_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
December 01, 2003
Free Will Starts...Now
Two decades ago, a neuroscientist named Benjamin Libet published a classic experiment on conscious will. He had his subjects rest a finger on a button as they stared at a specially designed clock. It had only one hand, which swept through a revolution once every 2.5 seconds. Libet would ask his subjects to push the button at their own choosing. In some runs, he asked them to note the position of the clock hand when they actually pushed the button. In other runs they had to note its position when they first began to think about pushing it.
Libet measured the brain activity of his subjects with EEG, and also attached electrodes to their hands to monitor their muscle activity. His subjects turned out to be good at timing the moment when they pushed the button, with an accuracy within just a few milliseconds. But they were not so good with their own intentions.
Near the top of the brain there's a region known as the motor area where neurons fire to make the body move in particular ways. Libet found that EEG recordings from the motor area in his subjects' brains began to shift into a new pattern 1.5 seconds before the subjects pressed the button. Libet interpreted this as the mental preparation that goes into initiating an action. But his subjects consistently claimed that they began thinking about moving their hand about half a second after the EEG recordings began to change. In other words, they had already started preparing to make a voluntary movement for half a second before they felt like they were making a voluntary movement.
A lot of scientists have questioned different parts of Libet's experiments over the years, but the results have held up pretty well. It seems that we only become conscious of our will after we begin to do something.
This week a team of European neuroscientists published a fitting tribute to Libet for the twentieth anniversary of his experiment. They ran Libet's experiment again, but some of the people they chose as their subjects had damage to certain parts of the brain. As they report in Nature Neuroscience, some kinds of brain damage make no difference to people's performance. But something fascinating happened to people who suffered damage to the parietal cortex, located at the back of the head. Like the healthy controls, they could nail the moment they actually pressed the button, to within a few milliseconds. But they also noted that they intended to press the button just around the time they actually did press the button. In other words, they were completely unconscious of their action until the action was already taking place.
What's most fascinating about this experiment is that the subjects with a damaged parietal cortex do not act all that strangely. They are perfectly able to carry out their will--to press buttons when they want to, to say what they want to say, to walk down the street where they want to go. Their very ordinariness offers a hint about the nature of both consciousness and will.
The evidence from many different studies suggests that intentions, plans, and similar thoughts are born within the prefrontal cortex at the front of the brain. These prefrontal neurons send out branches into a number of other regions of the brain, where models of intentions can be created. These models create predictions--if I do this, I should expect this sensory feedback. If I don't get that feedback, I've screwed up.
Some of these models may be completely unconscious--they may offer quick checks between what we expect to feel when we do something and what we really feel. There's also a model in the parietal cortex, but the authors of the new study suggest it does something different. It works at a higher level, seeing whether our actions match up with their desired goals. Normally the intentions formed in the prefrontal cortex trigger a model in the parietal cortex as the brain prepares the action in the motor area. Our conscious experience usually depends on the model, and not the intention itself.
Posted by Carl at 8:50 PM
I’m a computer programmer, and the issue of “conscious” thought is fascinating from an AI perspective, mostly because we aren’t very. In the early days, programmers found they could get computers to do all sorts of things that humans found “hard” – mathematical proofs, various sorting tricks, and so on. There was a great deal of optimism about computer speech, vision, and so on. After all, if these hard things (for people) were so easy on a computer, then mustn’t easy things for people be even simpler for a computer? Just work out the simple algorithm, and away you go!
What they didn’t understand is that we are unconscious of much of what our brain does, and vision, among other things is in fact a very complex process and requires plenty of hardware to do the job. It makes sense in hindsight – “consciousness” is sort of like a computer’s operating system. We don’t need to know HOW our brains work, but we do need something to manage the resources and interface with the real world. It’s an evolutionary point really – understanding itself is useless to the species, but efficient use of its resources is vital.
It’s amazing how many people talk about the wonders of a conscious computer, but that is meaningless. What we really would like is a computer that is unconscious in the same way we are.
Couldn't the result of the Libet eperiment just mean that we are bad at observing an outside factor (the time on the clock) in relation to a more abstract factor (the intention of pushing the button), while we are much better at correlating two outside factors (the time of the clock and the sensory feedback of a digit touching a button)?
