Predictability, Determinism and Free Will

In ordinary language, the concepts of predictability and determination are taken to mean roughly the same thing: if something is predictable, then it has definite causes that determine it to be the way it is; conversely, if something has definite causes that determine it to be the way it is, then it is, in principle, predictable. In philosophy, however, these are distinct concepts. Something that is deterministic need not be, in principle, predictable, and again, conversely, something that is predictable need not be deterministic. I will use two examples to illustrate this point, remarking on the second statement first, as I think it is the less significant of the two.

First, we will examine quantum physics. We would like quantum physics to be deterministic and may even have good reason to suggest that it must be, but at this point, we cannot say that with any certainty that it is, in fact, deterministic. Still, even supposing that it is not deterministic, we can use probability based models to predict, with sufficiently high precision, what the results, or outputs, of a quantum system will be.

Second, and I think more importantly, we can look to the universe at large. If we assume that the universe is entirely deterministic—which again we cannot say with any certainty but have good reason to think that it is—then it does not follow that everything in the universe need be predictable, in principle. We could say that a super-being with all the information about every single particle and its momentum could, theoretically, predict the state of the universe at any given time, but if we add materialism to this deterministic universe, this suggestion becomes meaningless. So let us think of it this way: if we want to model a system, we can represent each part of that system in a computer program. In order to do this, we will need to map each bit of information onto its own bit of computer coding, in a one to one fashion. Put simply, if we want to model a system with 10 components, we will need 10 bits of computer code, each mapping one of the ten components*. But we cannot do this with the universe at large. By definition, we need to map every single particle in the universe onto its own bit of computer coding—how can we do this? We have already exhausted every single particle in the universe by defining our system to be modeled—we simply have no particles left that could make up the computer coding for our program. Going back to our system of 10 components: if our universe only contains 10 particles, then we cannot model this system except by using the system itself as the model, but then we aren’t really modeling it, we are just watching the original system play out naturally. In this way, we can see that, even if our universe as a whole is deterministic, we still cannot, in principle, predict everything that is going to happen, because we, in principle, lack the means to do so, excluding the existence of non-physical super-beings.

To drive this home, I am going to borrow a quote from Richard Feynman:

It’s again this chess game business. If you were in just a corner where only a few pieces were involved, you could work out exactly what’s going to happen. And you can always do that when there’s only a few pieces, so you know you understand it. And yet, in the real game, it’s so many pieces you can’t figure out what’s going to happen. So there was a kind of hierarchy of different complexities. It’s hard to believe—it’s incredible, in fact most people don’t believe—that the behaviour of, say, me, one yack-yack, and you, nodding and all this stuff is the result of lots and lots of atoms all obeying these very simple rules.

To conclude, in a way, I want to remark on the relation between determinism, predictability, and our naive conception of free will. Part of the naive conception of free will is that we can, in principle, act in unpredictable ways. It simply is not the case that someone external to me could predict my own own behaviour with perfect precision. Often, the view of determinism, and its lay-equivocation with predictability, is seen as an attack on this conception of free will. But using the argument above, we see this need not be the case. We will never be able to predict the state of the universe at large, and if we cannot do say, we may always be misdefining one of the variables that we use to predict a local, closed system (i.e., for the purpose of this example, a human brain). Determinism does, in fact, have profound implications for free will if it turns out to be true, but they are much more subtle than they might seem at first glance.

*This is an oversimplification. We would also need computer coding for each of the laws describing the relations between the different components, but we will see that we need not even invoke these to illustrate the point.

Indeterminism: What It Is, and What It Isn’t

I want to briefly remark on the concept of indeterminism:

It is sometimes stated that we have two choices: determinism in the strict sense, or probabilistic indeterminism. This could not be further from the truth. Simply because a system is not strictly deterministic does not mean that the only other option is probability, or “lawlessness,” as some have put it. Agent causation is another option (note: it is possible to redefine and subsume agent causation under one of the two former options, but it is not necessary to do so).

That said, it seems to me that the attempt to formulate the problem in this way is not mere carelessness, but, in fact, a deliberate attempt by strict determinists to belittle their opponents. Most anti-determinists do not propose that simple probabilistic indeterminism is the right way to go, but rather endorse some form of agent causation, as mentioned above. If you can convince your audience, however, that your opponents are arguing for nothing more than “simple indeterminism” (i.e., the probabilistic form of indeterminism), then you avoid having to actually take on your opponents arguments, seemingly strengthening your own position.  It is worth noting that some of the arguments that get labeled as indeterminism in this way are actually only arguments against the strictest form of determinism.

