It is a pleasure to welcome so many old friends and a few new ones to this beautiful place, kindly lent us by Douglas Schwartz of the School of American Research. We would like to hear your reactions to the proposal we are making for setting up the Santa Fe Institute and to hear your ideas about how to structure it, what kind of intellectual problems it should address, what kinds of arrangements should be made for its governance, and what should be the first steps in establishing it.

It is usually said that ours is an age of specialization, and that is true. But there is a striking phenomenon of convergence in science and scholarship that has been taking place, especially in the forty years since the Second World War, and at an accelerated pace during the last decade. New subjects, highly interdisciplinary in traditional terms, are emerging and represent in many cases the frontier of research. These interdisciplinary subjects do not link together the whole of one traditional discipline with another; particular subfields are joined together to make a new subject. The pattern is a varied one and constantly changing.

In order to discuss a few examples of diverse character, I shall start from subjects close to my own and then move further away. I hope you will forgive me for talking about matters far from my own area of expertise and will correct whatever howlers I make in the course of doing so. Also, I apologize for mentioning in this introduction, for lack of time and space, only some of the emerging syntheses about which we shall hear and only some of the distinguished speakers who will discuss them.

Elementary particle physics and the cosmology of the early universe are the twin pillars on which all the laws of natural science are, in principle, based. These two fundamental subjects have practically merged in the last few years, especially on the theoretical side. In the earliest fraction of a second in the history of the universe, if we look at time running backwards, we go from an easily comprehensible quark soup (a few moments after the beginning) to an earlier era in which the conditions are so extreme that, if we could observe them, they would test our speculative ideas about unifying all the physical forces including gravitation. I should mention that in the last few weeks these ideas of unification have become much more specific. The hope, the very bold speculation, that we might actually find a general theory of all the elementary particles and forces of Nature is encouraged by recent developments in superstring theory.

Many of the mysteries of the universe seem to be tied up with particle physics.

As Frank Wilczek will probably tell us, the mystery of the smallness or vanishing of the cosmological constant, which is the value of the energy density of the vacuum, is intimately connected with particle physics. The mystery of the dark matter in the universe, which must outweigh visible matter by at least a factor of ten, is now believed to be in the domain of particle physics, since much of the dark matter may consist of hypothetical new particles such as photinos or axions.

Meanwhile, the trend toward divorce between physics and the frontier of pure mathematics, which went on for decades after the end of the nineteenth century, has been reversed. The description of elementary particle interactions and the attempts to unify them connect with the central part of pure mathematics, where algebra, analysis, and geometry come together, as in the theories of fiber bundles, or Kac-Moody algebras, and so forth, Frank Wilczek may address that topic, too, and Felix Browder will probably touch on it, as he discusses important parts of mathematics that are applicable to science.

The other examples will be drawn from the study of highly complicated systems. First of all, in the life sciences, a transformation has taken place in recent years that has been so dramatic as to impress itself on everyone, scientists and the general public. Ted Puck will address some central themes in biology and medicine. Of course, a discussion of the revolution produced by advances in molecular biology needs no introduction, but I should like to quote some remarks made last spring by a distinguished worker in that field. A few months ago, the National Academy of Sciences gave a party to celebrate the winning by American citizens of several Nobel Prizes, as well as the Swedish Riksbank Prize in Economic Science awarded in memory of Alfred Nobel. I was invited to speak, along with three other members of the Academy, and I chose the same subject as today’s. In fact, at least three of the four speakers that afternoon had chosen independently to address related subjects. David Baltimore, who preceded me, and Herb Simon, who followed me, both discussed the remarkable trends in science with which we are concerned here, and I shall take the liberty of quoting from David Baltimore’s remarks.

“The first place to start is to look at what’s happened to biology in the last ten years. About ten years ago, the field went through a watershed, because up to then, the precise tools available for dissecting genetic structure and understanding biological organization were really only applicable to microorganisms, which provided useful model systems but couldn’t answer the pressing questions of the organization of systems in human beings and other mammals. “Then, about that time, a variety of new techniques were developed that allowed us to get at the molecular details of higher systems, and overnight what had been seen as impossible became eminently feasible, and the methods which we generally call the recombinant DNA methods changed our whole perspective on what we could think about and do, and that has had many consequences, one of which has been the focus of molecular biology on understanding mammalian systems and specifically as surrogate human systems.

