[LUCID LIVING #4]  Suspension of Judgment: The Cornerstone of Science

By Piet Hut

What is science?  In how far is it different from other human activities?  Several answers come to mind.  Some people may insist that the use of mathematical models is crucial.  Others may point to the use of experiments to give objective proof for the theories put forward by scientists.  Yet others argue that there is a body of accepted core practices, all of which together characterize what it means to do science, albeit in significantly different ways in different disciplines.

But if I were to choose one central aspect that is present in any form of science, I would point to suspension of judgment.  When we put forward a new theory, or refine an existing theory, in both cases we introduce a working hypothesis.  As the name suggests, such a hypothesis can then be tested by working with it.  From the theoretical side, we can deduce new consequences, which can then be tested experimentally.  But in doing so, two things are essential: we should not believe our own working hypothesis, and we should not disbelieve it either.

If we believe in our hypothesis, we are likely to fool ourselves, in a myriad possible ways, something humans are very good at.  We may look for theoretical consequences that affirm our hunch, while neglecting to look for implications that could rule our hunch out.  And when doing experiments, if we take a position as to what we think the outcome should be, we are very likely to introduce subtle biases.  By suspending any judgment as to whether our guesses will eventually turn out to be true or not, we maximize the chance that we will be equally likely to notice factors that may confirm as well as factors that may contradict our hypothesis.

I like to say that science is the ultimate bipartisan activity.  A scientist gets credit for coming up with an interesting new hypothesis that is not obviously wrong, and that, if true, would lead to interesting new consequences.  And a scientist also gets credit for showing that a given theory is wrong, by showing convincing contradictory evidence.  Coming up with an interesting new idea is a progressive move, politically speaking, while showing such a new idea to be wrong is a conservative move.

Science is the ultimate bipartisan activity.

The fact that both activities are equally awarded, shows the bipartisan nature of science.  There is no judgment as to whether it is better to come up with something new and interesting, as is the case in fashion, or whether it is better to preserve the existing wisdom by criticizing devations from the old, as in many forms of established religion.

This is not to say that all scientists live up to that ideal, far from that.  I had an eye-opening experience along those lines during a lunch conversation, not long after I got my PhD.  Let me first sketch the background.  As a young postdoc collaborating with a couple colleagues, we put forward a theory that the Sun might have a companion star, which we tentatively called Nemesis.  This was soon after the time that Luis and Walter Alvarez and coworkers had put forward the hypothesis that an asteroid impact had effectively killed off the dinosaurs.

We speculated that this star, moving on a very wide orbit, could have caused comet storms every 26 million years or so that would have increased the chance for impacts of comets on Earth, which in turn might have increased the chance for mass extinctions and cratering.  The star hasn't been found yet, and may or may not exist; it's certainly possible that the periodicities reported in both cratering and mass extinctions were statistical flukes.  And it's also possible that new observations may find the kind of star that we predicted.

Whatever the final verdict will turn out to be, for me personally it was a fun experience waking up during a visit to Berkeley, and seeing our hypothesis described prominently on the front page of the San Francisco Chronicle that morning, under the headline of "Death Star Threatens Life on Earth" or something along those lines.

What surprised me at that time, was that most of my colleagues expected me to defend the new theory that we had put forwards, in a strong and one-sided way, leaving it for others to attack the theory.  Instead, I repeatedly emphasized that I didn't have any axe to grind.  I would be happy if we, or any other team, would discover the star, and I would be happy if we, or others, would show clearly that the statistics would not hold up to scrutiny with new observations as well as more accurate analyses.

I remember one particular lunch conversation at the Institute for Advanced Study, where I was a postdoc at the time, where quite a few colleagues around the table chided me for sitting on the fence, and not playing the role of enthusiastically defending my theory.  To my delight, just around that time Martin Rees, already then one of the most famous living astronomers, joined our table, and took my side of the argument; that is, the side of not taking sides.  He emphatically agreed with me, and mentioned that he had been a co-author on two recent papers, one of them defending a theory that he had put forward with a few others, and one of them attacking that theory.

I was delighted to hear that a famous astronomer like him was not only able to suspend judgment concerning his own ideas, but even was proud of his bipartisan attitude in testing the mettle of an idea by walking around it, so to speak, to look at it from different directions, to spot both strengths and weaknesses.  Alas, in practice most scientists I have met tend to behave more like chess players, looking for winning moves for theories they have identified with, rather than doing research with a more open mind.  Fortunately, in the long run, science as a human institutionalized activity, turns out to be far more reliable than individual scientists are.

Piet Hut is President of YHouse, Professor of Astrophysics and Head of the Program in Interdisciplinary Studies at the Institute for Advanced Study in Princeton, and one of the founders of YHouse.

Comment