Causation-in-the-world: A Contribution to Meta-theory of IR

Introduction

In the study of world politics, and of history and society more generally, it is common to ask whether a given outcome, be it an event, set of conditions or recurrent phenomenon, is shaped more by ‘ideational’ or ‘material’ factors. The question I want to examine here is different and more basic. It concerns whether such ‘shaping’, or ‘causing’, is itself ‘ideational’ or ‘material’.

Since it is often unclear how these two terms are distinguished, let me say first that my question is whether ‘causing’, as such, is ‘material’ (as opposed to ‘ideational’) in the sense of existing independently of human thought and language. This sense of the word ‘material’ is captured by Wendt’s defence of ‘a rump materialism’. ¹ This is an idea that, although much of the social world is socially constructed, certain facts, such as technological or geographical facts or facts of human nature, which shape the social world, are ‘brute facts’. Searle popularised this term when he contrasted such facts to ‘institutional’ ones: according to him, while the former requires a language to state them, the facts thereby stated are not themselves humanly constructed realities; the latter, by contrast, require ‘special human institutions for their very existence’. ²

My question is whether ‘causing’ is a brute fact, a feature of mind-independent reality, or the notion of causation-in-the-world is just that, no more than an idea – albeit a socially prevalent one – which we employ to make sense of our experience. This question has puzzled many philosophers and a number of philosophically minded contributors to International Relations (IR) as well. It is an interesting question because, among other things, if causation turns out to be a brute fact, those in IR who are critical of Wendt’s ‘rump materialism’, and other similar views, will at the least need to accommodate the ‘materiality’ of causal processes as such in their thinking; one cannot be a thorough going constructivist if, and at least to the extent that, the process of construction, which is a form of causation broadly conceived, is mind-independent. ³ My aim here, therefore, is to scrutinise the philosophical (realist/idealist) debate concerning causation-in-the-world and discuss the implications of the position I eventually arrive at to the study of world politics.

In what follows, I first explain the two positions, causal realism and causal idealism. I use Collingwood’s essay, published in 1938, to explicate the causal idealist position while noting that he is not himself a thorough going causal idealist. Collingwood famously combines conceptual analysis with historical learning – a commendable, though not often practised, method of philosophical investigation – which gives his article a very special place in the literature; and for my purpose his article is especially helpful as a resource with which to gain insight into what causal idealism may mean. I then rehearse a common interpretation of Hume’s philosophy of causation as an example of causal idealism while also acknowledging that some revisionist philosophers present Hume as in fact a causal realist. I then introduce causal realism, propounded, among others, by a group of scholars known as ‘critical realists’, as a prima facie plausible ontological claim.

This is followed by a second section where I identify and evaluate some standard ways in which critical realists have objected to causal idealism. Noting that causal realists would in turn need to ground their ontological claim, or to explicate its argumentational structure and basis, I introduce, in the third section, the philosophy of causation in the thought of Bhaskar, a founding figure of critical realism and influential advocate of causal realism. Bhaskar’s argument reveals that his causal realist ontology is grounded in his understanding of the practice of knowledge production and application in science – a point he, in any case, explicitly acknowledges.

Noting that the philosophical issues surrounding the causal realist/idealist debate are complex, I acknowledge, in the fourth section, that the advocates of causal realism, from the critical realist camp, are in fact aware of the complexities and make important qualifications, though relatively inconspicuously, to their often staunchly expressed realist stance. In the fifth section, I go on to discuss how we may fruitfully extricate ourselves from the apparent philosophical deadlock. Arguing that the divide between the two positions can be bridged, I conclude with a discussion on what all this means to the study of world politics.

I should state here that, in terms of the method of engagement and style of argumentation, much of what follows may be seen as an exercise in analytical philosophy and I should ask for the readers’ patience in this respect. Needless to add, I consider philosophical analysis to be an important, though not the only, way to contribute to a meta-theory of International Relations. ⁴ All that I aim to do here, however, is to think carefully about one of the most basic concepts in the study of world politics. Labelling it ‘Philosophy’ is somewhat counter productive as this makes it appear as though ‘IR’, being conventionally a separate discipline, could do without careful thinking – a position which, surely, no one would accept. The borderland of Philosophy and IR is an extremely slippery terrain and we can easily get trapped into one position or another which then produces an illusion that either one’s own position is correct or that we must settle for pluralism. But, if I am right in the moves I make below, there may not be sharply opposing positions in the first place in the particular subject area of this article.

Causal Realism and Causal Idealism

Whether ‘causing’ is a feature of ‘the world’ will depend on what is meant by ‘causing’ and ‘the world’. Collingwood’s etymological observation is pertinent to the question of what ‘causing’ means. According to him, in the historically original sense of the word ‘cause’, ‘that which is caused is the free and deliberate act of a conscious and responsible agent, and “causing” him to do it means affording him a motive for doing it’. ⁵ Collingwood adds: ‘For “causing,” we may substitute “making,” “inducing,” “persuading,” “urging,” “forcing,” “compelling,” according to differences in the kind of motive in question’. ⁶ In other words, ‘to cause’ meant ‘to motivate’ (someone to do something) in a variety of modalities; it still does.

