A Gentle Introduction to Metaethics

In Plato’s Republic, Socrates and his companions examine whether ethics are a human construct or part of the fabric of reality. Thrasymachus argues that morality is an ideology that promotes the interests of the strong over the weak. Concepts such as repaying your debts or respecting property rights are useful for controlling people, rather than being objectively right. Glaucon has a more positive view of morality as a human construct. To him, ethics is a system for solving the problem of cooperation and living together. Socrates, however, rejects both positions, arguing instead that morality is an objective part of reality. 

What Socrates and his companions are debating is called metaethics, the study of second-order questions about ethics, such as “Are moral facts real?” or “What do ethical statements express: beliefs about the world or desires?”. This contrasts to normative ethics, and applied ethics. Normative ethics examines first-order questions about ethics, trying to discover general theories of deciding what is right and wrong. Examples of normative ethics include utilitarianism and virtue ethics. Applied ethics takes the systems developed in normative ethics and applies them to practical scenarios.


Socrates is a realist about ethics, holding the belief that there are real ethical facts in the world that are objective. Opposed to realism is antirealism, the belief that ethical facts don’t exist.

If ethical facts exist out there to be discovered, how do we discover them? This depends on what kind of facts they are. Philosophers have come up with three options for what kind of fact morals could be:

  • Natural facts: Moral facts are physical facts like the law of gravity, or reducible to natural facts in some way.
  • Supernatural facts: Moral facts are supernatural, for example, coming from God.
  • Non-natural facts: Moral facts are neither natural nor supernatural: they are their own kind of fact. 1

Dividing up facts into categories like these is known as ontology, the study of what is. As we’ll see, the ontology we commit to will influence how we discover and think about ethics.


Many philosophers embrace ontological naturalism, the belief that there are only natural facts that can be discovered through empirical investigation. Naturalists are guided by the principle of parsimony, which states we shouldn’t invent new ontological categories unless absolutely necessary. If one is also committed to realism, ethical facts must therefore be natural facts or reducible to natural facts.

A challenge for naturalists is to explain how we can discover ethical facts through science. How can an experiment tell us what is morally valuable? Philosopher David Hume famously argued that one cannot derive an “is” from an “ought”, and science is firmly rooted in the “is” camp2. There are a few ways we might derive an ought from natural facts that ethical naturalists have proposed.

Evolutionary Theories

Early modern philosophers such as Hobbes and Rousseau didn’t necessarily take morality as an eternal and unchanging matter - they theorized about a pre-moral period and asked how humans could have developed normative behaviors from that pre-moral period. While Hobbes’ and Rousseau’s theorizing was based on a mistaken picture of what early life was like, we can use the same methodology to trace morality from our animal ancestors to modern humans. Philosopher Patricia Churchland argues that changes in the mammalian brain starting 350 million years ago caused us to develop ethical systems. As mammals evolved to care for their young, they developed the ability to feel pleasure and pain based on the reactions of others in their group. This allowed mammals to learn from each other and develop a distinct kind of sociality that rewarded and sanctioned each other based on how they lived up to various norms; an important precursor to morality. To Churchland, ideas such as the Golden Rule, Utilitarianism and the Categorical Imperative are the end product of this evolution of cooperative norms.

If we satisfy ourselves that such an evolutionary story is right, and we think of ethical facts as what the evolved norms we developed are, then we can easily situate moral facts as a kind of natural fact. Of course, there are problems with this line of thinking: is what helps us survive really the same thing as what is moral? Evolution favors low-cost answers to problems of survival and reproduction. Evolution might cause us to eat other animals, but is that really right? There’s no reason to think evolved norms will always correspond with what we judge as moral. This argument, known as the evolutionary debunking argument challenges ethical naturalists to explain how humans could have evolved to make objective judgements about morality independent of our narrow interests and desires. To philosophers opposed to evolutionary ethical naturalism, there will always be a gap between the norms we evolved to have and what is truly moral.

