Friendly fire lights up the lab

Jon Turney doesn't need his flak jacket for this polite salvo in the science wars.

Science is under siege, or so it seems to many scientists. There they are, trying to save the planet, relieve hunger and cure disease - or just tease out some interesting detail about how the universe works. The response is grudging approval at best, open hostility at worst. Who are these people, alarming us about global warming, telling us to inject our children with dodgy vaccines, messing about cloning sheep or pigs?

Social-research evidence tells a different story from newspaper headlines, though. Most people approve of most of science, most of the time. And scientists remain much higher in public esteem than, say, politicians or journalists. But it is still fair to say that science no longer commands the cultural authority it did. The reasons are many, varied and complex. But a few academic scientists have picked on a simple one: fellow academics have failed to accord science the respect it deserves.

Worse, some of those academics have studied science closely, and claim expertise - not in technical matters but in how science actually works. The result has been an ill-tempered and unproductive academic dispute. In fact, as newspapers' occasional references to the "science wars" suggest, it has hardly been an academic dispute at all. There has been more name calling and less attention to what people say than proper academic argument requires. The casual observer would simply hear people talking across each other, one camp accusing the other of trying to tear down the great edifice of science, the other avowedly bemused by the paranoia and suspicion of a group that still commands awesome cultural and material resources - even if US taxpayers did eventually decide not to give some of them a multibillion-dollar superconducting supercollider to play with.

This carefully discursive volume aims to put that right, and largely succeeds. Sociologist of science Harry Collins and sympathetic chemist Jay Labinger have their contributors offer opening statements, replies and then comments on the replies. Everyone is very polite to everyone else. The only remotely wounding remark that makes it into print is sociologist Trevor Pinch's suggestion that Alan Sokal and Jean Bricmont are "the noisiest guests at the party". As Sokal hoodwinked the journal Social Text into publishing his celebrated spoof article on cultural study of theoretical physics, and then collaborated with Bricmont in a swingeing attack on a whole slew of postmodernists in Intellectual Impostures , this is probably fair comment. But it is out of keeping with the exchanges here, which are confined mainly to analysis of what people have said in other contributions to the book, not outside it.

So does taking the heat out of the science wars lead to illuminating conversation? Within the limits set here, yes, on the whole. Those limits are set largely by the choice of contributors. The book is basically a series of exchanges between physical scientists and sociologists of scientific knowledge, most of whom study physics. There is much more to science and to science studies than this, as various contributors acknowledge. But this degree of common interest was probably needed to get into close enough contact to clear away misunderstandings that impeded real discussion.

These can be confusions about outlook and motive. As Steven Shapin says:

"It is sometimes hard for scientists to understand how the description and interpretation of science could be anything other than coded prescription or evaluation: telling scientists what to do, or sorting out good from bad science." This, after all, is usually what scientists are about when they describe or interpret science. But science studies strives for neutrality. Other misunderstandings may turn on single words. Sociologists regard "negotiation" as a neutral descriptive term. Scientists tend to read it as denoting that something fishy is going on. Or take physicist David Mermin's assumption that "knowledge" means collective knowledge, whereas Collins tends to mean the knowledge available to a single individual. As Mermin relates, this made it much harder for them to discuss how knowledge becomes established when they were debating the early history of relativity theory.

Overall, the essays and commentaries here clarify what bothers scientists about the way sociology of scientific knowledge reads their work, and what they can live with, or even find interesting. There are perhaps three key tenets of recent science studies that cause trouble. One is the philosophical point, due to Duhem and Quine, of the underdetermination of theory by evidence. Logically, you can always dream up any number of theories that fit the data. The scientists here accept the logic, but argue mostly that its significance is greatly overrated. It is often hard to think of any theories that fit the data and are compatible with the existing web of scientific concepts - and researchers invariably experience theory formulation as highly constrained.

The second point is the rather simple one, though endlessly repeated and elaborated in the science-studies literature, that the phenomena never speak for themselves. Someone always has to articulate what nature tells us, in our own local dialect, before it is transformed into universal knowledge. To that extent, at least, all knowledge is socially constructed, though not in a sense that need make natural scientists uncomfortable. What does make them uncomfortable is that exactly who gets to speak on behalf of nature, and under what circumstances, then become sociologically interesting questions.

More troubling, I think - because it comes closer to simple notions of "scientific method" and is a key finding of sociological case studies - is the point that it is, strictly, impossible to repeat an experiment. Less starkly, as there are an unlimited number of variables that might be relevant to the results, there is an inescapable social process of deciding what counts as a replication. And as sociologists are drawn to cases, such as gravity waves or cold fusion, where the facts are incomplete or contradictory, that process usually looks pretty messy, and prey to persuasion as much as rational choice.

But it is possible to see all three points as adding up to a difference of emphasis about how science proceeds, rather than a radical challenge to its status as far and away our best way of finding out about the world. Yes, science studies emphasises contingency where scientists - Steven Weinberg is the most prominent example here - tend to see necessity. Sociologists and historians remain attached to an anti-Whiggish history, which Weinberg thinks misses the point of history of science. He and other scientists here spend a good deal of time taking issue with Thomas Kuhn, while the social scientists claim that his influence is scarcely felt in the field any more (something of an exaggeration) and that the classic Kuhnian question of whether science progresses is a non-issue.

So what is there still to argue about? Rather little, from some points of view. Mermin, initially notably hostile to sociology of scientific knowledge, now finds that "whether the construction of scientific knowledge should be viewed as a process of discovering how nature works or as a process of consensus building among scientists" is a minor issue, because any episode described in one set of terms can be reformulated in the other.

The scientists are still wary about the professionally agnostic attitude the sociologists invariably adopt toward truth, their so-called methodological relativism. Does it necessarily entail full-blown philosophical relativism? The science-studies contingent does not believe so, but not all the scientists are convinced.

But this begins to look more like an academic issue, in the less flattering sense, than fuel for heartfelt public disagreement. There remains, though, a concern for the public image of science. And here there is a puzzle, which this collection highlights but does not solve. The science-studies folks believe it would be healthier for science to be seen as more human, more fallible, the best way of reducing uncertainty in a complex world but not often a way to eliminate it. Some of them believe that lay publics already know as much. Some scientists agree with them. Many, it seems - though not necessarily the ones represented here - do not. But what they think they have to lose is definitely a question for sociologists rather than natural scientists.

Jon Turney is head of science and technology studies, University College London.