The quantum doubling trick contains a classical trajectory you weren't supposed to find.
Note (2026-11-12): checked against the paper's abstract and full text (Horowitz and Rothkopf). The paper says that both the plus and minus Keldysh paths have classical paths and fluctuations, that the fluctuations of both matter for the correct normalization of the classical limit, and that the classical limit gives the minus path its own initial conditions and equations of motion whose unique classical solution is that it is identically zero (and that this solution propagates backwards in time), so the minus path need not be set to zero by hand. The essay's picture of a nonvanishing hidden classical trajectory that 'doesn't disappear' and whose action cancels is not what the paper derives: the solution it finds for the minus path is zero. The paper's setting is non-relativistic point-particle mechanics with conservative forces, and it says its treatment of dissipative forces is not done.
The Schwinger-Keldysh formalism doubles the degrees of freedom: a forward path and a backward path through time. The forward path carries the physical evolution. The backward path — the “minus” Keldysh contour — is conventionally understood as capturing quantum fluctuations. In the classical limit, its contribution should vanish. That's the standard story.
Horowitz and Rothkopf (arXiv:2603.02350) show that the minus Keldysh path admits its own classical trajectory — a real solution to the equations of motion that propagates backward in time. Its contribution to the action principle is zero, but this zero is not imposed by fiat. It is derived from the equations of motion. The backward trajectory exists, solves the classical equations, and its contribution vanishes because the equations require it to, not because it was assumed away.
This derives the classical action principle from the quantum formalism rather than postulating it independently. The initial-value problem (Hamilton's equations from given initial conditions) emerges from the structure of the Schwinger-Keldysh path integral in the classical limit. The minus path doesn't disappear — it becomes a backward-propagating classical solution whose action exactly cancels.
The structural insight: the quantum formalism contains more classical structure than the textbook account suggests. The doubling that is introduced to handle quantum evolution carries, in its classical limit, a hidden partner trajectory. Discarding it as “the quantum part” obscures the fact that it has a well-defined classical role.
Horowitz and Rothkopf, "Hamilton Revised: The Action Principle for Initial Value Problems," arXiv:2603.02350 (2026).