Free will emerges

Free will and notes on The Beginning of Infinity, Chapter 5
Arguments against free will tend to make the reductionist assumption that, because there isn’t a physical explanation for human freedom, it must be an illusion. Broadly, reductionism is the idea that the world can only be explained as the sum of interactions of its fundamental parts. For example, the high-level properties of a substance (such as boiling point or state at room temperature) can be predicted from low-level atomic interactions. Lots of these sorts of ‘reductive’ theories are true and useful; they have reach. Newtonian mechanics was a ‘reduction’ of Kepler’s laws of planetary motion and Galileo’s theories of motion. But reductionism as a methodology is mistaken. Science doesn't progress by analysing high-level phenomena into low-level phenomena. There are lots of high-level phenomena that, while they ultimately consist of low-level phenomena, have patterns and laws that are present on the higher level but not, so far as we know, on the lower level.

Douglas Hofstadter gave the following as an example of the inadequacy of reductionist explanations for emergent phenomena. Imagine a set-up of millions of dominoes, placed closely together (so that they can knock each other over in the usual way) in a complex pattern of rows. The dominoes are spring-loaded and, if knocked over, will pop back up after a set time. If a row is knocked over it can be interpreted as a binary ‘1’, and if not knocked over, a binary ‘0’. The set-up is sufficiently rich and complex that it can perform computations, and in this instance it is set up to tell you whether the number keyed in (by placing a row of that number of dominoes at a specified position) is a prime. One domino in the set-up is the output domino: If it is knocked over at the end of the computation, it means a divisor was found and hence the input number is not a prime. If it stays standing, the input number is a prime.

Now, if the output domino stays standing after one of these operations, an observer may single it out and ask: “Why did that domino never fall?” The reductionist explanation would be: “Because the domino behind it never fell, because the domino behind that domino never fell, because…” – and so on, or: “Because none of its neighbours ever fell, because none of their neighbours ever fell,” and so on.  This answer is true, but it is not an explanation: It merely states the already obvious fact that no domino will fall unless one of its neighbours falls. To explain why the domino didn’t fall, we have to make reference to the non-physical concept of primality and to the dominoes’ emergent capacity to say whether a given number is a prime.
So reductionist 'explanations' are inadequate for some emergent phenomena. Even if the series of physical events described by the reductionist account did actually happen, such emergent patterns have to be explained on their own terms. If human freedom is emergent, it doesn’t conflict with physical determinism -- it just has nothing to do with it.

There is a reductionist idea that the mind cannot affect the physical world, on the grounds that only physical events can cause other physical events. But the idea of a ‘cause’ is abstract; at the purely physical level, cause and effect are interchangeable. The laws of motion can retrodict as well as predict. So a cause is just an explanation we infer for why something happened, and there is no reason to think that physical explanations are the only explanations we have.

For more, see David Deutsch, The Beginning of Infinity, Chapter 5.

Popper’s three Worlds
To describe the emergence of human knowledge, Popper posits three worlds -- World 1, the world of physical events; World 2, that of mental states or subjective experiences; World 3, that of products of the human mind: theories, scientific (or otherwise intellectual) problems, the information (if not the physical stuff) in books and libraries. It used to be popular to deny the existence of World 1 and to claim that only subjective experience exists. Now it’s more fashionable to claim that only the physical world exists and to deny the existence of experience. The once-popular denial of the reality of World 1 was refuted by Dr Johnson on the grounds that it ‘kicks back’. If the physical world is an illusion, and that illusion behaves exactly as though it were real, then it needs to be understood and explained in exactly the same way as would a real physical world. It kicks back, and so it might as well be considered real.

To explain the physical (World 1) presence of man-made objects -- say, skyscrapers, computers, nuclear reactors -- we have to refer to the theories people formed about how to produce them. These World 3  theories could also be mistaken. A miscalculation might lead to the World 1 event of a bridge collapsing. World 3 has a logical structure that exists independently of humans. For example, though natural numbers are a human invention, facts about them can be discovered, such as the existence of primes.

The ‘human’ dimension of World 3 also kicks back. We make judgements about people’s beliefs, values, habits, etc., and explain their actions as responses, based on those things, to a given set of circumstances. I might be convinced that my boss is going to sack an incompetent colleague, but then discover that, because my boss is the forgiving type, the colleague has instead been put on a training programme. Using the laws of physics to explain all of these things reductively would be absurd. It would be as much as to say that they don’t exist, even though they kick back. If someone thinks that 4+4=9, he will be kicked back by the laws of mathematics. Similarly, (and quite rightly), people who deny the existence of free will are kicked back by the contradiction inherent in proposing such arguments as things that people ought to favour (to choose) over other theories.

For more on the three Worlds, see Karl Popper, The Open Universe, Addendum 1: Indeterminism is Not Enough.

