A Moral Fine-Tuning Argument

Abstract

This paper develops Mark D. Linville’s brief description of “a sort of moral fine-tuning argument”. I develop the argument in four ways: I unpack the argument and give it a clear formulation, I unpack the theistic explanation of why a somewhat reliable moral capacity is expected, I point to the significance of not seeking to explain a perfect moral capacity, and I put the argument up against the recent work on non-theistic moral epistemology by Derek Parfit, David Enoch, and Erik Wielenberg.

1. The Epistemological Problem: Explaining Correlation

Any proponent of moral realism, the view that there are normative facts not constituted by us, is faced with the challenge of giving an account of human moral knowledge. Philosophers take up this challenge in a variety of ways, giving an account of how we can have access to an independent moral realm, of how our moral beliefs can be justified or reliable, or arguing that moral beliefs can count as knowledge (Enoch 2011, pp. 152–58). Following David Enoch, I hold that a fruitful and elegant way of formulating the epistemological problem is to do so in terms of explaining correlation. Such an approach does not rely upon lots of theoretical assumptions (such as definitions of knowledge) or vague metaphors (such as access). The challenge is to explain the correlation between moral facts and moral beliefs. Most moral realists usually hold that we have a significant amount of true moral beliefs, which means there is a correlation between our beliefs and moral facts.1 It would be a massive lucky coincidence if our moral beliefs just happened to match these moral truths; so, some explanation must be given. It is challenging to explain this correlation, especially for a non-naturalist conception of moral realism, because this correlation cannot be explained by one of the factors causing or constituting the other. It cannot be our beliefs that cause or constitute normative truths because moral realism states that normative truths are independent of our normative judgements. And it cannot be the moral facts that are causally responsible for our moral beliefs: while concrete objects like sticks and stones can enter into causal chains, abstract objects like moral truths or mathematical truths are causally inert (Parfit 2011, p. 488).

Much discussion has arisen in the last two decades regarding “evolutionary moral debunking arguments”. Though such arguments come in many forms, they all make the case that those who simultaneously accept moral realism and evolution have a defeater for the belief that their ability to form moral beliefs is reliable (I use “ability to form moral beliefs” and “moral capacity” interchangeably). The argument typically goes like this: Evolutionary forces have influenced our moral capacity and, in turn, our moral beliefs. These evolutionary forces aim at survival and fitness, not attitude-independent moral truths. As this moral ability would have been selected regardless of whether the beliefs it produces are true, we have a defeater for the belief that this ability is reliable (Wielenberg 2010; Vavova 2015).

The evolutionary moral debunking argument is often taken to challenge the justification of our moral beliefs (Joyce 2017, p. 108). Suppose natural selection has pushed our moral capacity in directions not related to moral truth. In that case, we have reason not to trust this ability and realize that our moral beliefs lack justification—and even “good reason to think that our moral beliefs are probably mistaken” (Vavova 2015, p. 108). However, these debunking arguments are also relevant for the quest to explain the correlation between our moral beliefs and moral facts. Sharon Street writes that

allowing our evaluative judgements to be shaped by evolutionary influences is analogous to setting out for Bermuda and letting the course of your boat be determined by the wind and tides: just as the push of the wind and tides on your boat has nothing to do with where you want to go, so the historical push of natural selection on the content of our evaluative judgements has nothing to do with evaluative truth. Of course every now and then, the wind and tides might happen to deposit someone’s boat on the shores of Bermuda.

(Street 2006, p. 121)

Street points out that our moral beliefs may correspond to moral facts, but given evolutionary influences, such a correspondence would be extremely unlikely—so unlikely that it “begs for an explanation” (Street 2006, p. 125). Street’s reasoning shows that given evolutionary influences, the correlation between moral facts and moral beliefs is not something to be expected but a rather striking correlation that requires an explanation (Jakobsen 2020a). In the next section, I will argue that Christian theism provides such a good explanation of this correlation that it is possible to formulate a sort of moral fine-tuning argument, namely a theistic argument from moral epistemology.2

2. The Argument

Given Christian theism, a reliable moral capacity is to be expected—and a reliable moral capacity would explain the correlation between moral facts and beliefs. According to Christian theism—as well as some other theistic traditions—God has created the cognitive abilities of humans, created humans for a certain kind of life, and calls humans to particular acts of service, which makes us understand why humans have a moral capacity. The sum of these theistic doctrines implies, if not entails, a reliable moral capacity: if humans’ cognitive capacities are created by God so that humans can hear God’s calling and receive vocations from God, then it is to be expected that humans have a reliable moral capacity and therefore a large number of moral beliefs that correspond to moral facts.

