In his 2002 book The Problem of the Soul, he devoted a chapter to free will. He says that free will has deep theological roots. It is the idea of a gift of god-like powers that permits us to circumvent natural laws. He cites Roderick Chisholm's idea of "agent causation," which claims that free agents are "prime movers unmoved."

The mind is not an immaterial substance, as Descartes claimed, but embodied as an evolved capacity of the brain. Men are simply human animals. As a result, Flanagan says, free will is a "falsehood" that we are better off without.

Flanagan describes the difficulties involved in the traditional problem of free will, that it is compatible with determinism.

The only conception of free will that has ever been entertained that deserves the name of free will is the Cartesian conception of a mode of mental processing, or a mental faculty, that is totally unconstrained, totally self-caused. The prime mover, itself unmoved. But there is no reason, none, to think that there could be any such thing. It is so conceptually puffed up that it is incredible, incoherent.

Consider what it would mean to have such a free will. When I make a choice I do so ex nihilo, by electing, without anything constraining my deliberation, a course of action. But if nothing constrains my choice, then reasons don't constrain my choice either. And if that is so, then ordinary introspection must be deemed wildly wrong. After all, it seems to most everyone that when they are deliberating among the options at hand that they are weighing pros and cons and that this information constrains the choice.

Second, and just as bad, if when I choose I do so for no reason (choice may create a reason for action but does not itself rest on any reasons) then my choice is either arational or irrational. Since one of the main things — perhaps the main thing — any conception of free will worth wanting is supposed to do is to explain how rational choice is possible, and so to explain how I can be held rationally accountable for my choices, the orthodox conception of free will is a miserable failure. It is conceptually incoherent, in the sense that it provides no coherent way of conceiving of what it wants to gain for itself.

If you were only able to say that the orthodox picture of free will makes no sense from the perspective of the scientific image, you could be rightly accused of begging the question. All you would then be saying would be that what I assume doesn't permit what you assume. But I am making a stronger claim. Upon examination, the orthodox concept of free will makes no sense in terms of the agenda it sets for itself — to explain rational deliberation and choice.

If this is true, then there is no problem of "free will and determinism" worth discussing. There is a problem in the vicinity worth discussing, but free will and determinism is not it. The problem worth discussing is how to make sense of freedom, deliberation, reason, and choice within the framework set out by the human sciences generally and by mind science in particular. This can be done.

But before proceeding, let me mark as clearly as I can where I am positioning myself. The texts that discuss the "free will–determinism" problem take two main positions:

• Compatibilism: Free will is compatible with causal determinism. Most compatibilists say that free will requires causal determinism in the sense that the state of my will (itself determined by prior and contemporaneous causes) must be a sufficient cause of any choice I make.

• Incompatibilism: Free will is incompatible with causal determinism. Incompatibilists take one of two roads. Libertarians claim that since we have free will, determinism is false. Libertarians employ the concept of free will as Cartesian agent causation, or those who sense its incoherence by a promissory hand wave in the direction of "something or other that does the trick but that is yet to be articulated or formulated to anyone's satisfaction."
Hard determinists claim that since determinism is true, there is no such thing as free will.

Given my argument that the normal way of framing the problem — as the problem of "free will and determinism" — makes no sense you will not be surprised to discover that I think all these answers are unsatisfactory and the reason is that the problem is ill-posed. If forced to comment on the three positions I would say this. Libertarianism is a nonstarter because the Cartesian conception of free will, the only conception that has received articulation within philosophy as deserving the name free will, is a nonstarter.

The compatibilist, meanwhile, if he thinks free will is compatible with determinism, must have changed the subject. He cannot be saying that the Cartesian conception of free will is compatible with determinism because, well, it isn't. And indeed if one looks at the literature one will see that compatibilists invariably mean something different by free will than what the orthodox concept says it is.

The hard determinist, unlike the compatibilist, accepts the terms of the exercise as they are set and sees correctly that determinism is incompatible with free will, as the Cartesian conceives it. But both the compatibilist and the hard determinist make the same mistake. They both claim to know that determinism is true. But if what I have said about causation — there being both deterministic and indeterministic causes — is plausible, then neither can sensibly be said to know that determinism is true. Causation is ubiquitous. Ours is a causal universe. But no one yet knows the exact range of deterministic and indeterministic causation — assuming the universe contains some of each.

What to do? My proposal is this: Change the subject. Stop talk about free will and determinism and talk instead about whether and how we can make sense of the concepts of "deliberation," choice," "reasoning," "agency," and "accountability" (scorecard items) within the space allowed by the scientific image of minds. This is, I hasten to admit, just what I accused the compatibilists of doing. Since they cannot be saying that free will is compatible with causation, either deterministic or indeterministic, they must be claiming that something else—hopefully something similar to free will—is compatible with causation.

It would be misleading to call my position compatibilism, however, since compatibilism seems to accept the terms of the standard debate about "free will and determinism." Since I have been trying to frame the pressing question in terms of the compatibility of "rational deliberation and choice and causation," or as the problem of the voluntary and the involuntary, it will be best to call my view neocompatibilism. I do claim that we can make sense of rational deliberation and choice in a causal universe.

(The Problem of the Soul, pp.124-127)

Flanagan calls himself a "neo-compatibilist" with a position similar to that of Daniel Dennett, John Martin Fischer, Harry Frankfurt, and Susan Wolf.

He very neatly summarizes the standard argument against free will.

