Source: A cognitive reinterpretation of Stanley Milgram's observations on obedience to authority. American Psychologist 45: 1384-1385 (1990).

Prologue : I still receive an occasional response to this brief essay, a decade after it has been published, perhaps because it has been reprinted in a popular undergraduate anthology. Most readers read into my essay a dismissal of Milgram’s work. Nothing could be farther from my original intention: I believe that Milgram’s is one of the most profound, original, and interesting psychological research programs of the 20th century. His interpretation may be wrong, at least in part, but his experimental work will stand the test of time. It has been unjustly criticized because it unmasks an unpleasant truth. It has been endlessly vilified, as well, by some so-called students of human nature who have themselves long ago given up any hope of throwing light on the human condition and who resent anyone who actually does. It has come under fire because some powerful academics are unable to separate wheat from chaff. Of course Milgram lied to his subjects, and this is unfortunate.  All the same, lies are common practice in psychology, so that does not explain his critics’ ire.   Moreover, Milgram gave his subjects one of the best lessons they ever had, probably, and yet he is accused of inhumanity. He tried, himself, to understand man’s inhumanity to man, and ended up suffering from the same pettiness of spirit, from the same snow blindness, he sought to identify and extinguish.  He threw some genuine light on such horrors as Nazi Germany, contemporary Guatemala, million+ Iraqi deaths, or "missile defense," and was never forgiven for this.  It was not only they  who are guilty, he dared to say, but we.

Another ignored aspect of my note is its practical implications. It suggests that closed-mindedness and resistance to conceptual change are of paramount importance in human affairs, making again a mockery of our claim to rationality, yet the subject receives little attention from psychologists (organizational experts know a bit better). Nor did my brief note merely put forward a new interpretation of Milgram's work, but a proposal for an exciting research program which could cast light on the human condition. The failure to follow up on this idea, it seems to me, is lamentable, given "the low wheat-to-chaff ratio in the glittering piles of research publications" in the social sciences (Sherif, 1979).

These then are my main reasons for posting this ancient paper: Dissociate myself from Milgram’s vilifiers, and invite, once more, a courageous social scientist out there (perhaps outside the USA, where such research may never be published) to test my cognitive reinterpretation and thereby enrich, amplify, and commemorate Milgram’s work.




milg.jpg (10345 bytes)

Stanley Milgram's observations on obedience to authority have exerted a great deal of influence on such diverse disciplines as social psychology, holocaust studies, and political science. In Milgram's basic paradigm, a subject walks into a laboratory believing that s/he is about to take part in a study of memory and learning. After being assigned the role of a teacher, the subject is asked to teach word associations to a fellow subject (who in reality is a collaborator of the experimenter). The teaching method, however, is unconventional--administering increasingly higher electric shocks to the learner. Once the presumed shock level reaches a certain point, the subject is thrown into a conflict. On the one hand, the strapped learner demands to be set free, he appears to suffer pain, and going all the way may pose a risk to his health. On the other hand, the experimenter, if asked, insists that the experiment is not as unhealthy as it appears to be, and that the teacher must go on. In sharp contrast to the expectations of professionals and laymen alike, some 65% of all subjects continue to administer shocks up to the very highest levels.

Milgram's classical experiments have come under severe attacks. Some critics argue that their validity hinges on the acting ability of the learner and experimenter and that most subjects were probably able to sense the unreality of the situation. Others question the relevance of these laboratory results to the larger world. Still others question the ethics of the basic experimental design. Professor Milgram, for his part, insists that these and other misgivings are traceable to the unsavory nature of his results: "Underlying the criticism of the experiment," he says, "is an alternative model of human nature, one holding that when confronted with a choice between hurting others and complying with authority, normal people reject authority." (Milgram 1974, p. 169) Milgexp.jpg (3173 bytes)

But while Milgram's observations attracted much criticism and praise, and while they have somewhat altered our views of the human condition, the interpretation he provides for his results has received scant attention--the debate focuses for the most part on the reality and extent of obedience, not on its underlying causes.

According to Milgram, "the essence of obedience consists in the fact that a person comes to view himself as the instrument for carrying out another person's wishes, and he therefore no longer sees himself as responsible for his actions. Once this critical shift of viewpoint has occurred in the person, all of the essential features of obedience follow." Thus, "the major problem for the subject is to recapture control of his own regnant processes once he has committed them to the purposes of the experimenter." (Milgram, 1974, pp. xii, xiii). In addition to this presumed agentic state, Milgram goes on to explain, a variety of factors lock the subject into the situation. These include situational factors such as politeness and awkwardness of withdrawal, absorption in the technical aspects of the task, the tendency to attribute impersonal quality to forces that are essentially human, a belief that the experiment serves a desirable end, the sequential nature of the action, and anxiety.

