Snowballing, but still very much in control

'Post-academic' science is characterised by accountability, regulation, contracts - and, most worryingly, lack of autonomy, regrets Philip W. Anderson

We have learnt how to investigate phenomena in ways unimaginable 50 years ago - not just with particle colliders and space telescopes, but with microscopes that see and manipulate individual atoms, scanners that can watch the neurons "thinking", gene-sequencing machines, incredibly sophisticated analyses of rocks - you name it. None of this comes free: the tyros who used to assume that they needed only a machine shop or equivalent to build their own apparatus now are asking for millions of dollars in "set-up funds", or are forced to join collaborations in which they may feel little more independent than assembly-line workers.

As John Ziman points out, one cannot expect society to support an enterprise on this scale, growing at this rate, indefinitely. It can no longer be supported as, basically, a byproduct of higher education and an object of philanthropy. This fact causes another less well-documented change - the one that impinges most heavily on scientists' working lives. While society recognises that it has benefited enormously from the growth of science in economic terms and in military strength, greatly improved medicine and also in less tangible ways from television spectaculars to national prestige, this leads to the assumption that these benefits should be channelled in directions chosen by those who pay the bill. This change is happening in two ways (at least): by an increasing bureaucratisation of publicly funded science, and by what Ziman calls an "industrialisation" of much of science. There is a growing tendency towards making government support conditional on attracting industrial cooperation. As Ziman also points out, the older industrial research laboratories, inhabited by scientists working within the academic mode, are being replaced by cooperative academic-industrial teams, with the academic scientists often co-opted into industrial goals or being themselves entrepreneurs.

Ziman began his career as a solid-state theoretical physicist with a well-regarded book, Electrons and Phonons . In mid-life he segued smoothly into what has become a distinguished career as a sociologist of science. His early books in this field, Public Knowledge (1968) and Reliable Knowledge (1978), were where many of us learnt what he here calls the "Mertonian" view of science (named after Robert Merton, the father of science studies) and the role of sociological norms in making science as "reliable" as it is.

With Real Science he continues and updates his examination of the nature of science and technology as social phenomena. He is very ambitious: he claims to reconcile the two sides of the "science wars", the quarrel between a vocal subset of sociologists and other "science studies" types versus all of the scientists who are aware of what those guys are saying about us. Ziman hopes as well to update and correct his earlier picture ("model") of the sociology of science.

The book begins by restating the view of science that Ziman ascribes to Merton. Ziman calls it the Legend (always capitalised) and designates it with the acronym Cudos. The letters stand for the five ideals of the scientist: communication and communality, universality, disinterestedness, originality and scepticism. He expounds what I call the "neoclassical" view as distinct from Merton's Legend. This view was described in Ziman's early books as well as by others such as Mike Mulkay. It is that real scientists are not individually disinterested in the results of their investigations or particularly sceptical - but that the above ideals are enforced by a system that rewards compliance with the Legend. The production of original and demonstrably reliable knowledge leads to publications, tenure and recognition by the elaborate system of honours that the profession has constructed for itself. Although Ziman makes this further point only in passing, the strongest argument for science as a unique social construction is its extraordinary rate of growth measured in almost any terms: people, pages of publication, influence in the world, budgets, and so on. What grows is not only the total amount of reliable scientific knowledge but also its rate of accumulation. In this it is unlike any of the comparable professions: law, religion, even business, except insofar as business exploits the technology that grows out of scientific knowledge.

Ziman makes it clear that the Legend and indeed everything he discusses applies in full force only to the subset of the subject that he calls "academic science": institutes, government installations and even some industries, all of which maintain extensive groups of scientists that are behaviourally indistinguishable from groups of academic scientists in universities and that indeed often exchange personnel with academe. Quite so, but it is academic science that delineates an image for science as a whole.

Next Ziman devotes a chapter to the ways in which the system that has had these enormous successes is under extreme strain - or perhaps has even broken down. In essence, these ways are mostly traceable to science's very success. The first is the cost and the scale of much modern scientific research. More subtle but equally important is the change in the organisation of science that I describe above.

