Proc. Natl. Acad. Sci. USAVol. 80, pp. 7679-7683, December 1983Applied Physical Sciences

Science, technology, and invention: Their progressand interactions

(research planning/pure and applied science interaction/innovation/science policy)

CHARLES H. TOWNESDepartment of Physics, University of California, Berkeley, CA 94720

Contributed by Charles H. Townes, September 9, 1983

The interaction between pure and applied science is most fre-quently thought of as a flow of ideas and contributions frombasic discoveries toward applied science and industry. As sci-entists we are both conscious and proud of such contributions.A simple documentation of this well-recognized type of flow isprovided by the list of Nobel prizes in physics since 1925; ap-proximately 40% of the discoveries involved led rather rapidlyto substantial industrial applications of the new ideas they pro-vided. Most of these discoveries can be said to have originatedin basic science, and rather soon thereafter many have had im-portant industrial applications; some have produced large in-dustries. One can point, for example, to the discovery of theneutron by Chadwick, coupled with Cockcroft and Walton's ar-tificial radioactivity and slightly later Lawrence's contributionsto nuclear processes. These led to the growth of a nuclear in-dustry that has many important applications in addition to theproduction of nuclear energy. There is of course the transistorand solid-state electronics. There is also quantum electronicsand the somewhat smaller but already substantial industry ofmasers and lasers and their applications. There are other dis-coveries that will probably hatch industries somewhat later, suchas the Josephson effect, which has wide potential applicationsbut it has not yet produced an industry. In addition to discov-eries that have rather specific applications, there are also ab-stract ideas such as the uncertainty principle, which may haveno obvious and specific application to industry but yet pervadesall our thinking and thus makes many indirect contributions toapplied work.While scientists are rather familiar with the flow of contri-

butions from basic to applied work, and the phenomenon is rec-ognized in generality by the broader public, I do not believeadequate recognition is given to the reverse flow-the contri-bution of technology to basic science. Convenient and sophis-ticated instrumentation is but one example. More broadly, suchcontributions encompass important scientific discoveries pro-duced by applied research, the development of industrial andcommercial products on which much basic research depends,and new technical possibilities that emerge from applied workin industry and in military and space programs.

Contributions of applied to basic sciences

Harwit has recently surveyed the growth of astronomy and,while his views may be somewhat controversial, I believe theymake an important point that is pertinent to understanding in-teractions between the sciences, basic and applied. He con-cludes that many of the most significant discoveries in astron-omy have been made shortly after a new technique has been

introduced, and generally by people from outside of the field.One can cite a number of cases of this type; for example, Jan-sky's discovery of radio astronomy in the early 1930s. Janskywas an electrical engineer at Bell Laboratories who constructeda system for the sensitive survey of noise in radio communi-cations. In the course of this work, he discovered and identifiedthe first radio waves from extraterrestrial sources. Other ex-amples are the initiation of x-ray astronomy by a group at theNaval Research Laboratory, including Towsey and Friedman,by the use of newly available V2 rockets; the subsequent dis-covery of x-ray sources outside of our own solar system by Rossiand Giacconi; and the first detection of y-ray bursts by the LosAlamos group using the Vela satellite, which had been put inorbit for military purposes. There is also the work of Hewishand Bell, who put into operation new radio equipment that coulddetect short pulses of radio waves and unexpectedly discoveredpulsars. Then there is the discovery by Penzias and Wilson ofthe microwave background radiation as a result of careful ex-amination of microwave noise with receiving systems of newsensitivity. Many recent astronomical discoveries have grownout of the space program, a program stemming largely from theefforts and interest of engineers and government officials, butpromoted to a certain extent by scientists.

Some astronomers appropriately say that, although these wereimportant discoveries, the most important work of astronomyis the understanding of such matters and their coherent inter-pretation. Nevertheless, in any list of discoveries that havefructified the science, these examples are ones of prime im-portance and several have initiated new branches of astronomy.