“In ten years, we have seen enormous advances in understanding the immune system, in understanding hormones and their action, in understanding cancer, in understanding evolution, and even in the beginning of an understanding of the nervous system. We have seen tremendous advances in the underlying generalities of how things are organized, how genes are made, how genes are duplicated, how genes are expressed; and a side effect, one that is very significant, has been a striking unification of the kinds of problems that people in biology think about.

“If we go back ten or twenty years, hormones were thought about by physiologists, the nervous system was thought about by specialists in the various branches of the neurosciences, cancer was talked about by oncologists and physicians, evolution was discussed by population biologists, the immune system was studied by immunologists; none of that is true any longer. If you look at the seminar board on our floor at MIT, you will find seminars that cover the range of all of the things I have just talked about as well as plant biology and even the beginnings of behavioral biology, and that is a very different perspective and has tremendous organizational consequences, and actually quite staggering implications for education and for the structure of the field.”

We shall return in a little while to these implications. But let us first look at some other places where crosscutting subjects are appearing, with emphasis on the study of surface complexity arising out of deep simplicity.

Al Scott will tell us about nonlinear systems dynamics, a very exciting branch of mathematics with applications to many parts of science. Nonlinear systems dynamics can be exemplified by first-order differential equations with several dependent variables xi that are functions of time t. The applications are numerous. Peculiar phenomena associated with such equations come up over and over again. In the case of dissipative systems, for example, the orbits (x as a function of time) can be “attracted,” at very large times, to fixed points where the x’s are constant, or to limit cycles where the x’s go around in periodic orbits asymptotically, or to “strange attractors” that give chaotic behavior, so that the x’s at large times become infinitely sensitive to the boundary conditions on x at the initial time Chaos turns determinate systems into effectively indeterminate ones. Attempts are being made to apply these ideas to elementary particle theory, to the fascinating question of the approach to hydrodynamic turbulence, to problems of plasma turbulence, to oscillating reactions in chemistry like the Belousov-Zhabotinsky reaction, to biological clocks, and to many other areas of science.

Biological clocks, for example, seem to be nonlinear systems, each one with a free-running frequency that is usually different from what is actually needed for the clock, but with environmental signals setting the frequency as well as the phase. We are familiar with the resetting of phase in recovering from jet lag, for example. This form of clock seems to provide the kind of robustness that biological systems need. Recently there have been attempts to identify the mathematical phenomena of nonlinear systems dynamics in population biology, in problems of the brain and the mind, in attempted explanations of schizophrenia, and in problems of social systems.

Many of these applications are highly speculative. Furthermore, much of the theoretical work is still at the level of “mathematical metaphor.” But, I think this situation should cause us to respond with enthusiasm to the challenge of trying to turn these metaphorical connections into real scientific explanations. For that purpose, one useful advance would be to know whether these mathematical phenomena really crop up in the solution of partial differential equations, and if so, where.

There are analogous phenomena also for discrete variables, time and x. There, we connect up with fundamental areas of computer science such as Steve Wolfram has studied, including cellular automata and Turing machines. We also encounter new insights into how to construct reliable computers out of unreliable elements. I was associated more than thirty years ago with the first attempts to solve that problem, by methods that were eventually analyzed correctly by Jack Cowan. These days people are talking about more sophisticated methods, based on attractors, for making reliable computers out of unreliable elements. Attractors are almost certainly involved in this way in pattern recognition and perhaps that is true in other kinds of mental activity as well. It may be that attractors are again providing the kind of robustness that biological systems require, this time in connection with phenomena that include human thought. We might even speculate that attractors might be connected with our human habit or getting stuck in a certain way of thinking and finding it extremely difficult to jump out of the rut into another way of thinking. It would be fascinating if that turned out to be so; and understanding the situation a little better may help us to design new ways to stimulate creative thinking.

