by Lee Smolin
As there are not many such people, it should not be hard to make room for them in the academy. Indeed, you would think that many institutes, colleges, and universities would be happy to have such people. Because they think clearly about the foundations of their subjects, they are often good, even charismatic teachers. Nothing inspires students like a seer on fire. Because they are not competitive, they are good advisors and mentors. After all, isn’t the main business of colleges and universities to teach?
Of course, there is a real risk. Some of them will not discover anything. I am talking in terms of a real lifetime contribution to science. But then most academic scientists, though they succeed in career terms—get grants, publish a lot of papers, go to a lot of conferences, and so on—contribute only incrementally to science. At least half our colleagues in theoretical physics fail to make a unique or genuinely lasting contribution. There is a difference between a good career and an essential career. Had they done something else with their lives, science would have gone on much the same. So it’s a risk either way.
The nature and costs of different kinds of risk are issues that businesspeople understand better than academic administrators. It is much easier to have a useful and honest conversation about this with a businessperson than with an academic. I once asked a successful venture capitalist how his company decided how much risk to take on. He said that if more than 10 percent of the companies he funded made money, he knew he was not taking on enough risk. What these people understand, and live with, is that you get overall a maximal return, which maps to a maximal rate of technological progress, when 90 percent of new companies fail.
I wish I could have an honest conversation about risk with the National Science Foundation. Because I’m sure that 90 percent of the grants they give in my field fail, when measured against the real standard: Do those grants lead to progress in science that would not have occurred if the person funded did not work in the field?
As every good businessperson knows, there is a difference between low-risk/low-payoff and high-risk/high-payoff strategies, starting with the fact that they are designed with different goals in mind. When you want to run an airline or a bus system or make soap, you want the first. When you want to develop new technologies, you cannot succeed without the second.
What I wouldn’t give to get university administrators to think in these terms. They set up the criteria for hiring, promotion, and tenure as if there were only normal scientists. Nothing should be simpler than just changing the criteria a bit to recognize that there are different kinds of scientists, with different kinds of talents. Do you want a revolution in science? Do what businesspeople do when they want a technological revolution: Just change the rules a bit. Let in a few revolutionaries. Make the hierarchy a bit flatter, to give the young people more scope and freedom. Create some opportunities for high-risk/high-payoff people, so as to balance the huge investment you made in low-risk, incremental science. The technology companies and investment banks use this strategy. Why not try it in academia? The payoff could be discovering how the universe works.
19
How Science Really Works
THE IDEA OF CHANGING the way science is done in universities will no doubt appeal to some, while horrifying others. But it’s probably in no danger of happening. To explain why, we need to inspect the dark underbelly of academic life. Because, as the sociologists tell us, it is not just about wisdom, it is about power: who has it, and how it is used.
There are certain features of research universities that discourage change. The first is peer review, the system in which decisions about scientists are made by other scientists. Just like tenure, peer review has benefits that explain why it’s universally believed to be essential for the practice of good science. But there are costs, and we need to be aware of them.
I’m sure the average person has no idea how much time academics spend making decisions about hiring other academics. I spend roughly five hours a week in committees discussing other peoples’ careers or writing letters to be read by such committees. I have been doing this for some time. It is a major part of the job of a professor, and many professors I know spend more time on it than I do. One thing is sure: Unless you embarrass yourself by a visible show of irresponsibility or prove yourself too unpredictable or too untrustworthy, the longer you are a scientist, the more time you will spend meddling in the careers of other scientists. It’s not just that you will have more and more students, postdocs, and collaborators who need letters written for them; you are also involved in the hiring decisions of other universities and institutes.
Has anyone in administration ever studied how much this system costs us? Is it really necessary? Could we spend less time on this and have more time for science and teaching? I have had only a small taste of the system, and it is daunting. No department or institute with aspirations can hire without consulting a network of visiting and advisory committees staffed by influential older scientists from other institutions. There are also panels set up by funding agencies in the United States, Canada, Europe, and around the globe. Then there are all the informal contacts, telephone calls, and conversations in which you are asked to frankly and confidentially evaluate lists of candidates. After a certain point, a successful scientist could easily spend all of his or her time on the politics of who gets hired where.
This is called peer review. It’s a funny name, because it differs markedly from the notion of a jury of one’s peers, which suggests that you are being judged by people just like yourself, who are presumably fair and objective. There are real penalties—prison—for jurors who conceal a bias.