A Posteriori Naturalism

Evolutionary metaethicists try to define morals in terms that science can study. That is an example of a priori investigation, where we deduce facts by reflecting on the meanings of words and concepts. But definition isn’t the only way we can reduce one kind of fact to another. We could use a posteriori investigation instead, where we discover facts through experiencing the world. For example, we talk about water in terms of its properties, such as clearness and wetness. However, thanks to chemistry and a posteriori investigation of the properties of water, we know that the structure of water is H2O. Similar to investigating the properties of water, we might investigate the properties of moral goodness in the world and reduce them to natural facts.

Of course, it is one thing to put water in a test tube and analyze its properties, but quite another to analyze goodness. A posteriori naturalists concede that ethics is going to be a lot more difficult and complex to analyze than water. But they would also say there are already natural concepts, such as being healthy and being alive, that we know exist but do not have precise definitions for. With further scientific investigation, we might be able to figure out what exactly we are talking about when we refer to moral goodness. This school of thought is also known as Cornell Realism, since many of its proponents taught or studied at Cornell.

An objection to this line of reasoning is the Moral Twin Earth Thought Experiment by Terrence Horgan and Mark Timmons. In this thought experiment, there are two earths: Earth and Twin Earth. On Earth, goodness is connected to utilitarian criteria, but on Twin Earth, goodness is connected to universalizable norms (Kantian ethics). However, the concept of goodness plays the same practical role in both worlds: people are motivated to do things because it is the right action, and people disapprove of people taking the wrong action. If the denizens of earth and twin earth were to talk to each other about what is right, they would conclude they were talking about different things, even though they used the same method to get to their understanding of ethics. Horgan and Timmons think this undermines Cornell Realism because goodness should mean the same thing in both worlds.


So far, we have assumed that natural facts are only facts that can be discovered by the scientific method. But there is another tradition stemming from Aristotle that requires significant metaphysical reflection. These are facts about things’ natures. For example, we speak about human nature or an animal’s nature. In this we consider things as functional kinds, classes of things defined by the purpose for which they are created. For example, a knife is a functional kind; it exists for cutting. A good knife is one that cuts well, so we can say that sharpness is a property that makes a knife a good knife.

Extending this idea to human morality, however, is difficult. Knives are designed by their makers to cut things, but what are people for? Were we designed for something, and do we all have a single function? Neo-Aristotelians answer this with the concept of a characteristic way of living. For example, oak trees characteristically grow to 65-130 feet, so the properties that make an oak tree a good oak tree are the ones that enable it to grow to its natural height, such as having a strong root system. These normative facts about how an oak tree should be come from natural facts about its biological nature.

Morality could be the properties that support humans’ characteristic way of living, a form of virtue ethics. But one might object that these facts are different from ethical facts. After all, eating other animals, sexual infidelity and waging war are also characteristically human traits, but are they ethical?


Another objection to neo-aristotelianism is that if we investigate different cultures and ways of human life, there may not be a universal answer to what is an “ethical good way of living.” Instead, there would be different virtues depending on what society you lived in. This is a form of relativism, where right and wrong are relative to something else. It is a bit surprising that relativism is included among realist theories of morality. If anything, shouldn’t it count as a form of antirealism? It depends on the form of relativism one is arguing for. Relativism denies there are objective, universal ethical facts, but there’s another sense in which ethical facts are out there to be discovered. In this sense, morality is just one of many sets of rules that exist, similar to how there are different legal jurisdictions in the world.

For example, David Copp argues that different societies have different non-moral values and needs, and these facts influence the justified moral code for that society, or moral community. For example, one society may need careful conservation of water because of frequent droughts, and another may need access to high-speed travel because of a far-flung population. The objection here is that while needs may vary between communities, the reasoning behind the ethics may not. This would give us a common foundation beyond an individual’s communities values and needs.

