Complexity's conjuror

The dust-jacket of this little volume is a bit of a shock for the knowledgeable reader. There, prominently, under the unmistakable portrait of Paul Dirac, is the name Pierre Gilles de Gennes. Then one notices the title, Soft Interfaces, which is certainly characteristic not of Dirac but of de Gennes. Finally, in much less noticeable type, "The Dirac Memorial Lectures" can be made out and one arrives at the appropriate connection between the two. On the flap one finds a perfectly good portrait of de Gennes, with an only slightly francophone thumbnail biography. So, in fact, it is not a very serious publisher's boo-boo. Nevertheless, it is as though an honour named after Paderewski were to be given to Earl "Fatha" Hines. It is not that Hines is not a great musician, but that he has something different in mind when he plays. Dirac and de Gennes earned well-deserved Nobel prizes 60 years apart, but that is about it in terms of commonality.

Dirac noted that he had provided a theory underlying "all of chemistry and much of physics", and thereafter refrained from the workaday business of applying it, as have his successors on Silver Street in the "DAMTP" at Cambridge, and now this unfortunate tradition seems to have captured Cambridge's Newton Institute too.

De Gennes, in his public persona at least, takes a very opposite view: "Compared with the giants of quantum physics, soft-matter theorists look like the dwarfs of German folk tales. These dwarfs were often miners: we, also, are strongly motivated by industrial purposes. We see fundamental problems emerging from practical questions."

This is a very interesting role reversal: Anglo-Saxon pragmatism in the hands of the quintessential Frenchman, as opposed to the fascination of the English school with deep, abstract, philosophical questions. Of course, Cambridge has its empirical, pragmatic physics, in the shape of the Cavendish Laboratory, especially under the direction of Sir Sam Edwards; but the Cavendish and the Newton Institute crowd, to my knowledge, have never found it easy to communicate.

De Gennes is a tall, handsome Frenchman whose aristocratic forbears supplied the "de", with a very Gallic manner and lifestyle, and with more than enough charisma to have gathered around himself groups of unusually able associates in each of the many fields he has transformed by his presence. In his native France he is even something of a television personality.

It may be because of all his personal glamour and charisma that he has been able deliberately to opt for successively less glamorous and more "chemical" and useful branches of physics. His (recently fashionable) early work on magnetism was followed by the "Orsay Superconductor Group", which hoovered up all of the problems involving complexities of alloys, surfaces, and so on, left open by the breakthroughs in fundamental superconductivity theory of the late 1950s and early1960s. This group had very much a "sixties" culture, publishing as a collective without separate attribution to its members. The net result, unintended I am sure, was that de Gennes tended to receive the lion's share of the credit earned; one wonders, to this day, whether some careers were not slighted along the way.

In the late 1960s de Gennes segued into liquid crystals, a then unfashionable field to which he attracted his usual group of collaborators as well as some exciting foreigners such as Phil Pincus, Bob Meyer and Bill MacMillan.

From liquid crystals he went on (with Sam Edwards) to the effort of retrieving polymers for theoretical physics, after a long hiatus of interest in that subject. From liquid crystals and polymers it was a natural step to take on the whole broad spectrum of the properties of molecularly complex matter: adhesion, wetting, gellation, lubrication, fracture, granular matter and so on. Hence the broad field of "soft matter physics" that de Gennes named, made his own, and has been immensely effective in making intellectually respectable among a younger generation of theorists. He is one of the great architects of the "opening towards complexity", which, to my mind, is the chief hope for rescuing physics from the dead-end represented by John Horgan's "ironic, post-empirical science".

As a missile in the continuing battle for the soul of theoretical physics the present volume seems at first very slight. The scientific content consists of four short chapters of about 100 pages, with some 40 figures and 80 equations, not a single one of which contains so much as an integral sign. The chapters are on "wetting", "slippage", "adhesion" and "polymer welding". What is remarkable is the style of treating the more than a dozen very complicated and interesting problems these subjects present. Why does a droplet sometimes wet a solid in layers, like an Aztec pyramid? How does the strength of adhesives - suggested, in a brief historical note, to have been the source of the successes of the Phoenician navy - greatly exceed the simple chemical binding of one layer to the next? And so on. De Gennes's method is invariably to use the tricks of perspicuous visualisation, estimation of scales and of variation with key parameters, and, above all, to clear away all irrelevancies. It is a successful scheme and style, and the piling up of one example after another is enormously impressive.

Thus this volume will be of interest not just to materials scientists and chemical engineers. Condensed-matter physicists as well as the intellectually curious of almost any stripe who are not put off by its conciseness and understated manner should find it instructive to see how one brilliant intellect deals with complexity in nature.

It is especially interesting that De Gennes seems to avoid any expression of general principles that might inform his approach. He is the opposite of such complexologists as Per Bak or the Santa Fe school, who tend to write books elevating a theory of one or a few complex phenomena into a general scheme for "how the world works". De Gennes has on more than one occasion been close to a broad generalisation but has seemed to shy away from expounding it. He provided, for instance, the impetus for the discovery of the general topological theory of defects in ordered systems, but left its actual formulation to Toulouse, Kleman, Volovik and Mineev, and the connection to dissipation and broken symmetry to Toulouse and myself. Again, he was the first to hint at the "n=0" or "replica" trick in disordered systems, but carried that brief paper no further, leaving the field to Edwards and others. If I had to criticise this small book in any way, it would be for its lack of hints at any broad vision that informs the magic de Gennes brings to bear. But to see such legerdemain performed so deftly is very much worth the price of entry. We should all be envious.

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