Neuro Milgram – Your brain takes less ownership of actions that you perform under coercion

In a series of classic experiments performed in the early 1960s, Stanley Milgram created a situation in which a scientist instructed volunteers to deliver what they believed to be painful and deadly electric shocks to other people. Although this now infamous research has been criticised at length, people continue to be unsettled by its main finding – that most of the participants were quite willing to harm others when ordered to do so.

These findings have since been used to explain why some people can commit heinous crimes – the fact that they are "only obeying orders" handed down by others may make it easier for them to deny responsibility for their actions.

Milgram's study was an investigation of people's readiness to bow down to authority and obey coercive instructions to inflict harm, but it did not explore how the participants felt during the experiments, nor what was happening in their brains.

A new study in Current Biology that used brain wave recordings shows that when a person is coerced into performing an action, his or her brain processes the outcomes of that action differently from how it processes equivalent actions carried out intentionally, suggesting that coercion does indeed diminish our sense of agency, or the sense that we are in control of our actions.

The new research is based on a phenomenon called temporal binding, first described in 2002 by neuroscientist Patrick Haggard of UCL. It refers to the observation that the brain compresses time during voluntary actions, but not involuntary ones, so that our actions and their consequences are perceived to occur more closely together, enhancing our sense of agency.

Haggard and his collaborators performed two experiments to determine whether coercion alters perception of the time interval between an action and its outcome.

In the first, each of 30 pairs of female volunteers took turns at being the "agent" and the "victim". In the coercive condition, a researcher stood over the table at which pairs of participants sat, stared intensely at the agent, and ordered her to either take money from the victim or give her a mild electric shock, outcomes which the participant initiated by pressing one of two computer keys. In the free-choice condition, the researcher was more detached, and told the agents that they could inflict a shock on the victim in order to earn themselves some money, or take money from the victim, or refrain from such actions, and that they were totally free to choose – again, the agents initiated the outcomes they wanted by pressing different keyboard keys.

In both the coercive and voluntary conditions, the agents' key presses caused an audible tone to occur, with a variable random delay of up to one second, and the participants had to estimate the interval between the two.

Under the coercive, but not the free-choice condition, the participants estimated the intervals to be significantly longer than they actually were – in other words they showed a reduced temporal binding effect, suggesting that they had a reduced sense of agency over their actions. This was the case for both the harmful and the non-harmful outcomes, showing that the effect was not related to whether or not the agents inflicted any harm, but was due instead to the coercive instructions they were given.

Before the experiment, all the participants had filled out a questionnaire measuring empathy and various personality traits, and the more empathetic ones experienced a more dramatic reduction in agency when their actions had a harmful outcome compared to the less harmful ones. But most of them acted somewhat vindictively, giving roughly the same amount of electric shocks when they played the role of agent as they had received when they were the victim.

In a second experiment, the researchers recruited 22 more volunteers and used electroencephalography (EEG) to examine whether coercion alters the brain wave patterns associated with action outcomes. Consistent with earlier work showing that one particular type of brain wave, called N1, is far bigger for outcomes of voluntary actions than for those of actions performed under (non-coercive) instruction, they found that the outcomes of coerced actions produced smaller N1 waves than the outcomes of actions performed in the free-choice conditions.

A questionnaire administered after the experiments further revealed that the participants felt more responsible for their actions during the free-choice than the coerced trials.

Thus, being coerced into doing something seems to reduce our sense of agency, not just psychologically, but also at the level of basic brain function – the neural processing of the outcomes of coerced actions resembles the outcomes of passive movements more closely than the outcomes of voluntary or intentional actions.

Fifty years ago, Milgram reported that ordinary people usually comply with coercive instructions, even if it means inflicting real or apparent harm on others. These new findings show that the effects of coercion on the sense of agency are universal, as opposed to being associated with any particular characteristic of personality.

They also suggest a reason why people can be coerced so easily – coercion may automatically reduce the link between an action and its outcome, emotionally distancing people from distasteful consequences and diminishing their sense of moral responsibility.

Haggard and his colleagues believe that their findings could have profound implications for legal responsibility and the criminal justice system. They do not legitimise the notorious strategy used by defendants at the Nuremberg war crimes trials, that they were just obeying orders. But the researchers argue that the law would do well to shift the focus from people who obey orders to those who give them out.

About the author

Post written by Mo Costandi (@Mocost) for the BPS Research Digest. Mo trained as a developmental neurobiologist and now works as a freelance writer specialising in neuroscience. He writes the Neurophilosophy blog, which is hosted by The Guardian, and is the author of 50 Human Brain Ideas You Really Need to Know, published by Quercus in 2013. His second book, Neuroplasticity, is due to be published by the MIT Press later this year.

Further reading

Caspar, E., Christensen, J., Cleeremans, A., & Haggard, P. (2016). Coercion Changes the Sense of Agency in the Human Brain Current Biology, 26 (5), 585-592 DOI: 10.1016/j.cub.2015.12.067