Kwame Anthony Appiah
Kwame Anthony Appiah; drawing by Siegfried Woldhek

Kwame Anthony Appiah is a writer and thinker of remarkable range. He began his academic career as an analytic philosopher of language, but soon branched out to become one of the most prominent and respected philosophical voices addressing a wide public on topics of moral and political importance such as race, cosmopolitanism, multiculturalism, codes of honor, and moral psychology. Two years ago he even took on the “Ethicist” column in The New York Times Magazine, and it is easy to become addicted to his incisive answers to the extraordinary variety of real-life moral questions posed by readers.

Appiah’s latest book, As If: Idealization and Ideals, is in part a return to his earlier, more abstract and technical interests. It is derived from his Carus Lectures to the American Philosophical Association and is addressed first of all to a philosophical audience. Yet Appiah writes very clearly, and much of this original and absorbing book will be of interest to general readers.

Its theme and its title pay tribute to the work of Hans Vaihinger (1852–1933), a currently neglected German philosopher whose masterwork, published in 1911, was called The Philosophy of “As If.”1 Vaihinger contended that much of our most fruitful thought about the world, particularly in the sciences, relies on idealizations, or what he called “fictions”—descriptions or laws or theories that are literally false but that provide an easier and more useful way to think about certain subjects than the truth in all its complexity would. We can often learn a great deal by treating a subject as if it conformed to a certain theory, even though we know that this is a simplification. As Vaihinger says, such fictions “provide an instrument for finding our way about more easily in the world.”

One of the clearest examples Vaihinger offers is Adam Smith’s assumption, for purposes of economic theory, that economic agents are motivated exclusively by self-interest—that they are egoists. Smith knew perfectly well that human motivation was much richer than that, as he demonstrated in his book The Theory of Moral Sentiments, a work less widely known than The Wealth of Nations. But as Vaihinger explains:

For the construction of his system of political economy it was essential for Adam Smith to interpret human activity causally. With unerring instinct he realized that the main cause lay in egoism and he formulated his assumption in such a way that all human actions, and particularly those of a business or politico-economical nature, could be looked upon as if their driving force lay in one factor—egoism. Thus all the subsidiary causes and partially conditional factors, such as good will, habit, and so forth, are here neglected. With the aid of this abstract cause Adam Smith succeeded in bringing the whole of political economy into an ordered system.

Vaihinger explored the phenomenon in a wide range of cases, from mathematics, the natural sciences, ethics, law, religion, and philosophy. Appiah’s range is equally wide, but his examples are different; he gives special attention to psychology, ethics, political theory, social thought, and literature. In general he defends the value of idealization, but he is also aware of its intellectual dangers. He emphasizes that it is essential to hold on to the contrasting concept of truth, and to keep in mind both the departures from truth that idealization involves and the specific purposes for which it is useful.

Appiah has packed into this short book an impressive amount of original reflection on a number of topics, so my discussion will have to be selective. He mentions some examples from the natural sciences, but in such abbreviated form that they cannot be understood by readers who are not already familiar with the theories in question.2 I shall discuss some cases where Appiah’s analyses of idealization are more accessible.

The contemporary theory of what is standardly referred to as economic rationality is descended from Adam Smith’s egoistic model of economic behavior; it is based on a much more sophisticated and quantitatively precise but still-idealized model of the psychology of individual choice. The modern discipline of decision theory has permitted a great increase in the exactness of what we can say about this type of human motivation, by introducing quantitative measures of subjective degrees of belief and subjective degrees of preference.

