I want to do the following path integral.

$$\mathcal{Z}=\int\mathcal{D}x e^{iS[\dot{x}]}$$

The action only denpends on $\dot{x}$. For some reason, I want to replace the integral measure $\mathcal{D}x$ by $\mathcal{D}\dot{x}$.

So I have

$$\mathcal{Z}=\int\mathcal{D}\dot{x}\mathrm{Det}\left(\frac{\delta x}{\delta\dot{x}}\right)e^{iS[\dot{x}]}.$$

The variable $x$ is related with $\dot{x}$ via the linear transformation


which implies

$$\mathrm{Det}\left(\frac{\delta x}{\delta\dot{x}}\right)\equiv 1.$$

Am I correct in the above derivation?

  1. For the corresponding problem with discretized time, the Jacobian determinant of the coordinate transformation $$(x^0,x^1,\ldots, x^N)\qquad\longrightarrow \qquad (x^0,v^{1/2},\ldots, v^{N-1/2}), $$ where $$v^{j+1/2}~:=~\frac{x^{j+1}-x^j}{\Delta t} ,$$ would be $\det=(\Delta t)^{-N}$, not unity.

  2. For continuum time, the velocity is $$v(t)~=~\frac{dx(t)}{dt}~=~\int \!dt^{\prime} x(t^{\prime}) \frac{d}{dt}\delta(t\!-\!t^{\prime}). $$ Whether the functional determinant $${\rm Det}\frac{\delta v(t)}{\delta x(t^{\prime})} $$ is unity (or not) depends on regularization scheme and boundary conditions. However, see also this related Phys.SE post.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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