3.2.3 - Anti-realism - error theory, emotivism and prescriptivism

The Anti-Realist Problem

Moral language often sounds as if it reports facts. If someone says "cruelty is wrong", they do not usually sound as if they are merely describing their own mood or giving a personal preference. They sound as if cruelty has a moral status that others should recognise too.

Moral anti-realism challenges that appearance.

A mind-independent moral fact would be true independently of what any person, society, culture, speaker or ideal observer thinks, feels, chooses or prescribes. A mind-independent moral property would be a property such as wrongness or goodness that belongs to an action, person or state of affairs independently of our attitudes.

That creates a problem:

If there are no mind-independent moral properties or facts,
what is going on when we make moral judgements?

The three AQA anti-realist theories answer that question differently.

Do not confuse the two distinctions:

The opening argument for taking these theories seriously can be reconstructed like this:

1. P1: If ordinary moral language seems to report mind-independent moral properties, a metaethical theory must explain that appearance.
2. P2: AQA moral anti-realism denies that mind-independent moral properties or facts exist.
3. P3: There are different ways to explain moral language without such properties: systematic error, emotional expression, or prescription.
4. Therefore, C: Error theory, emotivism and prescriptivism are distinct anti-realist strategies, not three labels for the same view.

Hidden assumption: Explaining moral language does not require accepting the mind-independent moral facts that moral language appears to involve.
Inference type: Conceptual and explanatory reasoning.
Main pressure point: Whether the anti-realist explanation preserves enough of ordinary moral practice.

The lesson's evaluative question is not simply "Are there moral facts?" It is sharper: can anti-realism explain the way moral language is used in reasoning, persuasion, disagreement and claims of progress, without turning morality into nothing at all?

Error Theory

Mackie's error theory is the most direct anti-realist theory in this lesson. It keeps the realist-looking surface of moral language but denies that the world contains the moral properties needed to make ordinary positive moral claims true.

An error theorist says that when someone judges "cruelty is wrong", they are not merely expressing a feeling or issuing a command. They are making a truth-apt judgement. The problem is that the judgement presupposes a property of wrongness, and Mackie denies that such objective moral values exist.

That makes error theory a cognitivist form of anti-realism:

The "positive" qualification matters. A claim such as "nothing is morally wrong" is not itself a positive attribution of wrongness. Error theory is usually aimed at ordinary first-order claims such as "stealing is wrong", "kindness is good" or "you ought to do this".

Mackie's Chapter 1, Sections 8 and 9 support this view through the argument from relativity and the argument from queerness.

The argument from relativity is an abductive argument. It begins with disagreement and variation in moral codes, then asks which explanation is better. A realist might say that some groups perceive objective moral truth better than others. Mackie's anti-realist explanation is that moral codes reflect ways of life, social patterns and human responses.

The argument from queerness has two sides. Metaphysically, objective moral values would be unlike ordinary natural properties because they would somehow be intrinsically action-guiding. Epistemologically, if such values existed, it is unclear how we would know them. Mackie does not need to show that objective values are logically impossible. His claim is that they are too metaphysically and epistemologically strange to be credible.

The core error-theory argument can be reconstructed like this:

Claim at issue: Ordinary moral judgements involve systematic error.

1. P1: Ordinary moral judgements are cognitivist: they present themselves as beliefs that can be true or false.
2. P2: Ordinary moral judgements purport to attribute objective moral properties, such as wrongness or goodness.
3. P3: There are no objective moral properties.
4. P4: If a positive judgement attributes a property that nothing has, then that judgement is false.
5. Therefore, C: Ordinary positive moral judgements are systematically false.

Hidden assumption: Ordinary moral language is committed to objective moral properties, not merely to attitudes, prescriptions or social conventions.
Inference type: Deductive once P1-P4 are granted, with P3 supported by abductive and conceptual arguments such as relativity and queerness.
Main pressure point: Whether P2 or P3 should be accepted.

Error theory is initially plausible because it explains why moral debate feels factual. When two people disagree about whether an action is wrong, the error theorist can say they are making incompatible claims. The theory also explains why moral judgements can appear in arguments: if they are truth-apt, they can function like other propositions in reasoning.

But the cost is high.

The exam judgement should be careful: error theory is not the claim that moral sentences are meaningless. It is the claim that they are meaningful and truth-apt, but systematically false because the moral reality they presuppose is absent.

Emotivism

Ayer's emotivism gives a different anti-realist answer. It does not say that ordinary moral judgements are truth-apt but false. It says they are not factual propositions in the first place.