If, for example, on a cloudy day you have to decide whether or not to take an umbrella when you go out, you face four possibilities: (1) rain and umbrella; (2) no rain and umbrella; (3) rain and no umbrella; (4) no rain and no umbrella. Obviously your decision will depend both on your estimate of the likelihood of rain and on how much you mind getting wet, or alternatively how much you mind carrying an umbrella when it isn’t raining, but decision theory makes this more precise. It says your choice is explained by the fact that you assign a probability p between zero and one to the prospect of rain, and (ignoring misty in-between states) a probability of one minus p to the prospect of no rain, and that you assign a desirability, positive or negative, to each of the possibilities (1) to (4). By multiplying the probability and the desirability for each of these outcomes, one can calculate what is called the “expected value” of each of them, and therefore the expected value of taking an umbrella and of not taking an umbrella. The rational choice is to do what has the higher expected value.3


Decision theory applies this kind of calculus to choices among alternatives of any complexity, with any possible assignment of subjective probabilities and desirabilities. With the help of game theory it can be extended to multiperson interactions, as in a market economy. What interests Appiah is that the theory assigns these supposed quantifiable psychological states to individuals only on the basis of an idealization. They are not discovered by asking people to report their subjective probabilities and desirabilities: in general, people do not have introspective access to these numbers. Rather, precise psychological states of this type are assigned by the theory itself, on the basis of something to which people do have access, namely their preferences or rankings (better, worse, indifferent) among alternatives.

This by itself does not imply that the states are fictional: real but unobservable underlying causes can often be inferred from observable effects. The fiction comes from the way the inference proceeds in this case. Given a sufficiently extensive set of preferences (rankings of alternatives) by an individual, it is possible, employing relatively simple laws, to assign to that individual a set of subjective probabilities and desirabilities that would account for those preferences, if the individual were rational in the sense of the theory. But since rationality in the sense of the theory involves such superhuman capacities as immunity to logical error, instantaneous calculation of logical consequences, and assigning equal probability and desirability to all possibilities that are logically equivalent, it is clear that no actual humans are rational in this sense. So if we use the theory of economic rationality to think about the behavior of real human beings, we are treating them as if they were superrational (“Cognitive Angels,” in Appiah’s phrase); we are employing a useful fiction, which allows us to bring human action under quantitative laws.

The fiction is useful only for certain purposes. If it is not to lead us astray, we have to recognize the ways in which it deviates from reality, and to correct for those deviations when they make a difference that matters. This is in fact the concern of the recently developed field of behavioral economics, which tries to identify the consequences of systematic deviations of actual human behavior from the standards of classical economic rationality. (For example, people often fail to count logically equivalent possibilities as equally desirable: an outcome framed as a loss will be counted as less desirable than the same outcome framed as the absence of a gain; an outcome described in terms of the probability of death will be evaluated differently from the same outcome described in terms of the probability of survival.) Appiah’s point is more general: if we try to formulate laws of human psychology, we will inevitably have to ignore a great deal of the messy complexity of actual human life. This is sometimes legitimate, provided that we recognize the idealization and are prepared to restore the complexity when necessary—when, for example, assuming the rationality of every free market would send us off an economic cliff.

Consider next a completely nontechnical type of idealization that is omnipresent in contemporary thought and discourse: racial and sexual categories such as “Negro” and “homosexual.” The thought that someone—oneself or another—is a Negro or a homosexual has great personal, social, and political significance in our society. Yet in light of the actual complexity and variety of people’s biological heredity and erotic dispositions these are very crude concepts; they do not correspond to well-defined properties or categories in the real world. Nevertheless, Appiah says, we may find it indispensable to employ them:

In earlier work of my own, for example, I have argued both that races, strictly speaking, don’t exist, and that it is wrong to discriminate on the basis of a person’s race. This can usually be parsed out in a way that is not strictly inconsistent: What is wrong is discrimination against someone because you believe her to be, say, a Negro even though there are, in fact, strictly speaking, no Negroes. But in responding to discrimination with affirmative action, we find ourselves assigning people to racial categories. We think it justified to treat people as if they had races even when we officially believe that they don’t.

