After seeing the Ledocs World blog on Searle, I thought I would continue our discussion on free will but with the luxury of not having to hear contrary ideas. It’s not my intent to convince you that you are wrong, for I don’t even know that you are wrong. I’m an agnostic on the topic. But I do hope to show how one (if not you) might not regard it inconceivable that you are wrong.
Here is my understanding of your argument:
- The mind has free-will.
- Free will cannot arise from a deterministic, or even stochastic, system.
- Therefore the reductionist approach of science cannot, in principle, explain the mind.
I have several comments, probably philosophically naïve, which I won’t try organize.
Topic I: My recollection is that you asserted that free will is an empirical fact because everyone has the personal experience of “I am making a decision freely.”
1. This is not an argument for free will. Rather, it is a description of attitudes about free will and of how brains work. Perhaps this attitude about free will is hard-wired into the brain. If this is the case, then the belief that one has free will might be the result of determinism at work, and only a person who has achieved a degree of free will is free to doubt the existence of free will.
2. Your assertion, as I remember it, is not entirely true. Many Eastern belief systems hold that the concept of “I” is illusory. Here, for example, is one statement of that view:
Sri Ramana Maharshi taught that every conscious activity of the mind or body, for example ‘I think’, ‘I remember’, ‘I feel’ ‘I am acting’, etc., revolves around the tacit assumption that there is an individual ‘I’ who is doing something, a common factor and mental fiction termed the ‘I’-thought (a translation of Aham-Vritti, which literally means ‘mental modification of ‘I’). He equated individuality with the mind and the mind with the ‘I’-thought which is dependent on identification with an object, and said that after Self-realisation there is no thinker of thoughts, no performer of actions and no awareness of individual existence. When the thoughts arise, he said, the ‘I’-thought claims ownership of them- ‘I think’, ‘I believe’, ‘I want’, ‘I am acting’, but there really is no separate ‘I’ that exists independently of the objects that it is identifying with, only an incessant flow of misidentifications, based on an initial assumption that the ‘I’ is individual and associated with the bodily form.
I confess that I don’t follow this completely, but I’ve also seen Buddhist writings of a similar ilk. My point is that this view of “I-thought” as an illusion seems to stem from personal introspection and meditation. That is, there are people who think “I deciding” is an illusion on the basis of their own personal experience.
Well, this Eastern view about the illusory nature of an I-making-decisions nor more disproves free will than the convincing experience of making a decision proves free will.
Modern neuroscience also suggests that our conscious and controlling I is an illusion of sorts. This view is not based on unscientific Freudian or Jungian concepts but upon observing various brain regions and networks. But neuroscience goes in a different direction from there than does Eastern thought. There is, for example, a subsystem that maintains one’s sense of uniqueness and separateness. Suppress the activity of that system, and the subject begins feeling a sense of oneness with all. Other parts of the brain can be manipulated to produce a sense of complete conviction or of intense religiosity. Brains scans show different networks within the brain coping with a situation or decision, then coming to a resolution, then the conscious mind becoming aware of the decision.
So both western experimental science and eastern mental disciplines support the idea of the conscious controlling I as an illusion. I’m not saying that the concept of the conscious mind as a story-telling observer negates the concept of free will. But I do think that free will, if it does exist, is going to prove to be something more deeply layered and complex than it is commonly viewed. For example, it may prove a mistake to associate free will with the conscious mind.
Topic 2: Free will cannot arise from a deterministic or stochastic system. This sounds obviously true. Yet it seems possible that a deterministic system could produce behavior that would be very difficult to distinguish from free will. For instance, it is possible to know the laws governing a deterministic system, yet have the future behavior of that system to be, in principle and fact, unknowable.
Before going further with that thought, I’d like to provide some really interesting (to me) background. You probably know that most of the fundamental laws in physics are expressed in terms of second-order (i.e., involving second derivatives) differential equations. For example, Newton’s 2nd law is F = ma. Here F is force, m is mass, a is acceleration, which is the second derivative of acceleration. The force, in turn, can be a function of position. For some problems, such as two bodies orbiting about their center of mass, the differential equation has a general solution in the form of equations with constants used to fit the general solution to a particular case. For the orbiting bodies the general solution is that a body follows an elliptical path, a parabolic path, or a hyperbolic path. The precise initial conditions (masses, positions, and velocities) determine which sort of path and the particular values for constants of integration. For example, if the conditions produce an ellipse, the particular initial conditions then lead to particular values for the semi-major axis, the eccentricity, and the orientation of the ellipse. Another part of the solution allows one to calculate where in its orbit each body is as a function of time. So you get a continuous description of the system as a function of time.