Given non-theism and the evolutionary influence on our evaluative tendencies, the correlation between moral beliefs and moral facts is not to be expected. The likelihood of natural selection giving humans moral beliefs that correspond to moral facts is rather low. As Street says, it must be treated as a remarkable coincidence (Street 2006, p. 132). Now, this is not to say that human moral beliefs are incorrect. Moral beliefs, directly or indirectly influenced by natural selection, could correspond to moral facts. However, pointing out that there could be a correlation does not raise the likelihood of a correlation. Various influences on our evaluative tendencies, such as evolutionary influences, allow for a wide range of possible valuations.

So, given non-theism and the influence of natural selection, the likelihood of a correlation between moral facts and moral beliefs is rather low. Consider the case that there is a God capable of creating humans with the ability to acquire moral beliefs and who intends to do this. In that case, a correlation between moral facts and moral beliefs is expected. Given the existence of God, then, the likelihood of our moral beliefs corresponding to moral facts is higher than the likelihood of the correlation occurring given natural selection. Now, let us put this more formally. Let P stand for probability, C for the correlation between moral beliefs and moral facts, Ns for natural selection, and the symbol | stand for “given”. As we have argued, P(C|Ns) is lower than P(C|God).3 So, a theistic explanation makes the correlation more probable than a non-theistic explanation, meaning that the theistic explanation is the likeliest.

Being the likeliest explanation has a noteworthy implication, namely that one can make an inference to the truth of this explanation. According to the so-called likelihood principle, if some matter of fact is more probable on one hypothesis than on another, this matter of fact will support one hypothesis over the other. So, if the correlation between moral beliefs and moral facts is more probable given God than given natural selection and non-theism, this correlation will support Christian theism over non-theism. In formal terms: P(C|Ns) < P(C|God) → P(Ns|C) < P(God|C). In plain English, one can conclude that the correlation between moral beliefs and moral facts supports the hypothesis of theism. I will call this a theistic argument from moral epistemology.

This formulation of a theistic argument may be familiar to some. It has the same structure as many fine-tuning arguments for the existence of God. Such arguments will often proceed by first showing that a fine-tuned universe is more probable given theism than given non-theism and then concluding that a fine-tuned universe supports theism over non-theism (Collins 2009). Moreover, Alvin Plantinga presents a theistic argument with a very similar structure. He argues that the probability of human rationality functioning properly is higher given theism than given naturalism. Therefore, human rationality supports theism over naturalism (Plantinga 1993, chp. 12; 2000, pp. 227–40; 2011, chp. 10). Formally, it can be put like this: P(R|N) < P(R|God). Now, I find my argument from moral epistemology to be stronger than Plantinga’s argument from rationality. It is possible to construct a quite plausible evolutionary explanation of human rationality. However, evolutionary explanations of our moral capacity that fit with moral realism are not as easily available. Accordingly, P(Moral C|Ns) is lower than P(Rational C|Ns), meaning that an argument from moral epistemology will more clearly support theism than an argument from rationality.

Note that there is no need to take a stance on exactly how probable or improbable P(R|N) or P(R|God) is or to fix the prior probability of Christian theism. The argument concerns comparative likelihood, which means that it is sufficient to argue that one is clearly more probable than the other, which can be done without assigning fixed values of probability. How can this be done? By asking which of the candidates, God or natural selection, best explains the phenomena in question. In his work on inferences to the best explanation, Peter Lipton argues that we judge likeliness precisely by asking which possible explanations provide the most understanding (Lipton 2003, p. 105). On the face of it, natural selection does not increase our understanding of this correlation but rather makes it unexpected, while Christian theism provides understanding. Further below, I will unpack the theistic explanation, showing how Christian theism increases our understanding of why there is such a correlation and makes the correlation expected, and consider possible naturalistic explanations.

