Milgram Revisited: “Only obeying orders”

Adolf Eichmann (below) was tried in Israel in 1961 for crimes against humanity. Eichmann’s crimes were in his handling of the logistics of transporting millions of Jews to concentration camps built for the purpose of their extermination during WW2.    His defence was ‘only obeying orders’.

only obeying orders

Milgram’s Experiments

Eichmann’s defence inspired Stanley Milgram (1933-1984) a psychologist at Yale University to perform one of the most infamous of social psychology experiments.   He wanted to find out how far a person would proceed in inflicting pain in obedience to the authority figure of the experimenter.

He chose people varying widely in age, occupation and education as subjects. From the subject’s point of view he and another person came to the laboratory to take part in a study of memory and learning.   They were given a scientific sounding rationale for the study.   One of them became a ‘teacher’, the other a ‘learner’.

The ‘teacher’ was shown an electrified chair and given a sample 45 volt shock. The ‘learner’ was then placed in the electrified chair, wired up with electrodes and told that he will be read lists of word pairs.    When he hears the first one again he is supposed to say the second word.   If he makes a mistake he will be given an electric shock.

The ‘teacher’ was then taken to a different room (linked by intercom) where he was placed in front of a control panel with thirty switches labelled 14 to 450 volts with descriptive designations from ‘slight shock’ to ‘danger: severe shock’ and finally ‘xxx’.

The experimenter in a grey lab coat starts the ‘teacher’ off with the word pairs.   He tells the ‘teacher’ to administer the next level of electric shock when the ‘learner’ gets the word pairing wrong.

In fact, the ‘learner’ is an actor who receives no shocks but acts as though he did. The experimenter unemotionally in the face of objections from the ‘teacher’ just encourages him to continue the experiment. When the learner starts to make mistakes the level of electric shock is stepped up. “At 75 volts, he grunts; at 120 volts he complains loudly; at 150 he demands to be released from the experiment… At 285 volts his response can be described only as an agonised scream. Soon thereafter he makes no sound at all.” (Milgram 1973)

Milgram solicited predictions of the result of his experiment from 14 colleagues.   They almost uniformly predicted that the ‘teacher’ would refuse to obey the experimenter at 150 volts where the learner asks to be released from the experiment.   In fact about 60% of the ‘teachers’ went to the end of the experiment administering the full 450 volts.

The subjects (‘teachers’) were usually agitated during the experiment – sweating, trembling, stuttering or laughter fits.   They were much relieved at the end of the experiment to find they had not hurt anyone – though some showed no emotion throughout. Variations of the experiment were tried to find what parameters influenced the result. When the ‘teacher’ was allowed to choose the shock level rather than being told to raise it to the next level, the average shock chosen was less than 60 volts – lower than the point at which the victim showed the first signs of discomfort.   Only 2 out of 40 subjects went as high as 320 volts.

When the experiment was altered so that the experimenter gave his instructions by telephone rather than being in the room with the ‘teacher’, the percentage of ‘teachers’ obedient to the 450 volt level fell to 20%. When the ‘teacher’ was relieved of the responsibility of pulling the lever that administered the shocks, and merely specified the level at which the shock should occur the percentage of ‘teachers’ going all the way to 450 volts went up to 92%.   In that case the subjects claimed that the responsibility rested with the person who actually pulled the lever.

Milgram concluded, “The essence of obedience is that a person comes to view himself as the instrument for carrying out another person’s wishes, and he therefore no longer regards himself as responsible for his actions… The most far-reaching consequence is that the person feels responsible to the authority directing him but feels no responsibility for the actions that the authority prescribes.  Morality does not disappear – it acquires a radically different focus: the subordinate person feels shame or pride depending on how adequately he has performed the actions called for by authority … the most fundamental lesson of our study [is that] ordinary people, simply doing their jobs, and without any particular hostility on their part, can become agents in a terrible destructive process. Moreover, even when the destructive effects of their work become patently clear and they are asked to carry out actions incompatible with fundamental standards of morality, relatively few people have the resources needed to resist authority.” (Milgram 1973)

The experiment has been repeated in various parts of the world with even higher percentages of obedience in some cases. Milgram gave the subjects  personality tests in an attempt to find those aspects of personality or character that would predict how far the subjects would go, but he found no correlation with any of the test results.

