Consciousness and Moral Responsibility, by Neil Levy (Oxford University Press), £27.50/$45
It may at first seem obvious that consciousness is required for moral responsibility. After all, how could an agent be blameworthy if she were unaware of the features that made her action wrong? But in recent decades, both philosophers and scientists have challenged our intuitive picture of the importance of consciousness. It sometimes seems appropriate, for instance, to blame negligent agents, who are unaware of the wrong actions that they are performing. And recent studies in cognitive science have suggested that unconscious processes are responsible for producing many of our behaviours, including some for which we might be morally responsible.
In Consciousness and Moral Responsibility, Neil Levy defends the moral importance of consciousness against these challenges. His primary aim is to argue for a claim he calls the Consciousness Thesis, according to which moral responsibility requires that agents be conscious of the facts that give their actions moral significance.
Levy uses the first four chapters to lay the groundwork for the Consciousness Thesis, appealing extensively to the empirical literature on consciousness. Among the most important results in this area are the dramatic cases of automatism, in which unconscious agents display sophisticated behaviour. Levy describes the example of Kenneth Parks, who drove several miles and stabbed his mother-in-law while sleepwalking.
Despite our emerging understanding of the importance of the unconscious mind, Levy argues that consciousness is nevertheless required for morally responsible action. According to the “global workspace” theory of consciousness, which Levy endorses, conscious states are just those states that are globally “broadcast” within an agent’s mind. When a state is globally broadcast, it is available to all of an agent’s mental subsystems, including those that are unconscious.
Because global broadcasting is what allows these unconscious subsystems to communicate and work collaboratively, Levy claims, it also enables a range of behaviours that would otherwise be impossible. It allows agents to deliberate, to evaluate their current intentions against their long-term values, and to display flexibility in response to a wide variety of stimuli. In short, it makes it possible for agents to act like agents, rather than automatons. Since conscious states are globally broadcast states, and globally broadcast states allow for the possibility of responsible agency, Levy concludes that consciousness is necessary for responsibility.
Thus, Levy provides an elegant explanation for why unconscious agents like Parks are not morally responsible for their actions. In Chapters 5 and 6, Levy defends the more ambitious claim that normal agents cannot be responsible for actions caused by unconscious attitudes. An agent who acted on unconscious sexist attitudes, for example, would not be blameworthy. Levy examines two kinds of accounts of responsibility – “real self” and “control” accounts – and argues that both kinds of accounts imply that agents are only responsible when they act on conscious attitudes.
On real self accounts of responsibility, two conditions must be met in order for an agent to be praiseworthy or blameworthy for an action. First, the action must reflect the person that the agent truly is (i.e. her “real self”). Second, the action must express an attitude that has some morally significant (good or bad) content. Levy argues that neither condition is satisfied by agents who act on unconscious attitudes. Levy proposes that an agent’s real self consists in the totality of her attitudes, rather than in any particular attitude considered in isolation; therefore, no particular attitude can reflect the agent’s real self unless it has been evaluated against the agent’s other beliefs and desires. Only conscious, globally broadcast content can be checked against the agent’s other attitudes for potential conflict and rejected or modified if necessary. So real self accounts of responsibility commits us to the Consciousness Thesis.
Levy also argues that unconscious, implicit attitudes lack the moral content required to ground praiseworthiness or blameworthiness. Consider the unconscious association that some people make between “men” and “career” (and “women” and “family”). As a belief about statistics, this association has no moral content. Although it can cause agents to act badly – by leading them, unconsciously, to prefer male job applicants over female ones – it doesn’t reflect badly on the moral character of the agents who possess it. Levy claims that other unconscious attitudes are also likely to be morally neutral beliefs about statistical associations, and, as such, agents cannot be blameworthy for the actions that they produce. Once again, real self accounts of responsibility commits us to the Consciousness Thesis.
However, the real self theorist could push back against both of Levy’s claims. In response to the first, the real self theorist might note that many unconscious, unbroadcast attitudes are nevertheless available for conscious broadcast. If an agent’s other attitudes demanded it, presumably, these unconscious attitudes would be pulled into the global workspace, where they could be consciously evaluated. And perhaps the potential to be broadcast is sometimes sufficient for an unconscious attitude to reflect the agent’s real self. If an agent were sufficiently concerned about equality, for instance, he would access the attitudes relevant to sexism and change them. The fact that he hasn’t done so might indicate that his other attitudes implicitly “approve” of his sexism, and thus that his real self is (to a degree) sexist. The same might be true of negligent agents. If they cared enough, they would access the relevant unconscious attitudes, and thus their negligence reflects their real self without the relevant attitudes being conscious.
In response to Levy’s second claim, the real self theorist might note that even if some unconscious attitudes are beliefs about mere statistical associations, others are likely to have much richer moral content. Imagine an agent who systematically pursues his own interests while ignoring the welfare of others. If this agent is not very reflective, he may have no conscious awareness of the attitudes that cause him to act. Nevertheless, it seems as though the attitudes that cause his selfish behaviour are likely to have morally significant content. They may incorporate, for instance, a morally objectionable stance towards others. If so, they would reflect negatively on the agent’s moral character to a much greater degree than would neutral beliefs about statistics. If unconscious attitudes can represent the real self while carrying morally objectionable content, then the real self theorist’s conditions for responsibility can be met, even for actions caused by unconscious attitudes.
Levy also applies his argument to control theorists, who believe that moral responsibility requires control. Because unconscious attitudes are not currently available for modification, Levy contends, they are not under the agent’s control in the relevant sense – neither the attitude nor the actions that result from it will be apt to change in response to reasons. So, on Levy’s view, consciousness is also necessary for responsibility on control theories. But, like real self theorists, control theorists might respond by appealing to the potential for unconscious attitudes to be accessed. Presumably, there are many unconscious attitudes that would be brought into the global workspace, and modified in response to reasons, if the agent cared sufficiently about them. Control theorists might argue that this kind of counterfactual responsiveness is sufficient for responsibility-level control.
Although both real self and control theorists have grounds to challenge some of Levy’s arguments, he provides an illuminating discussion of the connection between consciousness and responsibility and makes a strong case that entirely unconscious agents cannot be responsible. This brief book is a valuable contribution to the literature and will be fascinating to those interested in the intersection of cognitive science and moral psychology.