22/05/2009

Interview with David Byrne

The following is a brief interview I conducted with British Sociologist and Complexity Scientist, David Byrne.

Dr. Byrne is Professor in the School of Applied Social Sciences at Durham University, England, where he is also Director of Postgraduate Studies. Dr. Byrne is the author of several books and a long list of articles, including his 1998 book, Complexity Theory and the Social Sciences--the first book to critically review and explore the application of complexity science to sociological inquiry. His most recent book, edited with noted sociologist and methodologist, Charles Ragin is The SAGE Handbook of Case-Based Methods

Dr. Byrne is an expert in methods, urban planning, community health, social policy, social exclusion and complexity science.

-----------------------------------------------
INTERVIEW WITH DR. BYRNE

CASTELLANI: Dr. Byrne, thanks so much for taking the time to do this interview. Your research agenda is rather vast in its scope—ranging from the philosophy of complexity science to method to urban planning to health care inequality. If you do not mind, I am going to narrow in on method first, given its wider implications for those reading this blog—most of whom are students and researchers new to the field of complexity science and its practice within sociology.

A. Case-Based Research
CASTELLANI: For the last several years, you have been a major advocate of a case-based approach to research. You specifically endorse what you and Charles Ragin refer to as Qualitative Comparative Analysis (QCA). First, how do you define QCA?

BYRNE: It is a method which is ‘set theoretic’ i.e. it understands causal relations in the social world in terms of relationships in combination – sets, rather than the unique contribution of single variables. It is based on systematic comparison – essentially an extension of John Stuart Mills’ method of differences. It requires careful qualitative engagement with specific cases as the foundation of that comparison.

CASTELLANI: Of the three major types of QCA (crisp-set, multi-value and fuzzy set), which do you find most useful and why? Or, do you approach the distinctions within QCA a different way?

BYRNE: I generally work with crisp set techniques and actually almost never go beyond the truth table. So I use QCA as a kind of mix of exploratory / explanatory – often focusing on ‘contradictory configurations’ in which the assemblage of elements in the line of the truth table – the configuration – generates different outcomes. That makes me look at those cases for what else is different about them. I see multi-value QCA as an extension of crisp set but it is much more complex to use. I frequently use Cluster Analysis as a data reduction technique and binarize membership of a cluster. Fuzzy set is very interesting and I have thought about how we might use distance from a cluster centre as a fuzzying principle but I have never managed to bring it off.

CASTELLANI: For researchers and graduate students new to case-based research, what is your best argument (apologetic) for including QCA in their toolbox of techniques?

BYRNE: For me the crucial things about QCA are the following:

•It allows for complex causation – lots of things acting together to generate an outcome. Conventional statistical modelling can do this in a limited sort of way through interaction.

•It allows for multiple causation – different combinations – in QCA terms configurations – can generate the same outcome. More than one way to skin a cat.

•It really makes us think about ‘what is a case’ – what Charles Ragin calls the processes of casing – just as important to specify the character and boundaries of cases as to be careful about operationalizing in measurement of what I prefer to call attributes or variate traces rather than variables.

•It really does have qualitative phases – conventionally at the beginning because the researcher really does have to engage closely with cases using qualitative techniques in order to establish attribute values. If you start, as I have often, with a data set of pre-given measures, you often have to move on to qualitative investigation to explore further differences.

•That word – differences – QCA is founded on distinctions.



B. Epistemology

CASTELLANI: Your research agenda is grounded in what you refer to as a complex/critical realist approach. What is complex/critical realism?

BYRNE: The term comes from David L. Harvey and his collaborator Reed. It involves a synthesis of the critical realist perspective of Roy Bhaskar (but the early Bhaskar) and complexity theory. So it says most of the world is made up of complex systems – although see Paul Cilliers’ important work on how such systems are both real and the products of scientific construction – the complexity part. Then it endorses critical realism’s deep ontology of the real as generative mechanisms, the actual as the contingently and contextually expressed outcome of those mechanisms (I wish we had another word than mechanisms), and the empirical as what we as scientists make from those mechanisms in action in the actual. Note ‘make’. This is a constructionist position but one which says that the real also has a say.

CASTELLANI: Why should researchers consider your epistemological approach important enough to adopt?

BYRNE: I would say it is David L. Harvey’s and I adopted his approach because it enabled me to make sense of social causality and allows agency, including conscious and informed agency, into play with the potential for knowledge to actually be applied in a meaningful and useful fashion. Does that for me and I recommend the treatment to others for the same reason.

C. The Complexity of Place, Space and Health.

CASTELLANI: Our Q&A is situated within the larger theme that I have been blogging on for the past couple weeks: how to improve the community health science literature by adopting a complexity science perspective.

