R E F E R E N C E F R A M E

Why Do They Leave Physics?Philip W. Anderson

I would like to quote for you from anapplication letter that one of my stu-

dents sent to a well-known manage-ment consulting firm after completinghis second postdoc and contemplatingyet another as his most probable nextstep.

There are several reasons forwanting to leave academia, notthe least being the desire for a. . . work environment which . . .rewards those who take appro-priate risks. . . . The problems Ihave been attracted to . . . aremotivated by real world datawhere there is no . . . existingtheoretical model. Involvementin such projects . . . requires adesire for new challenges, anappetite for risk, and the abilityto be innovative yet humble inthe face of... data.

You will be pleased to know that theapplication succeeded and that thewriter is now happily earning as muchas I did at the end of 50 years inphysics. We are experiencing a seriousbrain drain in physics, with condensedmatter theory taking some of the worsthits. Many of my colleagues must haveseen similar letters.

The obvious simple answers to thequestions raised by this letter do notseem to be viable:O Is it simply that the less able andimaginative are being weeded out? No,quite the opposite, in my experience:Those who leave are in many cases thecream of the crop. I hear this againand again from many different men-tors. I find many of those hired forpermanent jobs in physics to be amongthe least creative.0 Is there a serious shortage of per-manent jobs? Well, to some extent, but1 hear again and again of universitiesthat initiate searches at the tenure-track level but postpone the hiringdecision indefinitely.t> Can we blame Wall Street for se-ducing our best and brightest withenormous financial temptations? Inthe first place, the jobs are not exclu-

PHILIP ANDERSON is a condensed mattertheorist, who is the Joseph Henry Professorof Physics Emeritus at PrincetonUniversity in Princeton, New Jersey.

sively in finance but in a wide varietyof fields; in the second, it is my expe-rience that most of our lost physicists—certainly the one I quote above—havebeen willing to endure considerablehardship to stay in physics if they canbe assured of a reasonably permanentjob at an acceptable salary.

Frankly, I think that the youngwriter of this letter has put his fingeron the problem. "A work environmentwhich rewards those who take appro-priate risks" and the desire for "newchallenges" are what thrilled me aboutscience many decades ago, and whatwe no longer have in much of academicphysics.

It would be easy to blame the Na-tional Science Foundation and theother funding agencies, or the contrac-tion of fundamental science in industry,for this situation. Both contribute; NSFhas become steadily more bureau-cratic, requiring more rigidly format-ted proposals and more unanimityamong referees, and there is a newemphasis in industry on applicable re-search. But industry is hiring youngpeople, and my impression is that it isdoing a better job in pinpointing crea-tivity than are academia or govern-ment laboratories.

My own diagnosis is pretty muchidentical to the ones implied in theletter. The members of our profession—and particularly those who have re-sponsibility for hiring and for fundingresearch—are infected with "Horgan-ism," the belief that the end of science(or at least of our science) is at hand,and that all that is left to do is to grubaway at Kuhnian "normal science" fol-lowing the accepted paradigms. Theybelieve that there are no more scientificrevolutions possible, and that we are

now in pursuit of nothing but the nextdecimal place—hence, by the way, thefunding prejudice in favor of heavycomputer use and the existence of theoxymoron "computational physics."

In such a world, as in all "normal-science" periods, the institutional re-sponse that is occurring would makesense. Any proposal, to be funded,should have essentially unanimouspeer approval. Any new appointmentshould have unanimous approbationfrom all senior figures; any such seniorfigure who disagrees with the consen-sus is bound to be a crackpot and maybe ignored. When, in fact, it turns outthat in the real world, physics, evencondensed matter theory, is full of cri-ses and controversies, and reputablesenior scientists have deep disagree-ments, the administrator throws uphis hands in bewilderment and post-pones filling the slot. Another optionis to settle for the most orthodox-seem-ing appointee, who has the most easilycomprehensible (read "simple-minded")program in mind. Finally and worst ofall, the administrator feels constrainedto assure himself that the candidate is"fungible"; that is, that he or she canattract the requisite unanimity—fiveout of six "excellents" from the NSFreferences. (The flexibility of our fund-ing system is a thing of the past, now,with other agencies tending to followNSFs lead rather than to think forthemselves; this practice may be asbad as following the idiosyncratic andoften fallible judgment of the grantofficer himself.)

