1: \begin{abstract}

2: A rigorous implementation of the Wheeler-Dewitt equations was derived 

3: in the context of Loop Quantum Gravity (LQG) and was coined Quantum Spin

4: Dynamics (QSD). 

5: The Hamiltonian constraint of QSD was criticised as  

6: being too local and to prevent ``propagation'' in canonical LQG.

7: That criticism was based on an algorithm developed 

8: for QSD 

9: for generating solutions to the Wheeler-DeWitt equations. The 

10: fine details of that algorithm could not be worked out because the QSD 

11: Hamiltonian constraint makes crucial use of the volume operator which 

12: cannot be diagonalised analytically. 

13: 

14: In this paper, we consider the U(1)$^3$ model for Euclidean vacuum LQG 

15: which consists in replacing the structure

16: group SU(2) by U(1)$^3$ and otherwise keeps all properties of the SU(2) theory

17: intact. This enables analytical calculations and the fine details 

18: of the algorithm {\it can} be worked out.

19: 

20: By considering  one of the simplest possible non-trivial class of solutions 

21: based on very small graphs, we show that 1. an infinite number 

22: of solutions {\it exist} which are 2. generically {\it not normalisable} 

23: with respect to the inner product on the space of spatially diffeomorphism 

24: invariant distributions and 3. generically display {\it propagation}.

25: 

26: Due to the closeness of the U(1)$^3$ model

27: to Euclidean LQG, it is extremely likely 

28: that all three properties hold also in the SU(2)

29: case and even more so in physical Lorentzian

30: LQG. These arguments can in principle 

31: be  made water tight using modern numerical (e.g. ML or QC) methods 

32: combined with the techniques developed in this paper which we reserve for 

33: future work.    

34: \end{abstract}

35: