [1890]: 312), James makes two points, which anticipate his later pluralistic humanism of Pragmatism: A New Name for Some Old Ways of Thinking (1987a [1907]). First, he advances the idea of several rival theories which in some sense possibly can do ‘equally well’; as he will later say, ‘no theory is absolutely a transcript of reality, but … any one of them may from some point of view be useful’ (James, 1987a [1907]: 511). Second, he describes the success of the theory that will assert itself as depending on its use, be it in terms of a satisfactory account of experience, interest or needs. The second point leads to his pragmatist theory of truth as verifiability (see James, 1987a [1907]: 572ff.). If we focus on the first point, for a moment, that is, on how James conceived of theories, we will the more easily see how the two points hang together and how they give rise to the tension to which we wish to allude.
Speaking of words (e.g. ‘God’, ‘Matter’, ‘Reason’) as much as of theories, James emphasizes that one must bring out their ‘practical cash-value’ and that they are ‘less a solution, than … a program for more work’ (James, 1987a [1907]: 509). He writes (James, 1987a [1907]: 509f. italics in original):
Theories thus become instruments, not answers to enigmas, in which we can rest. We don’t lie back upon them, we move forward, and, on occasion, make nature over again by their aid. Pragmatism unstiffens all our theories, limbers them up and sets each one at work.
It is following this statement that James cites Papini’s hotel metaphor. 12 In Papini’s metaphor, the projects pursued in the different hotel rooms cannot all be ‘true’ in any of the conventional senses of the term. Yet, James seems to approve of all the projects in the rooms seeking to set ‘each one at work’, given that they must pass through the corridor which stands for practice. What, then, is the status of pragmatism in all this? James does not consider pragmatism to be one of the theories, but a ‘method’, although he does speak of it, following Papini’s notion of a ‘teoria corridoio’ (Papini, 1961 [1905]: 405), as a ‘great corridor-theory’. 13
Method is a potentially misleading term and requires qualification. James uses it in a different sense from Rorty, who opposes the term if by it is meant a ‘scientific method’ to which is attached the belief that there is something such as Nature’s Own Language out there to be discovered, and that this ‘discovery will give us a “method”, and that following that method will enable us to penetrate beneath the appearances and see nature “in its own terms”’ (Rorty, 1982: 192). For James, as we have seen, nature can be made over again. Method, for James, is ‘an attitude of orientation … of looking away from first things, principles, “categories,” supposed necessities; and of looking towards last things, fruits, consequences, facts’ (James, 1987a [1907]: 510, italics in original). James does not mean to say that one should not think about ‘first things’ or ‘principles’ (see Putnam, 2010: 189). Remember that in the fifth room someone is thinking about no less than a system of idealistic metaphysics! The point is rather — and this is where the pragmatist theory of truth comes in — that one cannot assert the truth, say, of that idealistic metaphysics without leaving the hotel room and passing through the corridor to (social) practice, where alone its usefulness shows. For James, pragmatism is not a theory, as it ‘has no dogmas, and no doctrines save its method’ (James, 1987a [1907]: 510).
Here is where our tension is finally out in the open. The more one makes pragmatism into a theory, in the case of IR into a social or political theory (regardless of whether into one beside others or one more fundamental than others), the more it must itself be subjected to the test of practice. The truth or usefulness of a theory cannot be shown on purely conceptual terms. And this is, it might be argued, where James somehow failed his own pragmatist commitments, as he erected elaborate bodies of theories (about psychology, religious experience, etc.), while hammering down his message about practice as the criterion of truth. Obviously, James might be taken to advance theories of his own without further truth-claims, as if he himself were then sitting in one of those Papini hotel rooms (although his rhetoric sometimes tells another story 14 ). Yet, what about his own pragmatism? James is explicit in not attempting to make pragmatism into a full-blown theory. On his account, it has no predictive power whatsoever, and ex ante ‘stands for no particular results’ (James, 1987a [1907]: 510).
