views updated


Noncognitivists (or nondescriptivists) hold that the function of normative judgments is not, or not primarily, to describe or state facts and that because of this, these judgments lack a truth-value. A strong form of ethical nondescriptivism says that moral judgments have no descriptive function, but weaker forms say only that their nondescriptive function is primary or dominant.

Differing accounts of the nondescriptive function of moral language generate a variety of nondescriptivisms. Moral judgments have been said to express emotions, feelings, attitudes, or stances; and they have been characterized as tools for performing other nondescriptive tasks such as commanding, requesting, endorsing, or commending. A. J. Ayer, whose position is called emotivism, said that "ethical terms" express emotions or feelings and that they "are calculated also to arouse feelings, and so to stimulate action" (1952, p. 108). C. L. Stevenson, whose metaethical theory is called noncognitivism, argued that the major use of "ethical statements" is dynamic rather than fact stating. They are not, he said, primarily used to describe interests or attitudes but rather to change or intensify attitudes and to influence behavior. What Stevenson called the emotive meaning of ethical terms makes this dynamic use possible and also explains why ethical judgments, unlike factual ones, are capable of moving us to action.

From the thought that moral judgments are exclamations and disguised commands Ayer concluded that they "have no objective validity whatever" and that "it is impossible to dispute about questions of value" (1952, p. 110). Stevenson tried to show that there is a place for ethical arguments, but he did not go beyond the claim that a reason is "relevant" when it is likely to influence some attitude. This means, at least to the critics of Stevenson, that the relation between the premises and the conclusion of an ethical argument is psychological rather than logical and that there is no clear distinction between ethical argument and propaganda.

Both Ayer and Stevenson were in the positivist tradition, but by the 1950s an interest in ordinary language also led increasing numbers of analytic philosophers to nondescriptivism. These thinkers acknowledged that moral language can be used descriptively, but they insisted that its "primary" (basic, fundamental) use is to perform any of a number of nondescriptive speech acts. R. M. Hare argued that the primary function of the word good is to commend and that when we commend anything "it is always in order, at least indirectly, to guide choices, our own or other people's, now or in the future" (1952, p. 127). Words such as right and ought are used for giving advice or, as he said, for prescribing. According to Hare, the claim that something is good has both descriptive and prescriptive meaning. The descriptive meaning of the word good changes as it is applied to different things, but the prescriptive meaning remains constant because good is invariably used to commend. This is why the prescriptive meaning is primary.

Hare described his own position as nondescriptivism, but he was more positive than Ayer and Stevenson about the role and value of logic in ethical arguments. Moral judgments, he said, are a subclass of "prescriptive" rather than "descriptive" languagethey are "universalizable prescriptions." Unlike attempts to persuade or to influence attitudes, a judgment that something is good or right is a prescription that is complete in itself, even if no change is brought about in the hearer's attitudes or behavior. Hare believed that there could be logical relations among prescriptive judgments, even commands; and he developed a logic of prescriptive discourse to account for those relations. In the end he concluded that while we can argue logically about what to do, a complete justification of a moral decision will always require the adoption, without justification, of some basic principle or principles as a part of a freely chosen "way of life."

P. H. Nowell-Smith offered a form of nondescriptivism he called multifunctionalism. He said that evaluative language is used "to express tastes and preferences, to express decisions and choices, to criticize, grade, and evaluate, to advise, admonish, warn, persuade and dissuade, to praise, encourage, and reprove, to promulgate and draw attention to rules; and doubtless for other purposes also" (1954, p. 98). Though his position is more complex than Hare's, he does agree that "the central activities for which moral language is used are choosing and advising others to choose" (p. 11).

After the contributions of Ayer, Stevenson, Hare, Nowell-Smith, and others, nondescriptivism was neglected as interest in applied ethics flourished and as those who did think about metaethics developed naturalistic forms of descriptivism. The new naturalists conceded that normative language has nondescriptive functions, but they then pointed out how those functions are compatible with simultaneous descriptive intent and therefore with the possibility of evaluating normative pronouncements in terms of truth and falsity. In the 1980s interest in metaethics was stimulated by new forms of nondescriptivism developed by Simon Blackburn and Allan Gibbard. The dominant issue at that time, however, was the dispute between moral (or ethical) realists and antirealists. Nondescriptivists are more likely to be antirealists, and descriptivists are more likely to be realists, but there are complications.

Formerly, both intuitionists and naturalists were descriptivists. Intuitionists identified moral facts with nonnatural facts, and naturalists identified moral facts with natural facts. If one who believes that moral facts are natural facts can be said to be a moral realist, then both naturalists and intuitionists were moral realists and were in a position to say that moral judgments are true when they correctly describe some natural or nonnatural reality. But there is a way to combine descriptivism with antirealism and another way to combine nondescriptivism with at least the practices of the realist. J. L. Mackie develops a descriptivist account of much normative language, but he argues that judgments of moral obligation, which are thought to be both objective and prescriptive, and judgments of "intrinsic" value are always false. One who says that something is "good in itself" is always speaking falsely because nothing is good in itself.

