Are compliant subjects morally blameworthy?

The Milgram experiments lead quite naturally to the depressing reflection that human nature is much more readily disposed to wrongful obedience than we might have expected or hoped.

The Inculpating View holds that no matter how widespread wrongful obedience is, and no matter how deep its roots within human nature, wick­edness remains wickedness; the fact that wickedness is the rule rather than the exception excuses no one. Suppose that experimenters were to demon­strate that two out of three people will walk off with someone else’s $100 bill if they are sure they can get away with it. The experiment suggests that greed has roots deep within human nature, but that creates no excuse for theft. The temptation to obey is like greed or any other temptation. It is perfectly natural to give in to it - that’s why they call it temptation! - but being perfectly natural excuses nothing.^[466]

It might be objected that the analogy between Milgram obedience and greed is a bad one. No matter what his rationalizations, the thief knows that theft is wrong, or so we may suppose. He simply allowed his baser drives to override his moral judgment, and that is why we don’t allow his greed to excuse him. If our earlier corruption-of-judgment explanation of Milgram obedience is correct, however, the drive to obey operates at a deeper level, undermining our very capacity to distinguish right from wrong.

But this objection overlooks the fact that we generally do not excuse wrongful behavior because it resulted from bad judgment - if anything, the fact that the wrongful choice was the product of judgment rather than of passion or pathology condemns it even more.

Or does it? Try a thought experiment. Suppose a group of high-school seniors is given a test of judgment, such as the familiar multiple-choice analogies test. And suppose that the test’s difficulty is calibrated so that every student in a control group passes it. This time, however, the test is admin­istered under extraordinary conditions: throughout the test, a large-screen television in the testing room broadcasts a video of a good-looking couple making enthusiastic, noisy, and improbably athletic love. Under these con­ditions, we will suppose, two-thirds of the students fail the test.

Clearly, we should conclude that passing the test under such distracting conditions is really hard. The numbers prove it.^[467] We would be foolish to blame the students for failing; and we would be cruel to punish those who failed, for example by refusing to admit them to college because of their bad scores. Surely we would blame the situation, which obviously undermined their capacity to judge.

The analogy to Milgram is straightforward. When people had the Milgram experiments described to them, they all passed the “test” of moral judgment: without exception, they predicted that they would break off the experiment well before the 450-volt maximum (and it should be clear that their prediction is in reality a moral judgment that complying to 450 volts would be wrong). But in the actual experiment, two out of three failed their test. Pursuing the parallel, we would be foolish to blame them for failing, and cruel for punishing them. The situation excuses their compliance. This is the Exculpating View.

In short, the Inculpating View holds people responsible for their wrongful obedience, regardless of how common wrongful obedience is, or how deeply rooted it may be in human nature. The Exculpating View excuses wrongful obedience whenever it is the statistical norm, because that fact shows how unreasonably difficult it must be to disobey under such circumstances.

How are we to decide between the Inculpating and the Exculpating Views? I propose approaching the problem indirectly, by looking at parallel puzzles in the treatment of psychologically based defenses in the criminal law. Admittedly, criminal responsibility raises different issues from moral responsibility, and the psychological defenses the law recognizes do not include the deep-seated propensity to obey. Despite these obstacles, there are enough suggestive parallels that examining the criminal law issues will allow us to triangulate toward our own question.

Consider the “heat of passion” or “extreme emotional disturbance” defense in homicide cases, which reduces murder to manslaughter.^[468] In its formulation in the Model Penal Code, the defense is available whenever a “homicide which would otherwise be murder is committed under the influ­ence of extreme mental or emotional disturbance for which there is reason­able explanation or excuse.”^[469] The canonical situation is a husband murdering his wife and her lover when he finds them in bed.

Surprisingly, however, this cliched bit of melodrama is not the typical situation in which the defense actually arises. Victoria Nourse examined every reported heat-of-passion decision in US courts between 1980 and 1995, and discovered a disturbing pattern. The paradigm case for heat of passion turns out to be men angry at women for exiting a relationship: boyfriends upset that their girlfriends have left them; long-separated husbands whose wives finally file for divorce; long-divorced husbands who learn that their ex­wives are remarrying; and men served with protective orders forbidding them from approaching wives or girlfriends they have battered. In other words, the typical heat-of-passion “provocation” turns out not to be infidelity, but a woman’s attempt to lead her own life free from her killer’s dominion; and the killer’s “passion” seems not to be sexual jealousy so much as the over­whelming desire to control and own a woman.^[470]

The Model Penal Code aimed to reform the criminal law by taking a scientific approach to human psychology.