It seems as if James’s pragmatist method indeed might open up the possibility of viewing several rival theories being of use alongside each other (in a sense which we will still have to establish). Surely, he is reflective of the consequences of his own theorizing about theories. 15 Applying a Jamesian view of theories as tools to IR would perhaps mean to take neo-realist, liberal and constructivist theories as equally well including the actual order in their respective scheme, while each perhaps is best put to use for certain purposes and under certain circumstances, which is the only way to establish anything like their truth. Theorists of IR may of course object that we are blatantly inconsistent and blind to the requirement of our own ‘method’, for surely what there is in positive doctrine about the pragmatist method, even if it is less than a theory (James speaks of words and theories), must itself be put to the test of practice. For what difference does it make whether one understands IR theories as handy tools or as primacy-seeking? Is this issue not a prime example of the sort of metaphysical quagmire of no use according to pragmatists?
James’s pragmatism indeed is if not a full-blown theory at least a meta-theory which advocates viewing theories as tools. The pragmatist method is not endorsed by James wholly without any relation to doctrine; what he claims is merely that pragmatism has ‘no prejudices whatever, no obstructive dogmas, no rigid canons of what shall count as proof’ (James, 1987a [1907]: 522). Yet, as James himself relates, pragmatism is ‘first, a method; and second, a genetic theory of what is meant by truth’. Arguably, James’s theory of truth is therefore to this extent part and parcel of the pragmatist method, as the latter is based on the commitment to inquire only into problems the solution of which would make a ‘practical difference’ (or have a ‘cash-value in experiential terms’) — itself the criterion of truth if the difference is considered a useful one (James, 1987a [1907]: 573). Above, we have invoked something like James’s verifiability theory of truth when insisting on the possibility of viewing several rival theories being of use alongside each other. It is thus important to explain our understanding of that theory, not least because it has been the subject of much criticism and misunderstanding.
James on truth as usefulness
Drawing on Peirce, James claims (1987a [1907]: 512, italics in original):
Any idea upon which we can ride, so to speak; any idea that will carry us prosperously from any one part of our experience to any other part, linking things satisfactorily, working securely, simplifying, saving labor; is true for just so much, true in so far forth, true instrumentally.
For James, those beliefs are true which are corroborated by experience in the long run. This to him is obvious for instance when one expects at the end of a path a house and upon having walked down the path comes to see the house. This sort of verification is what has triggered the correspondence theory of truth, and there is no fault with it so far. But James is quick to note that in practice ‘indirect as well as direct verifications pass muster’; in fact, if most of the truths which we accept are of the indirect kind, ‘truth lives … for the most part on a credit system’ (James, 1987a [1907]: 576). As regards ‘mental objects’, when true they are called definitions or principles, and they pose no further problem for verification, as their relations are ‘perceptually obvious at a glance’; ‘great systems of logical and mathematical truth’ establish the terms ‘under which the sensible facts of experience eventually arrange themselves, so that our eternal truths hold good of realities also’ (James, 1987a [1907]: 578). Here, however, is where James parts company with those who defend a correspondence theory of truth, for ‘to copy a reality is … one very important way of agreeing with it, but it is far from being essential’. It is not that correspondence theories of truth are mistaken or useless per se, but they are too narrow in thinking that an idea is true by virtue of a property inherent in it. Rather, correspondence theories of truth work for certain, but not for all, purposes. 16
For James, there is a reality against which all theories measure in the sense that new experiences and new facts might challenge our established body of truths, that is, of ideas and practices that have hitherto worked (for a linguistically turned way of putting this, one might think of it in terms of crises and routines). In that event, one aims to maximize the conversation of the old beliefs while restoring consistency. Our theories, James writes, ‘must mediate between all previous truths and certain new experiences’ (1987a [1907]: 580). Future events might always prompt us to junk heretofore established truths, as experience ‘has ways of boiling over, and making us correct our present formulas’ (James, 1987a [1907]: 583, italics in original). This is why in James’s version of pragmatist fallibilism no theory in the absence of its being ‘realized in rebus’ (1987a [1907]: 581, italics in original) can be true or false. 17
There are no a priori concepts, categories or principles which are true in the sense that they can be maintained in the face of all future experience. Clarence Irving Lewis (1883–1964) has similarly argued that a priori principles, being mere stipulations, instruments of the mind, impose no definite restriction on experience. They could easily be changed in the event that new experience shows them to be ‘infelicitous’ instruments (see Lewis, 2006). It might be interesting to note that when James speaks of experience boiling over, his examples are theories such as ‘Ptolemaic astronomy, euclidian space, aristotelian logic, scholastic metaphysics’, each of which was expedient for centuries and each of which later thought of as false (James, 1987a [1907]: 584). This choice of examples underlines James’s opinion that ‘within the evolving system of our opinions, most of the “old order” remains standing’ (James, 1987a [1907]: 513), but it should serve us also as a reminder that our current expedient theories, philosophies and truths (e.g. quantum physics, homo oeconomicus, even the linguistic turn) might one day become thought of as of little use, that is, as false.
