Critical Realism and the Philosophy of Social Science: Some Recent Interventions

Some participants to these discussions have also published monographs developing critical realist approaches to social explanation (e.g. Elder-Vass 2010; Porpora 2015), while critical realism itself has come under critical scrutiny from other versions of realist philosophy.

Daniel Little is an analytical philosopher who has created a ‘blog' to engage with philosophy of social science on line. His superbly clear and rigorous New Directions (2016) developed some of the themes in the blog into a systematic approach to the discipline. He endorses many of the philosophical ideas of critical realism (though having reservations about the ‘unnecessary obscurity' of some of the work of Roy Bhaskar), defining his own position as ‘scientific realism’. In developing this position Little reaches out also to features of both the meaning-centred (‘interpretivist') tradition in philosophy of social science, and what he calls the ‘assemblage theory' associated with Deleuze and Latour (see Chapter 8).

At the core of Little's approach is his recognition of social actors as constitutive of social reality. However, he strongly distances himself from methodological individualism, arguing that, first, social actors are themselves formed by the social contexts in which they develop and act, and, second, that the norms, institutions, organizations and social structures which are the aggregate outcome of their actions have a sufficient degree of stability to figure as real, with causal powers which reciprocally affect the conditions of action of individual actors. Despite the reality of higher-level structures, Little insists on what he calls Thicrofoundationalism' - ultimately it is the situated activity of social actors that is constitutive of the higher levels of social life. This suggests a number of heuristic priorities for substantive social scientific research, such as a focus on the local, everyday practices of people, and a reluctance to postulate effects of larger scale social structures in the absence of an account of how they might have arisen from the activities of social actors.

Part of the value of this book is the way Little carefully works his way through the different ways in which a range of realist philosophers have defined the levels (or level) of social structures above that of the combined individual actors. He cites Kaidesojas (2013) discernment of three different conceptions of the ‘emergence' of causal powers at higher levels of social structure, arguing that only a more ontologically ‘weak' version is defensible. It is agreed that social structures, processes, etc., may have properties/causal powers that are different from those possessed by the agents that compose them, but these emergent powers must themselves by capable of being understood in terms of those agents - that is, a microfoundational account must be (in principle?) possible. There is another convergence with critical realism in Little's rejection of a Humean (event regularities, statistical correlations) account of social causality in favour of a requirement to specify plausible mechanisms at the appropriate level.

There are some variations in the way Little defines both ‘microfoundations’ and the related approach of ‘actor-centred’ sociology. Early on he says ‘[S]ociological theory needs to recognize and incorporate the idea that all social facts and structures derive from the activities and interactions of socially constructed individual actors’ (Little 2016: 37). Elsewhere we get ‘[S]ocial phenomena are constituted by the actions of individuals, oriented by their own subjectivities and mental frameworks... their thought processes and mental frameworks are developed through myriad social relationships and institutions. So the actor is a socially constructed individual’ (pp. 40-1). Later in the book we have a definition of ‘methodological localism’ as a ‘social ontology grounded in the actions and social relations of socially constituted individuals’ (p. 188).

There are several equivocations here, but the shift between ‘activities and interactions’ and ‘actions and social relations’ is significant. Are the social relations linking individuals part of the ontology, or is it merely their ‘interactions’. If the former, then the social is already incorporated into the professedly individualist ontology. If the latter, then it is hard to see why (presumably) ad hoc and contingent ‘interactions’ could provide the ‘foundations’ of the wider patterns that constitute societies. The problem becomes clearer if we consider the social ‘constitution’ of the individuals. If social actors are so thoroughly constituted by their formation through participation in social relations and institutions, then what is the case for seeing individual activities as foundational relative to social structures? There seems to be a perfect symmetry of reciprocal causality. Social actors are thoroughly formed by their participation in social structures, and, in turn, social structures are thoroughly produced by the activities of individuals. There seems to be no basis in this account for giving ontological primacy to individuals or their activities, as is required for the ‘actor-centred’ approach.

Reference to individual actors as socially ‘constituted’ or ‘constructed, together with the further exposition in the second quotation strongly suggest a sociologically reductionist view of individual social actors. There is no mention of any respects in which our social being may be shaped by the physical or non-human contexts of our development, or, indeed, of pre-social dispositions or ‘given’ personal attributes. It could be that Little’s thinking is influenced by unconscious recognition of the extent to which our social being is built upon, enabled by but also limited by, our embodied existence as needy and demanding evolved primates.

It might seem that this re-thinking of the nature of the individual actor could provide an alternative basis for the ‘actor-centred’ approach - actors may contribute something to the social mix on their own account, independently of their formation by participation in social structures. But even this way to ground the actor-centred approach seems difficult to sustain, as social relations and structures themselves are more than forms or combinations of the activities of individuals. Some views of social relations and structures reduce them to discursive processes, as if the whole of social life were composed of verbal or quasi-verbal communication between persons, but this is not Little's view. Many of his examples mention the geographical distribution of aspects of social life, sociological accounts of inequality, technological innovation and so on. Material practices of immensely varied kinds involve the incorporation of materials, substances, forms of energy, animals, plants and micro-organisms into social relations. It would be impossible to give an account of an economic system, a craft, a form of inequality, a food system without recognizing the constitution of social structures as involving not just relations among (socially situated) human agents but also relations between them and non-human terms, and among them in relation to non-human beings, and so on.

