Cognitive science’s search for a common morality
Consider the following dilemma: Mike is supposed to be the best man at a friend’s wedding in Maine this afternoon. He is carrying the wedding rings with him in New Hampshire, where he has been staying on business. One bus a day goes directly to the coast. Mike is on his way to the bus station with 15 minutes to spare when he realizes that his wallet has been stolen, and with it his bus tickets, his credit cards, and all his forms of ID.
At the bus station Mike tries to persuade the officials, and then a couple of fellow travelers, to lend him the money to buy a new ticket, but no one will do it. He’s a stranger, and it’s a significant sum. With five minutes to go before the bus’s departure, he is sitting on a bench trying desperately to think of a plan. Just then, a well-dressed man gets up for a walk, leaving his jacket, with a bus ticket to Maine in the pocket, lying unattended on the bench. In a flash, Mike realizes that the only way he will make it to the wedding on time is if he takes that ticket. The man is clearly well off and could easily buy himself another one.
Should Mike take the ticket?
My own judgment comes down narrowly, but firmly, against stealing the ticket. And in studies of moral reasoning, the majority of American adults and children answer as I do: Mike should not take the ticket, even if it means missing the wedding. But this proportion varies dramatically across cultures. In Mysore, a city in the south of India, 85 percent of adults and 98 percent of children say Mike should steal the ticket and go to the wedding. Americans, and I, justify our choice in terms of justice and fairness: it is not right for me to harm this stranger—even in a minor way. We could not live in a world in which everyone stole whatever he or she needed. The Indian subjects focus instead on the importance of personal relationships and contractual obligations, and on the relatively small harm that will be done to the stranger in contrast to the much broader harm that will be done to the wedding.
An elder in a Maisin village in Papua New Guinea sees the situation from a third perspective, focused on collective responsibility. He rejects the dilemma: “If nobody [in the community] helped him and so he [stole], I would say we had caused that problem.”
Examples of cross-cultural moral diversity such as this one may not seem surprising in the 21st century. In a world of religious wars, genocide, and terrorism, no one is naive enough to think that all moral beliefs are universal. But beneath such diversity, can we discern a common core—a distinct, universal, maybe even innate “moral sense” in our human nature?
In the early 1990s, when James Q. Wilson first published The Moral Sense, his critics and admirers alike agreed that the idea was an unfashionable one in moral psychology. Wilson, a professor of government and not psychology, was motivated by the problem of non-crime: how and why most of us, most of the time, restrain our basic appetites for food, status, and sex within legal limits, and expect others to do the same. The answer, Wilson proposed, lies in our universal “moralsense, one that emerges as naturally as [a] sense of beauty or ritual (with which morality has much in common) and that will affect [our] behavior, though not always, and in some cases not obviously.”
But the fashion in moral psychology is changing.
A decade after Wilson’s book was published, the psychological and neural basis of moral reasoning is a rapidly expanding topic of investigation within cognitive science. In the intervening years, new technologies have been invented, and new techniques developed, to probe ever deeper into the structure of human thought. We can now acquire vast numbers of subjects over the Internet, study previously inaccessible populations such as preverbal infants, and, using brain imaging, observe and measure brain activity non-invasively in large numbers of perfectly healthy adults. Inevitably, enthusiasts make sweeping claims about these new technologies and the old mysteries they will leave in their wake. (“The brain does not lie” is a common but odd marketing claim, since in an obvious sense, brains are the only things that ever do.)
The appeal of the new methods is clear: if an aspect of reasoning is genuinely universal, part of the human genetic endowment, then such reasoning might be manifest in massive cross-cultural samples, in subjects not yet exposed to any culture, such as very young infants, and perhaps even in the biological structure of our reasoning organ, the brain.