In the academic world, with few exceptions, the people who evaluate you are older than you are, and more powerful. This is true all the way up the ladder, from your first college course to the applications you make for grants when you’re a professor. I do not want to disparage the hard work done by so many in the service of peer review. Most do it sincerely. But there are big problems with it, and they are relevant to the state of physics today.
An unintended by-product of peer review is that it can easily become a mechanism for older scientists to enforce direction on younger scientists. This is so obvious that I’m surprised at how rarely it is discussed. The system is set up so that we older scientists can reward those we judge worthy with good careers and punish those we judge unworthy with banishment from the community of science. This might be fine if there were clear standards and a clear methodology to ensure our objectivity, but, at least in the part of the academy where I work, there is neither.
As we have discussed in detail, different kinds of scientists contribute to theoretical physics, and they all have different strengths and weaknesses. There is, however, little acknowledgment of this; instead, we speak simply about who is “good” and who is “not good”—that is, peer review is based on the simplistic and obviously faulty assumption that scientists can be ranked as on a ladder.
When I read my first batch of letters of recommendation as a new assistant professor at Yale in 1984, I couldn’t believe it. While a good letter might convey a lot of information and make some attempt at nuance, a great deal of attention was given to the final paragraph, which usually offered a comparative ranking of the candidate: “X is better than A, B, and C, but not as good as E, F, G, and H.” I’ve read thousands of letters of recommendation by now, and at least half contain some version of that sentence. In the early years, there were a few instances when I was person A, B, or C. I recall agreeing that candidate X was indeed better than I was—and in fact some of those Xs have gone on to do very well. But I would not be surprised if research showed that these rankings are, on average, poor predictors of genuine success in science. If we were really concerned about making good hires, we would carry out such research. Certainly, there are not a few cases in which the top-ranked postdoc or assistant professor does not end up doing much and doesn’t get tenure.
What makes this practice even more problematic is that there are no
sanctions for bias. A professor will shamelessly write letters slanted toward his or her own students, or for people who are following his or her particular research program, or even for people of his or her own nationality. We may notice (and laugh about) the really blatant instances, but no one thinks it unusual. It’s just part of the system.
Here’s a basic rule for predicting the kind of junior scientist that senior scientists will recommend: Does the junior scientist remind them of themselves? If you see a younger version of yourself in X, then X must be really good. I know I am guilty of this, and I say so frankly. If you want to hire more people like me, I am great at picking them out. If you want to make fine distinctions among people very different from me, who are good at things I am not so good at or don’t value, don’t trust my judgments.1
Even for those of us who strive to be fair, there is no guidance or training on how to be objective. I have never been given any advice about how to write or interpret letters of recommendation, nor have I ever seen guidelines on how to recognize signs of prejudice or stereotyping in your own or others’ views. I have served on many committees for hiring, tenure, and promotion, but I have never been instructed, as jurors are instructed, on how to best weigh the evidence.
Once at a dinner party I asked people in other lines of work if they were trained in such matters. Everyone who was not an academic but who had responsibility for hiring or supervising other people had been given several days of training in recognizing and combating signs of unfairness or prejudice, in discounting the effects of hierarchy, and in encouraging diversity and independence of thought. They knew all about “heeding all the voices in your organization” and getting “a 360-degree look at job candidates” by seeking evaluations from people the candidates had supervised as well as from people who had supervised the candidates. If lawyers, bankers, television producers, and newspaper editors are assumed to need guidance in how to make wise and fair personnel decisions, why do we academic scientists assume we can do it automatically?
It is even worse than that. Behind the formal letters of recommendation is a network of confidential, informal conversations with experts: “What do you think of so and so? Whom do you think we should hire?”
These conversations are frank. The gloves are off. And this is not all bad. Many people try to rise to the challenge and be helpful, but the average level of objectivity is shockingly low. And here, especially, there is no price to pay for gaming the system in favor of your friends and the students of your friends. It is common for established experts to push their own students and postdocs, praising them unreasonably over others, especially the students of rivals.
Even in these frank exchanges, you seldom hear really negative comments. When people have nothing good to report, they will often just say, “Let’s move on. I’d rather not comment” or something mild like “I’m not excited.” But there are times when the mere mention of a name invokes an “Absolutely not!” or “Don’t go there” or “Are you kidding?” or the definitive “Over my dead body!” In my experience, in every such instance the candidate fell into one and often two of the following three categories: They were (1) female, (2) not white, and/or (3) someone inventing his or her own research program rather than following the mainstream. There are of course women and nonwhites who elicit no objections. But, again in my experience, these are cases where the candidate hews tightly to an established research program.
