Emotive Theory of Ethics

views updated


The term emotivism refers to a theory about moral judgments, sentences, words, and speech acts; it is sometimes also extended to cover aesthetic and other nonmoral forms of evaluation. Although sometimes used to refer to the entire genus, strictly speaking emotivism is the name of only the earliest version of ethical noncognitivism (also known as expressivism and nondescriptivism).

Classical noncognitivist theories maintain that moral judgments and speech acts function primarily to (a) express and (b) influence states of mind or attitudes rather than to describe, report, or represent facts, which they do only secondarily if at all. For example: To say "Stealing is wrong" is not primarily to report any facts about stealing but to express one's negative attitude toward it. Emotivists also deny, therefore, that there are any moral facts or that moral words like good, bad, right, and wrong predicate moral properties; they typically deny that moral claims are evaluable as true or falseat least in respect of their primary meaning. The attitudes expressed by moral judgments are held to be "conative" (that is, they have a motivational element) and not "cognitive" (that is, they are not beliefs/do not have representational content). Species of noncognitivism are differentiated by the kinds of attitude they associate with moral thought and discourse: emotivism claims that moral thought and discourse express emotions (affective attitudes, sentiments, or feelings) or similar mental states, typically of approval and disapproval, and is therefore sometimes called the "boo-hurrah" theory of ethics.

To understand emotivism, it is important to contrast it with subjectivism, the view that moral judgments and utterances represent, report, or describe someone's attitudes (for example, that we can translate "Stealing is wrong" as "I disapprove of stealing"). Noncognitivist theories deny that moral expressions of attitude take the form of report or description: They are often vague about the expressive mechanism, but it is supposed to bear a family resemblance to that of ejaculations (for example, uttering "Ouch!" to express being in pain) and performatives (for example, saying "Thank you" to express gratitude). Saying "Stealing is wrong" is therefore like saying "Boo to stealing!"

The significance of this difference is apparent, to the advantage of noncognitivism, when one examines what the strategies have to say about moral disagreements. Subjectivists must acceptwhereas noncognitivists denythat moral claims are made true or false by facts about people's attitudes. If A asserts "Stealing is wrong," and B responds "Stealing is not wrong," it is possible, from a subjectivist view, for A and B to be expressing compatible judgmentsif they are reporting the attitudes of different peopleand therefore not actually to be disagreeing at all. Although noncognitivism does not portray A and B as disagreeing about any fact, it does claim a "disagreement in attitude": A opposes stealing, and B does not.

According to emotivists, we engage in moral discourse in order to influence the behavior and attitudes of others. They claim, therefore, that moral utterances have a psychological function of arousing emotions in others, based on a human susceptibility to emotional influence by exposure to the emotional expressions of others. Charles L. Stevenson even identifies a statement's emotive meaning with this causal tendency.

Almost all emotivist theories acknowledge that moral judgments possess some content that is descriptive and truth-apt. Consider first "thick" evaluative terms such as the names of virtues or vices (for example, brave ) and pejoratives (for example, geek ); here it is easy to distinguish a descriptive meaning and an emotive meaning. But most emotivists also ascribe descriptive content to "thin" evaluative terms like good and right. One common account of this content (Stevenson 1944, Edwards 1955, Hare 1952, Dreier 1990, Barker 2000, Gibbard 2003) is that the property predicated of an object T by wrong, for example, is the property for which the speaker disapproves of T. Suppose Elizabeth declares "Stealing is wrong" and disapproves of stealing because she believes it typically causes misfortune to its victims; then the descriptive meaning of her utterance is that stealing typically causes misfortune to its victims. However, this meaning is deemed secondary because (a) it depends upon the emotive meaningthe descriptive meaning of wrong will differ from context to context, speaker to speaker, and even occasion to occasion, according to what arouses speakers' emotions, and (b) it has little or no moral significance. A and B will argue over whether stealing is wrong if they differ in attitude toward stealing but not if they differ only with regard to which properties arouse their disapproval of stealing or over whether stealing has some particular property.