How far have these technologies come in teaching us new truths about our moral selves? How far could they go? And what will be the implications of a new biopsychological science of natural morality? “The truth, if it exists, is in the details,” wrote Wilson, and therefore I will concentrate on the details of three sets of very recent experiments, each of which approaches the problem using a different method: an Internet survey, a cognitive study of infants, and a study of brain imaging. Each is at the cutting edge of moral psychology, each is promising but flawed, and each should be greeted with a mix of enthusiasm and interpretative caution.
* * *
Mike, the man we left sitting at the bus station, is in a particularly bad moral predicament: he must choose between two actions (stealing and breaking an obligation), both of which are wrong. Moral psychologists call cases like these “moral dilemmas.” Over the last half century, batteries of moral dilemmas have been presented to men and women, adults and children, all over the world. The questions at the heart of these studies are these: How do people arrive at the moral judgment that an action, real or contemplated, is right or wrong? What are the rules governing these moral calculations, and from where do they come? Which, if any, of the fundamental components are universal?
All of them, answered the eminent psychologist Lawrence Kohlberg. In the 1970s and 1980s, Kohlberg argued that moral reasoning is based on explicit rules and concepts, like conscious logical problem-solving; over the course of an individual’s development, the rules and concepts that he or she uses to solve moral problems unfold in a well-defined, universal sequence of stages. These stages are biologically determined but socially supported. In early stages, moral reasoning is strongly influenced by external authority; in later stages, moral reasoning appeals first to internalized convention, and then to general principles of neutrality, egalitarianism, and universal rights. It may be that what makes one culture, one sex, or one individual different from another is just how high and how fast it manages to climb the moral ladder.
To test this hypothesis, moral dilemmas were presented to people of varying ages and classes, both sexes, and many cultures (including people in India, Thailand, Iran, Turkey, Kenya, Nigeria, and Guatemala; communities of Alaskan Inuit; Tibetan Buddhist monks; and residents of an Israeli kibbutz). Kohlberg’s key methodological insight was to focus not on the answers that people give to moral dilemmas but on how they justify their choice. A seven-year-old and a white-haired philosopher may agree that Mike should not steal the ticket, but they will differ in their explanations of why not. The seven-year-old may say that Mike shouldn’t steal because he will get caught and punished, while the philosopher may appeal to an interpretation of Kant’s categorical imperative: act only on a principle that you would wish everyone to follow in a similar situation.
Kohlberg’s claims were deeply controversial, not least because the highest stage of moral development was accorded almost exclusively to Western adults, and among those, mostly to men. Critics attacked everything from the specific dilemmas to the coding criteria to the whole philosophy of monotonic universal moral development. The psychologist Carol Gilligan, for example, argued that women justify their moral choices differently from men, but with equal sophistication. Men, she claimed, tend to reason about morality in terms of justice, and women in terms of care: “While an ethic of justice proceeds from the premise of equality—that everyone should be treated the same—an ethic of care rests on the premise of non-violence—that no one should be hurt.” Similar arguments were made for non-Western cultures—that they emphasize social roles and obligations rather than individual rights and justice. On the whole, this emphasis on group differences won the day. Kohlberg’s vision was rejected, and the psychological study of moral universals reached an impasse.
Very recently, though, the use of moral dilemmas to study moral universals has reemerged. Marc Hauser of Harvard University and John Mikhail of Georgetown University are among the cognitive scientists leading the charge. The current theorists take as their model for moral reasoning not conscious problem-solving, as Kohlberg did, but the human language faculty. That is, rather than “moral reasoning,” human beings are understood to be endowed with a “moral instinct” that enables them to categorize and judge actions as right or wrong the way native speakers intuitively recognize sentences as grammatical or ungrammatical.
