Bad News for Free Will
The Libet Experiments showed we have no control over our actions. Or did they? Alfred Mele still managed to write a critique of them
No matter what view you take of the exact nature of free will, it is difficult not to see it as indissolubly linked to the idea of conscious choice. But suddenly that timehonoured association has been challenged. American psychologist Benjamin Libet has been exploring free will in his laboratory and come to the conclusion that none of us ever consciously decide anything. Our brains decide for us, before we are conscious of the decision. However you look at it, this seems like bad news for free will. Libet's work has attracted a lot of attention. In the journal Consciousness and Cognition, after reporting that "many of the world's leading neuroscientists have not only accepted our findings and interpretations, but have even enthusiastically praised these achievements and their experimental ingenuity" and naming twenty such people, Libet adds: "most of the negative criticism of our findings and their implications have come from philosophers and others with no significant experience in experimental neuroscience of the brain." Of one of his critics, Libet says: "As a philosopher, [Gilberto] Gomes exhibits characteristics often found in philosophers.
He seems to think one can offer reinterpretations by making unsupported assumptions, offering speculative data that do not exist and constructing hypotheses that are not even testable." (Gomes's articles didn't look like a philosopher's work to me, so I emailed him about this: turns out he's a member of a psychology department.)
What is Libet's evidence for his bold thesis about decisions? (Is it better than his evidence about Gomes's profession?) Libet needs to relate measurable brain events to decisions. One of his findings is a noticeable ramplike shift in electrical readings taken from the scalp about half a second (550 milliseconds) before muscle motion begins in subjects instructed to flex their right wrists whenever they wish. (A subject 'spontaneously' flexes many times during a session, an electromyogram detects the onset of muscle motion each time, and an EEG produces the electrical readings from the scalp.) Libet claims that the electrical shifts indicate decisions, and it isn't hard to see why he does. He believes that actions are caused by decisions, and the electrical shifts seem to be correlated with causes of actions because they regularly precede the onset of muscle motion by about half a second.
How does consciousness come into this? Libet's subjects are also instructed to look at a clock face and recall where a rapidly revolving spot on it was when they first became conscious of things that Libet variously describes as "intentions", "urges", "wanting", "decisions", and "wishes" to flex. On average, the beginning of the electrical shift precedes what subjects identify as the time at which they become conscious of a decision, urge, or whatever ("time W") by 350 milliseconds (ms). Time W precedes the beginning of muscle motion by about 200 ms. (Diagram 1).
If, as Libet claims, the decisions were made when the electrical shifts begin, and if the subjects' 'consciousness' reports were accurate, their brains made their decisions hundreds of milliseconds before the subjects became conscious of them. This obviously is a disturbing prospect for believers in free will, given the important place conscious deciding has in it. If you don't make your decisions consciously, you're not in the driver's seat; so you're not exercising free will.
A lot can happen in the brain in half a second. Is it absolutely clear that the onsets of electrical shifts indicate decisions rather than indicating noteworthy events that precede decisions?
Philosophers, like most people, recognize that deciding to do something differs from having an urge or wanting to do something. For example, you might have an urge to scream at an irritating coworker but decide not to. Might it be that, in Libet's laboratory, the brain generates urges rather than decisions of which the person is not conscious? If so, would this have implications about free will?
Since I'm pretending very poorly, I know to be a journalist rather than an academic philosopher, and I want to present both sides of the story without editorializing, I'll invent a philosopher to give you his personal take on these questions. His name is Phil; like most philosophers, he's not also a neuroscientist.
Phil observes that two additional experiments are relevant. In experiment P, Libet's subjects are instructed to watch the clock and flex when the revolving spot reaches a prearranged time. Electrical readings from the scalp show a distinctive ramplike shift that begins earlier than the shift in the main experiment.
In experiment V, his subjects are instructed both to prepare to flex at a targeted time and "to veto the developing intention/preparation to act . . . about 100 to 200 ms before the prearranged clock time." They receive both instructions together. The ramp produced in this experiment is very much like the ramp in experiment P except that it peters out about 150250 ms before the prearranged time.
Possibly, the onset of the ramp in experiment P represents a decision to flex soon. And because the ramp in experiment V looks the same as the one in experiment P until subjects perform their vetoing feats, it might be thought that the onset of that ramp also represents a decision to flex. "But it can't," Phil says. Here's why. Suppose we're playing a game resembling 'Simon says', and I say to you, 'One second from now, hop forward and hop backward at exactly the same time.' You know that no one can do that. And because you know you can't do it, you can't decide to do it. Similarly, a subject who is instructed, in effect, not to flex at a certain future time and is decided on complying, can't simultaneously be decided on flexing at that future time. So whatever the onset of the ramp in experiment V represents, it is not a decision to flex.
"Assume that the two ramps represent the same thing until they start looking different," Phil urges. The thing they represent can't be a decision to flex at the prearranged time, as we just saw. So what might it be? "An urge to flex," Phil suggests. The subjects in experiment P would seem to have such an urge. How about those in experiment V? Perhaps their wanting to comply with the instruction to prepare to flex at the appointed time together with their recognition that the time is approaching produces a growing urge to flex, or a simulation of such an urge, or the motor preparedness typically associated with such an urge. And the petering out of the ramp around 150250 ms before the targeted time might indicate a consequence of their 'vetoing' the urge. Here's Phil's punchline: "Since the onsets of the ramps in these two experiments do not represent decisions to flex, the same may be true of the ramp in the main experiment."
Libet says that "the final 'act now' process" in the brain "produces the voluntary act," and, he adds, "If the 'act now' process is initiated unconsciously, then conscious free will is not doing it." Imagine that, in Libet's subjects, an unconscious urge to flex 'soon' helps to generate a decision to flex 'right now' that helps to produce a flexing. Does the 'act now' process start with the onset of the urge, the onset of the decision, or something else? Suppose we say it begins with the unconscious emergence of an urge to flex about half a second before the muscle moves and the urge plays a role in producing the decision many milliseconds later. We can then agree with Libet that, given that the "process is initiated unconsciously, [...] conscious free will is not doing it" that is, is not initiating the 'act now' process. But who thinks that conscious free will is in the business of producing urges? In the philosophical literature, Phil observes, free will's primary locus of operation is deciding, and for all Libet has shown, his subjects make their decisions consciously after they become conscious of an urge to flex.
Libet asks, "How would the 'conscious self' initiate a voluntary act if, factually, the process to 'act now' is initiated unconsciously?" As Phil points out, Libet himself says that "conscious volitional control may operate . . . to select and control ['the volitional process'], either by permitting or triggering the final motor outcome of the unconsciously initiated process or by vetoing the progression to actual motor activation." In triggering the final motor outcome, a conscious decision to flex one's wrist right now would be initiating an action in a more direct way than does the unconscious urge that initiated the process that issued in the decision.
So, according to Phil, even if we take Libet's data from the main study at face value, there is a reasonable way of interpreting them that doesn't threaten free will at all. (Diagram 2).
It may be that in Libet's laboratory, urges to flex arise before his subjects are conscious of them. Apparently, the subjects are free to 'veto' these urges and free to act on them. It's up to them whether or not to act on their urges, and some report spontaneously deciding not to act on a felt urge to flex soon. The observation that urges like this arise unconsciously is absolutely no cause for worry about free will. "Even expert neuroscientists should understand that," says Phil, who is still irked by Libet's remark about philosophers.