This is the second post in a series discussing some key ideas from Why Free Will is Real (Harvard University Press, 2019). Many thanks to John Schwenkler and the Brains Blog for giving me this opportunity.
As noted in my previous post, I assume that free will requires intentional agency, alternative possibilities among which we can choose, and causal control over our actions. I take these three properties to be each necessary for free will and together sufficient. That is to say: an entity that lacks intentional agency could not be a bearer of free will. Non-agents, from rocks and tomatoes to armchairs and stoves, are not candidates for the ascription of free will. Similarly, an entity for which there are no alternative possibilities to choose from could not qualify as a bearer of free will either. Intuitively, I would not be free if, in my choices, I could never have acted otherwise. Finally, an entity lacks free will if its behavior is not caused by its intentional, mental states. If, whenever I do something, it is solely my brain that makes me do it, then my choices cannot be attributed to my own free will. On the other hand, if all three properties are present – that is, someone or something is an intentional agent, has alternative possibilities to choose from, and has causal control over the resulting actions – then this is enough for free will; nothing more is needed, no further mysterious ingredient.
Characterizing free will in this way has two advantages. First, it captures a robust understanding of free will. This matches the “libertarian” intuitions that many people have about free will before they encounter the skeptical arguments that often lead philosophers to adopt a weaker definition (for instance, one that drops the “can do otherwise” requirement). Second, by disaggregating free will into three properties – intentional agency, alternative possibilities, and mental causation – it clarifies what is at stake in the debate and takes some of the heat out of it. The difficult and often emotionally charged question of whether humans have free will is replaced by a set of more tractable and somewhat less charged questions: whether humans are intentional agents, whether they have alternative possibilities, and whether their actions are caused by their mental states. This gives us a checklist of things we need to consider if we wish to find out whether there is free will.
How, then, can we establish that humans have all three properties? The first thing to note is that there are two very different ways in which we can think about human beings. We can either think of them as purely physical systems, consisting of gazillions of interacting particles, and insist that human behavior is to be understood as nothing but a physical process. Or we can think of humans as not just physical but also psychological, as beings with mental states and cognitive processes that underpin their behavior. Call the first way of thinking the “reductionistic” one, and the second the “non-reductionistic” one.
It should be clear that if we adopt the reductionistic way of thinking, we may not find support for intentional agency, alternative possibilities, and mental causation. Intentionality may not seem to be a feature of physical systems; alternative possibilities may seem to conflict with physical determinism; and mental causation seems to go against the principle that all physical events must be attributable to physical causes. So, the reductionistic way of thinking leads directly to the free-will skepticism I have described in my previous post. However, the human and social sciences – from anthropology and psychology to sociology and economics – support the non-reductionistic way of thinking, which represents humans not as mere physical systems, but as agents with goals and purposes, beliefs and desires, and explains human behavior on that basis. It would be impossible to make sense of human behavior in its breadth and richness if we did not understand humans in this way. And this understanding, in turn, vindicates agency, choice, and mental causation as central features of human beings – features that emerge from (and “supervene” on) physical processes in the brain and body but do not lend themselves to a reductionistic description in physical terms alone.
Let’s begin with intentional agency. However much the different human and social sciences – such as anthropology on one side and economics on the other – disagree about how to explain human behavior, the one thing they all have in common is that they take an “intentional stance” towards human beings (to use a term from Daniel Dennett, 1987). That is, they explain human beings as agents who perceive the world and cognitively represent it, who act in pursuit of goals, and who respond to their situation in ways that are at least partly rational. Whether you consult anthropology or micro-economics, psychology or sociology, you will find this intentional mode of explanation as a common feature. By contrast, if we tried to make sense of human beings solely as heaps of interacting particles, or as complicated neural networks, we would at most be able to explain some details of the brain and body or some specific aspects of physiology and cognition – for instance, how the visual cortex implements certain perceptual tasks. We would not be able to explain the rich patterns of human behavior in their breadth and flexibility.
To give a simple example, if I ask a taxi driver to take me to Victoria Station on one day, and I ask another taxi driver to take me to King’s Cross Station on the next day, and each time I successfully reach my destination, it would be extremely hard — perhaps impossible, in practice — to explain in purely physical terms what the two events have in common. We would have to cite the incredibly complicated neural and other physical processes in each driver’s brain and body as well as in the car. Contrast this with the intentional mode of explanation. Once we recognize the two taxi drivers as intentional agents who understand where I wish to go, form the intention to drive me there, and have an intelligible reason to do so, we can easily explain what’s going on and make predictions on that basis. The assumption that the drivers are intentional agents is vindicated by its explanatory success. Generally, the ascription of agency to people is indispensable for a satisfactory explanation of their behavior. This point should be fairly uncontroversial.
