Harry Collins, Expertise, and Elective Modernism

I have attended a talk at Tel Aviv University by sociologist Harry Collins, one of the leading scholars in Science and Technology Studies today. In the talk, Collins presented the current state of his and Robert Evans‘ theory of expertise and his new notion of elective modernism. I want to share with you some thoughts about them, particularly about why, despite their promise and appeal, they only partly work.

In a nutshell, Collins and Evans distinguish between three waves of science studies. The first wave was traditional philosophy of science and functionalist sociology of science. Philosophers studied science as a logical system, while sociologists tried to identify the norms governing successful knowledge production. The second wave was after the publication of Kuhn’s Structure of Scientific Revolutions, where sociologists started to study science descriptively, adopting a relativistic framework, under the assumption that science is just politics by other means. The third wave, to which Collins and Evans want to steer the field of Science and Technology Studies, brings back the normativity of the first wave without giving away the insights of the second wave. They argue, for example,  that a relativistic second-wave sociologist has no intellectual resources to counter an African man’s claim that according to his local knowledge system, having sex with a virgin is the best way to cure AIDS.

At the heart of Collins and Evans’ theory is a taxonomic theory of expertise, where expertise is conceived as a social category. The idea is that there are many forms of expertise. By identifying the kind of expertise that is relevant to a given case, science studies scholars can make normative recommendations for how to deal with different cases without resorting to first-wave philosophical notions of truth and rationality, which Collins and Evans reject as misguided or unhelpful.

Collins and Evans’ “Periodic Table of Expertise”

In the talk, Collins addressed a central tension that I had always thought existed in his work. Collins argues that certain situations normatively call for scientific expertise, rather than other forms of expertise. A paradigm case would be a community of scientists that research an esoteric subject that requires special training and skills. In such a case, Collins and Evans argue that scientists should have the autonomy to pursue research as they see fit, and be shielded from external political influences as much as possible. They further argue that if the outcome of scientists’ research is relevant to some policy questions, policy makers should defer to the scientific consensus in the scientific community. At the same time, however, Collins is committed to his theory of the experimenter’s regress, which states that the outcome of a scientific consensus is determined by contingent local circumstances. Had the local conditions been different, a scientific consensus would have been formed around another position. That is, there is no direction for the growth of scientific knowledge. In particular, it does not advance toward the truth.

These two commitments seem in tension with one another. Why should policy makers defer to a scientific consensus if its outcome does not reliably tell the truth, but is rather an accidental by product of particular local social circumstances in a particular scientific community at a given time?

In his talk, Collins gave a reply to this question. The reply is his new theory of elective modernism, which states that we should not trust science because its outcomes are true or rational, but rather because it is committed to values of truth and rationality, regardless of whether it successfully produces them.  The justification here is not epistemic. Rather – for the lack of a better  term – it is religious. Science and its values of modernity are the best system of values among the available alternatives to run our lives by. Therefore we should choose it.

I find this justification troubling, however, for several reasons. First, even religious justifications have an epistemic dimensions. Many people do not choose a religion simply because it fits their personal values. If they take their religion seriously, they also want to know if it’s true. And not just for the sake of curiosity, but for a very important purpose – they need to know if they are going to hell or not.

Second, some people, such as Brian Wynne, would say that the question of which value system is preferable is contextual. Scientific values are good in some epistemic contexts but not others. I am not sure Collins would necessarily object to this, as he and Evans argue that different contexts call for different forms of expertise.

Third, I think that Collins and Evans’ claim that there are case in which  lay people do not have a choice but to blindly defer to a scientific consensus is misguided, politically and epistemically. This is not the right way to deal with the problem of democratizing and legitimating scientific knowledge. As I have written here before, and in a published paper as well, policy makers should not simply defer to a scientific consensus, when such exists. They can do better. They should only defer to it if the consensus has certain properties that they can detect.

This failure of Collins and Evans’ normative theory of expertise stems, in my view, from their insistence not to use the terms truth, evidence, and rationality, which are simply unavoidable when dealing with normative epistemic questions regarding the role of science in policy.

Philosophers, or more precisely, social epistemologists, have not shun away from these notions. And they have in fact been successfully doing third-wave science studies for a while now. Take for example, Heather Douglas – a frequently mentioned philosopher in the Bubble Chamber. Her solution to the problem of the role of science in policy is distinguishing between an illegitimate direct role for social values in evidential reasoning, and a legitimate indirect role for values. While her solution seems more conservative and traditional than Collins and Evans’, as it uses the traditional language of truth and evidence, appearances can be misleading. In Douglas’ account, as opposed to Collins and Evans’, social values may and should legitimately influence the formation of a scientific consensus as well as the final position the scientific community ends up accepting. Douglas allows the final outcome to be both epistemically justified and still dependent to some extent on local social circumstances. With Douglas’ account, proponents of the democratization of  knowledge have a legitimate basis from which to critique and influence the inner-workings of science. This is something that Collins’ elective modernism does not give them, and I doubt if it can.