None of the discoveries cited above was the direct or im-mediate result of the availability of some commercial product,but they do come from instrumentation that largely dependedon applied development and interests. The contributions of moreor less standard commercial products are also exceedingly im-portant to scientific discovery. Obvious modern examples arecomputers and solid-state electronics. Another is the devel-opment of a wide variety of commercial lasers, which are rap-idly penetrating scientific laboratories and providing new ex-perimental potentialities. Many types of specialized materialsare very important to research. In my own present work I relyon infrared detectors developed largely for the military estab-lishment, on very fine wire mesh gauzes developed for com-mercial filtering, on sensitive solid-state amplifiers, and on veryfast circuitry. These are all available with a quality that I couldnot practically produce myself. If they were not available, I wouldhave to do a different kind of astrophysics and could move onlyat a much slower pace.When Millikan undertook his famous oil drop experiment,

he personally built about 1,000 lead cell batteries in order tohave a high, stable voltage. We now get such a voltage supplyoff the shelf; it is simple, small, cheap, and quickly available.

7679

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertise-ment" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Aug

ust 1

5, 2

020

7680 Applied Physical Sciences: Townes

Some of the contributions of industry come about becauseof a particular industrial drive and interest; some are possiblesimply because of the availability of money. Industry can un-dertake developments and explorations that people in basic sci-ences and at universities simply cannot afford. This allows avery much broader base for developments that are difficult,expensive, and yet important enough to industry that they arevigorously pursued. In the long run, they are often also im-portant to basic science.

There are many interesting cases of the effect of adequateresources in the military or the space programs as well as thoseprimarily associated with industrial interest. In the early dayswhen integrated electronics were just being worked on, I re-member a telling answer when I asked an acquaintance met inWashington as to what he was doing there. He was approachingofficials in the Pentagon to get more money for his company towork on integrated electronics. He explained that integrated-solid-state circuitry was so far off and chancy that his companywould not think of putting any of its own money into such de-velopments but, since the military was willing to support thistype of research, the company was glad to be occupied withwhat would otherwise be a poor research investment. Such ajudgment was not true of all companies, but it was not an un-common attitude. Without governmental interest and a mas-sive infusion of money and effort, the integrated electronicsthat now means so much to both science and commerce mighthave been delayed.

I have emphasized the contributions from applied to basicscience partly because of their importance, but partly becauseI believe they are often not adequately recognized. The flowof ideas from basic to applied science is not less significant, butit is more frequently noted.

Importance of interactions between sciences and ofconcentration of effort

In addition to the development of ideas and of technology, theprogress of science is much affected by the connections andflow of information between fields and between scientists. Onecan see this in a positive sense from fortunate interactions thathave stimulated ideas and in the negative from delays in im-portant developments because of the lack of appropriate con-nections.

Consider, for example, the imaginative work of Gamov, Al-pher, and Herman that suggested and examined isotropic ra-diation produced by an initial universal explosion. The impactof this work was limited, I believe, because of such miscon-nections. Gamov, Alpher, and Herman apparently did not ap-preciate the fact that we were very close to being able to detectthe isotropic microwave radiation which they predicted theo-retically; appropriate experimentalists who might have been in-terested, on the other hand, were not close enough to this the-oretical work to become interested. It hence remained for Dicketo later reinvent the whole thing; he fortunately was knowl-edgeable about both the theoretical and experimental aspects.While Dicke was developing equipment to look for this radia-tion, it was found by Penzias and Wilson-one might say byaccident, but of course an "accidental" great discovery dependson necessary scientific sensitivity and ability. Initially, Penziasand Wilson were not in touch with Dicke and his already-de-veloped ideas. Their understanding of this discovery might havebeen slower than it was if they had not very quickly made suchcontact.

Another illustration of the importance of contacts comes fromsome of my own work. I apologize for so many personal illus-trations, but of course I am a bit more sure as to what happened

in my own field than I am in others. While musing on ways tomake real my general idea for a maser, I was profoundly helpedby a conversation with Paul, a German scientist who had visitedColumbia University and talked about an interesting new wayof producing molecular beams of especially high intensity. Hismethod was fresh in my mind, which made it seem practicaland possible for a maser to work. Without this conversation withPaul, I might possibly have dismissed the idea for some time.

