Decomposing the Will: Meeting the Zombie Challenge

The belief in free will is firmly entrenched in our folk understanding of the mind and among the most deep-rooted intuitions of western culture. The intuition that humans can decide autonomously is absolutely central to many of our social institutions from criminal responsibility to the markets to democracies and marriage. Yet despite the central place free will occupies in our commonsense understanding of human behaviour, the nature of this very special human capacity remains shrouded in mystery. It is widely agreed that the concept of free will has its origins in Christian thought (Arendt 1971, part 2, ch.1) and yet from the earliest days philosophers, theologians, and lawyers have disagreed about the nature of this capacity. Many have doubted its existence, some goes as far as to charge the idea of free will with incoherence. Nietzsche famously declared the idea of free will “the best self-contradiction that has been conceived” (Nietzsche BGE, 21). The philosophical controversies surrounding free will roll on to this day (see e.g. Kane 2004; Baer et al 2008), while the folk, for the most part, continue to employ the concept as if its meaning were wholly transparent, and its reality beyond question.

Into this happy state of affairs the cognitive sciences dropped a bomb. All of a sudden it seemed as if the folk’s happy state of ignorance might have come to an end, and the questions surrounding free will might now be settled once and for all using the experimental tools of the new sciences of the mind. The truth the scientists told us they had uncovered was not pretty: some claimed to have decisively demonstrated that there is no free will, and any belief to the contrary was a relic of an outdated and immature folk understanding we must leave behind once and for all (Libet 1985; Roth 1994; Wegner 2002; Prinz 2003; Lau, Rogers et al. 2004; Soon, Brass et al. 2008). [1] The scientists told us that we are the puppets of our unconscious brain processes which “decide” what we will do quite some time before we know anything about it. Some of the scientist’s statements were clearly intended as polemical provocations. However, they were also giving expression to an increasingly well-founded scepticism that free will as it is ordinarily understood might be a particularly powerful and striking illusion on a par with the magic tricks of the best conjurers. Unsurprisingly, the response from the intelligentsia was swift (see e.g. Wolfe 1996; Horgan 2002; Brooks 2007) and the findings have also generated a wave of interest within the academic world with a number of excellent monographs and collections of essays emerging aimed at evaluating the scientist’s findings (see e.g. Pockett et al 2006; Baer et al 2008; Vierkant 2008; Mele 2010). However so far what has been missing from the debate is a serious attempt to use the knowledge we undoubtedly gain from the experiments to enhance our understanding of human behaviour and its causes.

This collection aims to fill that void; it is designed to work as a tool for anybody who is interested in using the advances in the sciences of the mind to better understand the hitherto mysterious capacity for free will. Our talk of Decomposition in the title of our collection should be understood as an explanatory aim: we propose to take the conscious will, the original home of the homunculus, and to explore some of the ways in which scientists think it can be broken down into simple mechanisms whose interactions make us free agents. In doing so we aim for a better understanding of the relationship between psychological mechanisms and the experience we have of authoring and controlling our own actions. We embrace the scientific advances as an opportunity for a deeper scientifically informed self-understanding. We leave it as an open question, to be settled through a careful dialogue between philosophy and science, the extent to which such a self-understanding will turn out to be consistent with the folk belief in free will.

Category mistake?

Before we turn to the real substance of the book we must pause to consider some objections to such an explanatory enterprise. Many philosophers have pointed out that the sciences might be well placed to help us understand volition better, but there are important limits on what can science can tell us about autonomy (see e.g. Roskies (this volume)). Science might be able to help us with the “will” part of “free will”, but there are limits to the help it can offer us with regards to the “free” part. When we say that a person has acted autonomously or freely we are attributing responsibility to the person for the action, and hold the person accountable for the consequences of their action. It is far from obvious that the concept of moral responsibility answers to anything relating to psychological mechanism. On a natural enough understanding of responsibility, it refers to a magical ingredient that transforms clever cognitive mechanisms into autonomous agents.

The reason why some philosophers think that the sciences can’t really help us with the big question of what free will consists in, is that this question, at least in the philosophical discourse, has centrally been understood as the question of whether or not our actions are fully causally determined by nature, and what implications this has for our status as autonomous agents.

In philosophy the debate has been mainly played out between parties that believe free will exists. Libertarians (who believe that determinism and freedom are incompatible and that we are free) have argued with compatibilists (who also believe that we are free, but who also believe that we are determined and that freedom and determinism are fully compatible). The position that many neuroscientists seem to favour of hard determinism also exists in the philosophical debate, but is much less prominent.[2] Hard determinists agree with the libertarians that freedom of the will and determinism are incompatible, but they agree with the compatibilists that determinism is probably true and therefore they conclude that we do not have free will.

It is easy to see why the participants in the dispute between libertarians and compatibilists were not very impressed by the findings from cognitive science. At best, neuroscience can help with the question whether or not human behaviour and decision making really is determined. (Even here you might wonder whether that really is possible, given that scientific experiments seem to simply assume the truth of determinism.) It is difficult to see how the sciences should be able to help us to settle the conceptual question, whether or not freedom of the will is compatible with determinism.[3]

As our contributors show in some detail it is difficult to see how the advances of the neurosciences will help us with the metaphysical question of whether or not free will is compatible with the truth of determinism (Roskies). It is an interesting sociology of science fact that even though most philosophers agree broadly on this, it still does not seem to stop the publication of more and more books on the question. John Baer and colleagues (Baer et al. 2008) have edited an excellent volume on free will and psychology and the question of the relationship between determinism and free will is very prominent within the book.

