The cognitive basis of theism

Part Of: Religion sequence
Related To: The Logic of Mindreading
Content Summary: 3000 words, 30 min read

Religion as Natural

Most human beings in recorded history have participated in religion

There is no consensus definition of religion; but we can still put our finger on a family resemblance. Theistic religions tend to…

  1. Include a belief in spiritual beings, such as ghosts, angels, ancestor spirits, and so on. These entities often have mental lives, but no physical form. 
  2. Posit an afterlife.
  3. Affirm the purposeful creation of the universe, including humans and other animals. 

You are not going to find a place, anywhere, where such notions feel absurd. Yet despite the ubiquity and family resemblances, the content of religious traditions vary immensely across cultures and across history. 

Perhaps religion is similar to language. All societies have at least one language; all societies have at least one religion. Also like language, religion is not present at birth. It develops instead through immersion in a social environment. The specific language or religion that a child develops is determined by the culture in which the child is raised, not by genes or the physical environment. But there are universals of language. Every language has words and sentences, as well as principles of phonology, morphology, and syntax. And as we saw above, there are also universals of religion. 

In sum, the ubiquity and family resemblance of religious traditions point to a shared cognitive basis, but theodiversity reminds us these intuitions support constructs subject to cultural evolution.

Today, we will be exploring the cognitive basis of religion. Unlike language, a unified faculty which directly promotes biological fitness, we will learn how religious instincts are byproducts of human social machinery. 

Mind perception enables God perception

There is a cartoon depicting an alarmed husband telling his visibly upset wife, “Of course I care about how you imagined I thought you perceived I wanted you to feel.” While most social interactions are fortunately not as convoluted, social life does require mindreading. You can think of the mindreading faculty as a little machine in your brain, which guesses at the beliefs and desires of others.

Across cultures, supernatural agents are described as having minds. Contrary to some theological doctrines that cast God as an abstract universal force, Ground of Being, or the totality of everything, God has a remarkably human-like mind in the natural religion of the faithful. God can get angry when His followers do not act in accordance with His desires. Ancestor spirits will be pleased when they receive the proper signals of fealty and allegiance. Zeus has plans for humans that often upset his wife Hera.

Representing the mind of a being that your sense organs don’t perceive may require particularly potent mindreading abilities. Americans, Czechs, and Slovaks with better mindreading abilities are more likely than others to believe in God (Willard et al, 2020 but see Maij et al, 2017). 

Women are more religious than men. Ever wonder why? Well, women are on the average better at mentalizing and empathy. Once we adjust for men’s inferior abilities, women and men don’t differ in their belief in supernatural agents (Norenzayan et al, 2012).

If religiosity relies on mindreading, we might expect damage to its circuitry to affect religiosity. And indeed, autistic individuals are on average much less religious (Norenzayan et al, 2012). 

If religiosity relies on mindreading, we might expect theological reasoning to emerge at the same life stages as ordinary social reasoning. 3yo reasoning falls prey to a reality bias— in verbal reasoning, these children fail to distinguish between the state of reality and peoples’ (sometimes inaccurate) mental states. 4.5yo children become capable of reasoning about the false beliefs of other agents.

Do children intuitively treat supernatural agents as earthly? Lane et al (2010) answers in the affirmative. As children begin to appreciate human beliefs as fallible, they initially attribute similar limitations to all agents, including agents with exceptional powers (in the experiment, God and Mr. Smart). Only later do children learn to overcome their intuitions and affirm depictions of these “super-agents”.

Adults show this same penchant for viewing God as having an essentially human mind. Barrett and Keil (1996) show that even adults, reading a story of God in transcendent terms, think of God as having human-like mental limitations. 

If religiosity relies on mindreading, we might expect thinking about, and praying to, God should activate brain regions associated with mindreading. Indeed, that is precisely what was found in e.g. Schjoedt et al (2009)

Attachment to God

For obvious evolutionary reasons, babies don’t just care about their physical needs, but also the relationship with their caregivers. The attachment system is responsible for promoting these crucial relationships. The Strange Situation task nicely illustrates how it works:

An infant and its caregiver are taken to a room full of toys.  A stranger enters the room, interacting with the caregiver.  The caregiver leaves the room for a few minutes.  Then, after a while, the caregiver returns.  The infant’s behavior is carefully recorded.

