The Propagator

--- title: The Propagator updated: 2026-09-05 -updated_at: 2026-09-05T12:57:47.273Z +updated_at: 2026-09-05T14:35:46.074Z updated_via: api-get updated_ip: visitor-99c4 updated_token: f5edb1216383 @@ ... # The Propagator -The propagator is the answer to a question you should not be able to ask but can. +A propagator is the amplitude for a particle to go from point x to point y. That's the one-line definition. Everything else — the math, the Wick rotation, the pole structure — flows from that simple physical statement. -What is the amplitude for a particle to go from position $x_i$ at time $t_i$ to position $x_f$ at time $t_f$? +## The Two-Point Function -Not the probability. The amplitude. This distinction matters because amplitudes interfere and probabilities do not. The propagator $K(x_f, t_f; x_i, t_i)$ is a complex number. Its squared magnitude gives the probability density. But it is the number itself � the full complex value with magnitude and phase � that contains the physics. The phase encodes the action. The action encodes the dynamics. The propagator encodes everything. +Mathematically, the propagator is the two-point correlation function: -Feynman's insight was that the propagator is a sum over paths. More precisely, it is the path integral over all paths that start at $(x_i, t_i)$ and end at $(x_f, t_f)$: +Δ(x - y) = ⟨0| T{φ(x)φ(y)} |0⟩ -$$K(x_f, t_f; x_i, t_i) = \int_{x(t_i)=x_i}^{x(t_f)=x_f} \mathcal{D}x(t) \, e^{iS[x(t)]/\hbar}$$ +where T is the time-ordering operator. If x⁰ > y⁰, the field at x comes after the field at y, so you create a particle at y and destroy it at x. If y⁰ > x⁰, the order reverses. The time-ordering just makes sure the math respects causality. -The symbol $\mathcal{D}x(t)$ means integration over all possible functions $x(t)$ � all possible trajectories consistent with the boundary conditions. This is not a Riemann integral. It is a functional integral, an integral over an infinite-dimensional space of functions. Mathematicians have spent decades trying to put it on a rigorous foundation, and they have made progress, but the physics does not require the math to be finished. The propagator is defined by this formula, and it works. +For a free scalar field of mass m, the propagator in momentum space is: -For a free particle, the propagator can be computed exactly. The result is: +Δ(p) = i / (p² - m² + iε) -$$K_0(x_f, t_f; x_i, t_i) = \sqrt{\frac{m}{2\pi i\hbar(t_f-t_i)}} \, \exp\left(\frac{im(x_f-x_i)^2}{2\hbar(t_f-t_i)}\right)$$ +That +iε is not a typo. It's the mathematical device that tells you how to navigate the poles at p² = m². Without it, the integral is undefined. With it, you get the Feynman propagator — the one where positive-energy particles propagate forward in time and negative-energy ones backward. -The magnitude falls off as $(t_f-t_i)^{-1/2}$ � the probability spreads out as time passes, which is the familiar dispersion of a wave packet. The phase is proportional to the classical action for a free particle, $S_{cl} = \frac{m(x_f-x_i)^2}{2(t_f-t_i)}$. The quantum propagator carries the classical action in its phase like a fossil carries DNA. +## What the Propagator Means -This is the pattern. For any system with a Lagrangian that is at most quadratic in position and velocity, the propagator takes the same form: a prefactor times $e^{iS_{cl}/\hbar}$. The classical action appears in the exponent, and the prefactor captures the quantum corrections from integrating over fluctuations around the classical path. For non-quadratic Lagrangians, you cannot compute the path integral exactly. But you can expand around the classical path, compute corrections order by order in $\hbar$, and this expansion is the foundation of perturbation theory in quantum field theory. +The propagator is not a probability. It's an amplitude. You can square it to get probabilities, but the propagator itself can be complex, and it can be negative, and it can be infinite. It's a Green's function — a response function. It tells you how the field at one point is correlated with the field at another. -The propagator has a group property that is both simple and profound: +Think of it this way: you have a quantum field. You poke it at point y, creating a particle. What's the amplitude that the field will respond at point x? The propagator answers that question. -$$K(x_f, t_f; x_i, t_i) = \int dx \, K(x_f, t_f; x, t) \, K(x, t; x_i, t_i)$$ +In position space, the propagator for a massive scalar field looks like a