ABSTRACT Frankfurt’s ‘Alternate Possibilities and Moral Responsibility’ broke with the tradition of understanding the kind of freedom required for responsibility in terms of alternative possibilities. At the same time, it inspired and motivated a new family of views in its place: views that focus exclusively on actual sequences or the actual causes of behaviour. But, what exactly does that ‘exclusiveness’ claim amount to? At first sight, it may seem natural to interpret it as the claim that the only facts that (...) are relevant to an agent’s freedom are certain facts about actual causes. This would imply that any non-actual facts are simply irrelevant to the freedom of agents. This paper argues that this interpretation is mistaken, and proposes a better one. It also discusses the related but more general question of the type of project that we are invested in when giving a theory of freedom: Are we interested in the bottom-level grounding facts, or are we interested in some higher-level facts? (shrink)