But if ‘causing’ is equivalent only to ‘affording a motive’, as Collingwood suggests was the case initially, then it is an anthropomorphic metaphor to remark that an apple was ‘caused’ to fall, a ball was ‘caused’ to roll and so on. Indeed, he gives a brief historical account of how the commonplace ideas of ‘reluctance’ and ‘compulsion’ in interpersonal relations came to be incorporated anthropomorphically into Newton’s explanation of the movement of physical objects in terms of ‘inertia’ and ‘force’. ⁷ ‘Causing’, in such contexts, being a metaphorical expression, cannot, in Collingwood’s understanding, be what literally goes on in the world. If he is right, ‘causing’ is literal in relation to people being caused (or motivated) to do something but metaphorical and therefore imaginary in other contexts of what we normally call causation. Provocatively, he urges us to abandon metaphorical uses of the word ‘cause’ as they have, in his view, created much confusion in philosophical discussions about causation ‘from the time of Kant onwards’. ⁸ That subject need not detain us here, however; I want instead to consider briefly the issue of literalness.

Let us suppose (i) that we believe we have a good reason to claim that A afforded B a motive to do X and (ii) that we express our thought by stating, ‘A caused B to do X’. Observing this, Collingwood will note: ‘It is a literal use of the verb “to cause” to remark here that “A caused B to do X”’. There is a difference, however, between (1) that one sentence – ‘A caused B to do X’ – means literally the same as another – ‘A afforded B a motive to do X’ – and (2) that either sentence literally represents what happened because the former (the sentence) corresponds to the latter (what happened). Clearly, it does not follow from the fact that the two sentences in question are interchangeable that either of them corresponds to what, if anything, went on in the A/B relation. Of course, A/B may have confirmed to us that, actually, that was how it was. But that will in turn be A/B’s own explanatory construction; it is a causal statement, expressing their understanding of what happened within the vocabulary available to them which enables them to articulate their experience. ⁹

Importantly, Collingwood believes that causing someone to do something (or being caused by someone to do something) is a process of which we have a direct experience – so that, at least in such instances, what we say corresponds (if we get it right) literally to what happens or happened; we know ‘causing’ took place. ¹⁰ In such cases, ‘causing’ is not just an idea but also real, according to his way of thinking. Collingwood, therefore, is not a thorough going causal idealist.

But it may still be asked: does it really make sense to say that a causal statement literally corresponds to what happened in the world; are we not saying this because we have already represented what happened in the world in our heads as a case of causing someone to do something (or being caused by someone to do something)? If we have already articulated in our thought what happened, it is of course easy to see that there is a literal match between what happened and the explanatory statement we offer. This makes me suppose that causal idealism may have a point; causation may be an explanatory concept, a category of understanding, all the way down. In any case, it certainly does seem to have a point – the very point that Collingwood is making – that ‘causing’ is not what we can have a direct experience of where, as in an apple being made to fall or a ball roll, the effectuation is not experienced internally by human consciousness.

Hume went further. He equated experience with sensory observation and thought that we have no experience of causation even between motives and actions as we cannot feel any necessary connection between them. ¹¹ This has led him, according to a common interpretation, to argue for a thorough going causal idealism. Kim, for instance, states: ‘Hume’s celebrated critique of “necessary connexion” as an objective relation characterizing events themselves was perhaps the first – clearly the most influential – expression of a systematically articulated irrealist [i.e. idealist] position on causation’. ¹² This common, if contested, interpretation of Hume’s view runs as follows. ¹³

When we say that one event caused another, what we have in mind is that the first event brought about, or necessitated, the second event. Hume accepted this; according to him, as Mackie rightly notes, the idea of bringing about, or necessitating, was integral to our concept of causation. ¹⁴ Hume thought, however, that this idea has no correspondence in our observations; we have no sensory experience of causing or causal powers realising themselves. This created a puzzle for him since he also believed that our concepts were imprints of our experiences (or ‘impressions’). His task, therefore, was to give an account of how our idea of causal powers or necessity arose. Hume’s solution was to offer an ingenious causal story. According to Hume, we develop an idea of causal powers/necessity because, and to the extent that, we encounter a regular sequence of events, which psychologically conditions us into expecting a relevant effect-type event to follow when we see a cause-type event. Our idea of causal necessity is nothing but this conditioned expectation transposed to the world. For Hume, causal necessity is solely an idea and a causal relation is nothing but a regularly observable event-sequence. This is the gist of the so-called ‘Humean’ theory of causation, also known as a regularity theory. Hume’s position, presented in this way, is one case of ‘causal idealism’ inasmuch as, for him, causation, in the sense of the operation of causal powers or causal necessitation, is not a feature of the real (mind-independent) world.

There are, however, dissenting opinions on what Hume had in fact argued. ¹⁵ Indeed, if, as stated above, Hume’s concern was purely to explain the psychological process whereby we come to form the idea of causal necessity, his argument cannot establish whether or not causal necessity is (also) real. Of course, his empiricist epistemology will make him say that we have no experience and therefore no knowledge of causal necessity. But it does not follow from this that, in Hume’s own view, we know causal necessity to be only an idea. I will return to this point later in the article.

But here arises a troublesome question: whether Hume himself had argued for it or not, how is it even plausible to suggest that ‘causing’, or the operation of causal powers, is not a feature of the world? I ask this because I am inclined to think, as with most other people that I know, that ‘causing’, understood in this way, is a feature of the world. Are they (myself included) all living a philosophically unsound, illusory life? After all, if I bang my head hard against a stone wall, it will hurt regardless of what I think or say. So, how could all this square with the view of causal necessitation as only an idea?

