Critical Realism and Causal Analysis in International Relations
Milja Kurki
Introduction
Drawing on the anti-positivist philosophy of science of Roy Bhaskar, ‘critical realism’1 has sought to challenge some of the core assumptions theorists hold on the nature of explanation and science in IR theoretical inquiry.2One important area in which critical realists challenge disciplinary conventions in IR is the issue of causal analysis. Causation has been a contested notion in much of twentieth-century philosophy of science and social science and, since the late 1980s, has also been debated in International Relations, where the causal approach of the positivists has come under increasing criticism from a lection of post-positivist ‘constitutive’ theorists. Critical realism eks to reformulate currently dominant understandings of the role and nature of causal analysis in the social sciences and in IR. This short contribution to the forum focus on examining the critical realist intervention to the debates on causal analysis in IR. Critical realism, it is argued, opens up important new avenues in IR theorists’ and rearchers’ conceptions of causal analysis: avenues previously hidden from view by the dominance of a positivist view of science in IR.
____________
1. This paper focus on ‘critical realism’ associated with Bhaskar’s ‘critical naturalist’ philosophy of science. Roy Bhaskar, A Realist Theory of Science (Hassocks: Harvester Press, 1978); Roy Bhaskar, The Possibility of Naturalism: A Philosophical Critique of the Contemporary Human Sciences (Atlantic Highlands, NJ: Humanities Press, 1979). While the term ‘scientific realism’ is ud by many other IR theorists, such as Alexander Wendt, the term ‘critical realism’ is preferred here becau of its clo association with Bhaskar’s thought, the focus adopted here. There are many different strands of philosophical realism and some strands of scientific realism do not adopt a strongly non-positivist stance as critical realists do. For example, while Wendt’s work has been ground-breaking in IR, it does not follow Bhaskar’s philosophy of science in certain important respects.
2. Heikki Patomäki and Colin Wight, ‘After Post-Positivism? The Promis of Critical Realism’, International Studies Quarterly 44, no. 2 (2002): 213–37; Heikki Patomäki, After International Relations: Critical Realism and the (Re)Construction of World Politics (London: Routledge, 2002); Colin Wight, Agents, Structures and International Relations: Politics as Ontology(Cambridge: Cambridge University Press, 2006). Some critical realist themes are also developed in American ‘sci
entific realist’ scholarship. See, for example, David Dessler, ‘Beyond Correlations: Towards a Causal Theory of War’, International Studies Quarterly 35, no. 3 (1991): 337–55; Alexander Wendt, Social Theory of International Politics(Cambridge: Cambridge University Press, 1999).
I will proceed by, first, analysing some of the key tenets of the critical realist approach to causation and its contributions to our understanding of causation in IR. A critical realist account of causation suggests that the positivist model of causal analysis is not the only way to engage with the complexity of causal relations in world politics. On the other hand, the critical realist model provides a framework for ‘constitutive’ IR theorists to re-engage with causal analysis. For it is argued that, despite their misgivings on causation, ‘constitutive theorists’ associated with various ‘post-positivist’ approaches in IR already implicitly do causation in ways that provide IR with rich knowledge of caus and conquences in world politics.
The latter part of the paper will engage with some of the criticisms that the critical realist understandings of causation may encounter from both the positivists and the post-positivists in IR. I will show that, while the criticisms are interesting, a defence of the critical realist position is possible and, in fact, helpful in clarifying certain misunderstandings held in the discipline about critical realism.3 Indeed, the issue of causal analysis demonstrates well that important aspects of the
critical realist critique have been misunderstood in IR and that, as a result, important lines of convergence between critical realism and other IR theoretical approaches have been left inadequately explored.