This kind of rhetoric is highly counterproductive, and should be attacked whenever it is identified.

Determinism, or Indeterminism: That is not the question.

In my Metaphysics class today, the following argument was put up for scrutiny:

1) If determinism is true, then no one acts freely, ever.

2) If indeterminism is true, then no one acts freely, ever.

3) Either indeterminism is true, or determinism is true.

4) Therefore, no one ever acts freely, ever.

5) If no one ever acts freely, ever, then no one is ever responsible for their actions.

Premise 1, in brief, relies on that assumption that if the world is deterministic, then everything that happened today was a necessary consequence of what happened millions of years ago. If everything that happened today was a necessary consequence of events in the distant past, then no person has any control over the present—it is all set in stone, as it were. Free will dictates a certain amount of control over present actions, so if this control is absent, then so is free will.

Premise 2, on the other hand, relies on a purely probabilistic definition of indeterminism. If events are indeterministic, which is to say that they are merely an odds game with event A having a 40% probability, and event B having a 60% probability, then we still lack any sort of “control” over the situation. Which event occurs is largely arbitrary, relying only on some unknown odds, written in the sky or otherwise.

This is not to say that these are the only ways in which premises 1 and 2 can be formulated, but this is how they were presented in this case.

Most of the objections raised, both in my class and in the literature, from what I’ve seen,  have attempted to disprove either premise 1 or 2. That is, there can be free will under determinism, or there can be free will under indeterminism. Most of these amount to some re-formulation of free will. I will not be taking either of these positions. Instead, I will attack premise 3: That the world is either deterministic or indeterministic.

The core of my argument rests on the claim that premise 3 presents a false dilemma. It is either determinism, or it is indeterminism, but not both. I assert that it is, indeed, both, or at the very least, we are not in a position to rule this possibility out. Current physics, which is where most of these theories claim to have their support, does not itself claim to have sorted this issue out. We know that under certain circumstances, such as when the scale is microscopic, that the world behaves in an apparently indeterministic way. Under other circumstances, such as when the scale is macroscopic, the world behaves in an apparently deterministic way. Many propose that we can link these two, and show that it is really one, and not the other, in virtue of a fundamental property of nature: namely parsimony—or, that the universe is, at its most fundamental, simple (simple in the sense that it all can be reduced to more or less the same thing). But, what they miss, is that it does not have to be this way. There is, in fact, no law that says that the universe must be simple. It may very well turn out that the universe is complicated, perhaps even too complicated for us to understand it, in the proper sense of the word.

(The following is mere speculation, I have absolutely no empirical basis for the ideas that follow; however,  I still, personally, find a great deal of plausibility in them, but you have been warned, nonetheless!)

Building off of this, and the fact that most of the arguments that place free will either in a purely deterministic or a purely indeterministic light typically have to resort to a reformulation of free will itself, I now assert that free will is only a coherent construct in a world that is both deterministic and indeterministic. What I propose is the following, which relates this more specifically to the theme of this blog: free will can only exist in conscious creatures. This may seem unnecessary to state in so many words, but the following should provide reasons for it. Complex brains are, in a general sense, specialized organs for planning and deliberation. Given that the microscopic events of this world are largely indeterministic, and that the macroscopic events are largely deterministic, we can postulate the following: brains serve to make sense of a vast multitude of indeterminacy. Through the process of evolution, and, to steal a phrase from a neuroscientist I once knew, thanks to the goddess of molecular evolution, they came to be in a position to turn underlying indeterminacy into coherent, conscious actions. This is not an appeal to a “collapse-of-the-wave-function” view of consciousness, to be clear. Rather, it is an attempt to reconcile the disparate aspects of reality into one coherent framework.

