
0:00
1:03:40
Heat a paperclip past a little over a thousand kelvin and it stops being magnetic. Let it cool and the magnetism returns — pointing somewhere. Nothing in the laws of physics prefers that direction over any other, and yet the iron picks one. Not "settles into the lowest energy state": there is no lowest state, there are infinitely many, all identical, and it takes one.
The first conversation in this series argued that symmetry breaking is a genuine large-number effect, using a sugar molecule. This one does the version a molecule cannot do: a sharp transition at a definite temperature, a continuously degenerate family of choices rather than two, and — the consequence nobody expects — free waves and a stiffness that no single molecule can possess.
The method is Landau's, and it is almost insultingly simple. Near the transition the order is small, so expand the free energy in powers of it. Symmetry kills every odd term. The quadratic coefficient changes sign at the critical temperature — that is the entire content — and a positive quartic term stops the runaway. The bowl becomes a wine bottle, the marble rolls off the hump into the trough, and the order grows as the square root of the distance below the transition. Landau wrote it in nineteen thirty-seven at twenty-nine, a year before the Lubyanka.
Then the two consequences that matter. Because the trough is a continuous circle, you can rotate the choice slowly across the sample and the cost falls without limit — that is a Goldstone mode, a wave that costs arbitrarily little, and in a magnet it is a magnon you can watch directly with neutrons. And run the same sum in three dimensions instead and the cost of twisting a sample *grows* with its size — that is rigidity, which is why a solid resists shear, why a magnet holds its direction against any local disturbance, and (next time) why a superfluid flows forever. One stiffness: ask it statically and you get a modulus, ask it dynamically and you get a wave.
The honest ending is that the method is wrong where it matters most. Landau theory predicts the order grows as a square root — an exponent of one half. Real magnets give about a third; the two-dimensional Ising model gives one eighth; helium's heat-capacity exponent has been measured to four decimal places aboard a space shuttle. Mean-field theory assumes each part feels the average of its neighbours, and close to the transition there is no meaningful average — the sample becomes a froth of ordered patches of every size at once. Why that froth appears, and why utterly different systems share the same wrong-looking numbers, is a later conversation.
———
Series: Physics - Tutorial · Grad level. Every equation spoken in full, no chalkboard required. Built by Fermi AI.
More episodes from "Fermi Podcast"



Don't miss an episode of “Fermi Podcast” and subscribe to it in the GetPodcast app.








