‘Moral naturalism’ is a term with a variety of meanings inethics, but it usually refers to the version of moral realismaccording to which moral facts are natural facts. That is the subjectof this entry.

Naturalistic approaches to ethics are as old as moral theory itself.Both Aristotelian and Confucian ethics, for instance, containrecognizably naturalistic elements. But moral naturalism wasn’tarticulated as a distinctive metaethical doctrine until 1903, withG.E. Moore’sPrincipia Ethica.Principiaestablished metaethics as a branch of moral theory distinct fromfirst-order ethics, and did so by arguing against moral naturalism asa metaethical doctrine. This rejection of naturalism shaped moraltheory for nearly a century. Beginning in the 1980s, however,metaethicists began developing new ways of articulating and defendingmoral naturalism. It is now one of the most popular views inmetaethics, perhaps the most popular.

This discussion will be organized in three parts. The first part, inSections 1–2, will characterize moral naturalism together withtwo other views that are closely associated with moral naturalism,descriptivism and reductivism. The second part, in Sections 3–4,will look at arguments for and against moral naturalism. And the thirdpart, in Sections 5–7, will examine the three most popularcomprehensive naturalist doctrines in detail.

• 1. What is Moral Naturalism?
• 2. Descriptivism and Reductivism
  • 2.1 Descriptivism
  • 2.2 Reductivism
• 3. Why be a Moral Naturalist?
  • 3.1 Arguments from Supervenience
  • 3.2 Arguments from Anti-Skepticism
• 4. Objections to Naturalism
  • 4.1 The Open Question Argument
  • 4.2 The Normativity and Triviality Objections
• 5. Neo-Aristotelian Naturalism
• 6. Cornell Realism
• 7. Jackson’s Analytic Functionalism
• Bibliography
• Academic Tools
• Other Internet Resources
• Related Entries

1. What is Moral Naturalism?

Moral naturalism is the view that moral facts are stance-independent,natural facts. It can also be characterized as the view that moralproperties are stance-independent, natural properties.

Fact Naturalism: Moral facts are natural facts.

Property Naturalism: Moral properties are naturalproperties.

These two views are subtly different because there are different kindsof moral facts. Particular facts are instantiations of some property;particular moral facts are instantiations of some moral property;particular natural facts are instantiations of some natural property.So Fact Naturalism entails Property Naturalism: if all moral facts arenatural, then all particular moral facts are instantiations of naturalmoral properties. But not all moral facts are particular facts.General moral facts, like the fact that genocide is wrong, consist ineither patterns of instantiations of moral properties or underlyinglawlike moral principles which explain those patterns ofinstantiations. Underlying moral principles do not consist ininstantiations of moral properties; it is possible that genocide iswrong but never committed, in which case there would be noinstantiation of a moral property of wrongness to correspond to theprinciple that genocide is wrong. So while Property Naturalism entailsthat allparticular moral facts are natural facts, PropertyNaturalism does not entail Fact Naturalism because it’s possiblethat moralprinciples are non-natural. For ease ofexposition, I’ll mostly ignore this complication in whatfollows, and treat Fact Naturalism and Property Naturalism asinterchangeable while focusing on Fact Naturalism as the strongerthesis. But moral principles will cause trouble at a few points.

A fact is stance-independent just in case it doesn’t depend onthe beliefs or attitudes of any agent. Moral realism is the view thatmoral principles are stance-independent; this is a more precise way ofcapturing the common idea that, for a realist, morality is a matter offact rather than a matter of opinion. But what does it mean for a fact(or property) to be natural? The usual answer is that naturalfacts/properties are the kinds of facts/properties that scientificmethods tell us about (see, e.g., Shafer-Landau 2003, Chapter 3; Smith1994, 17; Copp 2007, 27–28). To say that moral facts are naturalfacts, then, is to say that moral facts are part of the naturalisticpicture of the world that is revealed by empirical science.

Moral naturalism is opposed to moral supernaturalism, which holds thatmoral facts are a kind of supernatural (typically, divine) facts. Itis also opposed to moral non-naturalism, which holds that moral factsare neither natural nor supernatural, but are rather their own, thirdkind of thing. And as a version of moral realism, moral naturalism isopposed to all varieties of anti-realist views in metaethics,including constructivism, relativism, expressivism, and error theory.While anti-realists are usually naturalists in the sense that theywant to account for morality in naturalistic terms, anti-realists holdthat stance-independent moral facts can’t be reconciled with anaturalistic worldview. Moral naturalists hold that they can be.

For the naturalist, moral facts are the kinds of facts that can beinvestigated by (broadly) scientific methods. But what kinds of factscan be investigated by scientific methods? There is much lessagreement on how to answer this further question. One common view isthat facts can be empirically investigated just in case they arecapable of entering into causal explanations of our experiences, andthereby be investigated through those experiences (e.g., Bedke 2009;Enoch 2010; Lutz 2020). And while most (perhaps all) moral naturalistswould say that moral facts can feature in causal explanations, thismay not be a good characterization of natural facts, sincesupernatural facts could, perhaps, also feature in causal explanations(Sturgeon 2009; McPherson 2015). We could try to solve this problemwith brute force, by defining a natural fact as any fact that featuresin causal explanations that is not a supernatural fact. But this is adhoc; and now we need to characterize supernatural facts. We might alsoworry about what it means for a fact to feature in a causalexplanation, which is itself a vexed and controversial issue. Perhapsit is best, then, to just leave it at “natural facts are factsthat are known scientifically.”

But even this minimal characterization faces difficulties. For onething, it is extremely controversial which methods count as“scientific.” The “demarcation problem” ofdistinguishing science from non-science (or“pseudoscience”) is a perennial puzzle in the philosophyof science, with many philosophers doubting that there is any generalway of characterizing scientific methodology (Feyerabend 1975). Asecond problem is that this is a methodological or epistemologicalcharacterization of the natural; rather say what kinds of facts moralfactsare, moral naturalism would be a view about how moralfacts areknown. This is, in fact, how some characterizemoral naturalism and non-naturalism. InMoral Realism: ADefense, Russ Shafer-Landau affirms a number of metaphysicaltheses that are commonly associated with moral naturalism, mostnotably saying that moral facts are entirely constituted by morefundamental natural facts (Shafer-Landau 2003, 74–8). Yet hestill calls himself a non-naturalist, because he thinks that we havemoral knowledge by intuition and reflective equilibrium rather than byscientific methods (Shafer-Landau 2003, 56–58ff). Some criticshave argued that Shafer-Landau is actually a naturalist in virtue ofhis metaphysical commitments (Bedke 2012, fn 1). But if amethodological characterization of the natural is best, thenShafer-Landau has labeled his view correctly.

So far, we have characterized moral naturalism in terms of fourclaims:

Moral Realism: There are stance-independent moralfacts.

Fact Naturalism: Moral facts are natural facts.

Property Naturalism: Moral properties are naturalproperties.

Methodological Naturalism: We know about moral claimsin the same way that we know about claims in the natural sciences.

But moral naturalism is sometimes associated with a fifth,linguistic claim, about the nature of moral language. Thatclaim is:

Analytic Naturalism: Our moral claims aresynonymous with certain (highly complex) claims in thenatural sciences.

It might be tempting to say that moral naturalism is the conjunctionof these five claims, but there is a strong tension betweenMethodological Naturalism and Analytic Naturalism. If AnalyticNaturalism is true, then it should be possible to analyze moral claimsand show what claims in the natural sciences they are synonymous with.But if this is possible, then substantive moral principles areknowable not via scientific methods but via a priori conceptualanalysis. Analytic Naturalism therefore seems to entail thatMethodological Naturalism is false, and vice versa. Naturalists whoaccept Analytic Naturalism are called, appropriately enough, analyticnaturalists. Naturalists who reject Analytic Naturalism are syntheticnaturalists.

This distinction between analytic naturalism and synthetic naturalismcreates further trouble for our ability to characterize moralnaturalism. If moral naturalism is the view that moral facts are knownby empirical methods, then analytic naturalism is not a version ofnaturalism at all (Flanagan, Sarkissian, and Wong 2016). Yet manywould resist the suggestion that analytic naturalists aren’tnaturalists, since the term “moral naturalism” was coinedto refer to analytic naturalism; synthetic naturalism is a relativelyrecent entrant to the metaethical scene.

Perhaps we can resolve this conflict by recalling the differencebetween particular moral facts and moral principles. Analyticnaturalists agree with synthetic naturalists that we should be able toinvestigate particular moral facts empirically. An analytic naturalistwho held that ‘good’ was synonymous with‘pleasurable’ (e.g.) would agree that whether or not aparticular action instantiates goodness/pleasure is the sort of thingwe can investigate empirically. And synthetic naturalists agree withanalytic naturalists that certain claims about the relations betweenour moral concepts, like “if it is morally required to helpothers, then it is not morally forbidden to help others,” can beanalytic and knowable a priori (although see Howard and Laskowski2021). What is primarily at issue in the debate between analytic andsynthetic naturalists is the status of claims about moral principles.Analytic naturalists hold that these claims are analytic; syntheticnaturalists hold that they are not. So if we define moral naturalismin terms of Property Naturalism and remain neutral on Fact Naturalism,this sidesteps the question of the nature of moral principles; perhapsthis can mend the rift between analytic and synthetic naturalists. Butquestions about the nature of moral principles seem awfully close tothe metaphysics of morality, so perhaps we don’t want tosidestep them. And Property Naturalism doesn’t help answer theearlier worries we had about how to characterize moral naturalism:What properties are the kinds of properties that can be investigatedby scientific methods, and what methods for investigating thoseproperties are the scientific ones?

One more characterization of moral naturalism is worth mentioning. Itis increasingly popular to think that appeals to the notion of“grounding” can help clear up what is at stake in avariety of metaphysical debates, and so some believe that we canexplain naturalism in terms of grounding:

Grounded Naturalism: All moral facts are fullygrounded in non-normative facts (Rosen 2017a).

But it’s not clear if this draws the line between naturalism andnon-naturalism in the right place. One problem is that it’scontroversial what the relationship between “naturalfacts” and “non-normative facts” is (see Section 4.2below). Another problem is that “grounding” is itself ahighly contested notion. “Grounding” refers to some sortof asymmetric determination relation which is supposed to be“almost maximally intimate” (Rosen 2017a, 157). It is“almost” maximally intimate because the maximally intimaterelation is identity, but grounding is not identity; unlike identity,grounding is asymmetric. Non-naturalists would reject the idea thatmoral facts are identical to non-normative facts. But grounding isnot identity; so whycan’t non-naturalists saythat the relationship between the moral and the non-normative isalmost maximally intimate? Some self-describednon-naturalists say that moral facts are entirely grounded innon-normative facts (Moore 1942; Bader 2017; compare Leary ms –see Other Internet Resources). What’s wrong with that? What doesit mean for a relation to be “almost” maximally intimatein the first place?

Sometimes philosophers clarify the notion of grounding by saying thatthe grounded facts are “nothing over and above” thegrounding facts (Rosen 2017a, 157). But putting things this waythreatens to show too much. If moral facts are “nothing over andabove” the non-normative facts, then moral facts wouldapparently have to be non-normative facts; normativity seems to countas something “over and above” non-normativity. But manymoral naturalists say that moral facts are normative (again, seeSection 4.2). Perhaps it is coherent to say that the normative is“nothing over and above” the non-normative; many goodphilosophers speak this way and we shouldn’t write them off asconfused. But “nothing over and above” is no more clearthan “almost maximally intimate.” So we lack a clearpicture of what grounding consists in, and it’s consequentlyunclear why non-naturalists can’t say that moral facts are fullygrounded in non-normative facts but naturalists can.

We could say that naturalism is the view that moral facts are groundedinnatural facts rather than innon-normative facts(Chilovi and Wodak 2022). It’s (arguably) clearer why this isthe sort of thing that naturalists would want to say andnon-naturalists wouldn’t, but this formulation doesn’thelp us understand what natural facts are in the first place.

All moral naturalists will agree with the claim that morality isnatural. But there’s depressingly little agreement on how tounderstand the “naturalness” of morality. It’s all abit of a muddle.

2. Descriptivism and Reductivism

This section will look at two views closely associated withnaturalism, descriptivism and reductivism. Many philosophers who callthemselves descriptivists or reductivists also call themselvesnaturalists (or at least end up taking the same side as self-describednaturalists in many metaethical debates). Perhaps we can usefullycharacterize naturalism in terms of either of these views.

Naturalism as Descriptivism: Moral facts are naturalfacts if and only if (and because) descriptivism is true.

Naturalism as Reductivism: Moral facts are naturalfacts if and only if (and because) reductivism is true.

More modestly, even if Naturalism as Descriptivism and Naturalism asReductivism don’t pan out, we might prefer to just focus ondescriptivism or reductivism as interesting views that capture many ofthe same ideas that incline philosophers toward naturalism.

2.1 Descriptivism

Many philosophers have found it intuitive that there is a distinctionbetween terms like ‘good,’ ‘bad,’‘right,’ ‘wrong,’ or ‘reason,’ onthe one hand and terms like ‘pink,’ ‘shiny,’‘bigger than a bread box,’ or ‘pleasurable’ onthe other. The terms on the first list are callednormativeterms while the terms in the second list aredescriptiveterms. It is (mostly) uncontroversial that moral terms arenormative terms; so, if we refer to moral properties using moral terms(contra some expressivists and some error theorists), some normativeterms refer to moral properties.Moral descriptivism is theview that some descriptive terms refer to moral properties as well(Heathwood 2009).

