ABSTRACT

Introduction Traditionally, IR theorizing has been, and in all likelihood will continue to be, a fairly divisive line of study. As Holsti (1985) has noted, IR is not only a divided but essentially a dividing discipline.1 Often the study of IR is seen as taking place within the confines of certain paradigms – a perspective that was perhaps too hastily borrowed from philosopher of science Thomas Kuhn in the 1970s (Kuhn 1962/1996; for the applicability of the concept in the context of IR, see the discussions in Hakovirta 2008; Smith 1995: 14-16; and Guzzini 1998: 119-129). In essence, the paradigmatic approach has pitted the different schools of thought in the study of IR against each other (see Schmidt 2002). This is often done in a rather crude manner whereby hastily construed ‘straw men’ are juxtaposed against each other with the aim of showing the superiority of the approach chosen and preferred by the author. Nor is this approach confined to mainstream IR theorizing alone, as the more specialized sub-fields have portrayed the same characteristics as well. For example, Puchala (1999: 318) has argued how theorizing on the European Union ‘has recently evolved into a full-scale, hard-fought debate . . . with contenders jumping upon one another’s attributed weaknesses while disregarding one another’s insights’ (see also Ojanen 2007: 55). But IR is not alone in this tendency: political science and indeed the social sciences in general as well as the humanities have suffered from the same syndrome (see Almond 1990; Chalmers 1990; Jokisipilä 2007: 34; King et al. 1994: 4; and Ryan 1972: 89). In sum, it seems more a case of the basic human condition than some peculiar pathology specific to the study of IR that we are dealing with. This state of affairs largely stems from the post-positivist notion that all observation is essentially theory-laden: we do not have access to reality without the help of the concepts we employ to approach it (for classic treatments of the issue, see Hanson 1958: ch. 1; and Quine 1960: chs 1 and 2). Accepting this raises difficult questions concerning the role of research vis-à-vis notions of truth (even with a small t) and the objective nature of the reality we are seeking to grasp and explain with our theories. As Imre Lakatos (1972) has argued, we can never test theories directly against the world, but only against other theories in light of the world (for a discussion, see Wendt 2004: 260; and Motyl 1990:

20-21). As a consequence, we have to engage ourselves in a ‘three-cornered fight’, where theories duel with each other and with the world in light of the empirical evidence we can muster (Lakatos 1972: 115; see also Laudan 1977; and Van Evera (1997: ch. 1). These inherent difficulties related to conducting research in the first instance have often been seen as deleteriously affecting our chances of making informed choices and comparisons between rival theories. Here the field of philosophy of science is, however, strongly divided, with one school of thought arguing against the possibility of comparing different theories, or at least ranking them in any meaningful way in relation to a notion of truth (Quine 1960; Feyerabend 1975, 1993; Kuhn 1962/1996, 2000), and another school of thought, which through a notion of truthlikeness, or verisimilitude, sees a fruitful way of comparing theories and making informed choices between them (Popper 1959/1972, 1963/1972; Niiniluoto 1999). A middle ground between the two polar oppositions is a stance which, while accepting epistemological relativism, nevertheless argues that the referent object of theories (‘the world’) is ontologically real, therefore giving some grounds to meaningful comparison between rival theories (Lakatos 1972; Laudan 1977; Bhaskar 1986, 1998; see also the useful and level-headed discussion in Wight 1996). This is also the path followed in this study. Although perhaps understandable from a human and social point of view, the paradigmatic approach in IR is nevertheless to be regretted as it contributes – in a rather artificial way, one might add – to the already wide dispersion and disintegration of different ideas and strands of theorizing within the discipline. It has also largely prevented scholars from building theoretical models of multicausal mechanisms in International Relations: instead of looking for complementarity, different theories with different causal mechanisms have been repeatedly pitted against each other (Buzan and Little 2001b). The result has been the present self-image of the discipline based on a succession of ‘great debates’ between the different schools of thought (Smith 1995: 14; Fearon and Wendt 2002: 67). But as Steve Smith (1995: 17) has noted, often the debates were ‘really only a series of statements of faith, with political or sociological factors determining which voice was heard’ (for a similar critique, see Halliday 1994: ch. 2; Schmidt 2002: 12-16). This was the case in the debate between positivists and post-positivists in the 1980s and early 1990s and more recently in the debate between rationalist and constructivist (meta-)theoretical stances on the study of IR as well. This has had the unfortunate result of shifting the emphasis away from empirical works and practical problems to crippling ontological and epistemological battles within the discipline. As a consequence, what was previously a study of concrete issues – questions of war and peace, as Holsti (1985) has put it – has to a degree become an exercise in bad, or at least shallow, philosophy of (social) science (Wight 2002: 26; Monteiro and Ruby 2009). None of this is meant to deny the basic relevance of philosophy of science for the study of IR, however; without theories and their conscious application we would not be able to understand what we see. This criticism should not be taken to mean that the proliferation of different strands of theorizing is an altogether

