Path Integrals — Mathematical Foundations

Path integrals define quantum amplitudes and statistical partition functions as sums over all trajectories, weighted by

e^{i S /ℏ}

(real time) or

e^{- S /ℏ}

(imaginary time). The central result: the propagator equals a functional integral

\int D x e^{i S [x] /ℏ}

, which is rigorously a measure only in the Euclidean (Wiener) case. Applies across quantum mechanics, QFT, statistical mechanics, and stochastic analysis. Formulated by Dirac (1933) and Feynman (1948), made rigorous by Kac (1949).

A path integral assigns to a quantum transition amplitude the formal expression $⟨ x_{f} ∣ e^{- i H T /ℏ} ∣ x_{i} ⟩ = \int D x e^{i S [x] /ℏ}$ , an integral over the infinite-dimensional space of paths connecting fixed endpoints. It matters because it makes symmetries, perturbation theory, and the classical limit manifest, and because it unifies quantum theory with diffusion and statistical mechanics. The surprising fact is that the real-time oscillatory integral is not a well-defined measure-theoretic object — no countably additive complex measure of total finite variation exists on path space reproducing it — yet its imaginary-time (Euclidean) sibling is the perfectly rigorous Wiener measure. The mathematical foundations are thus a story of analytic continuation, oscillatory-integral regularization, and the Feynman–Kac formula bridging operator semigroups and stochastic processes.

The Core Insight

Quantum evolution over a finite time is the composition of infinitely many infinitesimal evolutions. Inserting complete sets of position eigenstates at each time slice converts the operator $e^{- i H T /ℏ}$ into a multi-dimensional integral over intermediate positions; the continuum limit of these slices is the path integral. The phase of each path is its classical action $S [x] = \int L d t$ , so quantum mechanics becomes "democratic interference among histories," with the classical trajectory emerging by stationary phase as $ℏ \to 0$ .

Formal Statement

Time-Sliced Propagator

Let $H = \frac{p ^{2}}{2 m} + V (x)$ act on $L^{2} (R^{d})$ . The propagator (integral kernel of $e^{- i H T /ℏ}$ ) is defined as the $N \to \infty$ limit of the time-sliced approximation with $ϵ = T / N$ , $x_{0} = x_{i}$ , $x_{N} = x_{f}$ :

K (x_{f}, T; x_{i}, 0) = N \to \infty lim (\frac{m}{2 π i ℏ ϵ})^{N d /2} \int k = 1 \prod N - 1 d^{d} x_{k} exp [\frac{i}{ℏ} k = 1 \sum N ϵ L (\frac{x _{k} - x _{k - 1}}{ϵ}, x_{k})]

⟳Insert

N - 1

resolutions of identity

\int ∣ x ⟩ ⟨ x ∣ d^{d} x

between factors of

e^{- i H ϵ /ℏ}

, then use the Trotter formula

e^{- i H ϵ /ℏ} \approx e^{- i T ϵ /ℏ} e^{- iV ϵ /ℏ}

and the Gaussian Fourier transform of the free kernel.

Here $L (\overset{x}{˙}, x) = \frac{1}{2} m \overset{x}{˙}^{2} - V (x)$ is the Lagrangian, $ℏ$ is the reduced Planck constant, and the prefactor $(m /2 π i ℏ ϵ)^{d /2}$ per slice normalizes the free-particle Gaussian.

Formal Functional Integral

In continuum notation one writes the same object as

K (x_{f}, T; x_{i}, 0) = \int_{x (0) = x_{i}}^{x (T) = x_{f}} D x (t) exp [\frac{i}{ℏ} S [x]], S [x] = \int_{0}^{T} L (\overset{x}{˙}, x) d t

⟳This is shorthand for [1];

D x

absorbs the limit of the product measure together with the

N

-dependent normalization prefactor. No genuine

σ

-additive measure

D x

on path space exists for real

ℏ

— the expression is defined by its time-sliced limit.

Trotter Product Formula

Let $A, B$ be self-adjoint operators on a Hilbert space such that $A + B$ is (essentially) self-adjoint on $D (A) \cap D (B)$ . Then for all $ψ$ in the Hilbert space,

e^{- i t (A + B)} ψ = N \to \infty lim (e^{- i t A / N} e^{- i tB / N})^{N} ψ

⟳For bounded operators expand both sides to first order in

t / N

; the error per step is

O ((t / N)^{2})

so the total error is

O (t^{2} / N) \to 0

. The unbounded case is Kato's theorem, proved via the Trotter–Kato approximation theorem for strongly continuous semigroups.

This is the rigorous engine converting the operator identity into equation [1].

