Taking stock of situationism
The Stanford Prison Experiment seems to portray a world in which the very idea of personal integrity seems absent - a Goffmanesque world where there are no selves, only selves-in-roles, selves who slide frictionlessly from role to role, in each case conforming to the expectations of the role and whatever principles of right behavior come attached to its script.
Behind the mask, another mask; behind all the masks, a vacuum; beneath the vacuum, a mask once again. Indeed, some theorists have drawn conclusions very close to the claim that when it comes to character, there’s no there there. John Doris, in his recent book Lack of Character, concludes from a careful examination of the experimental literature that human character, defined as a fixed set of dispositions toward certain behavior, is largely a myth.49 Gilbert Harman, drawing on the same experimental literature, agrees.50 Both draw on an interpretation of the experiments I have been reviewing called situationism, defended most cogently by Lee Ross and Richard Nisbett. At this point, I want to detour briefly to discuss the merits of situationism, in order to distinguish the argument I am offering from a version of situationism that I think is wrong.The thesis of situationism is, quite simply, that differences in situations account for much more of the observed variation in human behavior than do differences in personality. Our tendency to believe otherwise, that is, to ascribe people’s behavior to their personality or character - or, for that matter, their voluntary choices - rather than the situation they are in, is what situationists criticize as the “fundamental attribution error.”51 Situationists
to make conditions better for their boys. When one mother told me she had never seen her son looking so bad, I responded by shifting the blame from the situation to her son.
‘What’s the matter with your boy? Doesn’t he sleep well?’ Then I asked the father, ‘Don’t you think your boy can handlethis?’ He bristled, ‘Of course he can-he’sarealtoughkid, aleader.’ Turningto the mother, he said, ‘Come on, Honey, we’ve wasted enough time already.’ And to me, ‘See you again at the next visiting time.’” Ibid.49 See generally John M. Doris, Lack of Character: Personality and Moral Behavior (2002).
50 Gilbert Harman, Moral Philosophy Meets Social Psychology: Virtue Ethics and the Fundamental Attribution Error, 99 Proc. Aristotelian Soc’y 315 (1999).
51 Ross & Nisbett, supra note 4, at 29-89. point, for example, to some of Milgram’s findings.[533] In one variant of his shock experiment, Milgram placed the subject on a team with another subject - actually, of course, a confederate of Milgram. When the confederate uncomplainingly obeyed orders to continue the shocks, 90 percent of the subjects went along, but when the confederate refused to administer high- level shocks and walked away from the experiment, compliance by subjects plummeted to 10 percent. Clearly, it would be implausible to assume that the subject population in one variant differs radically in propensity to comply from the subject population in the other. Instead, the conclusion must be that situational differences generate this dramatic swing from near-universal compliance to near-universal rebellion. Similarly, Doris points to experiments by Isen and Levin that showed that people who find a dime in the coin return of a pay telephone were vastly more likely to help a stranger pick up papers she has dropped than people who find no dime.[534] Again, the power of the situation appears to dominate or even dwarf the power of personality and character in determining behavior. The lesson of situationism, drawn by Harman and Doris, seems to be that we have no character that disposes us to behave consistently across situations.[535] The disconcerting picture seems to be a near-determinism of situations, in which minuscule differences in the situation - a dime or no dime, the presence or absence of other people in the room - turn into major differences in behavior, and individual idiosyncracy explains very little of the differences.
(It is a near-determinism, not a strict determinism, because individual idiosyncracy still plays some explanatory role.[536])I believe that caution is in order about what conclusions to draw from these observations, however. There’s no denying the situationists’ point that minute changes in situation can dramatically affect the proportion of people exhibiting a given behavior. But the situationists have a hard time explaining why different people behave differently in the same situation. In Milgram’s basic experiment, two-thirds of the subjects complied, but one-third did not. The point is more general. Throughout the experimental literature of social psychology we find striking and statistically significant correlations between experimental variables and subjects’ responses - but, significant as they may be, the correlation coefficients seldom exceed 0.5, which by the standards of physics is a low correlation, signifying that the manipulated variable accounts for only one-fourth of the variance in behavior. People differ, and the question for situationism is how these differences should be explained.