Another philosopher, Gilbert Harman, argues that ethical statements only make sense relative to agreements in a moral community. One community might agree to bury their dead, and another agree to cremate them. We can only make sense of the statement “It is ethically wrong not to bury the dead” relative to an agreement made in a moral community. A problem here is defining “moral community.” We have lots of cross cutting agreements between different people in our lives. I may have a different agreement at my workplace than with my neighbors. If we are in multiple moral communities, what do we do when one agreement contradicts another? How do we draw the line between insiders and outsiders of a given moral community? An extreme answer is that we’re all a moral community of one, known as simple subjectivism. But if that’s the case, there’s no need for agreements at all - you don’t need to make an agreement with yourself.

Another objection to relativism is how to make sense of moral disagreement. According to relativism, two people from different moral communities are talking past each other when they discuss morality. But perhaps we could find common ground between our ethical systems, a sort of least common denominator between our moral systems. If so, that would undermine the relativists’ claim that there are no universal values.

Divine Command Theory

A common assumption between naturalistic theories of ethics is that morality is a purely human phenomenon. If we are after universalizable, eternal ethical truths, naturalism is an unsatisfying approach. Isn’t it possible that what we talk about as right and what is actually right are different things? Theistic philosophers seek morality in God’s commands, known as divine command theory. This grounds morality in an objective supernatural fact rather than the “man is the measure of all things” of naturalism. For example, G.E.M Anscombe argued that we talk about morality as law, and that presupposes the existence of a lawgiver. Since only God could bind all humans to a universal command, if we stop believing in God, moral statements become a kind of nonsense.

Another kind of divine command theory is response-dependent metaethics. Roderick Firth imagines that facts about what is right or wrong are determined by how a perfect ethical judge would react to an action. Such a judge would be omniscient, impartial and dispassionate. While Firth doesn’t explicitly suggest God, God is an obvious choice for such a judge.

A challenge to divine command theory is the Euthyphro Dilemma, from the platonic dialogue Euthyphro. In it, Socrates asks:

Is the pious loved by the gods because it is pious, or is it pious because it is loved by the gods?

In other words, is something good because God wills it, or does God will it because it’s good? If we take the first horn, it would seem that if God commanded cruelty, that would be what is right. Doesn’t this make God arbitrary and capricious? If we take the second horn of the dilemma, there is a standard for morality that is external to God, undermining divine command theory.


If we’re dissatisfied with natural and supernatural explanations of ethics, but want to stay moral realists, that leaves only one answer: morality is a non-natural fact. In philosophy jargon, moral facts are sui generis, of their own kind. Ethical facts, after all, seem much different from natural facts. For one thing, ethical facts prescribe the way something ought to be rather than how the world is; therefore we have to investigate ethics differently from the way we investigate empirical facts.

A naturalist might object that we’re creating new ontological commitments that we’re not justified in believing. But there’s at least one other kind of plausible non-natural fact: mathematical facts. While math is undoubtably real, the infinite lines and points of geometry are different from anything that we encounter in the real world. We also pursue math in a different way than we do science. We start with axioms and use proof to discover theorems, rather than make observations and generalize. Of course, just because we concede the existence of one kind of non-natural fact doesn’t prove that ethical facts exist and are non-natural. However, we can construct analogous arguments. Quine and Putnam argued that we are justified in believing in platonic mathematical entities because they are indispensable to our scientific theories.

A similar argument can be made for normative facts and their role in practical reasoning. We often have to make practical decisions about our lives (for example, should I keep reading this article, or read some other internet article) and weigh the pros and cons of such a decision. But in order to have such pros and cons, we have to have some normative belief about what is good and what is not. Normative facts are indispensable here, and we need to ontologically commit to them to make practical decisions. This doesn’t get us all the way to ethical facts, just that we must have normative reasons when choosing from various different options. However, isn’t ethics one of the things we deliberate over when making a decision? Ethical facts could be a subset of broader normative facts.

Beyond the naturalist’s criticism from the principle of parsimony, there are objections to the non-naturalist worldview. One objection is from epistemology: how do we get to know ethical facts if they’re non-natural? Unlike mathematics, there is no established way to get to the truth of ethics. Non-naturalists have answered this concern in various ways. One is to argue that we have an ethical intuition, a form of moral perception akin to our ability to perceive facts about the world. Another is to appeal to the role of reason in empirical science. It’s not enough to make observations, we have to weave these observations into a logically consistent worldview, a coherentist epistemology. If we use our reason combined with observations of the world, perhaps we can create a coherent web of ethical facts.