However, second order differential equations do not necessarily have general solutions. In gravitational problems, once you have three or more bodies, there is no general solution that you can write in the form of a formula. Instead, for each set of initial conditions, you have to use numerical computational techniques to produce essentially a table of values representing the state of the system at particular times. That is, you would get values for the positions and velocities of the bodies at one particular moment of time, then at another moment of time, and so on. Instead of a continuous description, you get a discontinuous, discrete description. Another limitation is that the accuracy of the result is limited by the precision of the calculations. Also, if you wish to change the initial velocity of one of the bodies, you can’t just adjust the values of some integration constants. Instead, you have to go through the entire computation routine again.
Ok, perhaps that wasn’t all that interesting. But let’s go a step further. Suppose you do go back and change one initial condition, a location, say, by a small amount. You probably would expect that the new numerical solution you calculate would be different from the original solution by a small amount, probably a difference that grows with time. In fact, that often is the case, but not always. For instance, a couple of French astronomers recently calculated planetary positions in the solar system for the next 5 billion years. They did 2501 trials, each time changing the initial size of Mercury’s orbit by about 1/64th of an inch. The vast majority of the trials produced future orbits not much different from the present. But a few produced wildly different results, such as Mercury falling into the Sun or crashing into Venus or Mars colliding with Earth.
This is an example of chaotic behavior, as the term is used in physics. This means that the behavior is extremely sensitive to initial conditions. How extreme? I don’t know if what I’m about to say applies to this particular case, but here’s how extreme. Consider a point on, say, the x-axis representing the numerical value of a particular initial condition. Imagine marking off a range that includes that point. In a truly chaotic system, no matter how small you make that range, it will include values that lead to solutions that are wildly different from one another.
So the only way to correctly predict the future of a chaotic system would be to measure every initial condition exactly. Even without invoking the Heisenburg Uncertainty Principle, this impossible. If you have a technique to measure a position to a million decimal places, your are still clueless about the 1,000,002nd decimal place. And you would have to do the numerical calculations to infinite precision. Thus the future behavior of the deterministic, yet chaotic, system is unknowable.
Let’s apply this to brain theory. Suppose there is no free will and, even a stronger supposition, that the brain is deterministic. Suppose, unlikely as it may seem, that we develop a way to very precisely measure every relevant physical parameter of the brain—the location of every ion, the location and activity of every neuron, and so on. And suppose we have a mathematical theory capable of extracting the brain state from this information and calculating the subsequent physical parameters and brain states. It’s extremely likely that such a complex system has elements of mathematical chaos, hence, even with all this information, the information is not infinitely precise, and it will be impossible to predict the future behavior of the brain. Most likely the predictions would work pretty well for a while, but eventually a chaotic situation would mess things up.
Now assume the brain has free will and we go through the same exercise. It’s likely that most of the time our behavior is pretty automatic, so the behavior prediction will work pretty well for a while, but eventually a major free-will event will occur and mess things up.
So my expectation is that the behavior one would see in a deterministic yet chaotic brain would be sufficiently complex and unpredictable that it would be very difficult to distinguish it from a free-will brain.
Topic 3: Science is overreaching itself
My recollection is that you expressed the third point (in my summation of what I think your argument was) by saying you thought science is overreaching itself in its attempts to explain the mind. I don’t know if intended so, but the phrasing suggests that science has an element of hubris. I’d make the following points. First, while some scientists may have an agenda of demonstrating a materialistic explanation, others think of themselves as trying to solve a puzzle and think that some aspects are difficult even to think about now, let alone explain. Second, the argument that science cannot explain the mind in principle relies on a principle that has not yet been demonstrated by science. Science would need to learn much more about the workings of the brain and mind before recognizing that principle. Third, even things said to be impossible in principle in physics, such as perpetual motion machines, refer to principles that have, so far, been verified (but not proven) experimentally. Your “in principle” is not of that sort. For science, identifying what is impossible in principle is something to be determined empirically, not by pronouncement. Is philosophy overreaching itself if attempts to impose non-scientific constraints upon what science can accomplish?