These cases do not start out as idealizations. “Negro” and “homosexual” became important social identities because it was widely believed that they were essential properties possessed by some people and not others, and that they had behavioral, social, and moral consequences. Appiah maintains that when someone who does not share these beliefs goes on using the terms, this is not just the verbal acknowledgment of a misguided but tenacious social illusion; it is an example of fictional thinking. We do not truly distance ourselves from these categories and perhaps should not:


Identities, conceived of as stable features of a social ontology grounded in natural facts, are often…assumed in our moral thinking, even though, in our theoretical hearts, we know them not to be real. They are one of our most potent idealizations.

This invites the question: When are these idealizations indispensable, and when on the contrary should we resist them, by appealing to the more complex truth? Appiah addresses this and related questions with great insight in an earlier book, The Ethics of Identity,4 but not here.


Appiah considers another type of idealization that he calls “counter-normative”: thinking or acting as if a moral principle is true although we know it isn’t. He believes we do this when we treat certain prohibitions—against murder or torture, for example—as moral absolutes. His view is that strictly, there are exceptions to any such rule, but it may be better to treat it as exceptionless. In that way we will be sure to avoid unjustified violations, without countervailing risk, since “it is remarkably unlikely that I will ever be in one of those situations where it might be that murder was permissible (and even less likely that I will ever be in one where it is required).” Appiah adds that sometimes the advantage of the fiction will depend on its acceptance not by an individual but by a community. Perhaps the strict rule against making false promises would be an example, since even if it is not universally obeyed, the general belief that it is generally accepted encourages people to trust one another.

Which moral rules one regards as fictions or idealizations will depend on what one believes to be the basis of moral truth. Appiah does not take up this large topic, but his discussion seems most consistent with the view that the ultimate standard of right and wrong is what will produce the best overall outcomes. Counternormative fictions then become useful if we will not achieve the best overall outcomes by aiming in each case at the best overall outcome: it is better to put murder and torture entirely off the table. This is an area of perennial controversy, but those who think the prohibitions on murder, torture, and false promises have a different source, dependent on the intrinsic character of those acts rather than overall outcomes, may be less prone than Appiah to attribute their strictness to idealization.

Appiah concludes with a topic of great philosophical interest, that of idealization in moral theory itself. There is some possibility of confusion here, because he is talking about idealization in a sense somewhat different from that discussed so far.

Every morality is an ideal; it enjoins us to conform to standards of conduct and character that we are often tempted to violate, and it is predictable that ordinary human beings will sometimes fail to conform, even if they accept the morality as correct. This by itself does not involve idealization in Appiah’s sense. The moral principles need depend on no assumptions that are not strictly true. A morality describes not how people do behave but how they should behave; and it has to assume only that they could behave in that way, even if at the moment many of them do not.

The idealization that interests Appiah occurs when political thinkers or philosophers theorize about morality. In developing their accounts, they will often imagine situations or possibilities that differ from what is true in the actual world, as an aid to evaluating moral or political hypotheses. One type of idealization consists in evaluating a moral or political principle by considering what things would be like if everyone complied with it. But as Appiah points out, this is far from decisive:

Consider a familiar kind of dispute. One philosopher—let us call her Dr. Welfare—proposes that we should act in a way that maximizes human well-being. What could be more evident than that this would make for the best world? Another—Prof. Partiality—proposes instead that we should avoid harm to others in general but focus our benevolence on those to whom we have special ties. There is every reason to doubt that this will make a world in which everyone is as well off as could be. But a world in which everyone is succeeding in complying pretty well with Prof. Partiality’s prescription might be better (by standards they share) than a world where most of us are failing pretty miserably to comply with Dr. Welfare’s. And given what people are actually like, one might suppose that these are the likely outcomes.

An ideal that cannot be implemented is futile. The question is, how much of a drag on moral ideals should be exercised by the stubborn facts of human psychology? How far can moral ideals ask us to transcend our self-centered human dispositions without becoming unrealistically utopian? As Appiah says,

Some aspects of human nature have to be taken as given in normative theorizing…, but to take us exactly as we are would involve giving up ideals altogether. So when should we ignore, and when insist on, human nature?