We must not let The End of Sciencebecome a self-fulfilling prophecy. Thereason that Horgan's pessimism is sowrong lies in the nature of scienceitself. Whenever a question receives ananswer, science moves on and asks anew kind of question, of which thereseem to be an endless supply. The kindof people we most need are not thosewho are good at answering well-posedold questions, but those who are capa-ble of posing new ones. The best wayto prevent the end of science is toprovide opportunity in abundance forthe most creative and original of ouryoung people. This is not happening.But it needs to. •

© 1999 American Institute of Physics, S-0031-9228-9909-210-8 SEPTEMBER 1999 PHYSICS TODAY 11

© 2000 American Institute of Physics, S-0031-9228-0001-220-5 JANUARY 2000 PHYSICS TODAY 15

I enjoyed reading Philip Anderson’s“Reference Frame” column in your

September 1999 issue (page 11). Thesituation facing young scientistsneeds continuing attention if we are to keep the best and brightest of them in the field.

I agree wholeheartedly with An-derson’s answers to the column’stitle question: “Why Do They LeavePhysics?”: Horganism (that is, thebelief propagated by science writerJohn Horgan that there are no newscientific laws to discover) is a dan-gerous recurring infection in the sci-entific community, one that most cer-tainly predates Horgan himself. Andit is undoubtedly true that the cur-rent funding paradigm does little topromote creative, curiosity-drivenresearch.

However, I believe there is alsoanother causal factor that has led tothe relative ease with which theyoung, talented members of our com-munity leave, a sociological one thatemerged with the end of the coldwar. Government funding cuts, in-dustry downsizing, the influx of tal-ented foreign scientists, and othercontemporaneous events led to irre-versible changes throughout sciencein the early 1990s and shattered thedreams of many young scientistswho had promising scientific careersahead of them. Not only could theynot get the kinds of jobs they want-ed, they could barely find any jobsfor which they were appropriatelyqualified. Many left physics withremorse and reluctance to take jobson Wall Street, in management con-sulting, or with what would now becalled Internet start-ups.

Although those employment diffi-culties have eased considerably, thesociological paradigm shift (to abusea Kuhnian term) that resulted re-mains with us. Graduate studentswho began their studies in the 1990s

came to physics with their eyes wideopen about their job prospects. Nolonger did they dream of eventuallywalking in the shoes of their thesisadvisers. They were content with the idea of doing some interestingresearch for a few years and thenmoving on to careers elsewhere. Therecruiting of graduate students alsochanged gears: Aware that applica-tions for admission were falling,especially from Americans, manydepartments began trying to sellgraduate physics education as goodtraining for many nonscientificcareers. Consequently, graduate education has shifted from a kind of guild/apprenticeship model tosomething more akin to preprofes-sional training. In many graduateprograms, students are offered non-credit courses to enhance their com-munications skills. Many of themstep out of the lab a couple of nightsa week to take computer science ormarketing classes. Few are still deadset on the idea of a career in physics.

I do not view this shift as a badthing. It was necessary, given thecircumstances, and the apprentice-ship model was doomed to implode, a Ponzi scheme that presupposedsustained exponential growth inresearch. But perhaps we as a com-munity have not fully made thetransition. Graduate students inmany departments and researchgroups are still often treated as sec-ond-class citizens, as cheap laborwhose every waking hour should bedevoted to research. But in the cyni-cal and savvy 1990s and in the midstof a record economic boom, that is asurefire way to drive away our mosttalented students.

So, although I agree with Ander-son that we need to provide creativeopportunities for young physicists asan incentive for them to stay in thefield, we also need to address the cul-ture of physics (and of all of science),to create a community in which theycan feel welcome, encouraged, andrespected.

JONATHAN R. FRIEDMAN(jonathan.friedman@sunysb.edu)

State University of New Yorkat Stony Brook

Philip Anderson’s essay containsthe seeds of the exact problem

that he is concerned about—namely,that too many of the best and bright-est are leaving physics. In fact, heassumes that physics is currentlybeing practiced mostly within theconfines of academia and govern-ment laboratories, and presumablywithin only a few narrow specialties.Unfortunately, the same attitudeseemed prevalent last March at theAmerican Physical Society’s centen-nial meeting, where there were vir-tually no symposia covering suchtopics as fluid dynamics or optics.Recently, though, the special every-day-physics issue of PHYSICS TODAY(November 1999) has served as awelcome tonic in this regard.

The fact is, physics is where youfind it, be it in academia, industry,or a government lab. It is unfor-tunate that there are people whowould segregate physics and have itbe a small, parochial, and inapplica-ble branch of science. If the goal ofphysics is truly to explain how andwhy the world (or entire universe) iswhat it is, then it is a fool’s game torestrict physics to a vanishinglysmall subset of subjects to study.