We all know that computers are not only tools for calculation, but also increasingly for symbolic manipulation, which means they can be used for doing theoretical scientific work. In many cases they can also serve as a kind of theoretical laboratory for experiments in the behavior of model systems. In addition, they are objects of study as complex systems of enormous interest in themselves. Since World War II, in a great deal of interesting theoretical work, they have been compared with neural nets and even with human organizations. A subject embracing portions of linguistics, psychology, neurobiology, neurochemistry, computer science, and so forth, has grown up, that some people call cognitive science. We all know that in most situations, theory has to advance along two tracks: the fundamental search for dynamical explanations on the one hand, and on the other, the phenomenological search for pattern in the laws of Nature. There are associated experimental domains in each case. This is true of the study of the brain, where phenomenological aspects are covered under the rubric of mind and involve the study of behavior, and sometimes, in human beings, even the study of introspection. There is always a reductionist bridge between these two kinds of explanation, the fundamental and the phenomenological. (I assume all of us are in principle reductionists.) But it often takes a very long time to construct such a bridge, such as the one between the brain and the mind, even though great strides are being made. While the construction is going on, it is necessary to pursue both approaches, which means in this case to study both the brain and the mind.

New interdisciplinary subjects are growing not only out of brain science, but also out of mental science, that is to say, psychology and psychiatry. One that I think will be of particularly great interest in the future is the scientific study of human mental processes outside awareness, what is sometimes called the unconscious mind, long dealt with by psychoanalysis, but needing to be incorporated into regular science. Pathways out of the “unconscious” are available, not only in the areas of free association, slips of the tongue, dreams and so on, but also in hypnosis and other altered states of consciousness. Hypnosis, conditioning, and, perhaps, subliminal perception may provide pathways into the “unconscious.” Here, not only mental science is involved, but also physical science. With improved SQUID devices and all sorts of other tools from physical science, one may be able to discriminate by objective means among different states of consciousness, so that when we study them psychologically it will not be a circular process. As progress is made on the brain-mind bridge, the panoply of brain science or cognitive science will also be increasingly applicable. Mardi Horowitz and Jerome Singer will discuss these matters.

There is a striking theoretical resemblance between the process of learning and the process of biological evolution. The field of evolutionary and population biology is one to which sophisticated mathematics has been applied for a long time, with benefit to both biology and mathematics. Much has been learned and much is still not understood. We will have a brief discussion of the state of this extraordinarily important field by Mark Feldman, and we may all reflect on the benefits of future interactions among students of computers, learning, and evolution. Manfred Eigen will tell us about a laboratory system that exhibits evolutionary behavior and may be related to the chemical reactions that produced the first life on Earth; he will thus introduce us to the subject of pre-biotic chemical evolution.

Now cognitive scientists and students of various kinds of evolution are beginning to get together. A new subject is taking shape, which has roots in cognitive science, in nonlinear systems dynamics, and in many parts of the physical, biological, and even the behavioral sciences. Some people call it self-organization, others complex systems theory, others synergetics, and so forth. It tries to attack the interesting question of how complexity arises from the association of simple elements. A conference is being planned at the Center for Nonlinear Studies in Los Alamos on at least part of that new subject—the study of evolution, learning, and games, with emphasis on the theory of adaptation. The conferees will listen to reports on game theory strategies in biological evolution, the coevolution of genotype and phenotype in biological evolution, theoretical and experimental results on chemical or prebiotic evolution, the development of foraging strategies in ant colonies, strategies for the evolution of new algorithms in artificial intelligence (using crossing-over and natural selection in computer programs), models of human learning, the mathematical theory of regeneration in the visual cortex, discoveries on cellular automata and Turing machines, stability of deterrence and stability of the U.S.—U.S.S.R. arms competition, spin-glass models of neural networks, and other diverse topics. Yet the discussion is to be general, with physicists, mathematicians, population biologists, neurophysiologists, social scientists, computer scientists, and engineers all trading questions and comments. Many common threads are already evident, especially the nearly universal importance of adaptation, the need for random inputs in the search process, the importance of high dimensionality, the efficacy of recombination, and the importance of attractors and, in many cases, of numerous attractors.

Let me pick out just one topic, out of many excellent ones, to highlight as an example. John Holland, Professor of Computer Science at the University of Michigan, will describe the present state of his method for getting computers to evolve strategies for solving problems. He has a sort of community of instructions, with competition and natural selection, and variability produced by random crossing over, as in chromosomes. Lo and behold, clever new strategies emerge from his computer. So far his genetic analogy is with haploid organisms. He has not yet introduced diploid genetics—just think how much better it will be with sex.