Causal Analysis, International Relations Theory, and the Critical Realist Intervention
Causation has been one of the most contested concepts in philosophy of science and social science. In philosophical circles key debates on causation have centred around: (1) the meaning of the concept; (2) the reality of caus; and (3) the methods of causal analysis. Various positions have been taken on the issues during the last two thousand years and a full account of the positions is not possible here.4 For our purpos it suffices to state that during the last three hundred years causation has been understood predominantly in accordance with an empiricist philosophy of causation. This approach, to summari it crudely, studies causality through obrved patterns of facts: causality is something we identify through studying general patterns of obrved events, whether it is the tendencies of heavy objects to fall to the ground or the tendency of democracies not to fight other democracies. For empiricists we have grounds for speaking of causal connections or casual laws only when ____________
3. Some influential critiques include: Friedrich Kratochwil, ‘Constructing
a New Orthodoxy? Wendt’s ‘Social Theory of International Politics’ and the Constructivist Challenge’, Millennium: Journal of International Studies, 29, no.1 (2000): 73-101; Fred Chernoff, ‘Scientific Realism as a Meta-theory of International Politics’, International Studies Quarterly, 46 n.2 (2002): 189–207; Maja Zehfuss, Constructivism in International Relations: The Politics of Reality(Cambridge: Cambridge University Press, 2002).
4. However, for an interesting history of causation e Wallace, Causality and Scientific Explanation I & II (University of Michigan: Ann Arbor, 1972).
strong empirical regularities have been identified.5 For many empiricists, regularities can also be ud to predict (at least probabilistically): for example, given an empirical ‘causal law’ of gravity has been identified, we can predict that when a rock is dropped from a height it will fall. As a conquence of the dominance of empiricism, the bedrock of positivist philosophy of science6, causal analysis has entailed the prioritisation of methods focud on systematically obrving patterns of facts or ‘laws’.
The empiricist assumptions on causation have also held sway in the minds of many IR scholars throughout the twentieth century. I cannot here investigate in detail the different forms that empiricist
assumptions have taken in IR. I will merely state that they em to have informed, in various permutations, both the classical behaviouralist approaches in IR and many ‘post-behaviourist’ methodological prescriptions for social scientific IR, for example, tho of Michael Nicholson and Gary King, Robert O. Keohane and Sidney Verba (which have influenced the methods of study of many contemporary liberal, realist and even constructivist theorists in IR).7 While IR theorists from different schools of thought have historically disagreed in their substantive analysis of caus of war and peace, for most of the history of the discipline the idea of causal analysis, conceived of in a positivist fashion, has been largely unquestioned.8 This was until the 1980s and 1990s, that is, when many reflectivists in the discipline started to reject causal notions altogether and came to favour explicitly a ‘non-causal’ form of theorising often referred to as ‘constitutive theory’.9 With the solidification of the terms of ____________
5. Although some positivists would even question the validity of causal language and e relations of patterns simply in terms of correlations. Indeed, the very status of causal language is in question for some logical positivists. See, for example, R. Carnap, Philosophical Foundations of Physics: An Introduction to the Philosophy of Science (London: Basic Books, 1966), 204.
6. Positivism is of cour a contested term. Here I will take it to refer to a philosophy of science that i
s informed by empiricist theory of knowledge and that emphasis systematic gathering of empirical facts, empirical testing of hypothesis, value of instrumental knowledge and fact–value distinction.
7. John A. Vasquez and Marie T. Henehan (eds), The Scientific Study of Peace and War: A Text Reader , xx–xxii; D.J. Singer, Human Behaviour and International Politics; Contributions from the Social-Psychological Sciences (Chicago: Rand McNally, 1965); Michael Nicholson, Caus and Conquences in International Relations: A Conceptual Study (London: Pinter, 1996); Gary King, Robert O. Keohane, and Sidney Verba, Designing Social Inquiry; Scientific Inference in Qualitative Rearch(Princeton: Princeton University Press, 1994).
8. Although some classical scholars in IR were sceptical of the scientific approach to causal analysis of the positivists. See Hans Morgenthau, Scientific Man and Power Politics (London: Latimer Press, 1947); Hedley Bull, ‘International Theory: Ca for Classical Approach’ in Contending Approaches to International Politics, ed. K. Knorr and J. Ronau (Princeton, NJ: Princeton University Press, 1969).