Anderson and I agree on twopoints—that financial compensationin science is significantly higher out-side of academia, and that industryis hiring young people while acade-mia is not. Where we disagree is onwhether this situation adverselyaffects the advance of physics (or ofall science). In a free market, talentgoes where the money is. Society hasprogressed to the point where thecontinuing advancement of technolo-gy is crucial to future development,and thus there are commercial re-search programs willing to pay ahigh price to bring in good talent.Therefore, it is ludicrous to claimthat commercially funded work in,say, the semiconductor industry, theoil industry, or the telecommunica-tions industry is detracting from theadvancement of physics. It may besatisfying to blame others—namely,those with responsibility for hiringand for funding research—but thatbegs the question as to how many

Letters submitted for publication shouldbe sent to Letters, PHYSICS TODAY,American Center for Physics, OnePhysics Ellipse, College Park, MD 20740-3843 or by e-mail to ptletter@aip.org(using your surname as “Subject”).Please include your affiliation, mailingaddress, and daytime phone number. Wereserve the right to edit letters. continued on page 75

L E T T E R S

More on ‘Why Do They Leave Physics?’:Money Matters, Research and Job Opportunities

LETTERS (continued from page 15)

JANUARY 2000 PHYSICS TODAY 75

young researchers the physicsdepartment at, say, Princeton Uni-versity has recently hired (on atenure-track level) and how muchfunding the department receivesfrom, say, the National ScienceFoundation.

ANDREW RESNICK(aresnick@cleveland.dynacs.com)

Dynacs Engineering CoBrook Park, Ohio

Although I enjoy well-taken (andwell-deserved) jabs at National

Science Foundation funding proce-dures and computational physics, Ihappen to think that Philip Ander-son has missed the core issue intoday’s physics (and science in general): money.

Overhead, as a form of virtuallyunrestricted funds, is an object high-ly coveted by university administra-tions. Consequently, the entire re-search infrastructure has been over-hauled to suit the purpose of secur-ing it. As a result, the ambiance ofphysics has changed such that, inlieu of scholarship, scientific achieve-ment, and reputation, we are nowassessed in terms of four so-calledobjective criteria: number of paperspublished, prestige of the journalsinvolved, number of invited talksgiven, and amount of grant moniesreceived. (One of the symptoms of this shift is the explosion in thenumber of papers that get intoprint—as reflected in the fact thatthe heft of Physics Review Letters isnow approaching that of PhysicalReview as it was a few years back,and yet, in condensed matter theory,which happens to be Anderson’sfield, progress in the last 15 years or so hardly deserves such a vol-uminous output.)

In short, one shouldn’t expectadventurous talent to be enamoredof and stay in a system run by bu-reaucrats and dominated by what I’ll call operators. The underlyingproblem is not in our minds and atti-tudes, as Anderson seems to believe.Rather, it is the corruptive influenceof federal funding, and it is very real.

ROSTISLAV A. SEROTA(serota@physics.uc.edu)

University of CincinnatiCincinnati, Ohio

Philip Anderson’s essay is a goodcommentary that outlines a real

problem, but I think the answer isclear. Twenty years ago, physics wasexciting, but now—largely as a result

of all the work that has been doneusing government funding to addressmany first-order questions—it isn’t.Thus the reason for studying physicstoday shouldn’t be mainly prepara-tion for doing physics research (thatis, for learning more about nature),but rather it should be in further-ance of acquiring a modern general,or liberal arts, education. So the sensible thing to do after getting aphysics PhD is to go on to systemsanalysis or the world of finance orsome other segment of industry orcommerce. After all, as Andersonpoints out, the better students doleave the academic world! Granted,one spell as a postdoc may be neces-sary to convince the graduate thatit’s time to move on, but it’s best notto linger. (Some 20 years ago, as aphysicist working for a major thinktank in the Washington, DC, area, Igot a call from a professor in a first-rate physics department: They weregetting a student out with a PhDafter 11 years, and would we be in-terested in interviewing him? I saidno; if this man was willing to workfor starvation wages as a graduatestudent for that long, he surely did-n’t have the attitude we were lookingfor—one conducive to grappling withreal-world problems under the in-evitable time constraints.)

