Contributions to moral philosophy, from Aristotle to Emmanuel Kant to John Stuart Mill and to John Rawls, are among the peak achievements of human intellect. Experiments in Ethics addresses The Question: how does moral philosophy relate to the effort in experimental psychology and behavioral game theory to model the contribution of moral values to individual choice and strategic interaction?
This engaging book does not fully settle the issue (does a philosopher ever fully settle the issue?), but it is packed with mordant insights and suggestive ideas for the behavioral scientist.
I start from the understanding that moral reasoning and moral behavior are an intimate part of the human behavioral repertoire. Unlike mathematics or video games, moral behavior is a part of even the simplest hunter-gatherer societies known to us, and the complexity of moral reasoning appears to be a feature of human society everywhere. Humans produce morality the same way spiders weave webs.
This understanding suggests a very simple answer to The Question: just as physics moved from Natural Philosophy to Natural Science in a previous era, and more recently just as the study of human speech moved from the Philosophy of Language to Linguistics, so now does ethics move from Moral Philosophy to Behavioral Science in the current era. If this were correct, moral philosophy in the current era would be relegated to the position of the interpretation of behavioral science, just as the philosophy of science has become the study of the practice of scientists and the interpretation of their intellectual products.
I very much like this simple answer, because I have always been hostile to the notion, shared by many moral philosophers, that they have some special access to Moral Truth, giving them the right to make value judgments concerning the moral behavior of the untutored man in the street. Just as the linguist records but does not determine the rules of grammar, so the moral philosopher reflects on ethics but does not produce moral truths.
However, this answer to The Question presupposes a rather thoroughgoing moral relativism: the institution of slavery is just in a given society if it is justified by a moral principle in that society. In particular, there is no conceptual position outside of society according to which we can judge the superior morality of one society over another. This is fine for the scientist, who is predisposed to observe without moral judgment, but it overlooks a key point: people treat a statement of the form "x is good" to be a fact about x, much as "x is blue" is a fact about x. Indeed, people can argue whether x is in fact good, and they are not arguing about how x is treated in any particular society.
It does no good to retreat the position that each individual has a personal morality, just as each has a personal genome, only partially shared with others. If we argue about morality, we are not arguing about what each of us believes, but rather what we should jointly believe.
I think this reasoning is decisive. We miss something fundamental about morality by saying merely that humans have moral rules the way they have food preferences. Behavioral scientists study the moral rules of people the same way they study the food preferences of individuals. But, there is more to morality than that. And that is why the simple answer to The Question is wrong. As Appiah puts it, "Does anybody really think a moral system is just merely because it has be accepted by most members of a society? "(150) The only reasonable answer is No.
An alternative answer is that moral philosophers specify criteria of goodness, justice, and moral worth. When people disagree with the moral values singled out by the philosopher, people are simply wrong. This is the typical stance of the various types of utilitarianism (Bentham, Sidgwick, Mill), as well as the deontological moralists, such as Kant. Since the criteria need not correspond to what people actually value morally, these schools have an ultimately intuitionist epistemology, although I imagine that it is possible to derive the guiding moral criteria from biological evolution or some other empirical regularity.
Appiah is dismissive of such theories. He says, "Astronomers have stars; geologists have rocks. But what do moral theorists have to work with? For centuries, they typically claimed to proceed from truths that were self-evident---or, more modestly, from our moral intuitions." (p. 73). However, he notes that "Where surveys show people have incompatible intuitions, we should predict that each intuition has some past philosopher who constructed a theory around it. "(191) His point is that the variety of intuitions of the philosophers has about the same breadth, if not depth, of that of hoi poloi.
For Appiah, this does not mean that the moral philosopher should not appeal to his intuitions. He quotes Sir David Ross, in The Right and the Good (1930), who proposed that "the moral convictions of thoughtful and well-educated people are the data of ethics just as sense-perceptions are the data of a natural science." (p. 75) The point is that the moral philosopher has the same data available as the ordinary citizen, but has spent more time and energy grappling with and trying to make sense out of this data. The evidence of the experimental psychologist and behavioral game theorist are thus welcomed by the moral philosopher, who can use this data to sort out moral issues more effectively.