Both Blackburn and Mackie begin with a Humean projectivism according to which the normativity we think we discover in nature is projected onto a value-free world by us. When we see and are moved by cruelty to the bull, we objectify our negative attitude, and promote it too, by saying that bullfighting is wrong. Projectivists are antirealists. Mackie combines his antirealism with descriptivism and takes this to result in an error theory. Blackburn begins with antirealism, adds his version of nondescriptivism or "expressivism," and emerges with what he calls quasi realism, the idea that the linguistic practices of the realistsaying that bullfighting is really wrong, for exampleare perfectly in order and that no error is made. One of his main concerns is to defend this quasi realism by showing how we "earn the right" to "practice, think, worry, assert, and argue" as though moral commitments are true in some straightforward way (1984, p. 257).

Blackburn's view is that we do not describe reality correctly or incorrectly when we make moral claimswe express "stances." He characterizes a stance as a "conative state or pressure on choice and action" but admits that we could also call this an attitude. But whatever we call it, "its function is to mediate the move from features of a situation to a reaction, which in the appropriate circumstances will mean choice" (1993, p. 168).

Gibbard also defends a nondescriptivist or "expressivist" account of normative judgments. Normative judgments, he says, take the form of saying that some act, belief, or feeling is "rational," or "makes sense." The point of making such a judgment is not to describe something, not to attribute a property to it, but "to express one's acceptance of norms that permit it" (1990, p. 7). A norm, according to Gibbard, is "a linguistically encoded precept," and the capacity to be motivated by norms "evolved because of the advantages of coordination and planning through language" (p. 57). There are norms of many kinds, but when we say that what someone did was morally wrong, we are expressing and endorsing norms that govern feelings of guilt by the agent and of anger by others.

Three arguments are traditionally deployed against nondescriptivists. According to the grammatical argument, since moral judgments are phrased in ordinary indicative sentences, there is a prima facie reason to treat them as statements and to treat those who make them as attempting to make statements. Nondescriptivists will reply that here the grammar is misleading, but they can then be asked to explain why this should be so. There is also a logical argument against nondescriptivism. If moral judgments lack a truth-value, then it is impossible for them to play a role in truth-functional constructions (implication, conjunction, and negation, for example) and in arguments. It is also difficult to know how they are to be interpreted when they occur embedded in complex constructions such as statements of belief and doubt. According to what has been called the phenomenological argument, not only do moral claims look and behave like descriptive utterances, they "feel" like them too. When we claim that something is good or right, we do not seem, even to ourselves, to be merely expressing ourselves or ordering others to do things. Nondescriptivists will try to explain why these judgments have this distinctive feel, but descriptivists will insist that the feeling is important data that cannot easily be explained away.

Starting with Ayer, each nondescriptivist has been forced to develop some reply to these, as well as to other, difficulties. Blackburn, for example, responds to the logical argument by developing an expressivist account of truth. He wants to show how it makes sense to claim moral truth even if there are no moral facts and even if our moral claims are no more than expressions of stances or attitudes. Gibbard sketches a solution to the embedding problem that exploits the idea that when we make a normative statement we are expressing a state of mind that consists in "ruling out various combinations of normative systems with factual possibilities." He develops a formalism that allows him to use this idea to account for "the logical relations that hold among normative statements" (1990, p. 99).

Owing to the work of Blackburn and Gibbard, nondescriptivism is alive and well, but its prospects are uncertain because it is truly difficult to develop convincing and definitive answers to the objections from grammar, logic, and phenomenology. Furthermore, nondescriptivism needs a fact/value distinction, and this is something about which philosophers have become increasingly nervous. The early descriptivists tried to reduce values to facts, or they accepted the fact/value distinction and then relegated values to a philosophically insignificant pragmatic limbo. Since then there has been a tendency to argue that many statements that appear to be safely descriptive must be understood to have nondescriptive elements. Nondescriptivists now point out that even if the line between facts and values is blurred or moved, we can still draw an important distinction between assertions and expressions. This claim, however, will continue to be challenged by those who are impressed by the descriptive nature of norms or the normative nature of descriptions.

See also Applied Ethics; Ayer, Alfred Jules; Hare, Richard M.; Mackie, John Leslie; Metaethics; Moral Realism; Projectivism; Stevenson, Charles L.


Ayer, A. J. Language, Truth, and Logic. New York: Dover, 1952.

Blackburn, S. "How to Be an Ethical Anti-realist." In Essays in Quasi-Realism. New York: Oxford University Press, 1993.

Blackburn, S. Spreading the Word: Groundings in the Philosophy of Language. Oxford: Clarendon Press, 1984.

Blackburn, S. "Wise Feelings, Apt Reading." Ethics 102 (1992): 342356.

Darwall, S., A. Gibbard, and P. Railton. "Toward Fin de Siècle Ethics: Some Trends." Philosophical Review 101 (1992): 115189.

Gibbard, A. Wise Choices, Apt Feelings: A Theory of Normative Judgment. Cambridge, MA: Harvard University Press, 1990.

Hare, R. M. Freedom and Reason. Oxford: Clarendon Press, 1963.

Hare, R. M. The Language of Morals. Oxford: Clarendon Press, 1952.

Mackie, J. L. Ethics: Inventing Right and Wrong. Harmondsworth, U.K.: Penguin, 1977.

Nowell-Smith, P. Ethics. London: Penguin, 1954.

Stevenson, C. L. "The Emotive Meaning of Ethical Terms." In Facts and Values. New Haven, CT: Yale University Press, 1963.

Stevenson, C. L. Ethics and Language. New Haven, CT: Yale University Press, 1944.

Urmson, J. O. The Emotive Theory of Ethics. London: Hutchinson, 1968.

Richard Garner (1996)