Skip to main content


Log in

No need to get up from the armchair (if you’re interested in debunking arguments in metaethics)

  • Published:
Ethical Theory and Moral Practice Aims and scope Submit manuscript


Several authors believe that metaethicists ought to leave their comfortable armchairs and engage with serious empirical research. This paper provides partial support for the opposing view, that metaethics is rightly conducted from the armchair. It does so by focusing on debunking arguments against robust moral realism. Specifically, the article discusses arguments based on the possibility that if robust realism is correct, then our beliefs are most likely insensitive to the relevant truths. These arguments seem at first glance to be dependent on empirical research to learn what our moral beliefs are sensitive to. It is argued, however, that this is not so. The paper then examines two thought experiments that have been thought to demonstrate that debunking arguments might depend on empirical details and argues that the conclusion is not supported.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Similar content being viewed by others


  1. Surveys of the debate include Vavova (2015) and Wielenberg (2016). This debate has spilled over to a number of additional domains, such as religion and metaphysics. For a comprehensive survey, see Korman (2019).

  2. Similar claims are made by Hopster (2018), Isserow (2019), Levy & Levy (2018), Mogensen (2016, 1812) and Parfit (2017, 3:286). In the context of a more limited debunking argument, Greene (2014, 726) expresses a similar attitude. Pölzler (2018), May (2018) and Sauer (2018) are examples of whole books devoted to engaging with detailed empirical science and debunking arguments in ethics of various sorts.

  3. Articles in this spirit include Chappell (2017), Kahane (2011, 111–12), Klenk (2017) and White (2010). Berker (2009) focuses on Greene’s dual process theory and Greene and Singer’s claim that it undermines deontological reasoning. I do not discuss these elements here. All of these authors’ claims are of limited scope. They do not explicitly claim that empirical details never matter. Nevertheless, I suggest that they are examples of a school of thought or philosophical leaning that tends to disregard the importance of empirical details for debunking arguments in metaethics.

    An interesting side note: Cuneo (2007) suggests that the division between these empirical vs. non-empirical inclinations underlies the division between metaethical naturalists and non-naturalists as well.

  4. David Enoch calls the intuition that motivates this view, the just-too-different intuition. For a clear and concise survey on how the just-too-different intuition challenges naturalism, see Paakkunainen (2018).

  5. Shafer-Landau (2003, 72) also endorses metaphysical autonomy as part of his non-naturalism. However, when he discusses the meaning of “non-naturalism” (p. 58) he prefers a disciplinary characterization rather than metaphysical. According to Shafer-Landau, non-naturalism means that morality does not belong to the same discipline as the natural sciences, primarily because the methodology for discovering moral truths is different from the methodology of natural science. Thanks to Preston Werner and to Russ Shafer-Landau himself for helping clarify Shafer-Landau’s views.

  6. Moral facts are facts about how we ought to act or about which states of affairs are good or bad. Therefore, pure moral facts, even though purely moral, will still be about non-moral circumstances, that is, a conjunction of non-normative properties. For instance, the claim “it is bad to cause pain” tells us that causing pain, a non-normative circumstance, has the moral property bad.

  7. Kit Fine (2002) argues that there is a unique normative necessity. Fine’s view is further developed by Gideon Rosen (forthcoming) who also nicely clarifies the relationship between this view and the widely accepted view that moral properties strongly supervene on non-normative properties. If all pure moral facts are metaphysically contingent, that means that moral properties do not strongly supervene on non-moral properties, though weak supervenience is not ruled out. (The distinction between weak and strong supervenience is made by Kim (1984)).

  8. Finlay (2007, n. 25) also notes this link between the two types of autonomy. Hume himself seems to have based his epistemic thesis on the metaphysical intuition, that “ought, or ought not, expresses some new relation…entirely different from [is, and is not propositions]” (Hume 1888, sec. 3:1:1).

  9. See also Shafer-Landau (2003, 63).

  10. To see this clearly, consider: Mind and body are very different according to dualism, but they do causally influence each other. Perhaps dualism is an implausible view, however, it is not inconsistent.

  11. I am assuming that the term “non-naturalism” that I use to characterize robust realism does not logically imply non-causality. Shafer-Landau (2003, 58–59) and Cuneo (2007, 856) think similarly. That is why I have to explain why, even though it is not a logical necessity, non-naturalists typically do endorse non-causality.

    Our beliefs about what we ought to do can of course affect our behavior. I am assuming that, according to robust realism, the belief-states are distinct from the facts themselves, such that this would not be considered a way of moral facts causing things.

  12. Some people think that certain religious authorities or texts fit the job as well. I do not, for predictable reasons, but I do not get into this debate here. Whether moral perceptualists should be understood as rejecting this thesis is an interesting question that I do not pursue here. For a presentation of the view and related discussion, see Bergqvist and Cowan (2018) and Werner (2016, 2017).

  13. See Paakkunainen (2018, sec. 3) for a brief survey. Tristram McPherson (2011, 232) distinguishes between robust normativity and merely formal normativity. Formal normativity means that actions (etc.) can be right or wrong according to some standard, where the standard can be conventional, made up or anything else. Robust normativity requires that the standard be authoritative. As John Mackie (1977, 40) put it with regard to moral norms, they must have to-be-doneness built into them. It seems plausible to conjecture that the autonomy theses, if true, are true of all and only robust norms. Thanks to David Enoch for the reference.

