SUPPOSE you approach a fruit bowl overflowing with apples, oranges and bananas, all perfectly ripe. On this occasion, you choose an orange. Or do you? Because although it may feel like you were free to pick an apple or a banana instead, many of those who contemplate such things insist that you aren’t at liberty to choose at all.
The same goes for all sorts of decisions we fret over, from the trivial to the consequential. If we could somehow rewind the universe, they say, you would behave in exactly the same way – because that is how your brain is made. Your sense of having free will is just an illusion.
Advertisement
The implications of such a claim are alarming because they force us to rethink many cherished assumptions. If our choices are somehow predetermined, there is no point agonising over moral dilemmas and less justification for punishing people for their crimes. This stuff matters. But can it really be true that none of us, as the poet William Ernest Henley put it in his stirring poem Invictus, is the captain of our soul?
While that question has long been the preserve of the physicists, philosophers and religious scholars, a growing number of biologists are now weighing in. Only this month, two leading neuroscientists have published books on the subject, both of which claim that a new understanding of the brain supports their case. So, with the debate having moved into new territory, can we finally figure out, once and for all, whether we have free will?
Determinism
The question has fascinated people since the time of the ancient Greeks, who debated how free will could be reconciled with what they saw as a “deterministic” universe, where everything that happens is governed by prior causes. The subsequent march of science seemed to support this deterministic world view, from Isaac Newton’s discovery of the laws of motion in the 17th century to the orderly arrangement of elements in the periodic table.
For many, free will is ruled out by classical physics, which tells us that everything that happens in the physical world has a cause, whether we are talking about snooker balls careening around a table or planets whirling around the sun, and is therefore predictable according to physical laws. Our decisions must surely be outputs of our brains, which are ultimately made of physical matter, regardless of how complex they are. And their inner workings must be subject to those same laws of classical physics as planets and snooker balls. If only we had a sufficiently powerful brain scanner – one that could read out the position of every neuron, every synapse, every molecule – our decisions could be predicted before we made them.
A possible lifeline for believers in free will came with the arrival of quantum physics in the early 20th century. This says that subatomic particles can occupy multiple different states at once, existing as a spread of probabilities that only “collapses” into something definite when measured, or observed. While the particles behave classically en masse, when single particles can be observed, they can take a spread of values within a certain range and these seem, as far as we can tell, to be quite random. In other words, the universe, at its most fundamental, is non-deterministic.
A second get-out clause comes from a branch of physics called chaos theory, which describes how complex systems that are highly sensitive to small changes end up behaving in unpredictable ways. This shows that, even with all the information to hand, we couldn’t, in fact, predict everything that happens in the universe.
Yet neither of these two lifelines seems very secure. Maybe our brains do have elements of randomness or unpredictable processes, but that doesn’t sound much like being the captain of your soul.
Which raises the question: should we really be thinking about free will at the level of fundamental physics? Some argue that our capacity for free will, should it exist, would operate not at the level of particles and atoms at all, but in complex networks of brain cells. Decision-making, in other words, is a biological phenomenon, which is where neuroscience enters the fray.
Is free will an illusion?
It began in the early 1980s with a simple experiment carried out by Benjamin Libet, a neurophysiologist at the University of California, San Francisco, that has been endlessly repeated, elaborated on and argued about ever since. The set-up was to ask a volunteer to sit with their hand on a button and to press it after a random interval of their choosing. They were asked to note, by means of watching a timer, the instant they chose to move. They also had an electroencephalogram, or EEG, with electrodes placed on their scalp to monitor changes in brain activity.
Unsurprisingly, Libet’s experiment found that people became aware of their decision to hit the button a fraction of a second before they actually pressed it. What was startling was that activity in the part of the brain that controls hand movements, started to rise before the person believed they had taken the decision – on average, about a third of a second before. The result shifted the free will debate on its axis because it suggests that whatever is making the decision isn’t our conscious experience – the thing that feels like the “captain” inside our heads.
That seems to tally with psychological research demonstrating that we can easily be mistaken about our decision-making processes. When people are tricked into thinking they made a certain choice that they didn’t, they are surprisingly willing to make up plausible “explanations” for their decision. When asked to choose the most attractive person from two pictures, for example, they are later able to justify their choice despite the fact that the experimenter has secretly switched the photos. This phenomenon, sometimes called confabulation, is often interpreted as evidence that just because we feel like we have freely made a decision, that isn’t necessarily the case.