In asserting that head-banging ‘causes’ pain, however, we are, I think, already assuming ‘causing’ to be a feature of the world. Wendt, among numerous others, claims that there are brute facts about the world regardless of our state of knowledge but this does not necessarily mean that something ‘causing’ something else is one such fact, as, for example, Wight seems to suggest. ¹⁶ The latter proposition, which cannot be shown, needs to be argued for rather than treated as if immediately obvious from our empirical observations themselves. ¹⁷

It is also often remarked in defence of causal realism that ‘[i]f humanity ceased to exist we have good grounds for assuming that sound would continue to travel and that heavy bodies would fall to the earth’. ¹⁸ However, since causal realism holds that causal powers are real despite their unobservability, the imagined absence of humanity to observe observable phenomena is irrelevant as a ground for accepting causal realism. The category of ‘unobservable but real’ and that of ‘observable but unobserved’ should not be conflated. ¹⁹

In my view, therefore, the two arguments just touched on do not allow causal realists to dismiss causal idealism without further ado; and, indeed, causal idealism – that ‘causing’ is not a feature of the world – is in fact quite plausible and I would say ‘correct’, if ‘the world’ is equated with the ‘empirical’ world. This is a simple point: since the operation of causal powers is not something we can experience or observe, ‘causing’, of course, is not an ingredient of the (empirical) world.

However, the idea that what we think of as ‘the world’ is exhausted by what we call ‘the empirical world’ can be countered by a broader notion of ‘the world’. Patomäki and Wight, following Bhaskar, thus speak of the world of which ‘the empirical world is only part’. ²⁰ Their view is that causal powers are an ingredient of this (wider-than-the-empirical) world. And it does appear that the idea of the world in which there are causal powers is regularly entertained in our way of making sense of how things work in nature and society (independently of our understandings). This will not mean that such an idea is integral to human thought in any timeless sense. ²¹ But the existence of such powers is widely taken for granted; and causal realism is defended by a number of philosophers, such as Harré and Madden and Bhaskar, and adopted by critical realists, some of them eminent contributors to the meta-theory of IR. ²²

Critical Realists’ Usual Objections to Causal Idealism

In social science and IR, causal idealism, conventionally associated with Hume, has been the dominant assumption. However, there is a determined opposition to this ‘Humean’ hegemony by critical realists. ²³ While I am sympathetic to their view that what we call ‘the empirical world’ is only one aspect of what we consider to be ‘the world’ and that causal powers are ingredients of this wider world, I am not entirely persuaded by some of the arguments advanced by critical realists against causal idealism. Three such arguments may be touched on here: (1) that causal idealism inevitably leads to the ‘anything goes’ stance; (2) that causal idealism lacks ontological grounding; and (3) that causal idealism is fallacious because it privileges epistemology over ontology. ²⁴

(1) The objection that causal idealism inevitably leads to the ‘anything goes’ stance may be resisted. Bhaskar is of course right to say that ‘it is only if the working scientist possesses the concept of an ontological realm, distinct from his current claims to knowledge of it, that he can philosophically think of a rational criticism of these claims’. ²⁵ But neopositivist social scientists, who follow the ‘Humean’ regularity theory of causation and subscribe to causal idealism, do possess the concept of an ontological realm, which, for them, is the empirical world. Within the neopositivist paradigm, empirically observed regularities do matter and circumscribe what can or cannot be treated as causal relations: hence, neopositivists’ stress on hypothesis-testing. ²⁶ Clearly, not ‘anything’ goes. Of course, we can debate about whether their (or any other) ontological thinking is adequate. The point being made here is not necessarily that the causal idealist ontology is adequately grounded but rather that the accusation that causal idealism inevitably leads to an ‘anything goes’ attitude requires a further articulation and defence. ²⁷

(2) In dealing with the second objection concerning ‘ontological grounding’, it is important to distinguish between an ‘ontological claim’ and its ‘grounding’. The former is an assertion, the latter argumentation, explicating the structure and basis of the assertion. Moreover, it should not go unnoticed that the causal idealist claim – that causal necessity is not a component of the world – is an ontological claim; it concerns what the world does or does not consist of. One other point should be noted: it does not follow from the fact that, according to causal idealism, causal necessity is not a component of the world that this (ontological) claim has no grounding (where the ‘grounding’ is rightly understood as argumentation explicating the structure and basis of the assertion).

The key issue is whether causal idealism has an adequate grounding or not, which is really about whether it is backed by a persuasive argument. It may or may not be possible to advance a persuasive argument for accepting this particular ontological claim. We may find that causal realism has a more convincing argument supporting it. Considering these issues is what constitutes ontology (understood as a branch of philosophical deliberation concerning ‘existence’). The fact that causal idealism claims causal necessity to be only an idea, and not a feature of the world – and here one might be lured into remarking ‘not grounded in the world’ – does not ipso facto deprive grounding (or argumentational backing) from that particular ontological claim. What we need is not ‘ontological grounding’ in the sense of grounding in an ontological assertion, but a particular ontological claim’s grounding. ²⁸

(3) Causal idealism is also said to be mistaken because it privileges epistemology over ontology. There is a sense in which causal idealism does this. It does this to the extent that it presents an epistemological claim – that we cannot have knowledge of causal powers because we cannot experience them – as a ground for making an ontological claim that such powers do not exist and that they are only an idea in our heads. This will be an instance of an epistemological claim being used as a basis for an ontological one. There are two problems that I see here: (1) the epistemological claim that we cannot have knowledge of causal powers because we cannot experience them may be opposed; and (2) the claim that what exists is reducible to what we can know to exist may too be opposed. Of these two oppositions, the former only challenges causal idealism for basing itself on an erroneous epistemology. It is the latter that challenges causal idealism for giving priority to epistemology over ontology.