Fine-tuning arguments tend to spend some time assigning a value to the probability of fine-tuning due to chance. The reasoning goes like this: our universe is governed by several fundamental physical parameters whose values cannot be predicted from any of our physical theories. In that sense, they are arbitrary. Moreover, they also fall within a very narrow range of life-permitting values. Had the values been slightly different, life could not exist. By calculating what values these constants could have had and the range of life-permitting values, one could come close to assigning a number to the probability of a certain constant being fine-tuned for life.4 Such a calculation cannot be performed when dealing with the correlation between moral beliefs and moral facts. But that is no major disadvantage. While formulations of the fine-tuning argument might include a prolegomena on physical constants, the argument itself is an evaluation of comparative likelihood.5 Calculations on physical constants do not have any other function than to make the case that a certain natural explanation is unlikely. That case can be made without any calculation. Just as fine-tuning arguments make the case that a constant can have a wide range of possible values while there is only a small range of life-permitting values, the argument from the correlation of moral facts and beliefs can make the case that the range of possible evaluative beliefs is wide, but the correlation can only take place within a small range of evaluative beliefs (Mogensen 2022, p. 69). Street appeals to natural selection, arguing that evolution allows for a very wide range of evaluative tendencies. If humans had evolved more along the lines of lions, baboons, or some social insects, they would have very different evaluative beliefs concerning sexual ethics, attitudes required towards outsiders, or the importance of one’s well-being (Street 2006, p. 120). So, just as fine-tuning arguments point to some single constant (such as the cosmological constant) that must be within a certain life-permitting region, we can point to some single evaluative topic (such as sexual ethics) that must be within a certain region to correlate with moral facts.

Moreover, just as fine-tuning arguments point to a finely tuned balance between physical variables (such as gravity and electromagnetism), we can point to a balance between evaluative variables. Take, for instance, the balance between egoism and universalism: to what degree can I prioritize my well-being at the cost of others, and to what degree must I sacrifice my well-being to increase the well-being of others? Or take the balance between morality and prudence: to what degree must I sacrifice my happiness to do what is morally good? There are numerous ways to finely balance these evaluative variables. The total range of possible parameter values is very wide, while the range that would reliably generate true moral beliefs is narrower. Just as the balance between the physical variables gravity and electromagnetism cannot fall outside of the narrower life-permitting range, the balancing of evaluative variables cannot fall outside of the narrow truth-permitting range, as that would lead to systematic errors in our ethical judgments. So, even though we cannot perform a calculation of how wide a range a certain constant could have, it is safe to say that there is a wide range of possible evaluative tendencies and ways of balancing evaluative variables, and a narrow range that all these factors must fall within to produce true evaluative beliefs. As such, this argument from moral epistemology is analogous to the argument from fine-tuning.

Note that the argument from the correlation between moral beliefs and facts does not require that P(C|Ns) = 0. That is, it does not require that the influence of natural selection on our moral belief leads to moral skepticism. The conclusion only requires that the correlation is more probable given Christian theism than non-theism. Accordingly, this argument is not as easy to refute as debunking arguments. When a debunking argument makes the case that your moral belief or moral belief-forming mechanism is not related to the truth of the subject matter and, therefore, is not justified, it is possible to refute this debunking argument by providing an alternative explanation. However, to refute this argument from reliable moral belief to the existence of God, it is not sufficient to come up with a possible non-theistic explanation of the correlation between moral beliefs and moral facts; one needs to come up with an explanation that is better than the theistic one. That is not an easy task, both because Christian theism gives a good explanation of why we are to expect reliable moral beliefs (more on this in the next section) and because Christian theism gives a good explanation of why the correlation is precisely the way it is; it is neither perfect nor marginal correlation (more on this in the Section 4).

