Last week, Slate inaugurated Human Nature, a feature that will cover the intersection of science, culture, and politics. The idea is to explore how science shapes culture and politics, as well as how culture and politics shape science. I’ve got a background in the philosophy of natural and social sciences, but my training in the practice of those fields is limited, so I’m going to make some mistakes. In the spirit of science—and the Internet—I’ll try to correct mistakes and rethink conclusions as I go along, with the help of smart readers like you.
Today I’m revisiting last week’s installment, which argued that the Stanford Prison Experiment, or SPE (a 1971 study conducted by psychologist Philip Zimbardo, in which upstanding young men assigned to be “guards” in a mock jail abused their “prisoners”), didn’t explain the abuse of Iraqi prisoners at Abu Ghraib. To make sense of the terms and arguments that follow, read the original piece first. Here are several objections readers made to that piece, along with some discussion of the issues they raised.
1. “Had Zimbardo not cancelled the experiment, it’s entirely reasonable to assume that the Stanford guards would have engaged in the exact same behavior” that happened at Abu Ghraib. Entirely reasonable? Why? Empirically, this didn’t happen. The notion that it would have happened if the experiment hadn’t been stopped seems, well, not very scientific. You could argue that the aggression at Stanford was heading toward violence and would have gotten there in days or weeks. But you could argue at least as plausibly that there’s a line between preliminary aggression and inflicting serious injury, and that most people don’t cross that line easily. On this view, the interruption of SPE prevented it from falsifying rather than confirming the conjecture that it would have become Abu Ghraib.
2. “At the outset of the study, the ‘guards’ were told in no uncertain terms that physical violence was absolutely forbidden. I think that if this prohibition had not been in place, violence would have occurred.” Again, it’s conjecture. Zimbardo says some incidents in the SPE indicate that intercession by supervisors was all that averted more serious violence. He was right, ethically, to draw this line. Still, by drawing it, he prevented falsification of the conjecture.
Look at it the other way. One reader, The Bell, explains why, in a real, no-holds-barred situation like Abu Ghraib, the Stanford guards might have been less aggressive, not more:
I am sure that most of them … began to understand–perhaps intuitively–that they had the opportunity to test the bounds of their role’s authority. … That would explain why Zimbardo’s guards thought up abuses every bit as sadistic as those imagined by the guards at Abu Ghraib yet always stopped short of real physical harm. Rather than losing sight of their situation as “just a game,” their (sub)conscious understanding that it was just a game–and with willing participants–enabled them to give themselves permission to act in ways that would normally be anathema to them.
The just-a-game factor is even more likely to account for the “good guards” at Stanford who failed to stop the “bad guards” from abusing prisoners. Speaking of participants in such an experiment, The Bell observes, “Those who feel uncomfortable and rather silly about the whole business–such as myself–tend to act very blandly and passively. We are looking to do just enough to satisfy the instructor/observer.” This is exactly what most of the “guards” at Stanford did. Zimbardo supposes they would have done the same, or worse, at Abu Ghraib. But at Abu Ghraib, knowing the stakes were real, they might have felt compelled to intercede.
The argument, in a nutshell, is this: What happens in a “situation” inside the lab doesn’t predict what will happen in a “situation” outside the lab, because inside the lab, everyone knows it’s just a “situation.” The best rejoinder to this argument is the Milgram experiment, which invites a different set of objections. But that’s a debate for another day.
3. Stanford and Abu Ghraib are more similar than different. This is impossible to settle scientifically. Zimbardo screened out some variables (principally, differences in personality) in order to focus on others. I don’t reject his claims 30 years ago that the SPE showed the effects of the variables he included. What I reject are his suggestions today that those variables explain Abu Ghraib. The effects in the latter case were far more serious, and the intuitive explanation for this difference is the role of variables that were excluded from the SPE.
4. Exclusion of variables is necessary to science.Slate reader Mpento writes, “Factors that are left out of an experiment are generally done so as to minimize confounds to the results (which can occur if there are either too many variables to control for in an experiment of limited size, or variables that cannot be controlled for, and thus could affect the results of the variables you are attempting to study).” I understand that this exclusion helps the scientist get cleaner data on the variables he’s isolating. But what’s being excluded is most of reality. That’s why the SPE explains what happened at Stanford far better than it explains what happened at Abu Ghraib.