However, whether causal realism triumphs over causal idealism on these two points will depend on the persuasiveness of the argument advanced. And whether the argument produced in defence of the realist ontological claim is sound or not is a question that needs to be scrutinised independently of the realist complaint that causal idealism privileges epistemology over ontology in the sense just explained. Even though causal realism may be free of this mistake, its ontological claim must in turn be backed by some further argument and such an argument (or the causal realist ontology’s grounding) may also turn out to be problematic – perhaps in a different way. In short, arguing that the grounding of the causal idealist ontology is defective does not necessarily show that the grounding of the causal realist ontology is adequate.

Bhaskar’s Grounding of the Causal Realist Ontology

The critical realist claim that causation, understood as the operation of causal powers, is a feature of the world stems from Bhaskar’s philosophy of science. ²⁹ Bhaskar formulates his idea of the world – his ontological claim – on the basis of his interpretation of the practice of scientific experiments. Experiments are, of course, what scientists conduct regularly to test their causal hypotheses. Bhaskar interprets the practice of experimentation as one where the scientist produces x to see if it, in turn, produces y in conformity with a hypothesis being tested in an environment which they try their best to keep free of interference by other causal factors. Such an interpretation is not Bhaskar’s idiosyncratic one but conforms to scientists’ own understanding of what they are doing when they engage in experiments. It is on the basis of such an understanding that scientists engage in what they do; and a key element of this understanding is the concept of a cause as that which has the power to produce its effect – so that, in order to let the cause (x) produce its effect (y), the scientists need to prevent extraneous causes from interfering with their experiments.

Bhaskar’s argument (which he calls ‘transcendental realism’) is that without an assumption that causal powers exist and are present in the wider-than-the-empirical world, what scientists do when they conduct experiments would not be intelligible as rational and that therefore it would not be possible for the scientific community to continue to engage in what they do. He twists his argument somewhat here and suggests that because science is possible, given that it occurs, the assumption that causal powers are present in the wider-than-the-empirical world must hold, leading him to the conclusion that while causal powers cannot be shown, they can be known, to exist. ³⁰

Bhaskar’s argument therefore involves the following three or four moves: (1) causal powers cannot be shown to exist; but (2) what cannot be shown to exist may yet be known to exist; (3) causal powers are known to exist in the sense that the assumption that they exist can be argued to be valid (given what scientists do); and (4) the statement that causal powers exist is true in the sense that they are known to exist. It seems to me either that Bhaskar is making the first three moves only or that he is adding the fourth. If he is only making the first three moves, his claim is an epistemological, not an ontological, one, pointing out, basically, that valid assumptions count as knowledge. But if he is adding the fourth move, he must be seen to be transposing a statement about being into a statement about our knowledge of being, which, according to his own definition, is a case of ‘epistemic fallacy’. ³¹ However, it does not strike me that step (4) actually involves any ‘fallacy’. It might be fallacious to suggest that X does not exist because it cannot be known to exist; but to assert that X exists on the grounds that we know it to exist does not strike me as a fallacy at all. In short, Bhaskar is either making an epistemological claim or transposing an ontological assertion into an epistemological one, though not fallaciously on that account.

However, there is a problem in Bhaskar’s representation of experiments, which Collingwood would alert us to. Are scientists right in their interpretation that not only do they produce x (or cause x to happen) in an environment they try to control, but that x in such an environment in turn produces y (or causes y to happen)? They may claim to have an immediate (i.e. unmediated or direct) experience of producing x. Bhaskar, for his part, considers it ‘clear that we are aware of ourselves as causal agents in a world of other causal agents; and that unless we were so aware we could not act intentionally, or come to know ourselves as causal agents at all’. ³² Thus far, his position is similar to Collingwood’s, which I outlined earlier. But Bhaskar, as Collingwood would warn, could not, without further explication, make a parallel claim regarding the x-y sequence. Bhaskar, in incorporating scientific practice as a ground on which to build his causal realist argument, seems to make an important assumption here: for him, it is not simply that, under experimental conditions scientists create, x is found ‘actually, in fact’ ³³ to be regularly followed by y but, crucially, that, in those conditions, x, which they produce, in turn produces or necessitates y. But how can Bhaskar claim to know that, in this x-y sequence, a causal power is in operation, such that, in the circumstances, x (‘actually, in fact’) brings about y ? He offers two lines of argument here.