On the other hand, these are contrived circumstances, says Adina Roskies, a philosopher at Dartmouth College in New Hampshire. “There are certainly cases in which you can fool people into thinking they’ve done something when they haven’t. But that does not suggest that all the time, when we think we’ve done something, we haven’t done it,” she says. After all, we don’t take the success of visual illusions as a sign that no one can really see.
Conscious choices
The Libet result has been challenged in other ways, too. Some critics argue that it is an artefact of the instructions given in the experiment. Subjects are told to “let the urge to act appear at any time without any pre-planning or concentration on when to act”. In fact, activity in the relevant part of the brain tends to randomly fluctuate up and down. On being effectively told to act on a whim, people might be interpreting a higher level of unconscious brain activity as their cue to press the button, says , a geneticist and neuroscientist at Trinity College Dublin, Ireland. “The set-up is designed to take any deliberation out of it.”
Perhaps the most compelling criticism, however, comes from a 2019 study that tweaked the Libet experiment, asking subjects to make a more meaningful decision: to press a button to decide which of two charities would receive $1000. This found before people felt they had made the decision.
That suggests the Libet finding applies only to arbitrary decisions, which could be driven by random fluctuations in brain activity, but is irrelevant for deliberative decisions, says at Chapman University in California, who was involved in the study. “Let’s say I can predict when you [press a button] – who cares? But if there’s a burning car with a baby inside, are you going to save the baby? Those are the kinds of decisions we care about.”
While Maoz’s team saw no evidence of the Libet-style brain activity in the movement part of the brain, it is possible that a similar pattern of activity could be happening elsewhere. His team is continuing to look for such a signature using functional MRI scanning, which should be more able to detect activity in deeper regions of the brain, he says.
But Maoz believes that more light will be shed by investigating this question as it relates to animals, even very simple ones. “Free will has really been studied in humans, but there are now a lot of things we can study in animals because we have technology for recording and stimulating [brain] activity in animals,” he says.
The behaviour of simple animals like fruit flies is usually assumed to be highly deterministic, involving reflexive responses to external stimuli. In fact, it is anything but, says at the University of Regensburg in Germany. Brembs and his team have studied the flight patterns of fruit flies placed in a white drum with no visual cues – effectively a sensory deprivation chamber – to remove all possible stimuli. They found that the flies’ zigging and zagging inside the drum was neither deterministic nor random, but bore mathematical hallmarks of chaotic systems and was impossible to predict. More recently, in 2021, another team found similar signatures of chaos in the outputs of neurons that govern when Aplysia sea slugs .
Brembs says similar systems could be at work in human brains when we make decisions. If so, it wouldn’t prove we have free will, but it would at least scupper the argument that, given enough information, it would be possible to predict all our decisions – we couldn’t even predict the twists and turns of a fruit fly. “It could provide some component of choice,” he says.
Yet not all those who study animal brains are so confident that it will release us from the shackles of determinism. In fact, the latest argument against free will comes from Robert Sapolsky, a neuroscientist at Stanford University in California. In his new book, Determined: Life without free will, Sapolsky argues that the more we discover about how the brain works, and the many different influences on human behaviour, the less room there is to squeeze in the capacity for independent decision-making.
Inescapable influences
To take just a handful of examples: geneticists have found hundreds of DNA variants that influence our tendency to be depressed or impulsive; epidemiologists have shown that how well a fetus is nourished during pregnancy can cause enduring “epigenetic” changes to DNA with lifelong consequences for mental health; and social scientists have demonstrated how experiences in our early life may mould our tendencies to be anxious or to commit crimes. “It’s one big seamless arc,” says Sapolsky. “There’s no aspect of behaviour that’s not being influenced by a zillion of these factors.”
We may not yet know everything about how the brain makes decisions, says Sapolsky, but one day we might understand near-enough everything. And even now, he adds, we know enough to suggest we won’t make any startling new discoveries to indicate that the brain can avoid being influenced by such factors. “It would require some findings that defy most of biology to find a place to shoehorn [free will] in. There’s no crack where you could discover that here’s where our neurons, instead of communicating with neurotransmitters, start whispering to each other – and that’s why someone decided to rob the liquor store.”
