In a news item in the September 1 issue of Nature, “Taking aim at free will” (free online), Kerri Smith recounts the latest findings of neuroscience about how and when we make “decisions,” and how that bears on philosophical issues of free will. The two-page piece is worth reading for its exposition of the latest research (some not yet published), and how philosophers are reacting to it.
The research, as we’ve discussed before, largely involves experiments that force participants to make decisions, and doing simultaneous brain scans that can a) “predict” the decision (albeit not with perfect accuracy) and b) find out when the brain actually takes action. Those studies, pioneered by Benjamin Libet and continued in more sophisticated form by John-Dylan Haynes, involve scanning the brains of subjects who are forced to make choices, and comparing when the brain registers a choice with when the subject becomes conscious of having made that choice. All the studies find that brain scans can predict, sometimes with high accuracy, which decision will be made, and that the brain activity occurs up to several seconds before the subject records having made a decision.
Here’s an example of Haynes’s recent findings:
Haynes. . . has replicated and refined his results in two studies. One uses more accurate scanning techniques to confirm the roles of the brain regions implicated in his previous work. In the other, which is yet to be published, Haynes and his team asked subjects to add or subtract two numbers from a series being presented on a screen. Deciding whether to add or subtract reflects a more complex intention than that of whether to push a button, and Haynes argues that it is a more realistic model for everyday decisions. Even in this more abstract task, the researchers detected activity up to four seconds before the subjects were conscious of deciding, Haynes says.
Another study by Itzhak Fried, a scientist and neurosurgeon:
He studied individuals with electrodes implanted in their brains as part of a surgical procedure to treat epilepsy. Recording from single neurons in this way gives scientists a much more precise picture of brain activity than fMRI or EEG. Fried’s experiments showed that there was activity in individual neurons of particular brain areas about a second and a half before the subject made a conscious decision to press a button. With about 700 milliseconds to go, the researchers could predict the timing of that decision with more than 80% accuracy. “At some point, things that are predetermined are admitted into consciousness,” says Fried. The conscious will might be added on to a decision at a later stage, he suggests.
More than 80% accuracy!
The experiments show, then, that not only are decisions made before we’re conscious of having made them, but that the brain imagery can predict what decision will be made with substantial accuracy. This has obvious implications for the notion of “free will,” at least as most people conceive of that concept. We like to think that our conscious selves make decisions, but in fact the choices appear to have been made by our brains before we’re aware of them. The implication, of course, is that deterministic forces beyond are conscious control are involved in our “decisions”, i.e. that free will isn’t really “free”. Physical and biological determinism rules, and we can’t override those forces simply by some ghost called “will.” We really don’t make choices—they are made long before we’re conscious of having chosen strawberry versus pistachio ice cream at the store.
We’ve discussed this issue before, and have seen how some philosophers like Daniel Dennett, and many of the commenters here, aren’t bothered by this: they simply redefine “free will” as something more sophisticated than what I see as the common idea (i.e., were we to relive a moment of decision, we could have decided the other way). Nevertheless, the neuroscience clearly perturbs the philosophers. Here’s how Walter Glannon, a philosopher at the University of Calgary, reacts:
And if unconscious brain activity could be found to predict decisions perfectly, the work really could rattle the notion of free will. “It’s possible that what are now correlations [he's referring to the correlations between specific areas of brain activity and the decision that's made after that activity occurs] could at some point become causal connections between brain mechanisms and behaviours,” says Glannon. “If that were the case, then it would threaten free will, on any definition by any philosopher.” . . . If neuroscientists find unconscious neural activity that drives decision-making, the troublesome concept of mind as separate from body disappears, as does free will. This ‘dualist’ conception of free will is an easy target for neuroscientists to knock down, says Glannon. “Neatly dividing mind and brain makes it easier for neuroscientists to drive a wedge between them,” he adds.
In other words, Glannon recognizes the problem that pre-conscious “decisions” pose for free will. But although he says the threat is to “any definition by any philosopher,” not all philosophers agree. The “compatibilist” school, for instance, manages to reconcile complete physical determinism of decisions with some notion of “free will.”