Bhaskar, as a philosopher, dealing with concepts and assumptions, and not with processes of nature, cannot claim to know that this is what happens. But he argues that the ‘putative causal or explanatory “link”’ (showing how x brings about y) is given in scientists’ theory ‘containing a model or conception’ of that linkage. ³⁴ He refers to such a model or conception of the putative causal or explanatory link by the term ‘postulated mechanisms’ and adds that ‘[u]nder certain conditions some postulated mechanisms can come to be established as real’. ³⁵ Bhaskar does not say what these conditions are. According to my reading of him, however, ‘established as real’ must mean ‘accepted as real by the existing scientific community’; the ‘reality’ he speaks of is relative to a scientific community’s understandings of nature. ³⁶

The other line of argument is perhaps the more crucial. Bhaskar points out that causal laws identified under experimental conditions are (thought by scientists to be) also operative outside the experimental space; otherwise they would not resort to experiments to identify causal laws which they then use to try to explain phenomena in open systems. This means, argues Bhaskar, that causal laws are not on the same plane as de facto regularities observable under experimental conditions; rather, they should be interpreted as depicting the ways causal powers/mechanisms operate in closed, as well as open, systems, even though it is only in closed systems that scientists manage to make their operations manifest themselves as regular patterns. ³⁷ In other words, scientists’ causal laws are not regularities observed under experimental conditions (let alone mere ‘empirical generalisations’); rather, the laws state the ways of workings of nature which underlie and generate the regularities observed.

I find Bhaskar’s characterisation of science quite persuasive. ³⁸ However, it is important to note that his causal realist ontology is not a description of the world as it really is independently of our knowledge claims but is grounded in – and in fact replicates – (what he interprets as) the ontological assumption underlying scientific knowledge production and application. Bhaskar states that his central question is ‘what must the world be like for science to be possible?’ ³⁹ But he effectively translates this into: ‘What concept of the world can philosophical analysis reveal scientists to presuppose in their activities in order that scientific activities can be understood as rationally intelligible and therefore possible activities?’ And Bhaskar claims that there is nothing untoward, and everything right, about this move. He states:

Ontology, it should be stressed, does not have as its subject matter a world apart from that investigated by science. Rather, its subject matter just is that world, considered from the point of view of what can be established about it by philosophical argument. The idea of ontology as treating a mysterious underlying physical realm … has done much to discredit it; and to prevent metaphysics from becoming what it ought to be, viz. a conceptual science. ⁴⁰

Hence his claim, at the very beginning of the book, that his primary aim is ‘the development of a systematic realist account of science’. ⁴¹ He also states: ‘I do not claim in this book to solve any general problems of philosophy. It is my intention merely to give an adequate account of science’. ⁴²

The foregoing discussion alerts us to some considerable complexities in this terrain of philosophical argumentation and the need to be very careful in formulating our views in this area. It may be unsurprising therefore to find that some of the leading critical realists have made, or have felt the need to make, important, though perhaps often unnoticed, qualifications to their main assertions, as will be noted briefly in the next section.

Critical Realists’ Caveats

We may begin with Kurki. Having argued strongly against the ‘Humean’ regularity view of causal relations, she comes towards the end of her book to make an important qualification: ‘the question of causation is not a problem that can be solved: it is merely a problem that can be solved in various different ways’. ⁴³ It is interesting to encounter this seeming concession to pluralism at the end of the book, which argues strongly against the ‘Humean’ regularity theory of causation on philosophical grounds. But such a qualification suggests that perhaps an even more careful philosophical analysis is required. Another example of an easy-to-miss qualification is found in Wight’s book: ‘All substantive ontological claims require epistemological justification’. ⁴⁴ This almost looks like a hidden disclaimer, given his stress on the need to avoid privileging epistemology over ontology. But, if Wight is thinking here that a causal realist ontological claim must be backed by some argumentation, addressing, among other things, the issue of how, and in what sense of the word, we can claim to know that there really are causal powers, I entirely agree with his statement.

At this point, I should quote a passage, which appears in the last two pages of Bhaskar’s book:

There is no way in which we can look at the world and then at a sentence and ask whether they fit. There is just the expression (of the world) in speech (or thought). Science is an activity, a process in thought and nature which attempts to express in thought the natures and constitutions and way of acting of things that exist independently of thought.… We express [our understanding of] [sic] nature in thought.

Science, then, is the systematic attempt to express in thought the structures and ways of acting of things that exist and act independently of thought. ⁴⁵

A number of key ingredients are intertwined in this passage: ‘the world’, ‘the structures and ways of acting of things that exist and act independently of thought’, ‘our understanding of nature’, ‘(to) express [our understanding of] [sic] nature in thought’, ‘expression (of the world) in speech (or thought)’. My way of sorting out this cluttered passage is to consider Bhaskar to be saying that science is a systematic attempt to express in speech (or language) our understanding concerning the structures and ways of acting of things that exist independently of thought. I believe Bhaskar will not object to this paraphrasing; I agree with this characterisation of natural science; and I accept that the idea of causal powers is integral to science, which is a way of expressing our understanding of nature that exists independently of our thought.

Bhaskar’s position, however, may now begin to sound like causal idealism inasmuch as the idea of causal powers is treated as integral to a way of expressing our understanding of the world. The main difference I can see between this and Bhaskar’s causal realist position is that, according to the latter, causal powers are not just an idea but this idea accurately captures what happens in the world. But, of course, as Bhaskar rightly acknowledges in the above passage, there is no way in which we can look at the world and then at a sentence (‘causal powers are in the world’ would be one such sentence) and ask whether they fit. In fact, Bhaskar’s ‘world’, where he thinks ‘causal powers’ are located, not being the ‘empirical world’, is not somewhere we can ‘look’ at all.

