Social causation and biographical research: philosophical, theoretical and methodological arguments

Giorgos Tsiolis and Michalis Christodoulou have written a deeply theoretical book arguing that we should see reconstructive biographical research as a method for constructing causal accounts of social lives and generalizing beyond the particular cases under study. They deploy a critical realist ontological perspective to back up their case, contrasting it with the refusals of positivists and interpretivists/poststructuralists to see biographical research as providing causal explanations. The book is simultaneously ambitious in the depth of philosophical argument it attempts but also quite narrowly focused on biographical research as its object. I would expect this book to be useful for students of methodology at an advanced level: it provides valuable material for researchers developing causal arguments using qualitative research in general and biographical interviewing in particular. In this review I will outline the core argument, pausing along the way to make a few comments on its strengths, on possible concerns and on promising lines of development.

They position their argument by contrasting critical realism’s account of singular causation (in the sense that causality can be identified in individual cases) with positivist understanding of causation as a pattern that can only be detected as a regularity across a large number of cases. Given that biographical research is conducted by interviewing small numbers of people in depth, it does not provide the large-n samples required for positivists to see its results as having causal implications. The classic interpretivist response is to substitute interpretation for causal explanation as the objective of qualitative research. The authors do not abandon interpretation but stress that ‘even if human actions are meaningful, they can still be causally explained’ (5): with a critical realist ontology that sees causality as operating uniquely in every case as a result of the interaction of different mixes of mechanisms in each case, there is no reason to abandon the search for causal explanations in small-n samples.

This leads us towards mechanism-based accounts of causal explanation, but not all mechanism-based approaches are equal. Tsiolis and Christodoulou reject the flat ontology of the sort of mechanism-based explanations that are characteristic of pragmatism and analytical sociology (5, 13) and instead advocate a multi-layered approach justified in terms of ‘the emergent properties of people’s social forms’ (8).

To provide support for an emergentist perspective, they dive into the murky waters of the philosophy of mind. This is not unreasonable, since this is the primary literature on emergence in analytical philosophy (also see Elder-Vass 2010, 28–33). However, this literature is challenging, and although they have done well to understand it this nevertheless leads to one of my concerns about the details of the book. The trouble with using the philosophy of mind as a foundation for discussions of emergence in the social realm is that the putative emergence relation that these philosophers have been discussing is not a relation between the properties of a whole and the properties of its parts. Philosophers of mind are essentially engaged in a debate that is defined by what I have called ‘residual Cartesianism’ (Elder-Vass 2014a). Emergentism is employed in this debate as a way to assert that mental properties – or mind – are somehow independent of the physical brain and its properties. It would be perfectly reasonable to argue that our mental properties are emergent properties of a whole person or their brain, which arise from the interactions between their interconnected lower-level parts, notably their neurons. But these philosophers are trying to do something much more obscure: they are debating whether one set of properties of a whole person (mental properties) can be emergent from another set of properties of the same whole person (so-called physical properties). This is not consistent with realist accounts of emergence in the social world, but this is not terribly clear from the debate in the philosophy of mind literature because it is conducted largely in terms of properties without ever discussing what they are properties of (Elder-Vass 2014a).

On this basis, one of the most popular arguments in the philosophy of mind is Fodor’s account of functional emergence, which, to simplify, asserts that there can be causal regularities at a higher level even though the mechanisms operating at a lower level are different in different cases. This is known as multiple realizability of the lower level, or, in the cases where there are many different versions of the lower-level mechanisms, wildly disjunctive multiple realizability. Tsiolis and Christodoulou are tempted by Keith Sawyer’s argument that the argument from wildly disjunctive multiple realizability can be transferred to the social realm (Sawyer 2001; Sawyer 2002; Sawyer 2003; Sawyer 2005). Unfortunately, Sawyer’s argument is deeply problematic. As Ylikoski points out, the higher level explanation that Sawyer regards as emergent takes the form of a covering law, or an empirical regularity, rather than an explanation in terms of causal mechanisms (Ylikoski 2009, 529), and is thus thoroughly incompatible with a critical realist ontology. Sawyer is also an ontological individualist: for him emergence only produces a functional similarity across a group of cases, while every individual case can be reductively explained in lower-level terms. As I have argued elsewhere, Sawyer’s approach is incompatible with the sort of relational approach to emergence that is called for by a critical realist ontology (Elder-Vass 2014b). Tsiolis and Christodoulou, by contrast, see it as not only compatible with, but supportive of, the critical realist approach (e.g. 27–28, 39, 111–113). Still, they do not really build on Sawyer’s account of emergence, with the result that their invocation of Sawyer does not undermine other parts of their argument.

Their primary ontological commitment is to critical realism, to which they devote a chapter, and they go on to argue convincingly that the critical realist approach to causality is also compatible with the technique of process tracing, and with case-based methods that seek to develop generalized theory from the analysis of carefully selected individual cases. Partly perhaps because it is something I have been thinking about myself, I found their discussion of generalizability particularly interesting. They contrast a positivist approach, in which generalizations are developed in terms of the application of variables to a particular class of individuals, with a process-oriented approach, in which generalizations are developed in terms of causal mechanisms and extension of the argument is driven by consideration of what contextual factors are required for the mechanism to operate (50–51).

As I understand it, their argument is that with a critical realist ontology we have implicit grounds for generalizing from case studies available to us that are not available to either positivists or interpretivists. Because critical realists are committed to the argument that events are produced by mixes of mechanisms that are themselves available to operate whenever the structures that have these mechanisms are present, we have a ready-made route to generalisability: if we find a mechanism at work in a case, we are entitled to believe that it could operate in other cases where the structure with the mechanism is present (subject to potential interference from whatever other mix of mechanisms is present). So we can generalize, cautiously, from case studies without the need for quantitative evidence from large samples (58–62). They then go on to argue that this affects appropriate sampling strategies in research design.

The later chapters are more oriented to reconstructive biographical research as a technique for identifying latent meaning structures. They show a strong awareness of the difficulties of treating biographical interviewing as evidence of life experience, but ultimately take the view that this is a viable means for uncovering the mechanisms at work in social cases. Taking the book as a whole, it does a good job of showing that qualitative research in general and biographical research in particular can make important contributions to developing causal explanations in social research. It is at its strongest when dealing with methodological issues rather than philosophical issues, but it remains quite theoretically focused throughout: this is very much a book about methodology rather than about research methods as such. It would be particularly useful for research students employing qualitative interviewing who need to provide support (notably in their methodology chapters) for the argument that we can validly draw causal conclusions from this type of research.