The results of these changes, Ziman says, are so far reaching as to constitute a new mode of knowledge production that he names "post-academic science", characterised by such words as contracts, regulation, accountability. This is a good coinage, but the word he does not use, or rather uses only by implication, and that more than any other strikes me as characteristic of the changed face of science, is absence of autonomy : the academic scientist had autonomy, the post-academic has much less. "What is remarkable," Ziman writes, "is that academic science survived so long as a distinctive culture... It would be even more remarkable if the historical clock were to be turned back," because in a way, he says, the post-academic scientist is more in keeping with "modern" modes of organisation.

In the next five chapters, Ziman asks what is happening at present to each of the five Cudos ideals of science. Throughout he keeps in mind the question of whether science is still "reliable knowledge". I found these chapters rather discursive and, occasionally, questionable. For instance, he seems to find the communication system of science little changed, or at least he spends much space discussing why the old model was reliable. (As he points out, this is the core of science: science may be thought of as a system, with unique characteristics of communicating, recording, and vetting information.) In my experience, the most visible outcome of the structural change in science has been a tendency for the traditional processes of peer review and refereeing to disintegrate. Specialised journals of low quality have proliferated, publishing essentially anything they receive, while the few prestigious journals have become overworked, relying too much on professional editors who are not scientists and have to choose the peer referees either arbitrarily, by computer, or using criteria based on editorial convenience. The result is reports that are arbitrary, irrelevant or venally competitive. Many fields have abandoned the conventional journals for scientific communication, relying instead on unrefereed web-based "archives". The usual journals and the international meetings serve primarily as window dressing for the decoration of CVs. Still, the system works after a fashion and in the long run. But this book does not tell me why.

In the next chapter, on universality, I found much to agree with and some to disagree with, but little in the way of a central argument. The argument would appear to be that, along with other changes, science has become so esoteric that our so-called universal knowledge is fully comprehensible only to scientists themselves, and indeed only to subcultures of scientists. The idea is neither new nor completely true. It is true that anyone who has tried to teach first-year physics knows that the population contains two species, those who can get rid of their "folk (or Aristotelian) physics" intuitions, and those who cannot. (Of the latter, some seem to become philosophers, critical of science, in revenge.) However, once barriers of this kind have been overcome, and an individual has become sufficiently open-minded to be comfortable with the abstractions and the generalisations of science, personally I have not found him or her to be "epistemically" (Ziman's favourite word) blind to other scientific ways of thinking. At the Santa Fe Institute, for instance, good scientists, in the physical and human sciences, welcome ideas from other fields of science. Of course, there are entrenched establishments in every field, but the nature of science is such that these establishments are born to be overturned by the next discovery.

The ideal of disinterestedness and objectivity is the main point of attack on science in the "science wars" and the point at which post-academic science seems most unreliable on the face of it. But, as Ziman points out, "the (direct) effects of social interests on knowledge production are so weak and self-cancelling that they mostly get lost in the noise". In other words, it is still hard directly to bias scientific results on one side or the other of socioeconomic questions.

There are, however, two ways in which the public perception of a scientific issue can be skewed. First, as in (Ziman's example) the early studies of "mad cow disease" or in the perennial battle over missile defence, a government can choose to listen to one side of a technical debate and not the other. Second, and more serious, the agenda will tend to be set within the concerns of the dominant culture of the time, such as studies of the diseases of white males rather than of women or blacks. Both of these are ways in which, as Ziman points out, post-academic science is in principle - but, mysteriously, not necessarily in practice - less reliable than its predecessor. He mentions that it is well within the post-academic era and within academic science that the concept of "public interest science", carried on by such organisations as the Federation of American Scientists in the US, has grown up. The barricades against the misuse of science are not in fact manned by the deconstructionists but by academic scientists themselves.