We can draw three predictions from the theory that morality operates as language does. First, just as each speaker can produce and understand an infinite number of completely original sentences, every moral reasoner can make fluent, confident, and compelling moral judgments about an infinite number of unique cases, including ones that they have never imagined confronting. Second, cross-culturally, systems of moral reasoning can be as diverse as human languages are, without precluding that a universal system of rules, derived from our biological inheritance, underlies and governs all these surface-level differences. Finally, just as native speakers are often unable to articulate the rules of grammar that they obey when speaking, the practitioners of moral judgment may have great difficulty articulating the principles that inform their judgments. Hauser, Mikhail, and their colleagues have tested these predictions with a set of moral dilemmas originally introduced by the philosopher Phillipa Foot in 1967 and now known collectively as the Trolley Problems. To illustrate the category, let’s begin with Anna, standing on the embankment above a train track, watching a track-maintenance team do its work. Suddenly, Anna hears the sound of a train barrelling down the tracks: the brakes have failed, and the train is heading straight for the six workers. Beside Anna is a lever; if she pulls it, the train will be forced onto a side track and will glide to a halt without killing anyone. Should she pull the lever?
No moral dilemma yet. But now let’s complicate the story. In the second scenario, Bob finds himself in the same situation, except that one of the six maintenance people is working on the side track. Now the decision Bob faces is whether to pull the lever to save five lives, knowing that if he does, a man who would otherwise have lived will be killed.
In a third version of what is clearly a potentially infinite series, the sixth worker is standing beside Camilla on the embankment. The only way to stop the train, and save the lives of the five people on the track, is for Camilla to push the man beside her down onto the track. By pushing him in front of the train and so killing him, she would slow it down enough to save the others.
Finally, for anyone not yet convinced that there are cases in which it is wrong to sacrifice one person in order to save five, consider Dr. Dina, a surgeon who has five patients each dying from the failure of a different organ. Should she kill one healthy hospital visitor and distribute the organs to her patients in order to save five lives?
By putting scenarios like these on a Web site (http://moral.wjh.harvard.edu) and soliciting widely for participants, Hauser and his lab have collected judgments about Trolley Problems from thousands of people in more than a hundred countries, representing a broad range of ages and religious and educational backgrounds. The results reveal an impressive consensus. For example, 89 percent of subjects agree that it is permissible for Bob to pull the lever to save five lives at the cost of one but that it is not permissible for Camilla to make the same tradeoff by pushing the man onto the track.
More importantly, even in this enormous sample and even for complicated borderline cases, participants’ responses could not be predicted by their age, sex, religion, or educational background. Women’s choices in the scenarios overall were indistinguishable from men’s, Jews’ from Muslims’ or Catholics’, teenagers’ from their parents’ or grandparents’. Consistent with the analogy to language, these thousands of people make reliable and confident moral judgments for a whole series of (presumably) novel scenarios. Also interestingly, Hauser, Mikhail, and their colleagues found that while the “moral instinct” was apparently universal, people’s subsequent justifications were not; instead, they were highly variable and often confused. Less than one in three participants could come up with a justification for the moral difference between Camilla’s choice and Bob’s, even though almost everyone shares the intuition that the two cases are different.
So what can we learn from this study? Has the Internet—this new technology—given us a way to reveal the human universals in moral judgments?
We must be cautious: Web-based experiments have some obvious weaknesses. While the participants may come from many countries and many backgrounds, they all have Internet access and computer skills, and therefore probably have significant exposure to Western culture. (In fact, although the first study included just over 6,000 people from more than a hundred countries, more than two thirds of them were from the United States.) Because the survey is voluntary, it includes a disproportionate number of people with a preexisting interest in moral reasoning. (More than two thirds had previously studied moral cognition or moral philosophy in some academic context, making it all the more surprising that they could not give clear verbal justifications of their intuitions.) And because subjects fill out the survey without supervision or compensation, sincerity and good faith cannot be ensured (although Hauser, Mikhail, and their colleagues did exclude the subjects who claimed to live in Antarctica or to have received a Ph.D. at 15).
Also, this is only one study, focused on only one kind of moral dilemma: the Trolley Problems. So far, we don’t know whether the universality of intuitions observed in this study would generalize to other kinds of dilemmas. The results of the experiment with Mike and the bus ticket suggest it probably would not.