There is heated debate among physicists over why there are not more women or blacks in physics, compared with other fields just as challenging, such as mathematics or astronomy. I believe the answer is simple: blatant prejudice. Anyone who has served, as I have, on decades of hiring committees and hasn’t seen naked prejudice in action is either blind to it or dishonest. There are rules and ethics of confidentiality that prevent me from giving examples, but there are several detailed studies that tell the story.2
Perhaps it’s to be expected that prejudice is fierce in this field. How many leading theoretical physicists were once insecure, small, pimply boys who got their revenge besting the jocks (who got the girls) in the one place they could—math class? I was one of these, at least until I figured out what the jocks knew—that it is all about confidence. But I still recall feeling smug about my abilities in algebra, and I can report that, at least for me, the identification of math skills with maleness runs very deep. But then, why do women have less difficulty getting hired as pure mathematicians than as physicists? Because it is clearer in mathematics when you have done something good. A theorem is either proved or not proved, while the judgments that go into ranking theoretical physicists are much more diffuse, which gives more room for bias. It is not always easy, for example, to distinguish a good theorist from one who is just assertive. Note that whereas there have always been talented women musicians, the number of women hired by orchestras rose significantly when candidates began auditioning behind a screen.
This is why there is affirmative action. In all my experience, I have never seen a woman or an African American hired through an affirmative-action program who didn’t strongly deserve it—that is, who wasn’t already arguably the best applicant. When hiring committees are no longer composed only of white men and we stop hearing expressions of open prejudice, then we can relax affirmative action. As it stands, people who are different—who, for one reason or another, make powerful older male physicists uncomfortable—are not hired. There is affirmative action for people who are visibly different, like women and blacks. But what about people who just think differently—who reject mainstream approaches in favor of their own ideas? Should there be affirmative action for them, too?
Many of us participate in peer review with the best intentions of choosing ethically and objectively. And when all else is equal, the more deserving candidate is chosen. That is, when you get down to comparing white men of about the same age and background, who are all pursuing the same research program, the system will generally pick out the one who is cleverest and works the hardest. But the problem is that you have to do a lot of winnowing before you reach the point where everything is equal. Up until that point, the process is political. It is the primary mechanism by which older and more powerful scientists exert power over younger scientists.
This makes for a process of enforced consensus, in which older scientists ensure that younger scientists follow their directions. There are some simple ways in which this power is exerted. For example, a candidate for a faculty position needs letters from a great many people, all more powerful than the candidate. One not-so-positive letter will normally kill an appointment. When I first encountered this plethora of recommendation letters, I was puzzled. Certainly you could get a good picture of a candidate from three or four letters. Why ten or fifteen, the number often required by highly ranked universities?
One reason is that the goal is not only to hire good scientists. Hiring committees, chairs, and deans often have another goal in mind, which is to raise (or in fortunate cases, preserve) the status of the department. By this I mean something more measurable than a young scientist’s promise, for measures of status are given by numerical rankings. These are made by external evaluators, who combine their impressions with numbers like total grant funding and numbers of citations. Department chairs and deans have to be concerned with this, because such matters have brute financial repercussions relevant for their own careers as administrators. It is, first of all, important to hire people who are likely to win generous grant support. This immediately favors members of large established research programs over initiators of new programs. By asking for many letters, you can measure how well the potential hires are already perceived by senior scientists who matter. The goal, then, is not to hire the scientist most likely to do good science but the scientist whose acquisition will optimize the status of the department in the short term. This is why hiring committees don’t agonize over long-term issues, like which candidate is most likely to have original ideas that will matter in twenty years. Instead
, they want to know that ten to fifteen senior scientists think the candidate is a high-status member of their community.
But to elicit such a large number of positive letters, you have to be part of a large research program. If you are in a small program with fewer than ten senior people in a position to judge you, you may well be forced to ask for evaluations from people who disagree with what you do or whose programs are in competition with yours. So there is safety only in numbers. No wonder the big research programs dominate!
There is no doubt that this system has benefited string theory and made it more difficult for people who pursue alternative research programs. As a recent New York Times article noted, “[S]cientists have yet to develop more than fragments of what they presume will ultimately be a complete theory. Nevertheless, string theorists are already collecting the spoils that ordinarily go to the experimental victors, including federal grants, prestigious awards, and tenured faculty positions.” David Gross, now the director of the Kavli Institute for Theoretical Physics at UC Santa Barbara, is quoted in the same article as saying, “Nowadays, if you’re a hotshot young string theorist you’ve got it made.”3