Next consider alternative possibilities. Just as we wouldn’t be able to explain human behavior without recognizing people as agents, so we wouldn’t get very far in explaining behavior if we didn’t view people as making choices in which alternative actions are open to them. The idea that humans face choices between different options, deliberate about them, and select one option among the possible ones is no less important for the human and social sciences than the idea of agency itself. This means that we represent humans not as deterministic machines, but as beings for whom different courses of action are possible. I call this idea “agential indeterminism”. Even in a field like decision-and-game theory in economics, which is sometimes (mis)interpreted as representing humans as nothing more than utility-maximizing automata, the notion of a decision tree with choices between several possible options is central. Having different options does not mean that they are all equally likely to be chosen. After deliberation, a decision-maker may well find some options more rational or more attractive than others.
I argue in my book that the assumption of agential indeterminism is a key presupposition of the intentional mode of explanation itself. Without that assumption, our explanations of people’s behavior in the human and social sciences would not get off the ground. My conclusion here is similar to that reached by Helen Steward (2012), who argues that the very idea of agency requires some form of indeterminism. Now one may legitimately ask whether this doesn’t conflict with physical determinism. As I will explain in more detail in my next post, agential indeterminism is compatible with physical determinism – an initially surprising point which, despite sounding counterintuitive, can be established in a formally precise way.
Finally, let’s turn to mental causation. Skeptics argue that it is not our conscious intentions that cause our actions, but physical states of the brain. On the theoretical side, they cite Jaegwon Kim’s “causal exclusion argument” (1998), which asserts that if we attribute our actions to anything other than a physical cause, this will breach some central tenets of a scientific worldview, such as the principle that there are no physical effects without physical causes or the principle that we should not postulate more causes than strictly necessary. If a physical cause, such as a neural state of my brain, suffices to account for the movement of my arm, for instance, then we should not postulate any further mental cause. On the empirical side, skeptics cite a series of experiments conducted by Benjamin Libet and his co-authors (1983), and subsequently others, showing that, when subjects are asked to perform voluntary movements of their limbs, one can detect some preparatory brain activity – a neuronal readiness potential – before the subjects experience the conscious intention to act. Libet took this to show that our intentions are only passive byproducts of the real physical causes.
My response, which I can here summarize only briefly, is that both the theoretical and the empirical arguments against mental causation can be rebutted if we are careful in defining what we mean by “causation”. If we look at how causation is understood in the special sciences, this points towards a definition of causes as systematic difference-making factors for the resulting effects (see, for instance, the interventionist theory of causation defended by Judea Pearl, 2000, James Woodward, 2003, and others). Roughly speaking, causal regularities are difference-making regularities that remain in place when we control for confounding factors. Peter Menzies and I (2009, 2017), but also others (such as Panu Raatikainen, 2010, and Adina Roskies, 2012), have argued that the most systematic difference-making factors for human actions are often at the intentional, psychological level, not at the sub-intentional, physical one. It is our intentional, mental states that most robustly co-vary with the resulting actions, not their precise physical realizers in the brain, which are too fine-grained to qualify as difference-makers – a phenomenon that Peter and I called “realization-insensitivity”. In Libet’s experiments, the neuronal readiness potentials measured prior to a subject’s formation of a conscious intention are, arguably, not difference-making causes of the actions, among other things because subjects can still abort an initially intended action after the neural activity has begun. The neuronal readiness potentials are best understood as belonging to the physical implementation mechanism of voluntary action. The intentional, psychological level remains a site of causal regularities, all the more when we move away from the simple motor actions studied by Libet and consider more complex actions that involve sophisticated planning. And so, the idea of mental causation is entirely compatible with a scientific worldview.
Now you might say: my arguments for agency, choice, and causal control only show that viewing people as intentional agents with alternative possibilities and mental causation is explanatorily useful: a convenient theoretical construct or fiction. But that doesn’t imply that this is what human beings are really like. Explanatory usefulness doesn’t imply reality, and the picture of human beings as choice-making agents conflicts with the more reductionistic picture given to us by the fundamental sciences. In reality, people are nothing but heaps of interacting particles.
