My dissertation maintains that the best theory of free will is an agent-causal view of the sort proposed by Timothy O'Connor. The signature of agent-causation is that an agent acts freely only if there is an event having a cause, where this cause is not some other event but rather the agent of the act in question. Like O'Connor, I reason that nothing can efficiently cause, though something may contribute causally to, someone's directly free action. I propose a theory of causal contribution in the spirit and on the shoulders of Roderick Chisholm. I attempt to get very clear about several pre-theoretic features of free action. Although widely discussed at the pre-theoretic level, scholars rarely use the notion of ultimate origination when constructing rigorous arguments for, against, or in defense of particular theories of free action. Intuitively, an agent acts freely only if she is self-determining in that she is an ultimate source or underived originator of change. I argue only agent-causation obviously satisfies a precise formulation of the origination condition. Accordingly, I fault Carl Ginet and Stewart Goetz's simple indeterminist theories as well as Robert Kane's causal indeterminism. I defend agent-causation from popular objections, for example, objections advanced by Bernard Berofsky, C.D. Broad, Baruch Brody, John Bishop, Randolph Clarke, Roderick Chisholm, Donald Davidson, Alan Donagan, Carl Ginet, Ted Honderich, Robert Kane, John Thorp, and Gary Watson. I undermine the following two objections in detail. First, Peter van Inwagen argues that since an undetermined free action is impossible, no agent-causal theory of undetermined free action is correct. Second, Andrei Buckareff and Richard Feldman contend that agent-causalists cannot account for rational free action' that is, for an agent's acting freely for a reason. Though I defend Ted Warfield's novel argument for incompatibilism from Dana Nelkin and Samuel Rickless's recent criticism, I argue that Warfield's case fails for independent reasons. Nonetheless, I underscore the point that my agent-causal view best explains our pre-theoretic commitment that no directly free action could be determined.