Feynman–Kac Formula

Let $V$ be bounded below and continuous, and let $(b_{t})_{t \geq 0}$ be $d$ -dimensional Brownian motion started at $x$ under Wiener measure. Then the imaginary-time (Euclidean) semigroup $e^{- t H /ℏ}$ has kernel given rigorously by an expectation:

(e^{- t H} ψ) (x) = E_{x} [exp (- \int_{0}^{t} V (b_{s}) d s) ψ (b_{t})]

⟳Apply Trotter [3] to

H = - \frac{1}{2} Δ + V

; the free factor

e^{t Δ/2 N}

generates the Brownian transition density (heat kernel), and the product over slices converges to the time-ordered Wiener integral

E_{x} [\dots]

by dominated convergence using boundedness of

V

from below.

Why It Holds / Derivation

From Operator Evolution to the Path Integral

Step 1 — Slice the evolution. Write

T = N ϵ

and use the group property

e^{- i H T /ℏ} = (e^{- i H ϵ /ℏ})^{N} .

Step 2 — Insert position resolutions. Between each pair of factors insert

1 = \int d^{d} x_{k} ∣ x_{k} ⟩ ⟨ x_{k} ∣

, giving

K = \int k = 1 \prod N - 1 d^{d} x_{k} k = 1 \prod N ⟨ x_{k} ∣ e^{- i H ϵ /ℏ} ∣ x_{k - 1} ⟩ .

Step 3 — Approximate the short-time kernel. By the Trotter formula [3], to leading order in

ϵ

⟨ x_{k} ∣ e^{- i H ϵ /ℏ} ∣ x_{k - 1} ⟩ \approx ⟨ x_{k} ∣ e^{- i T_{kin} ϵ /ℏ} e^{- iV ϵ /ℏ} ∣ x_{k - 1} ⟩ = e^{- iV (x_{k - 1}) ϵ /ℏ} ⟨ x_{k} ∣ e^{- i \overset{p}{^}^{2} ϵ /2 m ℏ} ∣ x_{k - 1} ⟩ .

Step 4 — Compute the free kernel. Insert momentum states and complete the Gaussian:

⟨ x_{k} ∣ e^{- i \overset{p}{^}^{2} ϵ /2 m ℏ} ∣ x_{k - 1} ⟩ = (\frac{m}{2 π i ℏ ϵ})^{d /2} exp [\frac{im ( x _{k} - x _{k - 1} ) ^{2}}{2ℏ ϵ}] .

Step 5 — Assemble the action. Multiply the

N

factors:

K = (\frac{m}{2 π i ℏ ϵ})^{N d /2} \int k \prod d^{d} x_{k} exp [\frac{i}{ℏ} k \sum (\frac{m ( x _{k} - x _{k - 1} ) ^{2}}{2 ϵ} - ϵ V (x_{k - 1}))] .

The bracket is the Riemann sum of

\int_{0}^{T} (\frac{1}{2} m \overset{x}{˙}^{2} - V) d t = S [x]

. Step 6 — Continuum limit. Taking

N \to \infty

formally yields equation [2]. Rigor requires either Wick rotation

t \to - i τ

(yielding the convergent Wiener integral and [4]) or the theory of oscillatory integrals on infinite-dimensional spaces (Albeverio–Høegh-Krohn).

■

Richard P. Feynman | 1918–1988

Feynman developed the path-integral formulation in his 1942 Princeton thesis and the 1948 Reviews of Modern Physics paper "Space-Time Approach to Non-Relativistic Quantum Mechanics," inspired by a cryptic 1933 remark of Dirac that $⟨ x_{t + ϵ} ∣ x_{t} ⟩$ "corresponds to" $e^{i L ϵ /ℏ}$ . Feynman recast all of quantum mechanics as a sum over histories, gave the diagrammatic perturbation theory that bears his name, and — crucially for foundations — emphasized that the dominant paths are nowhere differentiable, with $∣Δ x ∣ \sim Δ t$ , anticipating the Brownian roughness that Kac would make precise. He shared the 1965 Nobel Prize with Schwinger and Tomonaga.

Building Intuition

Imagine evolving a wavefunction through time as a relay of infinitely many tiny hops. At each hop the amplitude to go from $x_{k - 1}$ to $x_{k}$ is a complex number of unit magnitude whose phase is the little bit of action accrued. To get the total amplitude you multiply phases along a path (adding actions) and then add the resulting complex numbers over all paths. Most paths are wild and their phases scramble to zero by destructive interference — except near the classical trajectory, where $S$ is stationary and neighbouring paths share nearly equal phase, so they add coherently. As $ℏ \to 0$ this stationary-phase peak sharpens to a delta function on the classical path: that is how Newton emerges from quantum democracy. In imaginary time the phases become real exponential weights $e^{- S_{E} /ℏ}$ , turning interference into ordinary probability and the path integral into a genuine average over Brownian paths.

