Reading Time: 12 minutes

Atheists have an image problem.

Two previous articles covered one potential reason for the image problem—namely, that atheism is a concealable identity, and the most visible atheists differ from typical atheists in a number of ways. Some differences are subtle, reflected in slightly different personality profiles between typical atheists and very online atheists, for example. Other differences are quite stark, such as the vast gulf in social and political attitudes between most atheists and the occasional culture war musings of the world’s most prominent celebrity atheists. I pointed out that these mismatches can drive misperceptions of atheists. People may form impressions of atheists based on the most visible examples—even though these visible atheists aren’t especially representative of atheists on the whole.

This article argues that although public (mis)representation of atheism might contribute a small amount to atheism’s image problem, the main drivers of anti-atheist sentiment come from much more fundamental psychological and cultural forces. Mismatches between the most visible atheists and the most typical atheists are surface ripples that can conceal a much deeper current that drives suspicion of atheists. This deep and powerful current? Our intuitions about morality, and religion’s role in it. Let’s take a quick tour of a decade’s worth of research on anti-atheist prejudice.

Anti-atheist prejudice, and what social psychology doesn’t easily explain about it

In a landmark paper, Penny Edgell and colleagues conclude that “atheists are less likely to be accepted, publicly and privately, than any others from a long list of ethnic, religious, and other minority groups.” People report that they wouldn’t like to vote for an atheist candidate for President, that they wouldn’t like their kid to marry an atheist, and that atheists are the group who least agrees with their vision of the USA.

Why are atheists so disliked?

The main drivers of anti-atheist sentiment come from fundamental psychological and cultural forces.

A classic social psychology approach to prejudice and stereotyping stems from social identity and realistic conflict. The basic idea here is that we humans are great at sifting ourselves into groups. We categorize ourselves and others by our numerous social identities. Personally, I’m simultaneously a father, a dog lover, a psychology professor, a fan of the Colorado Avalanche hockey team, a born Coloradan, an expat living in London, and numerous other things. These are my social identities. We all have them. Over time, we tend to affiliate with people who belong to the same categories as us. We might start to avoid or even denigrate people belonging to other, conflicting, social identities. Problems get amped up when there are tangible or symbolic stakes that different social groups might be in realistic conflict over. Social identities can easily become allegiances in intergroup conflicts. The general prediction of this approach is that people will come to like similar others and dislike dissimilar others, leading to ingroup love and (potentially) outgroup hate.

This social-identity-then-conflict model is a pretty standard social psychological approach, and it does a great job explaining lots of things about lots of types of prejudice. But it doesn’t obviously explain anti-atheist prejudice. For one, atheism isn’t much of a social identity. Religious groups are salient and obvious social identities. Catholics or Sunni or Sikhs or Shaivists are all clear religious groups with shared norms and practices and beliefs. Atheists? Not so much—learning that someone’s an atheist doesn’t tell you much about what they do, just one thing they don’t believe. To paraphrase a Ricky Gervais quip, saying atheism is a social identity is like saying not going skiing is a social identity. It’s just something that people don’t believe and tells us little about their other beliefs or practices or values. Atheism is a negatively defined category, not an identity for most people. Because atheism isn’t a salient social identity, social identity processes are somewhat nonstarters for explaining anti-atheist prejudice.

From prejudice to prejudices

Starting in the 2000s, some evolutionary psychologists (like Steve Neuberg and Cathy Cottrell and Mark Schaller) began arguing that it doesn’t make much sense to flatten all prejudice down to a single dimension running from ingroup love to outgroup hate. Instead, we might associate different sorts of people with different sorts of specific functional threats. Different functional threats require specific functional responses. We see some people as threats to physical safety and react to them with fear. Other people might be seen as pathogen risks, and we react to them with disgust and avoidance. The key insight here is that we might not have some unitary psychological process that leads to prejudice, so much as distinct and nuanced reactions that produce myriad different prejudices against different groups, commensurate with the threats each is assumed to pose.

According to this view of prejudice, if you want to understand why a given group is disliked, you first need to consider what specific threats that group might be perceived as posing. Then you can form hypotheses about what particular form prejudice against them might take.

