Comment three on the July issue of Science, Technology, & Human Values (July is volume 37, number 4).

Recently, I claimed that STHV had produced what I thought was one of the best issues in a while.

I will, in a series of posts over the next month or so, comment on each paper in the special edition on context with comments, criticisms, and occasional tangents.

Third paper is “Context and Culling” by Law and Moser.

Original Abstract:

This article asks how contexts are made in science as well as in social science, and how the making of contexts relates to political agency and intervention. To explore these issues, it traces contexting for foot-and-mouth disease and the strategies used to control the epidemic in the United Kingdom in 2001. It argues that to depict the world is to assemble contexts and to hold them together in a mode that may be descriptive, explanatory, or predictive. In developing this argument, it explores how contexts are assembled in a series of different descriptive and explanatory narratives in epidemiology, policy, critical social science, and (feminist) social studies of science

Reaction and Commentary:

This is a good lesson piece in this new irreductive camp; however, some elements of this piece will madden readers. The piece is about the foot and mouth disease outbreak in the UK during the early 2000s. Much of the empirical material comes from a government document about the early events and a retrospective account of why the epidemiological predictions were so far off the mark either before, after, or during the events.

The best part of the article will have to wait until the end this review (it is about how epidemiological modelling might need to account for context a bit more, and in new irreductive ways), but some basic points are still worth rehashing. First, the chapter takes for granted this insight: contexts don’t really exist because contexting is done (RE: active). As it happens, contexts are constructed, and they tend to be treated as (unnecessarily) coherent constructions when observed first-hand in the wild, or when crafted second-hand by sociologists or scholars in STS. To show evidence for this, Law and Moser provide a number of accounts (which seem to layer on one another nicely), but which are unfortunately called “contexts” (rather than accounts or explanation generation techniques). These accounts become “contexts”, I will grant them, in that they begin to populate the setting and describe the relationships between relevant actors and entities (keep in mind, this is foot and mouth disease among sheep). They begin with a generically objective account of how epidemiological modelling is imperfect and takes many forms, and then take to comparing the two forms that make their way into the case study. After that context, they layer another context (RE: account) onto it, namely, one that STSers will like: a critical account where political connections among leading scientists leads to a corruption of the policy implications of the otherwise scientific models, which results in a culling of many sheep that otherwise might have lived, but that seem to have fallen under the axe of political interests too powerful and broad to protect them. <lingering pause> Then the authors backpeddle a bit, showing how none of these accounts is sufficient, and find a critical examination of both accounts, similar to accounts by Singleton about policy, where we ask: was any policy towards the culling of sheep with foot and mouth disease ever really enacted in full? (thus, making any explanation quite difficult to support) … the concluding remarks suggest that any policy being enacted is incomplete, the policy itself was different in different locations and differently applied in similar/proximal locations … there was, in fact, no one policy to assess for effectiveness and instead a plexus or cacophony of variations on the recommendations were enacted. The policy was multiple; a differentiated singularity. This is quite important because it is as easy to make errors like this (RE: treat a policy as non-multiple) when telling Whiggish retrospective accounts as it is when trying to establish prospective accounts — and this applies to both scholars as well as those persons in the public that do this sort of contexting (or account-making). In the end, and here is the best part, Law and Moser suggest that this is one way to make epidemiological modelling even more effective; erase our bias for coherent accounts where policy adherence and contextual coherence are assumed.

One serious pick at the piece, though, must be the style of writing; without being too critical or traditional, the conversational tone, which had the cadence of a bright teacher talking down to a dull student, was difficult at times to stomach…