Mathematical perspective

In mathematics, the assumption of a single global state-space is being expressed by the concept of state-space representation of a dynamic system. There are two assumptions here, both are correct as a special case (see above):

1. First, that the single 'best' state-space description of a system is possible;
2. The description of a dynamic system in terms of a single 'best' stable state-space representation helps to understand / use / create systems better;

Armed with this irresistibly powerful (but, in our view, limited) method, people study dynamic systems in terms of their trajectories in a stable state-space, which is globally defined (as if by some external to the system omniscient agent).

If we drop this assumption of a single global state space and the possibility of omniscient observer, yet still want to understand a dynamic system, we have to take an 'internal' view of it. The only way to come up with a better understanding of the whole system when being a participant of it is to reach an agreement / coherency / coordination with other participants of the same system which also have limited, but probably non-overlapping observations of it.