Justin Cruickshank is an interesting commentator on the philosophical underpinnings of critical realism. Critical realism was developed initially by Roy Bhaskar in A Realist Theory of Science and The Possibility of Naturalism: A philosophical critique of the contemporary human sciences, and has been further elaborated by a number of philosophers. The theory is now playing a lively role within sociology and sociological theory. Cruickshank’s key ideas are developed in several papers, “A tale of two ontologies: an immanent critique of critical realism” (2004) (link), “Knowing Social Reality: A Critique of Bhaskar and Archer’s Attempt to Derive a Social Ontology from Lay Knowledge” (2010) (link), and “The positive and the negative: Assessing critical realism and social constructionism as post-positivist approaches to empirical research in the social sciences” (2011) (link). Fundamentally Cruickshank takes issue with the nature of the arguments that critical realists have offered for their specific ideas about ontology.
Cruickshank regards the doctrines of critical realism as expressed by Bhaskar and his successors as fundamentally a philosophical theory rather than a highly general and abstract social theory; and he finds that the theory is justified on several lines of philosophical argumentation. The arguments that he criticizes involve apriori philosophical reasoning and inference from lay concepts about the natural and social worlds.
“A tale of two ontologies” highlights the philosophical presuppositions and language of critical realism — assumptions about the variants of ontology (transitive and intransitive), absolute metaphysical knowledge, transcendental metaphysical knowledge, conceptual science, immanent critique. Cruickshank finds that Bhaskar embraces the idea that critical realism is a philosophical theory rather than a scientific theory, and that this places the theory on shaky ground:
In support of the differentiation of philosophy from science, and contrary to the claim made about the historical transitivity of ontology made in response to Chalmers, Bhaskar says he avoids the epistemic fallacy by producing a philosophical ontology. He argues that if we conflate scientific and philosophical ontologies then we commit the epistemic fallacy, by remaining confined within questions about knowledge. (573)
The transcendental method that Bhaskar uses, according to Cruickshank, is based on Kant’s philosophical theories:
Against empiricism, Bhaskar’s transcendental realism (which was later renamed ‘critical realism’) holds that the condition of possibility of science is the explanation of causal laws which are different from the changing contingent observable regularities we may perceive outside experiments. The ontological turn advocated in RTS is meant to render explicit the ontological presuppositions implicit within the practice of science. In doing this, Bhaskar argues that the condition of possibility of science is the existence of underlying causal laws in open systems (i.e. systems characterised by change with no observable constant conjunctions), rather than causal laws being observed constant conjunctions within artificial closed laboratory systems. (569-570)
But this method leads to a conundrum:
The version of ontology required to allow critical realism to fulfill its hegemonic project rests on a dogmatic metaphysical claim to know a stratum of ultimate reality beyond knowledge. Critical realists try to avoid such explicit dogmatism by defining ontology in terms of the transitive domain rather than the intransitive domain. However, defining ontology in terms of the transitive domain commits the epistemic fallacy, and precludes any possibility of the ontology being used as the basis for an hegemonic project, as the ontology would be fallible and hence open to revision (unless dogmatically privileged). [my italics] (581)
So Bhaskar et al have painted themselves into a metaphysical corner: they require that ontology should be about reality as it really is (intransitive); they retreat from the implication of a dogmatic philosophical position; and they wind up in the position of conceptual relativism (transitive domain) that they sought to avoid.
Cruickshank plainly prefers to deal with these issues in a way that is not so dependent on purely philosophical arguments. Here is the position that Cruickshank thinks is most reasonable:
We may accept the view that ontological questions are important questions, and argue that we ought to regard ontological theories as fallible interpretations of reality. In other words, the focus in this article is on the status claimed for ontology, and not the issue of wether one or other substantive social ontology is the definitively correct or incorrect definition of social reality. The emphasis is on continually developing ontological theories through critical dialogue, rather than arguing that an individualist, or structuralist, or praxis based ontology, etc., is the correct definition of social reality. (568-569)
In contrast to foundational epistemology which defines reality to fit a subjective, mentalistic foundation, we may adopt an anti-foundational approach that rejects the starting point of epistemology as the separation of the lone mind from the world. We may instead hold that our beliefs are engaged with the world and that we need to revise and replace our theories in the course of our engagement in the world. (582)
As regards social ontology this means that social scientists need to become engaged in an on-going debate about the ontological theories currently existing in the transitive domain. This debate needs to turn not just on the use of immanent critique, to assess the internal coherence of a position, but also on the usefulness of an ontology in informing empirical work. (583)
And in fact, this seems like an entirely defensible way of thinking about the role of ontology: not as a set of philosophical truths to be established by a priori arguments, but rather as a revisable set of ideas coherently related to the best scientific conceptual systems we have developed to date.