A special subject is the evolution of human behavior, where it is evident that biological evolution has been overtaken by cultural evolution. This field has recently been enlivened by controversy between some sociobiologists, who have underestimated the cultural transformation of the biological roots of human behavior, and some cultural anthropologists, who have tried to minimize the role of biology in the explanation of human behavior. I am sure that a synthesis will emerge from this dialectic process. However, the field goes far beyond such a controversy and has contributions from paleontology, primatology, archaeology, psychology, and so forth. To consider a layman’s example, some day we might be able to choose between two popular models of the evolution of organized human violence, which threatens all of us so dramatically in this era. According to one model, there has always been a tendency towards occasional intraspecific violence from early man up through the hunter/gatherer stage of culture and on to the present. As people have formed larger and larger groups, and of necessity have become organized more tightly and on a larger scale, and with improved weapons as well, the scale of violence has correspondingly grown. According to another model, somewhat different, there was a qualitative change at a certain time, perhaps at the time of the invention of agriculture, or a little later at the time of the development of hydraulic agriculture, when relatively peaceful hunter/gatherer societies were replaced by competitive societies with the concept of property. They supposedly initiated real warfare, albeit on a small scale by today’s standards. Marxists tend to adopt the second model, but so do a number of thinkers who are not Marxist. It would be very interesting to be able to choose between these two ideas, or find another way of looking at the whole question. We shall hear some interesting observations from Irven DeVore and Richard Wrangham, whose studies of primate behavior bear on the possible validity of the first hypothesis.

In general, the study of prehistoric cultures now involves an intimate association of archaeology, cultural anthropology, and ecology, but physics, chemistry, botany, and many other scientific subjects are also contributing through what some people call archeometry. They mean the study of old objects, especially artifacts, by advanced technical means that can yield information not only about dates and authenticity, but also patterns of use, methods of manufacture, provenience, and, therefore, mines, trade routes, and so forth. We could in the future throw new light on the mystery of the classic Mayan collapse, for instance, which was the subject of a series of discussions here at the School of American Research some time ago that resulted in a fascinating book. Or we could understand better the successive extinctions of Pueblo cultures here in the Southwest. (I hasten to add that at the time of each extinction, some Pueblo cultures survived, and some survive to this day.) Probably, with the aid of the various disciplines working together, one can to some extent resolve these and other mysteries and, thus, understand better the conditions for the survival of human culture. On this subject, Douglas Schwartz will have some interesting insights to share with us.

In many of the areas of research we are discussing, a common element is the explosive growth of computer capability and of computer-related concepts. We have mentioned that the computer is a marvelous tool for calculations, for theoretical experimentation, and for symbolic manipulation. It is not only an aid to thinking, but a system to be studied and compared, for instance, with the brain. Undergraduate students are choosing computer science as their major subject in record numbers, they are flocking to it like lemmings. Nevertheless, some of us believe that the emerging subjects of information science and artificial intelligence are not providing a broad enough scientific and cultural foundation for research and education in the computer field. Closer ties with many fields of natural and behavioral science and with mathematics would seem to be desirable, as at the conference being planned at CNLS.

Furthermore, it is important to teach students to avoid the pitfalls of reliance on massive computer facilities. Most of us are familiar with these pitfalls. The tendency to calculate instead of to think is an obvious one: “I’ll run it for you Tuesday,” rather than “I’ll think about it for a minute.” Another tendency is to neglect essential qualitative and synthetic aspects of many systems under study in favor of mere analysis of easily quantifiable concepts. Avoiding such neglect is of great importance, because we are concerned here not only with complex physical and chemical systems and computers, but also with such subjects as language, the brain and the mind, ecosystems, and social systems and their history, for which exclusive emphasis on the analytic and quantitative aspects can be disastrous. Many of our topics link natural science, behavioral science, and the humanities, and the contribution of certain subjects in the humanities, such as history and possibly applied philosophy, may be crucial.

