The following `must have' been considered already. It is possible to define homotopy groups for a simplicial set. How about for an n-category? or anyway a 2-category? Naively, the problem seems to be how to `add' 2-cells, but now that we have good pasting theorems.... ================================