The classic problem of free will is to reconcile an element of freedom with the apparent determinism in a world of causes and effects, a world of events in a great causal chain.

Determinists deny any such freedom.

Compatibilists redefine freedom as the fact that our wills (determined as they are by prior events) are nevertheless part of the causal chain and so we can feel responsible for our actions.

Libertarians think the will is free when a choice can be made that is not determined or necessitated by prior events. The will is free when alternative choices could have been made with the same pre-existing conditions.

Freedom of the will allows us to say, "I could have chosen (and done) otherwise."

In a deterministic world, everything that happens follows ineluctably from natural or divine laws. There is but one possible future.

In the more common sense view, we are free to shape our future, to be creative, to be unpredictable.

From the ancient Epicureans to modern quantum mechanical indeterminists, some thinkers have suggested that chance or randomness was an explanation for freedom, an explanation for the unpredictability of a free and creative act. A truly random event would break the causal chain and nullify determinism, providing room for human freedom.

Freedom of human action does require the randomness of absolute unpredictability, but if our actions are the direct consequence of a random event, we cannot feel responsible. That would be mere indeterminism, as unsatisfactory as determinism.

Moreover, indeterminism appears to threaten reason itself, which seems to require certainty and causality to establish truth, knowledge, and the laws of nature.

Most philosophers in all ages have been committed to one or more of the dogmas of determinism, refusing to admit any indeterminism or chance. They described the case of "indeterminism is true" as a disaster for reason. They said chance was "obscure to human reason." They found "no medium betwixt chance and necessity."

Many scientists agreed that science was predicated on strict causality and predictability, without which science itself, considered as the search for causal laws, would be impossible.

For those scientists, laws of nature would not be laws if they were only statistical and probabilistic. Ironically, some laws of nature turn out to be thoroughly statistical and our predictions merely probable.

Fortunately, for large objects the departure from deterministic laws is practically unobservable, probabilities become indistinguishable from certainties, and we can show there is an "adequate determinism" and a "soft causality."

We review the history of the free will problem.

We then summarize the requirements for free will, and propose a working solution based on the past and current ideas of those philosophers and scientists who have addressed the free will problem.

Recent debate on the free will problem uses a taxonomy of positions that has caused a great deal of confusion, partly logical but mostly linguistic. Let's take a quick look at the terminology.

At the top level, there are two mutually exclusive positions, Determinism and Indeterminism.

Under Determinism, two more positions conflict, Compatibilism and Incompatibilism.

And under Indeterminism, Robert Kane in the Oxford Handbook of Free Will distinguishes three positions recently taken by Libertarians - Non-Causal, Agent-Causal, and Event-Causal.

Instead of directly discussing models for free will, the debate is conducted indirectly.

Is free will compatible with determinism? is a frequently asked question. Most philosophers answer yes and describe themselves as compatibilists. They call libertarians "incompatibilists."

Is determinism true? is another frequent question. The answer, at least in the physical world, is now well known. Determinism is not "true." The physical world contains quantum randomness - absolute chance.

Chance does not mean that every event is completely undetermined and uncaused. And it does not mean that chance is the direct cause of our actions, that our actions are random in any way.

Nevertheless, the typical argument of determinists and compatibilists is that if our actions had random causes we could not be morally responsible.

To avoid the obvious difficulty for their position, most compatibilist philosophers simply deny the reality of chance. They hope that something will be found to be wrong with quantum mechanical indeterminism. Chance is unintelligible, they say, and thus there is no intelligible account of libertarian free will. Some dismiss free will (as many philosophers denied chance) as an illusion.

Recently, professional philosophers specializing in free will and moral responsibility have staked out nuanced versions of the familiar positions with new jargon, like broad and narrow incompatibilism, semicompatibilism, hard incompatibilism, and illusionism.

Awkwardly, the incompatibilist position includes both "hard" determinists, who deny free will, and libertarians, who deny determinism, making the category very messy.

Broad incompatibilists think both free will and moral responsibility are incompatible with determinism. Narrow incompatibilists think free will is not compatible, but moral responsibility is compatible with determinism.

Semicompatibilists are narrow compatibilists who are agnostic about free will and determinism but claim moral responsibility is compatible with determinism.

Hard incompatibilists think both free will and moral responsibility are not compatible with determinism. Illusionists are incompatibilists who say free will is an illusion.

Soft incompatibilists think both free will and moral responsibility are incompatible with strict determinism, but both are compatible with an adequate determinism.

Soft causalists are event-causalists who accept causality but admit some unpredictable events that are causa sui and which start new causal chains.

For those who know indeterminism is the case, at least in the microphysical world, many deny that chance and quantum randomness can be important for free will. Oddly, this includes agent-causalists, who postulate a non-physical origin for causes (like reasons in the agent's mind), and non-causalists, who claim volitions and intentions are simply uncaused.

For the "event-causal" theorists of free will, we can distinguish six increasingly sophisticated attitudes toward the role of chance and indeterminism. "Event-causal" theorists embrace the first two, but very few thinkers, if any, appear to have considered all six essential requirements for chance to contribute to libertarian free will.

1. Chance exists in the universe. Quantum mechanics is correct. Indeterminism is true, etc.

2. Chance is important for free will. It breaks the causal chain of determinism.

3. Chance cannot directly cause our actions. We cannot be responsible for random actions.

4. Chance can only generate random (unpredictable) alternative possibilities for action or thought. The choice or selection of one action must be adequately determined, so that we can take responsibility. And once we choose, the connection between mind/brain and motor control must be adequately determined to see that "our will be done."

5. Chance, in the form of noise, both quantum and thermal, must be ever present. The naive model of a single random microscopic event, amplified to affect the macroscopic brain, never made sense. Under what ad hoc circumstances, at what time, at what place in the brain, would it occur to affect a decision?

6. Chance must be overcome or suppressed by the adequately determined will when it decides to act, de-liberating the prior free options that "one could have done."

Of those thinkers who have considered most of these six aspects of chance, a small fraction have also seen the obvious parallel with biological evolution and natural selection, with its microscopic quantum accidents causing variations in the gene pool and macroscopic natural selection of fit genes by their reproductive success. Biology affords other examples of two-stage processes, with first chance, then adequately determined choice. For example, the immune system.