Moral descriptivism is closely related to moral naturalism. Let usassume, as seems plausible, that the things that scientistsinvestigate are things that can be referred to with descriptivelanguage. Good science doesn’t moralize, it just says how thingsare. If this assumption is correct, then analytic naturalism entailsdescriptivism. If moral claims are synonymous with scientific claimsand scientific claims are made using descriptive language, then moralclaims are synonymous with descriptive claims. Moral terms refer tomoral properties; any synonymous terms would refer to those sameproperties; therefore, there are descriptive terms that refer to moralproperties.

Synthetic naturalists can be descriptivists, too. According tosynthetic descriptivism, there are descriptive terms that refer tomoral properties, but this is not so in virtue of any analyticrelation between normative and descriptive terminology. Rather, thesynthetic descriptivist argues for the existence of synthetic propertyidentity claims. The paradigm synthetic property identity is“water is H₂O.” ‘Water’ and‘H₂O’ are just two different labels for thesame thing. But ‘H₂O’ is not the definition of‘water;’ that water is H₂O was not, and couldnot have been, discovered by pure conceptual analysis. It is asynthetic truth known a posteriori. Synthetic descriptivists hold thatdescriptive terms which refer to moral properties reveal syntheticproperty identities, just as in the case of water andH₂O.

However, even though analytic naturalism entails descriptivism andsynthetic naturalism is consistent with descriptivism, Naturalism asDescriptivism is false. The first problem is that many supernaturalterms are descriptive. So it’s possible that moral propertiesare supernatural (e.g. divine) properties and descriptive terms referto those moral, supernatural properties. In this case, descriptivismwould be true and naturalism false (Sturgeon 2009).

Another serious problem for Naturalism as Descriptivism is that, whileanalytic naturalism entails descriptivism, synthetic naturalismdoesn’t. There’s no reason why facts that can beempirically investigated and fit comfortably with a scientificworldview must be referred to with descriptive terminology. Anaturalist can say that goodness is a natural property that is knownvia empirical methods and features in causal explanations, and theonly term that refers to goodness is ‘goodness.’ JonathanDancy calls this non-descriptivist naturalism “one-termnaturalism,” since one-term naturalists say that there is onlyone term – a moral term – that refers to moral properties(Dancy 2006).

Incidentally, because naturalism doesn’t entail descriptivism,one common challenge to naturalism fails: “You say that goodness(e.g.) is a natural property, but which natural property is it?”For a naturalist may reply: “Goodness is a natural property. Theonly natural property that goodness is identical to is itself:goodness.” This is a coherent and complete answer to thechallenge (Sturgeon 2003). The challenger seems to be demanding thatthe naturalist pick out the property of goodness using descriptiveterminology. This is an apt challenge to descriptivism but not tonaturalism. One-term naturalists are not descriptivists (Väyrynen2021).

But while Naturalism as Descriptivism seems like a non-starter,perhaps descriptivism is an interesting view in its own right. Wemight even prefer to talk about descriptivism rather than naturalismin light of our difficulties with characterizing naturalism.Unfortunately, there are also problems with characterizingdescriptivism.

One problem is that it’s actually rather easy to refer to moralproperties with descriptive terminology. Goodness was the primarysubject of G.E. Moore’sPrincipia Ethica. So the phrase‘the subject ofPrincipia Ethica’ refers togoodness, and that phrase uses only descriptive terminology. But thisisn’t a proof of descriptivism; descriptivists want somestronger relation between descriptive language and moral properties.But it’s hard to say what that stronger relation might be.

A second problem is the existence of “thick” normativeterminology, which apparently expresses both descriptive and normativeconcepts, or perhaps concepts that are both descriptive and normative.Common examples include ‘brave’ (and other virtue terms),‘cruel,’ or even, perhaps, ‘pleasurable’(Roberts 2017, 212). Frank Jackson (1998) proposes that thickterminology, and any other terminology about which there could bedoubt as to classification, should be considered normative. Perhapsthat solves the problem.

A third problem concerns our method for characterizing normative anddescriptive terminology. We started out with two contrasting lists ofterminology, saying that one of the lists was a list of normativeterminology and the other list, the “everything else”list, was a list of descriptive terminology. But our list of normativeterminology was incomplete. Normative terminology includes everythingon that list and all other terminology that is similar. But what doesall this normative terminology have in common in the first place? Whatmakes a term normative? One way of characterizing normativeterminology is broadly functional: normative terminology expressesthoughts which regulate our actions in certain ways. But embracing afunctional account of normative terminology cedes much ground toexpressivists, who hold that the meaning of normative terminology justis to express action-regulating states. And characterizing normativeterminology in functional terms also deprives moral realists (and thusmoral naturalists) of one of their main arguments for realism, sincethe existence of moral facts may no longer be read directly off of thesemantics of moral terms (Eklund 2017). For that reason, some realistsprefer to characterize normative terminology in referential terms:normative terminology is terminology that refers to normativeproperties (Hernandez and Laskowski 2021). But if we give areferential characterization of normative terminology, it’s nolonger obvious how to draw our distinction between normative anddescriptive terminology. If a bit of terminology that we wouldintuitively put on the list of descriptive terminology ends upreferring to a moral property (which happens, according todescriptivists!), then that terminology is ipso facto normative. So ona referential account, which terms count as normative and which countas descriptive is not discoverable by simple reflection on themeanings of the terms, but must instead involve substantial ethicaland metaethical theorizing. A referential account of normativeterminology would also threaten to render descriptivism incoherent,since any terminology which referred to a moral property would benormative rather than descriptive.

2.2 Reductivism

Another view that is often closely associated with naturalism is“reductivism.” The reductivist says that moral propertiesreduce to some other kind of property. Again, we might be interestedeither in Naturalism as Reductivism as an ambitious attempt tocharacterize moral naturalism, or just in reductivism as aninteresting view in its own right. Unfortunately, it is unclear whatit means for one property to reduce to another. ‘Reduce’is a philosopher’s term of art, but it is used in very differentways by different philosophers. We’re in for another muddle; butlet’s try to clear things up a bit.

We should note at the outset that there is a paradigm example of areductive view, which is used almost every time a metaethicistdiscussing reductivism needs a toy example to play with. That paradigmis the hedonic reduction: that goodnessis pleasure andthereforereduces to pleasure. This introduces a keydesideratum of any account of reduction: that it should count thehedonic reduction as a reductive view.

That said, let’s begin with one of the clearest things writtenabout reduction. Here’s Nicholas Sturgeon:

Naturalism is in one clear sense a ‘reductionist’ doctrineof course, for it holds that moral facts are nothing but naturalfacts. What I deny, however, is that from this metaphysical doctrineabout what sort of facts moral facts are, anything follows about thepossibility of reduction in another sense (to which I shall henceforthconfine the term) more familiar from the philosophical literature:that is, about whether moral expressions can be given reductivedefinitions in some distinctive nonmoral vocabulary, in which anyplausible moral explanations could then be recast (Sturgeon 1988,239–240).

Sturgeon here distinguishes two distinct senses of the term‘reduction.’ One is linguistic. A term can be reduced toanother term just in case the first term can be defined in terms ofthe second. In this sense, reductivism is just another name foranalytic naturalism. In later work, Sturgeon mentions the possibilityof “synthetic reductions” (Sturgeon 2009), so I thinkit’s more accurate to say that Sturgeon’s reductivism isanother name for descriptivism. Note that the hedonic reduction is adescriptivist view since it states that the descriptive term‘pleasure’ refers to the moral property of goodness. Sothe linguistic sense of ‘reduction’ as another name fordescriptivism satisfies our key desideratum.

Let’s pause for a second to clear up another perennial source ofconfusion about how metaethical terminology is used. Sturgeon and theother Cornell realists (see Section 6) call themselves“non-reductive” naturalists; but when they say this, theyaren’t describing any distinctive metaphysical commitments oftheir view. They are simply rejecting descriptivism. Sturgeon, atleast, is perfectly clear about this in the above quote. There aremetaphysical doctrines associated with Cornell realism, of course,(again, see Section 6). But as the Cornell realists used that term,“non-reductive” naturalism is just naturalism withoutdescriptivism, i.e. one-term naturalism.

If we understand ‘reduction’ in its linguistic sense,reductivism is synonymous with descriptivism, and so we have aredundant bit of terminology. That’s a good reason to understandreductivism as a distinct, metaphysical doctrine. So: what is ametaphysical reduction? We might look to general metaphysics for helphere. Unfortunately, metaphysicians use the term ‘reduce’in lots of different ways; there is no one standard account for us toborrow (Stoljar 2010, 161–162; McPherson 2015). But perhaps wecan get a sense of the general idea. In recent years, there has been aproliferation of terms within metaphysics that all seem to be aimed atcapturing the same idea: we can say what something is composed of, orwhat it is grounded in, or what it reduces to, or what the realdefinition of a thing is, and all of those locutions seem to be aimedat capturing the same thought (Rosen 2015). All of these locutions areways of articulatingwhat it is to be some thing by sayingwhat it is made of (compare Schroeder 2005; Schroeder 2016).One key idea here is that if A reduces to B, then A is lessfundamental than B (Laskowski 2020). Let’s call this theFundamentality Platitude.

If we understand reductivism as a metaphysical thesis aboutfundamentality, per the Fundamentality Platitude, we can make aplausible case for Naturalism as Reductivism. Any naturalist will saythat the fundamental entities are the sorts of objects and propertiesthat are the subject matter of fundamental physics: bosons and leptons(etc.) with their spin and mass (etc.). Goodness is not one of thosefundamental physical properties; goodness is a higher-level propertywhich is ultimately composed of fundamental physical properties, thesame as any other natural property. Naturalists are, in this way,metaphysical reductivists. And because non-naturalists deny that moralproperties are natural, their account of the nature of moralproperties is not constrained by the physicist’s account offundamental properties. A non-naturalist might well say that somemoral properties are metaphysically fundamental. So naturalists say,whereas non-naturalists may deny, that all moral properties reduce tonon-moral, physical properties. That’s Naturalism asReductivism.

However, there are problems with this argument. First, supernaturalproperties cause trouble again. If moral properties are reducible tosupernatural properties, then reductivism is true and naturalism false(Sturgeon 2009). As before, we could try to finesse this by talkingabout “reductive naturalism,” the view that moralproperties reduce to natural properties, in particular. Manyphilosophers do speak this way. But we can’t characterizenaturalism as the view that moral properties reduce to naturalproperties unless we already know what natural properties are. So talkof reductive naturalism means giving up on Naturalism asReductivism.

So let’s set aside the more ambitious Naturalism as Reductivismand just focus on reductivism as a potentially interesting andimportant view in its own right. Unfortunately, there are problemswith giving a useful metaphysical characterization of reductivism.Consider the hedonic reduction: that goodness reduces to pleasure invirtue of being identical to pleasure. Some philosophers defend theidea that pleasure is metaphysically fundamental. If we combinepleasure fundamentalism with the hedonic reduction, we end up with theview that goodness is both reducible and fundamental (McPherson andPlunkett forthcoming). So depending on what we say about themetaphysics of pleasure, the hedonic reduction may not be consistentwith the Fundamentality Platitude.

There’s a larger problem lurking here. We’ve characterizedtwo different ideas as key to contemporary discussions of reduction inmetaethics: the Fundamentality Platitude and the hedonic reduction.The hedonic reduction counts as a reductive view because of anotherplatitude about reduction: that identity entails reduction. Call thatthe Identity Platitude. Here’s Graham Oddie:

Exactly what is reduction? Take two kinds of entities, Type-A entitiesand Type-B entities. Although reduction is a contested notion, here isan undeniablysufficient condition for the reducibility oftype-A entities to type-B entities: every type-A entity isidentical to some type-B entity. (Oddie 2005, 15, italics inoriginal)

The larger problem is that the Identity Platitude and theFundamentality Platitude are inconsistent. Identity is a symmetricrelation, but being-less-fundamental-than is an asymmetric relation,so nothing can satisfy both of those platitudes. If reduction isasymmetric, then reduction is not entailed by identity, and so theIdentity Platitude is false. If reduction is not asymmetric, then thefact that A reduces to B doesn’t entail that A is lessfundamental than B, and so the Fundamentality Platitude is false.

Here’s another way to demonstrate the point. Consider a closecousin of the hedonic reduction, the view that rightness is identicalto conducing-to-pleasure. Conducing-to-pleasure is a complex property,composed of pleasure, conducing, and perhaps some abstract structuralelements. On this view, rightness is identical toconducing-to-pleasure, and rightness is composed of and lessfundamental than conducing and pleasure. So rightness stands in adifferent relation to the complex conducing-to-pleasure than it doesto the simpler elements. Which of those two relations is the reductionrelation? ‘Reduction’ is a term of art; it means whateverphilosophers mean by it. And, unfortunately, philosophers use the term‘reduction’ to refer to both of those relations. So notonly is ‘reduction’ ambiguous between a linguistic and ametaphysical concept, there’s a further ambiguity in themetaphysical use of the term.

Let’s sum up the discussion thus far. We were looking for an aptcharacterization of moral naturalism, and had a hard time finding one.We then looked at two ways of characterizing naturalism in terms ofother theses associated with naturalism: descriptivism andreductivism. We found that there are substantial problems withcharacterizing naturalism in terms of either descriptivism orreductivism, and that there are further problems with trying tocharacterize descriptivism and reductivism themselves. So whilephilosophers frequently use the terms ‘naturalism,’‘descriptivism,’ and ‘reductivism’ whendescribing various metaethical views, there’s no way to definethose terms that isn’t highly controversial. In light of thisdepressing and distressing realization, there’s only one thingto be done: ignore all of this and press on as though we know whatwe’re talking about when we use these terms. That is,unfortunately, what philosophers in this area tend to do. So eventhough there’s little agreement about what naturalism is,let’s see what philosophers have said about naturalism.