On the other hand, the survey participants did include a fairly even balance of sexes and ages. And the fact that sex in particular makes no difference to people’s choices in the Trolley Problems, even in a sample of thousands (and growing), could be important. Remember, Carol Gilligan charged that Lawrence Kohlberg’s theory of multi-stage moral development was biased toward men; she claimed that men and women reason about moral dilemmas with equal sophistication, but according to different principles. Hauser and Mikhail’s Internet study lets us look at the controversy from a new angle. Gilligan’s analysis was based on justifications: how men and women consciously reflect upon, explain, and justify the moral choices that they make. It is easy to imagine that the way we justify our choices depends a lot on the surrounding culture, on external influences and expectations. What Hauser and Mikhail’s results suggest is that though the reflective, verbal aspects of moral reasoning (which Hauser and Mikhail found inarticulate and confused, in any case) may differ by sex, the moral intuition that tells us which choice is right and which wrong for Anna or Bob or Camilla is part of human nature, for women just as for men.
Still, the Internet’s critical weakness is intransigent. As long as people must have Internet access in order to participate, the sample will remain culturally biased, and it will be hard to know for sure from where the moral consensus comes: from human nature or from exposure to Western values. The only way to solve this problem is to investigate moral reasoning in people with little or no exposure to Western values. And cognitive scientists are beginning to do just that.
* * *
One group of experimental participants that is relatively free of cultural taint is preverbal infants. Before they are a year old, while their vocabulary consists of only a few simple concrete nouns, infants have presumably not yet been acculturated into the specific moral theories of their adult caretakers. Infant studies therefore offer scientists the chance to measure innate moral principles in mint condition. With this opportunity, of course, comes a methodological challenge. How can we measure complex, abstract moral judgments made by infants who are just beginning to talk, point, and crawl?
To meet this challenge, developmental psychologists who study all areas of cognition have become adept—often ingenious—at teasing meaning out of one of the few behaviors that infants can do well: looking. Infants look longer at the things that interest them: objects or events that are attractive, unexpected, or new. Looking-time experiments therefore gauge which of two choices—two objects, people, or movies—infants prefer to watch.
From just this simple tool, a surprisingly rich picture of infant cognition has emerged. We have learned, for example, that infants only a few days old prefer to look at a human face than at other objects; that by the time they are four months old, infants know that one object cannot pass through the space occupied by another object; and that by seven months, they know that a billiard ball will move if and only if it is hit by something else.
Only recently, though, has this tool begun to be applied to the field of moral cognition. The questions these new studies seek to answer include the following: Where do we human beings get the notions of “right,” “wrong,” “permissible,” “obligatory,” and “forbidden”? What does it mean when we judge actions—our own or others’—in these terms? How and why do we judge some actions wrong (or forbidden) and not just silly, unfortunate, or unconventional?
Not all transgressions are created equal; some undesirable or inappropriate actions merely violate conventions, while others are genuinely morally wrong. Rainy weather can be undesirable, some amateur acting is very bad, and raising your hand before speaking at a romantic candlelit dinner is usually inappropriate, but none of these is morally wrong or forbidden. Even a tsunami or childhood cancer, though both awful, are not immoral unless we consider them the actions of an intentional agent.
The psychologist Elliott Turiel has proposed that the moral rules a person espouses have a special psychological status that distinguishes them from other rules—like local conventions—that guide behavior. One of the clearest indicators of this so-called moral–conventional distinction is the role of local authority.
We understand that the rules of etiquette—whether it is permissible to leave food on your plate, to belch at the table, or to speak without first raising your hand—are subject to context, convention, and authority. If a friend told you before your first dinner at her parents’ house that in her family, belching at the table after dinner is a gesture of appreciation and gratitude, you would not think your friend’s father was immoral or wrong or even rude when he leaned back after dinner and belched—whether or not you could bring yourself to join in.
