Beginning a Dialogue on the Changing Environment for the Physical and Mathematical Sciences: Report of a Conference (1994)

Daniel Kleppner

Department of Physics and Research Laboratory of Electronics

Massachusetts Institute of Technology

Science has been regarded with such favor by the public and the government ever since World War II that basic research should not require much defending. In the colorful words of Representative George E. Brown, Jr., chairman of the House Committee on Space, Science, and Technology: “Science has come to occupy a place in American culture somewhere alongside baseball, Old Glory, and the right to avoid self-incrimination ” (American Journal of Physics, September 1992, p. 779). Times are changing and science, along with many other institutions, is being challenged. In May 1993 the National Science Board issued a statement titled “In Support of Basic Research, ” which reveals not only that the Board endorses the role of the National Science Foundation in supporting basic research, but also that it judges that the value of basic research needs reaffirming. Questions about basic research in the United States range from the reasonable —How much basic research should we do?—to the unreasonable—Why do any? There is general agreement that our national goals must include economic prosperity and a healthy environment, but there is no unanimity about the role of science in achieving these.

“Basic research” means scientific research whose primary goal is to understand nature rather than to address a particular problem. In some quarters, the term has become slightly suspect, carrying connotations of elitism and suggesting visions of scientists having fun at the expense of the public. Consequently, one sometimes hears euphemisms such as “curiosity-driven research” or “long-range research.” However, the fact of the matter is that basic research is elitist and scientists often have fun—the sort of fun that can come from exhausting labor year after year.

This rationale for basic research is sometimes called the “serendipity argument.” Its essence is that investments in basic research eventually pay handsome dividends, though one cannot predict how. There is lots of evidence for serendipity in science starting with Michael Faraday or even going back to the dawn of modern science. A recent summary appears in Science, Technology, and the Federal Government: National Goals for a New Era, a report of the Committee on Science, Engineering, and Public Policy (COSEPUP; National Academy Press, Washington, D.C., 1993).

History is rich in examples of scientific research that have led to practical applications in areas far removed from the original work. Fundamental

research on electromagnetism contributed directly to the development of modern communications. Investigations in solid-state physics enabled the invention of the transistor. The recombinant DNA technology that led to the biotechnology industry arose from studies of unusual enzymes in bacteria. Mathematics, often regarded as highly abstract, is at the core of applications as diverse as aircraft design, computing, and predictions of climate change. (p. 17)

The Task Force on the Health of Research of the Committee on Science, Space, and Technology, U.S. House of Representatives, in a report to the committee in July 1992, expressed doubts about the assertions of the serendipity argument:

A major failing of U.S. science policy has been the absence of institutional mechanisms designed to test the validity of these and related assertions. There are three fundamental reasons for this failure. First, the U.S. research system was designed primarily by and for scientists, and these assertions serve to preserve the autonomy and legitimacy of the research community. Second, it is the researchers themselves who are called upon to evaluate a system which they have little incentive to alter. Third, the economic performance of the United States in the 1950s, 60s and 70s has been construed as a vindication of these science policy principles. (It must be emphasized, however, that the economic preeminence of the U.S. after World War II was virtually a foregone conclusion because the production capacity of most of our economic competitors had been destroyed.)

The suggestion that our scientific enterprise is fundamentally self-serving and may even be a sham is bound to trouble U.S. scientists. Like most of my colleagues, I am well aware that I pursue my personal interest in science at the public's expense. From time to time—particularly after working on some idea that eventually turned into a dead end—I wonder whether the public is getting its money's worth. For every success there are numerous failures. However, recently I learned that research in which I participated years ago has paid off. The story provides one more piece of evidence that basic research can be an excellent investment.

