Turner applies this thought to science and expertise, using three examples: the Large Hadron Collider, the Homestake mine experiment, and climate change. He calls the unpredictable consequences in such expert systems "knoweldge risk". According to Turner, science in many areas has become more impact and policy oriented, it is under pressure to come up with solutions which involve various disciplines. Old quality controls which worked in mono-disciplinary settings no longer work. But the new system has not put anything comparable in its place. We have moved from a loosely coupled system to a tightly coupled system, and at the same time increased the complexity. When confronted with errors produced by the system, the initial response by the experts follows a pattern described by Perrow: confusion and a radical underestimation of the seriousness of the problem; the wrong outcomes are ‘‘not only unexpected, but are incomprehensible
for some critical period of time’’ (Perrow  1999, 9).
Here is is Turner's conclusion:
But complexity itself produces unexpected interaction effects that belong to no discipline. The expertise in question is about things that are not under any disciplinary domain. Consequently, the structures involve mechanisms of consensus production and control that differ from, and involve more overt use of authority, negotiation, and pidgin communication between specialties than in the model of disciplines with paradigms. The problem of aggregating multidisciplinary knowledge is not a problem that can be solved in any disciplinary way. The pressure to produce results leads to the de facto elimination or moderation of some of the means of correction provided by the traditional disciplinary model—what Robert Merton called organized skepticism is less effective when the objects of the skepticism are not disciplinarily defined. These new structures have many virtues. They produce results that can guide policy and innovation and produce objects that disciplinary science could not. But there is no epistemic free lunch. Each of these differences creates knowledge risk.Before people jump to the bits about climate change it is worth reading the whole piece.