The use and abuse of science
A sociologist of science, and one-time relativist, on the proper role of experts.
Harry Collins is a sociologist (a distinguished research professor at the University of Cardiff, as it happens). More precisely, he is a sociologist of science. And in this role, he has spent several decades observing physicists who study gravitational wave detection in an effort to understand how science, as a practical enterprise, works; how knowledge is generated and validated as such; and how truth is as much a social achievement as it is a purely scientific one.
The brickbats have never been far away. He has been charged with relativism, accused of postmodernism, and even implicated in the so-called Science Wars of the 1990s, which culminated in the Sokal Hoax of 1996, when Alan Sokal, a staunch scientific realist, successfully duped the lit-crit journal, Social Text, into publishing a paper of meaningless jargon and anti-science posturing.
But as his new book Why Democracies Need Science shows, Collins is no longer seeking to unmask science as just one perspective among others, but to consolidate its authority. The spiked review decided to speak to Collins about rescuing the status of expertise, and why, where and when, he thinks some people really do know best.
spiked review: In Re-thinking Expertise (2007), you talk of a change in social scientists’ attitude towards science in the 1960s, from an imperative to explain why science worked to a questioning of the very truth and validity of scientific knowledge itself. Could you say a bit more about that moment?
Harry Collins: You could say that the movement you’re talking about started in the 1960s, with an iconic book – Thomas Kuhn’s The Structure of Scientific Revolutions. At the time, people complained Structure was turning science into mob psychology. But a lot of what was in Structure had actually been anticipated in the work of a Polish-Jewish microbiologist called Ludwig Fleck in the 1930s. Nobody knew about Fleck’s work for quite a while, although Kuhn did. There were also similar ideas about the nature of scientific knowledge circulating among the positivists at around the same time.
But I would say that the really vigorous movement challenging the then high status of scientific knowledge actually started a little later with the advent of the sociology of scientific knowledge (SSK). ‘Wittgenstein, Mannheim and mathematics’, the first paper in this discipline, was written by David Bloor and published in 1973, but the first empirical work was probably carried out by me, which consisted of papers written in 1974-75. The first of these was about the tacit knowledge of scientists, and the second showed that scientific replication was not a straightforward business, because scientists could never be sure an experiment that they claimed failed to replicate another result had been done properly. And the same went for an experiment that claimed to replicate an earlier result. You could never know which were the competently performed experiments, because experiments are very difficult, and mostly they don’t work, so you have to keep on trying and trying to get them to work, and you never know whether you’ve tried hard enough.
So that meant that the very straightforward scientific claim that a result had been tested and confirmed (or refuted) was problematised. If you asked a scientist how he or she knows his results are correct, he or she would say that he or she could replicate his or her results, and we can’t replicate incorrect ones. But it becomes much more complicated than that, and the argument about whether a result has been replicated turns into an argument about who are the competent scientists and who aren’t. So the idea of competence makes the question of scientific knowledge far more sociological than purely scientific. And having set that argument out in a 1975 paper, I developed it in a 1985 book, Changing Order: Replication and Induction in Scientific Practice.
review: How would you characterise your sociological objective during that period? Were you trying to demystify scientific practice?
Collins: Yes, at that point, that’s exactly what we were doing. If you go back to that immediate postwar period, in the 1950s, science was riding high, having invented radar, the atomic bomb and so on. We had reached the point where scientists had been given too much authority, particularly in relation to issues or areas in which they had no real expertise. So a typical Horizon programme of the 1960s would feature someone in a white coat talking about something or other simply because they were a scientist. And we certainly wanted to demystify all of that. As I often put it, what we succeeded in doing was levelling the cultural playing field. At that point science stood way above other knowledge-generating enterprises, and we helped level it down. The trouble is that it has been levelled too far, and one has to find ways to give science some pre-eminence again.
review: In your attempts to level the playing field, and demystify science, did you relativise truth? Did you make it contingent upon its conditions of production?
Collins: From about 1975 until about 1981, I was a fairly radical philosophical relativist. But it then dawned on me that that position couldn’t really be proved by empirical work. So from the early 1980s onwards, I became a methodological relativist. And all that means is that if you look closely at scientific work, you cannot see where nature gets in in the short term. But we cannot know whether or not there is some sort of hidden-hand mechanism, which means that nature has its effect in the long-term. Nevertheless, if you wanted a good sociological analysis of why scientists come to believe one thing rather than another, you have to work on the assumption that it’s not simply because one thing is true and the other false. That just makes the analysis banal. There are other sociological factors at work.
review: Going back to Kuhn’s work briefly, it may not have been as original as perhaps it first appeared, but it was certainly influential among philosophers, social scientists and even the public. And it did look as if he was advocating a relativist argument, that science is a story not of progress exactly, but of revolution and rupture, a story of a succession of new, incommensurable paradigms of knowledge (ie, from the Ptolemaic to the Copernican). And that, as a result, there was no way of comparing these paradigms, because they don’t construct completely different objects of knowledge. Where do you stand on the relativist implications of what is Kuhn’s founding text?