Sapolsky’s argument goes to the crux of his interest in how we think about morality and criminal justice. As someone who, earlier in his research career, helped to uncover the way that stress hormones can change the brain, Sapolsky has previously acted as an expert witness in court cases, arguing that defendants should get shorter prison sentences because they were abused as children. “I say to the jurors: ‘If you had the exact same upbringing, with the exact same biology, you’d be sitting at the table as the defendant’, ” he says.
Such ideas have been challenged, however. As Roskies points out, a decision can be influenced by something without being determined by it. A gene variant or stress hormone could make a choice more likely without making it 100 per cent certain, for example. “If you think free will exists, that’s not to say that there aren’t influences and constraints on our behaviour,” she says.
And neuroscience isn’t the only branch of biology that has something to say about free will, says Mitchell. “If we want to understand how human beings do things – that is, where causal power [for our actions] comes from – then that question extends back to say: ‘Well, how does any organism do something?'” The answer, says Mitchell, lies in the evolution of biological agency, or the ability to act with intention or purpose, an argument he makes in his new book Free Agents: How evolution gave us free will.
As Mitchell sees it, when the first simple life forms appeared on Earth about 3.7 billion years ago, one of their most interesting essential qualities was that they did stuff. Before that, things happened: grains of sand tumbled around, chemicals reacted and volcanoes spewed out lava. But those were inert physical processes. The first life forms, however, used energy to work against the second law of thermodynamics – the principle that everything tends to become more disorganised over time – and hence stay alive. “Agency is a really core property of living things that we almost take for granted, it’s so basic,” says Mitchell.
It is here that we can find the source of our free will, he says. In fact, this is what the nervous system evolved for. “It is primarily a control system, the job of which is to define a repertoire of actions and choose between them. This control system has been elaborated over evolution to give greater and greater autonomy.”
Rethinking reductionism
For Mitchell, it is a mistake to think we can always reduce complex systems to their component parts and consider causation in nervous systems, say, at the level of their atoms. Contrary to this “reductionist” approach, complex systems like brains can only be understood by considering their interactions at higher levels of organisation.
This view has recently been put on a firmer footing by research that reconsiders the reductionist approach of seeking “microscale” causes. In 2022, Erik Hoel, then a neuroscientist at Tufts University in Massachusetts, analysed more than a dozen different kinds of causation in complex systems, proposed by researchers in fields from statistics to genetics and psychology. In every case, he found some form of “causal emergence” – where the causes of some phenomenon or behaviour emerge not at the microscale but a higher or more coarse-grained level of the system.
In the case of decision-making, then, a person’s thoughts and feelings and memories are as much genuine causal forces as what happens at the level of atoms and neurons. “The idea that every event has a cause is only a problem for free will if it’s taken to mean that every cause is at this lowest level, it’s all physics,” says Mitchell. “But the entire structure of a nervous system can be a cause of things – you can be a cause.” Which suggests we might just be the captains of our souls, after all.
Whether any of this will allow free will sceptics and believers to reach an accord is far from clear. The opposing sides can’t even agree on what it would take to provide clinching evidence either way. Sapolsky will believe in free will only if some aspect of human behaviour can be shown to be completely devoid of prior influences. “Here are the neurons that caused it to happen – show me that they would have done the exact same thing if all the surrounding neurons sculpted by the previous history of your life had been different,” he says.
Mitchell says that is setting the bar impossibly high. “What kind of a being would be behaving free from any prior causes? They wouldn’t have a reason for doing anything, because reasons are past causes – it would just be a random behaviour generator.” But he adds that it is difficult to conceive of anything that would convince him that his argument is wrong, because it isn’t a simple, testable hypothesis. “It’s hard to say there’s some particular experiment that could show we do or don’t have free will,” he says.
Roskies, for one, isn’t holding her breath. “There have been thousands of years of discussion about this issue,” she says. “If there were a simple answer, we would have figured it out by now.”
New Scientist audio
You can now listen to many articles – look for the headphones icon in our app