Obviously, many factors enter into scientific or technicalproductivity. An interesting aspect of this comes from Shock-ley's study some time ago of the distribution of numbers of pat-ents per staff member at Bell Laboratories and of numbers ofscientific papers published per staff member. Of course, sim-ple numbers of patents or papers do not necessarily measurecreativity. However, they represent some measure related toproductivity, and that measure shows a striking distribution. Afew individuals in this study had each produced 100 or morepatents, while most staff members had 1, 2, or only a few. Thedistribution was very peaked, as was the distribution of num-bers of scientific papers. It clearly did not correspond to anysimple gaussian function of, let us say, IQ, but was character-ized by an almost divergent peak, approximated by what is gen-erally known as a l/f distribution. That is, the number of in-dividuals who had produced a given number of patents or paperswas approximately inversely proportional to the number of suchpatents or papers.One can speculate about what makes certain individuals in

their particular situations so highly productive; a statistical dis-cussion by Montroll makes it very plausible that the combi-nation of a large number of different effects, each making itsown gaussian-type contribution, characteristically produces sucha l/f distribution. It also seems reasonable that many differentindividual factors affecting research productivity can have a

multiplicative effect. One of these probably is intensity of twodifferent kinds-the intensity of work on the part of the in-dividual and the intensity of concentration within an institu-tion, so that the cross-fertilization of information and ideas peakboth in the individual and at the institution. An individual's in-tuition is clearly related to his or her knowledge, experience,and rapidity of sorting out better solutions from poorer ones,using these qualities. Frequently, the efficiency of sorting outunfruitful routes for solution of a problem depends on easy ac-cess to a friend, who has relevant knowledge. Hence, beingclosely surrounded by other scientific and technical activity isoften important to new and exploratory work. When an inter-esting but nonstandard idea comes up, the quality of access toapparatus for quick experimentation or the quality of opportu-nities for discussion with-others who have useful knowledge or

experience can make its exploration either practical or impracti-cal.We need to consider the extent to which we in the United

States will lose the ready contact with others on the forefrontof their fields or the ready access to information and materialsneeded, as research and industry of high quality continue theirspread to additional countries and the United States plays a lessdominant role in science as a whole. We have in the past beenfortunate in having a remarkably central role in science andtechnology, in attracting to the United States many outstandingscientists from abroad, and in maintaining active industries inmany fields. The dilution of the world's scientific and technicalefforts can bring inefficiencies as -well as the welcomed in-creased participation of all nations in the scientific enterprise.Fortunately, today communication is rather efficient and travelis quick and relatively cheap. However, travels to many meet-ings can become oppressive and it is not clear that intercom-munications among scientists, particularly scientists in differ-

Proc. Natl. Acad. Sci. USA 80,(1983)

Dow

nloa

ded

by g

uest

on

Aug

ust 1

5, 2

020

Proc. Natl. Acad. Sci. USA 80 (1983) 7681

ent types of environments, will in the future be as favorable foranyone as they have been in the past for the U.S. scientific com-munity.

While the general environment has a profound effect on thedevelopment of science and technology, one must not neglectthe importance of individuals and of the microenvironment.This bears on the question sometimes asked whether scientificor technical innovation is really the product of some uniquelyinnovative individual in the right situation or instead is simplya series of discoveries whose time has come. The latter rep-

resents the view that, when the time is ripe for a particular de-velopment, it comes along; while there is an individual or a group

which produces it first, there are many others who might havedone so only a short while later. This view would make dis-covery not critically dependent on any individual or fortunateenvironment but rather a matter of general overall progress.There are certainly some ideas and scientific developments withthis characteristic, but one can also find those that were delayeda long time for lack of some key element or individual.

Discovery of the isotropic radiation (the "big bang") is one

of those that was probably delayed for want of the right dis-coverer. Available techniques and ideas would have allowed itsoccurrence much earlier if someone had made the appropriateapproach. The breakdown of parity is another case; effects ofnonconservation of parity were noticed as early as about 1930,and someone with the right idea might have suggested such an

effect much earlier, even if clearcut experimental proof wouldhave been more difficult than in the post-World War II period.The transistor, on the other hand, is a case in which effectivediscovery and development could probably not have occurredmuch earlier than when it did. There was at least one overt at-tempt before World War II to make a solid-state amplifier butit failed for reasons that were poorly understood at the time.Success depended, I believe, on the development of solid-statetheory and more understanding of the solid state than was avail-able at a substantially earlier time.The maser and laser represent an interesting case. I see no

component idea required for these that, when they were in-vented, was younger than about 25 years. There seems to beno reason the maser and laser could not have appeared about25 years earlier than they did if known ideas and techniqueshad been put together in the right way. In addition, one now

finds many naturally occurring astronomical masers; they ob-viously predated man. Some produce rather strong radiation-e.g., the water masers that radiate at a wavelength of 1.35 cm.

