A post on the consistently interesting British Psychological Society Research Digest blog has got my head spinning somewhat. It seems that a team of researchers in Berlin have used brain imaging technology to show that when it comes to certain motor decisions, human free will is, in the words of the Digest, “an illusion.”
The researchers scanned the brains of subjects who had been asked to decide whether to press a button with their left or right finger. The subjects indicated when they had made their choice. But the brain scans showed that activity in the frontopolar and parietal cortices roughly ten seconds before the subjects reported their decision was correlated with that decision. In other words, monitoring brain activity in these areas could indicate whether the subject would choose left or right—well before the subject was consciously aware of having made the choice.
Is this a blow for free will? Not necessarily. Ever since Hume, all but the most far-out defenders of the proposition that human beings have some form of free will have acknowledged that our account of free will must not contravene physical laws. Moreover, most acknowledge that a “free” action is not best understood as an “uncaused”—ie random—one. The difficulty for defenders of free will is to come up with an account of what a free action is that doesn’t contradict the laws of science, but that at the same time cannot be reduced to them—for if a fully realised scientific account of the world can explain every human action, it’s difficult to see what room is left for the idea of free will.
But defenders of free will don’t place human actions outside the realm of causality. When the subjects in the experiment described above make their decision to press the button with one finger or the other, their decision is not akin to that made by a random number generator. Their decision will have a cause, or more likely a vast number of causes. Our understanding of the neuroscience of decision-making is presumably nothing like sophisticated enough to be able to tell us what these causes are, but there’s no reason why we should be aware of them, as opposed to being aware of the decision we make as we make it. It just so happens that brain-scanning technology is now sufficiently advanced to be able to detect the first neurological stirrings of those parts of the brain that come into play when we make decisions like these before our conscious understanding. Yet if someone was to place you in this experiment and tell you that you are not making your choices between left and right freely, it wouldn’t be difficult to produce evidence suggesting otherwise.
And yet, and yet… it’s easy to see why evidence like that adduced by the Berlin team might be seen as some as weighing against free will. If a scientist monitoring my brain activity knows what I’m going to do before I do, in what sense can I be said to have made my decision freely? What all this really shows is how intractable the problem of free will is, and how, in the light of increasingly sophisticated levels of scientific, particularly neuroscientific, understanding, the defenders of free will need to produce a correspondingly sophisticated definition of freedom.
Prospect readers might like to note that the free will/determinism problem is given rather more exalted treatment by AC Grayling in the May 2008 issue of the magazine, published later this week.
I think that much of the confusion about the whole “free will” vs determinism issue is that we are not using an agreed definition of “free will”. As stated above, free will does not imply unpredictability - indeed, the consistency of behaviour that indicates the presence of a stable individual lends itself to predictability. Lack of freedom is better displayed by the ability to impose specific choices on a person, regardless of whether these choices are consistent with their preferences or character.
It is also important to see the neurological function as part of a broader human decision making process. I may well be standard human brain procedure to allow a period of inactivity after the decision is “made” in order to be sure that it is a stable decision. This period, if it were to exist, would surely have to be packaged up into the “making a decision” bundle whose freedom is being assessed. To do otherwise would simply prove that computers can be quicker to report results than humans, which, based on current computing power, is surely not that surprising.
The Berlin group’s findings replicate a much older similar discovery made back in the 1980s. The earlier work was based on analyzing brain electrical activity (as recorded by electrodes placed on the scalp’s surface) and was also interpreted by many as a blow against the idea of free-will. The debate emerged in the forum of the journal “Behavioral and Brain Sciences” (8:4, 1985), in an article written by Benjamin Libet which summarized a series of his experiments. The journal’s format includes detailed commentary on each of its articles, so the author’s scientific presentation and interpretation of the results were followed by lengthy discussions of the experiment’s implications by other neuroscientists, psychologists, anthropologists and philosophers. Naturally they tended to disagree with the author and each other, especially on the free-will issue.
The particular neuroscience finding of interest was that there is an approximate one-half second lag between “readiness potentials” (a specific, well-defined electrophysiolgical activity) that always precede conscious decisions and one’s awareness of making the conscious decision, signaled by the subject pressing a button. One interpretation of this is that non-willed aspects of brain activity always anticipate voluntary actions in some weird way, thus indicating that these actions are not fully voluntary in the way in which we believe them to be; the only salve for free-will coming from this particular experiment is that we retain a consciously-controlled veto function over our “conscious” decisions to act, which we can implement in the last two-tenths of a second before beginning the action after it was intended to begin.
A Danish popular science writer, Tor Norretranders, wrote a book called “The User Illusion – Cutting Consciousness down to Size” in which he made a great deal of this finding as it reflects not only upon considerations of free-will, but, more generally, of consciousness (the book was first published in 1991 and in English translation in 1998). The book, which departed from reporting into a kind of diffuse “New Age” advocacy and fairly wooly philosophizing, probably received more attention at the time than the scientific article and its responses did six years earlier. However, it introduced a variety of other neuroscientific evidence and raised some interesting questions about the relative roles of subconscious brain processing and conscious thought or decision-making, pointing out that the vast preponderance of what goes on in the brain and results in our behavioral responses (and our interpretation of these) takes place in a way that is seldom if ever accessible to consciousness (this may or may not say something about free will; that issue remains muddy).
What is of interest in the present connection is that although the brain scan results now move the subconscious “neural anticipation” of a specific decision back by ten seconds, the interpretation of what this means for free-will is still up for grabs. After all it could well mean that the instant of the freely made decision, if that is what is indicated in the scan, merely precedes the conscious act signaling the decision by a longer time than we think reasonable (Kate’s remarks are somewhat along this line and give a good functional reason — but not necessarily a true one — to boot); while it does seem to show that the actual decision is somewhat opaque and not immediate to our conscious minds, the result certainly does not remove the possibility of such free decisions being made “subconsciously” (unfortunately this idea gets us into a kind of annoying and perhaps infinite regress). As Nick indicates above, every last term in such verbal formulations has to be more clearly defined and perhaps, a la Wittgenstein, examined for its “perspicuity” – does it get us to the bottom of things or just lead us into another dead end where words deceive us?
I think the free will/determinism argument could in theory go on till the end of time - because time as we experience it is a one-way street and we can never go back for a do-over to find out, empirically, whether or not we really had a choice.
Thanks all for the useful and stimulating comments. We may have more on this fascinating area in the next issue of Prospect.