That is especially true in the case of policy studies. Policy studies constitute one of the most vital activities in our society, increasingly necessary for our survival. Not often discussed although widespread, policy studies concern the individual, the family, the community, the state, the nation, or even the world community. These studies consider what are the likely consequences of particular decisions; how uncertain are these consequences; how the consequences are likely to affect in some concrete way various systems of values. We have to take into account the enormous and increasing complexity of modern society. These days, much legislation, for example, accomplishes the opposite of what it sets out to promote, along with even larger and unexpected side effects. The same if often true of technical innovation, the side effects of which are notorious. A full-scale study of a local or national or world problem properly done, would have contributions from natural science, social science, applied philosophy (especially ethics and aesthetics), law, medicine, practical politics, and, of course, mathematics and computer science in order to handle the vast number of variables. It is very difficult to bring all these disciplines together, even in think tanks designed for that purpose. Our compartmentalization of learning is becoming more and more of a grave hazard. Here, too, it is especially important and challenging to combine mathematical sophistication in such matters with the proper consideration of value systems often difficult to quantify. Computers have exacerbated this problem, although they need not do so, and they are, of course, essential for huge studies. They need not do so, because with the aid of powerful computers, one can proceed in ways such as the following: devote great care, in any policy study, to finding really sensible surrogates or yardsticks for many of the important values involved, treating this as a major part of the work. Then, instead of assigning relative quantitative measures to the various values and simply optimizing, display in a multi-dimensional way how the different policy options affect all those surrogates and how sensitive the effects are to change in policy. We may find, then, for example, that minor sacrifices in one important value may allow large gains in another. It is important, of course, to estimate uncertainties as well, and even more important to use science, engineering, and general inventiveness to enlarge the sphere of policy options in order better to accommodate many important values.

Thus, we see one example of how computers can be used to render policy studies more humane. A suggestion of how mathematics teaching can accomplish something similar was the main thrust of a lecture I once gave to the students of the Ecole Polytechnique in Paris, a sort of military school of science and engineering that functions as a template of mathematics and mathematical science for the intellectual elite of France. At that time, an invited lecturer spoke not to a few of the students, but the entire student body, which was marched in, in uniform. I started by congratulating them on being privileged to get such a splendid technical education as was offered at the Ecole, then said that, of course, many of them would end up not as scientists or engineers but as managers of great enterprises in France, and that their firm grounding in mathematics would be just as valuable there, since many sophisticated mathematical theories had been developed in economics and management. Then, to the dismay of the students and the delight of the professor of physics and the professor of social science who had invited me, I explained that what I meant was that mathematics would be useful to them defensively, so that they would not be snowed by studies in which relatively trivial matters had been quantified and carefully analyzed, while dominant values were set equal to zero for convenience. We need a balanced and humane use of mathematics in these cases, and people who have not been trained in defensive mathematics will have difficulty defending their sound qualitative judgments against the onslaught of pseudo-quantitative studies.

In my remarks so far, I have tried to sketch, with the aid of some important examples, the revolution that is taking place in science and scholarship with the emergence of new syntheses and of a rapidly increasing interdependence of subjects that have long been viewed as largely distinct. These developments pose a difficult challenge to our institutions. In my remaining time, let me discuss that challenge and one or two possible components of the response.

We have an imposing apparatus of professional societies, professional journals, university departments of research and teaching, government funding agencies, and peer review committees or sections, all directed (at least in part) toward quality control in the traditional disciplines. In the past, it has been possible to accommodate, over time and with considerable difficulty and inconvenience, but in the long run with reasonably satisfactory results, the appearance of cross-disciplinary subjects like biochemistry or nuclear engineering. I believe, however, that the current developments in science and scholarship represent a much more rapid and more widespread rearrangement of subjects than we have experienced before and that it involves much of the most important new work in science. (But certainly not all. Let me make that perfectly clear, as one of our recent national leaders used to say. I am not trying to play down the importance of individual achievement in traditional fields, which remains vital to the health of the scientific and scholarly enterprise.) The apparatus we have described needs to change more rapidly and more radically than it is accustomed to doing, and we must understand what would be useful and appropriate changes and how they might practically be carried out.

Ways will have to be found of permitting and encouraging higher education suitable for the new widely emerging syntheses. Probably in many cases longer and more varied education, perhaps even formal education, will be needed, including years of postdoctoral study and apprenticeship; and we will have to learn to adjust to the personal and economic changes involved.

