Boundary Conditions and Dynamics of Collective Intelligence

Motivation

After Dr Page's first lecture Paul posted a project description for measuring diversity, particularly the point where enough diversity is present to be effective in producing collective intelligence. Later Amir said: Why just look at where enough diversity starts? Why not look at what happens among a diverse group when they are together that causes their diversity to be diffused into homogeneity and their value as a diverse group lost? Joe has been expressing similar thoughts in the realm of task-team formation and finding the optimal point of team dissolution, the point where most of the ideas have been shared and no more value remains for the group's continuance.

So lets throw some questions out there:

How do we define, measure and detect the boundary conditions for effective collective intelligence?

Is the future of such a group always decay into homogeneous redundancy or is some kind of evolution possible?