The Standard View in research ethics maintains that, under certain conditions, investigators may deceive subjects and may enroll subjects without their consent. In contrast, it is always impermissible to coerce subjects to enroll, even when the same conditions are satisfied. This view raises a question that, as far as we are aware, has received no attention in the literature. Why is it always impermissible to undermine the validity of subjects’ consent through coercion, but it can be permissible to undermine the validity of subjects’ consent through deception, and it can be permissible to enroll subjects without any consent at all? The present analysis suggests that the answer traces to the conditions on the appropriate treatment of subjects. This conclusion suggests that some requirements for human subjects research, and for valid consent more generally, trace not to the protection of subjects per se but to the proper behavior of agents.