1: \begin{abstract}

2: 

3: \vs{5}

4: 

5: There is a common expectation that the big-bang singularity must be resolved in quantum gravity but it is not clear how this can be achieved. A major obstacle here is the difficulty of interpreting wave-functions in quantum gravity. The standard quantum mechanical framework  requires  a notion of time evolution and a  proper definition of an invariant inner product having a probability interpretation,  both of which are seemingly problematic in quantum gravity. We show that these two issues  can actually be solved by introducing the embedding coordinates as dynamical variables à la Isham and Kuchar. The extended theory is identical to general relativity but has a larger group of gauge symmetries. The Wheeler-DeWitt equations describe the change of the wave-function from one arbitrary spacelike slice to another, however the constraint algebra makes  this evolution purely kinematical and furthermore enforces the wave-function to be constrained in the subspace of zero-energy states. An inner product can also be introduced having all the necessary requirements. In this formalism big-bang appears as a finite field space boundary on which certain boundary conditions must be imposed for mathematical consistency. We explicitly illustrate this point both in the full theory and  in the minisuperspace approximation.

6: 

7: 

8: \end{abstract}

9: