<p>Some recent discussions of mechanistic explanation have focused on control operations. But control is often associated with teleological or normative sounding concepts like goals and setpoints, prompting the question: Does an explanation that refers to parts or entities within mechanisms “controlling” each other thereby fail to be mechanistic? In this article, I introduce a distinction between open-ended and closed-ended control. I then argue that explanations that enlist control operations to do explanatory work can count as mechanistic in the New Mechanist sense only if such control operations are closed-ended, not open-ended. In certain scientific fields that incorporate control operations within their mechanistic models, for example, systems neuroscience, the state of the science is such that this requirement often cannot be met. We should therefore distinguish between models/explanations that are “mechanistic” in the strict (New Mechanist) sense and those that are “mechanistic” in a weaker sense employed by fields in the latter category. </p>