Collins: With any theory of this kind, you can take it in different directions. So if you want a radical interpretation, the way you’ve characterised it is correct. But Kuhn didn’t want a radical interpretation of his work. And more recently, from the early 2000s onwards, my colleagues and I have been arguing that the radical interpretation is philosophical, but it doesn’t have the practical implications that people tend to think it has. Because we don’t want to live in a society – a post-truth society – where anybody’s view on specialist matters is taken to be as good as anybody else’s. Such a society would be a dystopia. Vaccinations would no longer be deemed necessary; proven medical treatments would be abandoned; energy networks would become unreliable.
If I go to give a lecture somewhere, I always say that you can’t believe that everything is relative as that, that one person’s opinion is as correct as another’s. Because if you did believe that, instead of paying a lot of money to fly me to Germany or to the US to give a talk, you’d just ask the first person walking past the lecture theatre to come in and give the talk, and it would be just as good as mine. So, in actual fact, our society runs on the idea that some people know more about some things than others, and that’s a society we want to live in. We don’t want to live in a society where anybody’s view is as valuable as anybody else’s. So you have to start thinking about how you can reconcile the sort of society you want to live in with philosophical analyses such as those of Kuhn, or sociological analyses like mine.
And the way of reconciling it is to stop thinking so much about the truth of the matter, and start thinking about who has the expertise. Now you never know how the truth of the matter is going to turn out. That’s a very long-term business. But you can say straight away who’s got the most expertise, who knows what they’re talking about. And you can say that the best technical decisions you can make are going to depend on giving the views of people who know what they’re talking about more weight than the views of those who don’t know what they’re talking about. Now this can’t guarantee the right decision, but it can guarantee giving you the best decision in a moral sense.
So, obviously, if you want to know whether a vaccination is good or bad, you ask someone who’s studied it; you don’t ask your mum or the first person you meet in the supermarket. We don’t want to live in a society in which alternative facts are given the same epistemic status as those generated by those who’ve looked hard at what’s going on.
review: On the one hand, there is case to be made for expertise – that within their fields, be it physics, biology and so on, experts should be accorded authority. But on the other hand, politicians and policymakers have tended to use expertise, beyond its field of application, to justify certain policy decisions. Policymakers have almost leeched off the authority of science to lend their own decisions a certain authority. So is there a problem with the political use of science?
Collins: Yes, there is a problem. But first you have to ask two separate questions. The first is whether you can ever justify expertise used with integrity, and we’re arguing, yes, you can. And the second is to ask whether expertise is being used with integrity. And we see a lot of the time that it’s not used with integrity. And we don’t think scientific expertise used without integrity is a good thing.
Now, in Why Democracies Need Science, we have a suggested solution to this problem, which is the setting up of a new kind of committee. It’s a committee that more or less fills the current role of the government’s chief scientific adviser, or something of that sort. But this new committee would be made up of natural scientists and social scientists. And these would be natural and social scientists of a special kind. We call them the ‘owls’, because owls are birds that can turn their heads round 180 degrees. In other words, there would be natural scientists who have some comprehension of the social aspect of science. And there would be social scientists who know to an extent how science works. And what this committee presents on any given scientific controversy to the policymakers is, firstly, today’s scientific consensus, and secondly, the grade (A to E) of that consensus. So if the consensus is in physics, the committee would say this is the consensus of the mainstream scientists, taking out fringe scientists and so on, and it’s a grade A consensus. Whereas if it was economics, you would say that this is the consensus, and while there might be a pretty strong consensus among economists – think of the monetarists of the Thatcher and Reagan era – we’d say it’s only a grade E consensus. That is, you can’t rely on it.
The rule would be that policy, in the last resort, should be a matter of politics, and the policymakers should therefore be able to make any policies they like, but they can’t distort the nature of the scientific consensus to support those policies.
To give you an example from a few years ago, in South Africa, then president Thabo Mbeki decided not to distribute antiretroviral drugs to pregnant mothers. And he justified it on the grounds that they might be dangerous, telling his upper house to look on the internet for this great controversy about the safety and efficacy of antiretroviral drugs. And if you looked online, you would indeed find that there was a controversy about the safety and efficacy of antiretroviral drugs. However, if you went to the scientific community, its members would say, yes, there’s a debate on the internet, but within the scientific community there’s no debate at all. All those arguing against antiretroviral drugs are past their sell-by-date. So the scientific consensus is that there’s no problem with antiretroviral drugs.
Now, what Mbeki had done was to disempower the electorate, because he told them that there is a scientific problem with antiretroviral drugs when there wasn’t. He would have been perfectly entitled to say there isn’t a problem with antiretroviral drugs, but we’re not going to distribute them in South Africa because they’re too expensive, we’re better of spending our money on other things, and we don’t want to give the impression that South Africa is some sort of promiscuous, disease-ridden country. And, if he’d done that, he would have made a political decision and the electorate could have voted him in or out on the basis of that decision. But as it was, they were disempowered. If you had the owls saying what the consensus is, that would stop Mbeki doing things like that.