These could have been detected much earlier with availabletechnology; astronomy might thus have led us to develop ma-

sers and lasers earlier if we had happened to detect these as-

tronomical systems.Antibiotics provide another interesting case. Extensive use

of antibiotics came along only 10-20 years after their discoveryby Fleming; nevertheless, there was a dead period of about 10years with little interest in antibiotics until Florey began, in1938, to look into their potential. Within only a couple of years

after this, antibiotics became available for medical purposes.

Thus, important gaps in scientific developments do occur andprogress in science and technology is not a regular and steadymarch forward, produced automatically by the world's or a na-

tion's generally favorable regard.

The problem of planning and foresight in research

Two different routes toward scientific progress may be de-scribed; they cannot be completely separated, but for heuristicpurposes their separate consideration is useful. One route is thepursuit of clearly important science that is foreseen by a num-

ber of those who are expert in the field. In many cases, theimportance of a certain line of development seems clear. Fur-thermore, it may be known approximately how to pursue theappropriate line of investigation. What is then needed is onlythe appropriate combination of money, time, and people. Thosein the field recognize that the investigation is important, work-ing at it with adequate intensity and ingenuity will provide an-swers to important questions, and the research will be great funfor the scientist. This is a very significant way of making newscientific discoveries, and there is much of science that fits sucha mold.

Another route to discovery involves work not recognizableto everyone as important and perhaps even regarded as ratherunimportant or infeasible by everyone except a few individuals.Even the few scientists who think work in the field is inter-esting may not have the right idea or approach initially. Never-theless, the number of important surprises or discoveries fromsuch situations is significant. Developments of this type do notnecessarily involve more innovation than the first, more easilyunderstood route. However, they do come on us more sud-denly and are more surprising, and we need to appreciate andbe sensitive to their importance. The more obviously importantlines of development for science may be carried out with a highlevel of imagination and skill. Nevertheless, it is generally notsuch a surprise if they achieve important results. The unex-pected development from an obscure or apparently mundanefield and the obviously important line of investigation are sim-ply different routes to discovery; the latter are much easier totalk and reason about but to overlook the "unexpected" routewould be wasteful and stultifying.

Particularly in discoveries whose very occurrence is difficultto foresee, an interesting and valuable interplay between pureand applied science is likely. This is partly because the unex-pected, happening in either pure or applied work, may con-tribute most importantly to some field other than its own. Dis-covery of radioastronomy or of the big bang radiation in thecourse of applied work are such cases. I hope I can appropri-ately refer again to the development of quantum electronics,one of my own interests. The development of a very intense,coherent, and exquisitely controllable light beam depended,surprisingly, on intensive research on the interaction betweenmicrowaves and molecules. This is because the laser grew outof the maser, being a particular kind of maser adapted to IR andoptical wavelengths. The maser itself came from microwavespectroscopy. The general idea was developed independentlyat three places, by Weber at the University of Maryland, byBasov and Prokhorov at the Lebedev Institute in the SovietUnion, and by myself at Columbia University. All of us wereworking in microwave spectroscopy. Microwave spectroscopywas in turn a somewhat surprising development that came outof interplay between pure and applied science. It began largelyin our industrial laboratories because that was the locale of mi-crowave technology. However, the industrial laboratories werenot terribly interested and hence such research soon moved tothe universities where it proceeded for some years before lead-ing to invention of the maser and laser. Particularly after in-vention of the laser, industry moved vigorously back into thefield of spectroscopy and quantum electronics and has done aremarkable job in subsequent developments. However, thedozen or so earliest contributors to the field were almost en-tirely scientists who had grown up in the field of radio and mi-crowave spectroscopy. At a later stage it was developed rapidlyand imaginatively in industry by somewhat different personnelincluding, of course, many engineers. This development is inturn now supplying high-quality instruments that are invalu-able to basic scientific research.