Nourse is critical of the Model Penal Code’s approach, and I am as well. Her findings about the circumstances under which the heat-of-passion defense gets invoked provide a virtual reductio ad absurdum of the mechanistic treatment of provocation. Mitigations reflect judicial and legis­lative compassion for wrongdoers who have committed crimes under un­usually trying circumstances. Does a man who flies into a murderous rage because his wife dates someone else three years after they separated really deserve our compassion?^[473] Surely not; and surely it is appropriate to moralize about whether his murderous rage was justified.

In line with this thought, Nourse proposes a different approach to extreme emotional disturbance, based on the concept of a warranted excuse.^[474] Begin with the philosophically attractive idea that emotions can be appropriate or inappropriate - that they embody (or at least correspond with) evaluative judgments of objects and situations that can be true or false, warranted or not warranted.^[475] If a man flies into a murderous rage because his wife has been raped, his emotion reflects a warranted evaluative judgment about the rape - that rape is wicked and horrible.

If, on the other hand, the killer has become enraged because his wife is leaving him, his emotion corresponds with the evaluative judgment that she is not entitled to leave him - perhaps even that wives are never entitled to leave their husbands. This evaluative judgment deserves no endorsement from the law. Even assuming that he was in the grip of extreme emotional disturbance when he killed her, the heat-of-passion excuse should be unavailable to him, because the emotion is unjustified. In line with this reasoning, Nourse proposes a legal test to distinguish warranted from unwarranted extreme- emotional-disturbance excuses for homicide. If the killer’s emotional dis­turbance is provoked by an act, like rape, which the law condemns, the excuse is warranted; if it is provoked by an act that the law protects, like leaving a relationship, the excuse is unwarranted.

There is one way in which the “warranted excuse” terminology can be misleading. It is important to realize that what makes the excuse unwarranted is not just that the actor’s emotion corresponds with a false evaluative judgment. The excuse is unwarranted because the actor’s emotion corre­sponds with an evil evaluative judgment - one that reflects badly on the actor’s character. The excuse fails not because its underlying evaluative judgment is epistemologically unwarranted; the excuse fails because its underlying evaluation is morally detestable.

Admittedly, it runs deeply against the modern temper to moralize about psychological forces over which we arguably have no control. That is what the warranted-excuse approach does, inasmuch as it relies on moral judgments to distinguish causal explanations for behavior that mitigate lia­bility from causal explanations that do not.

Yet assigning responsibility in a world of causal explanations is what Compatibilism (the approach to the free-will problem that insists that moral responsibility is compatible with determinism) is all about - and the criminal law is Compatibilist through and through.

Viewed abstractly, then, the strategy for separating warranted from unwarranted heat-of-passion excuses amounts to this. First, we make explicit the underlying judgment that the emotion reflects. Second, we ask whether the judgment is warranted. Third, if the judgment underlying the emotion is unwarranted, we ask whether in addition it is morally condemnable. If so, the excuse is unwarranted.

How can we apply these ideas to Milgram obedience? Notice first that the propensity to obey is not an emotion. It is more like a hankering, like wanting to smoke a cigarette or scratch an itch. But even though the urge to obey is not an emotion, we can treat it along the same lines as the heat-of-passion defense: first, by making explicit whatever underlying judgments it corre­sponds with; second, by asking if they are justified; and third, if they are unjustified, by asking whether they are in addition morally condemnable.