3. Unpacking the Theistic Explanation

When Mark D. Linville provides a theistic explanation of the general reliability of our moral faculty, he appeals to the notion of God as the creator: God has designed human moral faculties to guide human conduct in light of moral truth (Linville 2009, p. 414). Connecting the general reliability of our moral faculty to the notion of God as creator is not an ad hoc move; seeing God as the creator is central to most conceptions of theism. However, the notion of a creator God as such does not give us reason to expect a general reliability of our moral faculty. Alvin Plantinga realizes this when developing a theistic argument from human rationality. He specifies the notion of God as a creator with a Christian notion of what God wanted to create. According to Christian theology, God created humans in his image. A central feature of being created in God’s image, says Plantinga, is the idea that humans are created so that they resemble God. Humans resemble God in being persons, which includes having intellect, will, beliefs, intentions, and affections (Plantinga 2000, pp. vii, 212; 2011, p. 4). Drawing on the theology of Thomas Aquinas, Plantinga is primarily concerned with how this is a resemblance of cognitive capacities so that humans imitate God “especially in this, that he understands” (Plantinga 1993, p. 236). If God wanted to create humans that resemble God, and resemblance includes cognitive capacities, generally reliable cognitive capacities are to be expected.

As Plantinga realizes, it might be an overstatement to say that human likeness to God primarily concerns intellect and understanding (Plantinga 2011, p. 269). In biblical theology, likeness to God is more easily related to morality than rationality. It is common among Old Testament scholars to hold that “the image of God” does not signify likeness in the sense that our faculties are parallel to God’s faculties but rather a vocation given to humanity to live as God’s representatives (Middleton 2005). In New Testament theology, the notion of being like God or like Jesus has moral significance. Jesus, who is God incarnate, urges people to “follow me” and to learn from his example (Mark 1:17, Matthew 11:29). The apostle Paul uses the notion of imitation, urging people to be imitators of the Lord (1 Thessalonians 1:6). While the notion of following Jesus and imitating him cannot be reduced to ethics, it certainly includes ethics. In Paul’s ethics, the key theme is precisely to imitate God after the model of Christ (Witherington 2016, vol. 1, p. 244). So, the notion of likeness to God has a stronger connection to morality than rationality and, therefore, gives a stronger reason to expect a generally reliable moral capacity than a generally reliable cognitive capacity.

Note that a moral theory that emphasizes following the exemplar of Christ—call it moral exemplarism (Zagzebski 2004, p. 232 ff.)—does not suggest that our moral capacity is flawless. If that were the case, no exemplar would be needed. Nor does it indicate that our moral capacity is entirely unreliable. If that were the case, no exemplar would be efficacious. Exemplarism indicates that our moral capacity is reliable enough to recognize and respond to proper exemplars. Our moral capacity is not perfect; it needs guidance and correction, such as by the example of Christ.

Other parts of Christian theology also connect to the human moral capacity and give us reason to expect that this capacity is generally trustworthy. Consider a Christian conception of the goal of human life, namely, a relationship with God. In the church’s history, different kinds of imagery have been used to describe this relationship, such as filiation, friendship, or union (Macaskill 2013, p. 48 ff.). In the same way as likeness to God cannot be reduced to ethics but includes it, relationship with God also cannot be reduced to ethics but includes it. Take the introduction to the Ten Commandments: “I am the Lord your God, who brought you out of Egypt, out of the land of slavery. You shall have no other gods before me” (Holy Bible 1996, Deuteronomy 5:6–7). What we see here is God expressing moral obligations situated in a covenant relation (Hare 2015, p. 148). Having God as Lord is connected to keeping God’s commands (Holy Bible 1996, Leviticus 18:4; Deuteronomy 5:6–7), and loving God is connected to keeping God’s commands (Holy Bible 1996, 1 John 5:3; John 14:15). If God created humans to be in a relationship with him, and this relation includes morality (keeping God’s commands), there is reason to expect that God created humans to have a generally reliable moral capacity.

