It's only the thought that counts

Discovering truths about the world is not a simple thing. Much of our understanding depends on context - what we already know. One of the central facets of human existence is our relationships with other people, and to the degree we accord others respect and trust, we also assign a higher likelihood of truth to the knowledge of the world that they share. This trust in people is not wrong or irrational - it's a perfectly valid way to economize on the time we need to spend trying to understand. Collectively we are far more intelligent then we would be if we tried to find everything out completely for ourselves. But too much trust in others is also what most frequently leads us astray. In the end, for scientific truths, the only thing that matters is the idea in itself, not the people who came up with it. Whether the original proponent was a gray eminence or a complete outsider, whether morally a saint or the reverse, no characteristic of the individual scientist should matter in the long-term acceptance of their contribution. Of course, scientists are human, and degree of eminence does have some effect on the speed with which an out-of-left-field idea becomes mainstream. And there is also the well-known Matthew Effect, where if a less-well-known person and a prominent person come up with the same idea, it's usually the prominent one who gets the credit. Nevertheless, in the long run and overall it's the thought, not the person, that counts, in science.

I write this note partly in the context of the Swift-Hack ("climategate") stolen email episode. This has been loudly trumpeted for something like a month now, including on various conservative media outlets, as somehow damaging to climate science. In reality, the email exchanges reveal the scientists discussing and debating the science in scientific terms, and there is absolutely no evidence from the emails that any "idea" associated with climate science was in any way fraudulently put forward or tampered with in some way. But what you do have is a collection of sometimes less-than-flattering portrayals of scientists at work, expressing their emotions and flaws. They are human, as we all are. But if you actually think you see any evidence of anything damaging to the science itself in the emails, I suggest you post on this thread at Michael Tobis' blog, where he's been looking for any climate naysayer to state their case clearly. Not much response so far there... As far as I can tell, the whole emphasis of the attack here has been on trying to elevate the status of individual people in regard to the science, and then attack their integrity unmercifully. This has already been repeated many times in the past with Al Gore, Michael Mann, Ben Santer, Jim Hansen, and now Phil Jones, who I'd personally never heard of before this past month. The truth is, the science does not depend on any one of these humans - they have flaws as we all do. But in science it is preeminently the thought, not the person, that counts.

There was a period of scholarship, under Scholasticism, where the focus was very much on what eminent people had written in the past. At a time when civilization had been in decline, looking to the writings of the far past was not a bad way to gain knowledge. But too often the approach placed far too much trust in what individual people had done or said. That may help you recover lost knowledge, but it's not science.

In contrast, consider the Ising Model. More than 12,000 papers have been written on this simple physical model of ferromagnetism. It consists of a square lattice (with generalizations to other dimensions and lattice structures) with magnetic particles at each vertex of the lattice that can be in an "up" or "down" state; the energy of the system is lowered if neighboring particles have their magnetic fields pointing in the same direction. At low temperature, almost all the particles point in the same direction and you have a ferromagnet. As temperature rises you reach a critical point where the net internal magnetic field goes to zero and the system becomes paramagnetic (induced moment in a magnetic field). Real magnets show this same critical point behavior, so this simple model captures some of the essential properties of the system. Most interestingly, near the critical point the behavior of some of the system properties is not "analytic" - it involves fractional powers of the difference between the temperature and the critical temperature, so that there is a real singularity in some properties at the critical point. This is a pretty rare and striking thing in a physical model - more importantly, that same singularity behavior actually also shows up in real materials whenever you have a second-order phase transition of this sort, though with differing power laws. So the physical properties of this mathematical model, seemingly quite different from any real material, in fact clarify and illustrate a number of very important features of the real world. It's not surprising it has been the focus of thousands of research articles in the 85 years since it was first proposed.

In the ordinary world you might suspect all that work means that Ernst Ising, who the model is named for, was a highly respected figure in the science of magnetism, critical phenomena, and general statistical mechanics that this relates to. He should have won many awards, had many students, written many papers himself, right? Shouldn't somebody whose name appears in tens of thousands of scientific papers be some sort of scientific celebrity?

Actually, no. Ising did end his life as a physics professor in Peoria, IL, but he did very little physics research after his 1924 PhD thesis, where the model was first presented. The model itself was most likely the idea of the now-even-more-forgotten Wilhelm Lenz, Ising's thesis advisor. Ising didn't actually solve the two-dimensional problem which has the interesting behavior, rather his solution was for the much more boring one-dimensional version. The two-dimensional model wasn't solved until 1944, in a paper written by the much better-known Lars Onsager; it had been given the name "Ising model" by the even-better-known Rudolph Peierls in a 1936 paper that proved it must have spontaneous magnetism at low temperatures.

Ising's career after his PhD is perhaps illustrative of the status of scientists in our world. Until he came to the US, he never held a position as a physicist. Part of the problem was his being a Jew in Germany in the 1930s - but consider his career path: a teacher at several schools, a headmaster, a shepherd, a railroad worker, and then forced to work for the German army. That is what it means to be a scientific celebrity with your name on tens of thousands of research papers.

In science it is the idea, not the person, that matters. Sometimes the result is almost cruel.

My second illustrative example is Hofstadter's butterfly. It comes from a physical model actually not far removed from the Ising model - the spectrum of quantum energy states for an electron on a two-dimensional square crystal under the influence of a magnetic field. Much like Ising's model leading to non-analytic behavior near a critical point, Douglas Hofstadter in this paper also showed something for the very first time that became very influential - the energy spectrum is self-similar, a fractal. Hofstadter's butterfly hasn't been as influential as the Ising model, but Google Scholar shows close to 600 follow-on papers using the term. Once again this was the work of a PhD thesis, in this case Hofstadter's at the University of Oregon, with the paper published in 1976. Everybody's heard of fractals, and Hofstadter's work on two-dimensional electron systems also presaged a great deal of interest in the "quantum Hall effect", which led to two Nobel prizes. So you'd think Hofstadter also would be some sort of physics celebrity.

Well, he is a bit of a celebrity, but not in physics. He moved on to be a professor in computer science and then cognitive science - in 1980 he won a Pulitzer prize for the rather stunning Gödel, Escher, Bach. Douglas Hofstadter is probably one of the brightest minds of our time, but that has essentially nothing to do with the importance of his little model of electrons in a magnetic field. Completely coincidentally, while I was a postdoc at Indiana University in 1990-1991 I ran into a problem that was very similar to Hofstadter's - actually taken to the next level - I was looking at the same problem for electrons on a Sierpinski gasket, itself a two-dimensional fractal structure. I had been working on the properties of electrons in quasicrystals, non-crystalline materials that were the subject of my own PhD work, and it looked like the techniques I'd developed could be applied to that system. I wondered what beautiful pictures could come out of doing the same thing on a system that started out as a fractal. I did get some very pretty pictures, and studied Hofstadter's original paper carefully trying to understand them. I never actually got anything publishable out of it though. I had even thought of dropping in on Hofstadter himself since he was also in Bloomington, but never quite got up the courage. He was too much a celebrity for other things then. I could figure out everything I needed from his old paper and what I knew already, so talking to him wouldn't have been much of a help in my work. It would have been fun though - ah well.

Anyway, once again, in science it's the thought, not the person, that counts.