New Experiment

Now a slightly different version of Milgram’s experiment has been performed by a group of ‘cognitive neuroscientists’ from University College London and the Free University of Brussels led by Patrick Haggard (Caspar 2016). They wanted to find out to what degree the participants felt ‘in charge’ when they knowingly inflicted pain on each other and when they knew the aim of the experiment.

In the new experiments the participants (all female) were tested in pairs.  They took turns being ‘agent’ and ‘victim’ thus ensuring reciprocity.    Each was initially given £20.  The agent sat facing the ‘victim’ and so could monitor directly the effect of her actions.    In a first group of participants, the agent could freely choose on each trial to increase her own remuneration by taking money (£0.05) from the ‘victim’ (financial harm) or not.   Money transfer occurred in 57% of trials.   In a variation of the experiment the financial harm was accompanied by an electric shock to the ‘victim’ at a level that was tolerable but not pleasant (the electric shock was administered in 52% of trials).

In both of these groups the experimenter stood by and in some cases told the agent to take the money (group 1) or shock the victim (group 2).   In the other cases the experimenter told the agent to exercise her free choice.   There were also a number of trials as controls where the experimenter asked the agent to press the space bar whenever she wanted (‘active’) and where the experimenter pressed the agent’s finger on the space bar (‘passive’).

In order to investigate the agent’s ‘sense of agency’ (“the subjective experience of controlling one’s actions, and, through them, external events”) the key presses caused a tone to sound after a few milliseconds (variously 200, 500 and 800 msec) and the participants were asked to judge the length of the interval. The rationale behind this is that action-result times are perceived as shorter when the person carries out the action voluntarily (such as raising one’s arm) than when the action is done passively (someone else raises the arm).   So if coercion reduces this sense of agency, interval estimates should be longer in the coercive than in the free-choice condition.

Thus there were several comparison sets of data: free choice versus coercion, financial harm versus physical harm and harm versus no harm, as well as the control conditions (active versus passive).   When they were ordered to press a particular key (producing either harm or no-harm), the participants judged their action as more passive than when they had free choice and they perceived the time interval from the tone as longer (p=0.001).   This did not change depending on whether there was a harmful outcome, though it did when the potential harm was greater (ie physical rather than financial).

So the conclusion was that the coercion rather than the severity of the actual outcome was the determining factor in the sense of agency.   The agent experienced less sense of agency when she was coerced than when she freely chose between the same options – regardless of whether harm was actually inflicted.   So the plea “Only obeying orders” might not be just an attempt to avoid blame “but may rather reflect a genuine difference in subjective experience of agency.”

The participants were also given personality tests prior to the experiments to see if there were any predisposing factors. It was found that those scoring higher on empathy showed a greater reduction in the sense of agency when their actions had harmful outcomes.

In a second experiment, the same procedures were used but the agents were also hooked up to an electroencephalogram (EEG) to investigate changes in brain activity associated with the free choice / coercive conditions.   When an unpredictable stimulus such as a tone occurs it is followed by a ‘negative response potential’ approximately 0.1 seconds later in the frontal part of the scalp (usually referred to as the N100).  The expectation was that the N100 would be larger in amplitude when the agent freely chose her action compared with that when she felt coerced. This was indeed the case (amplitude ratio approx 1.3).   So not only the subjective ‘sense of agency’ but also neurophysiological activity is reduced under coercion.

Haggard says people genuinely feel less responsibility for their actions when following commands regardless of whether they are told to do something evil or benign. So the ‘only obeying orders’ excuse shows how a person feels when acting under command.

