It’s OK if ‘my brain made me do it’: People’s intuitions about free will and neuroscientific prediction

Highlights

• Some have argued that perfect neuroscientific prediction conflicts with free will.

• We found most participants judge that neuro-prediction is consistent with free will.

• Judgments about free will varied with the presence or possibility of manipulation.

• Judgments about free will were mediated by judgments about ‘bypassing’.

• Results suggest people’s understanding of free will is minimally metaphysical.

Abstract

In recent years, a number of prominent scientists have argued that free will is an illusion, appealing to evidence demonstrating that information about brain activity can be used to predict behavior before people are aware of having made a decision. These scientists claim that the possibility of perfect prediction based on neural information challenges the ordinary understanding of free will. In this paper we provide evidence suggesting that most people do not view the possibility of neuro-prediction as a threat to free will unless it also raises concerns about manipulation of the agent’s behavior. In Experiment 1 two scenarios described future brain imaging technology that allows perfect prediction of decisions and actions based on earlier neural activity, and this possibility did not undermine most people’s attributions of free will or responsibility, except in the scenario that also allowed manipulation. In Experiment 2 the scenarios increased the salience of the physicalist implications of neuro-prediction, while in Experiment 3 the scenarios suggested dualism, with perfect prediction by mindreaders. The patterns of results for these two experiments were similar to the results in Experiment 1, suggesting that participants do not understand free will to require specific metaphysical conditions regarding the mind–body relation. Most people seem to understand free will in a way that is not threatened by perfect prediction based on neural information, suggesting that they believe that just because “my brain made me do it,” that does not mean that I didn’t do it of my own free will.

Introduction

In recent years, a number of prominent scientists have argued that free will is an illusion, claiming that research in neuroscience and psychology shows that our ordinary experience of and beliefs about free will are systematically mistaken (e.g., Bargh, 2008, Cashmore, 2010, Coyne, 2012, Greene and Cohen, 2004, Harris, 2012, Montague, 2008, Tancredi, 2007, Wegner, 2002). Following Nahmias (2009), we will call those scientists who think free will is an illusion “willusionists.” When willusionists challenge free will, they do so by explicitly claiming that we lack free will as it is understood by laypersons, that free will as it is ordinarily understood is incompatible with the naturalistic assumptions and the experimental evidence of the modern mind sciences. The willusionists advance a variety of arguments and evidence to defend this conclusion. In some cases, they suggest that neuroscience demonstrates that there is no non-physical soul or mind and assume that people think such a soul is required for free will (e.g., Cashmore, 2010, Montague, 2008). In other cases, they suggest that neuroscience is showing that people do not really make choices and assume that free will requires the ability to make choices (e.g., Coyne, 2012). In other cases, they suggest that neuroscience and psychology show that consciousness is epiphenomenal and assume that free will requires a causal role for conscious mental states (e.g., Libet, 1999, Wegner, 2002). And in some cases, they suggest that these sciences provide evidence for determinism or for mechanism and assume that free will is incompatible with determinism or mechanism (e.g., Bargh, 2008, Greene and Cohen, 2004, Tancredi, 2007). These arguments often overlap, but in every case, the willusionists refer to evidence demonstrating that information about brain activity can be used to predict behavior before people are aware of having made a decision (e.g., Bode et al., 2011, Libet, 1999, Soon et al., 2008, Soon et al., 2013), and they use these findings as evidence of a naturalistic worldview that allegedly challenges our ordinary beliefs about free will.

In contrast to these willusionists’ view, most philosophers hold that free will is compatible with the naturalistic worldview of science, even if the laws of nature turned out to be deterministic (e.g., Dennett, 2003, Fischer and Ravizza, 1998, Frankfurt, 1971, Wolf, 1990; see Bourget & Chalmers, 2013, for data on views held by philosophers). However, willusionists generally dismiss these accounts of free will as radical revisions to the ordinary understanding of free will. Coyne (2012) writes that compatibilists “have concocted ingenious rationalizations for why we nevertheless have free will of a sort. It’s all based on redefining ‘free will’ to mean something else.” Harris (2012, p. 16) writes, “the ‘free will’ that compatibilists defend is not the free will that most people feel they have.” Greene and Cohen (2004, p. 1780) write, “intuitive free will is libertarian, not compatibilist. That is, it requires the rejection of determinism and an implicit commitment to some kind of magical mental causation.” However, the willusionists’ assumption that compatibilists are simply “changing the subject” is questionable. Existing evidence on non-philosophers’ intuitions about free will suggests that many, perhaps most, people do not understand free will to be incompatible with determinism (Murray and Nahmias, 2014, Nahmias et al., 2006; cf. Nichols & Knobe, 2007). Furthermore, existing evidence suggests that most people do not understand free will to be incompatible with naturalism or physicalism (Mele, 2012, Monroe and Malle, 2010, Monroe et al., 2014, Stillman et al., 2011).

