When we walk into a coffee shop, we think it is up to us whether to have an espresso, cappuccino, or strawberry-toffee-flavored latte with soymilk. Similarly, when we think about more serious matters—which job to apply for, whether to get married, or whether to sacrifice our self-interest to do the morally right thing—we tend to think those choices are up to us, too. Of course, what we do is constrained by our environments, means, and habits. We are susceptible to subconscious influences and nudges, as psychologists and marketing experts know too well. But there still seems some room for choice. When I had my coffee this morning, I could have had a tea instead. This, in a nutshell, is the idea of free will: people have the capacity to choose and control their own actions.

An increasing number of popular science writers and some scientists are telling us that free will is an illusion.

Yet an increasing number of popular science writers and some scientists are telling us that free will is an illusion. The author Sam Harris and the biologist Jerry Coyne are just two prominent examples. When asked by Edge “What scientific idea is ready for retirement?” Coyne, for instance, volunteered “free will,” writing, “Our thoughts and actions are the outputs of a computer made of meat—our brain—a computer that must obey the laws of physics.” Recently this line of thinking has even made it into popular writings by scholars in the humanities, as well. In his latest book, 21 Lessons for the 21st Century (2018), the historian Yuval Noah Harari speculates that in the age of big data, “‘free will’ will likely be exposed as a myth”—and that this, in turn, has significant ramifications, among them that “liberalism might lose its practical advantages.”

According to the skeptics, human actions aren’t the result of conscious choices but are caused by physical processes in the brain and body over which people have no control. Human beings are just complex physical machines, determined by the laws of nature and prior physical conditions as much as steam engines and the solar system are so determined. The idea of free will, the skeptics say, is a holdover from a naïve worldview that has been refuted by science, just as ghosts and spirits have been refuted. You have as little control over whether to continue to read this article as you have over the date of the next total solar eclipse visible from New York. (It is due to take place on May 1, 2079.)

Such free-will skepticism may not yet be embraced by the general public. Nor is it new; the philosophical debate about whether free will is compatible with determinism stretches back centuries, and the modern scientific debate has been roiling at least since the famous neuroscience experiments on the alleged neural causes of voluntary actions conducted by Benjamin Libet in the 1980s. Still, this skepticism makes trouble for some deeply held views about ourselves. The idea of free will is central to the way we understand ourselves as autonomous agents and to our practices of holding one another responsible.

Lawyers, for example, are well aware of that, and the questions that neuroscience raises for the law have become a growing area of study in legal scholarship. How, for instance, could we blame and punish people for something they did not do out of their own free will? When an avalanche harms someone, it would not occur to us to blame the avalanche: unlike you and me (at least as most people see it), it is not a moral agent capable of responsibility. When a person harms another, we hold that person responsible. If the skeptics are right, this is a mistake. In both the human case and the avalanche, the skeptics say, the harm results from physical processes inside a heap of atoms and molecules.

The idea of free will, the skeptics say, is a holdover from a naïve worldview that has been refuted by science—just as ghosts and spirits have been refuted.

Some free-will skeptics—including Harris and Coyne but also the philosophers Derk Pereboom and Gregg Caruso—welcome these implications. They point out that, as a society, we are far too obsessed with responsibility, punishment, and retribution. Many of the world’s criminal justice systems are inhumane as well as counterproductive. The skeptics have a point here, but one can support criminal justice reform while holding on to one’s belief in free will. Human dignity and restorative justice should be reasons enough to focus more on rehabilitation and reintegration of offenders and on tackling the social background conditions of crime. Giving up on the idea of free will, by contrast, would have other unsettling implications, independently of anything to do with blame and punishment. For example, how could we sincerely deliberate about important choices if we didn’t take ourselves to be free in making those choices? The philosopher Immanuel Kant already understood this problem when he noted that we must view ourselves as free when we engage in practical reasoning.

It is important to ask, then, whether free will can be defended against the skeptical voices, or whether, instead, its defenders are clinging to a superstition. I think that science has not refuted free will, after all. In fact, it actually offers arguments in its defense.