History and Development

Although suggestions of emotivism can be found throughout the history of philosophy (David Hume and other early modern sentimentalists have particularly close affinities), the emergence of the theory is usually attributed to a series of short suggestions by British philosophers in the 1920s and 1930s (Ogden and Richards 1923, Barnes 1933, A. S. Duncan Jones as reported in Broad 19331934, Ayer 1936); however, earlier formulations appear in German/Austrian value theory from the late nineteenth century (Lotze 1885, Windelband 1903, Marty 1908, and see Satris 1987 for this influence on Anglo-American emotivism). The British emotivists were reacting, in part, to the metaethical theory of nonnaturalism (or intuitionism) advocated by G. E. Moore, H. A. Pritchard, W. D. Ross, and others.

Moore had persuasively argued that moral words could not be defined except in terms of other moral words and inferred (invalidly, as was revealed by the discovery that nonsynonymous terms could be coreferential) that moral words could not refer to "natural" or empirical properties and that moral sentences could not describe natural or empirical facts. Any such attempted definition left out something essential. (This claim is closely related to the alleged is/ought distinction, or "fact-value gap"). Emotivists were convinced by these arguments, but some, influenced by logical positivismthe doctrine that only sentences which are empirically verifiable are meaningfulbalked at the notion of "nonnatural," nonempirical moral properties and facts. In their diagnosis, the essential something that cannot be captured by any naturalistic analysis of moral language is the expression of speakers' emotions.

Emotivism found its greatest and most dedicated champion in the person of the American philosopher Charles L. Stevenson (1937, 1944) and enjoyed its heyday in the 1940s and 1950s (Nowell-Smith 1954, Edwards 1955) before being largely supplanted by forms of noncognitivism that were thought to be less vulnerable to objection (especially the prescriptivism of Hare 1952, 1963). To philosophers seeking to condemn the horrors of World War II in absolute terms, the claim that moral judgments merely express feelings appeared inadequate. Emotivism's legacy is a widespread recognition today of the significance of emotions for ethical thought, and the efforts of a number of contemporary philosophers since the 1980smost notably Simon Blackburn (1993, 1998)who continue to argue for its central tenets.

The Case for Emotivism

The philosophical stature of emotivism has risen from a number of solidly argued foundations: the apparent failures of efforts to give naturalistic definitions of moral words or to identify natural properties as their referents, epistemological scruples about the existence of nonnatural properties, and the reliable link between moral judgment and emotion. Philosophers still vigorously disagree about whether or not it is possible to find objective referents for moral terms, however, and there are alternative explanations of the connection between moral judgment and emotion: perhaps moral words name properties that reliably arouse emotional responses in us, perhaps they name the dispositional properties of reliably arousing emotional responses, or perhaps their use conversationally communicates speakers' approval and disapproval without in any strict sense "meaning" it.

Further, many philosophers maintain that it is possible and not very unusual for people to make sincere moral judgments without feeling or expressing the relevant emotion (this discussion centers on a figure known as the "amoralist") and that emotive meaning is, therefore, not an essential element of moral judgment. Emotivists commonly respond with the claim that these are not genuine moral judgments but are made in "inverted commas"i.e. that they merely mimic the practice of moral judgment. The case for emotivism is not bolstered by this claim, however, unless grounds can be found for accepting the "inverted commas" diagnosis that are independent of emotivist convictions themselves.

The emotivist explanation of moral language also provides simple answers to a number of puzzles in metaethics: First, it explains the fact that people are typically motivated to behave in accordance with their moral judgments. Cognitivists have some difficulty explaining this motivational connection because they identify moral judgments with beliefs. On an orthodox view, a belief is not enough to motivate action by itself; it needs to be combined with a desire or similar conative attitude. But, according to emotivism, moral judgments consist in favorable and unfavorable attitudes, and people are likely to perform the actions they feel favorably toward and likely to avoid actions toward which they feel unfavorably.

Second, emotivism explains the synthetic a priori character of moral judgment stressed by nonnaturalists: that is, that despite the fact that an empirical description of a state of affairs or action entails neither by logic nor by meaning the goodness or badness or rightness or wrongness of that state of affairs or action, its description alone nonetheless suffices for us to be confident in passing moral judgment on it. Although it may seem mysterious how anyone could know just from description of a state of affairs or action that it necessarily possesses some further, unspecified property, we have no such need for further information in order to respond emotionally.