negative phenomenon either. In fact, the plurality in theorizing can be seen as reflecting the multifaceted nature of international relations and, as such, can also be seen as a source of strength for the discipline as well (Smith 2007). Nor should the words above be taken as a denial that the different ontological and epistemological stances of different schools within IR do have real bearing and that at least some of the differences are genuine. This cannot be refuted. But this current dispersion should make way, and in fact calls for intra-disciplinary (or inter-paradigmatic) investigations – within certain limits. The Quinean/Kuhnian/ Lakatosian idea of the ultimately theory-laden nature of all observation need not necessarily result only in three-cornered fights where different strands of theorizing beat each other to a pulp. Instead, this idea should also make way for a search for commonalities between different strands of theorizing. Hence, no theoretical approach has a monopoly on the truth. Here the pluralism endemic to IR actually comes to our rescue: it is also a vast resource from which to draw different strands for our own thinking in our search for a common ground, points of complementarity and even potential convergence between the different strands of IR theorizing in order to arrive at richer, multi-causal accounts of the often unobservable social mechanisms residing behind the events at stake. Following Schelling (1998: 32-33), a social mechanism can be defined as ‘a plausible hypothesis, or set of plausible hypotheses, that could be the explanation of some social phenomenon, the explanation being in terms of interactions between individuals, or between individuals and some social aggregate’ (see also Elster 1999: 1). According to Hedström and Swedberg (1998), identifying social mechanisms serves as claiming the middle ground between identifying scientific laws and providing mere descriptions; they allow explanation but no prediction. Although the version of social mechanisms advocated by them is clearly an instrumentalist one (see the discussion especially on pages 13-15), their basic approach can also be applied in the scientific realist manner by placing the causal mechanisms at the ontological level (George and Bennett 2005: 135-136). In recent years, scientific, or critical, realism has gained momentum also in IR (see Kurki 2006; Patomäki 2002; Patomäki and Wight 2000; and Wendt 1999 as well as the contributions to Forum on Scientific and Critical Realism in International Relations in Millennium, 35, 2007). To cut a long story short, while accepting the theory-laden nature of the scientific enterprise, scientific realists nevertheless separate the scientific concepts we employ – which Roy Bhaskar (1998) has called the transitive objects of science – from the real structures and mechanisms in the world – the intransitive objects of science – that are the object of our work. As such, the concepts we employ do not ‘construct reality’, as conventionalists such as Kuhn have argued, but instead offer us a potentially truthlike description of the essentially mind-independent reality (Niiniluoto 1999: 92; Outhwaite 1983: 323). Scientific realism, unlike the strict positivism that preceded it, accepts that reality can contain unobservable structures and causal links that have to be theorized and deduced from other, observable phenomena (Archer et al. 1998; Keat and Urry 1982; Niiniluoto 1999). In the words of Keat and Urry (1982, 30):

For the realist, adequate causal explanations require the discovery both of regular relations between phenomena, and of some kind of mechanism that links them. So, in explaining any particular phenomenon, we must not only make reference to those events which initiate the process of change: we must also give a description of that process itself. To do this, we need knowledge of the underlying mechanisms and structures that are present, and of the manner in which they generate or produce the phenomenon we are trying to explain. In describing these mechanisms and structures we will often, in effect, be characterizing the ‘nature’, ‘essence’, or ‘inner constitution’ of various types of entity.