It seems reasonable to suppose that something like the constellation of factors above accounts for the subjects' obedience. At the time Milgram's fascinating observations were carried out, such an explanation would appear highly probable and fairly complete. Note however that, unlike Milgram's observations, the evidence in favor of this explanation is fairly circumstantial. The best that can be said, for instance, about Milgram's key postulate of the agentic state is that it makes sense and that, if true, it may account for the data.

Certain developments in cognitive psychology which came to their own after 1974 suggest the presence of another key causative factor. Before making the connection between obedience and cognition, we need to familiarize ourselves with these developments. For the sake of brevity, I shall only describe here a few recent experiments which seem most directly applicable in this context.

These experiments were patterned after, and provide striking confirmation for, earlier observations (Festinger, Riecken, & Schachter, 1964; Milgram 1974, 1984; Kuhn, 1974; Karmiloff-Smith & Inhelder, 1975; Ross & Anderson, 1982). In these more recent experiments, subjects were recruited to evaluate the efficacy of a self-contained instructional manual. Before they could provide the needed appraisal, they were told, they needed to acquire a first-hand experience of its content by studying it and following the instructions it provided for about four hours. At some point in the teaching process, the manual introduced a false volume formula for a sphere--a formula which led subjects to believe that spheres are 50% larger than they are. Subjects were then given an actual sphere and asked to determine its volume; first by using the formula, and then by filling the sphere with water, transferring the water to a box, and directly measuring the volume of the water in the box. The key question was: Would subjects believe the evidence of their senses and abandon their prior beliefs in the formula, the competence of the experimenter, and the legitimacy of the entire setup? Preliminary observations (Nissani, 1989a,b) suggested that the task was far more difficult than expected: no subject decisively rejected the false formula or declined to use it in subsequent tasks. In later experiments various attempts were made to ease the conceptual transition called for by this experiment. In one variation (Nissani & Hoefler, 1991), all subjects held a Ph.D. degree in a natural science and were employed as research scientists and professors in two major research universities. A special section--involving measurements of a second ball--was introduced and constructed with the deliberate aim of helping these scientists break away from the false formula. In another variation (Nissani & Maier, 1990), the discrepancy concerned the circumference of an ellipse, thereby ruling out the possibility that earlier results were ascribable to the difficulty of dealing with three dimensional concepts. But none of these variations substantially altered the initial results:

The preliminary observations reported here suggest that the importance of conceptual conservatism has been underestimated in the psychological literature and that the insistence that the phenomenon constitutes one of the major impediments to progress in the history of ideas could very well be correct. In particular, although conceptual conservatism has received the attention of experimentalists, although its importance in human affairs has been long recognized and although the results reported here are based on a small sample, the qualitative outcome of this study--all subjects clung in practice to an observationally absurd formula and none rejected it outright even on the verbal level--are surprising. Even when we deal with ideologically neutral conceptions of reality, when these conceptions have been recently acquired, when they came to us from unfamiliar sources, when they were assimilated for spurious reasons, when their abandonment entails little tangible risks or costs, and when they are sharply contradicted by subsequent events, we are, at least for a time, disinclined to doubt such conceptions on the verbal level and unlikely to let go of them in practice. (Nissani, 1989b; pp. 23-4).

These results poignantly suggest a rather counterintuitive conclusion, a conclusion which could not be fully appreciated by Milgram sixteen years ago: Transitions from one belief to another are not as smooth as common sense or intuition would seem to suggest. For instance, attempting to provide a retrospective explanation for his failure to reject the false formula of the sphere, one of our subject-scientists wrote: "It is difficult to imagine that one could be deliberately deceived in an exercise like this" (Nissani & Hoefler, 1991).