Applied Physical Sciences: Townes

Dow

nloa

ded

by g

uest

on

Aug

ust 1

5, 2

020

7682 Applied Physical Sciences: Townes

While individual scientists may at times be remarkably per-spicacious in evaluating a particular direction of scientific de-velopment, more generally the problem of foresight in researchand technological development needs to be approached with aclear recognition of the limits of our wisdom and with raremodesty. As support for this view, I want to comment on twostudies of the past that were directed toward foreseeing ourtechnological future. One was produced by a high-level com-mittee appointed by Roosevelt, which made its report in 1937on technological trends and national policy. In this report, thecommittee looked back at other predictions, commenting inparticular on those made in 1920 in the Scientific American,which undertook to predict the coming 75 years. The com-mittee's view was that the Scientific American's predictions werefairly good. Fifty percent of its predictions were said to havebeen correct, 25% were still doubtful, and only 25% were wrong.However, if one examines what was actually predicted, .I do notthink the results are especially impressive. On examining thisprediction, I noted in the Scientifwc American of the 1920s anarticle entitled "How science has now made gambling on pe-troleum prospects a thing of the past." Now to the predictions.

The specific predictions of 1920 made some very reasonablecomments-for example, in the field of genetics it said, "Weshall continue to pick up information as to how heredity works.-Doubtless the goal here is the human animal; whether that goalwill be obtained is a matter of guesswork." But it also said, onthe subject of extrasensory perception, "No careful person cancategorically deny the accumulating evidence that there is reallysome sort of communication between individuals widely sep-arated in space." With this kind of generality and batting av-erage, I am not highly impressed with a count of 50% cor-rectness. The 50% correct predictions tended to be things thatwere already under active consideration; it's what was missedthat is impressive. The misses included any mention, for ex-ample, of radio broadcasts (which came on stream with theopening broadcasts of KDKA only 1 month later) or of talkingpictures. However, predictions of the Roosevelt-appointedcommittee are more to the point because they were made byexperienced scientists and engineers of the type-the NationalAcademy of Sciences might assemble for this purpose.The report to President Roosevelt in 1937 was to allow sound

thinking about the future impact of science on society and wasdone by appropriate and well-known scientists and technolo-gists of the time. The Academy. had itself appointed anothercommittee that was interacting with the committee Roosevelthad appointed especially for this purpose, so that a still widergroup of scientific statesmen were represented. The resultingreport made a number of wise assessments; it said, forexample,reasonable things about agricultural research and the devel-opment of agriculture and about the development of better ro-tating machinery, which would make creation of electricity moreefficient. It also discussed synthetic gasoline from coal and syn-thetic rubber as having substantial potential. However, againwhat the report missed is more interesting. It missed almost allof the most exciting developments of the next few decades. Itmissed nuclear energy, though this possibility became obviousonly about 1 year after the report, with the work of Hahn andMeitner. There was no mention of the field at all but, if thecommittee did think about it, members might have referred toa statement a few years earlier by Lord Rutherford, who wasclearly the expert of the time. Rutherford said "Anyone whoexpects- a source of power from the transformation of these at-oms is talking moonshine." In the newspaper report of dis-cussion following Rutherford's statement, it was' essentially onlyE. 0. Lawrence who was brash enough to say that, while he didnot know how it could be done, there still might be a chance.

The committee missed antibiotics, which had been discoveredby Fleming some 8 years before that; Florey's work was only1 year in the future. It missed jet aircraft. It missed rocketry,space exploration, and any use of space. It missed radar. It missedcomputer development. It missed the transistor. It missedquantum electronics and, looking still a 'little further in the fu-ture, it missed genetic engineering.

If one lists the most interesting and exciting developmentsin technology-during the next decade or decades, the ideas thatthe committee missed approximate such a list.The above examples should give us all pause as wewrite our

committee reports. Having just finished participating in one onthe future of astronomy, I must share concern over the prob-able extent of our foresight. The problems with scientific de-cisions made by committees are, I believe, quite deep. Peerreview seems to me to have serious problems with respect toinnovation. Committees tend to be conservative as a group andto follow established routes. These may be good routes and re-sult in good science. However, for many of the developmentsthat bring important surprises one needs the spark of occa-sionally crazy enthusiasm that comes from an individual. Some-one needs to 'have the conviction that a new approach is rightwhen others, perhaps even the community as a whole, are-skeptical. Wise committee members, particularly senior states--men, all too easily recall "Yes, I thought of that 15 years ago;I looked at it and it's not going to work," without recognizingthat there are new aspects to the idea that may allow its success.