3. Why be a Moral Naturalist?

Moral naturalism is a very popular view (although it might be moreaccurate to say that “moral naturalism” is a popularlabel…). That’s because moral naturalism is theconjunction of two views that are themselves very popular. The firstis naturalism: that everything which exists is natural. The second ismoral realism: that stance-independent moral facts exist. Combinethese two, and we get the conclusion that moral facts (exist and) arestance-independent natural facts. Thus:

We’ll look at more technical arguments for moral naturalismshortly, but the Basic Argument is the reason why so manymetaethicists have a default sympathy for moral naturalism. It’sa popular conjunction of two popular views. Naturalism has proven tobe the most successful project, ever, for advancing human knowledgeand understanding. The success of naturalism doesn’t entail thatnothing exists which can’t be accounted for naturalistically,nor does it entail that there are no other projects that couldsuccessfully advance human knowledge and understanding. But manyphilosophers suspect that empirical methods have been so dramaticallysuccessful in shedding light on so many different areas of inquirybecause some version of naturalism is correct. (Non-naturalistsdisagree, of course!)

Similarly, moral realism is a popular view. There are many differentmoral claims – that rape and murder are wrong, for instance, orthat helping others in need is good – that have the status ofcommon sense. We treat these common sense moral claims, at leastimplicitly, as accurately describing reality and not being merematters of opinion. Moral realism is popular precisely because itpromises to do justice to this common attitude towards our moralcommitments. (Anti-realists disagree, of course!)

3.1 Arguments from Supervenience

Let’s turn to the more technical arguments for naturalism. Thereare two distinct arguments for moral naturalism that begin from thepremise that the moral supervenes on the natural.

Supervenience: There are no two metaphysicallypossible worlds where the natural facts are the same but the moralfacts are different.

As we saw in Section 1, any two worlds that are identical with respectto the moral facts are identical with respect to both moral principlesand instantiations of moral properties. So Supervenience entails therelated claim that no two metaphysically possible worlds have the samedistribution of natural properties and a different distribution ofmoral properties.

The first argument, the Direct Argument, comes from Frank Jackson(Jackson 1998, 118ff. See also Streumer 2008; Brown 2011; Clarke2019). Supervenience is the first premise of the Direct Argument. Thesecond is

Intensionalism: If it is metaphysically necessarythat something has propertyF if and only if it has propertyG, thenF andG are the same property.

Intensionalism and Supervenience together entail naturalism. Considerthe moral property of wrongness. Every wrong action has some naturalfeatures; call the features of a given wrong action \(N_1\).Supervenience entails that any action with \(N_1\) is necessarilywrong. Of course, there is more than one action that is wrong, but theargument generalizes. Call the natural features of a second wrongaction \(N_2\). Supervenience entails that any action with \(N_2\) isnecessarily wrong. And so on, for all natural features of wrongactions \(N_1 \ldots N_n\). Now consider the disjunctive property(\(N_1\) or \(N_2\) or … \(N_n\)). Let’s call thatdisjunctive property \(N\). Because \(N\) is a disjunction of naturalproperties, it is itself a natural property. Any action that has \(N\)is necessarily wrong. And because \(N\) is the disjunction of thenatural features of all morally wrong actions, any action that iswrong necessarily has \(N\). So, necessarily, an action is wrong ifand only if it has \(N\). Therefore, by Intensionalism, wrongness isidentical to \(N\). And because \(N\) is a natural property, wrongnessis a natural property. QED.

The most obvious way to object to the Direct Argument is to rejectIntensionalism (see, e.g., Bader 2017). A common example in the debateover Intensionalism is the pair of propertiesbeingtriangular andbeing trilateral. Necessarily, anythingthat has one property has the other: if a figure has three angles, italso has three sides (Bealer 1982). Enemies of Intensionalism say thetwo properties aren’t the same: one property is a propertyfigures have in virtue of the number of angles they possess, while theother is a property figures have in virtue of the number of sides theypossess. Defenders of Intensionalism say that these are the sameproperty: there’s no difference between being three-sided andbeing three-angled, and that’s precisely because thoseproperties are necessarily co-instantiated. This is an ongoing debatein metaphysics.

The Direct Argument also relies on the assumption that disjunctions ofnatural properties are natural properties, but that’s a dubiousassumption. Many would claim that disjunctions of propertiesaren’t properties at all, or at least not the right kind ofproperties (Kim 1993; Majors 2005; McPherson 2015; Klocksiem 2019).For other criticisms of the Direct Argument, see Wedgwood (2007, Ch.9) and Dunaway (2015).

The second argument from Supervenience is the Explanatory Argument.This argument was developed by RM Hare and Simon Blackburn (Hare 1952;Blackburn 1993), but the contemporary canonical formulation comes fromTristram McPherson (2011). According to the Explanatory Argument,supervenience is a striking phenomenon that must be explained. Aninability to explain supervenience counts strongly against a view.Naturalists are well-positioned to explain supervenience: if moralfacts are natural facts, then, trivially, there can be no differencein the moral facts without a difference in the natural facts. Butit’s not clear how other views – particularlynon-naturalism, but perhaps also quasi-realist expressivism (Sturgeon2009; Dreier 2015) – can explain supervenience.

The most obvious way to respond to the Explanatory Argument is tooffer an explanation of supervenience that doesn’t trade onnaturalist assumptions. For example, Russ Shafer-Landau (2003, Ch 4)and David Enoch (2011, Ch 6) have argued that supervenience can beexplained by the fact that there are moral laws. Ralf Bader (2017)argues that there is a distinct sort of normative grounding relationwhich both explains supervenience and doesn’t entail propertyidentity (but see Morton (2020)). And Stephanie Leary (2017) hasargued that there are hybrid normative properties whose essencesexplain supervenience (but see Faraci (2017); Toppinen (2018)).

A bolder response denies supervenience altogether. Supervenience iswidely accepted, but not uncontroversial (Roberts 2018). Sturgeon(2009) accepts supervenience, but his reasons for doing so are“parochial:” he is a naturalist and naturalism entailssupervenience. But Sturgeon sees no reason why anyone who’s nota naturalist should accept the doctrine. Several non-naturalists,including, most notably, Gideon Rosen (2017b) and Kit Fine (2002),have shared this assessment of the dialectical situation and rejectedSupervenience. Fine and Rosen argue that moral principles are notmetaphysically necessary but are, instead,normatively necessary, where normative necessity is weakerthan metaphysical necessity but still quite strong. It’s easyfor non-naturalists to explain the normative necessity of moralprinciples, so non-naturalists can explain everything that needs to beexplained. If Supervenience is false, then both the Direct Argumentand the Explanatory Argument fail.

3.2 Arguments from Anti-Skepticism

A second family of arguments for naturalism are epistemic. Scientificmethods (broadly construed) are very powerful tools for coming to haveknowledge of the world so, if moral facts are the kinds of facts thatwe can know about by scientific methods, then we have powerful toolsfor obtaining moral knowledge. Non-naturalists, by contrast, say thatwe know about moral facts via non-empirical methods, typicallyintuition. This seems like a much less firm footing for our moralknowledge than the scientific method. This is controversial, ofcourse. Intuitionist moral epistemology is popular (Huemer 2005;Chudnoff 2013; Bengson, Cuneo, and Shafer-Landau 2020). And it’snot clear how we can use empirical methods to learn about moral facts(although we’ll look at some proposals in Sections 5–7)(Shafer-Landau 2003, Ch 3). But it’s widely believed that moralnaturalists have an easier time answering moral skeptics thannon-naturalists, and that’s a reason to be a moralnaturalist.

Let’s sharpen this point by looking at one influential family ofskeptical challenges: explanationist challenges. Explanationistchallenges begin from the premise that our moral beliefs are explainedby something other than the moral facts. What else? Different versionsof the explanationist challenge focus on different things. Accordingto the traditional version of the challenge from Gilbert Harman, ourmoral beliefs are explained by our personal psychological attitudes(Harman 1977, Ch 1). Other versions of the explanationist challengefocus on more distant causes of our moral beliefs. Disagreementarguments note that moral beliefs tend to cluster along culturallines, which suggests that our moral beliefs are explained by culturalpressures (Mackie 1977, 37). Evolutionary debunking arguments positthat our moral beliefs are ultimately explained by evolutionarypressures (Joyce 2006; Street 2006). The common thought is that ourmoral beliefs are not explained by the moral facts; they’reexplained by a tangle of other naturalistic factors. Exactly why thiswould lead to a skeptical conclusion is controversial, but manyphilosophers think that explanationist challenges make a strong casefor skepticism. (For critical discussion, see, e.g., Enoch 2009;FitzPatrick 2015; Vavova 2015; Schechter 2017).

Naturalists have an easy way to answer explanationist challenges. Thenaturalist claims our moral beliefs are indeed explained by a tangleof naturalist facts, but they hold that some of those natural factsare the moral facts. For instance, Sturgeon argues that we thinkHitler was evil because Hitler was evil. Hitler’s evil is anatural feature of his character, and that feature both explains whyHitler did the things that he did and why we judge him harshly as aresult (Sturgeon 1988). Similarly, David Copp argues that to be goodjust is to be conducive to social cooperation. So when theevolutionary debunker claims that we are naturally inclined, byevolution, to think that things which are conducive to socialcooperation are good, this just amounts to the observation that wejudge that things are good because they’re good (Copp 2008).This is a compelling enough response that many advocates ofexplanationist challenges construe them as challenges fornon-naturalists, but not for naturalists (Joyce 2006; Bedke 2009; Lutz2018).

4. Objections to Naturalism

We’ve just seen that there is some reason to think that, ifthere are any moral properties, those properties are natural. In thissection, we’ll look at the most prominent objections to moralnaturalism.

4.1 The Open Question Argument

I said at the beginning of the article that contemporary metaethicsbegan with G.E. Moore’s rejection of moral naturalism. Theargument that he deployed against naturalism is known as the OpenQuestion Argument (OQA). The OQA is perhaps the most influentialargument in metaethics, although today it is widely held to havecrippling flaws. Inevitably, this has inspired many critics ofnaturalism to find ways to fix up or extend Moore’s argument inone way or another.

Explaining Moore’s argument is complicated by the fact that hisoriginal statement of the argument, in Sections 5–15(particularly 13–14) ofPrincipia Ethica, is famouslyunclear. Moore himself admitted as much, in an unfinished forward to asecond edition (Sinclair 2019). I won’t engage in Moore exegesishere. I’ll just give a version of the OQA that most contemporarymetaethicists would recognize as “the Open QuestionArgument.”

Moore’s OQA takes aim at theories of the form “Everythingthat isN is good,” where ‘N’ is abit of descriptive terminology that refers to a natural property, andwhere claims like this are true in virtue of the meanings of‘good’ and ‘N.’ Moore is concernedwith goodness because he regards goodness as the most fundamentalmoral concept, and holds that all moral inquiry is, in the end,inquiry into what sorts of things are good. And he’s concernedwith theories that are true in virtue of meaning because he believesthat there are true claims of the form “Everything that isN is good” (for some potentially rather complicatedN). He just denies that it is analytic that beingNis the same thing as being good. “If I am right, then nobody canfoist upon us such an axiom as that ‘Pleasure is the onlygood’ or that ‘The good is the desired’ on thepretense that ‘this is the very meaning of theword.’” (Moore 1903, Sec 6.)

The OQA relies on a test for theories of the form “AllAs areBs.” The first step is to turn thestatement of the theory into a question: “AllAs areBs” becomes “Given thatx isA,isx alsoB?” Some questions of this kind areclosed questions: one can know the answer simply by understanding thequestion. “Given thatx is a bachelor, isxunmarried?” is a closed question. Anyone who understands themeanings of ‘bachelor’ and ‘unmarried’ is in aposition to see that the answer to that question must be yes. Anyonewho expressed doubt about that question must be incompetent witheither the concept ofbachelor or ofunmarried. So“All bachelors are unmarried” passes the test. That claimis true in virtue of the meanings of ‘bachelor’ and‘unmarried,’ and so being unmarried is part of the essenceof what it is to be a bachelor. Now consider the theory that pleasureis goodness. We can ask “Granted thatx is pleasurable,isx good?” And this question is open. Someone canunderstand it perfectly and still reasonably doubt the answer. So thetheory that pleasure is goodness fails the open question test;pleasure cannot be identical to goodness. The argument generalizes:For any natural propertyN and any moral propertyM,“Granted thatx isN, isxM?” will be an open question. Therefore, no moralpropertyM is identical to any natural propertyN.

The most obvious problem with this argument is that it’s notreally an argument against naturalism at all. It is, instead, anargument against analytic descriptivism. That is, it’s anargument against the claim that there is some descriptive term‘N’ that necessarily refers to a moral propertyM in virtue of the fact that there is a normative term‘M’ that refers toM, and‘N’ follows from a conceptual analysis of‘M’ (Feldman 2005; Vessel 2020). But as we saw inSection 2.1, analytic descriptivism is not the same thing asnaturalism. Some naturalists are synthetic descriptivists, who holdthat ‘N’ necessarily refers toM, butthat this is established empirically rather than by conceptualanalysis (Brink 1989, Ch 6; Brink 2001). And one-term naturalists holdthat there is no descriptive term that necessarily refers to any moralproperty (Sturgeon 2003; Väyrynen 2021). The OQA has no forceagainst either of these views. For these reasons, the common verdicttoday among metaethicists is that – unless it can besubstantially modified or extended in some way – the OQA has noforce against naturalism.

