Over at Backreaction, Bee is running an advent calendar of her own, with amusing anecdotes about famous physicists. Apparently, it’s a good year for advent calendars.
A couple of days ago, her story was a famous one about Heisenberg nearly failing to get his Ph.D. because he disdained experiment:
Wien wanted to fail Heisenberg, but Sommerfeld, in whose exam on theoretical physics Heisenberg had excelled, put in a strong word for Heisenberg. Heisenberg passed the doctoral examination with the lowest possible grade
Between this, and my own advent calendar posts about historical physics, I got to thinking about the split between experiment and theory in physics, and wondering where it all went wrong.
Nowadays, we take a nearly absolute division between theory and experiment to be almost a given. Obviously, each side needs to know a little bit about the other– you can’t interpret data without some idea of what theorists are talking about, and you can’t make useful predictions without knowing what experimentalists can do– but the vast majority of physicists are primarily one or the other.
It wasn’t always this way, though. Newton was, in many ways, the father of physics as a mathematical science– he invented calculus in order to be able to do physics, after all– but he also did numerous experiments. Some of these are still pretty useful to think about– his optical experiments splitting and recombining white light, for example– while others seem more like the stuff of mad science and anything a reputable researcher would get up to, such as working a thin knife blade up behind his eyeball so he could describe the effects that he saw by pressing on different parts of the eye. Most of Newton’s contemporaries similarly dabbled in both experiment and theory.
At some point, though, the two began to separate. By the early 1900’s, you had experimentalists like Rutherford, who cherished a deep distrust of theorists (with the exception of Bohr, who got a pass for being a good soccer player), and theorists like Heisenberg and Pauli, whose negative effect on experiments was so infamous that one apparatus failure was jokingly blamed on the fact that he was changing trains in the same city at the time that it broke.
The question is, is there a clear point of divergence? If so, when? I suspect it would be sometime in the early 1800’s, because Faraday has a reputation as being a great experimenter but bad with math, but I don’t know enough about the history to have a complete picture. It could also be much later, though, because Fermi had a very god reputation in both experiment and theory. Or it could be one of those eternal struggles– some of the commenters on Bee’s post try to trace this back to Plato and Aristotle, but I don’t really trust any comparison going back that far. I’m not sure we have a complete enough picture of what they were up to to make useful analogies between ancient Greek philosophers and modern scientists.
(I also wonder if the Heisenberg story gets misattributed to Einstein, who was actually a good deal more practically minded than his popular image. His major contributions were in theory, but he also invented some clever experimental devices and techniques, including a type of refrigerator.)
So, I’m curious to hear what other people think, and if there are really good arguments one way or the other. Or, failing that, I’m always happy to hear more amusing anecdotes about experimentalists vs. theorists…
Interesting question – my inkling would be that the truly great physicists (eg Newton and Einstein) have a bit of both. This is similar to Lee Smolin’s idea in ‘The Trouble With Physics’ about seers versus craftspeople.
People that can do both are really few and far between, but they often wind up being exceedingly influential. Witness two real giants from the mid-20th century who made deep contributions to both disciplines: Fermi and Bardeen. As far as I can tell, the time of the dabbler in both realms is over, as the specialization required to reach the cutting-edge is considerably more significant, and that getting too deep into one way of approaching the world seriously inhibits the other. This is perhaps explained best by a quote from Yogi Berra:
In theory there is no difference between theory and practice. In practice, there is.
I did 2/3 of my thesis work on experiments before realizing I’m better at theory. That experimental background has served me well: The theoretical questions that I pose these days involve asking what the fundamental limits are to how precisely certain things can be measured, or how much information can be extracted in certain situations.
It is still possible to do both theory and experiment at the cutting edge, although admittedly not if you want to do particle physics. In quantum information there are several examples. Raymond Laflamme, currently director of IQC in Waterloo, started out as a theorist and made influential contributions to the theory of quantum error correction, but he also works on experiments in NMR implementations. Manny Knill is also a very good theorist, but he has recently started playing with ion traps. I guess the lesson is that, if you want to do both theory and experiment, then it is better to choose a field that has not been around for many decades, so that the amount of material that you need to master is manageable.
The theory-experimental divide in other fields is not as stark as it is in physics. For example, there is a reasonably large and influential number of experimental biologists who started their careers in theoretical physics or applied mathematics. Also, engineers seem to be able to bridge this divide far more easily.
Fermi is a bit of a unique genius. I don’t remember any other comparable scientist who was both good in theory and in experiment.
Other great physicists of that time were already pretty much divided: Bohr was a theoretic physicist and Rutherford was a great experimentalist. Einstein was a pure theoretic physicist.
Earlier scientists were a bit more mixed. Maxwell was a bit of both. Ditto for Lord Kelvin.
You should check out Ursula DeYoung’s “A Vision of Modern Science”. Robert Tyndall’s a useful case study on when and why the theory/experiment division occurred.
Thanks for the link :o)
I want to put in a good word for Heisenberg because I made the story very short. I too had to suffer through several “Praktika” which is how the experimental physics exercises are still called today in Germany. The problem with these Praktika is that they are often equipped with old, partly broken, missing or malfunctioning things, and they are terribly uninspiring to say the least. They leave you with zero room for creativity, unless you use your creativity to not make the experiment at all (then you learn a valuable lesson in how to fake data). Heisenberg too found himself in such a situation and wasn’t interested enough to track down the missing equipment. Of course that doesn’t really excuse his lack of knowledge about microscopes, but it does explain his lack of enthusiasm.
That having been said, the issue you talk about is yet another aspect of specialization. This is a natural thing to happen both in communities as well as in ecosystems. It becomes problematic only if those who don’t specialize can’t survive. This is problematic because you need them to keep things together.
Alex Besogonov: “Einstein was a pure theoretic physicist.”
For a pure theorist he was very good with experiments. He didn’t do much “scientific” experiments, because he was busy with commercial endeavours. Not just the Einstein-Szilard refrigerator, but also his gyrocompass, made by Anschütz. It was so good that by the beginning of WW2 all Navies used it, except those that spoke English. (Kind of ironic for a pacifist like Einstein…) The royalties from the compass, cleverly rerouted via a mailbox company in the Netherlands, were enough to keep him going in style after he was kicked out of Germany.
@Bee: Maybe the lesson from the “Praktika” has something to do with maintenance of equipment…
The self-regard of mathematicians is always a problem as they will always promote theory over reality. Unfortunately in the UK the replacement for Richard Dawkins as the Simonyi Professor for the Public Understanding of Science in Oxford is Marcus du Sautoy. He’s a nice guy and bright but he knows very little about science and doesn’t understand the core of the scientific approach. He thinks mathematics is a science – no, it isn’t, it’s a different way of working, it’s not based on observation and testing of the real world, its core is internal consistency and a certain aesthetic. And with that one-eyed view, he thinks the only proper science is that which a mathematician can understand. Some stuff is too complex for mathematics and mathematicians at present, that does not make it valueless, it makes it interesting!
It’s intriguing how the puff from physicists suggesting that their hyper-expensive machines are really important in getting them huge amounts of money while it is much more difficult to get funds for the softer life and earth sciences which are far more fundamental to survival of life on earth, health and happiness than the number of GeV in a Higgs boson.
Wouldn’t it be revolutionary to call for undergraduate programs in physics to ‘co-value’ formalism instruction and laboratory instruction?