Relating this line of argument to the question of what difference it makes whether we view theories as tools or as primacy-seeking, it becomes clear that we cannot prove or argue conclusively that the pragmatist approach advocated in this part of our article indeed makes a practical difference. Though we are happy to turn our theorizing about theories against our own stance, we can only offer reasons why viewing theories as tools would be useful in the social sciences for some purposes, that is, useful for just so much, useful in so far forth. To prove it or to argue the case conclusively, one would have to be able to predict the outcome of future practice(s) in the judgement of relevant communities, which would undermine the very pragmatist commitment towards fallibilism. If pressed as to what precisely usefulness means, pragmatists, as Rorty points out (1999 [1994]: 27f), are limited to fuzzy and at times unhelpful answers because they do not believe that the future conforms to a plan, but rather that it is open and potentially astonishing.
Hilary Putnam, who stands between Rorty’s neo-pragmatism and Quine’s scientism, comments along similar lines on James’s notion of usefulness. Defending James, Putnam writes (1995: 24, fn. 7):
Some critics even read James — against repeated statements to the contrary, explicit and implicit, in his writing — as holding that if the consequences of believing that p are good for you, then p is ‘true for you’. Let me say once and for all that James never used the notion of ‘true for me’ or ‘true for you’. Truth, he insists, is a notion which presupposes a community, and, like Peirce, he held that the widest possible community, the community of all persons (and possibly even all sentient beings) in the long run, is the relevant one.
Putnam furthermore takes away from the impression that James’s emphasis on usefulness is too absolute by heeding the qualifications James usually attaches to statements of his such as that the true ‘is only the expedient in our way of thinking’ (James, 1987a [1907]: 583). According to Putnam (1995: 9f), James held that there are different types of statements with different types of expediency, such as usefulness for prediction, conservation of past doctrine, simplicity or coherence. Simplicity and coherence are here not appraised as virtues per se or as copying the world in any sense, but as types of expediency, that is, as simply very useful ideas in the eyes of relevant communities.
This all comes down to efforts at persuasion by giving good reasons; but it is not that persuasion is only successful if it is accompanied by good reasons. Rather, one is in many instances simply more persuasive if one gives good reasons, particularly if those whom one addresses at least approximate Peirce’s ideal community of scientists or critical inquirers. Surely, and for good reasons itself, the practice of giving and asking for reasons is well established in the social sciences, as giving evidence or proof is in other sciences. The usefulness of a theory may be argued for by drawing on any of the following types of expediency as good reasons: policy relevance, simplicity (if that e.g. means an easy application of the theory), complexity (if that e.g. means comprehensiveness), context-sensitivity, context-independence, high degree of corroboration through past and present cases, potential to inspire further theories or new vocabularies and so on.