In Moore’s defense, even though he overlooked the possibility ofsynthetic descriptivism and one-term naturalism, those views were notyet developed at the time he first outlined the OQA. Indeed, they weredeveloped by naturalist philosophers who recognized the force of theOQA against analytic descriptivism and so set out looking for othernaturalistic views that could evade the OQA. Rather than thinking ofthese other naturalistic views as an oversight of Moore’s, wecould just as easily give him credit for pointing others in theirdirection.

Most philosophers think that the OQA does succeed in refuting analyticdescriptivism, yet analytic descriptivists have developed severalresponses to the OQA. One traditional response, from William Frankena(1939), says that the OQA begs the question. To an analyticdescriptivist who sincerely holds that, e.g., pleasure is thedefinition of goodness, the question “Given that X ispleasurable, is it good?” will seem closed. So whether thatquestion is open or closed is precisely what is at issue, so Mooreshouldn’t claim that it is open without further argument.Another traditional response appeals to the idea of the Paradox ofAnalysis: if all analytic claims are obvious, then it’simpossible for there to be an interesting, informative conceptualanalysis. Since some conceptual analyses are interesting andinformative (we philosophers tell ourselves, desperately), not allanalytic claims are obvious. The OQA is a test of obvious analyticity,but the correct definitions of moral terms might just be non-obvious(Smith 1994, 37–39). Another response begins from theobservation that Moore gives only a handful of examples ofnaturalistic analyses and generalizes to the conclusion that anyproposed analysis of moral terms must be wrong. That’s anawfully hasty generalization. It’s not clear that this worry canbe addressed by providing any finite number of additional failedanalyses. If ‘goodness’ (e.g.) can be analyzed, it canpresumably be analyzed in one correct way, and all of the infiniteother possible analyses are false. So the hypothesis that‘goodness’ can be analyzed predicts that there are aninfinite number of incorrect analyses of ‘goodness.’ Ahypothesis can’t be refuted by data that it predicts, so even aninfinite number of incorrect analyses doesn’t refute analyticdescriptivism (Finlay 2014, Ch 1).

4.2 The Normativity and Triviality Objections

Although Moore’s version of the Open Question Argument today hasfew defenders, there have been a number of recent attempts torefashion it into a more compelling form. One popular version of theOpen Question Argument, called the Normativity Objection (see, e.g.,Parfit 2011; Scanlon 2014), sidesteps questions about the cognitivesignificance of moral and descriptive terminology and appeals toconsiderations regarding the natures of natural and normative facts.Moral facts are normative. They concern what is good and what we havereasons or obligations to do. Natural facts – the kinds of factsthat scientists study – are facts about the innate physicalstructure of the universe and the causal principles that govern theinteraction of matter. Those are obviously just two different kinds offacts; moral facts are normative but natural facts are not. In tryingto give a naturalistic account of morality, naturalists forgot themost important thing: that moral facts aren’t purely facts aboutthe way the worldis; they are facts about whatmatters.

There are, generally, two ways in which a naturalist might respond tothis objection. First, a naturalist could say that moral factsaren’t essentially normative; it may be the case that wetypically have reasons to act morally, but reason-giving force is notpart of the essence of moral facts. That suggestion might have thefeeling of an absurdity – of course moral facts are the kinds ofthings that provide reasons; if an action is morally required,that’s a good reason to do it! But according to some“reforming definitions” of morality (Brandt 1979; Railton1986), this thought is a defect in our conception of morality. Itwould be more accurate and fruitful to define moral facts in termsthat aretypically reason-providing but notnecessarily reason-providing.

The biggest problem for reforming definitions is that reason-givingnormativityis (arguably) essential to morality. As Joyce(2000, Ch. 1) argues, normativity is a non-negotiable commitment ofour moral discourse. If the only kinds of facts that exist arenatural, and natural facts are not a source of reasons for everyone,then this amounts to a proof that moral facts don’t exist (seealso Luco 2016). Reformers may respond by calling into questionJoyce’s notion of “non-negotiable” commitments. Itis controversial whether any of our concepts have commitments that arenon-negotiable in this way (compare Prinzing 2017). Reformers may alsoreply by pointing out that they’re not entirely giving up onnormativity. If moral facts are the kinds of facts that we typicallyhave reason to care about, and there are strong forces in humansociety and psychology that lead us to care about moral facts, thenthey’re normative enough for all practical purposes.

The second way to respond to the Normativity Objection is to say thatnormativity is itself a natural phenomenon. The most popular strategyfor substantiating natural normativity proceeds in two steps. First,show that all normative concepts can be analyzed in terms of one,fundamental normative concept. Second, show that fundamental normativeconcept picks out a natural property. There are a number of ways thatsuch an account could proceed: here are two recent, influentialexamples:

• Mark Schroeder (2005; 2007) accepts the “buck-passing”or “reasons first” account of normativity (Scanlon 1998),which says that all normative concepts can be analyzed in terms of theconcept of a reason. Schroeder also accepts the “Humean”theory of reasons (as a substantive, synthetic truth), which saysthat, roughly, S has a reason to Φ just in case Φ-ing willpromote the satisfaction of one of S’s desires. If the Humeantheory is correct, then being a reason is a natural property. And, ifall other moral facts are to be analyzed in terms of reasons, then allmoral facts are natural facts.
• Phillipa Foot (2001) accepts a “value first” accountof normativity, which says that goodness is the fundamental normativeconcept. She also accepts a neo-Aristotelian account of goodness (seeSection 5 below for more on neo-Aristotelianism.) Ifneo-Aristotelianism is correct, goodness is a natural property. And ifall moral facts are to be analyzed in terms of goodness, then allmoral facts are natural facts.

This two-step strategy is popular but controversial. Those moved bythe Open Question Argument and the Normativity Objection are skepticalthat the second step of the naturalizing strategy could ever becompleted. Non-naturalists doubt that it could ever be shown that thefundamental normative concept picks out some natural property becausenormative properties and natural properties just seem to obviously bedifferent kinds of properties. Wittgenstein claimed to “seeclearly, as it were in a flash of light, not only that no descriptionthat I can think of would do to describe what I mean by absolutevalue, but that I would reject every significant description thatanybody could possibly suggest, ab initio, on the ground of itssignificance” (Wittgenstein 1965). David Enoch (2011) is morepithy, saying simply that natural properties and normative propertiesarejust too different for any natural account of afundamental normative property to be satisfying.

But it’s unclear what this “just too different”intuition amounts to (Laskowski 2019). In what sense are the naturaland the normative “just too different?” Schroeder thinksthat the intuition has force only if there is “a perfectlygeneral single truth about which any reductive view would be forcedinto error” (Schroeder 2005, 14). The existence of such a truthwould be the proof that the natural and normative are just toodifferent; without such a truth, the non-naturalist hasn’treally offered an argument against naturalism. But because allnormative claims conceptually reduce to claims about reasons, arguesSchroeder, there will be no such truth, provided that we have acoherent account of the foundational notion of a reason. So a coherentaccount of normative reasons can explain all normative claims. Thesefundamental normative facts about reasons are themselves explained bythe natural facts to which reasons reduce. But because this reductionof reasons will take the form of a synthetic reduction, there end upbeing no conceptual connections between the normative and the natural.This general lack of conceptual connections between the normative andthe natural is what explains the “just too different”intuition, and it is fully compatible with moral naturalism as asynthetic metaphysical truth (see also Copp 2020).

Schroeder’s response may succeed in providing anaturalist-friendly explanation of the “just toodifferent” intuition, but it seems strange to say that theexistence of numerous analytic relations between different normativeconcepts, combined with a lack of analytic relations between normativeand natural concepts, supports the thesis that the normative isnatural. The lack of conceptual connections between the natural andnormative may be better explained by non-naturalism (Enoch 2011).

Derek Parfit’s Triviality Objection (Parfit 2011) is anothercontemporary extension of the Open Question Argument. If moralnaturalism is true, says Parfit, then it will be possible to makemoral claims and natural claims and have those two claims contain allthe same information. And a statement of equivalence between any twoclaims that contain the same information must be trivial. But moralclaims that describe the relationships between moral facts and naturalfacts are not trivial at all – they are highly substantive. Somoral naturalism is false. However, it is not clear that thisTriviality Objection is any more forceful than the Open QuestionArgument. Parfit’s central motivating thoughts are thatnatural-moral identity claims are substantive rather than trivial, andtherefore moral claims concern a different kind of fact than naturalclaims. These are exactly the central thoughts behind Moore’sOpen Question Argument, and so naturalists to respond to thisobjection in largely the same way that they respond to the OpenQuestion Argument. Naturalists typically respond by adopting a versionof synthetic naturalism, and arguing that moral-natural identities areinformative in the same way that other synthetic property identityclaims (like water = H₂O) are (Copp 2017).

Having discussed moral naturalism in general terms, we’ll nowexamine three of the most popular naturalistic metaethical views indetail. As we go, it will be worth keeping the normativity objectionin mind. A version of the normativity objection will end up being oneof the most pressing objections to all of these views.

5. Neo-Aristotelian Naturalism

Aristotelianism has had adherents since… well, since Aristotle.Even in the dark times of the mid-20th century, when naturalism was indisrepute, post-Moore, Aristotelians like Elizabeth Anscombe kept theflame alive (Anscombe 1958). Neo-Aristotelianism has both ethical andmetaethical commitments. We’ll primarily be looking at themetaethical aspects of neo-Aristotelian naturalism, but first-orderethics will make an appearance as well.

We can characterize Aristotelian naturalism by four major ideas. Thefirst is that there is no one property of goodness. The goodness of agood toaster is different from the goodness of a good movie, which isdifferent from the goodness of a good time to dance, which isdifferent from the goodness of a good person. To be a good thing is tobe a good thing of its kind. The second is that this fact is mirroredin our language. The semantics for goodness is attributive rather thanpredicative (more on this shortly). The third is that the mostfundamental ethical concept is the concept of a good person. Goodactions are the actions a good person will do, good things are thethings that good people will desire or pursue, and good practicalreasoning is reasoning in the way that good people reason. And thefourth is that the standards that make something good are derivablefrom the nature of the kind of thing it is. So what makes someone agood person is derivable from the kind of thing a human person is.This implies that morality is derivable from biology – althoughthis biological conception of human goodness is quite controversial,even among Aristotelians.

Let’s start by looking what Aristotelians say about thesemantics of ‘good.’ Contemporary Aristotelians followPeter Geach (1956) in distinguishing between two different kinds ofadjective: attributive and predicative. The difference between the twocomes from whether they validate the following inference pattern:

1. X is anFG.
2. X is anH.
3. Therefore,X is anFH.

IfF is a predicative adjective, that inference will bevalid. IfF is an attributive adjective, then it won’t.To make this concrete, consider the following instances of thisschema:

4. Jerry is a brown mouse.
5. Jerry is an animal.
6. Therefore, Jerry is a brown animal.

4–6 is valid. If Jerry is a brown mouse, then he’s a brownanimal. Therefore, ‘brown’ is a predicative adjective.Predicative adjectives refer to the same property in all contexts.Compare this to:

7. Jerry is a large mouse.
8. Jerry is an animal.
9. Therefore, Jerry is a large animal.

7–9 is not valid. A large mouse is still a very small animal.Therefore, ‘large’ is an attributive adjective. Themeaning of an attributive adjective in any given context is determinedby the noun that it is modifying. “Large mouse” means(something like) “Large for a mouse.”

With this in mind, consider the following argument:

10. Frank is a good tennis player.
11. Frank is a husband and father.
12. Frank is a good husband and father.

This is clearly invalid. Frank might have a powerful serve and awicked backhand, yet neglect his family. This indicates that‘good’ is attributive (although see Szabó 2001).‘Good’ does not refer to the same property of goodness inall contexts. Instead, ‘good K’ means something like‘good-for-a-K.’

Judith Jarvis Thomson has influentially expanded on Geach’sbasic idea. According to Thomson, there is a property of being a goodK only ifK is a goodness-fixing kind (Thomson 2008,21). For instance, “knife” is a goodness-fixing kind;thus, there is the property of being a good knife. Goodness need notalways be understood in terms of goodness-fixing kinds. Things can begood in other respects, such as “good for use in makingcheesecake” or “good for Alfred” (Thomson 1997,278). Thomson calls things that are good in this way“good-modified.” There is no such thing as “a goodsmudge,” full stop, because “smudge” is not agoodness-fixing kind in the way “knife” is. But theremight be a smudge that is good for using in a Rorschach test, because“being used in a Rorschach test” is a way that smudges canbe good-modified (Thomson 2008, 21–22).

Thomson holds that it makes no sense to ask whether an act is good ornot, because “act”, like “smudge”, is not agoodness-fixing kind. Accordingly, acts can only be good-modified;they can be good in some respect – e.g., a moral respect. Actsare morally good when they spring from an agent’s morallyvirtuous traits. A moral virtue is any trait such that an individualof a kind K is as morally good as a K can be only if it has that trait(and it’s possible for to either have or lack that trait)(Thomson 2008, 79).

It’s quite plausible that goodness functions like an attributiveadjective sometimes: a good tennis player might be a bad father. Butthere might also be a predicative use of the word ‘good.’For instance, consider the idea of a “good state ofaffairs.” Thomson argues that this notion is incoherent becausea state of affairs isn’t a goodness-fixing kind; she takes thisto be a strong challenge to consequentialism (Thomson 2008, 62). Yetif the notion of a good state of affairs is intelligible – andit does seem to be – then perhaps ‘good’ sometimesserves as a predicative adjective that refers to a property of (moral)goodness (Sturgeon 2010).