You may disagree, but a major theme that I see in your work over the last decade is your rigorous and nuanced attempt to develop a methodological-epistemological framework researchers can use to develop better models of the complexities surrounding place, space and health. This includes the complexities of social exclusion, urban planning, spatial inequality, and the challenges surrounding the relationship between individuals and the communities in which they live. For example in your chapter, Complex and Contingent Causation—the Implications of Complex Realism for Quantitative Modeling (found in Carter and New’s Making Realism Work, 2004) you address one of the biggest challenges facing the community health science literature today: the inability of researchers to create a satisfactory way to address the relationship between micro-level health outcomes and aggregate level phenomena such as the neighborhood effect.

You state: “Multi-level modeling has been proposed as a way of resolving the difficulties of cross-level relationships among individually expressed health and social conditions. This interesting approach does represent a genuine effort to confront problems which are central to the relationship between the collective and the individual. However, this chapter will argue that the approach remains unsatisfactory, precisely because it ‘disembodies’ both aspects of the complex individual and aspects of the complex social systems through which individuals lead their lives” (p. 51).

CASTELLANI: What do you mean that researchers tend to “disembody” complexity?


BYRNE: Disembody is a specific kind of abstraction. Abstraction is necessary – I think Katherine Hayles is great on this in her How we became post-human but we also have to be very careful. I was using Chris Allen’s arguments – which I found interesting, well put and provocative – to frame my own argument. Chris was saying: don’t lets regard agentic human beings as physiological dopes ‘determined’ by the external and their own attributes in interaction. He pointed out that there is real variation in outcome – the reality of any probabilistic form of explanation of cause e.g. in a randomized controlled trial (RCT). I agree up to a point but think that we can move towards a better account if we think really hard about complex and contingent causation. I have written elsewhere about how I don’t have TB despite being exposed to cases in adolescence and having a very strong Heaf test reaction at that point. Too well fed, too well housed, and with parents who didn’t get the disease or die of it whilst they both had siblings who did and did so bred for resistance. But if I get AIDS or am starved in conditions like a WWII Japanese prisoner of war camp, then I will get TB. That is complexity expressed in my individual body and I want a modelling process which moves towards allowing for that.

CASTELLANI: As a solution, how do you think the methodological-epistemological framework you have developed helps researchers to preserve the complexity of their models?

BYRNE: First by making us think about it. Second, by looking for and using methods, quantitative and qualitative, which respect the complexity of the real as opposed to artificial (I owe this distinction to Elias Khalil) world. So always be skeptical about simplicity. It might be there but mostly it isn’t.

CASTELLANI: Related, you and others (such as Paul Cilliers and Charles Ragin) have criticized complexity scientists for making the same reductionistic mistake as multi-level researchers: complexity scientists still seem to reduce to an unnecessary level the complexity of systems. Why do you think complexity scientists fall prey to this reductionistic tendency? How do they get out of this trap?

BYRNE: See Morin’s excellent essay on this very point at: http://cogprints.org/5217/1/Morin.pdf

My take is that the kind of complexity which says we can always generate complexity from simple interactions following for example rules – note always, I have no quarrel with sometimes here – ends up with specifications which ‘look like’ the laws of Newtonian science although of course they are nothing of the kind. However, they are reductionist – you can do this if not in a white coat then in a techy sort of way which makes you look like a proper scientistic scientist. There is a real battle to be fought here although interestingly there are physicists – Peter Allen’s excellent work for example – and lots of eco centred biologists – as well as medics – who are beginning to recognize that they cannot deal with problems of explanation and action without dealing in what Morin calls general complexity.

D. The Future of Sociology

CASTELLANI: Without creating a straw-person, I think it is fair to say that sociologists, particularly those in the main-street of the profession have been slow to embrace or involve themselves in a critical dialogue with complexity science. What is your best argument for why sociologists should involve themselves in the new science(s) of complexity?

BYRNE: Because it allows us to deal with systems without falling into the Parsonian trap (although note that Parsons did have a sense of the complex from time to time). It also is a way towards agentic intervention. My first degree was in Sociology and Social Administration – we would usually but not necessarily correctly talk about Social Policy instead of administration today – and my Master’s was in that field rather than mainstream Sociology. I am an applied social scientist and complexity pushes towards action. It also is a way of getting past what frankly I see as the dead hand of much of contemporary sociological theory. Post modernism is a dead end but I am thinking here as much of Giddens and even of Bourdieu (and I have a deal of respect for Bourdieu). We need to engage empirically and get beyond the absolutely necessary preliminary task of empirical description into a serious and non-positivist engagement with social causality. That is what complexity lets me do.

---------------------
CASTELLANI: Dr. Byrne, thank you so much for your time. For more information on Dr. Byrne's work, visit his website by clicking here.

No comments:

Post a Comment