And the other side of the coin, as we saw when Ronald Reagan said that we’ve got to adopt these financial policies because there’s a consensus among economists to suggest this is the best way to go, the owls would have said that, yeah, but that consensus isn’t worth anything. You’ve got to ‘fess up and say, this is a political decision, not a scientific one. It would have stopped various politicians defending their policies on the basis that it was inevitable within the economic consensus. So that’s what we think the owls would do, and we hope it would stop expertise from being misused in the way, as you say, it often is.
review: But are you still not demanding too much of expertise, of the facts? Experts can say how things are, they can certainly say, for example, that vaccines do or don’t work. But, as you say, politics is the last resort here, and politics is not about how things are so much as how they ought to be, about how we ought to live. And science, experts or no experts, can’t tell us that.
Collins: No, science can’t answer that question. But the point is that when the public are being presented with a political choice, they should know what the scientific consensus is, or how the scientific consensus should affect that choice. It could still be overruled – if it couldn’t be overruled, you’d have a technocracy, a world in which policy is made by experts. And we don’t want that. We want policies made by politicians and by people. But politicians must not be allowed to distort or misrepresent the scientific consensus when they’re making their political decisions.
review: Let’s take the most obvious and controversial use of science and expertise to authorise policies, namely climate science. We’ve been told that there is a consensus. But to what extent do you think it has been used to justify policies or a policymaking trajectory that is not supported by the science?
Collins: Okay, yes, let’s look at the climate science debate. First of all, there is the question of whether global warming is caused by human activity – in other words, anthropogenic climate change. And it seems to me that there is now a pretty strong consensus that global warming is caused by human activity.
And then there’s the second question: what do we do about it? Things are much less clear here. Should we go for the immediate cessation of CO2 production and global-warming gas emissions, or should we go for long-term amelioration. Or should we, in fact, do nothing. And it’s a very difficult question. But that’s where the political debate should be. The political debate shouldn’t be about whether there is anthropogenic global warming; it should be about what we do, or don’t do, about it. The political debate shouldn’t be distorting the scientific findings.
The worrying thing is the post-truth approach to these matters, which involves making up anything you like, and surely that’s what we don’t want.
review: Are there not good reasons for distrusting the use of experts and science, even when their expertise is being used with integrity? Few economists seemed to anticipate the financial crisis in 2008 for instance.
Collins: Yes, that’s true. Mainstream economics is in a mess. It’s not the greatest of sciences, and there are extraneous reasons for this. In the UK, for instance, the quasi-marketisation of the universities means that departments are supposed to be judged on various measures, including the research assessment exercise, the research excellence framework and the number of papers published in the top journals and so on. Which is not necessarily a bad thing, but it can be misused, and I think that has happened in economics, because the gold standard for economics is to get published in one or two top journals, and those journals demand mathematical brilliance, rather than an accurate discussion of where the economy is taking us.
The internal criteria for judging excellence in economics, therefore, have very little to do with accurately describing the world. So there are massive problems with economics, and people are right to be suspicious of economists, but that doesn’t mean we should throw economics out altogether. We shouldn’t be rejecting experts; we should demanding they raise their game.
review: The case of economics is indicative of a broader issue in that a consensus on the facts of the matter might not be produced simply by a rational agreement among a group of expert practitioners. There are other extraneous factors involved: funding imperatives, the admissions criteria for academic journals and so on. So to what extent is the generation of a scientific consensus informed by extra-scientific factors?
Collins: Well that’s exactly why this committee, which we call the owls, will include social scientists to investigate and talk about those kinds of things. So it’s well established now in science studies that scientific knowledge coming from low-status sources is undervalued by scientists. There have been a lot of very good studies of this that show, for instance, that people who actually work with organophosphates, spraying them on crops, have long been saying that we can’t spray them under the safe circumstances that scientists describe. Those regulations are always violated. So organophosphates are not safe.
Now these people who are practically involved in the spraying of the organophosphates are experts. They’re experts on organophosphates. But they’re completely unqualified in terms of scientific, academic qualifications, and therefore their input tends to be undervalued. So you need the social scientists in there, saying, ‘No, you’ve got to value these people’s expertise just as much the expertise conferred on those with university educations’.
And vice versa. Let’s go back to Mbeki in South Africa, for example. The few people who were saying that the antiretroviral drugs weren’t safe had Nobel Prizes and a long list of academic titles. Nevertheless, their views were past their sell-by date. High-status expertise does become obsolete.
So it’s up to the social scientists to ensure that the application of scientific expertise, and the generation of scientific knowledge, is as little affected by these extraneous, unwanted social forces as possible.
Harry Collins is a sociologist of science at the School of Social Sciences, Cardiff University, and a fellow of the British Academy. He is the author most recently of Why Democracies Need Science, published by Polity Press. (Buy this book from Amazon(UK).)
Picture by: Imaginary Futures, published under a creative commons licence