The whole pattern of grants and peer review must evolve in ways that are hard to prescribe and even harder to carry out.

The journals and professional societies will have to evolve so that the establishment of standards and the conduct of refereeing can be carried out for the new transcendent subjects. All of that will be painful and difficult but exciting.

The universities will have to adjust their departmental structures and modify some of their traditional ways of selecting professors and planning curricula. Our first-class universities are in the hands of very clever people, and I am sure that gradually some suitable changes will come about, as in the other organization despite the existence of very considerable bureaucratic inertia. But the change may well be slow and, for a long while, not wholly satisfactory.

Let me describe, therefore, as one important contribution to the resolution of the crisis that we face, a new institution that could serve as an example and a challenge to the older ones.

The fact that natural and social science are redefining themselves seems to create the opportunity for a new kind of institution that would combine the advantages of the open teaching and research environment of the university with the flexibility of interdisciplinary patterns in national laboratories and other dedicated research institutions.

What we propose is the creation here in Santa Fe of such an institute for research and for graduate and post-doctorate education in selected areas, based on novel principles and responsive to the trends in science and scholarship that we have just been discussing. The typical American university must provide instruction in a wide variety of fields for its undergraduates. Even an institute of technology with emphasis on science and engineering has numerous departments, especially in the humanities, that give service courses. A relatively specialized institute, such as we envision here, cannot provide the kind of general coursework that an American undergraduate is supposed to require. Such an institute should not award a bachelor’s degree. Even elementary graduate instruction of the conventional kind would give rise to problems. Usually there are departments in the traditional disciplines, each offering master’s as well as doctor’s degrees, and each scheduling a variety of full-length lecture courses in a great many subdisciplines. Professorial staff have to be hired to attend to all those courses.

We propose a quite different structure for the new institute, and we would like to hear your comments on it. Full-scale lecture courses would not be emphasized; teaching would be accomplished mostly in seminars and short series of lectures, but, above all, by means of apprenticeship and research. Only the Ph.D. would be awarded, typically in interdisciplinary subjects forming part of the research program, although not necessarily always. Advanced graduate students would be easily accommodated in such an institution. Beginning graduate students and even occasional students without a bachelor’s degree would be welcomed if they could dispense with the traditional array of long lecture courses covering the ground of each subject and dealing with material already available in books. We would hope that many of our students would have acquired as undergraduates an elementary background in natural science, mathematics, the social sciences, and some parts of the arts and humanities.

In this we hope that the Institute can do without the usual departments. Faculty members trained in particular subfields, and with strong interdisciplinary interests of particular kinds, could be selected without worry about having all the other subfields of each particular discipline represented, because we would not try to offer a complete curriculum in that discipline.

Research groupings, which may change over time, would constitute themselves. Presumably, those research groupings would recommend to the faculty and administration highly qualified candidates for new appointments. We need your advice on how this might work. Interdisciplinary appointments, which are often so difficult to make at universities with a traditional structure, would be encouraged. At a typical university, for example, an archeometer with a Ph.D. in chemistry would have a very different time being appointed either to the chemistry or to the archaeology department, in one case because he is doing the wrong research and in the other case because he has the wrong degree. Most archeometers have taken refuge in other places, for example, in the basements of museums.

I had a very interesting experience a few months ago visiting a great university where there is a famous Russian research center. After a little while I found myself dragged off to see the very amiable President of the University. The Director of the Center wanted me to help him persuade the President that the University should appoint a distinguished expert in Soviet economics, who would be immensely useful to the Russian research center. He is a very good Russian scholar and a very good economist, but he was not doing what the economics department thought was its highest priority, and he was not doing what the Russian history department thought was its highest priority, and so, neither department would be able to appoint him. I believe that ultimately common sense won out in that case, but it does not always do so.

That kind of problem is apparently very widespread. It has its foundation in a real concern that lies behind the skepticism about academics seeking interdisciplinary appointments. Faculty members are familiar with a certain kind of person who looks to the mathematicians like a good physicist and looks to the physicists like a good mathematician. Very properly, they do not want that kind of person around. In fact, our organization into professions, with professional societies, journals, traditions, and standards of criticism, has much to be said in its favor, because it helps to safeguard excellence. Presumably some new patterns of setting standards are needed, and that is something we could well discuss.