Another reason to expect a generally reliable moral capacity comes from the notion of God as a giver of vocation. In the biblical material and the Christian tradition, God is seen as calling humans to a certain life or a certain service. Karl Barth, for instance, sees God’s vocations to humans as ad hominem and ad hoc, as to a specific person and a specific situation. God’s calling, then, is “something highly particular” (Barth 1961, p. 499; Biggar 1993, p. 164). This notion of vocation as something depending on a particular time, place, and person means that vocation does not belong to the work of God the creator but, to use Trinitarian terms, the work of the Holy Spirit (Barth 1960, p. 501). Now, the notion of God calling humans to a certain vocation implies that humans are the type of creatures that can receive a vocation. Responding to God’s call is something one ought to do; it is normative. However, depending on what sort of theistic metaethics one ascribes to, God’s call might be a moral requirement, or it might be seen as a wider category (R. M. Adams 1999, p. 303). In any case, the notion of God as a giver of vocation gives us reason to expect that humans can respond to normative matters of how one should live, which includes morality.

To sum up, my point here is that central parts of Christian theology warrant that P(C|God) is high. In relation to Plantinga’s argument from human rationality to theism, I have pointed out that non-theistic evolutionary explanations of rationality are more easily available than non-theistic evolutionary explanations for morality. Theological topics, such as humans being created in the image of God and God as a giver of vocation, are more tightly connected to human morality than general rationality, meaning that the Christian explanation of why P(Moral C|God) is high stands stronger than the explanation of why P(Rational C|God) is high.6 This means that the argument from moral capacity is even stronger than Plantinga’s argument from rationality.

4. The Goldilocks Enigma: Neither Perfect nor Marginal Correlation

Fine-tuning arguments make the case that the laws and constants of physics are finely tuned to exactly match the values required for a life-permitting universe. The match is perfect. When it comes to the correlation between moral beliefs and moral facts, the match is not perfect. A significant number of moral disagreements, both in our time and across history, indicates that, at best, some small minority of human beings have moral beliefs that correspond entirely with the moral truth. So, our moral beliefs do not entirely match moral facts. This is significant for the argument: possible explanations of the correlation between moral beliefs and facts are not good explanations if they point to a perfect correlation. Rather, they should provide an understanding of why the correlation is neither marginal nor perfect. Like Goldilocks, who found a bed that was not too big nor too small, we must find an explanation that grants not too much correlation nor too little.

The difficulty of explaining a correlation that is not perfect is evident in the work of Derek Parfit. Parfit thinks that Street’s premise, namely that natural selection has greatly influenced our moral tendencies, is formulated more strongly than it needs to be. He states that if “normative beliefs were even partly produced or influenced by natural selection”, their justification would be undermined (Parfit 2011, p. 517). However, he takes these evolutionary influences to be outweighed by the ability to reason. Parfit argues that we have gained a trustworthy ability to reason because this ability is evolutionarily advantageous. And the same cognitive ability that enables us to reason also enables us to engage in normative reasoning.7 Now, for our normative beliefs to be not even partly produced or influenced by natural selection or other distorting influences, Parfit puts forth an almost flawless cognitive ability to respond to normative reasons and thereby acquire true moral beliefs.

Parfit’s moral epistemology is very optimistic.8 He proposes a cognitive ability so potent in its task to acquire true moral beliefs that it can safeguard the moral agent from distorting evolutionary influences. While this safeguards his position from debunking arguments, it does not provide a good explanation of the correlation between moral beliefs and moral facts. As the epistemology is so optimistic, it falls short in explaining why the correlation is not at all perfect, namely, why humans are liable to persistent error. Parfit’s moral epistemology illustrates that one needs to do some delicate balancing when constructing a moral theory. On the one hand, recognizing evolutionary influences might lead to skepticism. On the other hand, safeguarding oneself from this skepticism might lead to an overly optimistic moral epistemology. I will now make the case that Christian theology can help us walk this fine line as the theological resources of creation and sin can help us understand moral fallibility without making us doubt whether any of our moral beliefs are true.

According to Christian theology, the human condition is not only determined by being created in the image of God. Humanity is also under the condition of sin. Some theologians have defined sin as “that which God does not want done” (Jenson 1999, p. 133), but that is not to say that sin only describes human acts. Sin is also a description of a state in which human beings find themselves. So, sin is an inescapable and encompassing reality that affects human beings. I want to draw attention to a particular effect of sin, namely what has been called “the noetic effect of sin”, the effect of sin upon the human mind.