Sum over histories between fixed endpoints; the classical path (gold) is the stationary-phase trajectory among nearby quantum paths.

Worked Examples

Free Particle Propagator

Given:

V = 0

, mass

m

, propagate from

x_{i}

x_{f}

in time

T

d = 1

. Step 1 — Stationary path. The classical path is the straight line; the classical action is

S_{cl} = \frac{m ( x _{f} - x _{i} ) ^{2}}{2 T} .

Step 2 — Gaussian fluctuations. Because

L

is quadratic, the fluctuation integral is exactly the product of slice Gaussians, giving the prefactor

m /2 π i ℏ T

. Step 3 — Assemble.

K = (\frac{m}{2 π i ℏ T})^{1/2} exp [\frac{im ( x _{f} - x _{i} ) ^{2}}{2ℏ T}] .

Numerical check: Take

m = 9.11 \times 1 0^{- 31}

kg (electron),

x_{f} - x_{i} = 1

= 1 0^{- 9}

T = 1 0^{- 15}

ℏ = 1.055 \times 1 0^{- 34}

J·s. Phase

= \frac{m ( Δ x ) ^{2}}{2ℏ T} = \frac{( 9.11 \times 1 0 ^{- 31} ) ( 1 0 ^{- 18} )}{2 ( 1.055 \times 1 0 ^{- 34} ) ( 1 0 ^{- 15} )} \approx 4.32

rad.

K \approx (\frac{m}{2 π i ℏ T})^{1/2} e^{i (4.32)}, prefactor magnitude \approx 3.7 \times 1 0^{10} m^{- 1/2}

Imaginary-Time Harmonic Oscillator Partition Function

Given:

H = \frac{p ^{2}}{2 m} + \frac{1}{2} m ω^{2} x^{2}

. Compute

Z = Tr e^{- β H}

via the Euclidean path integral over periodic paths

x (β ℏ) = x (0)

. Step 1 — Mode expansion. Expand in Matsubara modes

ω_{n} = 2 π n / (β ℏ)

. The Euclidean action's quadratic form has eigenvalues

m (ω_{n}^{2} + ω^{2})

. Step 2 — Functional determinant.

Z = n = - \infty \prod \infty \frac{1}{( ω _{n}^{2} + ω ^{2} ) / ω _{n}^{2}} \propto \frac{1}{2 sinh ( β ℏ ω /2 )} .

The product is evaluated using

\prod_{n \geq 1} (1 + x^{2} / n^{2}) = sinh (π x) / (π x)

. Step 3 — Final result.

Z = \frac{1}{2 sinh ( β ℏ ω /2 )} = \frac{e ^{- β ℏ ω /2}}{1 - e ^{- β ℏ ω}} .

Numerical check:

ℏ ω = 0.025

eV (typical phonon),

T = 300

K so

β = 1/ k_{B} T

k_{B} T = 0.0259

eV. Then

β ℏ ω = 0.025/0.0259 = 0.965

Z = \frac{1}{2 sinh ( 0.483 )} = \frac{1}{2 ( 0.502 )} \approx 0.996

matching the exact harmonic partition function, dimensionless.

Common Misconceptions

" $D x$ is a measure on path space." It is not, for real time. The Cameron theorem shows there is no complex measure of finite total variation on

C [0, T]

giving the Feynman integral; the would-be "Lebesgue measure on

R^{\infty}

" does not exist (no translation-invariant

σ

-finite measure on an infinite-dimensional Banach space). The symbol [2] is rigorous only as the time-sliced limit [1], as an oscillatory integral (Albeverio–Høegh-Krohn / Fresnel integrals), or after Wick rotation to the Wiener measure. "Discretization is unambiguous." It is not. For Hamiltonians with momentum-dependent or position-dependent mass terms, different discretizations (mid-point vs. end-point evaluation of

V

, ordering of

\overset{x}{^}, \overset{p}{^}

) yield genuinely different continuum limits differing by

O (ℏ)

quantum corrections. The mid-point (Weyl-ordering) prescription is required to reproduce the symmetric operator ordering. Path integrals "remember" operator ordering through the discretization. Non-existence of Feynman measure. Suppose a countably additive complex measure

μ

C [0, T]

existed with

\int e^{(i /ℏ) \int (\frac{1}{2} m \overset{x}{˙}^{2}) d t} d μ = K_{free}

. Cameron (1960) proved that any candidate constructed as a limit of finite-dimensional Gaussians with imaginary covariance has total variation diverging like

\prod_{k} ∣2 π i ℏ ϵ / m ∣^{- 1/2} \to \infty

. Hence no finite-variation measure reproduces the free propagator. The Euclidean analogue does exist — the Wiener measure — precisely because its covariance is real and positive, making the infinite product converge. This sharp contrast is the central foundational fact.