9. See, for example, K.M. Fierke, Diplomatic Interventions: Conflict and Change in a Globalizing World(Basingstoke: Palgrave Macmillan, 2005), 1–19; David Campbell. Writing Security: United States Foreign Policy and Politics of Identity, 2nd edn (Manchester: Manchester University Press, 1998), 4.
the ‘fourth debate’ in IR, which pitted the ‘positivists’ against the ‘post-positivists’10, it has become widely accepted that there are two options in IR theorising: causal theorising that studies general patterns11 and non-causal theorising that studies contingent role of rules or discours.
Critical realism points out that many treatments of causal analysis in modern philosophy of science, and in contemporary IR, are riously problematic in that they have been informed by an uncritical acceptance of a rather narrow positivist conception of cau.
Critical Realist Intervention in Debates on Causation
Detailed examinations of the critical realist philosophy of science can be found elwhere12 but here the core assumptions of the critical realist philosophy of causation are assumed to be the following:
1. Caus exist as (ontologically) real forces in the world around us
and caus are ubiquitous (‘nothing comes from nothing’).
2. Many caus are unobrvable and the empiricist obrvation-
bad approach to causal analysis is problematic.
3. Caus do not work in ‘when A, then B’ manner and always
exist in complex causal contexts where multiple caus interact and
counteract with each other.
4. Social caus are of many kinds: from reasons and norms to
discours and social structures. Interpretation is central to causal
analysis in social science.
What makes critical realism a distinctive approach to the philosophy of science is its particular ontological approach. Critical realist philosophy of science starts with the assumption that reality exists independently of human obrvers. Critical realism es this as the fundamental justification of the practice of science.13 It follows that the causal forces that sciences
____________
10. While some call this debate ‘third debate’, I follow Ole Wæver’s characterisation of the contestati
on between the positivist and post-positivists as a ‘fourth debate’. Ole Wæver, ‘The Ri and Fall of the Interparadigm Debate’, in International Theory; Positivism and Beyond,ed. Steve Smith, Ken Booth and Marysia Zalewski (Cambridge: Cambridge University Press, 1996).
11. Even though within the positivist tradition some, on the basis of empiricist logic, also reject the very notion of causality in favour of speaking of correlations of obrved facts. See footnote 5 above.
12. See, for example, Andrew Collier, Critical Realism: An Introduction to Roy Bhaskar’s Philosophy(London: Verso, 1994); Margaret Archer, Roy Bhaskar, Andrew Collier, Tony Lawson and Alan Norrie (eds), Critical Realism: Esntial Readings (London: Routledge, 1998); Bhaskar, A Realist Theory of Science; Bhaskar, The Possibility of Naturalism; Patomäki and Wight, ‘After Post-Positivism?’. 13. See, for example, Christopher Norris, On the Limits of Antirealism (Manchester: Manchester University Press, 1997). See also Colin Wight’s article in this forum.
study, from gravity to social structures, must also exist as real ontological forces (outside our obrvations): it is only by accepting that causal forces really exist ‘out there’ that we can make intelligible scientists’ efforts to explain why and how process around us work as they do and, indeed, our everyday efforts to negotiate the world around us where things em to give ri to other
things.14 Importantly, critical realists also believe in the ubiquity of causal forces. Critical realists, following Aristotle, accept that ‘nothing comes from nothing’, that is, they believe that all events, process, objects and agents ari out of some conditions and influences, if you like from a pre-existing causal context.
While advocating a causal understanding of the world, the critical realists reject the empiricist logic for identifying caus. Critical realists challenge the assumption that causal analysis is dependent on empirical obrvation of regular patterns of facts. For critical realists caus are often unobrvable and hence causal analysis cannot be dependent on perception alone: ‘to be [and to cau] is not to be perceived [but] to be able to do’.15 For critical realists caus are often unobrvable. What causal analysis eks to do is, not to describe obrved patterns, but to give an account of the underlying causal powers that explain why the patterns of facts we obrve exist. Critical realism argues that in engaging in causal explanation, scientists, both in the natural and social sciences, need to engage in ‘deep ontological’ inquiry,16 which involves conceptualisation of the nature of the unobrvable structures that lie beneath obrvable patterns. In conceptualising ontological reality, scientists always draw on the conceptual and metaphorical tools available to them and, in this n, science is always social process. Epistemologically, critical realists emphasi pluralism and ‘
opportunism’: the nature of the ontological object has an important role in defining which ways of knowing are appropriate to it.
