The Epistemic Authority of Expertise

When is it more rational to think for oneself or to defer to the relevant expert? Expertise is either closed-system oriented and lay-person oriented. The first sort is concerned primarily with controlling and manipulating a discipline's defining set of variables as a closed or relatively closed system. The second sort is simply in the business of "advising" clients. I argue that when expert claims are of the first sort, the layperson must defer to the experts; but when experts either extrapolate from their closed-systems, or if they are of the second sort, then the layperson should think for herself.