MandM header image 2

Erik Wielenberg and the Autonomy Thesis: Part Two Standard Objections to the Autonomy Thesis, Reasons to be Moral Without God

March 20th, 2017 by Matt

defaul1The autonomy thesis contends that there can be moral requirements to φ regardless of whether God commands, desires, or wills that people φ. In his monograph, Robust Ethics: The Metaphysics and Epistemology of Godless Normative Realism,[1] Erik Wielenberg offers arguably one of the most sophisticated defences of the autonomy thesis to date.

Wielenberg argues three things. First, I the most plausible alternative to the autonomy thesis, the divine command theory, is problematic because it cannot account for the moral obligations of reasonable unbelievers. Second, II  robust realism, the thesis that moral requirements are sui generis non-natural properties which supervene upon natural properties, can be formulated in a way that avoids the standard objections to the autonomy thesis. Third, III robust realism provides a better account of intrinsic value than any meta-ethical theory that identifies moral goodness with states of God.

In a previous post I discussed I, I argued Wielenberg’s critique of divine command meta-ethics fails. This post will begin looking at his second major claim II. As an alternative to DCM, Wielenberg proposes a position he labels Robust Realism, the thesis that moral requirements are sui generis non-natural properties which supervene upon natural properties. By itself, robust realism is compatible with both theism and atheism. Wielenberg refers to the conjunction of robust realism and atheism as Godless Robust Normative Realism (GRNR) Wielenberg contends that his formulation of GRNR avoids the standard objections to the autonomy thesis.  Here I’ll look at one such objection.

Reasons for being Moral without God.

One objection Wielenberg discusses is the perennial concern that, in the absence of God, people lack compelling reasons to comply with morality’s demands. Wielenberg takes the following comments by Craig as representative of this objection.

Even if there were objective moral values and duties under naturalism, they are irrelevant because there is no moral accountability. If life ends at the grave, it makes no difference whether one lives as a Stalin or as a saint…Why should you sacrifice your self-interest and especially your life for the sake of someone else? There can be no good reason for adopting such a self-negating course of action on the naturalistic worldview… Life is too short to jeopardize it by acting out of anything but pure self-interest. [2]

 Wielenberg initially suggests that Craig is arguing that “if people had moral obligations, but God did not exist, then people would have no normative reasons to carry out their obligations”.[3] He re-joins, plausibly, that people often do have normative reasons to refrain from wrongdoing. “the fact rape harms its victims is a compelling reason for me not to rape, regardless of whether refraining from rape benefits me” (emphasis mine).[4]

Wielenberg then suggests that Craig might be offering a different argument. He might be contending that “if God does not exist then people lack self-interested normative reasons to perform their obligations” Wielenberg suggests taken this way, there are two problems with the argument. First, even if it were true, it wouldn’t follow that people lack any normative reasons to perform their obligations.  Second, this claim isn’t true “people often do in fact do have powerful self-interested reason’s for caring about fulfilling their obligations”[5] Wielenberg cites various empirical studies which show that immorality tends to disrupt, and damage people’s social relationships and harm meaningful connections.[6]

However, it seems implausible to me that Craig is maintaining either of these positions. Elsewhere.  Craig clarifies his position as follows:

[I]f God does not exist, then prudential reason and moral reason can and often do come into conflict, in which case there is no reason to act morally rather than in one’s self-interest. That’s consistent with saying that in other cases it is, indeed, prudent to act morally.[7]

Here Craig affirms that if atheism is true, people often can have both moral and prudential normative reasons to carry out their moral obligations. So he doesn’t affirm, either of the positions, Wielenberg attributes to him.  He isn’t saying that no one ever has reasons to refrain from wrongdoing, nor is he claiming that no one ever has prudential reasons to refrain from wrongdoing.  Craig’s objection is rather that, if atheism is true, moral and prudential reasons can and do come into conflict. When they do people lack any reason to comply with what morality demands. “One has moral value pulling in one direction and prudential value tugging in the opposite, and no way to decide rationally which choice to make”[8].

Latter Wielenberg suggests a different way of formulating the argument:

[1] If GRNR is true then morality and self-interest sometimes diverge in the long run.

[2] But, morality and self-interest never diverge in the long run.