Appiah thus accepts a version of virtue ethics. He says, "Ethics is, in that formulation of Aristotle's, about the ultimate aim or end of human life, the end he called eudemonia." (164) We can translate eudemonia as "flourishing," the term used by Amartya Sen in his work on ethics and economic theory. It is clear why Appiah is so happy with behavioral game theory and experimental psychology: to decide what is best for humans requires that we know a lot about humans. "The end of philosophical ethics," he says, "is to make sense of the project of eudaimonia. It cannot do that on its own: it needs the assistance of all the moral sciences (204)
Appiah explains virtue ethics as follows: "The right thing to do is what a virtuous agent would do in the circumstances...A virtue is a character trait that a person needs in order to have eudaimonia---that is, in order to live a good life. (36) He recognizes that the study of real-life humans indicates that people are at best only inconsistently virtuous, often lapsing, and typically behaving virtuously only over certain idiosyncratic spheres of life. That's okay, he says. Eudaimonia is an ideal state, not one that can be achieved. He remarks that "Virtue ethics assumes that being virtuous is a character trait. In fact, people are virtuous in some situations and not in other situations. There is thus no such thing as a "virtuous disposition." We are just not built that way." (46) However, "individual moments of compassion and moments of honesty make our lives better, even if we are not compassionate or honest through and through." (70)
My own analogy to Appiah's view, which I find attractive, is that the moral philosopher is like the master chef, taking everyday ingredients at their freshest and producing a culinary delight that keeps the customer coming back for more, or perhaps buying and using the master's recipe book. One positive attribute of this analogy is that food preferences, like ethical preferences, have a unity in diversity. We can appreciate the culinary arts of many cultures, but there are some foods that are delicacies in one society that are an abomination in another. So it goes with moral values, which are universal, but have their jarring incompatibilities.
Virtue ethics has something deep to teach the behavioral scientist. For many decades, economists and biologists assumed that the natural state of a human being was selfishness, a characteristic relaxed only on behalf of one's immediate kith and kin. Indeed, biologists extended "selfish genes" to selfish individuals, and economists equated "rationality" with being "self-interested." The contrast with Appiah and other virtue theorists is quite striking. Virtue theorists believe that it may be difficult to determine what the virtuous act is, and it may be tempting to opt for an immediately rewarding act lacking virtue, but virtue is ultimately its own reward. Appiah says, "The claim that we ought to do what's in everyone's long-term interest...is a necessary truth underlies morality." (24)
Many experimental studies in recent years have shown that people often prefer to be virtuous, and are willing to sacrifice material reward to achieve moral ends. Appiah's claim, however, is far stronger. Not only do many people strive to be virtuous, but those who routinely perform virtuous acts have more satisfied lives than those who do not. For instance, James Konow and Joseph Earley, "The Hedonistic Paradox", Journal of Public Economics 92 (2008) note that people who volunteer for community helping or are generous givers in laboratory altruism games are happier, have higher self-esteem, life-satisfaction and even physical health than those who do not.
Appiah's analysis is flawed by his unwillingness to consider analytical models of human behavior. He is happy with experiments, but he never mentions that these experiments may indicate a certain analytical framework that radically illuminates human choice. I refer here to the rational actor model (which I prefer to call the beliefs, preferences and constraints (BPC) model) for understanding the notion of objectives and constraints, and game theory, for understanding strategic interaction. In the BPC, moral values are part of the individual's objective function, so there are quite naturally tradeoffs between material self-interest and moral values, and among competing moral values. It is incorrect to say that individuals are "imperfectly" virtuous, because it is the nature of tradeoffs that there is no unique first-best point, given our constrained circumstances. I am sure that we are, each of us, imperfectly virtuous because of our human failings, but the very framework of virtue ethics, according to which the virtuousness of an act is assessed independently from any other act or concrete circumstances, is just wrong. Appiah would have a much cleaner argument if he took the beliefs preferences and constraints model seriously, and if he analyzed social interaction in game-theoretic terms.