But let’s follow the Aristotelian in saying that the term‘good’ is attributive and that the central ethical use of‘good’ is to describe a good human person. So: what makesa person as good as a person can be? The fourth major Aristotelianidea is that the standards of goodness for a thing are determined bythe kind of thing it is. But how can a thing’skinddetermine the standards for that thing? The traditional answer, fromAristotle, says that kinds of things have a particular form, or shape.The form of hammers is just their hammer-shape. The form of a thingdetermines its function: the function of a hammer is to hammer becausehammers are shaped in a way that makes them good for hammering. Andfunction thus determines thetelos (purpose or end) of thething: hammers are for hammering, and a good hammer is one thathammers well. All of this can be understood in unmysterious,naturalistic terms. We can apply this same model to human beings. Theform of a human determines what the specific human function is. And agood person is, therefore, someone who performs that function well bysatisfying the characteristic ends of human life.

The most influential account of human ends comes from RosalindHursthouse. Hursthouse says that human life is characterized by fourends: survival, reproduction, characteristic enjoyment and freedomfrom pain, and the good functioning of the group. Survival andreproduction are common to all living things, including plants. Butthe manner in which these ends should be pursued are characteristic ofthe species. Penguins have one characteristic way of reproducing andlions have another (Hursthouse 1999, Ch 9). The characteristicallyhuman way of pursuing these four human ends is by the use ofrationality. Humans alone among the animals act not from mereinstinct, but from a rational capacity of deliberation and choice. Sogood, “flourishing” humans are those who pursue the fourends in accordance with reason. Like many other neo-Aristotelians,Hursthouse uses the word “flourishing” as an Englishtranslation of Aristotle’seudaimonia, since it nicelycaptures the analogy with other kinds of living organism that iscentral to Aristotelian naturalism. But ‘flourishing’might not be the best word, since it may misleadingly imply anuntroubled life (Foot 2001, Ch 6).

A common objection at this point is that the notion of a purpose ortelos, while central to Aristotle’s conception ofbiology, is more controversial in a Darwinian biological paradigm(Millum 2006; compare Moosavi 2022a). Now, many philosophers ofbiology will still be happy to say that the purpose of the heart is topump blood. But in a Darwinian paradigm, the purpose of a part of anorganism is the role that part plays that contributes to the survivaland reproduction of that organism (in the environment in which itevolved). So to say that “the purpose of the heart is to pumpblood” is just to say (something like) “pumping blood isthe activity of the heart that contributes to the survival of theorganism.” Accordingly, Darwinian teleology seems ill-suited toplay any role in ethics. It doesn’t really apply to organisms asa whole, and it contains not the slightest whiff of evaluativesignificance.

Because of this, some neo-Aristotelians prefer to avoid (or at leastde-emphasize) the idea of a human telos and instead talk in terms ofwhat Michael Thompson has called “Aristoteliancategoricals” (Thompson 2008, Ch 4). These are claims like like“Oak trees have deep, sturdy roots” or “Humans have32 teeth.” Aristotelian categoricals are not statisticalregularities. They needn’t be true of all, or even most, membersof a species; most humans have less than 32 teeth (Anscombe 1958).Rather, they describe something important about the nature, or thecharacteristic form of life, of different kinds of living things(McDowell 1998). They say what “should” be the case for amember of a species, or what will be true of an organism insofar as itis flourishing.

Aristotelian categoricals allow us to evaluate organisms in anaturalistic way. An oak tree with deep, sturdy roots is (at least inthis respect) flourishing, while one without deep, sturdy roots isdefective. Similarly, then, a good person is flourishing insofar asthey participate in characteristically human ways of life, and badpeople are defective in the same sense that an oak without deep,sturdy roots is defective. “A bad tree.” The onlydifference is that humans and oak trees participate in very differentways of life. The aptness of the comparison is better illustrated bylooking at other social animals, like wolves. Wolves hunt in packs;that is their characteristic way of life. A wolf that hunts in a packflourishes, while a wolf that hunts by itself is defective. A goodlife for a wolf is partly a matter of cooperating with others of itskind to achieve further ends which are constitutive of a flourishinglife (Foot 2001, 16). So, too, with humans.

But whether we evaluate human organisms in terms of a telos or acharacteristic form of life, it’s not clear how these kinds ofevaluations are relevant to ethics. Even if there is a characteristicway of human life, why should we live that way ourselves? What isnormative about “natural goodness” (Prinz 2009)? Living acharacteristic human life might be a very bad way to live. There aresome characteristic ways of human life that seem to be quite bad: forexample, destroying the environment, fighting ideological wars, andhaving a sense of altruism that is limited to a relatively smallcircle of family and friends. Aristotelians seem to assume a ratherrosy picture of human nature that’s hard to square with aserious study of history or sociology (Williams 1972; Millgram2009).

Neo-Aristotelians typically answer this worry by appealing to the roleof reason (Foot 2001, Ch. 2; Hacker-Wright 2009; Lott 2012; Jordan2020). The supreme human virtue that characterizes a flourishing humanlife isphronesis, the virtue of being able to reason wellabout how to act (Hursthouse 1999, Ch 4). So fighting ideological warsis not characteristic of human life because it’s not reasonableto fight ideological wars; that’s not how a good person, onewithphronesis, would act. The characteristic ways of lifefor any given thing are the virtues of that thing; so thedistinctively human virtues are the virtues that are particular tohuman life, which is a social life governed by rational deliberation.And those are just the traditional virtues of moral philosophy:courage, humility, generosity, and so on (Foot 2001, Ch 5). Humanswith those virtues are, in that respect, flourishing. Those that lackthe virtues are defective. This way of operationalizingAristotelianism throws out the idea that discerning natural goodnessis a matter of applied biology (Hursthouse 1999, 178–191). Thecharacteristic form of human life is defined ethically, notbiologically (Nussbaum 1995). This is also, perhaps, the way to savethe notion of atelos. The distinctive form of humanity isrationality, and so the purpose of life determined by that form isactivity in accordance with reason, i.e. ethical living. So humanteleology has nothing to do with biology, and everything to do withethics (McDowell 1998, Thompson 2008).

One might well worry that this abandons the naturalistic character ofAristotelianism. Aristotelians typically reply that to be a goodperson is just to be a certain sort of human organism with certainmental characteristics (which amounts to having a brain that isfunctionally organized in a certain way). This all constitutes anaturalistic account of human goodness. So (recalling our distinctionsfrom Section 1), while neo-Aristotelians generally accept PropertyNaturalism and Fact Naturalism, many end up rejecting MethodologicalNaturalism (Hursthouse 1999, Ch 10; Moosavi 2022b). John McDowell hasargued that the human virtues are “natural” only insofaras they relate to the “nature” of human beings, and humannature is not a subject for scientific investigation (McDowell 1998).Micah Lott has argued that neo-Aristotelians should give up onpretensions to naturalism (Lott 2020).

Three further problems linger. First, appealing to practical reason tosolve the problem of normativity seems to undermine the coreAristotelian idea that we can define a good person in terms of what isdistinctive of human life. Kant argued that Aristotle was wrong toground ethics in the form of a human life, since what matters forethics is the form of pure practical reason. ContemporaryAristotelianism seems to concede this point to Kant (Woodcock 2018;but see Jordan 2020 or McKracken 2021). Second, neo-Aristotelianismlooks circular. One of the main ideas of Aristotelianism is that thegoodness of a good person is the fundamental moral concept, but if wedefine a good person in ethical terms, we’ve gone in a circle.Brown (2016) responds that a good person is defined in teleologicalrather than ethical terms, but this just raises the question of how wedefine teleology. Biological definitions seem inadequate, butnormative definitions make the account circular once again. Third,even if the circularity concerns can be answered, we might worry thata larger epistemological problem remains. How can we identify virtuousagents? Abandoning Methodological Naturalism might create moreproblems than it solves (but see Jordan 2016).

6. Cornell Realism

If moral naturalism is the view that moral facts are the kinds offacts that can be investigated in a broadly scientific way, then noview captures the spirit of naturalism better than Cornell realism.Cornell realism was developed in the 1980s by Richard Boyd (1988),David Brink (1986), Nicholas Sturgeon (1985), and Peter Railton(1986a; 1986b); the view gets its name from the fact that Boyd, Brink,and Sturgeon were working or studying at Cornell University at thetime. It is a comprehensive metaethical system, with interrelatedlinguistic, metaphysical, and epistemological commitments, thatdeliberately mirrors scientific methodology in ethics as closely aspossible.

It might seem odd to suggest that we can know things about morality byusing scientific methodology. Scientific methods are all, ultimately,grounded in an epistemology of observation. And as Gilbert Harman(1977, Ch. 1) famously argued, it does not seem that we can observemoral facts in anything like the same way that we can observe otherkinds of natural facts. It’s rather obvious how we can haveempirical knowledge of natural properties such as redness orroundness; they are directly observable. But goodness doesn’tseem to be directly observable, and that looks like an importantdisanalogy between moral properties and natural properties.

But not all natural properties are directly observable. Some kinds ofnatural properties are complex, and knowable only through thefunctional role they occupy. Consider, for instance, the property ofbeing healthy. Being healthy isn’t like being red; there’sno one way that healthy people look. Of course, there may be somecharacteristic visual signs of healthiness – rosy cheeks, aspring in one’s step – but these visual signs are neithernecessary nor sufficient for healthiness. These directly observableproperties are only indications of healthiness. Healthiness is acomplex natural property, wholly constituted by an organism’sbody being in the “proper” configuration, with a robustcausal profile. There are many things that can cause or impede healthby their presence or absence: food, water, disease, etc. And there aremany things that will result from health in typical circumstances:energy, long life, etc. Rosy cheeks and a spring in one’s stepare indications of healthiness because these are properties that are– typically – caused by health. Our awareness of thecausal profile of healthiness thus gives us a way of figuring outwhich things have or do not have the complex property ofhealthiness.

Cornell realists hold that goodness is exactly like healthiness in allof these ways (Boyd 1988). Like healthiness, goodness is a complexnatural property that is not directly observable, but nonetheless hasa robust causal profile. Like “healthiness,”“goodness” is not synonymous with any simpler empiricaldescription. Instead, “goodness” describes thefunctionally complex natural property that is the effect of certaincharacteristic causes, and the cause of certain characteristiceffects. Many different things contribute to or detract from goodness– things like pleasure or pain, honesty or untruthfulness– and there are many things that will result from goodness intypical circumstances – things like human flourishing, orpolitical peace. Because goodness is a natural property with a complexcausal profile, the property of goodness can enter into explanatoryrelations. Thus, contra Harman, it is possible for goodness to explainour observations (Sturgeon 1985). We can, accordingly, observe whethersomething is good by looking for indications of goodness. This isexactly the same way that we observe whether something is healthy.

Another way that goodness is like healthiness is that both aremultiply realizable. We could, perhaps, give a more or less completedescription of what it takes for a human body to be healthy, althoughsuch a description might be quite long and complicated. But it’salso possible that aliens with a very different biology exist, and forthose aliens to be healthy, even though a healthy alien body isnothing like a healthy human body. But despite their radical physicaldifferences, both a healthy alien and a healthy human are healthy. Thesame might be true of goodness, or other moral properties. As anatural property, any particular instance of goodness is fullyconstituted by a complex of more fundamental physical properties. Butgoodness might be constituted by an indefinitely large number ofradically distinct physical configurations: the goodness of a promisekept is quite different from the goodness of a delicious andnutritious meal. This motivates the Cornell realists to callthemselves non-reductive naturalists (by which they mean that they areone-term naturalists rather than descriptivists). Goodness isrealizable in too many physical structures to give a usable definitionof morality in descriptive terms.

The great asset of Cornell realism is that it directly adopts widelyaccepted views about the nature of natural properties and scientificknowledge in order to answer the foundational questions of moralmetaphysics and moral epistemology. What are moral properties? Highlycomplex natural properties, multiply realizable in more fundamentalstructures, and individuated by their causal profiles. Are there,generally, properties like this? Yes; these are known as“homeostatic cluster properties.” Healthiness is one;moral properties are properties like that. How do we know about moralproperties? By looking for directly observable properties that arecharacteristically functionally upstream or downstream from the moralproperty that we are interested in (provided that we have justifiedbackground beliefs about the functional roles of moral properties). Dowe, generally, have knowledge like this? Yes: this is how we havescientific knowledge; moral knowledge is knowledge like that. In thisway, the theoretical resources of scientific realism also turn out tosupport moral realism (Boyd 1988).

A skeptic might object that it’s impossible to have justifiedbackground beliefs about the functional profile of moral properties.But this objection would prove too much. We are entitled to rely onbackground beliefs in moral theory development because we’reentitled to do so in science, generally (Boyd 1988, 189–191).Our theories and background beliefs are justified together, by theiroverall coherence and empirical adequacy, in both science andethics.