Influenced by sin, humans set aims that they ought not to set. According to a traditional Augustinian conception of sin, sin affects humans so that they no longer direct their lives towards God, the highest good, but instead love lesser worldly goods (Pannenberg 1994, pp. 243–45; Wood 2013, p. 20). Lower goods replace higher goods. Instead of living a life directed towards the highest good, humans live lives directed towards themselves, living self-oriented and self-directed lives. So, sin affects human aims. Humanity is in such a condition that self-centeredness comes naturally—as well as gluttony, greed, sloth, and pride—and that we take ourselves to have good reasons for having these aims (Wood 2013, p. 32). So, there is a noetic effect of sin upon human substantive rationality.

Moreover, there is a noetic effect of sin upon human procedural rationality. The seventeenth-century philosopher Blaise Pascal, belonging to the Augustinian tradition, writes about sin as self-deception. Deceiving ourselves has consequences for our theoretical reasoning. While it is often prudent to seek the truth, as that will help us carry out our projects, truth is often unwelcome when it opposes our desires, projects, or aims, which gives us the motive to reject the truth. We may embrace alternative beliefs that might be false but are more welcome because they support our agenda. So, we might deceive ourselves because we do not find truth attractive enough (Wood 2013, pp. 12, 38; Plantinga 2000, pp. 212–13). Such self-deception might have religious relevance; we might create a God in our image. It might have moral relevance. Parfit holds that everyone will be in moral agreement if they are careful in their reasoning and take into account all that is relevant (Parfit 2017, p. 309 ff.). The human tendency of self-deception casts doubts on this claim.9

Humans are fallen, but that is not the whole story. The Christian view is not that sin has eradicated human likeness to God. There is a duality to humanity. Humans are under the distorting influence of sin and, at the same time, display a likeness to God.10 Humans are not all good or all bad, but both great and wretched (Pascal 2008, L122/S155). Therefore, a distorting influence need not warrant moral skepticism, as Parfit worries. Seeing humans as created and called by God gives us reason to say that humans can recognize the good and get some basic moral beliefs right despite distorting influences. So, a Christian view of humanity as both great and wretched might neatly solve the Goldilocks enigma, giving an explanation of our moral capacity that grants not too much correlation nor too little.

5. Possible Natural Explanations: Third-Factor Responses

A popular response to evolutionary debunking arguments, which might also address the correlation problem, is the so-called third-factor response, where the idea is that a third factor, C, explains the correlation between A and B. It is not our moral beliefs that have caused moral facts or moral facts that have caused moral beliefs, but rather a third factor. David Enoch argues that this third factor is survival (Enoch 2011, pp. 168–75). Beliefs that lead to better odds of survival, such as the belief that ‘killing is wrong,’ would be evolutionarily advantageous. Moreover, Enoch argues that survival can be considered a moral good. Therefore, the evolutionary aim of survival (C) leads to a correlation between moral facts (A) and beliefs (B). Erik Wielenberg proposes another third factor, namely our cognitive faculties. He holds that cognitive faculties ground certain moral rights. Moreover, the cognitive faculties also generate the moral belief that we have certain rights. In that way, cognitive faculties (C) are responsible for the correlation between moral rights (A) and beliefs about moral rights (B) (Wielenberg 2014, p. 145).

A common criticism of third-factor responses is that they are circular or question-begging as they include normative premises such as “survival is good” or “beings with cognitive faculties have certain rights” (Crow 2016; Vavova 2015). While we believe that survival is good, it seems suspiciously question-begging to grant such moral beliefs as it is precisely such moral beliefs the debunker calls into question. Wielenberg defends third-factor responses from allegations of question-begging by clarifying who has the burden of proof (Wielenberg 2016, p. 5). The debunker wants to give us reason to think we might be in moral error. The burden of proof, then, lies on her. These third-factor responses aim not to provide fresh justification for our moral judgements but to undercut the debunker’s defeaters for our moral judgements. Now, the strategy of seeing third-factor responses as undermining debunking arguments might work when we consider the justification of our moral beliefs. But when it comes to explaining the correlation, a positive case must be given. The burden of proof, then, is back on the third-factor respondent.