Third, emotivism explains the supervenience of the moral on the empirical: why moral characteristics are such that if two states of affairs differ in any moral respect, they must also differ in some nonmoral or empirical respect. If a person is disposed to have a certain emotional response to some state of affairs, then he or she is disposed to have the same response to any qualitatively identical state of affairs. A person will be disposed to make the same moral judgment about two states of affairs, therefore, unless there is some difference between those states that arouses different emotions. While emotivism has an easier task offering solutions to these problems than most descriptivist theories, it must contend with noncognitivist rivals that offer similar explanatory resources.


Most of the objections to emotivism in particular are also objections to noncognitivism in general and focus on respects in which moral thought and discourse behave like ordinary, factual, truth-evaluable cognitive thought and discourse. These objections have been widely believed to refute noncognitivism of all varieties, and accordingly the emphasis in recent noncognitivist writing is on the "quasi-realist" project (Blackburn 1993) of explaining how nondescriptive thought and discourse can mimic ordinary descriptive thought and discourse. The treatment here focuses on the significance of these objections for emotivist theories.

the embedding (or frege-geach) problem

Emotivism purports to tell us the meaning of moral sentences; however as P. T. Geach (1960, 1965) and John Searle (1962) have pointed out, it and other forms of noncognitivism appear to succeed at most at explaining one kind of use of simple moral sentences: their use in direct assertion (for example, saying "Stealing is wrong"). Emotivism claims the descriptive form of simple moral sentences is merely a disguise. However simple moral sentences are also given many other uses in which they also behave like descriptive sentences and for which emotivist explanations seem inappropriate or impossible. Consider embedding of simple moral sentences into complex sentences and indirect contexts: disjunctions ("Either stealing is wrong, or Robin Hood was a saint"), belief ascriptions ("Elizabeth believes that stealing is wrong"), conditionals ("If stealing is wrong, then Joe ought not take Mary's lunch"), predications of falsehood ("It is not true that stealing is wrong"), and interrogatives ("Is it true that stealing is wrong?). In each case, a speaker uses the simple moral sentence "Stealing is wrong" but does not express emotions or unfavorable attitudes towards stealing. The emotivist proposal therefore is not helpful in understanding the simple moral sentence in these uses, which is reason to doubt whether it has captured its meaning at all.

It is possible to extend the emotivist account by assigning meanings in each of these contexts, but doing so introduces a further difficulty. Consider a simple moral argument: P1. If stealing is wrong, then Joe ought not take Mary's lunch; P2. Stealing is wrong; P3. Therefore, Joe ought not take Mary's lunch. This looks like a standard instance of modus ponens and therefore a straightforwardly valid argument. But if we attribute different meanings to "stealing is wrong" as it occurs in each premise, then the argument equivocates, and the conclusion doesn't follow. (Indeed, if P2 is interpreted as a mere expression of emotion without truth value, nothing can logically follow from it). Emotivism therefore casts doubt on the possibility of drawing inferences to or from moral claimssomething we do all the time.

Emotivists as early as Stevenson made use of minimalist theories of truth to argue as follows: to claim that p is true is simply to claim that p, so anyone who is disposed to claim "Stealing is wrong" is entitled to claim that "Stealing is wrong is true." But as the discovery of the embedding problem postdates emotivism's heyday, we do not find solutions to it from self-identified emotivists. Contemporary noncognitivists, however, devote much attention to the problem (especially Blackburn), and there are two broad strategies available: First, if some meaning can be found for the simple moral sentence that is common to these various embeddings and is compatible with emotivism, then arguably standard logic will allow moral inferences. There are two possibilities here. (a) Some seek to identify a noncognitive content that is common to all uses of moral sentences and that plausibly can be embedded in different sentential contexts. These efforts are characteristically found outside of the emotivist tradition (particularly in the work of Hare and Allan Gibbard), and the strategy does not seem so compatible with the emotivist doctrine that simple moral sentences express emotions; (b) Emotivists can turn to the supposed secondary descriptive content of moral claims to explain moral inferences. Because these descriptive contents have truth values, there is no difficulty in forming valid arguments with them. The success of any such explanation depends on the plausibility of the emotivist's claim to have identified the truth-conditional content of the premises and conclusions of moral arguments; it is also arguable that any success must come at the cost of abandoning genuine emotivism and noncognitivism.