Free will and creativity

So I was at a philosophy conference last weekend where a handful of undergraduates presented papers on various topics. One paper made replies to some recent criticisms of Frankfurt-cases as counterexamples to the notion that, for a person to be morally responsible for an action, it must have been possible for him to do otherwise. To this day generally accepted, the claim is known as the Principle of Alternate Possibilities. Frankfurt's argument is essentially that, even if sufficient conditions are met to ensure that a person acts in a certain way, they do not necessarily explain why the person acted in that way. Only the true explanation, which is to be found in the person's will, can tell us whether he acted freely. Frankfurt's main thought experiment runs thus: 

Jones has decided to kill Smith. Black also wants Smith to be killed, so he monitors Jones' mental state and will know if Jones is going to change his mind. If he does change his mind, Black has the power to make him kill Smith anyway. But Black doesn't want to get involved if he doesn't have to, so if Jones does go through with it, Black won't intervene. In this case, it seems intuitively true that Jones would be morally responsible for his decision to kill Smith, even though he couldn't have done otherwise.

The counterexample suggests that the real reason someone might be said to be morally responsible for an action is its explanation in terms of their will – their own inner reason for having so acted. Jones didn't kill Smith because Black determined him to – he did so because of his own decision. (There is the obvious objection that, if causal determinism is true, Jones's very thought process couldn't have gone any other way, so he could hardly be blamed for that either. I'll return to this shortly.)

Now, a recent criticism of Frankfurt-cases as counterexamples to the Principle disputes Frankfurt's concept of action. For something to count as a genuine action by an agent, the agent must have the ability to refrain from doing so at the time of the action. An 'action' is thus defined as an intervention into the course of nature that the agent need not have bought about. By this standard, Jones didn't even perform an action, let alone bare moral responsibility for it. But, as Moran points out, formalising this argument reveals that it's just a brute assertion of incompatibilism: No actual case is made for the claim that free will only occurs if the agent has the possibility to refrain from their chosen course of action. It has failed to show that there is any connection between an event depending on the agent's will and the agent's ability to refrain. So this mere definition of agency isn't a very successful criticism in itself.

Moran is critical of this and also of a second argument which says that, in the Frankfurt case, Jones does not have the right sort of control over his actions. Different sorts of control have been identified for the purposes of this debate, and Moran argues that a kind of hypothetical or 'conditional' control should suffice for agency. For example, Sally is driving a car and turns left. She was 'metaphysically determined' to do this, and in that sense she couldn't have done otherwise. However, it is still the case that if a certain set of events occurred, Sally would not have turned left. The ability to respond appropriately to events is part of her 'intrinsic properties': In principle she can refrain from turning left, and this seems to be the relevant sense in which she 'can' do that, even if she is destined to respond to events appropriately and is not 'able' to respond in any other way.

But even inanimate objects have this sort of conditional 'power' to some extent. If it rains, the grass can grow, but in the event it hasn't rained so the grass doesn't grow. So clearly the idea of free will conditional powers needs to be further explained. The most Moran says about this is that humans are "highly responsive to their environment," and that

we are the kind of creatures who are psycho-physiologically sophisticated in such a way that if any of a range of possible events occur, we are capable of the appropriate responses.

But a computer could be programmed to 'respond appropriately' to a range of possible events – so this isn't enough for free will. It also seems wrong to imply that inanimate objects, in their sheer obedience to the laws of physics, have a certain (if tiny) amount of the quality that supposedly makes human freedom. But free will isn't a matter of degree. This was brought up in the question period as someone asked where to draw the line between animals, etc., that aren't responsive enough to have free will and those which are. The reply was that more work needed to be done before a clear distinction could be made. 

In the second section of the paper it is argued that the future may be 'subjectively open' even if it's objectively determined. When we talk of a person considering 'options', we might only mean that those options are 'epistemically' rather than objectively possible. This looks like a restatement of Frankfurt's argument. But if something is only 'subjectively' the case it might just as well be an illusion, so this doesn't suffice either. It was brought up in the question period that our deliberations about the subjectively open future are themselves determined, so these Frankfurt-type arguments beg the question against incompatibilism. In any case, none of the arguments so far have shown that humans have intrinsic properties that result in free will.

Perhaps it has something to do with creativity. When we talk of people deliberating over various options, we refer more than anything else to the creative act of conjecturing and criticising various options and thereby also conceiving new options that didn't previously exist. Creative processes are inherently unpredictable, which is problematic for the requirements of hard determinism. Also, if I haven't misinterpreted it, the view of free will as creativity is embodied in Karl Popper's 'two-stage model' – mirroring his epistemology of conjecture and refutation. In a later post I'll look at this model and criticisms of it.

NB: I have referred to the sources below but if anyone wants citation for a particular sentence I'd be happy to give it.

Paper from the conference -- 'Agency, Frankfurt-Cases and the Compatibility of Determinism with Free Will and Moral Responsibility' in the BJUP
Alvarez, 'Actions, thought-experiments and the ‘Principle of alternate Possibilities’'
Frankfurt, 'Alternative Possibilities and Moral Responsibility'
Summary of Popper's two-stage model