And here is one final quotation that I have found most illuminating. This comes from Patomäki:

Realist ontological concepts such as ‘powers’, ‘producing’ and ‘mechanisms’ are themselves metaphorical, and should be assessed as such.… There is little that is strictly literal. The ‘literal’ meanings are either routinely reproduced metaphors or are based on immanent bodily and social experience, which ground metaphors. ⁴⁶

What I find interesting here is the fact that this is very similar to Collingwood’s position discussed earlier. It will be recalled that he is not a thorough going causal idealist insofar as he believes that causing (or motivating) someone to do something is what actually happens in the (empirical) world. He thinks that we have a direct experience of causing someone to do something, or being caused to do something, which we capture accurately by using precisely these expressions. And for Collingwood, in those cases of causing, where the effect-event is not an act of a free agent, the idea of ‘causing’ is metaphorical – as it is for Patomäki.

Bridging the Causal Realist/Idealist Divide

It will be useful at this point to consider where causal realists and idealists are in agreement with each other. They are agreed on, or can come to agree on, at least four points: (1) the idea of ‘causal powers’ is integral to our concept of causation; (2) we use this concept to make sense of our experience; (3) this concept is well embedded in our everyday thinking and action; and (4) it is also a necessary presupposition of the practice of scientific knowledge production and application.

However, the fourth point may be ambiguous. It may mean that an assumption of the reality of causal powers is a condition of the intelligibility of science as a rational activity (and hence a condition of the possibility of science). This, as I outlined earlier, is my interpretation of Bhaskar’s position. However, the fourth point may mean something more substantive, that is, the assumption that causal powers are a feature of the world is necessary (for scientists to make) precisely because they are a feature of the world – for, it may be asked rhetorically, why would it be necessary for scientists to assume that causal powers were real if they were not?

And this is where causal idealists begin to feel edgy. As I pointed out, Bhaskar acknowledges that there is no way in which we can look at the world and then at a sentence – ‘causal powers are in the world’ – and ask whether or not they fit. So far, so good, causal idealists will say. However, they may add: despite their welcome reassurance, causal realists may be thinking that it is precisely because causal powers are a feature of the world that it is necessary for scientists to assume that they are (even though their existence cannot be demonstrated directly). Causal idealists react to this somewhat nervously by thinking that, surely, we cannot know that causal powers are a component of the world. But there are a couple of arguments we can advance to assuage causal idealists’ anxiety.

First, even if we may be unable to know that causal powers are a component of the world, this does not mean that causal powers definitely do not exist. What is at issue here is not so much an epistemic fallacy as a logical error. It is not logically permissible to deduce from the premise that we cannot know that X exists a conclusion that we know that X does not exist (and that therefore X does not exist). Causal idealists, therefore, should refrain from thinking (if they are thinking this) that causal powers are definitely non-existent.

It is noteworthy in this connection that, according to Strawson, a leading revisionist interpreter of Hume’s philosophy of causation, Hume accepted the existence of causal powers as part of our natural beliefs and was not arguing that causal powers did not exist. ⁴⁷ The well-known ‘Humean’ regularity definition of causal relations was, according to Strawson, Hume’s attempt to state what he felt able to say within his empiricist epistemology. Hume wrote:

Our thoughts and enquiries are … every moment, employed about this [cause-and-effect] relation: Yet so imperfect are the ideas which we form concerning it, that it is impossible to give any just definition of cause, except what is drawn from something extraneous and foreign to it. Similar objects are always conjoined with similar. Of this we have experience. Suitably to this experience, therefore, we may define a cause to be an object, followed by another, and where all the objects similar to the first are followed by the objects similar to the second. ⁴⁸

Hume, in other words, knew that regular conjunction was ‘extraneous and foreign’ to a causal relation.

Second, causal idealists are happy to obtain causal realists’ acknowledgement that there is no way in which we can look at the world and look at the sentence ‘causal powers are in the world’ and see whether they fit. But neither party seems to notice an important point here: there is no way in which we can look at the world and look at any sentence (e.g. ‘a cat is on the mat’) and see whether or not they fit unless ‘looking at the world’ involves attention and articulation; we need to be attentive to the relevant aspects of the world and have articulated what it is that we are seeing when we look at it in order to be able to determine whether there is a fit between what the world presents to us and what the sentence states. Indeed, the world, even the empirical world, does not consist of ready-made sentences, looking at us and inviting us to confirm or disconfirm what we want to say about it. But if, because of this, we are afraid to make a statement about the world, we will not allow ourselves to say anything at all about it. Making a statement about the world means turning what is pre-discursive into discursive and we should allow ourselves to do this irrespective of whether we are dealing with the empirical or non-empirical features of the pre-discursive world – although, of course, a statement about the world needs to be confirmed, or disconfirmed, in different ways depending on whether it refers to the empirical or non-empirical features.

When causal realists make a statement about a non-empirical feature of the pre-discursive world and remark, for example, that causal powers exist, we should keep very clearly in our mind that they are not reporting on the match that they have found between this statement and what they see in the world as they look at it. Causal idealists should stop worrying therefore that this might be what causal realists are really saying (for it is not). Conversely, causal realists might try harder not to make themselves sound as if they were saying it. Instead of stating that causal powers cannot be shown but can be known to exist, they should perhaps say that causal powers can be understood to exist. I suggest this because ‘knowing’ is so closely intertwined with ‘seeing/showing’ in our everyday thinking that, when it is stated that X cannot be shown but known to exist, it is difficult to remove the notion that it is as if we were able to see X. In my view, the verb ‘to understand’ does not have quite the same connotation.