It could also be that it's much easier to pick a point on the clock and think 'this is when I push the button', than it is to pinpoint the start of a thought.
It?s amazing how many people talk about the wonders of a conscious computer, but that is meaningless. What we really would like is a computer that is unconscious in the same way we are.
I think it would be better the other way around. Namely, that we use computers to help make our thoughts more conscious.
Libet's experiments might be interpreted that, if I have brain activity indicating a certain action a second and a half before I think that I am initiating that action, then I can't really say that I have free will.
I understand this interpretation to imply something like this: actions take place between things and lead from one event to the next causally; so too my actions are, and have all been, determined by unending series of causal events; my consciousness has never and can never play a causal role in these determinations, rather it is a voyeur whose presence, which itself may or may not be attributal to the causal order of events, is only erroneously linked to the coming to be of certain actions; such erroneous links between consciousness and action are called instances of free will.
I take the authors comments about parietally damaged individuals to be tending toward this interpretation as well--again, that consciousness is a secondary on-looker and not essentially linked to the playing out of our actions.
However, I find it curious that he still refers to the action of the subjects as "voluntary." Is this a mere equivocation? Or is free will--i.e. voluntary action--present yet beyond the purview of conscious thought?
For Kant the essential factor in the question of free will was whether or not spontaneous events exist within the order of things. If "voluntary" is a mere equivocation in the article, then the subjects did not push the button "at will," but rather at a moment determined by the definite concurrence of an innumerable series of causal events (events which are spatial, neurological, hopelessly historical, etc., but which need not be arbitrarily predictable as such). If, however, no such concurrence determined each pushing of the button, but said event occured spontaneosly, that is, as a self-determined cause placed, at least to some genuine degree, arbitrarily within the greater series of events circumscribing it, then said event is indeed an instance of voluntary action and the subjects, by whatever agency, pushed the button at will. If this latter possibility be, in fact, the case, then we can interpret the results of Libet's experiments to show, not that free will does not exist, but that free will is independent of conscious thought. This interpretation would imply, as does the above interpretation, that our conscious selves are voyeurs to the events in reality; unlike the above interpretation, however, this latter interpretation leaves us to inquire into the agency of individual spontaneity, as well as into the ultimate relation the conscious self has to said agent. This latter inquiry is of particular concern given the experiential link between conscious thought and manifest action--what is normally called instances of free will.
It may be objected that if conscious thought is not the agent of spontaneous action then spontaneous action is not free will properly so called. I argue that spontaneous action need only be limited to the purview of life-energy (i.e. living things) to be free will properly so called and, further, that any objection which distributes spontaneity to all real things, animate and inanimate alike, is entailed, in light of the example of spontaneity furnished by Libet's experiments, by the notion that rocks and waterfalls have the proximate ability to behave like trees and humans. While it may be the case that nothing essential separates the behavior of animate and inanimate objects, it is my opinion that the drive to stay alive is essentially different from the interactions of inert entities (quantum or otherwise).
Thus, in my opinion, Libet's experiements do not controvert the possibility of free action, but they do point to a dividing line between what actions end up playing out in our lives and what exactly it is that consciousness brings to our lives.
Libet experiment has a fatal flaw, which was pointed out by Dennett: It assumes that there is a rigid threshold of consciousness.
That is, if something goes beyond a certain point that thing is said to be conscious. The error is fixing this point at an exact spot. Closer exmaination of consciousness reveals that: (from Can There be A Test for Cons.)
Dennett particularly stresses this "ill-defined boundary" of consciousness. The effect of this is that it is unavoidably arbitrary to set the threshold of emergent consciousness anywhere between oneself at one extreme, and, at the other, the sperm and egg of one?s conception, the earliest replicators, or the simplest phage. Moreover, Dennett's "multiple-drafts" model asserts that there is no definite time or place at which an event in the brain becomes conscious between unconscious-ness and consciousness, within an individual, there is another continuum, and it is thus equally arbitrary to set any kind of threshold of consciousness in one?s own brain.
And this makes sense; Everyone can re-count times when they've been more conscious than others. So what we saw in Libet's experiments is an emerging sense of consciousness. We still have Free Will; it's just really blurry and ambigious.