It could be the case, as Harris (2012) argues, that people do not really understand what the neuroscientific evidence or the naturalistic worldview suggests. He writes, “If the laws of nature do not strike most of us as incompatible with free will, that is because we have not imagined how human behavior would appear if all cause-and-effect relationships were understood” (p. 11). Harris here offers an error theory for why it may appear that most people judge free will to be compatible with our actions being governed by the laws of nature, though they would not if they appropriately envisioned what this would mean. Harris helpfully offers a scenario that he thinks would get most people to appreciate how human behavior would appear if it could be understood in terms of law-governed neural activity. His scenario involves the specter of a “perfect neuroimaging device”—a device that can perfectly predict in real time everything a person will think or do even before the person is aware of what she is about to think or do. The device can accomplish this feat through the information provided by brain scans. The device allows neuroscientists to predict with 100% accuracy everything a person will decide and do, including any attempts to try to trick them or to be unpredictable. We will refer to this ability to perfectly predict decisions and behavior based exclusively on neural activity as “neuro-prediction.” According to Harris, this neuro-prediction scenario is a useful way to help people understand that free will is “an illusion,” because neuro-prediction makes salient what it would mean if all of our actions were completely governed by law-like relationships between brain and behavior (2012, pp. 10–11; see also Greene & Cohen, 2004, p. 1781).²

On the view favored by many willusionists, the possibility of neuro-prediction should conflict with people’s view of free will for at least two reasons, though they are not always clear about whether they take these two reasons to be equivalent. First, willusionists often claim that neuro-prediction appropriately highlights the cause-and-effect relations entailed by the fully law-governed relationships between brain and behavior and claim that people take free will to be incompatible with behavior being fully law-governed in this way (Bargh, 2008, Cashmore, 2010, Coyne, 2012, Greene and Cohen, 2004, Harris, 2012, Tancredi, 2007). Second, willusionists often claim that neuro-prediction suggests the inexistence of a causally relevant non-physical mind or soul (Cashmore, 2010, Gazzaniga, 2011, Greene and Cohen, 2004, Libet, 1999, Montague, 2008), and these willusionists typically claim that people take free will to require a causally relevant non-physical mind or soul.

However, another possibility is that most people do not understand free will to be inconsistent with neuro-prediction, because they do not believe that there is an inherent conflict between free will and predictability based on neural activity. Many people may already assume that our mental activity is the product of (or based on, or identical with) brain activity, such that the possibility of such neuro-prediction is neither bewildering nor threatening. Others may assume that mental activity (perhaps especially conscious states) cannot be understood solely in terms of brain activity, but still have no beliefs that commit them to a metaphysical theory that rejects the possibility of prediction of decisions and behavior based on neural activity (e.g., their views are consistent with certain forms of non-reductive physicalism).

It is also possible that many people are what we might call “theory-lite” when it comes to commitments regarding the relationships among mind, brain, and free will (Nahmias & Thompson, 2014). If so, then people’s conception of free will may not be dependent on dualism or non-naturalism as many willusionists claim. To be clear here, we are not claiming that many people accept a physicalist theory of mind or reject the possibility of a non-physical mind or soul; we are only claiming that most people’s notions of free will and responsibility may not depend on dualism or non-naturalism, at least not in the strong sense claimed by many willusionists in which it is claimed that dualism is essential for free will (for relevant evidence, see Mele, 2012, Monroe et al., 2014). When it comes to free will, most people might be concerned primarily with whether or not it is “me” that causes my actions, regardless of whether that “me” turns out to be my brain and body or an emergent property of my brain or a distinct non-physical thing. On a theory-lite view, the possibility that my brain causes behavior need not conflict with me causing my behavior. Instead, the main concern posed by the possibility of neuro-prediction may be that it raises the threat of manipulation—another agent controlling me (whatever that me may be)—or the threat of “bypassing”—some process causing my behavior other than me, or the relevant processes within me. Manipulation clearly conflicts with people’s conception of free will and responsibility (Shepard and Reuter, 2012, Sripada, 2012), perhaps because manipulation is interpreted as bypassing the agent’s own mental states and reasons, whereas neuro-prediction need not involve such bypassing (Murray and Nahmias, 2014, Nahmias, 2014). On this view, perfect neuro-prediction alone would not be sufficient to undermine most people’s attributions of free will or responsibility; instead, it would only do so to the extent that it raises concerns about manipulation.

We conducted three experiments designed to test between these conflicting hypotheses about people’s ordinary views about free will and responsibility. Since practically all willusionists’ arguments against the existence of free will begin with evidence of neural-based prediction of behavior, people’s reactions to the possibility of neuro-prediction is a useful test of these willusionists’ claims about ordinary people’s views (see Harris, 2012 and Greene & Cohen, 2004 for explicit endorsements of neuro-prediction as a useful test). Thus, the focus of our experiments is to test whether neuro-prediction is sufficient to undermine attributions of free will (as the willusionists predict) or whether it only threatens free will to the extent that it raises threats of manipulation (as we predict).

The three experiments are also designed to shed light on whether the ordinary conception of free will requires a specific metaphysics of mind (e.g., dualism, physicalism) or is instead “theory-lite.” The neuro-prediction scenarios used in Experiments 1 and 2 strongly imply physicalism, with Experiment 1 describing scenarios that are compatible with non-reductive physicalism, while Experiment 2 describes scenarios that make physicalism even more salient by explicitly using the language of identity theory (i.e., the view that mental activity just is brain activity). The scenarios used in Experiment 3 are, instead, explicitly dualist in nature. Under the view that the ordinary understanding of free will requires a specific metaphysics of mind (e.g., dualism), one would expect to find different patterns of results among these experiments (e.g., low attributions of free will in the scenarios that rule out dualism). On the other hand, if the ordinary conception of free will is theory-lite, we would expect a similar pattern of results across all three experiments.