In this regard, some critical realists’ seeming failure to distinguish clearly between their category of the ‘unobservable but real’ (e.g. the existence of causal powers) and that of the ‘observable but unobserved’ (e.g. the existence of dinosaurs), touched on earlier, is worth recalling. This illicit conflation, however inadvertent, makes sense in relation to their strong claim that the former is ‘knowable’ – especially if this in turn is treated as if it were similar to ‘see-able’: ‘observable but unobserved’ is ‘see-able’ and therefore ‘knowable’ (if indeed ‘knowable’ is like ‘see-able’); ‘unobservable but real’, if it is not clearly distinguished from ‘observable but unobserved’, will then come to be treated as (or as if) ‘see-able’ and thereby also ‘knowable’.

Of course, these critical realists’ use of the ‘observable but unobserved’ as if it were similar to the ‘unobservable but real’ may not be a case of inadvertent conflation but intended heuristically. But it is better not to follow this heuristic path precisely because it makes them sound as if they are suggesting that the reality of causal powers is knowable as if it were visible. When we also hear them point out repeatedly that if someone were to shoot you in the head you would be lucky to be alive as if this undeniable fact, by itself, were sufficient to demonstrate the reality of causal powers, we may easily be misled into thinking that they are presenting the reality of causal powers as if it were a simple empirical truth.

However, to suggest that we should say we understand causal powers to be real is not at all to denigrate the status of our belief in their reality. We in fact do not (seem to) know of any other way of making sense of the world than to suppose, and understand, that causal powers are real. It is practically impossible to think or live without this understanding. This is an idea with which we make sense of important features of what we experience and we cannot translate this idea into a non-causal idea. ⁴⁹ I believe Kurki would agree with this view of the epistemological status of the causal realist ontological claim since, according to her, the position she endorses ‘argues that all scientific, and everyday, accounts of the world require the assumption of [causal] realism’. ⁵⁰ And, in my view, causal idealists can agree with causal realists on that point.

Furthermore, the critical realist interpretation that regularities are not an intrinsic feature of any causal relations but a phenomenon we understand as a manifestation of the operation of causal powers under appropriate conditions seems unanswerable; and, in my view, there is nothing in causal idealism as such that opposes this interpretation. This means that causal realism and causal idealism are not only not in contestation against each other, but can be united in their opposition to the regularity view of causation, which holds observable regularities to be an intrinsic feature of causation. ⁵¹

Conclusion: Implications for the Study of World Politics

Causal realism is a necessary presupposition of our everyday thinking and living. But if causal realism is such a natural belief, and everyone really believes in it, what difference does explicitly subscribing to it make to the study of world politics? There are two responses to this question.

One is to say ‘none’; we can just go on engaging in causal enquiries regarding anything that happens in world politics whose causes interest us. There is some truth in this. If we are inclined to engage in causal enquiries, we are already assuming that there is something that ‘brings/brought about’ the event or the situation that concerns us. That is what being curious about its causes means. So, reminding ourselves that causes are things that have the powers to ‘bring about’ their effects hardly adds anything to our enquiries.

However, critical realists tend to argue that this reminder is, in fact, highly significant. There is some truth in this, too. In particular, causal realism would remind us that when in search of a causal explanation of an event (y), we are not in search of an empirical generalisation that subsumes a particular sequence of events (x leading to y). Searching for empirical generalisations in open systems – and world politics is a primary example of an open system – is largely futile and even where one is found, it may not be a manifestation of one and the same causal process. From the viewpoint of causal enquiries in world politics (or in any study of an open field), a search for a ‘covering law’, in the sense of an empirical generalisation in the open system, is quite misguided; empirical generalisations simply aren’t causal laws. Here is a characteristically strong statement by Bhaskar:

practically all the theories of orthodox philosophy of science, and the methodological directives they secrete, presupposes closed systems. Because of this, they are totally inapplicable in the social sciences (which is not of course to say that the attempt cannot be made to apply them – to disastrous effect). Humean theories of causality and law, deductive-nomological [or ‘covering-law’] and statistical models of explanation, inductivist theories of scientific development and criteria of confirmation, Popperian theories of scientific rationality and criteria of falsification, together with the hermeneutical contrasts parasitic upon them, must all be totally discarded. Social science need only consider them as objects of substantive explanation. ⁵²

Thinking of causation in terms of causal powers scattered in the world, ‘exercised unrealised or realised unperceived (or undetected)’, ⁵³ rather than as having to do with discernible regular patterns, if any, in the empirical world has a significant impact, therefore, especially in a field, such as IR still dominated by the regularity view of causal relations. The regularity view is so prevalent, in fact, that even those professional historians who study the history of some aspects of world politics betray their acceptance of it when they claim not to be engaged in a specifically ‘causal’ enquiry (as they do not believe their explanations are based on covering laws or statistical generalisations). Conflating ‘theoretical’ with ‘generalising’, they also often say they are doing ‘history’, not ‘theory’.