I would suggest that to idealize in this context is not to ignore human nature but to regard it, rightly or wrongly, as capable of change. Only if the change is impossible or undesirable is the idealization utopian.

Appiah illustrates a different kind of reason to avoid excessive idealization with the example of immigration policy. To even pose the problem that faces us we have to take the existence of national boundaries as given, as well as the fact that some states treat their own citizens with flagrant injustice or are beset by chaos and severe deprivation. In thinking about what obligations such a situation places on stable and prosperous states, it is no use imagining a unified world without state boundaries, or a world of uniformly just states in which people are free to move from one to another. Such ideal possibilities do not tell us what we should do now, as things are.

Appiah’s response relies on the idea of fortunate nations each doing their fair share toward alleviating the plight of those seeking asylum, while acknowledging that many nations probably won’t meet this standard. This too is an ideal, but it doesn’t depend on imagining a world very different from the actual one.

Immigration is a special case, but Appiah deploys a more general form of the argument—unsuccessfully, in my view—to criticize the structure of John Rawls’s theory of justice. Rawls presents his most general principles of justice by the device of what he called “ideal theory.” That is, he tries to describe the structure and functioning of a fully just or “well-ordered” society, in which “everyone is presumed to act justly and to do his part in upholding just institutions.” Rawls held that ideal theory was the natural first stage in formulating principles of justice, before proceeding to a systematic treatment of the various forms of injustice and the right ways to deal with them—such as criminal law and principles of rectification. The latter enterprise he described as “nonideal theory,” and he held that it depends on the results of ideal theory.

Appiah objects that the description of a fully just society is no help with the problem we actually face, which is how to make improvements in our actual, seriously unjust society. He adds:

The history of our collective moral learning doesn’t start with the growing acceptance of a picture of an ideal society. It starts with the rejection of some current actual practice or structure, which we come to see as wrong. You learn to be in favor of equality by noticing what is wrong with unequal treatment of blacks, or women, or working-class or lower-caste people. You learn to be in favor of freedom by seeing what is wrong in the life of the enslaved or of women in purdah.

But this is misguided as a response to Rawls, whose method in moral theory is to begin precisely with intuitively obvious examples of injustice like those Appiah cites. Rawls’s philosophical project is to discover general principles that give a morally illuminating account of what is wrong in those cases by showing how they deviate from the standards that we should want to govern our society. Such general principles are needed to help us judge what would be right in less obvious cases. Both levels of inquiry are essential to the systematic pursuit and philosophical understanding of justice, and the whole aim of Rawls’s theory is to unite them. It is highly implausible to claim that an understanding of the general principles that would govern a fully just society will not help us to decide what kinds of social or legal or economic changes to our actual society will make it more just.

There is much more in this rich and illuminating book, including a fine discussion of our emotional response to fiction and drama. Appiah’s insight is that when we feel genuine sadness at the death of Ophelia, it is not because of what Coleridge called the “willing suspension of disbelief,” but because of the suspension of “the normal affective response to disbelief.” We react as if we believe an unhappy young woman has died, although we do not believe it, so this is another case of idealization.

The examples that Appiah discusses are interesting in themselves, but he also thinks they offer a larger lesson:

Once we come to see that many of our best theories are idealizations, we will also see why our best chance of understanding the world must be to have a plurality of ways of thinking about it. This book is about why we need a multitude of pictures of the world. It is a gentle jeremiad against theoretical monism.

It isn’t just that we need different theories for different aspects of the world, but that our best understanding may come from theories or models that are not strictly true, and some of which may contradict one another. This is a liberating outlook, though care must be taken not to let it become too liberating. As Appiah insists, we should not allow the plurality of useful theories to undermine our belief in the existence of the truth, leaving us with nothing but a disparate collection of stories. It is conscious deviation from the truth that makes a theory an idealization, and keeping this in mind is a condition of its value.