What underlying judgments correspond with Milgram obedience? That depends on what the explanation of Milgram obedience is. Here, I will assume that the corruption-of-judgment explanation I defended earlier is the right explanation. Subjects obey, according to the corruption-of-judgment account, because the experiment manipulates them into misjudging the point at which an electric shock violates the no-harm principle. The experiment begins innocuously, and each incremental step implicates the teacher a bit further in the project of shocking the learner. The experimenter’s repeated instruction - “The experiment requires that you continue” - reinforces the idea that every shock level is morally indistinguishable from those that went before. As a result, breaking off the experiment for moral reasons generates cognitive dissonance, because it suggests that the teacher has willingly par­ticipated in wrongdoing. The teacher cannot eliminate the dissonance by undoing what he’s already done. Instead, he eliminates the dissonance by gerrymandering the scope of the no-harm principle so that participating in the experiment does not appear to violate it. As one psychologist puts it, “Dissonance-reducing behavior is ego-defensive behavior; by reducing dis­sonance, we maintain a positive image of ourselves - an image that depicts us as good.”^[477] In other words, our judgment gets corrupted because only by corrupting our judgment can we continue to think well of ourselves. Con­science must be seduced into flattering our self-image.

On this analysis, the propensity to obey corresponds with the following line of (unconscious) reasoning: “If the next shock is wrong, the one I just administered was wrong as well. If so, I would have to believe that I had done something morally wrong; I would have to think badly of myself. That’s unacceptable. So the next shock can’t be wrong.”

That this line of reasoning is unsound goes without saying. It takes one’s own inevitable moral uprightness as a given, and our inevitable moral uprightness is never a given. But the reasoning is more than merely unsound. It reflects badly on our character. It reveals us as so childishly resistant to moral self-criticism that we will distort our sense of right and wrong to avoid admitting that we have done wrong. We are willing to electrocute the learner if the alternative is feeling a little bad about ourselves. Amour propre Uber alles!

The Milgram experiments demonstrate that two-thirds of us are fatally susceptible to this kind of unconscious reasoning, from which it follows that avoiding it must be rather difficult. On the Exculpating View, the difficulty of avoiding it mitigates our moral culpability. But the argument I have been elaborating leads to the opposite conclusion. Compliance originates in cor­ruption of judgment, and corruption of judgment in this case corresponds with the line of reasoning that I have summarized as amour propre Uber alles - a line of reasoning that is not only unsound but morally repugnant. A traditional theologian like St. Augustine would have labeled it, far more simply, as the sin of pride. Our susceptibility to self-corrupted judgment reflects badly on us, and no mitigation is warranted. In this case, at any rate, the Inculpating View seems closer to the truth.

It is important to understand what I am not arguing. I am not arguing that whenever a bad choice arises from fallacious unconscious reasoning that corrupts our judgment we bear full responsibility for making the bad choice. We bear full responsibility only when the unconscious reasoning is not only fallacious but morally reprehensible. Sometimes, fallacious unconscious reasoning casts no discredit on us, and in those cases the difficulty of avoiding it does mitigate our blame.

For example, cognitive psychologists have discovered that when we face risk-decisions we unconsciously employ quick-and-dirty heuristics that in trick cases can lead us to faulty probability-judgments. Presumably, natural selection bred these heuristics into us because they make up in ease and speed what they sacrifice in reliability. They are useful rules of thumb, and Mother Nature is a rule-utilitarian. The principle is the same as in optical illusions: our brain learns quick-and-dirty optical heuristics like “Small is far and big is near,” which can be exploited by illusionists to fool the eye. The rule “Small is far and big is near” is fallacious; but it does not reflect badly on us that we unconsciously follow it. Even if following it leads us to a fatal mistake, we aren’t to blame. In the same way, we aren’t to blame for mistakes arising from our quick-and-dirty cognitive heuristics, because it doesn’t reflect badly on us that we employ them. Finite creatures like us must and should employ them.^[478]

Milgram compliance is different, because the unconscious reasoning compliant teachers follow does reflect badly on them. What follows from these observations is that neither the Inculpating View nor the Exculpating View is entirely right, because each holds sway in some cases but not in others. Suppose psychologists discover that under some experimental con­dition C most people suffer a failure of judgment. The Inculpating View says that the large number of people suffering the failure doesn’t excuse the failure, while the Exculpating View says that it does. What we have dis­covered instead is that when susceptibility to C reflects badly on our char­acter, the Inculpating View is true; when susceptibility to C does not reflect badly on our character, the Exculpating View is true. In Milgram, the Inculpating View is true; compliant subjects are to blame for their wrongful obedience, even though it resulted from bad judgment, and their judgment was corrupted by dynamics they were unaware of. That is because their susceptibility to corruption of judgment, arising from the sin of pride, reflects badly on them.