Once freed of the regularity view, however, we begin to realise that there is something rather strange in the determination to look for regular patterns in international relations (where hardly any are found) and to try to explain them by a parsimonious causal theory (in a field where complexity appears to be the rule). Waltz’s contention that theories explain regular patterns begins to sound hollow and alien. ⁵⁴ He himself found no regular patterns to speak of, beyond what he treated as ‘recurrent’ patterns – though of the ‘rain would fall’ variety, for example, ‘war would recur’, ‘the balance of power would recur’. Where, exceptionally, this method is applied with apparent success, as in the case of the Democratic Peace Theory, its political impact can be pernicious, as Wight warns:

The attempt to construct a parsimonious theory of IR is not only flawed and doomed to failure, but also politically and ethically dangerous. It is dangerous … because such theories are apt to provide scientific legitimacy for particular forms of political practice. The promotion of western forms of democracy based on the scientific validity of a theory of democratic peace is but one example of this process. ⁵⁵

Leaving the regularity view behind does not, of course, tell us how we should or could construct a causal account of a significant singular event or condition, such as the outbreak of the First World War, the collapse of the Soviet Union, the current economic crisis or global warming, to name some obvious candidates in the field of IR. But at least we are reminded that such events likely require a combination of causal factors specific to particular cases. Thus Kurki remarks, quoting Collier, ‘a causal explanation is “a historical narrative in which a multiplicity of transitive verbs maps a complex causal sequence”’. ⁵⁶ She adds strikingly:

IR theorists should seek to understand the historical causal process in a holistic way, that is, concentrate on accounting for the complex interactions of various causes in specific historical contexts. ⁵⁷

And further:

Parsimonious framings of world politics [have to be] rejected as world politics [is] conceptualised as a complex web of interacting and counteracting causal powers and (structures of) social relations. ⁵⁸

Thus, the conventional disciplinary barrier between IR and historical studies of world politics, and more generally between theory and history, breaks down, as does the old, naïve – ‘second debate’ – distinction in IR between a ‘classical approach’ (openly embracing judgements) and ‘science’ (proclaiming to aspire towards scientific objectivity) as it is now freely admitted that ‘any explanation of a historical social process always involves a balance of judgement and that balances of judgement will remain contested’. ⁵⁹

Moreover, when we realise that causal powers are everywhere – exercised but not always realised – we begin to move beyond the study of ‘international relations’ narrowly conceived towards a study of ‘global politics’. Kurki again:

On the basis of the broad causal ontology accepted here, the ‘taken-for-granted’ nature of IR as a separate discipline must be questioned. Indeed, in order for IR to come up with explanatorily adequate causal accounts of complex global realities, it should open up to an analysis of social relations beyond the traditional scope of ‘International Relations’. ⁶⁰

In my judgement, such a critique of mainstream IR, focused on inter-state relations and involving a search for empirical generalisations and a parsimonious causal theory, in favour of a historical (or not ahistorical) study of global social relations, which pays attention to causal complexes, diversity of historical contexts and the contested nature of causal interpretations, is unanswerable. There is therefore a sense in which the anti-regularity stance is not oriented towards producing an ‘IR Theory’. Its interest is more in the direction of historical (or historically conscious) studies of various aspects of global social relations – although I should stress that an understanding of causal mechanisms plays an indispensable part in historical explanations. Beyond that, this stance does not, and should not be used to, sanction any particular substantive argument in the study of world politics. ⁶¹

And, finally, any approach to the study of world politics, which is explicitly claimed to be based on critical realism, or its central tenet, causal realism, must be clear about the nature of that philosophical doctrine as an assumption embedded in the practice of knowledge production and application, which includes the very approach being adopted. Like a rudder which is part of a boat, the causal realist ontology is part of the mode of enquiry which it steers, not a ‘philosophical foundation’ in the sense of something which philosophers prepare in advance of, and separately from, the production of substantive knowledge claims. ⁶²

However, since critical realism/causal realism is a philosophical doctrine, we must engage with it at that level. Given the apparent tendency of critical realists to try to demonstrate their philosophical correctness with considerable vehemence, it is important to engage with their philosophical argumentation in detail. We are familiar with the critical realists’ often repeated claims that their view is correct or superior to others because (1) it does not allow the ‘anything goes’ attitude, (2) it is ontologically grounded, and (3) it does not make the error of privileging epistemology over ontology. I have suggested that the validity or force of such arguments is dubious. Moreover, causal realism is not defended by merely pointing out, for instance, that if I try to fly unaided, I will fall or that a material object will fall to the ground even if no one is there to witness it. Such examples and the talk of ‘ontological grounding’ and ‘epistemic fallacy’ are hasty knock-down arguments which critical/causal realism need not employ to defend itself – for all that needs to be demonstrated and accepted is that the reality of causal powers, or their presence in the world, is an understanding required by scientific knowledge production and application as well as by our everyday thinking and living. On this point, causal realists and idealists can converge. They can, moreover, be united against the regularity view of causation which still dominates IR – with some considerable impact on the future study of world politics.

I began this article by suggesting that whether causal powers are real (or ‘material’) is an interesting question partly because if they are, we cannot be a radical constructivist. There may of course be other reasons why one would not wish to be a radical constructivist. But as far as causation goes, which has been the sole focus of this article, it would be difficult to do away with an understanding that causal powers are real, mind-independent and in that sense material. And I do not see why we should even want to do that.