Bessel function — it oscillates and decays. The heavier the particle, the faster the decay. A heavy particle is unlikely to propagate far. A massless particle (the photon, the gluon) has a propagator that decays slowly — 1/r in position space. That's why the electromagnetic force has infinite range but the weak force is short-ranged. -For any intermediate time $t$ between $t_i$ and $t_f$, the amplitude to go from start to finish is the sum (integral) over all intermediate positions of the amplitude to go from start to intermediate times the amplitude to go from intermediate to finish. This is the composition law of quantum mechanics. It says that the universe composes amplitudes the way you would compose functions, and it says that every moment is a possible measurement, even if you never measure. +## The Propagator in QED -In quantum field theory, the propagator is the Green's function of the field equation. The photon propagator tells you the amplitude for a photon to go from one spacetime point to another. The electron propagator does the same for an electron. These propagators are the building blocks of Feynman diagrams, and Feynman diagrams are the building blocks of the Standard Model. The entire edifice of particle physics rests on propagators. +In QED, the electron propagator is: -The propagator also reveals something about time. The quantum propagator is not time-symmetric in the way you might expect. The amplitude for a particle to go forward in time is not the same as the amplitude for it to go backward. Time has a direction built into the propagator, not through thermodynamics or initial conditions, but through the analytic structure of the amplitude itself. The $i\epsilon$ prescription that selects the correct propagator � that decides which pole to go above or below in the complex plane � is the mathematical expression of causality. The past influences the future, not the reverse. The propagator encodes this asymmetry. +S_F(p) = i(γ·p + m) / (p² - m² + iε) -The propagator is the most important function in quantum mechanics that no one thinks about until they need it. +The numerator (γ·p + m) projects onto positive and negative energy states. The denominator is the same pole structure. The photon propagator in Feynman gauge is: +D_F^μν(p) = -ig^μν / (p² + iε) + +No mass term — photons are massless, so there's no m² in the denominator. The gauge choice (Feynman gauge, in this case) determines the tensor structure. In a general R_ξ gauge, the propagator is messier, but physical results don't depend on ξ. That's gauge invariance doing its job. + +## Virtual Particles and Off-Shell Propagation + +Here's where things get interesting. The particles on internal lines of Feynman diagrams are *off-shell*. They don't satisfy p² = m². The propagator is evaluated at whatever momentum the integral assigns, and that momentum is generally not the mass-shell value. + +This is the source of the term "virtual particle." The particle exists in the mathematics of the propagator, but not in the classical sense. It's not a thing that could be detected. It's a term in a perturbative expansion. The propagator carries quantum information from one vertex to another, but that information isn't carried by a real particle. + +The uncertainty principle is often invoked here: ΔE·Δt ≥ ℏ/2, so a particle can "borrow" energy to be off-shell for a short time. This heuristic gives the right intuition but isn't the actual mechanism. The actual mechanism is that the propagator is a Green's function, and Green's functions don't care about on-shell conditions — that's the job of the external states, which are always on-shell. + +## The Propagator as a Sum Over Histories + +You can think of the propagator as a sum over all possible paths from x to y. In the path integral formulation, Δ(x - y) = ∫ Dφ exp(iS[φ]) φ(x)φ(y). Every field configuration contributes, weighted by its action. The classical path (where the action is stationary) gives the dominant contribution, but quantum fluctuations add corrections. + +In practice, you never evaluate this path integral exactly (except for free theories). You expand in the coupling constant, and each term in the expansion corresponds to a Feynman diagram. The propagator is the building block — every internal line is a propagator, and the whole diagram is just propagators multiplied together and integrated over internal momenta. + +The propagator is the atom of perturbation theory. Everything else is a molecule made from them. +