Moral judgments, in contrast, are conceived (by hypothesis) as not subject to the control of local authority. If your friend told you that in her family a man beating his wife after dinner is a gesture of appreciation and gratitude, your assessment of that act would presumably not be swayed. Even three-year-old children already distinguish between moral and conventional transgressions. They allow that if the teacher said so, it might be okay to talk during nap, or to stand up during snack time, or to wear pajamas to school. But they also assert that a teacher couldn’t make it okay to pull another child’s hair or to steal her backpack. Similarly, children growing up in deeply religious Mennonite communities distinguish between rules that apply because they are written in the Bible (e.g., that Sunday is the day of Sabbath, or that a man must uncover his head to pray) and rules that would still apply even if they weren’t actually written in the Bible (including rules against personal and material harm).
There is one exception, though. James Blair, of the National Institutes of Health, has found that children classified as psychopaths (partly because they exhibit persistent aggressive behavior toward others) do not make the normal moral–conventional distinction. These children know which behaviors are not allowed at school, and they can even rate the relative seriousness of different offences; but they fail when asked which offences would still be wrong to commit even if the teacher suspended the rules. For children with psychopathic tendencies (and for psychopathic adults, too, though not for those Blair calls “normal murderers”), rules are all a matter of local authority. In its absence, anything is permissible.
Turiel’s thesis, then, is that healthy individuals in all cultures respect the distinction between conventional violations, which depend on local authorities, and moral violations, which do not.
This thesis remains intensely controversial. The chief voice of opposition may come not from psychologists but from anthropologists, who argue that the special status of moral rules cannot be part of human nature, but is rather just a historically and culturally specific conception, an artifact of Western values. “When I first began to do fieldwork among the Shona-speaking Manyika of Zimbabwe,” writes Anita Jacobson-Widding, for example, “I tried to find a word that would correspond to the English concept ‘morality.’ I explained what I meant by asking my informants to describe the norms for good behavior toward other people. The answer was unanimous. The word for this was tsika. But when I asked my bilingual informants to translate tsika into English, they said that it was ‘good manners.’ And whenever I asked somebody to define tsika they would say ‘Tsika is the proper way to greet people.’”
Jacobson-Widding argues that the Manyika do not separate moral behavior from good manners. Lying, farting, and stealing are all equally violations of tsika. And if manners and morals cannot be differentiated, the whole study of moral universals is in trouble, because how—as Jacobson-Widding herself asks—can we study the similarities and differences in moral reasoning across cultures “when the concept of morality does not exist?” From the perspective of cognitive science, this dispute over the origins of the moral–conventional distinction is an empirical question, and one that might be resolvable with the new techniques of infant developmental psychology.
One possibility is that children first distinguish “wrong” actions in their third year of life, as they begin to recognize the thoughts, feelings, and desires of other people. If this is true, the special status of moral reasoning would be tied to another special domain in human cognition: theory of mind, or our ability to make rich and specific inferences about the contents of other people’s thoughts. Although this link is plausible, there is some evidence that distinguishing moral right from wrong is a more primitive part of cognition than theory of mind, and can exist independently. Unlike psychopathic children, who have impaired moral reasoning in the presence of intact theory of mind, autistic children who struggle to infer other people’s thoughts are nevertheless able to make the normal moral–conventional distinction.
Another hypothesis is that children acquire the notion of “wrong” actions in their second year, once they are old enough to hurt others and experience firsthand the distress of the victim. Blair, for example, has proposed that human beings and social species like canines have developed a hard-wired “violence-inhibition mechanism” to restrain aggression against members of the same species. This mechanism is activated by a victim’s signals of distress and submission (like a dog rolling over onto its back) and produces a withdrawal response. When this mechanism is activated in an attacker, withdrawal means that the violence stops. The class of “wrong” actions, those that cause the victim’s distress, might be learned first for one’s own actions and then extended derivatively to others’ actions.
Both of these hypotheses suggest a very early onset for the moral–conventional distinction. But possibly the strongest evidence against the anthropologists’ claim that this distinction is just a cultural construct would come from studies of even younger children: preverbal infants. To this end, developmental psychologists are currently using the new looking-time procedures to investigate this provocative third hypothesis: that before they can either walk or talk, young infants may already distinguish between hurting (morally wrong) and helping (morally right).