• • •

Contemporary free-will skepticism—at least of the kind that appeals to science—is part and parcel of a reductionistic worldview, according to which everything is reducible to physical processes. If we look at the world from the perspective of fundamental physics alone, then we will see only particles, fields, and forces, but no human agency, choice, and free will. Human beings, like everything else, will look like subsystems of a large impersonal physical universe. Of course, skeptics say that this is, in fact, what science implies. To suggest that human beings are anything beyond physical systems would be to revert to seventeenth-century metaphysics, the sort of mind-body dualism endorsed by René Descartes from which modern science has moved on.

Science has not refuted free will, after all. In fact, it actually offers arguments in its defense.

But it is a mistake to equate science with reductionism. Science does not force us to think of humans as nothing more than heaps of interacting particles. To the contrary, in the sciences of human behavior—from anthropology and psychology to economics and sociology—it is standard practice to think of people as intentional agents with a capacity for making choices and responding intelligently to their environments. Scholars in these fields explain human actions by depicting people as choice-making agents with beliefs and desires, goals and plans, on the basis of which they decide which actions to pursue. Different academic fields spell out the details in different ways—with different levels of emphasis on, for instance, the relationship between individual and structural factors influencing human actions—but the general supposition of intentional agency is nonetheless present in all of them. This explanatory practice does not assume anything supernatural. It just reflects the fact that agency, intentionality, and choice are essential postulates if we wish to make sense of human behavior. So, the first point to note is that science would have a hard time explaining human behavior if it didn’t view people as choice-making agents.

To illustrate, think about how we answer familiar questions about humans. Why does someone who has made an appointment normally show up? Why does a taxi driver take you to your specified destination? Why do consumers respond to price changes? Why do people support the political movements they do? In each case, the picture of humans as choice-making agents helps us to give the answer. The behaviors in question are readily intelligible if we think of people as having agency, intentionality, and choice. They are faced with different options, look at these options from their perspective, and select one of the options in a goal-directed and more or less intelligible manner, even if the resulting choices are not always fully rational. If we thought of people as mere physical machines, we would miss the intentional, goal-directed nature of their actions and get overwhelmed with physical details. We wouldn’t see the forest for the trees. It would be like trying to explain investors’ market transactions, voters’ electoral choices, or people’s cultural activities from the perspective of particle physics. Physics and even physiology are not the right approaches for explaining human behavior in its full range—holistically, we might say. At most, they can give us some insights into the mechanisms by which agency is generated in physical organisms. This is not to belittle those insights. Human agency and choice are among the most remarkable phenomena the physical world has produced, and as scientists and philosophers will acknowledge, there is much more to be explained. But this does not justify a reductionistic approach according to which the phenomena themselves are overlooked and get to be discounted.

Science does not force us to think of humans as nothing more than heaps of interacting particles.

Now, once we think of human beings in this nonreductionistic way, we are actually presupposing some form of free will, though liberated from supernatural undertones. That there is such a presupposition in our explanations of human behavior is seldom acknowledged, perhaps because free will is such a controversial concept and the practitioners of the relevant sciences may be reluctant to get drawn into metaphysical debates unless strictly necessary. However, free will, soberly speaking, can be defined as the capacity for intentional agency, choice among alternative possibilities, and control over the resulting choices. This capacity—it should be clear—is presupposed when scientists depict people as choice-making agents, whether in anthropology, psychology, economics, or sociology. 

• • •

The skeptics will object that all this is at best a useful fiction, at worst a harmful one. At any rate, they will say, the free-will presupposition is not literally true. But consider how scientists settle questions about what is and is not real.

Why do scientists accept gravity and electromagnetism as real, but not ghosts and spirits? The answer is that science must refer to gravity and electromagnetism to explain physical phenomena, and these properties are indispensable ingredients of a coherent theory of the world, while postulating ghosts and spirits is not only useless but also prone to introducing all sorts of incoherencies. Generally, to figure out whether some entity or property is real, scientists ask two questions: first, is postulating the entity or property necessary for explaining the world, and second, is it coherent with the rest of our scientific worldview? If the answer to both questions is “yes,” then the entity or property meets the reality check, and scientists feel ready to include it in their inventory of the world, at least provisionally.