What threat might atheists be perceived to pose?

No God? No good

To understand why atheists might be seen as a threat, it’s worth taking a conceptual step back to ask what functions religion might serve at a cultural and evolutionary level. When I give talks to atheist groups about this research, this is the point where people start to squirm. How can I propose that religion might be functional in an evolutionary sense? Isn’t religion just a toxic memetic mind-virus that hijacks credulous minds and must be rationally vanquished?

I’ve previously debunked some of these ideas using our best available science, so I won’t retread that ground. Instead, I recommend a bit more evolutionary curiosity.

Religion does seem to be a core part of human nature. Every known culture has had some form of religious beliefs, and until relatively recently, almost all people did too. Religion is as much a product of evolution as are rhinoceroses and robins and ribonucleic acids. The fact that rhinoceroses and robins and ribonucleic acids are all finely tuned evolutionary products doesn’t make them good, or bad—they just exist, and have specific functional adaptations that let them continue to exist. Similarly, religions being part of human nature does not make them good, or bad, or necessary, or inevitable. But it should spur scientific curiosity.

One prominent account of religion’s evolutionary success (outlined in books like Ara Norenzayan’s Big Gods, Joe Henrich’s the WEIRDest People in the World, or David Sloan Wilson’s Darwin’s Cathedral) is that religions are so widespread in part because some of them have proven quite successful at promoting cooperation within groups. Religious practices and beliefs and norms might have helped our ancestors scale up from small bands of people who largely cooperate with close kith and kin to large ultrasocieties that see widespread impersonal and even anonymous cooperation among people who’ve never met and will never meet again. This is a drastic shift in cooperative psychology, taking place over a relatively short timespan (maybe just a dozen or so millennia)! And large-scale cooperation is no easy evolutionary nut to crack. Within groups, there’s always the temptation to defect and freeride. In a largely cooperative group, solitary moochers and schemers can extract benefits from others’ labor without kicking in their fair share. This gives them a competitive advantage, and without mechanisms for curbing selfish freeriding, large-scale cooperation can’t get off the ground. Modern innovations like courts and police can’t be the answer—it takes cooperation to create these institutions, leaving our original puzzle still unsolved.

Big Gods

Some religions might’ve helped ease cooperative friction. Religions differ a lot in terms of their beliefs, norms, and practices. Some religions imagine local amoral gods; others imagine universal and morally concerned Gods. Some religions impose few practical costs on adherents; others mandate celibacy or dietary requirements or costly and painful rituals. Religions that include gods who are morally concerned and vigilant—Big Gods, in Norenzayan’s titular terminology—might give cooperation a boost. Big God religions tend to have firm moral commandments about what one ought and oughtn’t do. They also come with a built-in surveillance and enforcement system: God, after all, is always watching and occasionally quite wrathful. People might be willing to cheat a neighbor who will never find out, but will they similarly cheat if they believe in an all-seeing God with the power to smite them for their misdeeds?

Groups that practice Big Gods religions might’ve gotten a head start when it comes to solving cooperative problems. As a result, Big God religions rapidly spread and seem to have outcompeted little god religions in the arena of cultural evolution. One upshot of this is that most people on earth these days belong to a Big God religion. After all, these religions were the “winners” of a cultural evolutionary arms race where cooperation was the most coveted weapon.

The line between religion and morality gets blurred in Big Gods societies. Eventually, they may even be seen as synonymous – what one ought to do is often inseparable from what God wants. And here, I think, is the deep psychological and cultural root of anti-atheist prejudice.

Belief in Big Gods doesn’t just help individuals inhibit their own impulses, it can become a badge of who is morally on the up-and-up. To an adherent of a Big God religion, coreligionists can obviously be trusted. They share one’s beliefs and norms, and share the same fears of being smitten by the same vengeful God. Even believers in other religions can be extended some trust. They might not believe in my God, so the thinking goes, but at least they have a religiously-prescribed moral code and fear punishment by their own morally-concerned deity. Trust, morality, and religious belief, all intertwined.