The upshot of the above is simply that there seems no reason why viewing theories as primacy-seeking only should be the established and still largely dominant practice in the social sciences in general and IR in particular. A Jamesian view of theories allows continuing the practice of specialized research that is firmly committed to, say, a constructivist theory of IR and that seeks to establish its primacy over against all other theories on offer. But, at the same time, it loosens these theories up by introducing a theoretical pluralism that sets the test of practice against all theoretical elevatedness and radically leaves it to the scientific community to discuss over and again the types of expediency accepted as indicators for the usefulness of a theory.
Conclusions
We conclude by bringing forth some more general reflections on why viewing theories as tools and an additional emphasis on practice in terms of the epistemological status of theories would pay off in the study of international relations. Were we to follow Rorty, who has sought to dismiss once and for all the traditional project of epistemology because he saw no use in it, we would simply have to propose giving up the view of theories as primacy-seeking. 18 There are two reasons, however, why we think it inadequate to do so. The first reason is that it is not up to us to decide whether something is useful or not, but up to the relevant communities. This is where Rorty at times seems presumptuous as if it were up to him to decide over the usefulness of an issue, all by himself and on paper only. James was well aware of this problematic, as is manifest in his treatment of the truth of the hypothesis of God having been pragmatically so successful (James, 1987a [1907]: 522, 618). By analogy, how could we possibly deny the usefulness of the idea of primacy-seeking theories, given that it seems pragmatically so successful? At least, many in many relevant communities think it is. The question therefore rather is for what purpose and under what circumstance is that view useful? Hence, the second reason is that we, too, think that the competitive view of theories as primacy-seeking is useful, if and only if understood with that important, but still little acknowledged, Jamesian qualifier: just for so much and in so far forth.
We ourselves have assumed such a view in the first part of our article, delineating two strong points of a pragmatist theory of IR, as if we were in a Papini hotel room, trying to outdo all other theories by means of specific notions of abduction and practice. It is an open question whether one could actually hold or defend views if one were not to think of them as ‘true’, justified or, at least, useful. One straightforward use of seeking primacy, for instance, is to make oneself heard, especially if what one has to say is judged by many others to be a marginal rather than a central concern. We may indeed say that our pragmatist theory is ‘truer’ (or ‘better’) than all others. But this, admittedly, is in a non-deprecatory sense rhetoric and, more strictly considered, always said with a certain purpose in mind other than the ‘truth’ of the theory. What we might want to say is more like our theory being better than others in that regard, for that purpose, under that circumstance and so on.
As to IR theories we, consequently, submit that — contrary to the belligerent rhetoric in the discipline — neo-realists or constructivists are not wrong per se. Their theories have uses, though maybe not under all circumstances and for all purposes. All we claim is that the same holds for pragmatist theories of world politics. Viewing different theories as different tools and instruments for dealing with the social world would thus make possible different research agendas and designs, perhaps not so much concerned with showing that a given theory is wrong per se, but with arguments that it is not of much use when applied for the purpose x to a problematic situation y, where this or that theory might be of more use. Obviously, all such argument — whether it pertains to abduction as research strategy, the sequentiality of practice, creative decision-making or anything else — remains subject to an unpredictable future practice and to judgements of that practice by relevant communities. Not verification or falsification, but arguments for prospects and limitations of a theory in relation to well-defined problematic situations would be the aim of such research. Not least, such research would serve as an important reminder that the tension which we have identified in pragmatism cannot and perhaps should not be resolved, lest theories be stiffened into paradigms or mere apologetics: our intellects shall not come to rest before any presumed finalities (including the ones we have relied on in this article).
Footnotes
Acknowledgements
An earlier version of this article was presented and discussed at the general conference of the European Consortium for Political Science (ECPR) in Potsdam in September 2009. We would like to thank all participants who, by their questions, comments and suggestions, have contributed to the present version. We also express our gratitude to the anonymous reviewers and to Ulrich Roos.