Begley: Why Scientists Need to Change Their Minds

When politicians do it, they're tarred as flip-floppers. When lovers do it, we complain they're fickle. But scientists are supposed to change their minds. Having adopted their views on scientific questions— What killed the dinosaurs? Is the universe infinite?—based on a dispassionate evaluation of empirical evidence, they are expected to willingly, even eagerly, abandon cherished beliefs when new evidence undercuts them. So it is remarkable that so few of the essays in a new book in which scientists answer the question in the title, "What Have You Changed Your Mind About?", express anything like this ideal.

Many of the changes of mind are just changes of opinion or an evolution of values. One contributor, a past supporter of manned spaceflight, now thinks it's pointless, while another no longer has moral objections to cognitive enhancement through drugs. An anthropologist is now uncomfortable with cultural relativism (as in, study the Inca practice of human sacrifice non-judgmentally). Other changes of mind have to do with busted predictions, such as that computer intelligence would soon rival humans'.

Rare, however, are changes of mind by scientists identified with either side of a contentious issue. No one who rose to fame arguing that Alzheimer's disease is caused by sticky brain plaques and who has now been convinced by evidence that the plaques are mostly innocent bystanders, not culprits. No one who once pushed hormone replacement therapy to prevent heart attacks in menopausal women who now realizes that the drugs increase the risk of heart attacks (as well as stroke and breast cancer). No one who cast his lot with the theory that a killer asteroid sent the dinosaurs into extinction who now reads the impact-crater evidence as implicating worldwide volcanism instead. But really, we shouldn't be surprised. Proponents of a particular viewpoint, especially if their reputation is based on the accuracy of that viewpoint, cling to it like a shipwrecked man to flotsam. Studies that undermine that position, they say, are fatally flawed.

In truth, no study is perfect, so it would be crazy to chuck an elegant, well-supported theory because one new finding undercut it. But it's fascinating how scientists with an intellectual stake in a particular side of a debate tend to see flaws in studies that undercut their dearly held views, and to interpret and even ignore "facts" to fit their views. No wonder the historian Thomas Kuhn concluded almost 50 years ago that a scientific paradigm topples only when the last of its powerful adherents dies.

The few essays in which scientists do admit they were wrong— and about something central to their reputation—therefore stand out. Physicist Marcelo Gleiser of Dartmouth breaks ranks with almost every physicist since Einstein, and with his own younger self, in now doubting that the laws of nature can be unified in a single elegant formulation. Gleiser has written dozens of papers proposing routes to the unification of gravity and quantum mechanics through everything from superstrings to extra dimensions, but now concedes that "all attempts so far have failed." Unification may be esthetically appealing, but it's not how nature works.

The most fascinating backpedaling is by scientists who have long pushed evolutionary psychology. This field holds that we all carry genes that led to reproductive success in the Stone Age, and that as a result men are genetically driven to be promiscuous and women to be coy, that men have a biological disposition to rape and to kill mates who cheat on them, and that every human behavior is "adaptive"—that is, helpful to reproduction. But as Harvard biologist Marc Hauser now concedes, evidence is "sorely missing" that language, morals and many other human behaviors exist because they help us mate and reproduce. And Steven Pinker, one of evo-psych's most prominent popularizers, now admits that many human genes are changing more quickly than anyone imagined. If genes that affect brain function and therefore behavior are also evolving quickly, then we do not have the Stone Age brains that evo-psych supposes, and the field "may have to reconsider the simplifying assumption that biological evolution was pretty much over" 50,000 years ago, Pinker says. How has the view that reproduction is all, and that humans are just cavemen with better haircuts, hung on so long? "Even in science," says neuroscientist Roger Bingham of the University of California, San Diego, "a seductive story will sometimes … outpace the data." And withstand it, too.

Let's end on an up note. Like every other psychology researcher, Harvard's Daniel Gilbert believed that people are happier when they can change their minds. But in 2002 he and a colleague discovered that people are generally happier about irrevocable decisions: once you are locked in to a decision, you tend to focus on its positive aspects and ignore the negative ones. But if you are allowed to change your mind, you ruminate on both the positive and negative aspects of the choice, which makes you less happy. Inspired by his findings, Gilbert proposed to his girlfriend. Since the "till death" vow makes marriage an (almost) irrevocable decision, the result is that "I love my wife more than I loved my girlfriend."