In the 1950s I studied undergraduate physics at Cambridge University courtesy of a Fulbright fellowship. My tutor, Kenneth F. Smith, was a pioneer of atomic beam research in Great Britain. To provide a little enrichment to the syllabus, Smith explained the workings of magnetic resonance. He mentioned that it could in principle be used to make an atomic clock, possibly one accurate enough to see the effect of the earth's gravity on time, the “gravitational red shift.” The thought that gravity alters time struck me then, as now, as fantastic. The following year I entered graduate school at Harvard University, and when Norman Ramsey proposed that I try out a new idea for a superaccurate atomic clock, I lost no time in getting to work. The art of making an atomic clock fundamentally involves

observing some internal motion of an atom for as long a time as possible. According to the uncertainty principle, the natural fluctuations in the frequency measurement decrease as the observation time increases, provided that nothing else interferes with the measurement. Ramsey 's proposal was to increase the observation time by letting the atoms rattle around in a “bottle,” a chamber in which they might collide with the walls without feeling unduly upset. A rough analogue would be a football game, with the ball replaced by a clock that is expected to keep good time. After a few years of work Ramsey's suggestion bore fruit in the form of a device known as the hydrogen maser.

Shortly before the maser first operated, however, the gravitational red shift was detected by two other physicists at Harvard, R.V. Pound and G.A. Rebka, Jr., using a totally different technique (the Mossbauer effect) that did not involve an atomic clock. Nevertheless, the maser effort bore scientific fruit in a variety of precision measurements of fundamental constants and studies of atomic phenomena. Also, as anticipated, the maser turned out to provide quite a respectable atomic clock. Many years later it was actually used not only to observe the gravitational red shift but also to measure it precisely. In 1979 Robert F.C. Vessot and his colleagues at the Smithsonian Astrophysical Observatory launched a hydrogen maser in a rocket and tracked its frequency as it traveled upward through a distance roughly the radius of the earth, and then fell back. During the flight the clock was stable to a precision corresponding to about one second in ten million years.

Let me now jump to the present. A new consumer product will shortly make its debut, a device that will let you know your location almost anywhere on the earth to an accuracy of about 30 meters. It is based on the global positioning system (GPS), a navigational system originally developed by the military that uses time signals from a series of satellites that carry atomic clocks. The system is kept synchronized by a central station that employs hydrogen masers. The GPS system is currently used for commercial navigation and some consumer purposes. GPS receivers with built-in atlases are available for yachts at a cost of roughly $1,500, and somewhat cheaper versions are available for campers.

Plans are afoot for mass producing the GPS receivers on microchip and integrating it with a sophisticated storage and display system. In a few years it should be possible to buy a low-cost device for your car that will indicate your position on a map of the nearby streets, wherever you happen to be. The potential market is enormous, for practically everybody will want one. The potential economic benefit from the new industry is also enormous, not to mention the benefit of time saved from being lost.

A number of advanced technologies are crucial to the GPS. Foremost, of course, is the ability to employ reliable satellites. Ultraprecise techniques for transmitting time signals from space to ground are also essential. These techniques were originally developed for the hydrogen maser red shift experiment. Advanced geodetic analysis, microelectronics, and sophisticated data-processing methods are all essential. Nevertheless, the heart of the global positioning system is the atomic clock.

General relativity would rank low on any list of scientific priorities based on potential economic return. The atomic clock research took years to start paying off: more than two decades for military use, more than three decades for a consumer industry. However, the financial return from this single project is likely to considerably exceed the cost of all of the basic research in atomic physics ever done in the United States, not including the benefits of

simple and precise air and sea navigation, the savings to drivers and truckers from the costs of being lost, and the numerous emergency applications.

The research also paid off in ways whose value cannot be so easily measured. There were substantial scientific dividends. These included advances in precision measurement (Norman Ramsey received the Nobel Prize in 1989 for his general contributions to precision measurements), crucial contributions to the development of very long baseline radio interferometry, and applications to deep-space navigation. A number of hydrogen masers were used to navigate the Voyager fly-by of Neptune a few years ago. The most important dividends, however, were in human resources. Students trained in the hydrogen maser research have gone on to careers in universities, colleges, government laboratories, and industry. Some stayed in the general area of atomic physics; others went into astrophysics, biophysics, surface physics, and electrical engineering. Some have devoted their careers to education in colleges. Others have become internationally recognized leaders in research.

