The Critique of Practical Reason includes Kant's attempt to solve the problem of free will. Critics of Kant's solution for freedom, in the face of Newton's deterministic laws of nature, complain that it requires a transcendental realm that Kant called noumenal, which was in some ways an extension of Descartes' mental realm. Both Kant and Descartes locate human freedom in their noumenal/mental realms and consider the dualistic phenomenal/material/bodily realm to be completely governed by deterministic laws.

But there are readings of Kant that make human freedom, for him a prerequisite for morality, something entirely inside us, as seen in his famous distinction between the "starry skies above us and the moral law within." Bok focuses on a view of free will and responsibility that is something entirely within us and in some sense independent of natural causal explanations.

Bok applies some Kantian concepts, notably the important distinction between pure reason (theory) and practical reason (actual practice), to explain how humans can be free and responsible agents even if determinism is "true." She separates the question of freedom from that of responsibility and does not discuss the conditions under which we should hold persons morally responsible for their conduct. She also wisely separates (and does not discuss) the question of punishment from moral responsibility.

For Bok, practical reasoning does not look for causal chains, but looks for the grounds for choices not yet made, and not in a cause-and-effect temporal sequence. The purpose of practical reason is to determine the will, to answer the (Kantian) question "What should I do?" (Theoretical reason answers the question "What can I know?")

Bok says that libertarians cannot explain "self-determination" by discovering a theoretical causal chain of events. In any cases, there are many (perhaps infinite numbers of) causes for any event. Which cause is the most important is often a subjective matter.

Following Galen Strawson, even if some events in that chain are indeterministic (which she correctly notes means that our actions are not pre-determined from before our birth), discovering all the causes theoretically leaves our selves out of the picture. Quoting Thomas Nagel, she says in causal accounts the 'self' simply disappears.

Bok faults libertarians for using theoretical causal chains including indeterministic events to explain free will. They use indeterministic natural events to make our freedom "entirely independent of our purposes, our interests, our reasons and our selves." Only then would freedom exist for them, she says,

"Libertarians [should] want to show not only that antecedent events do not determine our conduct but that we do; that our choices are not just indeterministic but self-determined."

This is a reasonable criticism. The challenge for libertarians is to show that indeterministic events can break causal chains, as Bok sees clearly, but that in no way do they directly cause our actions.

This requires that the indeterminism be limited to the first stage in a two-stage model, where it simply expands the number of alternative possibilities available for evaluation and selection in the second, adequately determined, stage.

Bok concludes that the most rudimentary form of freedom is this: "to be free, our wills must be determined by ourselves and not by the external world." Theoretical reason is concerned to provide causal explanations of events. In this case "determined" must be interpreted as "caused." Our choices must be interpreted as "events."With such a theoretical formulation, the problem of free will cannot be solved. We cannot make sense of the problem as long as we assume that the kind of determination in question is causal.

Whereas theoretical reasoning is concerned with the causes of events, practical reasoning is concerned with the grounds on the basis of which we choose to act as we do. Practical reasoning constructs a system of reasons that allow us to determine which actions we should perform.

Bok says that when we engage in practical reasoning we must use a particular concept of freedom. She argues that "when we act freely in this sense we choose among genuine alternatives, that we should hold ourselves responsible for those acts we freely perform, that we should likewise hold others responsible for their conduct, and that this sort of responsibility differs in kind from mere causal responsibility."

She says that the truth of her claims is independent of the truth or falsity of determinism, or any hypothesis about the causes of human action. She says that the problem with libertarian accounts of self-determination derives from their attempt to construe self-determination as a causal property.

The practical reasons by which conduct is determined are not temporal, she claims. Thus she does not need to cite earlier and earlier events in a causal chain, though of course such events and their causes are considerations for practical reason. This agrees with the two-stage model view that the alternative possibilities (free thoughts) for willed action are all in existence in the mind at the same time (some indeterministically generated and in that sense "uncaused") for evaluation and selection at the later second stage.

Moreover, although the events that cause an action may be external to the mind, the practical reasons that are determining the action cannot be. Justifications for the action (reasons) are independent of any ultimate external natural events (whether deterministic or indeterministic) considered as causes. Bok says we must consider not causal origins but rational grounds, not the causes leading to the action but the reasons that convince her an action is sound.

Even if a particular action, the consequence of our practical reasoning, could be described theoretically as the effect of natural causes, we should not regard ourselves as obeying the dictates of some original natural event that controls our lives. She says that the coexistence of theoretical and practical views explains how we can be both fully natural beings and genuinely free and responsible beings.

Bok concludes:

On the one hand, we want to see our freedom and our dignity as moral agents written into the fabric of the universe for all to see. But on the other, we are unwilling to regard ourselves as worthy of being its sources or its guarantors. We may feel that as purposive agents we are tainted or fickle or untrustworthy; that our freedom is at risk to the extent that it depends on us. We may believe that any conception of ourselves that depends on our practical reasoning could only be a self-aggrandizing distortion or a way of projecting our own desires onto the universe. In either case we would think it an unworthy foundation for our freedom. If we adopt this view, and look to theoretical reasoning for our dignity as free agents, we will find no dignity worth having. But that is only because, in devaluing its source in our practical reasoning, we have already denied it.

Bok nicely summarizes the standard argument against free will, if determined, not free, if undetermined, not responsible.

The fundamental problem is as follows. Gazzaniga and Libet both assume that if our choices are causally determined by something, then we are not free. If they are right, then claiming that our choices are due to something nonnatural, like a ‘ghost in the machine’ or a spirit, cannot solve the problem unless that claim allows us to argue that our choices are not causally determined by anything. If, however, our choices are not causally determined by anything, then our choices are not under our control and we are not morally responsible for what we do. Moreover, the idea that our freedom consists in a tendency to make decisions in an indeterministic, uncontrollable way makes it unclear either that we are free, or why we would want to be.

"The implications of advances in neuroscience for freedom of the will." Neurotherapeutics 4.3 (2007): 557.