dice

This Still Isn’t Free Will!

The BBC recently published a story about some research about free will, and I am yet again struck by the naive conceptions of free will so many of us have. The article discusses research that as far as I can tell claims that animals, even simple ones, have a range of options available to them, and that the behaviors they engage in are more probabilistic rather than deterministic. Now, this is basically as much hard data as I was able to glean from the article, and yet the article is for some reason about free will. Having not read the original paper that this article is based on (link below), I’m left to draw my own conclusions about what’s gone horribly wrong here, so I’ll focus solely on what’s explicitly stated in the article.

The article discusses how rather than having one set deterministic behavior that will result from a given stimulus (if A then B, every time we have A, B will follow), animals actually have a range of options available to them (if A then B or C or D). Which action gets picked is somewhat probabilistic (B – 30%, C- 50%, D – 20%). It’s argued that this probabilistic behavior is evolutionary beneficial. Being able to behave in an unpredictable variable fashion is likely to confer survival advantages for predators or prey whose actions cannot be predicted by their opponents. The researchers also found that though unpredictable, the behaviors do seem to come from a fixed set of options.

I find every bit of that completely uncontroversial, but have to ask…what does any of this have to do with free will? You know what also behaves in probabilistic ways? Electrons. Does that mean electrons have free will? Probabilistic behavior is not a trump card in the free will debate, but simply an interesting aspect of behavior. I think there are two higher level conceptual reasons for these kinds of misunderstandings regarding free will, that are worth exploring before discussing the content of the article any more.

The first stems from a tricky aspect of human consciousness. Because of our human ability to reflect, to be self aware and self conscious, we are able to contemplate various possible choices in any decision. While we are making a decision we can choose among options, before we make a decision we can contemplate which option we will choose, and after we’ve made a decision we can contemplate which other options we could have chosen and how different things could have gone. This ability to imagine alternate sequences of events and the results that stem from them allows us behave in infinitely more intelligent ways than most other species. We can avoid novel mistakes and learn from past mistakes and do all sorts of other snazzy things. People often want to insert free will into this situation. The fact that we can contemplate a different course of action, these people say, means we could have chosen differently. Thus our choice was free. But these arguments fail to consider though is that we didn’t choose differently. We only chose one behavior. We did one thing. And just because we had the conscious experience of being able to contemplate other options in no way necessitates that the action chosen was itself not deterministic or that it was free in any way. We impose free will on this “could have” nature of conscious decision making, and there’s no good scientific reason for it. Even a computer can compare among choices and pick the one it determines to be best. Picking a choice among alternatives is not incompatible with determinism.

For the second conceptual problem I’ll use lower animals as an example. Take the fruit fly brought up in the article. Given our neuroscientific knowledge of how behavior is initiated in organisms, who or what are we actually bestowing the honor of free will on? Behavior is the result of various sub systems, all engaging in their own tasks. There is no little guy in the brain of the fly who gets all the input and decides, “lets go with behavior C!”. And similarly, there’s no little guy who flips a coin and decides what to do based on the result. Even if we grant that the behavior that emerges from the organism is truly probabilistic, a free action implies someone is actually making a decision to behave in a certain way. But we can program a robot to behave probabilistically, and no one would say it has free will. It’s not even conscious. Let’s not forget that the locus of evaluation of a being making a free choice must be a mind able to choose. And by grounding free will in the evolutionary adaptive probabilistic behavior of animals, the authors may have inadvertently argued against the possibility of free will in humans. The author even talks about “choices” basically just fitting a complex probability, but being perceived as conscious decisions. This is exactly right. Perceiving is not causing, and perceiving can be in error. Remember my post about the weirdness of consciousness? We perceive all sorts of illusions and hallucinations that don’t exist, might we also perceive free will when it is also just a trick of our brains?

All animals have some sorts of boundary conditions and constraints on their behavior that limits their freedom to some subset of ‘possible’ actions. I can’t fly, I can’t run 1000 mile an hour, I can’t morph into a lion, I can’t walk into a maximum security facility, I can’t walk through walls, I can’t experience pain as an orgasm, I can’t not feel sad when I see a dead animal, I can’t not love the people I love. I am not free to do these things. Whether it is because of my genetics, or my biological make up, or my experiences, I am constrained by my psychology and neurophysiology. But within these constraints I have a range of possible behaviors. This research suggests that science is currently only able predict with some probability what my action will be (well, not mine, but simple animals). But so what? That doesn’t imply that behavior isn’t deterministic, just that we can’t predict it yet. Whether the behavior that was chosen had a 80% chance of being chosen or a 20% chance, the choice happened for a reason. What was that reason? Why one choice rather than another?

The possibly not so obvious point that I am trying to make here, is that for an animal to evolve a way to choose among possible actions in a probablistic way, there has to be a neuronal mechanism that underlies this. A random number generator of sorts. But what does random mean? We have computers that have random number generators, but they aren’t truly random. What would that even mean in a computer? It’s not possible. when you roll a die, it might seem like the outcome is not predictiable, but that’s just a matter of complexity. The die obviously follows the laws of physics. And if you know it’s velocity, its rotation, its height, its physical properties, how much wind there is, etc…if you know all the variables, you could conceivably predict how the die will land. Any behavior, whether underlied by a choice or not, must have a reason for itself. That reason must have a causal explanation. If it has a causal explanation it’s deterministic (even if highly complicated). If it doesn’t have any causal explanation then we have to resort to quantum mechanical probabilistic explanations. And then it was random. Either way, the choice was not free. I did not exert free will. My behavior was simply unpredictable by science. And I had an experience of making a choice. If that’s the best science can do, it has yet to make a defense of free will.

————————-

The original research paper: http://rspb.royalsocietypublishing.org/content/early/2010/12/14/rspb.2010.2325