How well a third-factor response explains the correlation between moral beliefs and moral facts (A and B) depends on how well C is connected to A and B. The correlation is very well explained if C entails both A and B. If C merely suggests A and B or makes their occurrence more probable, the correlation between them would still, to some degree, be a lucky coincidence, meaning that the explanation is not satisfactory. Here, I suggest, lies a difficult trade-off for a naturalistic explanation of the correlation. A strong entailment connection between A and B does not seem to be plausible, while a weaker connection leaves the correlation unexplained.

Enoch holds that the aim of evolution, namely survival, gives rise to certain moral beliefs without entailing them. He argues that natural selection has shaped our cognitive faculties to value survival appropriately. However, evolution could have proceeded differently. Had the aim of evolution been different, so would our normative sensitivity. Had the forces shaping our cognitive faculties been different, so would our normative sensitivity, in which case we would have been systematically mistaken in our normative beliefs (Enoch 2011, p. 173). So, Enoch holds that the third-factor response grants some understanding as to why our moral beliefs correlate with moral facts. The correlation is not an entirely lucky coincidence. However, given the evolutionary contingency, we must consider ourselves epistemically lucky to have evolved in the way we have, meaning that “some brute luck may remain” (Enoch 2011, p. 173).

Wielenberg suggests that the moral realist can get rid of the luck by getting rid of the contingency in how C is related to A and B. He argues that there is no contingency in how our cognitive faculties (C) are related to moral truth. Basic moral truths are necessary truths, such as the truth that beings with certain cognitive faculties have certain rights. What, then, about how our cognitive faculties (C) are related to moral beliefs? Given Street’s reasoning, the evolutionary forces forming our cognitive faculties could have given rise to a wide range of moral beliefs. Wielenberg doubts that this is correct. First, he asks the reader: could the evolutionary forces, operating within the constraints of the actual laws of nature, have produced rational beings with moral beliefs radically different from our own? The answer is no. Street’s hypothetical evolutionary scenarios might be metaphysically possible but not possible given the actual laws of nature (Wielenberg 2014, p. 170). Now, if the laws of nature are contingent, the correlation would still be contingent. Wielenberg, then, goes on to argue that the laws of nature are not contingent but necessary. In that case, our cognitive faculties (C) and true moral beliefs have a very stable connection.

Exactly how stable is this connection? In an article from 2010, Wielenberg seems to suggest that the relation amounts to entailment. If the laws of nature are necessary truths, the correlation between moral facts and beliefs is also “a necessary feature of the universe” (Wielenberg 2010, p. 461). This eliminates any lucky coincidence: “Where there is no contingency, there are no coincidences” (Wielenberg 2010, p. 461). However, in his Robust Ethics from 2014, he clarifies that he does not believe that the laws of nature would grant us the cognitive capacity to form all and only true moral beliefs (Wielenberg 2014, p. 168). Rather, he argues that evolution could not have produced rational beings with moral beliefs radically different from our own. While Street argues that evolution could have given rise to a wide range of moral beliefs, Wielenberg holds that the possible range of moral beliefs is “much more limited than Street suggests” (Wielenberg 2014, p. 173).11

Wielenberg’s reasoning illustrates a difficult trade-off for the naturalistic explanation of the correlation. A stronger entailment connection grants a perfect (perhaps too perfect) correlation between moral facts and beliefs but is not that plausible. A weaker relation is more plausible but makes the correlation luckier. While Wielenberg, in the end, seems to hold that there is some degree of luck involved (Wielenberg 2014, pp. 173, 175), I still find third-factor explanations to be among the strongest contestants for an explanation of the correlation between moral facts and beliefs. However, a theistic explanation has some clear advantages over this naturalistic explanation. Let me mention three points.