Free actions, if there are any, are not deterministically caused nor are they caused by random processes of the sort countenanced by quantum physicists or complexity theorists. Free actions need to be caused by me, in a nondetermined and nonrandom manner...

(The Problem of the Soul, p.135)

I am open to there being genuine ontological indeterminacy at both the quantum level and the level of neural processing. But the attempt to gain free will from indeterminacy at the quantum level or at the level of global brain processes is a bad idea The last thing anyone wants is for free will to be the result of random causal processes.

In the two-stage model of free will, our actions are adequate, determined by all these causes

Give me the choice between my actions being strict causal outcomes of my genes, my history, my personality, the state or my mind/brain, and the current environment, and I will take that every day of the week, over the view that when I deliberate or act, I do so randomly, because my will has flown the causal coop and moved arbitrarily into a new, unpredictable place in the causal nexus.

(The Problem of the Soul, p.128)

Flanagan sees two visions of mind, variations on the ideas of Wilfrid Sellars (who set his student Robert Kane, to work on the problem of free will over 50 years ago). Sellars' manifest image and scientific image have become Flanagan's humanistic image and scientific image. Like Sellars, Flanagan is a compatibilist though he prefers "neocompatibilist."

This book is about the conflict between two grand images of who we are: the humanistic and the scientific. The humanistic image says that we are spiritual beings endowed with free will—a capacity that no ordinary animal possesses and that permits us to circumvent ordinary laws of cause and effect. The twentieth-century philosopher Roderick Chisholm sums up the main idea this way: When we act freely we exercise "a prerogative which some would attribute only to God: each of us, when we act, is a prime mover unmoved. In doing what we do, we cause certain things to happen, and nothing—or no one—causes us to cause those events to happen." The scientific image says that we are animals that evolved according to the principles of natural selection. Although we are extraordinary animals we possess no capacity that permits us to circumvent the laws of cause and effect. The question is this: Which is it? The two images, at least as depicted in these terms, are incompatible. The answer can't be both. Or, if it is, there is a lot of explaining to do.

We want to see ourselves truthfully, and we also want our stories to depict life as if it really means something. But we live in a world in which two distinct self-images, vying for our allegiance, disagree about human nature and about the ground of meaning. One image says humans are possessed of a spiritual part—an incorporeal mind or soul—and that one's life and eternal fate turn on the state of this soul. The other image says that there is no such thing as the soul and thus that nothing—nothing at all—depends on its state. We are finite social animals. When we die, we—or better: the particles that once composed us—return to nature's bosom, not to God's right hand. The humanistic image, embraced by most laypersons, scientists, and intellectuals, claims to be uplifting and inspiring. We create ourselves by exercising our free will. If we will well, when we die we reap eternal reward. From the perspective of the scientific image, this idea is extremely implausible, excessively flattering, and self-serving. If it provides meaning, it does so at cost to the truth.

But the scientific image, from the humanistic perspective, is dehumanizing— it drains life of meaning. Life has no transcendent purpose, and the quest to live morally becomes just one among several quirky features of our kind of animal. Defenders of the scientific image claim that their image need not be seen as depressing or inhospitable to a dignified, moral, and meaningful life. Humanists are skeptical. It is part of the humanistic perspective to deem science, especially the mental and human sciences, as a threat. Science is reductive and materialistic, and it offers no resources to help us find our way in the high-stakes drama that is life.

Perhaps the truth about human nature is eternally incompatible with an uplifting story about the meaning of life. The truth can sometimes be painful. Perhaps honestly acknowledging the truth about human nature would necessarily undermine any sense of purpose and meaning, and bring ennui and nihilism in its train. We tell our children stories about Santa Claus and tooth fairies. These stories are false, but they please the kids. Could we be acting like grown-up children who fabricate false stories for our own comfort, for the sake of meaning? Possibly. Some defenders of the scientific image think this is exactly the case.

But there is another possibility. Perhaps the mythic stories we are used to telling about our nature are beloved not because they are indispensable to a meaningful life but only because we have been historically conditioned to think so. Perhaps, too, there is sufficient room in the scientific image for mind, morals, and meaning that it can preserve much of what it means to be a person. If this is so, then the stark inconsistency between the two images is—at some level at least—more apparent than real. This is what I think.

In my experience, most defenders of the scientific image either ignore the dominant humanistic image or deem it silly and misguided, while defenders of the humanistic image simply assert that the scientific image is de-meaning. But both images share a common aspiration: to maintain a robust conception of what it means to be a person, a being possessed of consciousness, with capacities for self-knowledge and the ability to live rationally, morally, and meaningfully. No advocate of the scientific image has yet made an adequate effort to explain carefully, patiently, and explicitly how the scientific image can do this. That is the task I set for myself.

"The problem of the soul" is a shorthand way of referring to a cluster of philosophical concepts that are central components of the dominant humanistic image. These concepts include, for starters, a nonphysical mind, free will, and a permanent, abiding, and immutable self or soul. It is the survival of these concepts that ordinary people fear are at risk from scientific progress, and this fear is at the root of the deep-seated resistance to the scientific image. Ordinary, intelligent people have a (somewhat inchoate) view that nothing less than the meaning of life turns on how these concepts fare. If the nonphysical mind, free will, and the soul are not real things but are mere appearances, then, well, it is the end of the world—at least the end of the world as we know it.

(The Problem of the Soul, pp.ix-xi)