In perusing WordPress over the past few days I have found that the majority of philosophical entries touch on the issue of free will in some way, such that I have decided that today I will add to this dialogue. My interest in free will is for the most part centered on making room for the possibility of moral responsibility, such that I have pursued an answer to the question: what conditions are necessary and sufficient for moral responsibility? I doubt that compatibilism (the view that free actions are caused by one’s will, even if your will was determined by prior events to want to take that action) provides an adequately rich moral responsibility, such that its proposed conditions are not sufficient for moral responsibility, so I tend to think that agent-causal views of free will are necessary for moral responsibility. Agent-causal free will agrees with compatibilism that free actions are caused by one’s will, but counters that outside factors cannot causally determine the choice of the will.
One of the main objections to agent-causal views are that they do not fit into our naturalistic world-view — that is, the view that the physical and social sciences can tell us everything there is to know about the world (this is a very basic definition of naturalism, but there has yet to be a consensus on ‘naturalism’). A particularly pointed objection to agent-causal views is that the social sciences and agent-causal free will are incompatible, and given the social sciences’ reputable placement in our naturalistic world-view, agent-causal views are not naturalistic. I am normally wary of crediting the social sciences with giving us robust and certain knowledge (mainly because I think their methodology is incoherent), but it certainly seems that the social sciences have shown that given certain environmental factors, it can reliably be predicted how persons will act. Agent-causal views seem to entail that we should not be able to predict the actions of human agents if human agents can rise above the causal effects of their environment and choose freely.
Lynne Rudder Baker develops this objection quite well. Baker posits that agent-causal libertarians hold that in order for an act to have been freely willed in a manner sufficient for them being morally responsible for it, the choice must not have its origin in anything outside of the agent’s control. If this is the case, says Baker, then when an agent acts with agent-causal free will, no antecedent causes are sufficient to determine that agent’s action. For Baker, this is absurd because it entails that agent-causal libertarians are committed to the position that we cannot predict peoples’ behavior using social sciences and psychology. On this matter she writes that,
The libertarian holds that if an action is free, it is not explainable by the sum of the kinds of conditions that psychologists and social scientists appeal to – such as how the agent sees her situation, what she takes her options to be, her other beliefs, her desires, her character, her experience, her genetic inheritance, her environment. Taken together, these cannot be sufficient for a libertarian free action. Libertarian free action awaits exercise of an ability that no natural being has: the ability to rise above the complex mix of causes (heredity, environment, beliefs, desires, etc.) and interject an unexplainable X factor, over which the agent has ultimate control, and which renders theretofore-insufficient causes sufficient for the choice or action (Baker, Moral Responsibility Without Libertarianism, pages 313-314).
Here, Baker posits that if causes external to the agent are not sufficient to determine an agent’s actions then past events cannot explain agents’ behavior. For Baker this consequence is absurd; surely past events can and do explain agents’ behavior – the achievements of social scientists and psychologists are proof of this. Yet if past causal conditions do explain agents’ behaviors then antecedent causes are sufficient to determine an agent’s actions. But if this is the case, then agent-causal free will is impossible. Baker’s argument against agent-causal free will can be put thusly,
1. If agent-causal free will is true then past events cannot be used to predict the acts that agents will perform.
2. But past events can be used to predict the acts that agents will perform.
3. Therefore, agent-causal free will is not true.
Indeed, Baker’s argument seems to be troubling for libertarianism as it appears that if agents have the causal power to rise above past causal conditions then it should be impossible to predict the behavior of agents. As a great deal of research has proven and has been made possible by the fact that persons’ behaviors are predictable because of antecedent causes, it appears that libertarianism entails throwing out much of what we know to be true. To survive, the libertarian must show that it is possible to have free will of the agent-causal sort and still allow that past events in an agent’s life can explain their freely willed acts.
The Agent-Causal Libertarian’s Response:
Randolph Clarke and Timothy O’Connor are proponents of versions of agent-causal libertarianism that take seriously the reality that choices are influenced by past events and choices. Clarkes’ view is that past events influence but do not necessitate choices, such that, when an agent acts freely, they are the source of acting for those reasons, even if past events have caused the current set of choices, and caused some choices to have a higher probability of being chosen. O’Connor holds that an agent acts freely when they cause the internal state of having an intention to carry out an act, but other internal factors of deliberation make it that those reasons do not determine that that action be done. As a result, “our prior reasons can explain our actions without causally producing them”.[i] Given, then, that agent-causal libertarianism allows the possibility that past causal conditions can explain why an agent acts as they do, and yet still allow that agents have free will of the libertarian variety, Baker’s argument against libertarianism fails. Let us now tend to how Clarke and O’Connor’s versions of libertarianism avoid Baker’s objection.
