Let's Unriddle The Is–Ought Problem (Meta-Ethics)

The Problem

Hume’s law states that we cannot derive ‘ought’ (normative statements) from ‘is’ (descriptive statements) because there is a fundamental difference between how the world factually is and how it morally ought to be. Consider these two examples:

The sociobiological fact that men are genetically predisposed to be sexually aggressive does not permit the conclusion that men ought to rape women.
The zoological fact that people of all races are human beings does not permit the conclusion that we ought to treat people of all races equally.

The is–ought distinction is usually mentioned to discourage people from moralizing scientific facts. But what makes the derivation of ‘ought’ from ‘is’ a logical fallacy? Why do normative statements differ logically from descriptive statements?

The Solution

Core Argument

The purpose of morality is to facilitate social cooperation through principles of action.
A principle of action is moral if it facilitates social cooperation.
A statement is moral if it refers to a moral principle of action.
A principle of action can facilitate social cooperation only if it implies socioemotional consequences.
A statement is moral if it refers to socioemotional consequences.
A moral statement can refer to socioemotional consequences either implicitly (normative statement) or explicitly (descriptive statement).
Normative and descriptive statements that refer to the same socioemotional consequences are true under the same conditions.
If two statements are true under the same conditions, they are logically equivalent.
“You ought (not) to X” and “If you (do not) X, there will be socioemotional consequences” can be logically equivalent.
The derivation of ‘ought’ from ‘is’ need not be a logical fallacy.

Annotations

ad 1. As a sociobiological hypothesis, which is the best hypothesis we currently have, moral values and principles exist to facilitate cooperation between moral agents in a society. This does not imply that science can define the content of morality, which would beg the is–ought question, but it does imply that science can define the scope and concept of morality. Like language, morality serves a social function and is thus neither a theological nor a private matter; the purpose of morality is neither to foster man’s relationship with God nor to guide a socially isolated individual’s actions. Even if an action seems to be morally right or wrong beyond social relations, that intuition is still based on human biology and social conditioning.

ad 2. From #1 it follows that a principle of action is moral if it facilitates social cooperation. For example, “Generosity is good” refers to a certain action pattern and constitutes a principle that may facilitate cooperation in a society. By contrast, consider the following two statements:

“I know a way out of here” facilitates social cooperation, but it is not a principle.
“Oxytocin strengthens pair-bonds” relates to the facilitation of social cooperation, but it is not a principle of action.

ad 3. By definition, moral statements refer to moral principles. A paradigmatic moral statement has the form “You ought to X.”

ad 4. A principle of action cannot facilitate social cooperation without implying socioemotional consequences, which are either emotions shaped by social conditioning or social standing resulting in emotions. Typical socioemotional implications of moral behavior include shame, guilt, pride, sympathy, and embarrassment as well as the pain of social exclusion (ostracism) and the pleasure of social elevation (respect, honor, power). We know from cognitive science that morality would not be possible without emotions, because pure reason cannot induce human behavior.

ad 5. From #3 and #4 it follows that moral statements refer to socioemotional consequences.

ad 6. A normative statement can imply the same socioemotional consequences that a descriptive statement can express. For example, “Lying causes negative socioemotional consequences” expresses a fact that can be implied by “You ought not to lie.” Consider also how a person might respond to the normative statement “You ought not to lie”:

He might say, “Ok” if he understands and accepts the implied socioemotional consequences.
He might ask, “Why?” if he does not (or if he questions the metaphysics implied by #1).
He might retort, “Or what?!” to challenge the consequences.
He might reply, “Fuck off!” to voice his disapproval of the consequences.

ad 7. Normative and descriptive moral statements are true under the same conditions if they refer to the same socioemotional consequences:

Both the descriptive statement “If you eat with your left hand, you will experience negative socioemotional consequences” and the normative statement “You ought not to eat with your left hand” are true, for example, if the recipient was socially conditioned to feel embarrassed when he eats with his left hand, or if he loses respect when he does, or if he gets shamed for it.
Both the descriptive statement “If you eat with your right hand, you will experience positive socioemotional consequences” and the normative statement “You ought to eat with your right hand” are false if uttered in a social context where nobody cares about what hand people use to eat.

Whether an ‘ought’ statement has a truth value in the first place depends, of course, on the pragmatic context. For example, an ‘ought’ used for moral conditioning (e.g., teaching a child how to behave) or moral outrage (e.g., signaling disapproval of a behavior) cannot be true or false, whereas an ‘ought’ used to indicate a moral/cultural fact (i.e., that a behavior has socioemotional consequences) does have a truth value.

ad 8. By definition, two statements are logically equivalent if their truth table columns are identical under their governing operators.

ad 9. From #7 and #8 it follows that the normative command “You ought (not) to X” is logically equivalent to the descriptive conditional “If you (do not) X, there will be socioemotional consequences” if both statements are used in the same pragmatic context where they have truth values and refer to the same socioemotional consequences.

ad 10. From #9 it follows that ‘ought’ can be derived from ‘is’. In particular, socially embedded normative statements can be derived from descriptive statements about sociomoral conditioning. In other words, we can deduce what we ought to do in a social group from facts about the group’s moral values.

Final Remarks

We can now review the two examples mentioned at the beginning. Given that all the following sentences have truth values, we can say:

“You ought not to sexually assault women” is logically equivalent to “Rapists experience negative socioemotional consequences, and so will you if you sexually assault women.” ‘Ought’ is ‘is’, and our sociobiological knowledge about genetic predispositions is relevant to the extent that it influences those consequences.
“You ought to treat people of all races equally” is logically equivalent to “Racists experience negative socioemotional consequences, and so will you if you do not treat people of all races equally.” ‘Ought’ is ‘is’, and our zoological knowledge about the unity of mankind is relevant to the extent that it influences those consequences.

With the is–ought gap thus bridged, can we now derive moral principles from scientific insights? Only if we maintain our modest view of what a moral principle is. We can discover moral principles as sociological and anthropological facts, but we cannot say that something is good because it is natural. Other facts of science may be used not as moral principles, but to manipulate moral conditioning.

“How ought we to use science to manipulate moral conditioning?” That typical philosopher’s question makes no sense because it transcends the natural scope of morality. The question must rather be “How do we want to use science to manipulate moral conditioning?” Morality cannot act as a substitute for a personal or political will. ‘Ought’ cannot replace ‘want’.

Lastly, everything stated thus far seems to imply moral relativism, especially since the socioemotional consequences of actions obviously differ between social groups. But can there be no normative standards beyond social conventions? Can moral judgments never be universal?

Moral judgments can be universal to an extent, namely, to the extent that there is cooperation between social groups. When the moral principles of groups clash, people need more universal guidelines for cooperation, need ‘meta-moral’ principles, need normative ethics. However, every ‘ought’ on that meta-moral level is again built upon the purpose of cooperation, now inter-group cooperation with sociopolitical consequences.

Update: According to #2, a principle of action is moral if it facilitates social cooperation. “Use past tense when you talk about something that happened yesterday” is a principle of action that facilitates social cooperation. That would make it a moral principle. Either we accept that using proper grammar is a moral matter, or there is more to morality than social cooperation and premise #1, from which #2 is derived, is false. One solution to this problem could be to extend the definition of morality to include a limitation on self-interest, because grammar does not regulate egoism like moral principles do, and to the extent that individuals do not pursue selfish interests, they do not need morality to begin with.