There are two things to be said in response. First, science does not mandate adopting the reductionistic picture of human beings. To the contrary, the special sciences, from biology to the social sciences, support the alternative, non-reductionistic picture, and this picture is entirely compatible with the “physicalist” assumption that everything in the world is the result of underlying physical processes. Scientists recognize that even if everything is grounded in physical processes, many phenomena would be impossible to explain through the lens of fundamental physics alone. Higher-level explanations, such as those we find in fields ranging from biology to the social sciences, are indispensable. (The theoretical point which tends to be missed by proponents of radically reductionistic approaches is that supervenience does not imply explanatory reducibility. I discuss this point further in List 2018.)
Secondly, from a scientific perspective, our best guide to any questions about which entities or properties are real is given by our best scientific theories of the relevant domains. If we wish to find out whether electrons or neutrons are real, we must consult particle physics. Similarly, if we wish to find out whether the patterns of the climate are real, we must consult meteorology and climate science. This idea (defended by W. V. Quine, 1977, and Arthur Fine, 1984, and central to scientific realism) is sometimes called the “naturalistic ontological attitude”. In line with it, I suggest that if we wish to find out whether human agency, choice, and mental causation are real, we must consult our best scientific theories of human behavior, and as noted, these theories give a positive answer.
I can conclude this post by summarizing my overall argument – an indispensability argument for free will. The argument has two premises:
Premise 1: Our best explanations of human behavior depict humans as choice-making agents: agents with goals and purposes, alternative possibilities to choose from, and causal control over their actions. This depiction is indispensable and compatible with the rest of science.
Premise 2: If postulating certain properties or entities is indispensable in our best explanations of a given phenomenon and compatible with the rest of science, then we are (at least provisionally) warranted in taking those properties or entities to be real.
Putting these two premises together, we arrive at my conclusion:
Conclusion: We are (at least provisionally) warranted in taking intentional agency, alternative possibilities, and causal control over one’s actions to be real phenomena.
This, in a nutshell, is the core of the naturalistic case for free will. It is important to note that the naturalistic case for any ontological commitments in the sciences is normally provisional. New scientific developments might render certain postulated properties or entities dispensable even when they were previously considered indispensable. In the case of free will, if new scientific developments were to undermine the first premise of my argument – that our best explanations of human behavior depict humans as choice-making agents – then I would no longer be able to uphold my conclusion. It is good scientific practice to acknowledge this point.
In my next post, I will say more about why the sort of agential indeterminism that underpins the present picture of agency is compatible with physical determinism.
Dennett, D. (1987). The Intentional Stance. Cambridge, MA (MIT Press).
Fine, A. (1984). “The Natural Ontological Attitude.” In Scientific Realism, edited by J. Leplin, 83–107. Berkeley (University of California Press).
Kim, J. (1998). Mind in a Physical World: An Essay on the Mind-Body Problem and Mental Causation. Cambridge, MA (MIT Press).
Libet, B., C. A. Gleason, E. W. Wright, and D. K. Pearl (1983). “Time of Conscious Intention to Act in Relation to Onset of Cerebral Activity (Readiness-Potential): The Unconscious Initiation of a Freely Voluntary Act.” Brain 106: 623–642.
List, C. (2018). “Levels: descriptive, explanatory, and ontological.” Noûs, online early (preprint at PhilSci Archive).
List, C., and P. Menzies (2009). “Non-reductive Physicalism and the Limits of the Exclusion Principle.” Journal of Philosophy 106(9): 475–502.
List, C., and P. Menzies (2017). “My Brain Made Me Do It: The Exclusion Argument against Free Will, and What’s Wrong with It.” In Making a Difference: Essays on the Philosophy of Causation, edited by H. Beebee, C. Hitchcock, and H. Price, 269–285. Oxford (Oxford University Press).
Pearl, J. (2000). Causality: Models, Reasoning, and Inference. New York (Cambridge University Press).
Quine, W. V. (1977). Ontological Relativity and Other Essays. New York (Columbia University Press).
Raatikainen, P. (2010). “Causation, Exclusion, and the Special Sciences.” Erkenntnis73(3): 349–363.
Roskies, A. L. (2012). “Don’t Panic: Self-Authorship without Obscure Metaphysics.” Philosophical Perspectives 26: 323–342.
Steward, H. (2012). A Metaphysics for Freedom. Oxford (Oxford University Press).
Woodward, J. (2003). Making Things Happen: A Theory of Causal Explanation. New York (Oxford University Press).