There are many things that I like about Tuukka Kaidesoja’s new book. It is systematic, thorough, carefully argued, well judged, and ambitious while remaining measured and humble. But what I like most about it is the larger intellectual project it seeks to implement: the project of naturalising critical realist social ontology, or in other words reconstructing critical realist philosophy and its account of the social world such that they are continuous with the empirical sciences and sensitive to their practices and findings. No doubt the book could be read as an attack on original critical realism – the early philosophical work of Roy Bhaskar that provides the central arguments of contemporary critical realism – but it is clearly intended as constructive rather than destructive criticism.
Kaidesoja pursues this project by carefully examining the rhetoric that Bhaskar uses to frame and develop the argument for his particular version of scientific realism. As Kaidesoja makes clear, Bhaskar deploys a terminology, in A Realist Theory of Science (Bhaskar 1975) and to a lesser extent in The Possibility of Naturalism (Bhaskar 1989 ), that seems to invoke the authority of Kant in the analytical tradition, but in the process introduces a series of unfortunate associations. The most central of these is Bhaskar’s use of the concept of a transcendental argument to describe his argument for the existence of real causal powers, which he derives from the practice of scientific experimentation. This Kantian terminology has prompted critical responses from scholars who read into it a claim to infallible certainty for his conclusions, and who take it to imply that Bhaskar’s argument proceeds from a priori premises (Cruickshank 2004; Little 2013).
But these scholars, I suggest, are throwing the baby out with the bathwater. The bathwater, here, which does deserve to be discarded, is the pseudo-Kantian rhetorical form of Bhaskar’s argument. It could be pointed out in his defence that Bhaskar redefines the concept of a transcendental argument so that it no longer carries the same implications (e.g., he says that transcendental arguments are fallible: Bhaskar and Hartwig 2010, p. 63). Nevertheless, it was surely a poor judgement call to use the term in the first place, given its indelible association with claims to certainty, and there are moments when Bhaskar does seem to use this and related terminology to clothe his argument with some sort of Kantian authority. The baby, however, which really should not be discarded, is the substantial content of the argument: the case it makes for the existence of persistent causal powers. Kaidesoja’s great merit is that he sees beyond the superficial form of the argument, and while he dismisses that form, he also recognises that this does not entail that the substance of Bhaskar’s argument is wrong. What makes his book particularly valuable is that he goes on to reconstruct Bhaskar’s argument on a naturalistic basis. He rescues the baby, dries it off, and dresses it in a smart new set of clothes.
The baby that Kaidesoja recognises and rescues is that Bhaskar’s argument is an inference from successful scientific practices (experimentation in the physical sciences) to ontological claims. Unlike Kant’s transcendental arguments, there is nothing necessary about the specific conclusion that Bhaskar draws from the inference, though it’s certainly a plausible one. There are many ways in which the inference could in principle go wrong, as Bhaskar himself recognises, and as Mervyn Hartwig, for example, has recently made clear in his defence (Hartwig et al. 2013). And unlike Kant’s transcendental arguments, there’s nothing a priori about Bhaskar’s minor premise, which is a simple empirical claim: that experimental science succeeds in telling us useful (though fallible) things about how the world works beyond the laboratory. This claim, like any empirical claim, may turn out to be false, but it is surely highly likely to be true and it is a perfectly reasonable non-foundational premise for an argument. Kaidesoja argues that the logical form of Bhaskar’s argument would be better reformulated, not as a transcendental argument, but as an inference to the best explanation: a scientific form of argument from empirical premises.1 Hence it is entirely reasonable to conclude, with Bhaskar, that one of the reasons that experimental science is able to tell us useful things about the world beyond the laboratory is that we live in a world of consistently reproduced causal powers.
While there are many other excellent arguments in Kaidesoja’s book, there are also some I would wish to question. Let me focus here on one that he particularly stresses in his conclusion, alongside the argument covered above. This is an argument that seems to be somewhat less successful in saving the related baby. In it, Kaidesoja criticises Bhaskar’s depth ontology, and in particular his separation of reality into three domains that he labels the empirical, the actual, and the real. Kaidesoja describes the third of these as a “transcendental realm of being beyond concrete material objects” (Kaidesoja 2013, p. 204). For Bhaskar, the empirical is that part of reality that is experienced by perceiving individuals. The empirical is a subset of the actual, which consists of everything that exists and all the events that occur to everything that exists. And the actual is in turn a subset of the real, which also includes some non-actual elements. Kaidesoja characterises the non-actual elements of the real as abstract universals, and insists that abstract universals do not exist, and hence that we should reject the notion of a non-actual real.