The obvious and difficult problem in supporting offbeat ideasis how to sort out the real nuts from the apparent nuts, becausethere are plenty of people who are enthusiastic about some-thing with which no one else agrees. To this problem, there isno straightforward nor completely reliable solution. We mustsearch for ways in our national planning to see that there is enoughlooseness to allow both error and support of promising but un-certain research. Some of the necessary looseness is, unfor-tunately, much dependent on the supply of money and, at pres-ent, that is especially difficult. As the supply of money becomestighter, committees and decision makers tend to become moreconservative, feelingobligated to put the limited supply of moneyon those things that are most clearly going to pay off. They alsoknow that money for their own personal work is in short supply,which produces an almost unavoidable tendency toward con-servatism. When financial support is more abundant, there ismore taste for directing some money toward the longer re-search chances or less obviously rewarding approaches. Thisdoes not mean, however, that such a policy is necessarily lessshortsighted 'when money is tight than when it is plentiful.

Summary

By way of summary, I have a short list of some social requisitesfor scientific and technical innovation.

First on my list is a general interest throughout society inintellectual ideas and a sense of the excitement and value ofdiscovery.

Second is a diversity of approach, a diversity of types of in-stitutions, and of people. This means support of research, typesof research, and individuals that are not obviously in the main-stream of important science.

Third, along with diversity and widespread diffusion of ef-fort, there must be at least a modest number of institutions wherethe research environment is very intense. That is, there mustalso be concentrations of excellence, implying concentrationsof skillful people who maintain high standards and interact ina supportive and inspiring way.

Proc. Natl. Acad. Sci. USA 80 (1983)

Dow

nloa

ded

by g

uest

on

Aug

ust 1

5, 2

020

Proc. Natl. Acad. Sci. USA 80 (1983) 7683

We need the diversity that comes from research efforts in awide variety of institutions with various traditions and ap-proaches. However, we cannot rely on diversity alone. Theremust also be institutional peaks that are particularly creativeand may well appear to dominate the scene. To combine thiswith diversity requires a particular overt openness with respectto funding so the less well-recognized and the offbeat work canbe supported. This may require overlap of the charters of sev-eral sources of funding. As an example, I am skeptical that thenation would be well served if the National Aeronautics andSpace Administration and the National Science Foundation wereto map out too carefully the type of astronomy each should ex-clusively support rather than allowing substantial overlap. Ifonly one organization is responsible for making decisions in agiven field, over a long period of time there can easily be a stag-nation of ideas and approaches. More than one kind of au-thority to which researchers can appeal when turned down im-proves the chances of breaking down overly rigid decisions.My fourth condition is the encouragement of interaction be-

tween sciences and, in this interaction, I would emphasize theinteraction between pure and applied scientists. Easy and com-

mon contacts between individuals in universities and betweenscientists in universities, industry, or government establish-ments is an important phenomenon that needs our attention.Unfortunately, during the recent past our universities, indus-try, and governmental establishments became badly separated.They are now slowly becoming reconnected. This brings somedifficulties and conflicts of goals but also great values. One ofthe more obvious present examples of this important aspect ofour scientific and technological success is in bioengineering,where both the value and the uncertainties in connections be-tween academia and industry are obvious. One can regret thedistraction that such interactions with the larger society bringsto universities. One can regret the administrative complica-tions. One can perhaps even regret that there are concomitantopportunities for personal wealth that may stand in the way ofnormal scientific values and procedures. Nevertheless, suchnew drives and new connections will no doubt be very fruitfulfor innovation and, in the long run, the nation and its sciencewill benefit from maintaining and encouraging such interac-tions between the nation's wealth of scientists, pure and ap-plied.

Applied Physical Sciences: Townes

Dow

nloa

ded

by g

uest

on

Aug

ust 1

5, 2

020