  14. There is a different style of argument developed by Hartry Field (1989, 25–30) as an argument against mathematical Platonism, and applied more recently to robust moral realism (Enoch 2011, chap. 7; Street 2008). According to this argument, robust moral realism should be rejected because it implies that the reliability of our moral beliefs, which is a fact that calls for explanation, cannot be explained. This argument is not premised on the Conditional Debunking Conclusion. I engage with Field’s style of argument in depth elsewhere and provide reasons to think that it too does not depend on empirical details. I therefore allow myself to set it aside here. I engage with Field’s style of argument in depth elsewhere and provide reasons to think that it too does not depend on empirical details (Baras 2017).

  15. See Ruse (1986, 254); Joyce (2001, 163); Sinnott-Armstrong (2006, 43); Street (2006); Bedke (2014); Joyce (2016); Hill (2016); Braddock (2017); May (2018, chap. 4) and Sauer (2018, chap. 2). For an influential response, see Clarke-Doane (2016, sec. 2.2). Clarke-Doane does so on the basis of a principle which he calls “modal security”. For discussion of that principle, see Clarke-Doane & Baras (2019).

  16. As you can tell from this claim, my inclinations are internalistic. Others might be more attracted to a kind of externalism according to which what matters is whether my belief is sensitive, not whether I know it to be so. According to this view, the parenthetical additions should be omitted.

    What happens if you learn that your beliefs are insensitive, but you also justifiably believe that insensitivity and even knowledge of insensitivity don’t undercut justification? I believe this is an interesting question, one I cannot explore it here.

  17. Does that mean that the beliefs of most of us won’t be undermined in this way, because we lack detailed knowledge of the mechanisms and genealogy of the processes that produce our moral beliefs? Not at all. I remind you that, I’m about to argue that all that is needed here is a very minimal acquaintance with science, the kind that the vast majority of us, definitely the readers of this article, do have.

  18. I do not imply that the second premise does not need examination as well, only that empirical details do not even initially seem relevant to this examination. For discussion of versions of the second premise, see White (2010, 580–82).

  19. Note that the target of this paper is independent from and sometimes confused with a different project. We have quick gut feelings and it is a great empirical project to find out when and why our quick judgments match up with our slower more thoughtful judgments. I think this is a charitable interpretation of Woodward and Allman (2007, 2008), Railton (2014) and some of Greene’s (2013, 2014) work. The former two defend a more optimistic view of intuitions; Greene gives a more mixed verdict. However, the kind of debunking that is the target of this paper is the kind that undermines our moral judgments as a whole, including our slow and more thoughtful judgments.

  20. The qualification “of the right type” is needed to exclude deviant causal chains.

  21. Benacerraf’s original argument against mathematical Platonism was based on this kind of causal condition, applied to knowledge. The causal condition has since significantly fallen out of favor, which is part of why philosophers of mathematics tend to focus more on Field’s development of the argument, which does not rely on a causal condition on knowledge or justification. Regarding Field’s style of argument, see above footnote 14.

  22. For a recent development of a debunking argument along these lines, see Lutz (forthcoming).

  23. Sturgeon (1984, 2006) famously argues in favor of moral explanations. However, Sturgeon is not a robust realist; hence his view lies outside the scope of this paper. For a survey, and doubts about the claim that moral non-naturalists should reject moral explanations, see Majors (2007).

  24. Faraci’s main concern is with epistemic coincidence. However, in a footnote (n. 32) he suggests that there is a tight connection between epistemic coincidence and epistemic justification. In particular, that evidence of epistemic coincidence is undermining. This implies that a lack of explanatory connection between belief and truth is undermining, as Appropriate explanatory connection says. This idea, in connection to debunking arguments, has recently been further explored by Korman & Locke (forthcoming).

  25. Schafer (2010) pursues this line in defending moral realism from an insensitivity based argument.

  26. See for example Barkhausen (2016).

  27. Fraser’s (2014, sec. 3.1) environment condition for evolutionary debunking arguments is based on this sort of reasoning. Sauer (2018, 34) calls this kind of debunking argument “obsoleteness debunking”.

  28. This negative conclusion is slightly weaker than Greene’s in the quoted text. Greene makes a further inferential jump to the positive conclusion that some cases of incest should not be condemned.

  29. Greene’s claim that consequentialist reasoning does not fall prey to his own argument is criticized by Tersman (2008), Berker (2009) and Kahane (2011).

  30. Priest (1997) makes a similar point about the idea of unnatural sex. Thanks to Jessica Issarow for the reference.


Download references


I am indebted to David Enoch, Jessica Isserow, Arnon Levy, Christiane Merritt, Said Saillant, Matthew Scarfone, Daniel Telech, Preston Werner, Shlomit Wygoda, two anonymous referees and to participants at my presentation at the Ethics and Biology conference at the Center for Moral and Political Philosophy at the Hebrew University of Jerusalem for very helpful comments and discussion on previous drafts. I’d also like to thank Michael Klenk, Josh May and Hanno Sauer for discussing their related work.


I wrote this paper during a postdoctoral fellowship at the Hebrew University’s Center for Moral and Political Philosophy.

Author information

Authors and Affiliations


Corresponding author

Correspondence to Dan Baras.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Baras, D. No need to get up from the armchair (if you’re interested in debunking arguments in metaethics). Ethic Theory Moral Prac 23, 575–590 (2020).

Download citation

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: