Right, this is what I thought of. I think, mathematically, those are the same things and, thus, problematic.
So, all in all, it seems you don't need the path integral, but an integral over all particles.
Have you, or anyone else, thought about lattice and/or finite-dimensional variants of the...