How would you anticipate children to respond? It turns out that responses fell into three categories:

  1. Some children did not acknowledge the caregiver, seemingly only interested in play. 23% of all children exhibited this avoidant attachment 
  2. Some children clung to the caregiver, refusing to let go & resume play for some time. 15% of all children exhibited this anxious attachment.
  3. Some children came to hug the caregiver, and then peacefully resumed play. 62% of all children exhibited this secure attachment. 

These attachment styles have been shown to correlate with the emotional availability of the caregivers. A parental style of consistent care promotes security; inconsistent care promotes anxiety; consistent lack of care promotes avoidance. 

The attachment system doesn’t just facilitate parent-child relationships. It is also the substrate of romantic love, and generally friendships of all kinds. Which leads to a question: do believers use their attachment system to relate to God?

Granqvist et al (2010) adduces evidence for the affirmative. Consider this quote from Mother Teresa:

Since age 49 or 50 this terrible sense of loss—this untold darkness—this loneliness, this continual longing for God—which gives me that pain deep down in my heart—Darkness is such that I really do not see . . .—the place of God in my soul is blank—There is no God in me—when the pain of longing is so great—I just long & long for God—and then it is that I feel—He does not want me—He is not there—. . . God does not want me—sometimes I just hear my own heart cry out—“My God” and nothing else comes. 

It seems at least some believers experience relational attachment to gods, ancestors, etc. Just as people seek out other human minds in time of loneliness and stress, these are the times they are uniquely likely to affirm their religious beliefs. 

  1. Many unpredictable negative events (including illness, injury, fatigue, frightening events, and separation from loved ones), activate the attachment system and activate a search for minds. These same situations cause people to turn to God (Gray & Wegner, 2010).
  2. Subtly primed religious concepts buffer religious participants against the negative consequences of laboratory induced social isolation (Aydin et al 2010).  

Supernatural attachment is a bit unusual, however. Generally speaking, egocentrism describes the phenomena of, when we don’t have much knowledge of other people’s views, we often attribute our own values & beliefs onto them. Epley et al (2009) show that this phenomena happens even more frequently for people’s relationships with gods. 

Dualism, Disembodied Minds, and the Afterlife

Mind-body dualism represents the notion that minds and bodies are completely separate substances. Western thinkers often associate dualism with Rene Descartes. But cross-cultural research suggests that adult intuitions about disembodied minds are strikingly similar across societies (Cohen et al., 2011; Roazzi et al, 2013). Recent work has even found dualist thinking in ancient Chinese texts (Slingerland & Chudek, 2011). 

Further, children’s belief in the afterlife (at least in Western cultures) gets weaker with age, not stronger (Bering, 2006). While enculturation often does weaken these intuitions, mind-body dualism re-emerges in adults under cognitive load (Forstmann & Burgme, 2015).

Dualistic intuitions are a human universal. But why? Why should movies about Freaky Friday feel intuitive?

Bloom (2007) posits the independent system hypothesis. On this view, our dualism is a natural by-product of the fact that we have two distinct cognitive systems, one for dealing with material objects, the other for social entities. These systems can operate independently, and have incommensurable outputs. Hence dualism emerges as an evolutionary accident.

  • For example, Kuhlmeier et al (2004) tested 5-month-olds’ ability to reason about the law of continuous motion as it applies to the human body.  For inanimate objects, infant are surprised (i.e., look longer) when the object disappears from behind one barrier and then seems to reemerge from a separate barrier. Not so for humans!  This suggests that babies interpret humans with radically different assumptions, than they do other objects.
  • Indeed, our ability to imagine a disembodied mind often emerges early in childhood. Taylor’s (1999) research on children’s propensity to maintain social relationships with imaginary friends suggests that by age 3 to 4 years, children are already equipped to sustain vivid mental representations of the wants, opinions, actions, and personalities of such an agent. Of course, children do understand their imaginary friends as fictive (not so for adult beliefs about God). 

Most social interactions require physical perception and mindreading processes to execute in parallel. But mindreading supports offline processing; that is, we can think of others not in the room. On the independent systems hypothesis, offline processing makes sense – only one system is operational.

Sigmund Freud once said,

Our own death is indeed unimaginable and whenever we make an attempt to imagine it we can perceive that we really survive as spectators. 

The consensus view of mindreading is that it relies, at least in some ways, on simulation to attribute mental states to others (hence, the egocentric bias). This mechanism is uniquely poorly equipped to contemplate death. In line with this simulation constraint hypothesis, Bering (2006) found that that most undergraduate students, who later claimed to believe that consciousness stops at death), nevertheless stated that the dead person knew he was dead. 

The most dramatic demonstration of dualism concerns the development of afterlife beliefs. Bering and Bjorklund (2004) told children of different ages stories about a mouse that died, and asked about the persistence of certain properties. When asked about biological properties of the mouse, the children appreciated the effects of death, including that the brain no longer worked. But when asked about the psychological properties, most the children said that these would continue – the dead mouse can think thoughts, and hold desires. The body was gone, but the soul survives.

Death is thus processed biologically, and does not terminate social inference systems. This fact will be familiar to the bereaved.

Thus, the afterlife and immaterial agents are innately-endowed intuitions. But these disembodied minds are felt to have all-too-human mental lives. 

Mythology and MCI Theory

Innate dualism is explained by the independent system hypothesis. This in turn relates to the disjoint core knowledge hypothesis: there seem to be distinct ontologies built into the nervous system: objects vs organisms vs minds. Each ontology is associated with a distinct set of modules and the intuitions they generate. Most social interactions engage both biological and psychological modules; intuitions about disembodied minds express themselves when only the latter modules are active.

The disjoint core knowledge hypothesis allows us to better understand mythology. Artifacts of Paleolithic religion contain counterintuitive entities: stuff that blurs ontological boundaries:

Folklore like Grimm’s Fairy Tales are also rife with such counterintuitive entities (e.g., a talking wolf).

Too many counterintuitive entities produces a confusing story. But why have them at all? The simple answer is memorability: intuitive concepts are easy to grasp, but are quickly forgotten; a small number of counterintuitive entities increases the salience (and the memorability!) of the myth.

The disjoint core knowledge hypothesis explains why so many of our myths are seasoned with minimally counterintuitive (MCI) ingredients.

Two Teleological Instincts

American 4- and 5-year-olds differ from adults by finding the question “what’s this for?” appropriate not only to artifacts and body parts, but also to whole living things like lions (“to go in the zoo”) and nonliving phenomena like clouds (“for raining”). Kelemen (2004) shows how, around this age, children adopt a design-based teleological view of objects with increasing consistency. 

This trend seems to be innate rather than cultural. A study of responses young children receive when asking questions about nature indicates parents generally favor causal rather than teleological explanation. (Kelemen et al 2002). While physical explanations are culturally favored, teleological thinking does re-emerge in adults under cognitive load (Kelemen et al, 2009).

This form of object teleology (“what’s it for?”) is theorized to be a byproduct of artifact cognition. Great apes are toolmakers. Part of our ability to understand such artifacts is the capacity to reverse-engineer their purpose: to infer specific goals and motivations of their creators.

In contrast with object teleology, autobiographical teleology (“purpose in life”) ascribes purpose to significant life events. We know that counterfactual reasoning is used to construct a sense of purpose (Kray et al, 2010). The stories we tell ourselves about these events coalesce into our narrative identity during late adolescence, and casts a strong influence on life satisfaction (McAdams & McLean, 2013). There is a correlation between autobiographical teleology and religiosity, but the nature of this relationship is unclear to me. 

Anthropomorphism generate paranormal beliefs

In 1976, NASA’s Viking 1 was orbiting Mars, exploring the surface for possible landing sites. Here’s one of its pictures, in the Cydonia region:

Anthropomorphism, the penchant for hallucinating agents in natural environments, was recognized by David Hume as a pervasive human weakness:

There is an universal tendency among mankind to conceive all beings like themselves, and to transfer to every object, those qualities, with which they are familiarly acquainted, and of which they are intimately conscious. We find human faces in the moon, armies in the clouds; and by a natural propensity, if not corrected by experience and reflection, ascribe malice or goodwill to everything, that hurts or pleases us.

Daniel Dennett has made much of this instinct, dubbing it the hyperactive agency detection device (HADD) module. But Willard et al (2020) found anthropomorphism had zero effect on belief in God. Indeed, they were weakly negatively correlated, presumably because of the Abrahamic faith’s emphasis on aniconism. However, a person’s anthropomorphic bias does predict that individual’s beliefs about the paranormal (ghosts, aliens, etc).

The Four Pillars of Religiosity

Mindreading plays a central role in religious instincts. Its influence is largely mediated by more specific intuitions about mind-body dualism, promiscuous teleology, and anthropomorphism.  The specific relationships between these factors has been mapped:

Dual-process theory posits human cognition can be understood and implicit intuition vs explicit reflection. Since religion is rooted in social intuitions, we might predict that people who rely more extensively on their intuitions are more religious. And that is precisely what we find. 

This post locates the cognitive basis of religion in three mindreading-based intuitions, together with the penchant to take such intuitions at face value. But we will see later, there are at least two other drivers of religiosity. First, religion is promoted by cultural transmission via CRedibility Enhancing Displays (CREDs). Second, religion tend to flourish under conditions of existential insecurity, which is well illustrated by the proverb “there are few atheists in foxholes”.

These four elements promote religious belief. If any of them are missing, weakened or overriden, individuals will tend towards atheism (Norenzayan & Gervais, 2012).

But these factors are not created equal. The size of their contributions have also been extensively researched. While the cognitive factors discussed today inarguably matter, cultural factors have more clout. It is to these factors we will turn next time.

Until then!



  1. Aydin et al (2010). Turning to god in the face of ostracism: Effects of social exclusion on religiousness
  2. Granqvist et al (2010). Religion as Attachment: Normative Processes and Individual Differences
  3. Gray & Wegner (2010). Blaming God for Our Pain: Human Suffering and the Divine Mind


  1. Barrett & Keil (1996). Conceptualizing a Nonnatural Entity: Anthropomorphism in God Concepts
  2. Gervais (2013). Perceiving Minds and Gods: How Mind Perception Enables, Constrains, and Is Triggered by Belief in God
  3. Maij et al (2017). Mentalizing skills do not differentiate believers from non-believers, but credibility enhancing displays do 
  4. Norenzayan et al (2012), Mentalizing deficits constrain belief in a personal God
  5. Schjoedt et al (2009). Highly religious participants recruit areas of social cognition in personal prayer 
  6. Taylor (1999). Imaginary companions and the children who create them
  7. Willard, Norenzayan (2013) Cognitive biases explain religious belief, paranormal belief, and belief in life’s purpose 
  8. Willard et al (2020). Cognitive Biases and Religious Belief: A Path Model Replication in the Czech Republic and Slovakia With a Focus on Anthropomorphism


  1. Bering (2006). The folk psychology of souls 
  2. Bering and Bjorklund (2004) The natural emergence of reasoning about the afterlife as a developmental regularity.
  3. Bloom (2007). Religion is natural. 
  4. Carruthers (2020). How Mindreading Might Mislead Cognitive Science
  5. Cohen et al (2011). Cross-Cultural Similarities and Differences in Person-Body Reasoning: Experimental Evidence From the United Kingdom and Brazilian Amazon 
  6. Forstmann & Burgmer (2015). Adults Are Intuitive Mind-Body Dualists
  7. Kuhlmeier et al (2004). Do 5-month-old infants see humans as material objects? 
  8. Lane et al (2010). Children’s Understanding of Ordinary and Extraordinary Minds. 
  9. Roazzi et al (2013). Mind, Soul and Spirit: Conceptions of Immaterial Identity in Different Cultures 
  10. Slingerland & Chudek (2011). The Prevalence of Mind–Body Dualism in Early China 

MCI Theory

  1. Norenzayan et al (2006). Memory and Mystery: The Cultural Selection of Minimally Counterintuitive Narratives
  2. Lindeman et al (2015). Ontological confusions but not mentalizing abilities predict religious belief, paranormal belief, and belief in supernatural purpose


  1. Kelemen et al (2002). Why things happen: teleological explanation in parent-child conversations. 
  2. Kelemen (2004). Are Children ‘‘Intuitive Theists’’? Reasoning About Purpose and Design in Nature 
  3. Kelemen et al (2009) The human function compunction: Teleological explanation in adults. 
  4. Kray et al (2010). From what might have been to what must have been: Counterfactual thinking creates meaning.
  5. McAdams & McLean (2013). Narrative Identity.


  1. Norenzayan & Gervais (2012). The origins of religious disbelief
  2. Epley et al (2009). Believers’ estimates of God’s beliefs are more egocentric than estimates of other people’s beliefs 
  3. Riekki et al (2013). Conceptions about the mind-body problem and their relations to afterlife beliefs, paranormal beliefs, religiosity, and ontological confusions

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s