Second, even if it is granted that there are no truth relations between the premises of moral arguments and between the contents of moral judgments, it is arguable that there are relations of coherence or consistency between the judgments or states of mind that express those contents. Blackburn accordingly proposes and develops a "logic of attitudes," a system of norms governing the consistency of combinations of attitudes. The conditional premise P1 above, on this view, expresses approval of disapproval of Joe's taking Mary's lunch in the circumstance that one disapproves of stealing. A's attitudes are then allegedly inconsistent if A holds both this second-order attitude and the attitude of disapproval towards stealing expressed by P2 but does not also disapprove of Joe's taking Mary's lunch, the attitude allegedly expressed by P3. Accused by a number of critics of conflating logical inconsistency with pragmatic incoherence (Hale 1986, Schueler 1988, Brighouse 1990, and Zangwill 1992), Blackburn suggests that we can expand the concept of consistency to encompass pragmatic and logical forms. Critics argue that this strategy is not successful: because there is no form of merely pragmatic incoherence that exactly mimics logical inconsistency, Blackburn must claim that some apparently valid moral arguments are actually inconsistent (Hale 1993 and Van Roojen 1996), but noncognitivists have not been deterred.

reasons and justification

Emotivism is charged with being unable to accommodate the important role of rational argument in moral discourse and dispute. Although it emphasizes moral discourse's function of influencing others' behavior, it is thought to characterize this efficacy wrongly, as similar in kind to that employed in manipulation, intimidation, and propaganda. According to emotivists, we engage in moral argumentation with the immediate aim of arousing emotions in others, and moral utterances accomplish this by direct psychological causation. Their opponents object that genuine moral discourse involves furnishing others with reasons, as rational agents, to recognize as correct and thereby accept one's moral views (Hare 1951 and Brandt 1959).

It is true that conscientious moral debaters offer factual considerations as evidence or justification for their positions, and emotivists do not deny it. According to Stevenson, moral argument can take both "rational" and "nonrational" (or "persuasive") forms. On Stevenson's view, by a "reason" for a moral judgment we mean any factual consideration that might influence someone's emotions in the direction of that judgment, and therefore "rational" means of moral argument consist in offering such considerations. Protagonists in a debate over the morality of legalized abortion, for example, might dispute the facts about its consequences. "Persuasive" argumentation, on the other hand, consists in the use of emotive language for its direct psychological effects.

One line of objection, spearheaded by Richard Brandt, observes that it is possible to be emotionally influenced by considerations that are morally irrelevant, and argues that emotivism cannot accommodate the distinction between what is morally relevant and morally irrelevant. Stevenson's reply exhibits a typical noncognitivist strategy: he insists that we can meaningfully distinguish between morally relevant and irrelevant influences on people's attitudes but that when we do so, we are making further moral (and hence emotive) judgments. To judge a consideration morally irrelevant is therefore to express disapproval of being emotionally influenced by it.

other objections

Clearly not just any emotional response constitutes a moral judgment. Emotivists therefore distinguish moral judgments from other kinds of affective or conative reaction by appealing to a distinctive kind (or kinds) of moral emotion. Some critics object that moral approval and disapproval cannot be adequately differentiated from other kinds of affective and conative states without invoking the very moral concepts that emotivists seek to explain by themand therefore that moral emotions are in fact cognitive attitudes. Moral approval, for example, can arguably only be adequately characterized as the attitude of judging something to be morally good. If this is correct, then emotivism puts the cart before the horse in attempting to explain moral judgments by appeal to emotional states. However, if moral attitudes are not cognitive and are simply affective or conative responses, then it is questionable whether they have the sort of first-person authority that moral judgments purport to possess. If Gary's judgment that homosexuality is morally wrong rests on nothing more than a disposition to have an unpleasant feeling when he contemplates homosexuality, then he may have as good or better reason to resist, suppress, or work to change his emotional sensibilities as he has to oppose homosexuality.