The empiricists assume that, given they have adequate evidence of patterns of facts, we can assume, at least probabilistically, that ‘when A, then B’.17Critical realism rejects mechanistic18and predictive understandings of causal relations. It emphasis that caus always exist in open systems where multiple causal forces interact and counteract in complex ways and where individual caus cannot be isolated as in a laboratory experiment. Critical realists prefer ontologically holistic ____________
14. Bhaskar, Realist Theory of Science, 21.
15. Bhaskar, Possibility of Naturalism, 16.
16. Deep ontology as a term highlights the difference between empiricists’ obrvable ontology where obrvable events are what form the basis of ‘existence’. Deep ontology emphasis that causal forces exist on deeper levels of reality and hence can be, while unobrvable, nevertheless real.
17. An assumption Bhaskar has termed ‘regularity-determinism’. Bhaskar, Realist Theory of Science, 70–1.
18. While some critical realists u the notion of causal mechanisms, the are not understood in a classically mechanistic manner. Many critical realists, including this author, prefer to avoid the u of the metaphor mechanism precily becau of its mechanistic connotations.
accounts that explain how and why complex causal factors come together. They also prefer non-deterministic metaphors in their causal accounts: caus for them are tho things that produce, generate, create, constrain, enable, influence or condition.19 Recently, critical realists have sought to develop a causal terminology that recognis the differentiated and pluralistic nature of caus. Different kinds of caus (e.g. material, agential or structural) are recognid to cau effects in very different ways: while some caus may ‘push and pull’, others ‘constrain and enable’.20
Recognising the wider meanings of the concept of cau has been especially important in the social sciences where ontological objects and caus are complex and dynamic. Critical realism recognis that ontologically social caus are very different from the kinds of causal powers that natural sciences study. In the social sciences, causal factors include a variety of ontological forces:
material resources, social structures, social rules and norms, discours and, controversially for the interpretivists, also ‘reasons’ that agents have for their actions. Engaging with the sorts of causal factors – and the causal complexes21 that they form – entails non-empiricist epistemological tools: it entails interpretation and recognition of the ‘double hermeneutic’ relations between the inquirer and their objects of study. Following the empiricist method of studying patterns of behaviour we can identify some descriptive patterns of behaviour (for example that being working-class tends to be associated with voting socialist). However, we can only causally explain when we engage with the reasons actors give for their actions and the possible social structural caus that condition them to act in certain kinds of ways.22
Through its critique of positivism, critical realism forces us to recogni that positivist causal explanations in IR, while not necessarily without their us and insights, can be characterid by certain ontological and methodological limitations. In their focus on relations of independent and dependent variables positivist models can lack holistic ontological (conceptual) engagement with complex causal environments, even ____________
19. For a summary of various causal words e Andrew Sayer, Method in Social Science: A Realist Approach, 2nd edn (London: Routledge, 1992), 104.
20. See also Heikki Patomäki, ‘How to Tell Better Stories about World Politics’, European Journal of International Relations,2, no.1 (1996): 105–33; Paul Lewis, ‘Realism, Causality and the Problem of Social Structure’, Journal for the Theory of Social Behaviour, 30, no. 3 (2000): 249–68; Paul Lewis, ‘Agency, Structure and Causality in Political Science: A Comment on Sibeon’, Politics, 22, no. 1 (2002): 17–23; Milja Kurki, ‘Caus of a Divided Discipline: Rethinking the Concept of Cau in International Relations Theory’, Review of International Studies, 32, no. 2, (2006): 189–216.
21. This concept is developed by Heikki Patomäki; Patomäki, ‘How to tell Better Stories’, 105–33.
22. Qualitative rearch is en as ‘intensive’ causal rearch, while quantitative is en as an ‘extensive’ taxonomical model. See Sayer, Method in Social Science, 243.
