The PhD Standard


I came across this twitter thread in which a few physicists argue about what knowledge makes a Physics PhD.

Jon Dowling argues that some of the more advanced QM topics (WKB, specifically) are antiquated and not useful knowledge, now that we have the computational resources to solve these equations numerically to arbitrary precision. Rod van Meter chimes in with a slightly different take, that it should be possible to get a PhD in Quantum Computing without taking courses in Statistical Mechanics, and E&M, for example. However, he would require a colossal amount of other knowledge, mostly systems architecture stuff from what I can glean.

I confess to a strong emotional reaction when I went through this thread initially. It seemed insane to remove the eternal pillars of physics education from PhD requirements. Of course, then I remembered that my PhD courses were almost totally irrelevant to my dissertation work. Jackson’s E&M actually ruined electricity and magnetism for me. What had been a beautiful, elegant subject when presented by Griffiths (for undergrads, admittedly) turned into an shitstorm of arcane vector calculus identities that obscured whatever intuition I was expected to gain from the problem sets. Stat Mech I mostly slept through, and classical mechanics is always a drag. Quantum mechanics was great and fun. Unfortunately, spending an enormous time on scattering does not exactly prepare one for a career in quantum information!

Even though my dissertation was on a topic considered to be ‘quantum’ (it has quantum in the name!), there was very little quantum mechanics at all. Part of this is because the microscopic theory of superfluid helium was and is, in my view, totally inadequate. Indeed, I would have been better served by a graduate fluid dynamics course, which was offered only in the Mechanical Engineering department.

Evaluating a PhD program in the context of the usefulness of the coursework seems a little bit bizarre. I’d never argue that the core graduate physics curriculum should include organic chemistry or critical literary theory over electricity and magnetism, even though I’ve had to use Green’s functions about as often as I needed to know about organic molecules or whatever the hell literary theorists study.

What’s the point of the grad school curriculum?

  1. Further develop physical intuition in more complex physical systems
  2. Introduce advanced mathematical techniques in the fields where they are most prevalent
  3. By 1. and 2. create a common knowledge foundation among physicists.

In this framework there’s still room for modernization. The whole physics curriculum can be redone to focus much more on numerical techniques, now that computational power is cheap and ubiquitous. This is a far cry from allowing folks to get a physics PhD without taking quantum mechanics. Obviously the only physics course that shouldn’t extend beyond its Physics 101 form is Classical Dynamics.