[3] Therefore, GRNR is false[9]

Wielenberg grants the truth of [1], the empirical studies he earlier cited show only that people “often” have powerful prudential reasons to do what is right, not that they always or necessarily do. Wielenberg states that “an important different between a theistic universe and a godless universe” is that “without God, there is always the possibility that we will face a deep conflict between what is in our self-interest and what morality requires of us”. On the other hand, if God exists “there is a perfect correlation between morality and self-interest”[10]

Wielenberg’s rebuttal, therefore, focuses on denying [2], his treatment of [2] consists of a single paragraph:

It is hard to imagine a convincing non-question begging rationale for (2) that wouldn’t at the same time tell against (1). One might appeal to the existence of God to support (2) but this obviously begs the question. Alternatively, one might find a secular ground for (2)-but to the extent that such a ground is convincing it undermines the first premise…Craig often proceeds as if (2) were a datum for which any plausible moral theory must account. But such an approach has nothing to commend it. It is no more plausible than the falsity of (2) as a datum and arguing against Craig’s view. Such arguments get us nowhere.[11]

Wielenberg provides two reasons for rejecting [2]. These are (a) that there is no non-question-begging reason for affirming [2] which does not undermine [1], and (b) that [2] is not a datum to be explained by a moral theory.  Neither is compelling.

Regarding first (a), the problem is that several people, including Craig, have offered reasons for affirming [2] which neither assume theism or tell against [1].

Robert Adam’s, for example, has appealed to the intuition that moral judgments “have an action- and preference-guiding force that they could not have unless everyone had reason to follow them in his actions and preferences.”[12]  Adam’s argues that “if happiness will, in the long run, be strictly proportioned to moral goodness, that explains how virtually everyone does have an important reason to want to be good.”[13] However, if this is not the case, it’s hard to justify the conclusion that “everyone does have reason always to be moral.”[14]

Adam’s here focuses on the idea that everyone has a reason to be moral.  That if an action φ is morally wrong for a person P to perform, then P has a reason to not φ. Stephen Layman has offered a similar line of argument, focusing instead on the idea that people always have decisive reasons to do what morality demands.  Layman refers to what he calls the “The Reasons Thesis: The strongest reasons always favor doing what is morally required.” The idea that if something is obligatory, we not only have a reason to do it but that this reason is always decisive. Other reasons we may have for not complying such as reasons of self-interest or economics do not override it. Layman provides several examples which suggest that “If there is no God and no life after death, then there are cases in which morality requires that one make a great sacrifice that confers relatively modest benefits (or prevents relatively modest harms).” Layman argues that if such cases obtain,  reasons of prudence will override moral reasons we have doing the action in question.[15]

Neither of these arguments begs the question by assuming God exists. They appeal not to God’s existence, but theses about the authority of moral requirements, that they provide virtually everyone with decisive reasons for acting.  Nor do these arguments give us reason for questioning [1]. Both, of them, in fact, concede and incorporate [1] in their thinking. So, Wielenberg’s claim, that it’s hard to imagine a non-questioning begging rationale for [2] that doesn’t undermine [1], is false. Several such rationales have been offered in the literature which he does not even mention let alone address.

Moreover, Craig himself provided a similar rationale.   Craig claimed that:“[I]f God does not exist, then prudential reason and moral reason can and often do come into conflict, in which case there is no reason to act morally rather than in one’s self-interest[16]. Elsewhere, Craig has stated:

“I agree with Layman that on atheism, what he calls the overriding thesis ( namely that moral value always trumps prudential value) is not true, for one can have extremely strong prudential reasons for not acting morally, and there seems to be no common scale in which to weigh moral against prudential considerations”[17]

This brings us to  (b), Wielenberg’s objection that [2] is not a datum to be explained by a moral theory. While he is correct that [2] itself is not a datum of moral theory, the claim that moral demands are authoritative so that everyone always as a decisive reason to be moral is something that, at least prima facie, a meta-ethical theory should explain.[18]  Wielenberg himself seems to accept this. He states “as I suggested in Chapter 1, to have an obligation just is to have decisive reasons to perform a certain action.” Therefore, according to Wielenberg, it is a necessary truth that we always have decisive reasons to do what we are moral required to do, and it is impossible, for people to not have such reasons or them to be overridden by other reasons such as self-interest”. Adam’s, Layman and Craig, therefore, appeal to something that is, on his view, a datum a meta-ethical theory should explain.

