抄録
Libertarianism, as a theory of free will and voluntary action, presupposes the existence of free will and the possibility of agent causation, which is essentially different from event causation. Many contemporary physicalist-minded philosophers believe that agent causation is the most serious difficulty for the project of naturalizing libertarianism. From their point of view, the only possible causality is event causation, and the concept of agent causality is no less mysterious and incoherent than the concept of an agent. But this opinion arises through ignorance of the difference between causation and the physical processes which realize causal powers. As we can suppose that all mental abilities supervene upon brain processes, any kind of causation should be seen as supervening on some physical process.
There remains one more difficult problem. In the libertarian tradition the agent is thought to be a kind of “unmoved mover” or a special kind of cause, i.e. “causa sui (cause of itself)”. How is it possible to naturalize this mysterious cause ? To this hard problem I propose a solution: when physical systems (neural networks) are selfformed by an agent, they can realize that special kind of cause. Consequently it is possible to introduce the agent causality into the materialist framework without presupposing the existence of mysterious entities.