In our volume we start from the assumption that the cognitive sciences will not be able to help us directly with the discussion between libertarians and compatibilists. However, a simple dismissal of the scientific findings as irrelevant to the free will debate would be premature. Even though these findings might not have any bearing on the truth or falsity of compatibilism or libertarianism, it doesn’t follow that the scientific debate about volition is unrelated to the philosophical one about free will. To see why this is the case imagine for a moment that compatibilism is correct, and freedom of the will is consistent with a belief in determinism. Obviously this does not mean that everything that is determined would also be free. Compatibilists are keen to work out which kind of determinants are the ones that make us free and which ones don’t. What does it take to be an agent and for an instance of behaviour to count as a free action (see e.g. Velleman 1992)? If compatibilism were true, human autonomy would consist in behaviours being produced by the right kinds of mechanism, but which are those? Is it crucial that we can evaluate action options rationally? Is it important that we can initiate actions in a context (or stimulus) independent manner? Does it matter whether we can think about our own mental states? Must we be able to think about thinking? Do we need to control our behaviour in a conscious manner in order to be responsible for it (see Roskies)? These are all important philosophical questions about necessary conditions for human autonomy. But the answer to the question of whether or not we have any of these abilities, and to what extent we employ them in everyday action and decision-making is at least partly an empirical question. If we think that mental agency or conscious control is necessary for free will, the question of whether we are free or not will depend on whether we have these capacities and can exercise them in going about our everyday business. Hence, the question of whether we are free or not, turns out to crucially depend on a better scientific understanding of the machinery of our minds.

Now, take away our initial assumption about the truth of compatibilism, suppose libertarianism was the right answer to the determinism question. It seems not a lot would change. If you want to know the conditions for libertarian human autonomy you look at the compatibilist ones and add the crucial indeterminist extra. Libertarians will make use of exactly the same or at least very similar ingredients (rationality, action initiation, self knowledge) to give the necessary conditions an agent needs to fulfil, before adding ’true’ freedom of choice that turns merely self-controlled behaviour into truly free action.[4] If we don’t have or rarely exercise self-knowledge and rationality in generating our behaviour, this would seem to spell trouble for a libertarian account of free will just as much as for a compatibilist account.

Still one might worry that an account of free will needn’t be interested in questions about mechanisms. Philip Pettit (2007) for has for instance argued that neuroscience is only threatening to commonsensical free will notions that are built on an “act of will” picture.[5] Such a picture might attempt to identify specific cognitive mechanisms that causally enable “acts of will”. According to Pettit this would be to underestimate the social dimension of the free will discourse. Whether somebody is free or not crucially depends on their ability to justify their actions according to the normative rules set by society. The idea of free will works because we take ourselves to be free agents and have normative ideas about what the behaviour of a free and responsible agent should look like. As we all strive to fulfil the normative ideal of a free agent our behaviour begins to resemble the ideal more and more. This assimilation to the ideal is not dependent on their being a prior independent mechanism in the agents that would lead to such behaviour naturally without normative guidance.

This move familiar to the philosopher of mind from the discussion on folk psychology ((Dennett 1987)) seems particularly appealing in the case of free will and autonomy. Free will is such a loaded concept, a concept on which so much human high minded ideals are based that it seems rather likely that it is an idealisation, an abstraction, that is not reducible to actual mechanisms.

If Pettit’s suggestion turns out to be along the right lines, it wouldn’t follow that our interest in mechanisms is futile. It does nothing to undermine the thought that there may be specific functions of the machinery of mind, which are essential for being an autonomous agent. If it were really the case that free will is important only as a self-attribution, this would still invite the question why this self-attribution is important. It seems very likely that it can only be important if the self-attribution does in some way exert an influence on behaviour. Whether the will, understood in the way Pettit proposes, does exercise a causal influence on our behaviour will then be a question that will be answered in part by looking to the empirical sciences. It might well turn out that even though free will might be conceptually quite coherent and even though it would be quite possible to have the mechanisms to be free, cognitive science reveals that the machinery of the mind works in way that are inconsistent with our being free agents. We will label this worry the zombie challenge.

The zombie challenge

The zombie challenge is based on the amazing wealth of findings in recent cognitive science that demonstrate the surprising ways in which much of our everyday behaviour is controlled by processes that are automatic and unfold in the complete absence of consciousness. One of the key aims of this volume is to see whether and how these findings are relevant for our thinking about free will and even more importantly to give examples of how these findings might form the basis for empirically informed accounts of autonomy.

What we are calling the zombie challenge is quite different from the debate in the metaphysics of mind (Chalmers 1996) about the logical possibility of creatures physically and functionally like us that lack phenomenal consciousness. Our question is about the actual world and whether consciousness plays as significant a functional role in the causation of our behaviour as is normally assumed. The zombie challenge suggests that the conscious self takes a back seat in the control of our behaviour. The functional machinery that is responsible for the initiation and control of much of our everyday behaviour does all of its work without involving the conscious self. If the zombie challenge is effective, consciousness will turn out to be epiphenomenal. Such a finding would seem to be bad news for any belief in free will, even if we grant the correctness of compatibilism. Most compatibilists think that some form of control is what is special about freedom.[6] Exactly what the control consists in is up for grabs, but in most accounts it seems to be an assumption that this control is consciously exercised. If the zombie challenge is upheld, the capacity for conscious control would turn have been revealed to be idle. We may well have a capacity for conscious control, but if it doesn’t do any work in the production of our everyday behaviour this would seem to have major implications for views that take conscious control to be necessary for the exercise of free will. It would show that one of the conditions the compatabilist identifies as necessary for free will does little or no work in generating our behaviour. This increase the temptation to conclude the same is true for free will. We believe it is something like the zombie challenge that motivates scientists to claim that free will is an illusion. The experiments that motivate free will scepticism have very little to do with the truth or falsity of determinism.[7] The experiments all seem to point to the conclusion that the conscious self is an epiphenomenon.