It is not essential to Cornell realism that goodness be identifiedwith any particular complex natural property – different Cornellrealists have different first-order normative commitments. The mostinfluential of these accounts is due to Railton (1986a; 1986b).Railton, like Thomson, holds that moral goodness is defined in termsof what is non-morally good for agents. Whereas Thomson, as aneo-Aristotelian, defines what is good for a human in terms of humanbiology, Railton defines non-moral goodness for an agent as what afully-informed counterpart would advise us to desire, or, (perhaps)equivalently, what a fully-informed counterpart of ourselves woulddesire if they were in our actual position (see also Brandt 1979;Smith 1994). To illustrate, Railton asks us to imagine a traveler,Lonnie, who feels terrible because he is badly dehydrated. Lonnie doesnot know that he is dehydrated, and so is not taking appropriate stepsto make himself feel better. But imagine a fully-informed version ofLonnie – “Lonnie-Plus” – who knows about hisdehydration and knows that drinking clear liquids will make him feelbetter. Lonnie-Plus, who (like Lonnie) desires to feel better but who(unlike Lonnie) knows the best means to that end, would recommend thatLonnie drink clear fluids, and would drink clear fluids himself if hewere in Lonnie’s position. The fact that Lonnie-Plus wouldadvise Lonnie to want to to drink clear fluids means that drinkingclear fluids is good for Lonnie. This is not a relativist view ofgoodness, because the fact that Lonnie-Plus would choose to drinkclear liquids is determined by Lonnie’s circumstances andconstitution, and facts about Lonnie’s circumstances andconstitution are objective facts. In general: the complex naturalproperty of being good for an agent is identical to the complexnatural property that agent’s fully-informed counterpart wouldchoose. This makes goodness “objective, though relational”(Railton 1986b, 167).

This account is highly controversial. Railton’s case ofLonnie-Plus presents a counterpart who knows just a few more salientbits of information than Lonnie. But a counterpart who is trulyfully-informed will knowvastly more than weourselves do, and would, accordingly, be quite different from us. Afully-informed counterpart might be suicidally depressed, or madeneurotic by their full awareness of all the oddities and dangers inthe world, or might in some other way have been driven quite mad bythe process of becoming omniscient. It’s far from clear what afully informed person would desire, and, given that fully-informedindividuals are so radically different from ourselves, it’s notclear why we should think that the advice of such an agent would haveanything to do with our own personal well-being (Loeb 1995). It mightalso be impossible to be fully-informed in this way. It’splausible that our knowledge affects the nature of our experiences,and so a “fully-informed” agent would be incapable ofknowing what it is like to have experiences in a condition ofignorance, and thus could not be truly fully informed (Sobel 1994;Rosati 1995). We might also worry that full information accounts getthings backwards; advice attends to our interests, our interestsdon’t attend to advice, even ideal advice (see Risberg 2018 andcitations therein). Railton heads off this objection by denying thatpersonal gooddepends on an ideal advisor’s advice:“the existence of an individual’s objective interest canexplain why his ideally informed self would pick out for hisless-informed self a given objectified subjective interest, but notvice versa” (Railton 1986b, fn 17). This answers theworry, but leaves us somewhat adrift. If the advice of afully-informed counterpart doesn’t make a certain complex ofnatural properties count as good for an agent, whatdoes?

With this account of personal, non-moral goodness in hand, Railtonthen defines a moral standard as a standard that takes allagents’ interests equally into account. To act rightly is to actin accordance with this standard. Railton’s moral standard is“consequentialist, aggregative, and maximizing, [but it] is notequivalent to classical utilitarianism,” since Railton does notassume that personal, non-moral goodness consists in happiness(Railton 1986, fn. 31). This moral standard has explanatory power; forinstance, if a society does not take the interests of all intoaccount, it might be more prone to revolution, and so a revolution inan immoral society might be explained by that immorality. Because themoral standard is one that takes the interests of all into account,the right thing for an agent to do may fail to serve her interests.Assuming that practical reasons are grounded solely in self-interest(as Railton does), people who don’t care about the moralstandard might have no reason to act morally. As we saw in Section4.2, Railton accepts a “reforming definition” of moralitythat sacrifices the intrinsic normativity of moral standards. As wealso saw in Section 4.2, many consider this to be an unacceptableconsequence of the view.

There may, in practice, be substantial overlap between themetaphysical commitments of neo-Aristotelianism and Cornell realism.Cornell realists say that the good is a certain higher-order naturalproperty. Neo-Aristotelians say that goodness has something to do withhuman flourishing, which is also a higher-order natural property. BothCornell realists (Boyd) and neo-Aristotelians (Thomson) have found theanalogy with healthiness illuminating when explaining the nature ofmoral properties (see also Bloomfield 2001). So Cornell realists andneo-Aristotelians do tend to be similar in the way that they conceiveof normative properties. But they differ with respect to what they sayabout language.

While neo-Aristotelians favor an attributive semantics for‘good,’ Cornell realists accept causal regulationsemantics for moral terms, which is the view that moral terms refer towhatever property causally regulates their use. Adopting causalregulation semantics is a sensible thing for the Cornell realists todo, for two reasons. First, it continues their foundational commitmentto treating moral properties as natural properties that we know andtalk about in the same way that we know and talk about other naturalproperties. Causal regulation semantics are the standard semanticaccount for natural kind terms. And second, it helps them evade theOpen Question Argument (Brink 2001). According to causal regulationsemantics, moral terms cannot be defined in any verbal way. Theysimply refer to the (complex higher-order natural) property thatcausally regulates their use. This makes Cornell realism a form ofsynthetic naturalism. As we saw in 1.2, Moore’s Open QuestionArgument shows, at most, that analytic descriptivism is false.

But the Cornell realist’s semantics is also the source of themost influential objection to Cornell realism. According to thisobjection – Horgan and Timmons’s Moral Twin EarthObjection (Horgan and Timmons 1991) – we do not use moral termsin the way that the Cornell realist predicts. To understand the MoralTwin Earth Objection, we need to first understand how causalregulation semantics are supposed to work. The following thoughtexperiment, from Putnam (1975), has been highly influential: Imagine aworld – Twin Earth – where there is no H₂O, butthere is another substance called XYZ. This substance XYZ, whiledistinct from H₂O, fills the rivers and lakes, is clear andtasteless, etc. XYZ even has the property of being called‘water’ – when residents of Twin Earth fill up aglass with XYZ from the tap, they will say “I have a cup ofwater.” Yet for all this, we would not say that Twin Earth is aplanet where water is XYZ. We would say that this is a planet wherethere is no water. There is, instead, another substance – XYZ– that plays the same functional role. Yet when the person onTwin Earth fills a cup from the tap and declares “I have a cupof water”, we shouldn’t say they’re mistaken. Whatcould account for that? Putnam argues that the word‘water’ just means something different on Twin Earth thanit does in the actual world. In the actual world, ‘water’means H₂O because, in the actual world, H₂O isthe stuff that causally regulates our use of the term‘water.’ But on Twin Earth, ‘water’ means XYZbecause, on Twin Earth, it is XYZ that causally regulates TwinEarthlings use of the term ‘water.’ That is why the word‘water’ literally means something different on Twin Earththan it does on Earth.

According to the Moral Twin Earth Objection, things don’t workthis way for moral terms. Imagine a world – Moral Twin Earth– that is exactly like the actual world, except on Moral TwinEarth, people’s use of moral terminology is causally regulatedby different properties from those in the actual world. People usemoral terminology to praise and blame and to guide action, but theytake different kinds of actions to be worthy of praise or blame andthey guide their actions in different ways. Thus, if causal regulationsemantics is true for moral terms, the word ‘right’literally means something different on Moral Twin Earth than it doesin the real world – this is the intuition that drivesPutnam’s original Twin Earth case. But we do not have the samejudgment about how people use moral language on Moral Twin Earth! Ifpeople on Moral Twin Earth take different actions to be worthy ofpraise or blame, we don’t conclude that our words‘right’ and ‘wrong’ mean different things. Weconclude that there is a substantive moral disagreement between thedenizens of Moral Twin Earth and the people in our world, and thatsuch disagreement is possible only because we and the Twin Earthersmean the same thing by our moral terms. Horgan and Timmons argue thatthis proves that moral terms like ‘right’ and‘wrong’ do not refer to whatever it is that causallyregulates their use. So much the worse for Cornell realism. See Copp(2000), Dowell (2016), Dunaway and McPherson (2016), and Väyrynen(2018) for recent criticisms of this argument. Wisdom (2021)criticizes the whole debate on the grounds that our intuitions aboutTwin Earth and Moral Twin Earth are not stable enough to draw anymeaningful conclusions one way or the other.

The deep problem here is that causal regulation semantics implies akind of relativism. Indeed, some relativists, like Prinz (2007) andWong (2006), have argued that their view follows from the Cornellrealists’ empirical methodology and moral semantics. If wordsrefer to whatever causally regulates their use, and the same word isregulated by different things in different communities, then that wordrefers to different things in different communities. So if‘wrong’ refers to whatever causally regulates its use, andif two communities have a word ‘wrong’ that is causallyregulated by different things, then ‘wrong’ meansdifferent things in those communities. This saddles Cornell realistswith a classical problem for relativism. If ‘wrong’ justmeans something like ‘whatever my community disapprovesof,’ as the relativist says, then “My communitydisapproves of this act, but they shouldn’t, becausethere’s nothing wrong with it at all” isself-contradictory. But that’s not self-contradictory; so muchthe worse for relativism (see, e.g., Schroeder 2010, Ch 1). Similarly,if words refer to whatever causally regulates their use within acommunity, then “This action regulates the use of the term‘wrong’ in my community, but it shouldn’t, sincethere’s nothing wrong about that action” would beself-contradictory. But that’s not self-contradictory; so muchthe worse for Cornell realism (Sinhababu 2019).

One might think that Cornell realists can respond by offering a newaccount of moral language. But this response faces difficulties on twofronts. First, causal regulation semantics is the Cornellrealists’ way of avoiding a commitment to analytic naturalismand the accompanying Open Question Argument. If the causal referencetheory is false for moral terms, the naturalist may be forced toaccept a theory of moral semantics that would re-introduce the OpenQuestion Argument. Second, the causal reference theory for morallanguage is essential to the Cornell realists’ methodologicaland epistemological commitments. If moral goodness is the thing thatregulates our use of moral terminology, then the terms of ourempirical investigation are simple: we just investigate what regulatesthe use of our moral language. But if the moral facts are not(necessarily) the facts that regulate the use of our moral language,then we need a new methodology for moral investigation.

Because of this, the Moral Twin Earth objection is often presented asa kind of extension of Moore’s Open Question Argument. The OpenQuestion Argument is, in Moore’s formulation, an attack on theidea that there can be analytic natural-normative property identityclaims. But in a larger sense, the Open Question Argument picks out akind of epistemic shortcoming for moral naturalists. Moral naturalistsare committed to the idea that moral facts are a kind of natural fact– but which natural facts are the moral facts? More simply, howdo we know good from bad? If moral claims are synonymous with certainnatural claims, we can know by conceptual analysis which things aregood. And if moral facts are the facts that causally regulate our useof moral terminology, that provides us with another way ofinvestigating which facts are the moral facts. But if neither of thosestories is available, it seems that we have no way of identifyingwhich natural facts are the moral ones (Huemer 2005, chapter 4; Bedke2012). This would be a serious methodological problem for moralnaturalists, and would also mark a critical disanalogy between moralepistemology and scientific epistemology. This disanalogy would becrippling to Cornell realists, who hold that there is no suchdisanalogy.

7. Jackson’s Analytic Functionalism

The final version of naturalism that we’ll look at is FrankJackson’s moral functionalism. Jackson is an analyticdescriptivist. While his is not the only contemporary version ofanalytic descriptivism (see Smith 1994; Finlay 2014; Rawlette 2020),it is the most influential. This view is typically described asJackson’s, and the canonical statement of the view comes fromhis bookFrom Metaphysics to Ethics, though many of the ideasin that book were developed with Philip Pettit (Jackson and Pettit1995).

As we saw in Section 3.1, Jackson endorses the Direct Argument fromsupervenience to naturalism. Because moral facts supervene on naturalfacts, they are identical to natural facts. (Jackson talks about“descriptive” facts rather than “natural”facts, but I’ll continue to use “natural” for thesake of consistency.) The problem now is to say which natural factsthe moral facts supervene on and are therefore identical to. Jacksoncalls this the “location problem.” The location problemisn’t unique to ethics. Jackson expresses sympathy forphysicalism, the view that we could, in principle, give a completeaccount of everything that exists in terms of the kinds of things thatphysics does (or could) tell us about. Everything else that existssupervenes on (and is therefore identical to) complexes of thesefundamental physical things. But which complexes of fundamentalphysical things are the various non-fundamental things identical to?Location problems abound. Solving them is the task of “seriousmetaphysics” (Jackson 1998, 1–8).

So how can we solve location problems? What is the method of seriousmetaphysics? The first step is conceptual analysis. If we want to knowwhat the Fs are, we need to have some idea of what we’re talkingabout when we’re talking about “the Fs,” and thisunderstanding will frame our subsequent investigation. The first stepin figuring out that water is H₂O was understanding thatwater is the clear, tasteless stuff that fills the rivers and lakesand falls from the sky as rain, rather than the intoxicating stufffound in beer, wine, and liquor. If you thought that water was theintoxicating stuff in wine, you’d end up drawing a verydifferent – and incorrect – conclusion about thecomposition of water (Jackson 1998, 28–31ff).