Thus far, we have two different strands of research: one line of research says that distinct prejudices form in response to distinct functional threats, the other line says that cultural evolution led to a gradual blurring of the lines between religion and morality. Let’s combine these two strands.

The centrality of religion to cooperation and morality over cultural evolutionary time suggests that atheists might be seen as specifically cooperative and moral threats. Atheists aren’t big and scary. They aren’t gross and disease-ridden. They are moral wildcards, unworthy of trust.

If moral distrust is, as this story goes, central to anti-atheist prejudice, we might make at least two different predictions. First, we could predict that people would spontaneously associate immorality with atheism. Second, we could predict the same general pattern, but running in the other direction: atheists might be spontaneously seen as immoral.

Without God…everything is permitted?

In Dostoevsky’s The Brothers Karamazov, a character opines that without God and the afterlife, everything is permitted. “Everything” here happens to include grisly immoral deeds like murder. A cultural linking of religion and morality might make people intuitively assume that people who do immoral things are not religious believers. A properly God-fearing man simply wouldn’t do immoral things, after all (so goes the intuition, at least). To the extent that people see belief in Big Gods as a handy tool for inhibiting our basest urges, they may assume that people engaging in immorality must not believe in a God. In short, people hearing about immorality might spontaneously infer atheism.

To study this possibility, my colleagues and I have turned to some work by Nobel Laureates Daniel Kahnemann and Amos Tversky on what they call the representativeness heuristic. Their classic demonstration of the representativeness heuristic gave people the following puzzle:

Linda is 31 years old, single, outspoken, and very bright. She majored in philosophy. As a student, she was deeply concerned with issues of discrimination and social justice, and also participated in anti-nuclear demonstrations.

Which is more probable?

A)  Linda is a bank teller.

B)   Linda is a bank teller and is active in the feminist movement.

Most people tend to pick option B, which turns out to be incorrect. Option A merely says that someone is a bank teller. Option B describes someone who is both a bank teller and is active in the feminist movement. Every single person who belongs to Option B also belongs to Option A, which means that A is necessarily more probable than B. On this problem, people get led astray because Option B just sounds like someone who might fit the initial description of Linda. They view the description of Linda as representative of the target group membership implied in Option B. It’s not a conscious or rational thing, just an intuition. But it only happens when there’s an intuitive fit between the provided description of Linda and the group membership implied in Option B.

To make this concrete, consider the following results: given the description of liberal Linda above, something like 60% of people pick Option B if it implies that Linda is a feminist, but precisely 0% of people pick Option B if it implies Linda is an avid big game hunter. There’s just no intuitive linkage between antinuclear demonstrations and big game hunting, whereas there is a nice intuitive fit between liberal activist Linda and feminism. By separately experimentally manipulating the description given and group membership implied by Option B, clever researchers can infer the degree to which people intuitively find different descriptions representative of different groups of people.

We’ve capitalized on this pattern by giving our research participants lots of versions of this task in which we describe someone doing something immoral and then plug in lots of different potential group memberships as Option Bs. In doing so, we can see which groups of people are intuitively associated with different sorts of moral violations.

The results? Time and again, people intuitively assume that the perpetrator of immoral deeds is an atheist. Someone kicking a puppy? Atheist. Dabbling in light cannibalism? Atheist. Cheating at cards? Atheist. Having sex with a dead chicken? Atheist. Incest? Oh, definitely an atheist. One time we even described a priest who sexually abused children. Our participants’ intuition was that the priest was secretly an atheist!

One time we even described a priest who sexually abused children. Our participants’ intuition was that the priest was secretly an atheist!

In a big cross-cultural project, we sought to test whether people intuitively find even extreme moral violations representative of atheists. To do so, we gave people in 13 countries the following vignette:

When a man was young, he began inflicting harm on animals. It started with just pulling the wings off flies, but eventually progressed to torturing stray cats and other animals in his neighborhood.

As an adult, the man found that he did not get much thrill from harming animals, so he began hurting people instead. He has killed 5 homeless people that he abducted from poor neighborhoods in his home city. Their dismembered bodies are currently buried in his basement.

