Accepting complexity as a paradigm (shift) is based on a number of epistemological choices. Just as much as a more classical control driven paradigm of management is equally based on (other) assumptions that are never expressed, and seldom known. Unfortunately (at least in my opinion), not enough academic work (and certainly not enough PhD work) gives due (or even worse, gives any) attention to those choices.
I would like to suggest two interesting works that might help to position the problem and to progress in it. A paper of Marie Joelle Browaeys "Complexity of epistemology", comments on the consequence of considering epistemology as a "theory of knowledge" or rather as " a philosophy of science". Not as irrelevant as it might appear. A book written a number of years ago by Paul Cilliers "Complexity and Postmodernism" argues very clear and strong for a post-structural epistemological view as base for the complexity paradigm and in particular for connectionist approaches.