It is important to recruit for the faculty of the institute some of those rare scholars and scientists who are skilled and creative in a variety of subjects. We hope, too, that among the graduates of the Institute there would be more of this kind of person. Of course, not all the graduates would be genuine polymaths, but we would hope to turn out graduates capable not only of solving particular problems, but of thinking and analyzing and especially synthesizing in a wide variety of contexts.

Ways will have to be found of encouraging teamwork among people of the most diverse backgrounds interested in the same emerging syntheses. Here it will be important to have some scholars with synthetic minds who can grasp the similarities, especially theoretical parallels and common applicable techniques, among the many subfields under discussion and also specialists (in a few remarkable cases, the same people, but in most cases different people) who are responsibly familiar with the structure and the properties and the observational or experimental facts of each subject.

One of the challenges that we face, in tackling subjects that involve mathematics and natural science on the one hand and also social and behavioral science on the other, is that of marrying quite different intellectual cultures. The problem is exacerbated by the fact that many of the most mathematically sophisticated social scientists are those who are most attracted by the analyzable at the expense of the real. Fortunately, there are others who combine a concern with the crucial qualitative features of their subject matter with a receptivity to ideas from mathematics and natural science; and there are also natural scientists who are capable of learning about the complexities of human beings and their institutions.

There are some psychologists and pop psychologists who like to place people on a scale running from Apollonian to Dionysian, where, roughly speaking, Apollonians tend to favor logic, rationality, and analysis, while Dionysians go in more for intuition, feeling, and synthesis. In the middle are those tortured souls, the Odysseans, who strive for the union of both styles. The new institute would have to recruit a number of Odysseans to be successful!

You have read in our brochure about how we would have permanent faculty, tenure-track faculty, junior faculty, Ph.D. candidates, post-docs, visiting faculty, and non-resident fellows who would visit from time to time on a regular basis.

The research program of the Institute would include both experimental and theoretical work, which complement and reinforce each other. We would differ fundamentally, therefore, from the Institute for Advanced Study in Princeton, for example, which has no experimental work, does not award degrees (although I believe it is allowed to), and does not have very much collaboration among different kinds of scholars. Experimental and observational work of very expensive kinds, such as high-energy physics, astronomy, and oceanography, should probably not be undertaken, while use is made of cooperative arrangements with nearby observatories, laboratories, museums, industrial enterprises, and so forth.

I should mention that it is very tempting to consider adding future studies and policy studies to the material covered by the Institute. There is an urgent need to apply the skills of scholars and scientists to the problems facing communities, regions, nations, and the world. However, the nature of such policy studies, along with the mix of people necessary to do justice to them, is probably sufficiently different from that of the subject we have been discussing, that it would be better (and we need your advice and comments on this) to organize an autonomous and separately funded organization nearby that would concern itself with policy studies and speculation about the future. Such a nearby think thank, if it is created, could then employ selected faculty members, visitors, and students as consultants or part-time staff members, but it would also employ a number of distinguished full-time investigators experienced in policy studies and public affairs.

We describe in the brochure how after some five to ten years of growth, the personnel of the Institute would consist of so and so many professors and so and so many secretaries, and so and so many students, but we need your advice as to whether the numbers are reasonable and how to get from here to there.

The location of the Institute in this vicinity seems to provide a uniquely attractive cosmopolitan environment in a relatively unspoiled setting. (Of course, all buildings in Santa Fe look like this one, and the weather is always the way it is today!) Recruiting a superlative faculty and gifted students will be facilitated by this choice of location. George Cowan has described the proximity of Los Alamos, the radio and optical observatories of the Southwest, the museums and the Laboratory of Anthropology in Santa Fe. There is an emerging high technology research corridor in the Rio Grande Valley. It is also remarked in our propaganda that within a thousand mile radius lie the San Francisco Bay Area, the Rocky Mountain Region, Chicago, Minneapolis, St. Louis, and all of Texas. In any case, it may be, in an age of advanced communications and satellite television, that intellectual stimulation and the exchange of ideas will not require proximity to large urban agglomerations, and that we will be pioneering in that respect as well as others.

