They think it's all over

In a review in The New York Times of Bob Woodward's latest book about American politics, I found the following: "Mr Woodward's victims know they are faced with a choice I either they can refuse to cooperate, in which case they will be described I as they were described by their enemies I or they can cave in and tell Mr Woodward their version I Either way I they know that they are at his mercy because I the book is expected to be a bestseller before it is written."

John Horgan's book represents the arrival of such political reporting styles on the serious scientific scene - with the difference that we interviewees did not realise we were at his mercy until too late, having been pampered by respectful reporters lacking an agenda or platform of their own, hence obligingly furthering ours. Much of the fury this book has aroused in the scientific community is caused by the resulting sudden irruption of our sometimes imperfect personalities, our mannerisms, our casual backbiting, our unguarded boasts, and the like, into the public eye. Insofar as the scientists have eagerly angled for public notice, and in many cases for our own bestseller status, one might feel we have reaped a whirlwind of our own sowing; nonetheless where our remarks are shoehorned into favouring Horgan's personal agenda, we have a legitimate complaint.

Horgan, a staff writer for Scientific American, does occasionally catch our idiosyncrasies remarkably well. I laughed aloud at his interview with Tom Kuhn, taken back 50 years by his description of phrasing and manner to wartime Harvard (where Tom was junior marshal of Phi Beta Kappa and editor of the Crimson, and I a green scholarship kid from the Midwest). That said, often Horgan goes too far to put his victims at a disadvantage. It seems completely irrelevant to his thesis that the late Karl Popper's housekeeper believed erroneously that he was famous even among the local London cabbies.

The book's thesis is that science is the victim of its own success: that all the really interesting and soluble problems have been solved, or will be in a few years, so there is nothing for us to do but either accept our role as footnote writers to scientific history or to adopt an "ironic, postempirical" style. These are two useful terms adapted from the humanities: "ironic" from literature, the mode of echoing, reinterpreting or even mocking the Great Masters one cannot hope to surpass; and "postempirical", the mode of theorising about matters one cannot hope to approach experimentally: in both cases the enterprise may well be creative and exciting, but shares with philosophy and literature the possibility of endless reworking and reinterpretation - exactly the opposite of the clean, falsifiable truth which science claims to seek. The great masters to which Horgan refers in the case of science are the pioneers of quantum mechanics, relativity (which he broadens to include the hot big bang cosmology) and Darwinian evolution. Indeed, it is not possible to think intelligently about modern science without "playing off" from these great ideas, which inform everything we do - this is also true of the calculus, logic, the real numbers. Science intrinsically goes on from, rather than ironically negates, the great discoveries of the past.

There seem to be three more-or- less cogent arguments for the thesis Horgan advances, as well as some not so cogent. The first is in the words of the scientists themselves: the unfortunate propensity of senior scientists to declare that their field is now a wasteland of solved problems. Jim Phillips, who is often psychologically astute, remarked long ago of a column to this effect published in about 1959 by John Bardeen (coinventor of the transistor): "Of course, he has solved all the problems that were problems when he was young." Notorious in this mode was Brian Pippard's essay, "The cat and the cream", averring that academic research in condensed matter physics was futile, just at the time that two Nobel prizes were hatching in his own department, one for his boss, Neville Mott, and one for his student, Brian Josephson. Horgan gets considerable mileage out of a correspondingly depressed book about biology by Gunther Stent called The Coming of the Golden Age and of a symposium organised by Stent titled "The End of Science?" Second is Horgan's very clever use of the philosophical naivety of some of his subjects - he even remarks on it in the case of Ed Witten. These are people who really believe that the goal of science is the complete answer to some all-encompassing question; people who, like Witten and others in physics, Stephen Hawking in cosmology or Richard Dawkins (and possibly even Stuart Kauffman) in biology, believe there is some unifying, underlying principle or law whose consequences then need only to be explicated in order to understand everything. The position is called "naive reductionism" and is typified by Witten's curious remark, that "every exciting discovery in physics follows from string theory".

This naivety is reinforced by great reputations fed by the enormous appetite of the general public for speculative musings on the "mind of God", the ultimate theory of everything, or the fate of the universe, and seems to be shared by Horgan himself, who, let's face it, is not visibly deep scientifically. It is a regrettable fact that the same naivety is shared by generation after generation of able students who follow these pied pipers into the morasses of string theory, cosmology and other postempirical subjects. Horgan is, in a way, quite right in his description of this kind of work as ironic and postempirical, but wrong in seeing it as the essence of science; science itself is still an empirical subject. Kuhn's "normal science", in my mind, can be described as a search for answers, great science as a search for questions, the greatest science as a search for the form the answers may take. These last two types of search are sometimes hard to distinguish from postempirical science until someone has invented the apparatus or the type of argument necessary to check them out, but the scientific community itself is often able to tell the difference. One good test (which several of Horgan's interviewees fail to pass) is whether the scientist involved has any history of dealing successfully with empirical facts.

