A Lacking Category in Meta-Ethics : Non-Reductive Subjectivism
Let's get conceptually and semantically richer
1. Introduction
Meta-ethics, a branch of philosophy that investigates the nature, status, and foundations of moral claims, primarily concerns itself with questions about whether moral facts exist, how they are known, and what kind of reality—if any—they possess. It differs from normative ethics, which focuses on what we should do, and applied ethics, which addresses specific moral dilemmas. Meta-ethics asks deeper, foundational questions, such as whether moral facts are objective or subjective, and whether moral properties can be reduced to non-moral properties like psychological states or natural facts.
Within this field, a dominant division has emerged between two broad camps: moral realism and moral anti-realism. However, as we will see, there is a surprising assumption that a sub-domain of anti-realism, subjectivism, must be tied to certain reductive views, leaving a conceptual gap for a position that could hold both subjectivist and non-reductive commitments.
2. The Major Meta-Ethical Views
To better understand the gap in meta-ethics, we will first outline the landscape by discussing the major views, beginning with the primary distinction between realism and anti-realism.
2.1 Moral Realism
Moral realism is the view that moral facts and properties exist independently of minds. Realists argue that moral claims, such as "Torturing innocent beings for fun is wrong," are true or false regardless of how individuals or societies feel about them. Realism can be further divided into two main subcategories:
Reductive Moral Realism: Reductivists argue that moral facts are reducible to other kinds of facts, often those concerning well-being, desires, or psychological states. According to this view, morality can be understood through empirical means, much like the natural sciences. For example, a naturalist might claim that "torturing is wrong" is true because it decreases overall well-being or violates certain desires shared by most rational beings.
Non-Reductive Moral Realism: Non-reductionists reject the reduction of moral facts to other kinds of facts. They argue that moral facts are irreducible and cannot be fully captured by science or empirical investigation. Instead, these facts are sui generis—unique in their nature—and can only be apprehended through moral intuition or reason. Philosophers like G.E. Moore, who argued that moral properties are "non-natural," belong to this camp.
2.2 Moral Anti-Realism
In contrast, moral anti-realism denies the existence of objective, mind-independent moral facts. For anti-realists, moral claims do not correspond to external truths. Within anti-realism, there are several notable positions:
Non-Cognitivism: Non-cognitivists argue that moral statements do not aim to describe the world and therefore cannot be true or false. Instead, moral claims are expressions of emotions or prescriptions. For example, when someone says "Torturing is wrong," they are not stating a fact but expressing disapproval or urging others not to engage in such behavior. A.J. Ayer's emotivism is a classic form of non-cognitivism.
Subjectivism: Subjectivists argue that moral claims can be true or false, but their truth depends on individual or cultural attitudes. According to this view, "Torturing is wrong" means something like "I disapprove of torturing," or "My culture disapproves of torturing," making moral facts relative to the subjects making the claim.
Error Theory: Error theorists, like J.L. Mackie, assert that although moral language purports to describe objective moral truths, such truths do not exist. Therefore, all moral claims are systematically false. For the error theorist, when someone says "Torturing is wrong," they are making a false claim because there are no moral facts that could make such statements true.
3. A Critique of the Assumption
A common assumption in meta-ethical discourse is that subjectivistm must be reductionist. It is often taken for granted that if one adopts a subjectivist stance, they must also hold a reductionist view of moral properties, reducing them to individual preferences or cultural norms. This assumption is unwarranted. Why should subjectivism, which holds that moral facts depend on the attitudes of individuals or groups, necessarily entail a reductionist account of moral properties ?
This assumption overlooks a potentially rich category: non-reductive subjectivism, where moral properties could be irreducible yet properties of the minds of sentient beings. Consider desires and beliefs: they are mental properties that are plausibly not reducible, but we do not posit that they exist independently of minds. Similarly, one could argue that moral properties—while not objective in the realist sense—are irreducible aspects of our mental lives. Moral judgments would be another kind of propositional attitudes (see also this).
This form of non-reductionist anti-realism could thus recognize the unique nature of moral facts without requiring them to exist independently of minds. It would acknowledge that moral properties are not analyzable in terms of other properties, yet still hold that these properties are contingent on the existence of subjects. In this way, moral properties would be irreducible, but not objective.
4. Conclusion
Meta-ethics as a field seems to have made an implicit assumption that subjectivism must entail some form of reductivism. However, this assumption neglects the possibility of a non-reductive anti-realist position—one that maintains the irreducibility of moral properties while denying their mind-independent status. This category, currently underexplored, would allow for a richer, more nuanced understanding of metaethics. It is time to recognize this gap in meta-ethical discourse and explore the potential of non-reductive anti-realism as a legitimate category.


I have an easy time imagining a non-reductive non-cognitivism. You could say "The mental state expressed by moral utterances doesn't reduce to emotions, desires, prescriptions, endorsements, acceptance of norms, or plans. It's just in that general family of things."
I'm less sure that I'm grasping what would be meant for a non-reductive subjectivism. If we're talking about propositional attitudes, what are the relevant truth-makers? Perhaps "murder is wrong" means "[I have the moral intuition that] murder is wrong," where moral intuitions are taken as mind-dependent and primitive?
It seems like there are two different propositions being claimed here. One is that subjectivism could be non-reductive. The other is that moral anti-realism could be non-reductive. The two are clearly separable and arguably are actually easily separated. Error theory claims that all moral claims are false, it does not reduce these claims to anything else so it is clearly non-reductive (by one sense of reduction one could mean) and also clearly an anti-realist moral theory.
If saying moral claims are wrong is reductive because that would change the nature of moral claims (we presume people making them thought some subset were true and so it changes the judged nature of such claims) then that implies that at minimum any change, complication or nuance to a conventional unreflective moral claim is reductive and so the only meta-moral theories that are going to be non-reductive are going to be completely quietest (add or subtract nothing from conventional unreflective unphilosophical making of moral claims). Subjectivism would also be reductive if it has anything to add or subtract about moral claims.
We could also get into semantic debates about what subjective and objective even mean. Some would claim that reductionist moral realist theories are subjective because they make what is moral depend on some mental property of actual minds and so on.
I think a more suggestive take away is that the meta-ethical debate has been framed such that what non-reductive moral realists add to the moral discourse is somehow not foreign enough to it count as being reductive and what reductive moral realists add while reductive in a technical sense maintains reference to some mind independent essence so as to render the claims still objective. Whereas all forms of anti-realism add something (or subtract something) from the moral discourse such as to change the essential target of sense and so are reductive in that sense (hence error theory is reductive because even though it does not change the technical definition of a moral claim it changes their essential sense) on top of that are often reductive in a more technical sense (posit that moral elements can be reduced to some other kinds of elements such as personal approval and disapproval). Arguably this tilts the situation in favour of the moral realist, in so far as conventional unreflective morality is plausible. It might be useful to thus shift the meta-ethical debate to one about how we should even frame things in the 1st place and this might usefully focus debate on more clearly discussable areas of disagreement. I do sense that a lot of debates between meta-ethicists consist in talking at cross purposes about what even the person on the street even means when they say "oy, that's unfair" and the plausibility or soundness of the utterance.