We can use this argument to strike down some of the objections raised to both purely deterministic and purely indeterministic accounts of free will. One variety of the former asserts that if you could not have acted otherwise, then you could not have acted freely, as stated above. If there is some underlying indeterminacy, however, this is clearly not the case. There are, in fact, a multitude of different ways in which you could have acted. Aha! But this just reduces to a variety of the argument from indeterminacy—that actions are merely arbitrary instantiations of probabilities, right? But that is where the deterministic aspect of reality kicks in. Once the most basic underlying facts about the world are set, in a probabilistic fashion, then determinism takes over. For this, I draw on an idea put forth by John Searle: downward causality, but in no way do I claim to restate his argument. The higher-order functions of the brain, namely consciousness, do indeed have “causes” that exist as smaller, microscopic bits, but these higher-order functions also have the ability to rain down causation on these smaller bits, much in the way that higher-order theories of economics can influence the activities of lower level commodities. Neither of these can be “smoothly reduced,” as Searle puts it, to the other, but that does not imply that one or the other does not exist, or play a meaningful role. In fact, Searle says that typically, reduction of one thing to another serves the purpose of showing that one of those things does not exist, not the other way around, as is often claimed.

This may seem counter-intuitive, and in some ways, it does have to re-formulate the popular idea of free will. In particular, it draws a distinction between free will at its most basic on the one hand, and conscious will on the other. Conscious will, or the idea that you are consciously in control of all of your actions and thoughts, is inevitably false. A handful of psychological experiments demonstrating non-conscious biases and predispositions shows this very simply. But this is not what we are talking about when we say free will, or so I claim. Free will is much more general than the limited definition of conscious will. At its most basic, it requires that you be capable of acting in certain ways that rely on intentional stances. Even if you are not consciously aware of your decisions to act in certain ways, it is still you that is making them. You are your brain, and everything that comes along with it. Simply because something is non-conscious does not make it any less a part of you. It may clash with the popular account of who you are, but at the end of the day, you are made up of more non-conscious pieces than conscious pieces, so restricting our definition of free will to the conscious pieces seems to make little sense. Now, this is not to say that our conscious feeling of free will is irrelevant, but it is a different matter to bring up—specifically, it is more of an epistemic question than a metaphysical question.

On the time-scales of free will

In the past few decades, many empirically-bent workers in cognitive science, psychology, and philosophy have taken it upon themselves to correct our illusion that we have free, conscious will. The problem that many of these theorists overlook, however, is that the idea of free will can be separated into two relatively independent processes, differentiated by their respective time scales: short and long.

Short term physical activities are largely concerned with constructing the appropriate movements to match the current situation, like picking up and writing with your pencil, or walking down a street—both of which require incredibly complex movements from a basic systems level. The nature of our nervous system, however, allows such movements to be written into the system itself, such that the resulting structure of the system allows these movements to take place with little information processing needing to be done.

Correspondingly, in the realm of the brain, both empirical and theoretical viewpoints do, indeed, point to a remarkably diminished activity of free will—one need only think of split-second decisions as an example. A great many of these split-second decisions turn out differently than we would have preferred, given a chance to deliberate. This side of the issue makes sense from a biological/organismal standpoint: If we have incredibly complex brains that are capable of a great deal of things without conscious attention, why not take advantage of that when it comes to basic mental events? In this way, short term mental events and short term physical events are not so different. Such a method of functioning allows more resources to be devoted to other higher-order matters of cognition, such as planning and deliberation.

Planning ahead and considering possible alternatives is where the long term variety of conscious will starts to play a significant role. Planning and deliberation are not merely a system of inputs onto your nervous system with a more or less pre-set outcome, as is the case in many short-term scenarios. It involves assignment of value to certain outcomes, and a subsequent weighing of those values against one another. Note that I am not trying to say that long-term planning is not ultimately a result of physical actions in the brain, but rather that it is not so simple as its short-term counterpart. I am also not saying that these two time-scales are fully separate—if I were, then the long-term system could not influence the short-term system and people could not be help responsible for their actions. My view is instead that long-term systems work to influence and entrain the short-term systems towards a desired pattern of outputs. The details of this system need to be worked out, but the cursory description lines up more closely with experience than the one outlined at the top of the article.

The problem with those who use currently available empirical data to undermine free will, then, can be summed up quite simply. All of their data is built around short-term decisions. But even advocates of free, conscious will see reason to doubt short-term decisions as being fully controlled by consciousness. So, in effect, the present researchers are merely proving something true that most experts already thought true. The only objection to them, is this: short-term data do not reveal much of anything about long-term phenomenon. For a clearer example, the composition of a small pile of sand on a beach does not reveal much about the grander process of tidal erosion.