Clarke posits that an agent performs a free action when they cause an act to be performed for the reasons for which that act was performed. In other words, a free act is one in which an act is done for certain reasons, and the agent is the source of that act being done for those reasons. In such cases, even if the reasons are caused by past events, the act is freely done because the agent was the source of that act being done for those reasons.
That free acts are acts which are done for the reasons that the agent performs them, and that these reasons can be caused by past events, leads Clarke to the realization that past events cause but do not causally necessitate agents’ actions. By this Clarke means that, it is because of past events that certain choices are open to the agent, but it is precisely because multiple choices are open to the agent that it must be said that past events do not causally necessitate or determine the agent’s actions. If this is the case, says Clarke, then there is no sense in denying that past events were such that some actions have a greater probability of being performed than others. This is consistent with Clarke’s view of libertarianism as it only entails that from an objective viewpoint, and given past events, some actions have a greater probability of being performed due to their reasons. So long as the agent is the source of the act being done certain reasons, her free will in that choice is not any less if someone was able to predict her actions given a knowledge of past events. Thus, for Clarke, “probabilistic causation is not the threat to free will that causal necessitation is”[ii].
Yet if Clarke’s view of libertarianism allows that, “an agent’s behavior, besides being caused by her, is caused also by earlier events, among which are her having or coming to have certain beliefs, desires, preferences, aims, values, and so forth”[iii], and that these past events can be used to judge the probability that an agent will do a certain act, then past events can be used to explain actions meeting the conditions of libertarian free will. Therefore, Baker’s argument is not a bane to Clarke’s take on libertarianism.
O’Connor agrees with this general approach, and holds that an agent freely causes an act when that act was performed for reasons that the agent found motivating enough to exceed a threshold for action. The intention to act is caused by the agent in that the “intention was a direct consequence of the action-triggering intention brought about by the agent, and it causally sustained the completion of the action”.[iv] As such, the agent produces the desire to perform an action, but other factors, such as the character and beliefs of the agent, which are shaped by past events, influence the probability of that action being done. This leads O’Connor to posit that “the agent is the sole causal factor directly producing her intention to A, but her deliberation and activity take place within an internal context (including her total motivational state) that has probabilistically-delineated causal structure”.[v] The thought is, then, that past events, such as those that shaped the agent’s character and beliefs, can explain why certain actions were performed without it being the case that those past events caused those actions.
Thus, agent causal libertarianism, as argued for by O’Connor, allows that past events can factor into decisions without it being the case that those past events causally determined the action of the agent. Rather, those past events were involved in the probability of whether the agent would choose to perform that act. Therefore, past causal events can explain and be used to predict an agent’s behavior without it being the case that the agent does not meet the conditions for libertarian free will. It follows, then, that O’Connor’s version of libertarianism avoids Baker’s critique.
In conclusion, agent-causal libertarianism is not incompatible with the findings of the social sciences, such that, the view cannot be dismissed as not naturalistic on account of the social sciences. There may be other substantial objections to agent-causal libertarianism, either regarding its place in a naturalistic world-view or not, but the agent-causal libertarian will have to be content in answering them one at a time. I have endeavored to show how agent-causal libertarians, following Clarke and O’Connor, can respond to what I call the naturalism challenge posed by the social sciences. If enough progress is made responding to other objections, we can begin to embrace an agent-causal view that allows for the robust sense of moral responsibility that we experience.
[i] Timothy O’Connor, Persons & Causes: The Metaphysics of Free Will (Oxford: Oxford University Press, 2000), 107.
[ii] Randolph Clarke, “Toward A Credible Agent-Causal Account of Free Will,” Nous, 27.2 (June 1993): 193.
[iv] O’Connor, Persons & Causes: The Metaphysics of Free Will, 86.
[v] Timothy O’Connor, “Agent-Causal Theories of Freedom,” in The Oxford Handbook of Free Will, ed. Robert Kane (Oxford: Oxford University Press, 2011), 317.