But wait! Putnam’s Twin Earth thought experiment is supposed toshow that we can’t discover the meaning of ‘water’by a process of conceptual analysis; the people on Twin Earth have thesame concept of ‘water’ as we do, but their word meanssomething different. Jackson responds that Putnam and his followershave drawn the wrong lesson from the Twin Earth case. The Twin Earththought experiment doesn’t show us that the concept of watercan’t be analyzed; it shows that the concept can be analyzed,since the Twin Earth thought experiment is itself a nice bit ofconceptual analysis. Twin Earth – and all other thoughtexperiments – are nothing more than ways to elicit intuitionsabout how to deploy our concepts in a variety of counterfactualcircumstances. The intuitions that we elicit from thought experimentsare the stuff that conceptual analyses are made of. Putnam’sconclusion that ‘water’ refers to whatever causallyregulates the use of the term ‘water’ within a linguisticcommunity is itself a partial analysis of the concept of water(Jackson 1998, 39–79). Of course, merely having a conceptualanalysis like this in hand is not the end of our investigation. Nowthat we know what we’re looking for, empirical investigationtakes over and we investigate what the stuff is that causallyregulates the use of the term ‘water’ within ourlinguistic community. It turns out that that stuff is H₂O.That’s how we solved the location problem for‘water.’

Now we have a general outline for how to solve location problems.First, consider cases where we would, on reflection, deploy someconcept (e.g. ‘water.’) The result of this reflection willbe some set of platitudes about that concept. Next, we turn this setof platitudes into a “Ramsey sentence.” Take everyinstance of the word ‘water’ (or whatever), and replace itwith a variable, X. So instead of “Water falls from the sky asrain,” we will have a collection of sentences that say thingslike “X falls from the sky as rain.” This collection ofRamsey sentences is a description of a functional role; in this case,the water-role. Finally, we empirically investigate what, if anything,fulfills that role. We know that water is H₂O because wefound, through empirical investigation, that H₂O is thestuff that satisfies the water-role. In general, we know what Fs areby using conceptual analysis to figure out what the F-role is, andthen empirically investigating what, if anything, satisfies the F-role(Jackson 1998, Ch 2–3).

We can perform this process to solve the location problem in ethics.Begin with some moral concept, say “goodness.” Thenconsider a variety of cases where we would judge that something isgood. The result will be some set of platitudes about goodness. We canthen take those platitudes about goodness and turn them into a Ramseysentence which describes the goodness-role. Finally, we investigatewhat, if anything, satisfies the goodness-role. That thing is goodness(Jackson 1998, Ch 5–6).

What about multiple realizability? Jackson is attempting to provide aconceptual analysis of our moral concepts, but the Cornell realiststhought that no such analysis is possible in part because goodness(e.g.) is realizable in a large number of very different things.Jackson is unconcerned by this. If goodness supervenes on a largenumber of different descriptive properties, then goodness is identicalto the disjunction of all of those descriptive properties on which itsupervenes. Goodness might be identical to a disjunction –perhaps an infinitely large disjunction! – of other naturalproperties (see also Sinhababu 2018).

Many have found the suggestion that goodness is identical to aninfinite disjunction of natural properties to be metaphysically andepistemologically suspicious. Can an infinite disjunction of naturalproperties really be a property? And how could we have knowledge of anugly gerrymandered property? Jackson replies thatexcessivelydisjunctive properties are suspect, because they cannot play a causalrole (Jackson 1998, 106). But not all disjunctive properties –even infinitely disjunctive properties – are excessivelydisjunctive. Consider the property of baldness. Baldness supervenes on(and is therefore identical to) having a certain distribution of hairon one’s head. Well-known vagueness concerns about the predicate‘bald’ may prevent us from giving a more compactdescription of what it takes for someone to be bald. But there is somelarge – perhaps infinitely large – set of hairdistributions which we would, on reflection, definitely consider to bebald. Baldness is therefore identical to an infinitely largedisjunction of hair distributions. But that doesn’t mean thatthere’s anything metaphysically or epistemically weird aboutbeing bald. Many people are bald, many claims to the effect thatsomeone is bald are true, and in most cases it is not particularlydifficult to ascertain whether or not someone is bald. Jackson urgesus to say precisely the same thing about moral properties (124).

A second worry: When we assemble the platitudes about our concept of“goodness,” we will find that many of them implicate othermoral concepts. “If something is good, then you have a reason topursue it.” “Virtuous character traits are good.”And so on. So we can’t analyze the concept of goodness intoplatitudes and then investigate what satisfies those platitudes, sincemany of those platitudes will concern moral concepts that are equallyin need of analysis. Jackson’s solution to this problem is toanalyze all of the moral concepts together. Take all of the platitudesconcerning goodness or value or wrongness or virtue, etc. Replaceevery instance of “good” with a variable (g), replaceevery instance of “value” with another variable (v),replace every instance of “wrong” with another variable(w), etc. The result is the functional role description of thenetwork of our moral concepts. Goodness and value andwrongness are then the g, the v, and the w (etc.) that satisfy all ofthe platitudes concerning g, v, and w (etc.).

A final worry: Which platitudes are we to select as the platitudesthat appear in our analysis? The ones that we endorse on reflection?But who is the “we” who would endorse these platitudes onreflection? How much reflection must we undergo? Jackson answers thisworry by appeal to the concept of amature folk morality. Afolk morality is “the network of moral opinions, intuitions,principles and concepts whose mastery is part and parcel of having asense of what is right and wrong, and of being able to engage inmeaningful debate about what ought to be done” (130). This folkmorality is then revised by hunting out inconsistencies withinone’s moral commitments and engaging in moral discussion withothers, identifying disagreements, and hammering out any differences.The resulting mature moral theory will be both individually andinterpersonally coherent. That’s the final statement of theview: The moral properties are whatever satisfies the network analysisof the platitudes that make up mature folk morality.

Jackson’s appeal to a mature folk morality is his main wardagainst the Open Question Argument. To ask “X ispleasurable, but isX good?” is, in effect, to askwhether “Pleasure is good” is one of the platitudes thatwill feature in our mature folk morality. But we don’t have amature folk morality yet – we may not have one for a very longtime – and we have no way of knowing at this point whatplatitudes will feature in it. Of course the question remainsopen.

The fact that we don’t know what the mature folk morality lookslike also shows that Jackson’s idea of an infinitely disjunctiveproperty isn’t a core commitment of the view. If our mature folkmorality is a vast, heterogeneous collection of principles and cases,then only an infinitely disjunctive property could satisfy it. But ourmature folk morality could well be simple and compact. Whether we endup with a compact formulation or an infinite disjunction depends onthe outcome of a process we have not yet seen the end of.

Now to some harder objections. One objection holds thatJackson’s view isn’t really a version of analyticnaturalism, because the process of refining our mature folk moralitywill be the result of not just moral discussion and reflection, butalso our observations of the real world (Lane 2018). Relatedly, onemight worry whether the platitudes of mature folk morality are reallyconceptual truths. Someone might disagree with a mature consensus notas a result of conceptual confusion but as a matter of genuine moraldisagreement (Zangwill 2000).

A second objection concerns how we should understand“mature” folk morality. We don’t want any endpointof moral theorizing to count as “mature;” if mature folkmorality is just the theory we end up with in the future, whateverthat might be, then why should we care about it? Maybe people in thefuture will have the wrong moral beliefs! But if we define maturity innormative terms, as the outcome of a good process of deliberation (orsomething similar), Jackson hasn’t really succeeded in providingthe reductive naturalistic account of normativity that he was hopingfor (Yablo 2000).

We might also worry that no mature folk morality will ever emerge. Ifmoral disagreement persists into the indefinite future, just as it haspersisted throughout human history, there is no mature folk moralityand thus no set of platitudes that determines the moral facts. Jacksonconcedes that this is a real problem: if there is no convergence inmoral attitudes, his methodology would imply moral relativism. He canonly offer hope that convergence will occur some day (137–8).And so Jackson’s ambitious metaethical project turns out to reston an appeal to optimism.