In one study, conducted by Valerie Kuhlmeier and her colleagues at Yale, infants watched a little animated ball apparently struggling to climb a steep hill. A triangle and a square stood nearby. When the ball got just beyond halfway up, one of two things happened: either the triangle came over and gave the ball a helpful nudge up the hill, or the square came over and pushed the ball back down the hill. Then the cycle repeated. Later, the same infants saw a new scene: across flat ground the little ball went to sit beside either the triangle or the square. Twelve-month-old infants tended to look longer when the ball went to sit beside the “mean” shape. Perhaps they found the ball’s choice surprising. Would you choose to hang out with someone who had pushed you down a hill?
Another study, by Emmanuel Dupoux and his colleagues in France, used movies of live human actors. In one, the “nice” man pushes a backpack off a stool and helps a crying girl get up onto the stool, comforting her. In the second movie, the “mean” man pushes the girl off the stool, and picks up and consoles the backpack. The experiment is designed so that the amounts of crying, pushing, and comforting in the two movies are roughly equal. After the movies, the infants are given a choice to look at, or crawl to, either the “mean” man or the “nice” one. At 15 months, infants look more at the mean man but crawl more to the nice one.
These results are interesting, but each of these studies provides evidence for a fairly weak claim: by the time they are one year old, babies can distinguish between helpful actions and hurtful ones. That is, infants seem to be sensitive to a difference between actions that are nice, right, fortunate, or appropriate and ones that are mean, wrong, undesirable, or inappropriate—even for novel actions executed by unknown agents. On any interpretation, this is an impressive discovery. But the difference that infants detect need not be a moral difference.
These first infant studies of morality cannot answer the critical question, which is not about the origin of the distinction between nice and mean, but between right and wrong; that is, the idea that some conduct is unacceptable, whatever the local authorities say. Eventually, infant studies may provide evidence that the concepts of morality and convention can be distinguished, even among the Manyika—that is, that a special concept of morality is part of the way infants interpret the world, even when they are too young to be influenced by culture-specific constructions. So far, though, these infant studies are a long way off.
In the meantime, we will have to turn to other methods, traditional and modern, to adjudicate the debate between psychologists and anthropologists over the existence of moral universals. First, if Hauser and Mikhail’s Internet-survey results really do generalize to a wider population, as the scientists hope, then we might predict that Manyika men and women would give the same answers that everyone else does to the Trolley Problems. If so, would that challenge our notions of how different from us they really are?
Second, if Elliott Turiel and his colleagues are right, then even Manyika children should distinguish between manners, which depend on local custom, and morals, which do not, when asked the right kinds of questions. For example, according to Manyika custom, “If you are a man greeting a woman, you should sit on a bench, keep your back straight and your neck stiff, while clapping your own flat hands in a steady rhythm.” What if we told a four-year-old Manyika child about another place, very far away, where both men and women are supposed to sit on the ground when greeting each other? Or another place where one man is supposed to steal another man’s yams? Would the children accept the first “other world” but not the second? I have never met a Manyika four-year-old, so I cannot guess, but if so, then we would have evidence that the Manyika do have a moral–conventional distinction after all, at the level of moral judgment, if not at the level of moral justification.
Finally, some modern cognitive scientists might reply, we scientists hold a trump card: we can now study moral reasoning in the brain.
* * *
In the last ten years, brain imaging (mostly functional magnetic resonance imaging, or fMRI) has probably exceeded all the other techniques in psychology combined in terms of growth rate, public visibility, and financial expense. The popularity of brain imaging is easy to understand: by studying the responses of live human brains, scientists seem to have a direct window into the operations of the mind.