Reader Ben Kowal adds, “Science is based on the assumption of determinism (i.e., that a finite, and thus limited, number of laws are all that is necessary to perfectly predict and control nature). Therefore, it doesn’t seem to make sense to argue that Zimbardo’s attempt to limit the number of possible forces acting on humans in his experiment disqualifies his research from science.” I’m not arguing that Zimbardo’s exclusions make the SPE unscientific. I’m arguing that the exclusions, which are indeed scientific, show the limits of using laboratory social science to explain an event outside the lab.
5. Even if factors Zimbardo excluded (such as racism) account for Abu Ghraib, they vindicate his situationist approach, because they, too, are situational. “Race is a role like any other, and there are power differentials associated with every combination of race,” argues reader RG. Reader La Rana agrees: “The difference in severity of treatment can once again be ascribed to differences in the situations.” But if even an intuitively dispositional factor such as racism can be construed as situational, aren’t we defining “situational” so broadly that it means nothing? Doesn’t this make situationism unfalsifiable?
6. Situational factors don’t preclude personal responsibility. Reader Matt Murphy observes, “In both the Stanford and Milgram experiments, it is clear that not all subjects participate in all the abuse, which belies that idea that the circumstances are the only causal factor.” However, “If someone (say a police officer or my boss) comes up to me on the street and tells me to strip some men, pile them up, and take photos, I probably won’t do it. But if I am a solider, in a war, in a prison, temporarily assigned to be a guard, maybe I would do it.” I agree. Situational factors mitigate but don’t exclude dispositional factors.
The irony is that mechanical determinists don’t believe in logical inferences such as “Situational factors don’t preclude personal responsibility.” They believe in correlations. And there’s a high correlation between mechanical determinists and those who dismiss personal responsibility. As Slate reader LexP puts it, “Saletan seems not to grasp the distinction between the moral plane, where one is enjoined to act as best one can and is held responsible for one’s acts, and the natural, where there is only interdependence and mechanical cause and effect. The notion that we are free to act morally is perhaps a necessary illusion, but an illusion nonetheless. The Stanford experiment is simply a striking demonstration of this.” I don’t see why we should—or, as a determinist might put it, how we can—live by the rules of the “moral” plane, pretending that the “natural” plane is irrelevant.
7. Supervisors’ input is part of the situation. This is where Objection 7 undermines Objection 8 and complicates the prosecution of the abuse at Abu Ghraib. The more factors you define as situational, the less responsibility you can attribute to any specific participant, even a supervisor.
Several readers defend Zimbardo’s notion of responsibility. Reader joovin8 writes, “Zimbardo freely admits at www.prisonexp.org that he intervened in the experiment in counterproductive ways after getting caught up in the ‘situation’ he created. Does that absolve him of responsibility? Hardly, and I’m sure he’d agree.”
But Zimbardo’s destructive intervention didn’t begin after he was “caught up” in the situation. It began at the outset. The bags and chains his researchers introduced were two examples. Reader Stephen Oliver, citing a BBC study, points out another:
Thus one reason why the guards in our own research for the BBC did not behave as brutally as those in the Stanford study, was that we did not instruct them to behave in this way. Zimbardo, in contrast, told his participants: “You can create in the prisoners feelings of boredom, a sense of fear to some degree, you can create a notion of arbitrariness that their life is totally controlled by us, by the system, you, me - and they’ll have no privacy…. In general what all this leads to is a sense of powerlessness.”
Reader Zinya notes that Zimbardo has written, “I have been responsible for constructing evil barrels that produced many bad apples.” But Zimbardo also uses the barrel metaphor to depict himself as just another apple: “The most profound measure of the power of this situation was the extent to which it transformed me.” Another reader, psychout, diminishes Zimbardo’s responsibility with a similar metaphor: “Because of situational factors beyond the control of the participants, the guards, supervisors, indeed even Zimbardo himself became rats in their own maze.”
Barrels, apples, rats, mazes. All these metaphors confuse the extent to which supervisors are responsible for what happened at Abu Ghraib. Barrel-makers become apples; maze-makers become rats; situation-makers become role-players. If you hold them accountable for creating situations but not for playing the consequent roles in which they exacerbated abuse, will they bear the full extent of their responsibility? For that matter, will the supervisors of the Abu Ghraib abuse pass off their creation of that situation as the product of a larger situation in which they were only playing preordained roles? Where does situationism end? That’s the most important question. And it’s the one situationism can’t answer.