This line of argument takes us to a questioning of the sharp polarization between the natural and the social that is at the core of Little's ontology. It isn't always clear whether his strong claims for social and natural difference have to do with the characteristics of these ‘domains' themselves, or whether they derive from unquestioned caricatures of the sciences of nature. For example:

The social world has characteristics that fundamentally distinguish it from the natural world - heterogeneity, plasticity, and contingency, to name several. It is not a system of law-governed processes; it is instead a mix of different sorts of institutions, forms of human behaviour, natural and environmental constraints, and contingent events.

(Little 2016: 1)

The social world is not like the natural world. Nature is composed of things, forces and geometries that have strong determining regularities whose interactions can be formulated with mathematical precision.

(p. 3)

The implication of these statements seems to be that nature and society are different ‘worlds' which operate in radically different ways. One is homogeneous, unchanging and governed by mathematically specifiable regularities, while the other is heterogeneous, mobile, irregular and complex. I'll illustrate some of the problems with this ‘purification' of the two worlds with a quotation of a passage I wrote some time ago in a book about the ecology and behaviour ofbumblebees. Survey work discovered that one of the few remaining populations of two rare species had survived in a complex of habitats along the lower Thames estuary:

On both sides of the Thames, extending westwards into east London, and eastwards towards Southend and the Isle of Grain, lies a vast area of residential, industrial and post-industrial, retail, waste disposal, port, storage, and transport facilities...

(Benton 2006: 519)

At first sight this description invites comparison with the view of ‘entangled’ and contingent ‘assemblages’ of different beings, substances, processes and so on associated with versions of actor-network theory (see Chapter 8) or even ‘new materialism’ (see Chapter 10). There is certainly no clear separation of pure ‘social’ and ‘natural’ worlds, no contrasting domain of mathematical precision and predictability in opposition to a plastic, contingent and heterogeneous social one. However, a more nuanced critical realist materialism can make inroads into these entanglements and complexities. Geologists can explain the Cretacean deposits and much later overlay of sands and gravels, giving the estuary its fast-draining soils as well as the materials in a later epoch quarried for cement and for road construction. Relict pastures and grazing marshes and sea walls are reminders of a long pre-industrial process of reclamation of low-lying areas from the sea, while agricultural ‘improvement’ was avoided by the siting of dangerous and unsightly refineries, oil terminals and chemical factories among the poorer and less powerful working-class communities that formed along the estuary (see Corke 1984).

The questions posed here are not just how to integrate understandings of what are misleadingly abstracted as ‘nature’ and ‘society’ into coherent explanatory frames for complex realities. Also at issue is the possibility of assessing the truth-status of rival accounts. Since the now widely shared scepticism about empiricist and positivist theories of knowledge has loosened the imagined ties between empirical ‘data’ and research inferences, what is to restrain a free-for-all, ‘anything goes’ approach to knowledge-claims in the social sciences? In a recent paper, Rob Stones (2017) has sharply posed this question. Since the post-positivist ‘turn’, social scientists have a richly varied range of conceptual traditions, with associated social ontologies, many of which we have mentioned in this book. Stones argues that this emphasis on ontology has left open questions of epistemology - of criteria for establishing the cognitive status of substantive explanations that make reference to the causal powers and activities of these hypothesized realities. He offers a series of protocols that might guide evaluation of specific, contextualized explanations in terms of relevance, status and consistency with the presumed ontology. He then addresses the ‘hard case’ of John Law’s development of the actor-network approach, which, if accepted, would rule out the kind of rigorous epistemology advocated by Stones. Law’s approach does recognize important fields of social research to which ‘traditional’ research methods are appropriate, but also challenges the assumptions of stability, predictability and clarity of many aspects of social life. Instead, for Law, what counts as real is embedded in the forms of lay knowledge - the concepts and beliefs - of those involved in practices. As for many practices different participants will have different ways of understanding what they are doing, we are led to the conclusion of multiple realities, with no way of deciding between rival accounts.

Stones is sympathetic to Law’s enterprise, but challenges its rejection of epistemology on the basis of a critical realist distinction between transitive and intransitive objects of knowledge. Though ideas and beliefs are partly productive and constitutive of social reality, this does not rule out critical evaluation of the evidential basis, relevance of status of the beliefs concerned: in other words, beliefs, concepts, consciousness as transitive objects of knowledge (in lay discourse) can also figure as intransitive objects of knowledge on the part of social researchers, who may investigate both the role of the beliefs in practice, but also call into question their cognitive status. Stones also points out that Law has to treat beliefs and practices intransitively in making his own claims about the plurality of the contents of social life.

Elder-Vass (2017) complements Stones’s argument, postulating experience, debate within the research community, context-relevance and consistency with ontology as the basis for an epistemology for critical realist social science, whilst also providing a critique of the actor-network approach’s ontology.