her suffering, they won’t bother if they are free to leave. Only if they have to endure the sight and sound of her moaning will they prefer to get shocked themselves. As in Krebs’s experiment, the participant’s sympathy was manipulated by telling her either that she and Elaine had the same values and interests, or that they had incompatible ones (for example, if the participant read Newsweek, Elaine would be described as reading Cosmo and Seventeen). Sure enough, when participants felt themselves to be similar to Elaine, they relieved her of being shocked, whether or not they had to watch her suffer. If they felt themselves to be different, they took her place only when the alternative was to watch the suffering. Together with other studies, the experiment suggests that by default people help others egoistically, to relieve their own distress at having to watch them suffer. But when they sympathize with a victim, they are overcome by a motive to reduce her suffering whether it eases their distress or not.
Another set of experiments tested a second ulterior motive to helping, namely the desire to be seen as doing the socially acceptable thing.54 This time, rather than manipulating sympathy experimentally, Batson and his collaborators exploited the fact that people spontaneously vary in how sympathetic they feel. After the participants heard Elaine worrying aloud about the impending shocks, they were asked to indicate the degree to which they felt sympathetic, moved, compassionate, tender, warm, and soft-hearted. Some participants wrote high numbers next to these adjectives; others wrote low ones.
Once the procedure began, and long-suffering Elaine started getting zapped and was visibly unhappy about it, the experimenters used sneaky ways of assessing whether any desire on the part of the participants to relieve her distress sprang from pure beneficence or a desire to look good. One study tapped the participants’ mood with a questionnaire, and then either gave them the opportunity to relieve Elaine by doing well on a task of their own, or simply dismissed Elaine without the participant being able to claim any credit. The empathizers felt equally relieved in both cases; the nonempathizers only if they were the ones that set her free. In another, the participants had to qualify for an opportunity to take Elaine’s place by scoring well in a letterfinding task they had been led to believe was either easy (so there was no way to fake a bad performance and get off the hook) or hard (so they could take a dive and plausibly get out of being asked to make the sacrifice). The nonempathizers took the dive and did worse in the so-called hard task; the empathizers did even better on the hard task, where they knew an extra effort would be needed to allow them to suffer in Elaine’s stead. The emotion of sympathy, then, can lead to genuine moral concern in Kant’s sense of treating a person as an end and not a means to an end—in this case, not even as a means to the end of feeling good about having helped the person.
In these experiments, a person was rescued from a harm caused by someone else, the experimenter. Does sympathy-induced altruism dampen one’s own tendency to exploit someone, or to retaliate in response to a provocation? It does. In other experiments, Batson had women play a one-shot Prisoner’s Dilemma in which they and a (fictitious) fellow participant bid cards that could net them various numbers of raffle tickets, framed as a business transaction.55 Most of the time they did what game theorists say is the optimal strategy: they defected. They chose to bid a card that protected them against being a sucker and that offered them the chance to exploit their partner, while leaving them with a worse outcome than if the two of them had cooperated by bidding a different card. But when the participant read a personal note from her otherwise anonymous partner and was induced to feel empathy for her, her rate of cooperating jumped from 20 percent to 70 percent. In a second experiment, a new group of women played an Iterated Prisoner’s Dilemma game, which gave them an opportunity to retaliate against a partner’s defection with a defection of their own. They cooperated in response to a defection only 5 percent of the time. But when they were induced beforehand to empathize with their partner, they were far more forgiving, and cooperated 45 percent of the time.56 Sympathy, then, can mitigate self-defeating exploitation and costly