Another objection is from supervenience. When we say the ethical supervenes on the non-ethical, we mean that changes to the non-moral facts about a situation changes the ethical facts about that situation. Imagine that we were deciding the punishment for two criminals, and the non-ethical facts about both crimes were identical. If we punished one criminal with five years of imprisonment and let the other one off free, this decision would be arbitrary. The challenge here is to explain, if ethical facts are their own kind of fact, what is their relationship to non-ethical facts?

A final objection turns the argument from practical reasoning against itself. Consider the reasons that motivate people to act. Hume proposed the belief-desire model of motivation, where a belief and desire combined motivate you to act. For example, if I desire an apple and believe apples are in the kitchen, I will go to the kitchen to get an apple. Where do ethical judgements fit in here? If we state that ethical facts are non-natural facts, they seem to fall on the belief side of the divide. Some philosophers, known as motivational internalists, argue that ethical facts are in and of themselves enough to motivate someone to act. But if one accepts motivational internalism, why do beliefs about non-natural facts motivate us to act when beliefs about natural facts do not?


Opposed to moral realism are metaethical antirealists, who generally come in two flavors. The first kind of antirealism is the error theory, the argument that all ethical statements are based on an ontological error and that all ethical statements are actually false. A second kind of antirealism is fictionalism: the idea that ethics is convenient fiction. We know the sun doesn’t literally rise over the horizon, but it is sometimes useful to talk as if it does.

Error Theory

J.L. Mackie argued for error theory in Ethics: Inventing Right and Wrong. The first argument is from relativism, where he notes that moral disagreement is constant in the world. In these disagreements, either one side is right, and the other wrong, or there is no right/wrong at all. Mackie opts for the latter explanation, arguing that we approve of the ways of life in which we participate, and don’t approve of ones which we don’t participate in. Critics of this argument point out that being a realist-relativist is an option, and that Mackie’s argument doesn’t show that ethical discourse presupposes objective values and facts. Instead of being about objective facts, ethical statements could be an expression of commitment to our own, personal values.

Because of these objections, the argument Mackie makes that gets a lot more attention is the argument from queerness. Mackie argues that ethical values in ordinary discourse have two characteristics. One is that they are intrinsically motivating, meaning that recognizing that an action is good or bad would motivate us to do/not do it for no other reason than it being moral/immoral. Secondly, ethical facts are objectively prescriptive; they must generate reasons for us to do things that are independent of people’s particular desires. According to Mackie, torture must be bad irrespective of an agent’s desire not to be tortured; this ethical fact must give everyone an objective reason not to torture.

Mackie then argues that such facts don’t make sense. First of all, we’d need a special faculty to know such facts, which Mackie thought was implausible. Secondly, if intrinsically motivating and objectively prescriptive facts existed, they would be unlike any other fact we recognize; they have an ought built into them. But our modern scientific worldview tells us that such facts are, in his words, queer.

Objections to Mackie’s error theory generally object to how he constructs ethical facts as intrinsically motivating and objectively prescriptive. Some naturalists deny motivational internalism, and not all ethical judgments are motivating. One could judge, for example, that the Soviet Union invading Afghanistan in 1979 was wrong without being particularly motivated to do anything about it. There isn’t anything you can do about it now.

But that still leaves the argument that ethical facts are objectively prescriptive. Ethical facts may not motivate us to act but still provide us reasons to do things, known as justifying reasons. Internalists about justifying reasons argue that an agent can only justify an action if it connects directly to the agent’s desires or values. Externalists about justifying reasons argue the opposite: that an ethical fact could be a reason for us to act even if it doesn’t connect to an agent’s concerns. Mackie is arguing that ethics must provide objective, external, justifying reasons. Since he is an internalist about justifying reasons, he thinks such facts couldn’t exist. But one might argue that all that Mackie has shown is that externalism about justifying reasons is incorrect. It must have some connection to an agent’s desires in order to justify an agent’s action. Perhaps there are some general principles that, if we all lived by them, would do a better job promoting our concerns than acting only out of our particular concerns.