When a British acquaintance ofmine got a classics degree from theUniversity of Oxford some 30 yearsago, he became a civil servant, know-ing that nobody was going to payhim to write classical Greek poetry.One of his first assignments was tolay out and oversee land purchasesfor a superhighway linking the Lon-don area with Cambridge. Thatmeant having to tackle an array ofcomplex problems and also cope withserious constraints of time, money,and so forth. In other words, he hadto climb out of his professional sand-box and find ways of putting his edu-cation to work in dealing with somelarge-scale practical challenges. Helooked to specialists, such as survey-ors and purchasing agents, and hisfunction was to integrate their workso as to provide an acceptable overallsolution. He succeeded.

Similarly, I suggest, the presentfunction of a physics PhD should beto obtain a thorough general educa-tion and manifest a willingness totackle societal problems, rather thancontinue to play in the government-funded research sandbox.

ERNEST BAUER(ernestbaue@aol.com)Bethesda, Maryland

PHILIP ANDERSON REPLIES: I appre-ciate these letters and also the

many thoughtful ones that were sentto me directly, some of which con-tained positive proposals for dealingwith the situation and most of whichwere surprisingly (to me) supportive.But I would like to say a few wordsin response to the letters chosen forpublication.

I agree with Jonathan Friedmanthat the “cloning” sociology that grewup in the time of expansion wasunfortunate in many ways. I nevercaught it since my apprenticeshipwas served in England, where mostof my first students went on to jobslike road research or Israeli intelli-gence; academic or real research jobswere very scarce. Neither the stu-dents nor I seemed bitter, since wefelt our time together had been wellspent, and when I got a few genuine-ly original and imaginative students,there was not much difficulty inplacing them even in those scarcejobs, since there were people aroundthe world who were willing, eveneager, to take a risk with them. Suchrisk taking, in my diagnosis, is whatis now missing.

I find it strange to be accused byAndrew Resnick of academic snob-bery when I have been such an advo-cate of “the opening to complexity”(to quote the title of a symposium Iorganized) and of the “relevance,” atleast intellectual, of science. I amaware of much good science done forpractical ends, and I think the recordof Bell Laboratories, where I spent a35-year career, abounds with exam-ples. But in the end there are sub-jects in which the open-endedness ofacademic research is essential. Thecompanies, for instance, that standto gain the most from breakthroughsin cuprate superconductors are theleast willing to sustain the long-termeffort necessary for understandingthem—but perhaps that is a properfunction of government (after all, ourbasic understanding of the strengthof materials did not come from USSteel). Of course, the young men andwomen who join entrepreneurialcompanies are not wasting theirlives, but we need to keep enough ofthem in pure research and teachingto allow for renewal—and I will notgive up the principle that theyshould be the best and most creative.

Again, with Rostislav Serota Ihave a lot in common. Far be it forme to defend the overhead system oruniversity administrations. But Idon’t see them as quite as greedy ashe does. Perhaps he doesn’t know

78 JANUARY 2000 PHYSICS TODAY

that university overheads are lessthan half of those in industry. Gov-ernment agencies, particularly theNational Science Foundation, tryvery hard with minimal resources.Otherwise, I think he is simply rein-forcing my point.

Ernest Bauer seems to havecaught the Horgan infection. Ofcourse the questions that were cur-rent when we were in graduateschool have been mostly answered—as have, in my case, the ones afterthose. But science keeps throwing upnew questions, and it takes imagina-tion to see they are questions. I’malso a little tired of those Britishclassics students—they got the UKinto two world wars, messed up theBalkans permanently, and left theBritish people with four or five dif-ferent kinds of electric light plugs.At least physicists might have donebetter with the light plugs. Physics—unlike the classics—is a good educa-tion for such jobs as managementconsulting, and I don’t want to dis-courage those who take naturally tosuch careers. However, I deploreleaving only the least-imaginativeindividuals in physics to teach future generations.

PHILIP W. ANDERSONPrinceton University

Princeton, New Jersey

Fubini and SalamInspired Plans forMideast Synchrotron

The lead story of your “PhysicsCommunity” section in the Aug-

ust issue (page 54) is Toni Feder’saccount of new and hope-filled initia-tives for peace in the Middle East,based on scientific collaboration.Specifically, she tells of the plan forthe region to be given a decommis-sioned German synchrotron lightsource that would become the center-piece of a major international re-search facility—site still to be deter-mined—that would function as acenter where scientists from previ-ously hostile countries would inter-act, much as CERN brought togetherformer enemies after World War II.

