1 Reasons-Responsiveness and the Demarcation Problem Taylor W. Cyr and Andrew Law Forthcoming in Midwest Studies in Philosophy; please cite published version. Abstract: Standard reasons-responsiveness theories, such as Fischer’s and Ravizza’s (1998), tell us to look to other possible worlds in order to determine whether an agent is appropriately responsive to reasons. Carolina Sartorio (2018) has given a powerful critique of counterfactual accounts of reasons-responsiveness, what she calls the “demarcation problem,” and has given an alternative way of characterizing reasons-responsiveness, one that allegedly avoids the demarcation problem. While we agree with Sartorio that the demarcation problem is a serious one for standard counterfactual accounts of reasons-responsiveness, we argue that her own characterization of reasons-responsiveness faces a serious demarcation problem of its own. We conclude by sketching a characterization of reasons-responsiveness that, although inspired by Sartorio’s account, promises to avoid any demarcation problem. Introduction According to standard reasons-responsiveness theories, whether an agent was appropriately responsive (or sensitive) to reasons at the time of their action depends on how the agent would have acted in various possible worlds. 1 Suppose an addict takes the drug he is addicted to. According to standard accounts of reasons-responsiveness, the drug addict is not responsive to reasons when taking the drug because there aren’t a sufficient number of worlds (or patterns of worlds) where the agent refrains from taking the drug. For instance, in worlds where the addict 1 The locus classicus, which we discuss below, is Fischer and Ravizza (1998).