Physics needs a reality check: Review of ‘Lost in Math’ by Sabine Hossenfelder
Many physicists working on fundamental theories believe that “they are on the right crack,” says Sabine Hossenfelder in her already controversial new book “Lost in math: how beauty leads physics astray.”
I think this must be a typo, surely the author means “on the right track.” Then another possible interpretation comes to my mind. I email Hossenfelder asking if this means that the physicists who work on supersymmetric theories must be on crack.
No, Hossenfelder replies, the passage refers to the “promising cracks in the foundations” of physics, which persuaded many researchers and decision makers that the Large Hadron Collider (LHC) at CERN would soon find new fundamental particles.
But I remain with the impression that the reference to drugs isn’t entirely unintended. Perhaps physics is intoxicated with beauty? This is Hossenfelder’s main thesis: Many physicists are so much in love with their beautiful mathematical equations that physics is losing contact with the real world. Physics needs a reality check.
Costly experiments at LHC haven’t found any trace of supersymmetric particles. Hossenfelder explains that elementary particles come in two kinds, fermions that want to do their own thing, and bosons that are happy to group together. Just like, you know, libertarians and liberals. In supersymmetric theories, each fermion or boson has a heavier twin of the other kind.
Why? Because the equations of supersymmetry are beautiful. Does nature agree? No way. At least, the hypothetical supersymmetric particles haven’t bothered to show up in LHC experiments so far.
It’s worth noting that the LHC has found the new particle it was built to look for. In 2013, Peter Higgs and François Englert received the Nobel Prize in Physics for their theoretical prediction of the Higgs boson, which was eventually detected by the LHC in 2012.
But the theoretical work that led to the discovery of the Higgs boson was done in the sixties. “Since 1973 there hasn’t been any successful new prediction,” notes Hossenfelder.
The book is written in a clean and crisp conversational style. Using the language that we use every day, including the F word a couple of times, Hossenfelder covers a lot of modern physics at a level understandable by non-scientists and gives good conceptual explanations of advanced topics, such as quantum measurement, decoherence, symmetry breaking at low energies, and string theory.
Hossenfelder’s explanations are often simpler, clearer and at the same time more accurate than those found in other popular science books. For that alone, I wholeheartedly recommend “Lost in Math.” The book includes commented interviews with many other physicists, including Nobel laureates Steven Weinberg and Frank Wilczek.
Hossenfelder, who since 2006 writes the popular physics blog “Backreaction” and is on her way to social media stardom with 16,000 Twitter followers and counting, is likely to be criticized for what many will perceive as a harsh attack on the physics community. But her critique is mostly constructive, and I don’t find it too harsh. If physicists can’t take that, I think we have a problem.
Sabine’s main grudge is that contemporary research in physics is far too much driven by esthetics. “The conclusion I have drawn from my readings and the interviews is that our sense of beauty (for what the laws of nature are concerned) changes the more we learn,” she emails me. “Relying on beauty is putting the cart before the horse.”
Some current theories (unsupported by experiment) are motivated by a desire to explain the “fine tuning” of large numbers that almost exactly cancel each other. This is what happens when, for example, physicists try to match the observed expansion rate of the universe with theoretical predictions.
If you draw two numbers from an assumedly uniform random distribution between zero and one, and find that the difference between the two numbers is exceedingly small, you find this fine tuning unnatural and are tempted to look for a deep explanation.
But perhaps the random distribution was not uniform to begin with. Perhaps it was sharply peaked around the very numbers that you have found. If so, this is just the way things are and no deep explanation is needed. Hossenfelder explains to me:
“Naturalness arguments are good arguments if you do have statistics. E.g., it would be very ‘unnatural’ to throw a die and find it lands on a corner. You find that unnatural because you’ve thrown a lot of dice and never saw that happening before. i.e., you have statistics.”
“You could also go and calculate the probability of that happening based on a sample of initial conditions that you commonly encounter (statistics again) and would also find it’s exceedingly ‘unnatural’ or ‘fine tuned’ to find a die balanced on a corner. The problem is that when it comes to the constants in the laws of nature we have no statistics. We have only this one universe and this universe has these constants and that’s that. Making statistical inferences in such a case is just bad math.”
“I think what has happened here is that arguments which are perfectly sensible in certain situations have been carried over to situations where they don’t apply — and this has gone unnoticed because physicists are sloppy in stating their assumptions.”
Some could criticize Hossenfelder for focusing on the research programs that she considers as lost in math, like supersymmetry, instead of the research directions that she considers as most promising. “Personally I think (as I say in the book) that focusing on mathematical consistency and experimental guidance is most promising,” is Sabine’s reply.
“That makes me think that quantum gravity and dark matter are good research projects. Of course there could always be more to discover. Indeed you could never know that a theory really is the most fundamental one.”
“[But] if special relativity isn’t respected by the theory of quantum gravity, this has consequences which we should be able to observe on cosmological distances. People have looked for this but so far they haven’t found anything.”