If the human and social sciences must postulate intentional agency and choice to explain human behavior, then those properties pass the first part of the scientific reality test: they are explanatorily indispensable.

This test, a version of Occam’s Razor, can be applied not just to physics. It also supports the reality of “higher-level” entities and properties such as ecosystems, institutions, and poverty. These, too, must be accepted as real if we wish to explain our world, and they are ingredients of a coherent scientific worldview, even if fundamental physics does not speak about them. When we think about free will through the lens of this test, we get a new perspective. If the human and social sciences must postulate intentional agency and choice to explain human behavior, then those properties pass the first part of the scientific reality test: they are explanatorily indispensable.

What about the second part—coherence with the rest of our scientific worldview? Here, the skeptics will object that if the fundamental laws of physics are deterministic—like the mechanisms of a precise clockwork—then there is no hope of rendering the notion of choice-making coherent. At any point in time, there will be only one possible future sequence of events, given the physical past. Traditionally, physical theories—from Isaac Newton’s classical mechanics to Albert Einstein’s theories of relativity—have tended to represent the world this way.

Furthermore, even though indeterminism and randomness seemed to enter physics with the emergence of quantum mechanics (at least on the well-known Copenhagen interpretation), it is still an open question whether future, more advanced theories will retain this indeterminism. Einstein was famously unconvinced by the idea of indeterministic physical laws when he said, “God does not play dice.” Given that determinism has not been conclusively ruled out by science, therefore, we can’t count on quantum mechanics to defend free will—not to mention that quantum indeterminacies would probably be a farfetched source of free will anyway. Indeed, hard determinist skeptics insist that we never have any real choices. When you appeared to make a choice about whether to read this essay, only one option was genuinely available (reading it, as you are doing right now); the other option never existed.

These are subtle issues, but deterministic physical laws arguably do not preclude forks in the road within human agency. An agent’s future choices can be open at a psychological level even if the underlying physics is deterministic. Though this may sound counterintuitive, the distinction between determinism and indeterminism cannot be drawn independently of the level of description at which we are looking at the world. A system can behave deterministically at one level—say, the microphysical one—and indeterministically at another—say, the level associated with some special science: chemistry, biology, meteorology, and so on.

Deterministic physical laws arguably do not preclude forks in the road within human agency. An agent’s future choices can be open at a psychological level even if the underlying physics is deterministic.

Physicists themselves recognize this point in the field of statistical mechanics, which describes how indeterministic macro phenomena can result from deterministic micro processes. The weather, for instance, is a macro system that behaves indeterministically even though the atmosphere consists of a large number of individual molecules that each move around according to deterministic laws of motion. At a macro level, then, the Earth’s atmosphere can be thought of as indeterministic, despite being deterministic at a micro level. As the philosopher of physics Jeremy Butterfield puts it, a system’s micro- and macro-dynamics need not “mesh.” And, I would argue, such “emergent indeterminism” is not just apparent. It is best interpreted not as “epistemic”—due to incomplete information about the world—but as “ontic”: a feature of what the world is like. So, to cut a long story short, the sciences give us the resources to show that forks in the road in human decision-making can co-exist with determinism in physics. Of course, the openness of human choices is not just a phenomenon of statistical mechanics; it comes from option availability as described by our best explanatory theories of human decision-making.

• • •

For the time being, then, the hypothesis of free will is corroborated by the sciences of human behaviour. Free will, for the purposes of the human and social sciences, boils down to agency, intentionality, and choice, which are well-supported and indeed explanatorily indispensable ideas. Denying free will would be warranted only if these ideas weren’t needed for explaining human behavior or if they were somehow incoherent, which they aren’t.

To be sure, future science might vindicate a reductionistic approach and explain human behavior without representing people as choice-making agents. But science doesn’t seem to be heading that way. So far, psychology, broadly speaking, has resisted reduction and has been augmented but not replaced by neuroscience. Just as we wouldn’t deny the reality of ecosystems, institutions, and poverty merely because fundamental physics doesn’t refer to them, so there is no reason to deny the reality of agency, choice, and free will either. The skeptics’ mistake is to assume a reductionistic picture of humans that is neither mandated by science, nor adequate for understanding human behavior.