Bibliography

• Anscombe, G.E.M., 1958, “Modern Moral Philosophy”,Philosophy, 33: 1–19.
• Bader, R., 2017, “The Grounding Argument againstNon-reductive Moral Realism”,Oxford studies inmetaethics, 12: 106–134.
• Barker, Stephen J., 2000, “Is Value Content a Component ofConventional Implicature?”,Analysis, 60:268–279.
• Bedke, Matthew S., 2009, “Intuitive Non-Naturalism MeetsCosmic Coincidence”,Pacific Philosophical Quarterly,90(2): 188–209.
• Bedke, Matthew S., 2012, “Against NormativeNaturalism”,Australasian Journal of Philosophy, 90(1):111–129.
• Bengson, J., Cuneo, T., and Shafer-Landau, R., 2020,“Trusting Moral Intuitions”,Noûs, 54(4):956–984.
• Blackburn, Simon. 1993, “Supervenience Revisited.”Essays in Quasi-Realism. Oxford: Oxford University Press,130–148.
• Bloomfield, P., 2001,Moral Reality, Oxford: OxfordUniversity Press.
• Boyd, Richard, 1988, “How to be a Moral Realist”, inSayre-McCord 1988, 187–228.
• Brandt, Richard B., 1979,A Theory of the Good and theRight, Westminster, MD: Prometheus Books.
• Brink, David O., 1986, “Externalist Moral Realism”,Southern Journal of Philosophy, 24 (Supplement):23–40.
• –––, 1989,Moral Realism and the Foundationsof Ethics, Cambridge: Cambridge University Press.
• –––, 2001, “Realism, Naturalism, and MoralSemantics”,Social Philosophy and Policy, 18:154–176.
• Brown, C., 2011, “A New and Improved Supervenience Argumentfor Ethical Descriptivism”, in R. Shafer-Landau, ed.,OxfordStudies in Metaethics, vol. 6, Oxford University Press,205–218.
• Brown, E., 2016, “Aristotelian Virtue Ethics and theNormativity Challenge”,Dialogue, 55:131–150.
• Chilovi, S., & Wodak, D., 2022, “On the (in)significance of Hume’s Law”,PhilosophicalStudies, 179(2): 633–653.
• Chudnoff, E., 2013,Intuition, Oxford UniversityPress.
• Clarke, Christopher. “The Correlation Argument forReductionism”,Philosophy of Science 86(1):76–97.
• Copp, David, 1995,Morality, Normativity, and Society,New York: Oxford University Press.
• –––, 2000, “Milk, Honey, and the Good Lifeon Moral Twin Earth”,Synthese 124(1):113–137.
• –––, 2001, “Realist-Expressivism: ANeglected Option for Moral Realism”,Social Philosophy andPolicy, 18: 1–43.
• –––, 2008, “Darwinian skepticism aboutmoral realism”,Philosophical Issues, 18(1):186–206.
• –––, 2017, “Normative naturalism andnormative nihilism: Parfit’s dilemma for naturalism”, inSimon Kirchin (ed.),Reading Parfit, London: Routledge,28–53
• –––, 2020, “Just Too Different: NormativeProperties and Natural Properties”,PhilosophicalStudies, 177(1): 263–286.
• Dancy, Jonathan, 2006, “Nonnaturalism”, in David Copp(ed.),The Oxford Handbook of Ethical Theory, OxfordUniversity Press, 122–45.
• Dowell, J. L., 2016, “The Metaethical Insignificance ofMoral Twin Earth”, in Russ Shafer-Landau (ed.),OxfordStudies in Metaethics, Volume 11, Oxford: Oxford UniversityPress, 1–27
• Dreier, J., 2015, “Explaining the Quasi-real”,Oxford studies in Metaethics, 10, 273–298.
• Dunaway, B., 2015, “Supervenience Arguments and NormativeNon-naturalism”,Philosophy and PhenomenologicalResearch, 91(3): 627–655.
• Dunaway, B., & McPherson, T., 2016, “Reference Magnetismas a Solution to the Moral Twin Earth Problem”,Ergo,3(25).
• Eklund, M., 2017,Choosing Normative Concepts, OxfordUniversity Press.
• Enoch, David, 2010, “The Epistemological Challenge toMetanormative Realism: How Best to Understand It, and How to Cope withIt”,Philosophical Studies, 148(3): 413–438.
• –––, 2011,Taking Morality Seriously: ADefense of Robust Realism, Oxford: Oxford University Press.
• Faraci, D., 2017, “Hybrid Non-naturalism Does Not Meet theSupervenience Challenge”,J. Ethics & Soc. Phil.,12: 312.
• Feldman, F., 2005, “The Open Question Argument: What itisn’t; and What it is”,Philosophical Issues, 15:22–43.
• Feyerabend, P., 1975,Against Method, London: NewLeft.
• Fine, Kit, 2002, “The Varieties of Necessity”, inTamar Szabo Gendler and John Hawthorne (eds.),Conceivability andpossibility, Oxford: Oxford University Press, 253–281.
• Finlay, Stephen, 2014,Confusion of Tongues: A Theory ofNormative Language, Oxford: Oxford University Press.
• FitzPatrick, W. J., 2015, “Debunking Evolutionary Debunkingof Ethical Realism”,Philosophical Studies, 172:883–904.
• Flanagan, O., Sarkissian, H., & Wong, D. 2016.“Naturalizing Ethics”, in Kelly James Clark (ed.),TheBlackwell Companion to Naturalism, New York: John Wiley and Sons,16–33.
• Foot, Philippa, 2001,Natural Goodness, Oxford: ClarendonPress.
• –––, 2002,Moral Dilemmas and Other Topicsin Moral Philosophy, Oxford: Clarendon Press.
• Frankena, W., 1939, “The Naturalistic Fallacy”,Mind, 48: 464–477.
• Geach, Peter T., 1956, “Good and Evil”,Analysis, 17: 23–42.
• Hacker-Wright, J., 2009, “What is Natural about Foot’sEthical Naturalism?”,Ratio, 22: 308–321
• Hare, R.M., 1952,The Language of Morals, Oxford:Clarendon Press.
• Harman, Gilbert, 1977,The Nature of Morality, New York:Oxford University Press.
• Heathwood, Chris, 2009, “Moral and Epistemic Open QuestionArguments”,Philosophical Books 50(2):83–98.
• Hernandez, S., & Laskowski, N. G., 2021,. “What MakesNormative Concepts Normative”,Southwest PhilosophyReview, 37(1): 45–51.
• Horgan, Terence and Mark Timmons, 1991, “New Wave MoralRealism Meets Moral Twin Earth”,Journal of PhilosophicalResearch, 16: 447–465.
• Howard, N. R., & Laskowski, N. G., 2021, “The World isNot Enough”,Noûs, 55(1): 86–101.
• Huemer, Michael, 2005,Ethical Intuitionism, London:Palgrave Macmillan.
• Hursthouse, Rosalind, 1999,On Virtue Ethics, Oxford:Clarendon Press.
• Jackson, Frank, 1998,From Metaphysics to Ethics: A Defence ofConceptual Analysis, Oxford: Clarendon Press.
• Jackson, F., & Pettit, P., 1995, “Moral Functionalismand Moral Motivation”,The Philosophical Quarterly,45(178): 20–40.
• Jordan, J., 2016, “Ethical Naturalism and the Justificationof Claims about Human Form”,Dialogue, 55:467–492.
• –––, 2020, “Have Neo-AristoteliansAbandoned Naturalism? On the Distinctively Human Form of PracticalReason”,The Journal of Value Inquiry, 54(2):183–201.
• Joyce, Richard, 2001,The Myth of Morality, Cambridge:Cambridge University Press.
• –––, 2006,The Evolution of Morality,Cambridge, MA: MIT Press.
• Kim, Jaegwon, 1993, “Multiple Realization and theMetaphysics of Reduction”, inSupervenience and Mind,New York: Cambridge, 309–35.
• Klocksiem, J., 2019, “Against Reductive EthicalNaturalism”,Philosophical Studies, 176(8):1991–2010.
• Lane, A., 2018, “Are Moral Functionalism’s Moral APriori Commitments Really A Priori?”,Ethics, Politics &Society, 1: 15–15.
• Laskowski, NG, 2019, “The Sense of Incredibility inEthics”,Philosophical Studies, 176: 93–115.
• –––, 2020, “Resisting ReductiveRealism”,Oxford Studies in Metaethics, Volume 15,Oxford: Oxford University Press, 96–117.
• Leary, S., 2017, “Non-naturalism and NormativeNecessities”, inOxford Studies in Metaethics, Volume12, Oxford: Oxford University Press, 76–105
• Loeb, D., 1995, “Full-information Theories of IndividualGood”,Social Theory and Practice, 21(1):1–30.
• Lott, M., 2012, “Moral Virtue as Knowledge of HumanForm”,Social Theory and Practice, 38:407–431.
• –––, 2020, “Aristotelian Naturalism andthe Autonomy of Ethics” in M. Hähnel (ed.),Aristotelian Naturalism: A Research Companion, Cham: SpringerNature, 283–294.
• Luco, A. C., 2016, “Non-negotiable: Why Moral NaturalismCannot Do Away with Categorical Reasons”,PhilosophicalStudies, 173: 2511–2528.
• Lutz, M., 2018, “What Makes Evolution a Defeater?”Erkenntnis, 83(6): 1105–1126.
• –––, 2020, “The Reliability Challenge inMoral Epistemology”,Oxford Studies in Metaethics,volume 15, Oxford: Oxford University Press, 284–308.
• Majors, Brad, 2005, “Moral Discourse and DescriptiveProperties”,The Philosophical Quarterly, 55(220):475–494.
• McDowell, John, 1998, “Two Forms of Naturalism”, inMind, Value, and Reality, Cambridge, MA: Harvard UniversityPress.
• McPherson, Tristram, 2012, “Ethical Non-Naturalism and theMetaphysics of Supervenience”, inOxford Studies inMetaethics, Volume 7, Oxford: Oxford University Press,205–234
• –––, 2015, “What is at Stake in Debatesamong Normative Realists?”Noûs, 49(1):123–146.
• McPherson, T., & Plunkett, D., 2023, “Ground, Essence,and the Metaphysics of Metanormative Non-Naturalism”,Ergo, 9:26.
• Millgram E, 2009, “Life and Action”Analysis,69(3): 557–564.
• Millum J, 2006, “Natural Goodness and Natural Evil”,Ratio, 19: 199–213.
• Moore, G.E., 1903,Principia Ethica, Cambridge: CambridgeUniversity Press.
• –––, 1942. “A Reply to My Critics,”in P. Schilpp (ed.),The Philosophy of G.E. Moore. Chicago:Open court.
• Moosavi, P., 2022a, “Natural Goodness without naturalhistory”,Philosophy and Phenomenological Research,104(1): 78–100.
• –––, 2022b, “Neo-Aristotelian naturalismas ethical naturalism”,Journal of Moral Philosophy,1(aop): 1–26.
• Morton, J., 2020, “Grounding the normative: a problem forstructured non-naturalism”,Philosophical Studies,177(1): 173–196.
• Nussbaum, Martha C., 1995, “Aristotle on Human Nature andthe Foundations of Ethics”, in Altham and Harrison 1995,86–131.
• Oddie, G., 2005,Value, Reality, and Desire, ClarendonPress.
• Parfit, Derek, 2011,On What Matters (Volume Two),Oxford: Oxford University Press.
• Prinz, Jesse J., 2007,The Emotional Construction ofMorals, Oxford: Oxford University Press.
• –––, 2009, “The Normativity Challenge:Cultural Psychology Provides the Real Threat to Virtue Ethics,”Journal of Ethics, 13: 117–144.
• Prinzing, M., 2017, “The Revisionist’s Rubric:Conceptual Engineering and the Discontinuity Objection”,Inquiry, 2017: 1–27.doi:10.1080/0020174X.2017.1385522
• Putnam, Hilary, 1975, “The Meaning of‘Meaning’”,Minnesota Studies in the Philosophyof Science, 7: 131–193.
• Railton, P. 1986a, “Facts and values”,Philosophical Topics, 14(2), 5–31.
• –––, 1986b, “Moral realism”,ThePhilosophical Review, 95(2): 163–207.
• Rawlette, S., 2020,The Feeling of Value: Moral RealismGrounded in Phenomenal Consciousness, First Ebook Edition.(self-published).
• Risberg, O., 2018, “The Entanglement Problem andIdealization in Moral Philosophy”,The PhilosophicalQuarterly, 68(272): 542–559.
• Roberts, D., 2017, “Thick Concepts”, in TristramMcPherson and David Plunkett (eds.),The Routledge Handbook ofMetaethics,New York: Routledge, 211–226.
• –––,2018, “Why Believe in NormativeSupervenience?”,Oxford Studies in Metaethics 13:1–24.
• Rosati, C. S., 1995, “Persons, Perspectives, and FullInformation Accounts of the Good”,Ethics 105(2):296–325.
• Rosen, G., 2015, “Real definition”,AnalyticPhilosophy, 56(3): 189–209.
• –––, 2017a, “Metaphysical Relations inMetaethics”, in Tristram McPherson and David Plunkett (eds.),The Routledge Handbook of Metaethics, New York: Routledge,151–169.
• –––, 2017b, “What is a Moral Law”,Oxford studies in metaethics, Volume 12, Oxford: OxfordUniversity Press, 135–59.
• Scanlon, Thomas, 1998,What We Owe to Each Other,Cambridge, MA: Belknap Press of Harvard University Press.
• –––, 2014,Being Realistic AboutReasons, Oxford: Oxford University Press.
• Schechter, Joshua, 2017, “Explanatory Challenges inMetaethics”, in Tristram McPherson & David Plunkett (eds.),Routledge Handbook of Metaethics, London: Routledge,443–459.
• Schroeder, Mark, 2005, “Realism and Reduction: The Quest forRobustness”,Philosophers’ Imprint, 5(1),available online.
• –––, 2007,Slaves of the Passions,Oxford: Oxford University Press.
• –––, 2010,Noncognitivism in Ethics.London: Routledge.
• –––, 2016, “What Matters AboutMetaethics?”Does Anything Really Matter? Responses toParfit, Oxford: Oxford University Press, 213–230.
• Shafer-Landau, Russ, 2003,Moral Realism: A Defence,Oxford: Clarendon Press.
• Sinclair, N., 2019, “The Naturalistic Fallacy and theHistory of Metaethics”, inThe Naturalistic Fallacy,Neil Sinclair (ed.), Cambridge: Cambridge University Press,9–29.
• Sinhababu, N., 2018, “Ethical Reductionism”,J.Ethics & Soc. Phil., 13: 32.
• –––, 2019, “One-Person Moral Twin EarthCases”,Thought: A Journal of Philosophy, 8(1):16–22.
• Sinnott-Armstrong, Walter, 2006,Moral Skepticisms,Oxford: Oxford University Press.
• Smith, Michael, 1994,The Moral Problem, Oxford:Blackwell.
• Sobel, D., 1994. “Full Information Accounts ofWell-Being”,Ethics, 104(4): 784–810.
• Stoljar, Daniel, 2010,Physicalism, New York:Routledge.
• Street, Sharon, 2005, “A Darwinian Dilemma for RealistTheories of Value”,Philosophical Studies, 127(1):109–166.
• Streumer, B., 2008, “Are There Irreducibly NormativeProperties?”,Australasian Journal of Philosophy,86(4): 537–561.
• Sturgeon, Nicholas, 1985, “Moral Explanations”, inDavid Copp and David Zimmerman (eds.),Morality, Reason, andTruth, Totowa, NJ: Rowman and Allanheld, 49–78.
• –––, 2009, “Doubts about the Supervenienceof the Evaluative”,Oxford Studies in Metaethics,Volume 4, Oxford: Oxford University Press, 53–90.
• –––, 2010, “Normativity”,Analysis Reviews, 70(4): 744–753.
• Szabó, Z. G., 2001, “Adjectives in Context”, inIstván Kenesei and Robert M. Harnish (eds.),Perspectiveson Semantics, Pragmatics, and Discourse, (Pragmatics andBeyond New Series, 90), Amsterdam: John Benjamins,119–146.
• Thompson, Michael, 2008,Life and Action: ElementaryStructures of Practice and Practical Thought, Cambridge, MA:Harvard University Press.
• Thomson, Judith Jarvis, 1996, “Moral Objectivity”, inHarman and Thomson 1996, 65–154.
• –––, 1997, “The Right and the Good”,Journal of Philosophy, 94: 273–298.
• –––, 2001,Goodness and Advice,Princeton, NJ: Princeton University Press.
• –––, 2008,Normativity, Chicago: OpenCourt.
• Toppinen, T., 2018, “Essentially Grounded Non-naturalism andNormative Supervenience”,Topoi, 37(4):645–653.
• Vavova, Katia, 2015, “Evolutionary Debunking of MoralRealism”,Philosophy Compass, 10(2):104–116.
• Väyrynen, Pekka, 2017, “The Supervenience Challenge toNon-Naturalism”, in Tristram McPherson & David Plunkett(eds.),The Routledge Handbook of Metaethics, London:Routledge, 170–184.
• –––, 2018, “A Simple Escape from MoralTwin Earth”,Thought: A Journal of Philosophy, 7(2):109–118.
• –––, 2021, “Normative Naturalism on ItsOwn Terms”,Organon F: MedzinárodnýČasopis Pre Analytickú Filozofiu, 28(3):505–530.
• Vessel, J. P., 2020, “Moore’s Open QuestionManeuvering: A Qualified Defense”,The Journal ofEthics, 24(1): 91–117.
• Wedgwood, R., 2007,The Nature of Normativity, OxfordUniversity Press.
• Wielenberg, Erik J., 2014,Robust Ethics: The Metaphysics andEpistemology of Godless Normative Realism, Oxford: OxfordUniversity Press.
• Williams, Bernard, 1972,Morality: An Introduction toEthics, Cambridge: Cambridge University Press
• Wisdom, J., 2021, “A Dose of Reality for Moral TwinEarth”,Philosophical Psychology, 34(6):784–804.
• Wittgenstein, Ludwig, 1965, “A Lecture on Ethics”,Philosophical Review, 74(1): 3–12.
• Wong, David B., 2006,Natural Moralities: A Defense ofPluralistic Relativism, Oxford: Oxford University Press.
• Woodcock, S., 2018, “Aristotelian Naturalism vs. Mutants,Aliens and the Great Red Dragon”,American Philos Q,55(4): 313–328.
• Yablo, Stephen, 2000, “Red, Bitter, Best”,Philosophical Books, 41: 13–23.
• Zangwill, N., 2000, “Against Analytic MoralFunctionalism”,Ratio, 13(3): 275–286.

Academic Tools

	How to cite this entry.
	Preview the PDF version of this entry at theFriends of the SEP Society.
	Look up topics and thinkers related to this entry at the Internet Philosophy Ontology Project (InPhO).
	Enhanced bibliography for this entryatPhilPapers, with links to its database.

Other Internet Resources

• Leary, Stephanie, (ms),What is Moorean non-naturalism?

Related Entries

Aristotle, General Topics: ethics |cognitivism vs. non-cognitivism, moral |constructivism: in political philosophy |functionalism |Hume, David: moral philosophy |Moore, George Edward |moral epistemology |moral non-naturalism |moral realism |moral relativism |moral skepticism |naturalism |naturalism: in legal philosophy |supervenience

Acknowledgments

Many thanks to Spencer Case, Steve Finlay, Nick Laskowski, TristramMcPherson, Tim Perrine, Russ Shafer-Landau, and Pekka Väyrynenfor helpful discussion and comments. The structure of this article wasadapted from an earlier version by James Lenman.