Hossenfelder is acutely aware that scientific research is driven by human behavior and a complex interplay of cultural, social and political factors. To my question on whether she agrees with the opinion, expressed by cosmologist George Ellis in “Lost in math,” that the militant atheist attitude of some scientists is dangerously damaging the public appreciation of and support for science, Sabine replies with a short “Yes.”
So far this review has been positive, and I very much like “Lost in math” overall. But Hossenfelder begins to lose me near the end, where she launches a vigorous call for action to reduce human biases in scientific research. Earlier in the book, she suggests that it might be a good idea to outsource parts of the process of selecting promising theories for further work and funding to Artificial Intelligences (AIs) when the time is right.
But it can be proven that the AIs we could build in the foreseeable future can’t match human intuition. Perhaps, I guess, the appreciation of “beauty” (whatever that is) is one of those uncomputable processes that happen in our brain but not in current computing machines.
I’m also left with a nightmare vision of committees, whistleblowers and angry mobs shaming noncompliant researchers and purging their ideas. “Killing ideas is a necessary part of science,” says Sabine. “Think of it as community service.”
I tend to think that scientific research is best when it’s constrained least. Sabine’s proposal to improve research practices seems too ordered and geometric to me. Too mathematical. Lost in math.
Sabine replies:
“My proposal is the very opposite of constraining researchers. Point is, scientists are presently constrained because they are not free to follow their interests. Essentially I am making a free-market argument for knowledge discovery. Requiring scientists to work on what produces papers and attracts citations is a form of regulation that results in inefficiencies.”
“That you refer to this as ‘shaming’ reflects the very problem. Criticism is essential for science, but it doesn’t come to us easily. People take it personally all too often. But I think this is cultural and this culture can change.”
If you draw two numbers from an assumedly uniform random distribution between zero and one, and find that the difference between the two numbers is exceedingly small, you find this fine tuning unnatural and are tempted to look for a deep explanation.
But perhaps the random distribution was not uniform to begin with. Perhaps it was sharply peaked around the very numbers that you have found. If so, this is just the way things are and no deep explanation is needed. Hossenfelder explains to me:
“Naturalness arguments are good arguments if you do have statistics. E.g., it would be very ‘unnatural’ to throw a die and find it lands on a corner. You find that unnatural because you’ve thrown a lot of dice and never saw that happening before. i.e., you have statistics.”
“You could also go and calculate the probability of that happening based on a sample of initial conditions that you commonly encounter (statistics again) and would also find it’s exceedingly ‘unnatural’ or ‘fine tuned’ to find a die balanced on a corner. The problem is that when it comes to the constants in the laws of nature we have no statistics. We have only this one universe and this universe has these constants and that’s that. Making statistical inferences in such a case is just bad math.”
“I think what has happened here is that arguments which are perfectly sensible in certain situations have been carried over to situations where they don’t apply — and this has gone unnoticed because physicists are sloppy in stating their assumptions.”
Some could criticize Hossenfelder for focusing on the research programs that she considers as lost in math, like supersymmetry, instead of the research directions that she considers as most promising. “Personally I think (as I say in the book) that focusing on mathematical consistency and experimental guidance is most promising,” is Sabine’s reply.
“That makes me think that quantum gravity and dark matter are good research projects. Of course there could always be more to discover. Indeed you could never know that a theory really is the most fundamental one.”
“[But] if special relativity isn’t respected by the theory of quantum gravity, this has consequences which we should be able to observe on cosmological distances. People have looked for this but so far they haven’t found anything.”
Hossenfelder is acutely aware that scientific research is driven by human behavior and a complex interplay of cultural, social and political factors. To my question on whether she agrees with the opinion, expressed by cosmologist George Ellis in “Lost in math,” that the militant atheist attitude of some scientists is dangerously damaging the public appreciation of and support for science, Sabine replies with a short “Yes.”
So far this review has been positive, and I very much like “Lost in math” overall. But Hossenfelder begins to lose me near the end, where she launches a vigorous call for action to reduce human biases in scientific research. Earlier in the book, she suggests that it might be a good idea to outsource parts of the process of selecting promising theories for further work and funding to Artificial Intelligences (AIs) when the time is right.
But it can be proven that the AIs we could build in the foreseeable future can’t match human intuition. Perhaps, I guess, the appreciation of “beauty” (whatever that is) is one of those uncomputable processes that happen in our brain but not in current computing machines.
I’m also left with a nightmare vision of committees, whistleblowers and angry mobs shaming noncompliant researchers and purging their ideas. “Killing ideas is a necessary part of science,” says Sabine. “Think of it as community service.”
I tend to think that scientific research is best when it’s constrained least. Sabine’s proposal to improve research practices seems too ordered and geometric to me. Too mathematical. Lost in math.
Sabine replies:
“My proposal is the very opposite of constraining researchers. Point is, scientists are presently constrained because they are not free to follow their interests. Essentially I am making a free-market argument for knowledge discovery. Requiring scientists to work on what produces papers and attracts citations is a form of regulation that results in inefficiencies.”
“That you refer to this as ‘shaming’ reflects the very problem. Criticism is essential for science, but it doesn’t come to us easily. People take it personally all too often. But I think this is cultural and this culture can change.”
Post a Comment