Science critics will argue that the serendipity argument is unconvincing because it is anecdotal and anecdotal arguments are suspect. Furthermore, serendipity can happen anywhere: science has no monopoly on good luck. In practice, however, it is difficult to think of any institution in which serendipity pays off so relentlessly and so bountifully as it does in basic research. Technology-driven programs, for example, are not particularly good bets for serendipity. To cite one example, the spin-offs from NASA's manned space program have been paltry. None can compare, for instance, with the discovery of the buckeyball and the C₆₀ compounds that has created a new carbon chemistry and the possibility of fabulous materials. This discovery occurred in the search to learn how matter assembles itself in molecular beam clusters, with some inspired ideas from radio astronomy, research that would not rank high on most lists of potentially useful projects and whose cost, one might add, would be in the noise of a NASA budget.

Basic research needs to be defended today from policies and practices that have grown obsolete. The military threat that propelled much of our research agenda has disappeared, and the importance of basic research is being questioned. There are reasons for concern about the future. Fewer U.S. students are opting for research careers. University positions are losing their attractiveness as the competition for funding increases and the scientific infrastructure deteriorates. Basic research is a long-term investment, but the nation is preoccupied with short-term problems.

Various studies on science planning in the United States have been carried out in recent years, including Federally Funded Research: Decisions for a Decade (Office of Technology Assessment, May 1991), Renewing the Promise: Research-Intensive Universities and the Nation (President's Council of Advisors on Science and Technology, December 1992), the series of the Carnegie Commission on Science, Technology, and Government (Carnegie Commission, New York), and, of course, Science, Technology, and the Federal Government. In such a situation it would seem presumptuous to offer one's personal analysis. Nevertheless, the following three issues seem to me to be of particular urgency.

The discussion focuses on the physical sciences in the universities. Although this is only a small part of our total scientific effort, the universities are crucial to our future. They are responsible for more than half of the basic research and essentially all of the professional education in the physical sciences.

The phenomenal success of U.S. science in the decades following World War II is often cited to vindicate our pluralistic system of carrying out research in a variety of institutions with federal support coming from many agencies, all in a relatively uncoordinated manner. The system's flexibility allowed for rapid innovation and provided scientists with several avenues of recourse in finding a research sponsor. Sentimental attachment to this rather unconventional way of organizing science should not blind us to the fact that it has not worked well in recent years and that it is, in fact, failing.

Imposing a science policy on a diverse research community with differing agendas is a daunting task. The creation of the Federal Coordinating Council on Science, Engineering, and Technology represented one step in this direction. From time to time the creation of a Department of Science and Technology has been proposed for generating and implementing a coherent policy. Such proposals have generally been rejected due to concerns about the dangers of concentrating science-related decisions in one department. Nevertheless, Philip W. Anderson argues that this is an idea whose time has come (Physics Today, June 1993, p. 9). As evidence of the need for centralized science planning he cites the following. (The excerpts are paraphrased.)

• Direct government funding has emphasized big science projects while other types of research have had to rely on military and industrial funding. Industrial spending has plummeted and military research has been reduced, threatening areas of research that are essential to the development of new technologies and the creation of new industries.

• Different kinds of research are represented at vastly different levels of the executive branch. Energy research is at the Cabinet level; biological research is at a lower level in the Department of Health and Human Services. The National Science Foundation is many levels lower yet. Its appropriations are in competition with veterans affairs and welfare, and it is subject to continued pressures by Congress to alter its fundamental mission.

• Space Station Freedom is moving forward in spite of broad opposition by the scientific community. Although its official mission has become exploration, it is still represented to the public as an important scientific project.

• Pork barrel funding is seriously eroding science. The problem is due both to the power of Congress and the selfishness of scientists.

• Three national laboratories are concerned with nuclear weapons; one would be enough. Procedures are lacking for rapidly converting these laboratories to civilian use.

• We have no way to control the production of scientists. In the words of Anderson, “We are overproducing and undertraining young people. ”

The disadvantage of a Department of Science and Technology—aside from the obvious danger of adding a new layer of bureaucracy —is that it would decouple research programs from the agencies that need them. Nevertheless, unless we discover how to generate and implement a coherent science policy, problems such as these can only get worse.