The book's chapter on originality is full of insights into the way scientists think, such as the remark that scientists tend to follow Bayesian probability ideas intuitively, opting always for the simplest possible explanation and showing surprisingly high scepticism about standard statistical tests (see my own note in Physics Today , January 1992). And then, rather surprisingly, there is a lengthy passage that supports the claim to reconcile "constructivists" who believe all of science to be a social construct with the scientists themselves, "objective realists" to a man (or woman). In typical faculty-meeting fashion, Ziman offers eloquent arguments in favour of constructivism first, and then, with a firm "notwithstanding", demolishes them with the simple argument that our ordinary sense perceptions are no less "constructed" by our brains than such esoteric theoretical entities as quarks. I do not believe that Ziman's proposed reconciliation will take place.

The ideal of scepticism provides the framework for meditations on scientific mores, for instance on the relatively rare instances of fraud, as well as the common cases of self-deception. Academic science has never been seriously threatened by either in the long run, and that position has not much changed. It is in this ideal of scepticism that science differs from the pseudo-sciences such as parapsychology. Appended to this chapter is a discussion of an evolutionary model of science and technology. Even for the latter, I find it not totally satisfactory, while for science it seems quite inappropriate. Does Ziman really believe that the forms of the Standard Model, the genetic code or the Hodgkin-Huxley spike are historically determined by an evolutionary process rather than discovered? He places great emphasis on the difference between the concepts of "discovery", which is the classical process in science, as opposed to "construction", the postmodernists' concept. Both undoubtedly are part of science, but most scientists would agree that there is a solid core of the former in most subjects.

Finally, we come to the summary chapter,"What then are we to believe?", which is intended to bring all the above together. It is not clear to me that it does. Many pages are devoted to a discussion, so general as to belong in some philosophy textboook, of modes of perception and of cultural diversity. But when it comes to comparing science with other belief systems, such as religions, he bends over backwards so far as to get quite out of balance - claiming, for instance, that there is no more difference between animists and Quakers than between different scientists such as quantum cosmologists and experimental psychologists. Yet soon comes the "notwithstanding", after which he devastatingly shows that science is simply the extension of our sense perceptions. In the end, he lamely defends the "postmodern critique" by saying it is "strictly non-metaphysical". What does that mean: that it is physical? What is it, then? Just nonsense, as it appears to most scientists?

He affirms the neoclassical model of his earlier books. "At the coal face, science is not noticeably going postmodern" - though he warns against a likely decrease in creativity and a weakening of moral integrity. From the point of view of someone at the "coal face", I can only say amen, as I see my most creative students depart for Wall Street in disgust at the bureaucratic academic hiring scene and rigid funding guidelines. I see the "post-academic" mode as firmly in place in my field, and Ziman is to be praised for identifying the syndrome.

One peripheral cavil. The book has more than 1,000 numbered notes referring to 356 books and articles. To find any one of these references one needs to look in three separate places, and when you have done so, you find no explanation of its relevance or whether or not it is cited with approval. Many are simply cited as sources for words used in a commonly accepted sense, while others describe ideas I find perverse, yet both have no explanation. I am an inveterate footnote reader, and I confess to deep frustration. Perhaps less peripheral is that in spite of the massive bibliography, some seminal references are not there. For example, some of the discussion of the nature of scientific theories could have used ideas well expressed by Murray Gell-Mann in the excellent Santa Fe book Complexity (edited by George Cowan and David Pines, 1995). Section 10.8 is in large part based on my ideas in a quite widely referenced paper of 30 years ago. I get the impression that Ziman's coverage is enormous, but selective as to nationality, or perhaps as to scientific community.

In sum, this is an important book about the practice of science from someone who has been there. It is at times exasperating, but it must be read by anyone who wants to understand modern science as it is practised, because it brings together all aspects of the changed social system of science, and even gives it a name: "post-academic".

Philip W. Anderson, Nobel laureate, is professor of physics, Princeton University, New Jersey, United States.