Before Haggard did these experiments he had (along with the majority of neuroscientists and many modern philosophers) already espoused the philosophical viewpoints of physicalism¹ , epiphenomenalism² and reductionism³ . He claims that mind-body causation is dualist and “incompatible with modern neuroscience” since most neuroscientists believe that “conscious experiences are consequences of brain activity rather than causes.”    “Philosophers studying ‘conscious free will’ have discussed whether conscious intentions could cause actions, but modern neuroscience rejects this idea of mind–body causation.    Instead, recent findings suggest that the conscious experience of intending to act arises from preparation for action in frontal and parietal brain areas. Intentional actions also involve a strong sense of agency, a sense of controlling events in the external world.    Both intention and agency result from the brain processes for predictive motor control….” (Haggard 2005)

And again : “… the cause of our ‘free decisions’ may at least in part, be simply the background stochastic fluctuations of cortical excitability.” (Filevich 2013)


These experiments are interesting but care must be taken in their interpretation and in the consequences that may be claimed for jurisprudence. It is not clear whether the neurophysiological activity causes the subjective sense of agency or vice versa. What the experiments do reveal is that coercion causes both reduced sense of agency and reduced neurophysiological activity.

The experiments only concern what Elizabeth Pacherie terms ‘present-directed intentions’ ie those intentions which “trigger the intended action, …sustain it until completion, …guide its unfolding and monitor its effects”.   They do not touch upon ‘future directed intentions’ which are “terminators of practical reasoning about ends, prompters of practical reasoning about means and plans, and intra- and interpersonal coordinators” (Pacherie 2006).

One presumes that Haggard and his colleagues were motivated by future directed intentions when they decided to do the experiments and write their paper. They were not simply acting as the result of ‘stochastic fluctuations of cortical activity.’   If so, then the sweeping general conclusion loses its force.

The 18th century philosopher David Hume (1711-1776) thought that every object of the mind must be either an immediate perception or an ‘idea’ – a faint copy of some earlier perception.(Hume 1748)   This was criticised by his contemporary Thomas Reid (1710-1796) :“It seemed very natural to think that [Hume’s book] required an author and a very ingenious one at that; but now we learn that it is only a set of ideas that came together and arranged themselves by certain associations and attractions.” (Reid 1764)

According to Haggard and his colleagues not even ideas are now involved – only ‘stochastic fluctuations of cortical activity’.

The question of who bears personal responsibility is important to the rule of law. Certainly the person who gives the order to harm is culpable for the consequences.   But this does not absolve the person who actually carries out the order. The degree to which people feel responsible on average does not change the moral responsibility of any individual act.   Nor does it justify the inclusion of such ‘mitigating’ circumstances into criminal law.

Hannah Arendt (1906-1975) wrote a book on Eichmann’s trial (Arendt 1963), in which she coined the phrase “the banality of evil”.   It is not clear exactly what she meant by the phrase.   Milgram thought that she meant that Eichmann was not a “sadistic monster” but “an uninspired bureaucrat who simply sat at his desk and did his job“, and that she “came closer to the truth than one dare imagine.” (Milgram 1973)   It may well be true that in some situations evil is not perpetrated by fanatics and psychopaths but by ordinary people who see their actions as normal (banal = commonplace) within the prevailing conditions.   If so all of us are capable of committing horrendous crimes when the circumstances are right.

It is easy to see how ‘situationism’ (the philosophical belief that people act according to the situation in which they find themselves rather than by virtue of any moral or philosophical outlook they might have) is a credible paradigm.   But it predicts the actions of only 2/3rds of the subjects in the Milgram study. The new study suggests that there are character traits (eg ‘empathy’) that predict some aspect of the results (ie reduced sense of agency where there was a harmful outcome) more accurately.   But we do not excuse criminality on the grounds of character traits.

Evil was a common place in Nazi Europe, but for Arendt that did not render it excusable.   Whilst Arendt saw Eichmann as a cog in the machinery of the Final Solution she did not excuse his crimes nor fail to hold him morally responsible for his actions.   “If the defendant excuses himself on the ground that he acted not as a man but as a mere functionary whose functions could just as easily have been carried out by anyone else, it is as if a criminal pointed to the statistics on crime – which set forth that so-and-so-many crimes per day are committed in such-and-such a place – and declared that he only did what was statistically expected, that it was a mere accident that he did it and not somebody else, since after all somebody had to do it.” (Arendt 1963)

Despite the pressures some people do have the resources to buck authority even when the authority has far more clout than the man (or woman) in the grey lab coat. For example, the US GI Ronald Ridenhour forced the US congress to investigate the My Lai massacre in Vietnam where US servicemen massacred an entire village of 300 or more civilians in 1968. (Ridenhaur 1969) There were many people such as Raul Wallenburg (1912-1947) and Oskar Schindler (1908-1974) who protected Jews from the holocaust despite great personal risk.