[1] Erik Wielenberg, Robust Ethics: The Metaphysics and Epistemology of Godless Normative Realism (Oxford: Oxford University Press, 2014)

[2] Erik Wielenberg, Robust Ethics, 56-57, the citation is from William Lane Craig “The Indispensability of Meta-Ethical Theological Foundations for Morality” available at

[3] Ibid, 57

[4] Ibid

[5] Ibid

[6] Wielenberg’s examples all deal with Hume’s idea of the sensible knave, the individual who has a reputation for morality but engages in undetected wrongdoing when it’s in his self-interest. They don’t, however, address Kai Neilsen’s example of a “classist amoralist” who forms deep and genuine relationships with others within his class but ruthlessly exploits members of other classes to his advantage. See Kai Nielsen Why be Moral (Buffalo, N.Y: Prometheus Books, 1989) 295-296

[7] William Lane Craig  “Q&A 230 Is Life Absurd without God?” available at Accessed 6/2/2017

[8] William Lane Craig “This Most Gruesome of Guests” Is Goodness without God Good Enough: A Debate on Faith, Secularism and Ethics eds Robert K Garcia and Nathan L King (Lanham: Rowman & Littlefield Publishers, 2008) 182-183

[9] Wielenberg, Robust Ethics 59

[10] Ibid, 59

[11] Ibid, 59

[12] Robert Adams “Moral Argument for Theistic Belief” in The Virtue of Faith and Other Essays in Philosophical Theology, (New York: Oxford University Press, 1987). 158

[13] Ibid

[14] Ibid.

[15] See C Stephen Layman “God and the moral order”, Faith and Philosophy, 19: 3 (2002) 304-16.

[16] William Lane Craig “Q&A 230 Is Life Absurd without God?” available at Accessed 6/2/2017

[17] William Lane Craig “This Most Gruesome of Guests” 183

[18] Terence Cuneo makes this point, see  Terence Cuneo “Erik J Wielenberg Robust Ethics: The Metaphysics and Epistemology of Godless Normative Realism” Notre Dame Philosophical Reviews 2015.03.24 available at accessed 7/2/17

Tags:   · · · · · 1 Comment

Leave a Comment

1 response so far ↓

  • Hi Matt,

    I think clarification is required when it comes to assessing what it means for a reason to be “overriding”, or what’s “self-interest”.
    For example, if an agent values not being morally evil, then it seems from her own evaluative perspective, it’s in her interest, all other things equal, to refrain from immoral behavior – even if there might be sometimes reasons to engage in immoral behavior that are stronger from the perspective of the agent’s evaluative function.
    But that aside, I think Wielenberg’s problem here is the claim “to have an obligation just is to have decisive reasons to perform a certain action”. While I disagree that in normal humans, prudential and moral reasons often come into conflict (because as a matter of normal human psychology, humans tend to suffer and not be happy if they behave immorally), they might conflict in the case of normal humans in infrequent cases, and they very likely conflict more often in the case of psychopaths who are in a position of doing evil and almost certainly avoid punishment.

    As for Adam’s claim about the intuition that moral judgments “have an action- and preference-guiding force that they could not have unless everyone had reason to follow them in his actions and preferences.”, that can be explained by evolution: behaving immorally was overall in most circumstances negative for reproductive success, which resulted in an evolved motivation to not behave immorally, feelings of guilt if one behaves immorally, a desire to make amends if one does, etc. That may come into conflict with other motivations, also the result of evolution.

    But that aside, here’s a difficulty for DCT: what about people who do not have any such feelings?
    For example, a psychopath might not feel any guilt or anything like that. Now, they might learn right and wrong from other humans, and then have prudential reasons for not behaving immorally. But the problem is: Let’s say Jack is a psychopath who grew up in a society where immoral behavior X is widely regarded as morally acceptable or even praiseworthy.
    What prudential reasons does Jack have, on DCT, not to behave immorally?
    You might think a reason is to avoid afterlife punishment. Alas, Jack has no idea that X is immoral. He has no normally functioning moral sense that he can use to figure it out. And his society will not help. So, what’s the prudential reason?
    Another difficulty: Did God fail to communicate his command to Jack?
    After all, God did not pass on his command by means of a moral sense – Jack’s brain doesn’t work properly -, and society won’t help.

    Granted, DCTists might claim that a person like Jack is not metaphysically possible, that all personal agents have a functioning moral sense that allows them to figure out moral truth if they think about it, etc. But that’s extremely improbable. We do know that altering the brain alters pretty much any aspect of the mind. It’s very likely given the evidence about the connection between mind and brain, plus the evidence from observation of psychopaths, that someone like Jack is metaphysically possible (and some people similar to that are probably even actual).