At the same time that we will be seeking very substantial funds for the endowment and trying to work out how such an institution could best be structured and governed, we will be starting up a program of intellectual activity by establishing so-called research networks. Now research networks have a relatively long history, as exemplified fifty years ago by Delbruck and Luria, who were supported, I believe, by the Rockefeller Foundation. One does not really invent such networks; to some extent they already exist as invisible colleges, colleges without walls, but one can discover and assist them and develop them further. The MacArthur Foundation has been experimenting with such networks for the last few years, particularly in supporting research in scientific fields relevant to mental health. A subject is chosen (for example, the psychobiology of depression) and some research groups from different institutions are selected to participate in the network studying that subject. The groups and the individuals composing them represent a variety of disciplines, and the groups are chosen for the compatibility and complementarity with one another as well as for their excellence, so that they are able to function in a pattern of collaboration. The Foundation helps the groups to communicate with one another by telephone and by computer mail, by means of conferences and summer studies, and by exchanging post-docs as well as data, samples, information about methods, and so forth. It is hoped that the research network can then carry out an integrated attack on the problem it is studying.

In a somewhat analogous way, our Institute, if it can obtain operational funding, can start very soon to set up research networks for studying some of the emerging interdisciplinary syntheses we have been discussing. For each network, composed of individuals and research groups at various institutions, we will provide computer links and a budget for other kinds of communication, including meetings here in Santa Fe, probably short ones during the academic year and workshops lasting for weeks in the summer. The central headquarters here would be responsible for arranging the details.

During the early phase of operation of the Institute, there would be only a small faculty here. As academic members of the Institute begin to appear in Santa Fe (at first mostly non-resident fellows and others on leave from institutions elsewhere), there would be a few scientists and scholars representing each network locally and enhancing its cohesion.

What would be the relation of the network activity of the Institute to the existing academic and industrial organizations to which we all belong? The network activity could only be a benefit to those organizations and to their members and, if it proceeds as we hope, it would greatly facilitate the research of participants, wherever they are. At the same time, it would strengthen the nascent Santa Fe Institute. In fact, if we consider the two operations, building the networks and establishing the permanent Institute, we see that each is very valuable in itself and also that they are mutually beneficial.

As the permanent Institute gradually comes into being, there is no reason to believe that the networks will cease to operate. Assuming they are successful, they should presumably continue indefinitely and constitute one of the principal modes of operation of the Institute, adding strength to it, and also to many of the leading academic organizations in this country and to some abroad. In the long run, some of those institutions may be taxed by having one or two of their faculty members lured away, and an occasional bright student, but in exchange for that tax they would be provided with a very valuable service.

One of the most important questions that we have to address is this: Why not try to accomplish some of our objectives by adding to the activities of an existing university and saving the cost of creating a new institution? Well, I think that the national response to the challenge of the emerging syntheses will consist in great part of steps taken by the universities. They have already begun to respond to a considerable extent. But the form of the response, as I indicated before, is not likely to be adequate for a long time.

Let me poke a little fun at the universities and institutes of technology. The typical response of a university to the emergence of a new interdisciplinary subject is to set up a Center in an old Victorian house or a little shed left over from the First or Second World War, funded with soft money and treated to some extent like a stepchild. Wonderful results often emerge from these dilapidated structures, but some of the most talented researchers are not in permanent positions, have little influence on teaching policy, and are far removed from the centers of influence in the institution. Of course, a senior faculty member who has distinguished himself in a particular profession and made a great reputation can afford to shift to a new, interdisciplinary subject. He can sometimes get funding, although that is not very easy. However, the younger people who want to work on the new subject may have great difficulty furthering their careers, unless they wish to spend years becoming famous in some old-fashioned field.

It will be a slow and difficult process for each university to change from its old message, “Learn a traditional subject and stick to it,” to the new message, “It is all right to learn how to make connections among different subjects.” We would like to create here in Santa Fe a least one institution that is free from the drag exerted by past specialization and the tyranny of the departments, an institution that would encourage faculty, students, and young researchers to make connections. The message, that it is all right to think about the relations among different approaches to the world, may then spread more readily to the world at large: to the universities, the technical institutes, and even to the primary and secondary schools, where innumerable opportunities to point out connections are wasted every day. Thank you.