Physics Stack Exchange is a question and answer site for active researchers, academics and students of physics. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I've recently encountered a path integral of the form

$$\int \delta[a\phi+b\phi']\,L(\phi,\phi')\;\mathcal D\phi\mathcal D\phi'$$

(where $a$, $b$ are integers) and would like to eliminate one of the $\mathcal D\phi$ integrations.

Is this possible? A change of variables doesn't seem to work since e.g. for substitutions $u:=a\phi$ the path integral Jacobian isn't well-defined if $a\neq 1$.

share|cite|improve this question

If e.g. $b=1$, it is obvious that the delta-functional – you should really call it $\Delta$ and not $\delta$ – "cancels" against the integral, giving $$\int L(\phi,-a\phi) {\mathcal D}\phi $$ where I just substituted the right value for $\phi'$. Similarly for $a=1$. Now, for general $a,b$, the Jacobian is just the simple power $b^{-N}$ where $N$ is the (naively infinite) dimension of the integral ${\mathcal D}\phi'$. In a sensible regulation, $N$ will really become the Euler characteristic $\chi$ of the manifold on which $\phi'$ is defined – the regulated "number of points" on a manifold. So the result is $$\int L(\phi,-a\phi/b) {\mathcal D}\phi \cdot b^{-\chi} $$ If there are many, $k$ fields per point, you need $k\chi$, not just $\chi$. In practice, the factor $b^{-\chi}$ is just a normalization constant that you need to renormalize to a right value anyway. It doesn't depend on any dynamical fields so you may pretty much ignore it.

You shouldn't be afraid of such simple manipulations.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.