Another concern addresses whether emotivism has the resources to distinguish between accepting the negation of a moral claim and not accepting that moral claim. Believing that the next president of the United States will not be a woman is not the same mental state as not believing that the next president of the United States will be a woman; likewise it seems that accepting that abortion is not wrong is not the same mental state as not accepting that abortion is wrong. Critics charge, however, that emotivism has to explain both in terms of not feeling disapproval toward abortion.

See also Brandt, R. B.; Ethical Relativism; Ethical Subjectivism; Ethics, History of; Ethics, Problems of; Hare, Richard M.; Hume, David; Intuitionism and Intuitionistic Logic, Ethical; Logical Positivism; Moore, George Edward; Noncognitivism; Ross, William David; Searle, John; Stevenson, Charles L.; Value and Valuation.


Ayer, A. J. Language, Truth and Logic. London: Gollancz, 1936.

Barker, Stephen J. "Is Value Content a Component of Conventional Implicature?" Analysis 60 (2000): 268279.

Barnes, W. H. F. "A Suggestion about Value." Analysis 1 (1933): 4546.

Blackburn, Simon. Essays in Quasi-Realism. New York: Oxford University Press, 1993.

Blackburn, Simon. Ruling Passions. New York: Oxford University Press, 1998.

Brandt, Richard. Ethical Theory. Englewood Cliffs, NJ: Prentice-Hall, 1959.

Brighouse, M. H. "Blackburn's ProjectivismAn Objection," Philosophical Studies 59 (1990): 225233.

Broad, C. D. "Is 'Goodness' the Name of a Simple, Non-natural Quality?" Proceedings of the Aristotelian Society 34 (19331934): 249-268.

Dreier, Jamie. "Internalism and Speaker Relativism." Ethics 101 (1990): 626.

Edwards, Paul. The Logic of Moral Discourse. Glencoe. IL: Free Press, 1955.

Geach, P. T. "Ascriptivism." Philosophical Review 69 (1960): 221225.

Geach, P. T. "Assertion." Philosophical Review 74 (1965): 449465.

Gibbard, Allan. Thinking How to Live. Cambridge. MA: Harvard University Press, 2003.

Hale, Bob. "Can There Be a Logic of Attitudes?" In Reality: Representation and Projection, edited by J. Haldane and C. Wright. Oxford: Oxford University Press, 1993.

Hale, Bob. "The Compleat Projectivist." Philosophical Quarterly 36 (1986): 6584.

Hare, R. M. "Freedom of the Will." Proceedings of the Aristotelian Society 25 (1951): 201216.

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

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

Lotze, Hermann. Outlines of Logic and the Encyclopedia of Philosophy, edited and translated by G. T. Ladd. Boston: Ginn, 1885.

Marty, Anton. (1908). Untersuchungen zur Grundlegung der allgemeinen Grammatik und Sprachphilosophie. Halle: Niemeyer.

Nowell-Smith, P. H. Ethics. Baltimore: Penguin Books, 1954.

Ogden, C. K., and I. A. Richards. The Meaning of Meaning. New York: Harcourt, 1923.

Satris, Stephen. Ethical Emotivism. Dordrecht: Martinus Nijhoff, 1987.

Schueler, G. F. "Modus Ponens and Moral Realism." Ethics 98 (1988): 492500.

Searle, John. "Meaning and Speech Acts." Philosophical Review 71 (1962): 423432.

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

Stevenson, Charles L. "The Emotive Meaning of Moral Terms." Mind 46 (1937): 1431.

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

Van Roojen, Mark. "Expressivism and Irrationality." The Philosophical Review 105 (1996): 311335.

Windelband, Wilhelm. "Was ist Philosophie?" In Präludien: aufsätze und reden zur philosophie und ihrer geschichte. 2nd ed. Tübingen: J. C. B. Mohr, 1903

Nick Zangwill. "Moral Modus Ponens." Ratio 5 (1992): 177193.

Stephen Finlay (2005)