Although Feder does discuss thevital role played in this plan by theMiddle East Science Collaboration(MESC), as well as by various indi-viduals, she makes no mention ofSergio Fubini, who conceived thefirst Sinai Meeting on Physics, heldin Dahab, Egypt, in November 1995(see PHYSICS TODAY, February 1996,page 11), which led directly to the

founding of MESC in 1997. Fubinihas remained an active and essentialparticipant in MESC and its pushfor more science in the Middle East,as reflected in what happened at theParis meeting hosted by the UnitedNations Educational, Scientific andCultural Organization last summer.UNESCO’s director general FedericoMayor stated, “I wish to pay tributeto a group of scientists from the Mid-dle East and Europe who have work-ed together for five years now inorder to promote scientific coopera-tion in the Middle East and the Med-iterranean region. . . . We are fortu-nate to have with us today the per-son who initiated this movement ofscientists and did more than anyoneelse to make it the success it now is.He is Professor Sergio Fubini fromthe University of Turin and CERN.”

ROMAN JACKIW(jackiw@mitlns.mit.edu)

Massachusetts Institute of TechnologyCambridge, Massachusetts

One of the key individuals wholaid the groundwork for what

may turn out to be the Middle East’sfirst synchrotron light source and amajor international scientific researchcenter was Abdus Salam, a cowinnerof the 1979 Nobel Prize in Physics,the founder and long-time director ofthe International Centre for Theoret-ical Physics (recently renamed in hishonor), and “a humanitarian whodevoted much of his life to upliftingthe status of science and technologyin the third world” (PHYSICS TODAY,August 1997, page 75).

As noted in PT, Salam, born inwhat was then part of British-ruledIndia and is now Pakistan, “dreamedof creating 20 international centerslike the ICTP, spread throughout theworld.” As part of that vision, he ac-tively promoted the idea of advan-cing the cause of science and technol-ogy in the Middle East, not only byhaving researchers from the regionwork with their colleagues in thedeveloped world, but also by havingthe region develop its own facilities—including a synchrotron laboratory.

In May 1983, at the Symposiumon the Future Outlook of the Arabi-an Gulf University held in Bahrain,Salam delivered a paper entitled“The Gulf University and Science inthe Arab-Islamic Commonwealth,” inwhich he reminded his listeners that“We forget that an accelerator likethe one at CERN develops sophisti-cated modern technology at its fur-thest limit. I am not advocating thatwe should build a CERN for Islamiccountries. However, I cannot but feel

envious that a relatively poor coun-try like Greece has joined CERN, pay-ing a subscription according to thestandard GNP formula. I cannot re-joice that Turkey, or the Gulf coun-tries, or Iran, or Pakistan seem toshow no ambition to join this fount ofscience and get their men catapultedinto the forefront of the latest techno-logical expertise. Working with CERNaccelerators brings at the least thisreward to a nation, as Greece has hadthe perception to realise. . . .” 1

He then went on to make the fol-lowing points: “I have mentioned aninternational laboratory in materialsciences for Bahrain, with specialisa-tion in microelectronics and modernelectronic communications, includingspace satellite communication, tohelp also with the banking communi-cations needed at Bahrain. Such alaboratory was in fact proposed forthe University of Jeddah. The ideawas to emphasise science transfer inaddition to technology transfer andto create international laboratoriesin the fields of materials sciences,including surface physics and a labo-ratory with a synchrotron radiationlight source. The facilities createdwould have been of the highest pos-sible international order; the labora-tories would have been opened toteams of international researchers,who would congregate and work atJeddah, just as they congregate nowat the great laboratories in Ham-burg, Geneva or Paris. . . .” 2

Perhaps, in memory of Salam, Jed-dah (Jidda) in Saudi Arabia or a suit-able location in Bahrain should beadded to the list of possible sites forthe relocated German synchrotron.

References1. A. Salam, in Renaissance of Sciences in

Islamic Countries—Muhammad AbdusSalam, H. R. Dalafi, M. H. A. Hassan,eds., World Scientific, Singapore (1994),p. 45.

2. Salam, p. 53.SAMEEN AHMED KHAN

(khan@pd.infn.it)University of Padua

Padua, Italy

CorrectionNovember 1999, page 83—In PiotrWasiolek’s letter, a minus sign wasomitted from the conversion sequencefor the average indoor level of radon.The correct sequence is 1.3 pCi/L =1.3 × 10–12 Ci/L = 0.048 Bq/L =0.048 Bq/(0.001 m3) = 48 Bq/m3.

December 1999, page 59—CharlesTownes’s book is $29.95, not $49.95.