Which is more probable?

A)  The man is a teacher.

B)   The man is a teacher and…

Option B ended in different ways for participants in our two experimental conditions. Half of the participants got an Option B that ended “is a religious believer” and half got “does not believe in any gods.” Across all thirteen countries in aggregate, and within each country except Finland and (to a lesser extent) New Zealand, people intuitively assumed that the serial murderer was an atheist. Moral distrust of atheists was most pronounced in highly religious societies like India and the United Arab Emirates, but was still quite evident in secular strongholds like the Netherlands and Czechia.

When we analyzed the data for these studies, it was also possible to dig deeper and see which participants across countries were most suspicious of atheists. As one might expect, moral distrust of atheists on this task was most pronounced among devout believers. Perhaps surprisingly, however, even atheists intuited that serial killers must not believe in God.

Getting back to our Dostoevsky reference, people really seem to have the intuition that without belief in God, everything is permitted…up to and including animal cruelty, incest, necrobestiality, and serial murder.

The face of atheism

People seem to spontaneously infer atheism from immorality. What about the opposite intuition? People read atheism into immorality, but do people also spontaneously infer immorality when they merely imagine atheism?

To answer this question, my colleagues Dr. Jazmin Brown-Iannuzzi and Steph McKee devised a test using a nifty method called the reverse correlations procedure. This experimental psychology task lets researchers get a glimpse of how people spontaneously mentally represent different people. It gives scientists a quick sketch of how people view others. The task is an odd one, but the results are fascinating.

In a reverse correlation study, a bunch of participants are shown a series of side-by-side images. The images start out as fairly generic faces that are distorted with a bunch of random visual noise. Participants in these studies are asked to choose which of the two side-by-side images looks more like it belongs to a given category.

In our case, imagine that you’re a participant. You get shown the following two images and you’re asked to pick which one looks like someone who doesn’t believe in God.

You do this a couple hundred times. Each choice feels a little silly. The images barely look like faces, let alone specifically atheist faces. But after you’ve done this a couple hundred times, researchers can take all the images you selected as looking a bit atheist-ey and merge them together. Then they can merge your atheist face with the faces chosen by hundreds of other participants. This creates an aggregate image of what lots of people seem to think an atheist must look like.

We did exactly this, asking different sets of people to either pick the face that didn’t believe in God or the face that did. The end result was two different aggregate faces. One is the aggregate image depicting people’s spontaneous mental images of atheists; one is the aggregate image depicting a spontaneous mental image of a theist. Here are the striking results, atheist on the left:

Next, we asked a separate group of participants to rate the faces on a number of dimensions. Crucially, this second batch of participants was told nothing about how the faces were generated; they weren’t told anything about how one face supposedly represented an atheist and one a theist. This second batch of participants thought that the atheist face looked less religious than the theist face—this is a nice validity check for the procedure. But even more so they rated the atheist face as less trustworthy and less moral than the theist face.

The reverse correlation task lets researchers see how people spontaneously picture different people. When we generate an image of how people represent atheists, the resulting face is rated as less religious, moral, and trustworthy than a spontaneously generated theist face. Just imagining atheists, people spontaneously project an image of immorality.

Anti-atheist prejudice as moral distrust

Atheists get a bad rap, and their rap is considerably worse than standard social psychological approaches like social identity theory can easily predict. Instead, it might be useful to think of anti-atheist prejudice as a specific reaction to the unique moral threat atheists are perceived to pose. Religions may have helped scale up human cooperation, and one result of this is that people now conflate the concepts of religion and morality, assuming that belief in God is either synonymous with or at least necessary for morality. Within this intuitive worldview, atheists are potential moral wildcards who cannot be trusted. The result? People readily infer atheism from moral missteps, and people spontaneously “see” immorality and untrustworthiness stamped onto the faces of atheists they imagine.

Avatar photo

Dr. Will M. Gervais is a global leader in the scientific study of atheism and a Senior Lecturer of psychology at the Centre for Culture and Evolution at Brunel University London. Dr. Gervais’s research...