A basic MRI provides an amazingly fine-grained three-dimensional picture of the anatomy of soft tissues such as the gray and white matter (cell bodies and axons) of the brain, which are entirely invisible to x-rays. An fMRI also gives the blood’s oxygen content in each brain region, an indication of recent metabolic activity in the cells and therefore an indirect measure of recent cell firing. The images produced by fMRI analyses show the brain regions in which the blood’s oxygen content was significantly higher while the subject performed one task—a moral-judgment task, for example—than while the subject performed a different task—a non-moral-judgment task.
Jorge Moll and his colleagues, for example, compared the blood-oxygen levels in the brain while subjects read different kinds of sentences: sentences describing moral violations (“They hung an innocent”), sentences describing unpleasant but not immoral actions (“He licked the dirty toilet”), and neutral sentences (“Stones are made of water”). They found that one brain region—the medial orbito-frontal cortex, the region just behind the space between the eyebrows—had a higher oxygenation level while subjects read the moral sentences than either of the other two kinds of sentences. Moll proposed that the medial orbito-frontal cortex must play some unique role in moral reasoning.
In fact, this is not a new idea. In 1848 Phineas Gage was the well-liked foreman of a railroad-construction gang until a dynamite accident destroyed his medial orbito-frontal cortex (along with a few neighboring brain regions). Although Gage survived the accident with his speech, motion, and even his intelligence unimpaired, he was, according to family and friends, “no longer Gage”: obstinate, irresponsible, and capricious, he was unable to keep his job, and later he spent seven years as an exhibit in a traveling circus. Modern patients with similar brain damage show the same kinds of deficits: they are obscene, irreverent, and uninhibited, and they show disastrous judgment, both personally and professionally.
Still, the claim of a moral brain region remains controversial among cognitive scientists, who disagree both about whether such a brain region exists and what the implications would be if it did. Joshua Greene of Princeton University, for example, investigates brain activity while subjects solve Trolley Problems. He finds lots of different brain regions recruited—as one might imagine—including regions associated with reading and understanding stories, logical problem-solving, and emotional responsiveness. What Greene doesn’t find is any clear evidence of a “special” region for moral reasoning per se.
More broadly, even if there were a specialized brain region that honored the moral–conventional distinction, what would this teach us about that distinction’s source, or universality? Many people share the intuition that the existence of a specialized brain region would provide prima facie evidence of the biological reality of the moral–conventional distinction. The problem is that even finding a specialized neural region for a particular kind of thought does not tell us how that region got there. We know, for example, that there is a brain region that becomes specially attuned to the letters of the alphabet that a person is able to read, but not of other alphabets; this does not make any one alphabet a human universal. Similarly, if Western minds (the only ones who participate in brain-imaging experiments at the moment) distinguish moral from conventional violations, it is not surprising that Western brains do.
In sum, both enthusiasm and caution are in order. The discovery of a specialized brain region for moral reasoning will not simply resolve the venerable problem of moral universals, as proponents of imaging sometimes seem to claim. On the other hand, not every function a brain performs is assigned a specialized brain region. In visual cortex, there are specialized regions for seeing faces and human bodies, but there is no specialized region for recognizing chairs or shoes, just a general-purpose region for recognizing objects. Some distinctions are more important than others in the brain, whatever their importance in daily life. Cognitive neuroscience can tell us where on this scale the moral–conventional distinction falls.
* * *
One thing these cutting-edge studies certainly cannot tell us is the right answer to a moral dilemma. Cognitive science can offer a descriptive theory of moral reasoning, but not a normative one. That is, by studying infants or brains or people around the world, we may be able to offer an account of how people actually make moral decisions—which concepts are necessary, how different principles are weighed, what contextual factors influence the final decision—but we will not be able to say how people should make moral decisions.
Cognitive scientists may eventually be able to prove that men and women reason about Trolley Problems with equal sophistication, that African infants distinguish moral rules that are independent of local authority from conventions that are not, and even that the infants are using a specialized brain region to do so. What they cannot tell us is whether personal and social obligations should triumph over the prohibition against stealing, whether Mike should steal the ticket, and whether in the end it would be a better world to live in if he did.