Ayer's background in Language, Truth and Logic matters. Chapter 1 develops a verificationist approach to meaning: a factual claim must be connected to possible experience, while analytic claims are true by meaning. Chapter 6 applies this pressure to ethics. Normative moral judgements are not analytic truths and they are not empirically verifiable descriptions of the world. Ayer therefore denies that they state factual propositions.

The key move is this:

That distinction prevents a common mistake. Emotivism is not simple subjectivism. The subjectivist says "cruelty is wrong" means something like "I disapprove of cruelty", which would be a truth-apt report about the speaker. Ayer instead treats the moral utterance as an expression of attitude. Expressing disapproval is not the same as stating that one disapproves.

The emotivist argument can be reconstructed like this:

Claim at issue: Moral judgements are expressions of attitude, not factual propositions.

1. P1: A factual proposition must be either analytic or empirically verifiable.
2. P2: Normative moral judgements are not analytic.
3. P3: Normative moral judgements are not empirically verifiable.
4. P4: If a sentence is neither analytic nor empirically verifiable, it does not state a factual proposition.
5. Therefore, C1: Normative moral judgements do not state factual propositions.
6. P5: Normative moral judgements still have a practical role in expressing attitudes and influencing action.
7. Therefore, C2: Moral judgements function as expressions of approval or disapproval rather than truth-apt beliefs.

Hidden assumption: Ayer's verificationist criterion is the correct test for factual meaning.
Inference type: Deductive within Ayer's verificationist framework, plus linguistic explanation of moral use.
Main pressure point: Whether the verificationist criterion is acceptable, and whether expression of attitude captures enough moral language.

Emotivism has obvious attractions. It explains why moral judgements are closely connected to motivation. If someone sincerely condemns cruelty, we expect some practical stance against cruelty. It also explains persuasion. Moral language often tries to move hearers, not merely inform them.

The difficulty is that moral language seems to do more than express or arouse attitudes.

A precise AQA answer should avoid the classroom caricature that emotivism makes moral language meaningless in every sense. Ayer denies literal factual meaning for normative moral judgements, but he still gives them expressive and practical significance.

Prescriptivism

Hare's prescriptivism is also non-cognitivist, but it is not just emotivism with a different label. Hare thinks moral language is action-guiding, but he gives it more logical structure than simple emotional expression.

In The Language of Morals, Hare treats moral language as a kind of prescriptive language. A prescription tells someone what to do, or commits the speaker to a practical stance about what is to be done. Moral judgements are not primarily descriptions of moral facts. They guide conduct.

Two features matter most.

Universalizability does not mean that every moral principle must be vague and general. It means that the judgement cannot depend on a bare proper name or special pleading. If I say "it is wrong for you to lie to me, but fine for me to lie to you in exactly the same relevant circumstances", I have failed to universalize my prescription.

Hare also pays attention to value terms such as "good", "right" and "ought". AQA's set boundary includes The Language of Morals Chapters 1, 5, 7 and 10.2: Chapter 1 supports the account of prescriptive language; Chapters 5 and 7 support the distinction between descriptive and evaluative meaning; and Chapter 10.2 supports the treatment of "right" and "ought" as action-guiding without being entailed by descriptive facts alone. A word such as "good" can carry descriptive information in a context, but it also commends. Calling something a good strawberry may depend on features such as sweetness or freshness, but saying it is good does not merely list those features. It evaluates or commends the strawberry according to standards.

The prescriptivist argument can be reconstructed like this:

Claim at issue: Moral judgements are universalizable prescriptions.

1. P1: Moral judgements are characteristically action-guiding.
2. P2: Action-guiding moral language is better understood as prescriptive than as purely descriptive.
3. P3: Moral judgements are governed by consistency: one must make the same judgement in relevantly similar cases.
4. P4: A prescription governed by this consistency requirement is universalizable.
5. Therefore, C: Moral judgements function as universalizable prescriptions rather than as descriptions of mind-independent moral facts.

Hidden assumption: The rational structure of moral language can be explained through prescriptions and consistency without moral facts.
Inference type: Conceptual and linguistic analysis.
Main pressure point: Whether prescriptions can support enough moral reasoning without becoming truth-apt beliefs.

Prescriptivism improves on simple emotivism in two ways. First, it explains why moral judgements are practical: they guide action. Second, it gives moral reasoning a rational constraint: consistency across relevantly similar cases.

But it also faces objections.