The least plausible of Horgan's three arguments is that further progress in some fields, notably neurophysiology and the social sciences, is blocked by the sheer complexity and difficulty of the subject. He trots out, for instance, old and new speculations that he calls "mysterian" (not his neologism), to the effect that the human mind is incapable of understanding itself, but without giving any cogent argument except the opinions of various of his postempirical interviewees.

If we were to accept that all still-open questions are either too hard to solve or trivial, it would indeed appear we are approaching the end of science. But do we? Let us imagine ourselves back in the 18th century, contemplating the work of Newton and Descartes. If we were Horgan-minded, we should be in great despair: the continuum of space-time had been revealed and all the mathematics for discussing it had been discovered, in the shape of the calculus. Nobody has yet invented a way of expressing truths about space that does not use Newton's boring old calculus. All the rest is trivial. Well, the same role is played by quantum mechanics. Once one realises that it is necessary to use the symmetries of space and time to understand physics, one also realises that the quantum theory is probably unique as the mathematics we would have had to invent to express our physical ideas, and that we should be in desperate straits - reduced to the kind of stratagem proposed by David Bohm - if we did not have it. But the physics expressed during the past 75 years by this simple mathematical framework contains many new theoretical insights and many exciting and enlightening empirical discoveries that are not in any sense just a working-out of the consequences of quantum mechanics. This is true obviously when we consider the particle side, where such new ideas as the gauge principle that interactions follow from symmetries, and broken symmetry, are by no means mere consequences of quantum theory. What Horgan is missing in his simplified, layman's view of science is that "the devil is in the details". Darwinism does not solve all the problems of biology any more than the quantum theory solved chemistry or relativity solved cosmology. These theories are tools, just as mathematics, the telescope or statistical mechanics are tools. We would be crazy to try to be astronomers without a telescope or biologists without Darwinian concepts; each is a liberation, not a straitjacket.

This, then, is the general problem with the book. I shall pick out four instances of it, where Horgan nods, apparently omnisciently. First, he accepts the word of Linus Pauling that "I solved all of chemistry except maybe the sulfides" by 1931 (using, Pauling could have mentioned, the method of Heitler and London). But he does not mention that Pauling was misguided about metals (I cut my scientific teeth on his paper); and that the quantum chemistry of metals - the solved part, that is - was understood by another group of people entirely, including Eugene Wigner.

Horgan says that everything about cosmology is solved, the rest is dull. It hardly takes a sharp eye for instructive anomalies to realise that the two defects he finds in cosmological theory, the Hubble constant discrepancy and the fractal distribution of visible matter, are precisely the kind of anomaly that signals and informs the next step in science.

His interview with Per Bak reveals a similar failure of comprehension of science in his seemingly deliberate misunderstanding that a distribution of probabilities is one of the classic forms a theoretical result can take.

And his quote from Naomi Oreskes to the effect that numerical models are not "verifiable", which he applies by implication to the whole fabric of science, is a misunderstanding of the origin and thrust of Oreskes's thinking and of the nature of science. This essentially solipsistic argument (dating back to Bishop Berkeley) would deprive us of any "verifiable" knowledge of an external world whatever, even through our own senses. No sane person really accepts such solipsism. But worse is Horgan's implicit assumption that the goal of science is the Cartesian computer, a numerical prediction of everything by detailed calculation. Science (see above) is the search for understanding, and only in a fraction of cases do we have any need or desire (or capability) for a detailed numerical prediction. "Why?" and "How can I find out?" are at least as characteristically scientific questions as "How much?".

Finally, I cannot close without strongly objecting to Horgan's allegation that the some of the Santa Fe Institute's ideas derive from those of Ilya Prigogine, a statement that persuades me that Horgan's grasp of concept and of detail is just not up to the job he has set himself. Never mind the originality or the validity of any of Prigogine's ideas; they have nothing to do with concepts such as rugged landscapes, adaptive agents, complex adaptivity or self-organised criticality, which are current at SFI. The devil is in the details again. Horgan describes the serious reservations many of us have about Prigogine's ideas in one place, and in another accuses SFI of borrowing them without credit.

Despite all this I, like most of the reviewers I have seen, cannot consign The End of Science to the dustbin. Horgan's book is a good read, and we scientists very much need to have a few of the popular illusions about us dispelled. Some of us, few it is to be hoped, are cranks or charlatans, others may be visionaries, pointing the way to future; but you can never be quite sure who are the visionaries and who are bewildering themselves and misleading us. Too many of those who come to represent us to the public are mere publicists, with little record of empirical or mathematical achievement. Horgan could, in a better book, have exposed some of these pretensions to good effect without writing what, on rereading, leaves a lasting impression of deliberate destructiveness. If this review sounds somewhat harsh, it is because, in spite of all the author's protestations to the contrary, it seems to me he has most mischievously provided ammunition for the wave of antiscientism we are experiencing.

Philip W. Anderson, a Nobel laureate in physics, is professor of physics, Princeton University.