The answer, Ross and Nisbett tell us, lies in the fact that people construe situations differently, that “it is the situation as construed by the subject that is the true stimulus.”[537] Thus, the differences in response arise not from differences in human character but rather from differences in perception and construal.
Perhaps. Yet I find this explanation, according to which the one-third noncompliance rate in the Milgram experiment is explained by arguing that the noncompliant third perceived the situation differently from the two-thirds of compliers, to be both too convenient and too ad hoc, given that we don’t actually know anything about how Milgram’s subjects construed the situation. In Ross and Nisbett’s view, what I have called a near-determinism of situations becomes more like a true determinism. Ross and Nisbett localize individual idiosyncracy in the capacity for perception and construal, while accepting a version of stimulus-response determinism according to which the stimulus (the situation-as-construed) leads subjects to uniform responses.
But why? Why parse the individual this way, into a perception/construal capacity that exhibits idiosyncrasy and a responsive capacity that exhibits little or none? To do so seems arbitrary, and borders on downright inconsistency. After all, construing a situation is itself a kind of action, and one would suppose that a consistent situationist should posit that situations account for most of the variation in construals as well as in responses. In that case, however, the situationist is left with no explanation for variation among individuals placed in the same situation. Moreover, even if the situationist is right about individual variability in construing situations, one can reply that personality lies in large part in our habits of perception and construal, so at least some form of personality theory survives the situationist objection.I prefer to think of situations - the independent variables that experimenters manipulate - as sources of pressure or of temptation. Quite simply, the experiments demonstrate how difficult - but not impossible - swimming against the situational tide is. It’s so difficult that in the basic Milgram experiment only a third of the subjects are able to bring it off. Adding a compliant team-mate makes it more difficult still, so that only one subject out of ten was able to resist; while adding a noncompliant team-mate makes resistance easy enough that nine out of ten subjects were able to succeed at resisting the orders to continue administering shocks. In the terms of our initial metaphor, situational changes alter the relative gradient of both the high road and the low road. What the experiments do show, quite graphically, is that seemingly minor manipulations of the environment can cause astonishingly large changes in the ease or difficulty of action, the angle of incidence between the two roads.
Putting the situationists’ point in these terms - that is, that situations transform the ease or difficulty of certain courses of action - avoids the implication of determinism.
The situation sets conditions under which we choose, but the numbers strongly imply that these conditions do not render choice impossible. Notice that if every subject complied with Milgram’s experiment, it would provide evidence that the experiment had uncovered a mechanism akin to a physical reflex, over which we have no choice or control. And if only a few subjects out of the thousand complied, we might regard them as pathological cases, and excuse them from blame on the grounds that they have a screw loose somewhere. The actual two-thirds/one- third split precludes us from drawing either of these deterministic conclusions about the compliers.Because I resist situational determinism, I resist as well the radical suggestion that our deep-seated propensity to fall into predetermined roles (as in the SPE) means that we have no core self whose integrity matters, only a collection of selves-in-roles, each seeking its own harmony between behavior and principle, but without any larger unity of self. The experiments do show that we lack robust consistency across situations. This should not surprise us, however. After all, if some of my roles impose inconsistent moral demands - if, for example, with a wig on my head and a band round my neck I will be asked to do for a guinea what I would otherwise think it wicked and infamous to do for an empire - and my daily life leads me to occupy all these roles, and if, further, the actions I take in each role lead me to adopt beliefs that vindicate those actions, then dissonance theory predicts that I will preserve my conception of myself as a morally upright individual in the only way left: by abandoning the belief that my other beliefs should be consistent.[538] We purchase integrity, what Gerald Postema calls the “unity of practical consciousness,” at the price of logic, the unity of theoretical consciousness.[539] The experimental demonstration that we lack robust consistency across situations shows that integrity consists of a complex unity, stitched together with a great deal of self-deception that allows us to deny inconsistencies and the dissonance they induce. Integrity remains something that we seek. The problem, then, remains the one we began with: that the quest for integrity, manifested in all the psychological phenomena we have been reviewing, can drive us to behavior as disconcerting and morally repellent as that shown in the Stanford Prison Experiment or in Milgram’s demonstration.