If there are attempts to influence the law on the basis that these experiments prove diminished responsibility they should be dismissed.

The above contains passages extracted from the book Rethinking the Mind. Get the first volume here:


  1. Physicalism: the doctrine that everything is physical, ie all is matter and energy in its many forms and hence subject to the laws of physics.
  2. Epiphenomenalism: the doctrine that mental events are mere by-products of physical events and that mental events in themselves do not cause anything. In the classic description due to Thomas Huxley (1825-1895) consciousness is simply a collateral product of the working of the body in the same way that a steam whistle accompanies the work of a locomotive engine.
  3. Reductionism: the doctrine that explanations of phenomena are to be found in the smaller entities that comprise it eg) heredity in terms of DNA or in this case, human activities in terms of neural firings.


Arendt H (1963) Eichmann in Jerusalem: A Report on the Banality of Evil Penguin

Caspar EA, Christensen JF, Cleeremans A & Haggard P (2016) Coercion changes the Sense of Agency in the Human Brain Current Biology available at

Filevich E, Kühn S, Haggard P (2013) Antecedent Brain Activity Predicts Decisions to Inhibit PLOS 1 (February 13, 2013) available at

Haggard P (2005) Conscious Intention and Motor cognition Trends in Cognitive sciences vol 9(6) p 290-295 available at

Hume D (1748) An Enquiry Concerning Human Understanding Section II ‘Of the Origin of Ideas’ para 12.

Milgram S (1973) The perils of Obedience Harpers Magazine p62-77 available at

Pacherie E (2006) Towards a dynamic theory of Intentions in Pockett S, Banks WP & Gallagher S (eds) Does Consciousness cause behavior MIT Press p 145-167 available at

Reid T (1764) An Enquiry into the Human Mind chapter 2.6 (ed J Bennett) available at

Ridenhour R  (1969) Letter to US Congress available at

3 thoughts on “Milgram Revisited: “Only obeying orders”

  1. Very eloquent! Very persuasive. Highly convincing.

    Ironically before I did my psychology course and before I’d ever heard of Milgram I found my self in a relevant (though much more trivial) situation.

    It was in a biology unit and the lecturer arrived with a large bucket of live frogs. He said, “Cut a leg off.” Only two of us objected. A very clear briefing followed explaining that we would fail the unit if we did not carry out the current nerve experiment and so not get the degree. That left only me objecting and became notorious for saying, “The road from cutting off a live frog’s leg to running a concentration camp in Poland is long but it is very straight.”



  2. I used to speak of Milgram’s experiment to my students; they were regularly stunned by the results! It has been reproduced lately via a french TV and it went exactly as Milgram discovered! Some “teachers” were crying but still obeyed the hierarchy!

    I wasn’t aware of the new experiments you mention. It seems to me that the correlations between the choices and neurophysiological responses are very interesting. They make me ask if these protocols could be adapted to work with primates like Macaques maybe. They are interesting since different very close species (on anatomical grounds) have very different behaviors, depending notably on the environment in which they usually live (rich with food and a society rather free from hierarchy, or a poor environment in which a strong hierarchy is found).

    But maybe the results could be misinterpreted and lead to question the responsibility in humankind. We surely are a social species and therefore we must in some way be sensitive to some sort of role distribution that could even be part of some instinctive behavior. But we’re far from being simple primates without culture and responsive behavior! Aren’t we!


  3. The original experiment probably has some valid if somewhat scary implications.The later ones only meaningless justifications to justify
    greater levels of irresponsibility

Leave a Reply

Your email address will not be published. Required fields are marked *