Proliferation of PhDs. The number of PhDs awarded annually in the natural sciences and engineering grew from about 11,000 to 16,000 between the years 1980 and 1990 (Renewing the Promise: Research-Intensive Universities and the Nation , 1992, p. 32). Clearly one cannot sustain such growth indefinitely. To put these numbers in perspective, the PhD production in 1990 was about the same as in 1970, at the end of a decade of unprecedented scientific expansion. However, the growth rate in 1990 was much larger than in 1970. At present, many PhDs in physics are facing great difficulty in obtaining satisfactory positions. One could argue that the real problem is not overproduction but underutilization, but the fact remains that in the physical sciences there is currently an imbalance between work-force supply and demand.

The figures for total PhD production conceal the fact that the interest of U.S. students in scientific careers is dwindling: essentially all of the growth in recent years is due to the increasing number of foreign students. In the past, approximately one-half of these students have made their careers in the United States. The steadily increasing number of foreign students indicates that graduate education in the United States continues to be coveted abroad and that whatever the actual employment situation is in the United States, relative to opportunities in most countries, the United States remains a land of opportunity. Nevertheless, the issue is disturbing on several counts. First is the concern that having to rely on foreign students to meet our scientific needs indicates a serious inadequacy in our educational system. Second is the concern as to whether it is wise for our country to sap the scientific talent of foreign nations. Consider the former Soviet Union. U.S. universities now have their pick of the best Russian scientists. In the long run one can question whether it is in our national interest to help devastate Russian science. Finally, being dependent on foreign scientists makes the nation vulnerable to changes abroad. If the scientific standard of living abroad should rise enough to make careers back home attractive to foreign scientists, as it may well do in China, we could face a depletion of our scientific work force.

One might view the present employment predicament in physics as merely a symptom of a generally poor economic situation, which should improve. If our nation is to prosper, a stable supply of highly skilled scientists and engineers is essential. In the absence of a comprehensive science policy, however, the roller coaster of supply and demand for scientists is likely to continue.

Proliferation of research groups. There has been a steady growth in the number of physics research groups in the universities. Between 1984 and 1992, the number of research awards from the National Science Foundation's Directorate for Mathematical and Physical Sciences (astronomy, chemistry, materials science, mathematics, and physics) grew by 39 percent, from 3,621 to 5,021. The research budget has not kept pace. Although there was some growth in the early 1980s, the budget since 1985 has been essentially flat in constant dollars, in spite of the fact that the overall NSF budget increased substantially. The result, as one might expect, has been the erosion of our research capability. Two symptoms are of particular concern: researchers are spending inordinate amounts of time in attempting to obtain funding, and even the best U.S. research groups are finding themselves in competition with European groups that have vastly more resources at their disposal. As an example, the Max Planck Institute for Quantum Optics in Germany has a budget about two-thirds the size of the total NSF program in atomic, molecular, and optical physics. The average NSF grant size for experimental physics is about $120,000 per year, whereas the cost of a modest program in a private university—two graduate students, a postdoctoral researcher, some equipment—is about twice that amount.

A policy of stretching a relatively fixed budget to accommodate a steadily increasing number of research groups would be foolish. The problem, however, is not a poor policy. The problem is no policy. Without some overall plan for the future, the situation for university research in the physical sciences will continue to deteriorate.

The balance of resources between large projects organized around major facilities and support for the small research groups in the universities and elsewhere is an increasingly divisive issue in the science community. The major facilities are created in response to initiatives from the research community, but because of their magnitude, they inevitably acquire a political constituency. No such constituency is concerned with the health of small science. This problem is well known and widely discussed. However, in the absence of a comprehensive science policy, there is no way to assure a reasonable balance between these styles of research or that our overall scientific investments are reasonable.

The COSEPUP report Science, Technology, and the Federal Government presents national goals for science and technology. In the absence of a coherent science policy, the prospects that the nation will meet these goals, or any other scientific agenda that is deemed worthy, are poor. The COSEPUP report states, “For half a century, the federal government has strongly supported basic research in science and engineering. However, the government has never formulated an explicit policy for setting the level of that support” (p. 29). We can no longer conduct ourselves in this manner. We must recognize that our “no policy” is no longer working.