Another objection to error theory is the principle of charity. This principle means we should start out from the assumption that most of what people say is true, and if it seems false, we should reconsider if we understand what they are saying. A more sophisticated version of error theory that incorporates ordinary ethical discourse is fictionalism.

To understand the different kinds of fictionalism, let us examine the difference between declarative sentences and using them to make assertions. Consider the sentence:

The Starks live in Winterfell.

We’re not talking about a real family named the Starks or a real town named Winterfell; we are referring to the fictional Game of Thrones universe. There are two ways we could interpret this assertion:

  1. When you assert this sentence, you are expressing your belief that in the fiction of George R.R. Martin, the (fictional) Stark family lives in (fictional) Winterfell.
  2. You don’t really believe this sentence, and you are not asserting anything. Instead, you are engaging in a “pretense” of some sort.

The former is known as meaning fictionalism, and the latter as speech-act fictionalism. Now consider the sentence:

Murder is wrong.

According to meaning fictionalists, the actual assertion is that according to the fiction of morality, murder is wrong. So while the sentence is literally false, the assertion is true. A speech-act fictionalist might say that we are asserting something else - for example, that allowing murder would not be beneficial to us in general.

Philosopher Richard Joyce thinks that when we express moral assertions, we are not asserting universal moral truths but convenient rules of thumb. His reasoning is that it is generally beneficial for individuals to follow the rules of their community. But if people had to reason about these benefits every time they made a decision, they would frequently make the wrong decision. They might, for example, overestimate their ability to avoid getting caught or underestimate the costs of long-term lawlessness. Joyce thinks that to get individuals to do the beneficial thing, it is useful to have a fiction of objectively prescriptive moral values.

However, when most people make moral statements, they don’t think they are talking about a fiction. If we asked what fiction they were talking about when they say murder is wrong, they wouldn’t know what we were talking about. If fictionalists are to engage in the principle of charity, they need to explain why it doesn’t seem to most people they are engaging in pretense.


So far, we’ve assumed that ethical truths are a kind of fact about reality and that ethical thought is cognitive, or they are beliefs we have about the world. But perhaps we’re mistaken? Expressivists think that ethical discourse represents things more like desires or emotional reactions than beliefs about how reality is. Such a position allows us to hold onto a naturalist ontology while making sense of moral discourse. Motivational internalism also makes sense under expressivism: if morality is a desire, we don’t have to explain why morality motivates us to action when other facts don’t.

Earlier, I mentioned that one problem with non-natural realist metaethical theories is the problem of supervenience: why do ethical facts vary with non-ethical facts, if they are separate domains? R.M. Hare and Simon Blackburn think that expressivism explains supervenience. To them, supervenience is a constraint that our desires be consistent. Consider choosing apples at a grocery store. It would be odd indeed if we preferred apple A to apple B if both apples were the same in every respect. Ethical thoughts vary in a similar way: we must be consistent on pain of irrationality.


Emotivism, proposed by A.J. Ayer, argues that when we say “Stealing is wrong,” we are expressing a negative sentiment towards stealing rather than describing a specific property of stealing. This theory has four implications:

  1. Moral disagreement is nonsensical unless it takes place under shared values.
  2. There’s not much point to moral philosophy.
  3. Ethical statements are not truth-apt; they are neither true nor false.
  4. Ethical statements only have emotive meaning.

When Ayer argued there wasn’t much point to moral philosophy, you can imagine moral philosophers weren’t too impressed. And this isn’t the only view of expressivism out there. Some philosophers seek to be expressivists while at the same time taking morality seriously. Philosophers such as Alan Gibbard argue for a projectivist version of expressivism. In this theory, when we make ethical statements, we are expressing our commitment to certain values and norms. By converging on shared norms and values through ethical discourse, we find a way to live together. In this line of thought, ethics is still an expression of desires but still an important subject to consider.


