Eqs

Saturday, July 20, 2013

The curious case of the lethal publication

Disclaimer: I refuse to believe anything I wrote below. And so should you :-)

I have a confession to make. I think I may have been responsible for the demise of some very eminent scientists. Unwittingly, and unwillingly, of course. But the evidence is really quite damning, and I shudder to think what horrible deed I will commit next. Perhaps the only way for me to clear my conscience is by coming clean with the whole story. So bear with me here, as I free myself from this heavy burden.

It all began in 1986, when I started graduate school in Physics at Stony Brook University on Long Island. I had worked on solitons in nonlinear field theories already at Bonn University, but had never published anything about them. These solitons, called "Skyrmions" after their inventor Tony Skyrme, were supposed to be a good model of the nucleon. I wrote two papers about them in 1987. Shortly after publication of the first, Tony Skyrme died mysteriously (he was only 64). It said, at the time, that he died from an error in anesthesia while undergoing routine knee surgery, and I certainly believed it. Today I am not so sure anymore. I think I am to blame. When I write papers about somebody's life work, it seems like I seal their fate.

But I'm getting ahead of myself. Surely this was just a coincidence. It was just a pair of papers--and the second one was only published in 1988, because the publisher had lost the manuscript at one point. Yes, we sent in manuscripts by mail during the olden days. 

When I turned my attention from nuclear physics to theoretical biology, one of my first papers was on the concept of self-organized criticality, invented and championed by the physicist Per Bak. Per and I were  friends for the most part. He would tell me about how he handled submissions to the journal Physical Review Letters when he was editor there, and how that got him fired. (When the stack of manuscripts on the desk of a fellow editor got too tall, he just sent "accepted" notices out to all of the authors. Those were the days.) But the publication of my article "Self-organized criticality in living systems" was not something he seemed to be very fond of. I know that because I have good evidence that he was the referee on the submission to Science magazine that was ultimately unsuccessful. Anyway, he did not survive that publication. He passed away in circumstances that are quite shocking. I'm still sadddened by it. And, looking back, I feel even more guilty. 

But, I'm sure you understand, I had no inkling at that point that I wielded some sort of nefarious power. I was just following my career, publishing papers in the areas that interested me. And at that point in time, I had become interested in the theory of computation, mainly because I started to work in the field  of quantum computation. One of the seminal people in this area was Rolf Landauer, who showed that the only thing that costs you energy in a computational process is the act of erasure. I even sat next to him at a dinner once! But why did I just have to write this paper on "Complexity, Computation, and Measurement" in 1996, where I introduced the concept of "Landauer machines"? Rolf was a hero of mine, but unfortunately that paper was just too much. It was not long after it was published that Rolf succumbed. 

Naturally I could still ascribe all this to circumstance, and certainly I did. So when I wrote a paper about "Entropic Bell Inequalities" in 1996, I was not worried. After all, John Bell had died in 1990, and the other people possibly inolved in this work were those that created the concept of "EPR pairs" that were so fundamental to that paper. "EPR" stands for "Einstein-Podolsky-Rosen", of course. Einstein died in 1955, Boris Podolsky in 1966, and Nathan Rosen in.. Oh no! Rosen was still alive when we wrote our paper! But he was 86 already, and so it was just natural that he passed on even while our paper was still in revision. Or was it?

After this experience, I learned to be more careful. For example, I learned that if you write a paper together with the person whose seminal contribution you address, that person is safe. This is how Ismail Zahed and Gerry Brown survived the onslaught of my publications. And I felt safe writing papers on the classical theory of information, because its originator (the eminent Claude Shannon) was surely dead by the time I ventured on the scene. My article on the "Physical Complexity of Symbolic Sequences" that appeared in 2000 showed that you could use Shannon information as a proxy for the complexity of biomolecular sequences in general. But, I had no idea that Shannon was still alive. Nobody knew! He would go around to information theory conferences, listening to conversations between researchers during breaks, and then introduce himself. Of course, nobody would believe he was Shannon! 

So Shannon was alive after all, until this cursed paper of mine appeared! Another hero of mine slain! I was dumbfounded. Mortified. I vowed never to write a paper again. 

But you know that I could not keep this up for long. Like you yourself would have done, I shrugged this off as pure coincidence mixed with a pinch of paranoia. After all, Landauer and Shannon were old. It wasn't me, I repeated to myself. So I started on another project: to understand the abundance distribution of species. These distributions are fairly general: you can write down distributions of any taxon, as a function of how many subtaxa each taxon generated. So, for example, I wanted to know what is the distribution of number of species that each genus has created, or the distribution of genus sizes (number of genera) that each family generated, and so on. With my student Johan Chu, I developed a theory that would predict these disctributions based on only two (and sometimes just one) parameter, using the theory of branching processes. To test the theory, we wrote to John Sepkoski, a paleontologist at the University of Chicago, who as his life's work has created data sets (compendia) of marine animal families and genera. He kindly sent us his compendia, which we used for our paper that ultimately appeared in the Proceedings of the National Academy. John was 50 at the time, so surely he was safe.

But he was not. John Sepkoski passed away at the age of 50 of an apparent heart attack, just before my paper appeared in print. My heart broke a little too. What is this curse?

Could it get any worse? What if I wrote a paper on the growth of complexity in biology, a subject that is very near to the heart of eminent evolutionary biologist and paleontologist Stephen J. Gould? Could my paper "What is Complexity?" that appeared in 2002 have anything to do with his sudden passing? Perhaps I should not have addressed the concept of "punctuated equilibrium", one of hist best-known contributions?

I can't go on. I don't want to look at the list of my publications anymore, for fear of what I might find. Best not to think about it anymore. I hope that my upcoming paper in Nature Communications about a certain type of strategy in game theory (so-called ZD strategies) doesn't have any negative consequences for anyone. Bill Press, one of the co-authors of the original paper on ZD strategies is 65 and in perfect health, as far as I know. And Press's co-author, the eminent Freeman Dyson is... oh oh. Dyson is 89.