Seungchul Yang

The reasons-as-evidence view holds that a fact is a reason for an agent to φ iff it is evidence that the agent ought to φ. Several critics argue that this view overgenerates reasons: certain background factors—facts such as that the agent can φ, or that there is no reason not to φ—would count as reasons, even though it is widely accepted that they are not. I first reconstruct this problem and explain why a prominent reply, which treats such factors as reasons in at least some cases, does not adequately address the challenge. I then propose a simple revision: reasons are first-order evidence that one ought to φ, while background factors are higher-order evidence that a certain fact is evidence that one ought to φ. I argue that this revision preserves—indeed, reinforces—the main attraction of the reasons-as-evidence view while avoiding the counterexamples.

This paper addresses the problem of hard choices agents confront in scenarios referred to as “Hard Cases.” A key feature of Hard Cases is that an agent is presented with multiple options, each supported by reasons, but these reasons run out when considered against the reasons for other options. Another feature of Hard Cases is that it seems wrong for an agent to arbitrarily pick one option. Several views explain these features of Hard Cases, focusing primarily on the value relation between options, with discussions limited to the relata—namely, reasons for each option. However, reasons relevant to a choice encompass more than the set of reasons for individual options. After defining ‘choosing’ as distinct from ‘picking’ and introducing the notion of a ‘reason to choose,’ I argue that the reason to choose plays a role in what makes a case a Hard Case, and it is not necessarily grounded in any of the options but in the choice itself.