Representing the agent through second-order states

Some recent views of action have claimed that a correct conceptual account of action must include second-order motivational states. This follows from the fact that first-order motivational states such as desires account for action or mere behavior in which the agent's participation is lacking; thus, first-order motivational states cannot by themselves account for action in which the agent participates, so-called full-blooded action. I argue that representing the agent's participation by means of second-order states is bound to fail because it misrepresents what an agent is doing when acting in the full-blooded sense. I begin by characterizing full-blooded action and explaining the failure of first-order accounts to explain it. I next show that while second-order accounts have some success in explaining full-blooded action, they fail to distinguish it from action which exhibits motivational alienation. I then argue that even if this problem were resolved, the second-order accounts more fundamentally misrepresent full-blooded action by depicting such action in an introverted manner. I conclude by considering a sketch of agency and full-blooded action that does not rely on second-order states and by addressing a primary concern thought to favor the second-order accounts, the concern of agent causation.