It seems to me, however, that on this occasion Kaidesoja fails to see through the rhetorical form to the point of Bhaskar’s argument. What Bhaskar is arguing, as I understand him, is that there are true facts about the world, the referents of which are not actual things or events, and that we need some space in our ontology for these referents. In particular, there are facts of the form “entities composed of parts of the types l1, l2, … ln organised according to the relations r1, r2, … rn will possess property p”, and the truth of such facts does not depend on whether or not any such entities actually do exist. To restate the point a little differently, there are types of things of which it is true, irrespective of whether tokens of the type actually do exist, that tokens of the type would have certain causal powers if they did exist.
Let me offer a statement that illustrates the point: if a charged battery, a laser diode, a lens or two, a switch, some wiring, and a casing are linked in a configuration that falls within a certain range, the resulting object (which we sometimes call a laser pointer) will have the power to project a small point of coloured light onto a surface when the switch is depressed. For Kaidesoja, such objects, when they exist, do have causal powers. But there is something more about the world that is missed if we confine our understanding of causal powers to this recognition. It is not just a happy coincidence that particular laser pointers happen to have the power to project a point of light, but rather the consequence of a feature of the world that transcends the question of whether such objects actually exist at any particular point in time. At the time of writing, such objects do exist in the world, but the statement above about the causal powers of laser pointers would remain true if all such objects were destroyed, and the statement was already true before any such objects had been created or even thought of. The statement describes something about the world that is independent of the statement itself, independent of the existence of laser pointers and their parts, independent of the existence of human beings or of any entity capable of making or understanding such statements, and independent of the existence of beings capable of making or using such objects. Bhaskar calls that something a real causal power. Real causal powers do not exist in a material sense, but experimental science makes claims about them, and it is just such claims that make experimental science useful for understanding the world outside the laboratory.
We might go on to say that actual tokens of the type concerned have actual causal powers – i.e., particular laser pointers have the actual causal power to project a point of light – though Bhaskar does not use this term, but this is a different point. As Kaidesoja has pointed out to me (personal communication), Bhaskar at times invokes something like Harré’s account of powerful particulars, implying that actual things do have causal powers, but this is entirely consistent with his argument that an adequate ontology must also encompass real (but not actual) causal powers, and thus include elements that lie beyond the domain of the actual. This does not entail that mysterious abstract universals ‘exist’ in anything like the material sense that ‘exist’ entails in popular usage (and indeed Bhaskar avoids using the word ‘exist’ in connection with the non-actual real) but only that we need to recognise that there are facts about the world that are independent of corresponding existents. While we may not be able to learn such facts until the corresponding existents appear, that does not affect the ontological question.2 We need some way of including such non-actual facts about the world in our ontology, and this is the function of Bhaskar’s domain of the non-actual real.
Perhaps we could reformulate Bhaskar’s argument in clearer and more obviously naturalistic terminology, though I am not aware that he has ever called the domain of the real a “transcendental realm of being”, and he certainly does not usually refer to it in this way. If he has done, there is perhaps some rhetorical bathwater worth disposing of, but whether or not this is so, there is also an ontological baby worth saving. And it is a baby that in my view is thoroughly compatible with placing critical realism on a more naturalistic basis.
The use of transcendental rhetoric was perhaps understandable in the context of the analytical tradition in the 1970s. In the contemporary discursive context it suggests a foundationalism that is entirely incompatible with the critical realist project, and Kaidesoja has performed a valuable service for critical realism by explaining so clearly the problems this creates and the kind of route that realists should follow to avoid them. If we follow that route a little further we will find yet more value in Bhaskar’s early work.
Bhaskar, Roy (1975): A Realist Theory of Science. Leeds: Leeds Books.
Bhaskar, Roy (1989 ): The Possibility of Naturalism. London: Routledge.
Bhaskar, Roy and Mervyn Hartwig (2010): The Formation of Critical Realism. Abingdon: Routledge.
Cruickshank, Justin (2004): “A Tale of Two Ontologies: An Immanent Critique of Critical Realism”. In: Sociological Review 52. p. 567–585. [Crossref]
Hartwig, Mervyn (2007): Dictionary of Critical Realism. Abingdon: Routledge.
Hartwig, Mervyn, Dave Elder-Vass and Daniel Little (2013): “Replies by Elder-Vass and Hartwig to Cruickshank/Little”. http://understandingsociety.blogspot.com/2013/12/replies-by-elder-vass-and-hartwig-to.html, accessed 17 February 2014.
Kaidesoja, Tuukka (2013): Naturalizing Critical Realist Social Ontology. Abingdon: Routledge.
Little, Daniel (2013): “Reply to Elder-Vass, Hartwig, and Groff on critical realism”. http://understandingsociety.blogspot.co.uk/2013/12/reply-to-elder-vass-hartwig-and-groff.html, accessed 17 February 2014.
About the article
Published Online: 2015-07-04
Published in Print: 2015-09-01
©2015, Dave Elder-Vass, published by De Gruyter. This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License. (CC BY-NC-ND 3.0)