Consider now the typical subjects in Milgram's basic paradigm. They come to participate in a scientific investigation at an impressive, well-equipped laboratory at Yale University. They have every reason to believe that the experiment is conducted by responsible people. They have never before heard of tortures, killings, inhumanity, or immorality associated with modern scientific experiments. In fact, not only Milgram's subjects, but all of us, share this eminently reasonable belief. We know that university scientists are working under various legal and ethical constraints, and that barbarism of any kind is simply out of the question. Milgram's subjects walked into the experiment taking for granted the responsibility and basic morality of the entire setup. As in the case of subjects in a conceptual shift experiment, the experimental evidence contradicted this belief. Disobedience in such a setting presupposes a conceptual shift: Milgram's subjects had to discard their belief in the morality of the experimenter. "I knew you wouldn't let anything happen" to the learner, said one of Milgram's subjects in an effort to explain his obedience (Milgram 1974, p. 83). In contrast, one disobedient subject treated the experimenter "as a dull technician who does not see the full implications of what he is doing." (Milgram, 1974, p. 48).

If this conclusion is correct, Milgram's opinion that "people can't be counted on to disobey malevolent authority," and that "they obey as long as the command comes from legitimate authority" (Milgram, 1974, p. 89), is either incorrect or only partially correct. Rather, what people cannot be counted on is to realize that a seemingly benevolent authority is in fact malevolent, even when they are faced with overwhelming evidence which suggests that this authority is indeed malevolent. Hence, the underlying cause for the subjects' striking conduct could well be conceptual, and not the alleged "capacity of man to abandon his humanity . . . as he merges his unique personality into larger institutional structures." (Milgram, 1974, p. 188).

Some of Milgram's own data support this interpretation:

1. Milgram's results are surprising. Laymen and social scientists alike were unable to foretell the extent to which subjects would obey the experimenter. Likewise, laymen and social scientists were unable to predict the behavior of subjects in our conceptual shift experiments (Nissani & Hoefler, 1991). It is entirely possible that both inadequate forecasts are traceable to a single factor--underestimating the excruciating difficulty of abandoning a strongly held, eminently reasonable, belief.

2. In one of Milgram's variations, subjects were led to believe that the experiment was conducted by a private research firm. This single difference decreased obedience rate from 65% to 48% (Milgram, 1974, p. 69). The conceptual shift interpretation is consistent with this observation: because private research firms are less prestigious than Ivy League schools, it is easier under these conditions to abandon the belief in the experimenter's essential decency.

3. In another experimental variation, a single element of betrayal and patent injustice was introduced, leading obedience to drop from 50% to 40% (Milgram, p. 66). This result is again consistent with the conceptual shift interpretation proposed here.

A new experimental variation in Milgram's protocol could readily test the purely moral conflict Milgram's observations have so far failed to capture. As in other variations, the authority figure must be portrayed as legitimate. But, by the time the teaching session begins, this figure must been made to appear as highly callous and irresponsible. If successful, this variation would create a genuine conflict between willing obedience to malevolent authority and the voice of conscience. The data from Milgram's own experiments, the near-unanimous consent of Milgram's survey respondents, and cognitive data underscoring the difficulty of discarding reasonable beliefs, strongly suggest that obedience in such situations would be substantially lower than it was in Milgram's basic paradigm.

Hopefully, this crucial experiment will be undertaken by one or more readers of these lines.

REFERENCES (Author’s note: the references have been slightly revised and updated)

Festinger, L., Riecken, H. W. & Schachter, S. (1964). When Prophecy Fails. New York: Harper.

Karmiloff-Smith, A. & Inhelder, B. (1975). "If you want to get ahead, get a theory." Cognition, 3, 195, (1975).

Kuhn, T. S. (1974). The Structure of Scientific Revolutions. Chicago: University of Chicago Press.

Milgram, S. (1974) Obedience to Authority. New York: Harper & Row.

Milgram, S. (1984) Cyranoids. A paper presented at the 1984 meeting of the American Psychological Association.

Nissani, M. (1989) A hands-on instructional approach to the conceptual shift aspect of scientific discovery. Journal of College Science Teaching, 19: 105-107.

Nissani, M. (1989) An experimental paradigm for the study of conceptual conservatism and change.

Nissani, M.  (1994).  Psychological, historical, and ethical reflections on the Mendelian paradox. Perspectives in Biology and Medicine 37: 182-196.

Nissani, M. (1994). Conceptual conservatism: an understated variable in human affairs? (click here to read this article online) Social Science Journal 31: 307-318.

Nissani, M. & Hoefler-Nissani, D. M. (1992). Experimental studies of belief-dependence of observations and of resistance to conceptual change. Cognition and Instruction 9: 97-111.

Ross, L. & Anderson, C. A. (1982). In A. Tversky & D. Kahneman (Eds.) Judgment under uncertainty: heuristics and biases. Cambridge: Cambridge University Press, pp. 129-152.

Back to Moti Nissani's Home Page