This shows, as Kerri Smith points out, that philosophers are revising the definition of “free will” in light of these neuroscientific findings. This reminds me of how theologians redefine the meaning of Adam and Eve in light of genetic findings that we didn’t all descend from two ancestors, although I have a lot more respect for philosophers than for theologians.
There are conceptual issues — and then there is semantics. “What would really help is if scientists and philosophers could come to an agreement on what free will means,” says Glannon. Even within philosophy, definitions of free will don’t always match up. Some philosophers define it as the ability to make rational decisions in the absence of coercion. Some definitions place it in cosmic context: at the moment of decision, given everything that’s happened in the past, it is possible to reach a different decision. Others stick to the idea that a non-physical ‘soul’ is directing decisions.
This sounds to me very much like post hoc rationalization. What does it mean to “make rational decisions in the absence of coercion” if that decision has already been made? All it means is that our brains can cough up “rational” outputs in the face of diverse inputs, which of course is what they were evolved to do. And the second definition (which, by the way, is also my own), doesn’t solve the problem: if you’re a determinist (and what else is there besides molecules, genes, environments and physical forces?), there’s no possibility of deciding “otherwise” if all else is equal. Even the compatibilist commenters on this site don’t believe that, at any moment, with all conditions identical, we could make two different decisions. The third re-definition, of course, is bogus, since there’s simply no evidence for a non-physical “soul” that can guide our actions.
In the end, though, I think philosophers are bothered by the science. Al Mele, a philosopher who’s participating in a Templeton-funded study that involves scientists, philosophers and perhaps theologians (I hope not!) weighs in:
Imagine a situation (philosophers like to do this) in which researchers could always predict what someone would decide from their brain activity, before the subject became aware of their decision. “If that turned out to be true, that would be a threat to free will,” says Mele.
Well, this is only my feeling, but I think this is precisely where neuroscience is going. We can already predict some decisions with 80% accuracy. This will only improve as neuroscience becomes more sophisticated.
Beyond redefinition, there’s another way critics attack experiments like Haynes’s and Fried’s: they go after the the methodology:
Philosophers who know about the science, she [Adina Roskies, a philoospher and neuroscientist from Dartmouth] adds, don’t think this sort of study is good evidence for the absence of free will, because the experiments are caricatures of decision-making. Even the seemingly simple decision of whether to have tea or coffee is more complex than deciding whether to push a button with one hand or the other.
I find that criticism unconvincing. How, exactly, is deciding between coffee and tea more “complex” than deciding which button to press? And suppose you did the same experiment, but instead of using a button, just open a window in front of the subject behind which there is a cup of coffee and a cup of tea. If we could associate brain activity with their coffee vs tea preference, I’d bet you’d still get Fried-ian results: the brain would show a decision well before the subject was conscious of having made one.
Why is all this important, and not just a debate about philosophy? The answer is obvious: whether our actions are predetermined has obvious consequences for how and why we hold people responsible for their actions. As I’ve said several times before, the law already takes “responsibility” into account by treating criminals differently depending on whether their actions may have been caused by extenuating circumstances like mental illness. Nobody, I think, would refuse to consider the possibility that an act of aggression may have been caused by a tumor in the criminal’s brain.
The more I read about philosophers’ attempts to redefine and save the notion of “free will” in the face of the neurological facts, the more I think that they’re muddying the waters. I believe that the vast majority of nonphilosophers and laypeople hold a consistent definition of free will: that we really do make decisions that are independent of our physical make-up at the moment of deciding. If this isn’t the case, we need to know it. Yes, it may be depressing—Haynes admits that he finds it hard to “maintain an image of a world without free will”—but we can still act as if we had free will. We don’t have much choice in that matter, probably because we’re evolved to think of ourselves as choosing agents. But rather than define free will so we can save the notion in some sense (this is like substituting the word “spirituality” for “religion”), why don’t we just rename the concept we’re trying to save? Otherwise we’re just giving false ideas to people, as well as providing succor for religion, where the idea of real free will—the Holy Ghost in the machine—is alive and crucially important.
h/t: John Brockman