The objection to expressivism is that despite trying to be more charitable to ethical discourse than error theory, they are claiming that the use of terms like “true” and “fact” in conjunction with ethical statements is nonsense. That still seems rather uncharitable. Consider the statements:

  1. Torture is always wrong.
  2. That’s true.

When 2 says “That’s true,” it doesn’t seem like they’re engaging in nonsense. Yet if we deny that ethical sentences are truth-apt, they are engaging in nonsense. How can this be resolved?

Simon Blackburn answers this objection with a concept known as quasi-realism, saying that the problem isn’t with the statements but with how philosophers characterize concepts such as “truth.” The assumption philosophers make here is the correspondence theory of truth, where a proposition is true if it corresponds with reality, and that ethical statements are representational; they purport to represent reality. A competing view is the minimalist theory of truth, where truth is a kind of linguistic device that allows us to talk about statements that we assent to. In this system, one needn’t have a theory how some sentences correspond to reality, it only requires that we consider the proposition p and p is true as equivalent. This allows Blackburn to claim that ethical statements are truth-apt even though they don’t purport to represent reality. This earns the expressivist the right to use terms such as “truth” without committing to realism.

Frege-Geach Problem

Another objection to expressivism is the Frege-Geach Problem. Let us assume that expressivism is true and examine the following syllogism:

  1. Tormenting cats is wrong.
  2. If tormenting cats is wrong, then getting little brother to torment cats is wrong.
  3. Hence, getting little brother to torment cats is wrong.

This is clearly a valid inference. But according to the expressivist, (1) has emotive meaning, and (2) is descriptive. One can endorse (2) without thinking that tormenting cats is wrong. But if that’s true, then this syllogism commits the fallacy of equivocation, where the word “wrong” is used to have different meanings in (1) and (2). The challenge to the expressivist is to explain why this syllogism is valid.

My Thoughts

Now that we’ve seen a survey of different metaethical views, I can talk a bit about my thoughts on why we should consider moral realism. As mentioned earlier, it is difficult to derive an ought from an is. Instead, we can ask a simpler question: If I only considered my own narrow interests, how would I make decisions? Earlier, I talked about an argument for normative facts from practical reasoning. These norms are our preferences - the things we like and dislike, weighted by how much we like and dislike them. Such facts would fit into a naturalistic ontology - facts about our preferences are natural facts.

What if we extended that to consider the preferences of others? Such a system could contain only these normative, natural, facts without appeal to non-natural facts. Such facts are also desires about how the world should be. This is why we should consider other’s preferences when acting: we share the world between all conscious beings.

A difficult problem is what kind of fact a statement like “We should consider the preferences of others” is. I don’t think it’s a natural fact, as it’s not something that could be investigated scientifically; it comes from our own practical reasoning. Is having one non-natural fact in our ontology a good idea? On one hand, this nicely sidesteps the problem of supervenience since there are ethical facts to vary based on non-ethical facts. On the other hand, having a single non-natural fact seems to violate the principle of parsimony, especially when I’m advertising that the rest of the facts are natural.

Is this argument for a preference-based morality convincing? Not a lot, but perhaps there is no argument that would truly convince anyone there is an ought if they think there isn’t one. That’s not to say learning metaethics is worthless. It can teach us a lot beyond grounding morality in first principles. It can give us a greater understanding of our own conception of morality, where it comes from, and what it is.

Further Reading

What is This Thing Called Metaethics?. By Matthew Chrisman. 2023.

Foundations of Ethics: An Anthology. Edited by Russ Shafer-Landau and Terence Cuneo. 2007.

Moral Realism. Internet Encyclopedia of Philosophy.

Metaethics. Stanford Encyclopedia of Philosophy.

Divine Command Theory. Internet Encyclopedia of Philosophy.

  1. This nomenclature comes from G.E. Moore, which is a little confusing. Non-natural facts intuitively sounds like it could include supernatural facts. 

  2. How Hume should be interpreted on this point is controversial. 

<< Previous Next >>