First, arguing that the correlating is not as lucky as first assumed is not the same as explaining. Enoch, who recognizes that there is some luck involved in the correlation, admits that his way of coping with this epistemological challenge is not ideal (Enoch 2011, p. 175; Johnson 2023, p. 188). The reason why this way of coping with the problem is not ideal might become clearer when the epistemological challenge is compared to the fine-tuning argument. Fine-tuning arguments build on the thesis that there is an extremely wide range of possible universes but only a very narrow range of universes suited for life. Now, if one were to argue that the range of possible universes is not extremely wide but only somewhat wide, that would not explain the fine-tuning; it would only somewhat reduce the need for an explanation. In the same way, arguing that the possible range of moral beliefs is much more limited than Street suggests might make the correlation less lucky and less surprising, but there is still a correlation to explain.

Second, third-factor explanations are much more coarse-grained than theistic explanations. Enoch argues that moral truths have an evolutionary link to survival. Certain beliefs, such as “killing is wrong”, would push us to act in ways that promote survival. However, this moral belief is rather coarse-grained.12 A fine-grained moral truth would be Kant’s formula of humanity, stating that you should not treat persons merely as means. An evolutionary push towards survival does not explain these kinds of fine-grained moral truths. While Kant’s maxim will lead to less killing and more survival, so would lots of other moral beliefs, including beliefs where persons are regarded as means to some end (Van Eyghen and Bennett 2022, p. 128). Now, Enoch says that survival relates to a wide range of other normative truths so that this true normative belief enables us to construct a coherent system of more fine-grained moral beliefs (Enoch 2011, p. 169). However, such a solution might underestimate the degree of fine-tuned moral parameters mentioned above. Both Enoch’s belief that survival is good and Wielenberg’s cognitive faculties allow for numerous ways of balancing egoism and universalism, balancing morality and prudence, and cohere with a wide range of concrete ethical codes such as sexual ethics. Thus, these third factors cohere well within a wide range of moral belief systems, meaning the third-factor response only explains a coarse-grained correlation. A theistic explanation, however, has the resources to give a more fine-grained explanation by appealing to how God has created the cognitive abilities of humans with a goal in mind and revealed his will by calling humans to particular acts of service.

Third, the appeal to brute facts is not very illuminating. When Wielenberg makes the case that cognitive abilities (C) relate to moral facts in the way that cognitive abilities ground certain rights, he appeals to laws of nature and supervenience relationships being brute necessary truths. Moreover, when making the case that cognitive abilities (C) firmly relate to moral beliefs, he appeals to necessary laws of nature. Wielenberg might hold that introducing God does not add much to the explanation of the correlation. He states that one could argue that God established both the supervenience relationships and the laws of nature, but equally satisfactory, one could contend that the supervenience relationships and the laws of nature alone suffice (Wielenberg 2010, p. 460).13 Now, when Lipton considers how we identify the best explanation, the one that is most likely true, he asks us to consider which explanation provides the most understanding (Lipton 2004, p. 59). Appealing to brute facts does not do much to increase our understanding. To draw on the fine-tuning argument again: in the same way as stating “the fine-tuning is just a brute fact” does not provide much understanding, it does not provide much understanding to state that both moral facts and our moral beliefs are the way they are due to brute necessary facts. Christian theology, on the other hand, provides an understanding both as to why we can acquire true moral beliefs and why certain things are valuable. Particularly, Christian theism provides an understanding of the tug between being able to know the good while simultaneously being liable to persistent error (both regarding knowing and acting).

6. Conclusions

In this paper, I have developed a theistic argument from moral epistemology that mirrors the structure of the fine-tuning argument. The core of the argument is that Christian theism better explains the correlation between moral facts and moral beliefs than naturalistic alternatives. I take the argument to be stronger than Plantinga’s theistic argument from rationality, as an evolutionary explanation of human rationality is more easily available than evolutionary explanations of our moral capacity and also because a Christian theology concerning God as the creator and giver of vocation is more tightly connected to human morality than general rationality. When considering possible naturalistic explanations of the correlation between moral facts and beliefs, I argue that Parfit’s optimistic rationalistic approach falls short as it fails to account for persistent moral error. The third-factor explanations of David Enoch and Erik Wielenberg fall short as they still leave some room for the correlation being lucky, are more coarse-grained, and provide less understanding than a theistic explanation.