Connections

The path integral sits at the crossroads of several theories. Via Wick Rotation it connects to the Wiener Process and the Feynman–Kac Formula, identifying quantum amplitudes with diffusion expectations. Through stationary phase it reproduces the WKB Approximation and the Hamilton–Jacobi Equation in the classical limit. Its functional-determinant evaluation links to Zeta Function Regularization and the Heat Kernel Expansion. In field theory it generalizes to the Generating Functional $Z [J]$ , from which Feynman Diagrams arise by functional differentiation, and to Lattice Gauge Theory as a rigorous regularization. The measure-theoretic obstruction connects to Gaussian Measures on Banach Spaces and the Cameron–Martin Theorem.

Real-World Applications

Lattice QCD. The Euclidean path integral over gauge fields, discretized on a spacetime lattice and sampled by Monte Carlo, gives first-principles predictions of hadron masses and the QCD phase diagram.
Polymer physics (de Gennes). A flexible polymer chain's configurational statistics is literally the Feynman–Kac path integral for a Brownian particle; persistence length plays the role of $ℏ$ .
Option pricing (Black–Scholes / Dash). The Feynman–Kac formula expresses derivative prices as path integrals over the stochastic asset process; quantum-finance methods compute exotic options via Euclidean propagators.
Instantons and tunneling rates. Decay rates of metastable vacua (false-vacuum decay, the double-well splitting) are computed from Euclidean stationary "bounce" solutions $e^{- S_{E} /ℏ}$ .
Quantum Monte Carlo / Path-Integral Molecular Dynamics. The imaginary-time path integral maps a quantum particle to a classical ring polymer of $P$ beads, enabling simulation of liquid helium, proton transfer, and zero-point effects.
Polaron and condensed-matter problems (Feynman's polaron). Variational path integrals give the electron–phonon coupled ground-state energy across all coupling strengths.
Quantum gravity (Hartle–Hawking). The no-boundary wavefunction of the universe is defined as a Euclidean path integral over compact geometries, foundational to quantum cosmology.

Historical Development

1923 (Wiener). Norbert Wiener constructs rigorous Brownian-motion measure on continuous path space — the eventual Euclidean foundation.
1933 (Dirac). Dirac's paper "The Lagrangian in Quantum Mechanics" notes that the short-time amplitude "corresponds to" $exp (i L d t /ℏ)$ .
1942 (Feynman). Feynman's Princeton PhD thesis develops the least-action principle in quantum mechanics into a sum over paths.
1948 (Feynman). "Space-Time Approach to Non-Relativistic Quantum Mechanics" publishes the path-integral formulation.
1949 (Kac). Mark Kac, hearing Feynman lecture, proves the rigorous Feynman–Kac formula by Wick rotation to Wiener measure.
1957 (Gelfand–Yaglom). Method for functional determinants of Gaussian path integrals via an ordinary differential equation.
1960 (Cameron). R. H. Cameron proves the non-existence of a countably additive Feynman measure of finite variation.
1972 (Nelson). Edward Nelson rigorously establishes the Trotter-formula route and stochastic mechanics.
1976 (Albeverio–Høegh-Krohn). Rigorous theory of oscillatory ("Fresnel") integrals on infinite-dimensional spaces, giving the real-time integral a precise meaning.

Subtleties & Limits

The real-time measure problem remains unresolved in full generality. While Euclidean path integrals are rigorous for a wide class of potentials, and oscillatory-integral and analytic-continuation methods handle quadratic-plus-Fresnel-class perturbations, there is no universally accepted measure-theoretic foundation for real-time path integrals with generic interactions, especially in interacting relativistic QFT (the constructive field theory program succeeded for

ϕ_{2}^{4}, ϕ_{3}^{4}

but

ϕ_{4}^{4}

is believed trivial, and 4D Yang–Mills existence with a mass gap is a Clay Millennium Problem). Further open issues: rigorous treatment of fermionic Grassmann "integrals," the convergence/Borel-summability of the perturbation series defined by the integral, the proper measure for path integrals on curved configuration spaces and over geometries (quantum gravity), and the sign problem obstructing Monte Carlo for finite-density and real-time lattice computations.