Gods – 4 (An Anthropology of Religion Perspective — Why People Believe in Gods)

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.

by Neil Godfrey

Not all minimally counterintuitive concepts are good candidates for interesting story material. A potato that vanishes whenever someone looks is soon a forgotten idea; but a potato that can talk to you has potential for many creative plot lines.

In the previous post we saw how certain minimally counterintuitive concepts can have the potential to be strikingly memorable and shared in wider groups. By a “minimally counterintuitive concept” we mean a concept that violates only one or two attributes of that our basic mental tool kit would lead us to expect: e.g. a dog that talks (talks and thinks like a human) is more memorable and interesting as a focus of a story than a dog that experiences time backwards, is born or a rhino that mated with a bullfrog, sustains itself on graphite, speaks Latin and changes into cheese on Thursdays. The latter dog becomes a laundry list of oddities and even begins to lose any identity as a dog at all. A living animal that has the communication faculties of a human is distinctly notable, however.

But not all minimally counterintuitive concepts serve well as religious ideas. Justin Barrett’s chapter, “Gods” (in Religion, Anthropology, and Cognitive Science), proposes that the best religious belief systems focus around minimally counterintuitive concepts that are intentional agents. Recall our most relevant mental tool kit devices:

    • Naive Physics Device (physical objects cannot move through each other; need to be supported or will fall….)
    • Agency Detection Device (automatically tells us that self-propelled and goal-directed objects are intentional agents)
    • Theory of Mind Device (actions are guided by beliefs, by wishes to satisfy desires….)
    • Naive Biology Device (animals bear young like themselves; they are organic matter….)

Justin Barrett points out that researchers have demonstrated by means of experiments involving “moving dots on a computer screen and other artificial displays” that the “agency detection device” is “touchy” or “hyperactive”.

Anthropologist Stewart Guthrie has suggested that our tendency to find agency (especially human-like agency) around us has arisen for survival reasons. Historically, our best opportunities for survival and reproduction and our biggest threats were other agents. So, we had better be able to detect agents. Better to guess that the sound in the bushes is an agent (such as a person or tiger) than assume it isn’t and become lunch. If it turns out you were too cautious (about the wind blowing in the brush, for instance), not much is lost (Guthrie, 1993). This hypersensitive agency detection device (or HADD) produces non-reflective beliefs that agency is present and then the Theory of Mind tool generates non-reflective beliefs about beliefs, desires, perceptions, and so forth of the alleged agent (Barrett, 2004).

Imagine viewing simple movies of the following sort:

(1) Two small squares are sitting in a line, separated by several inches. The first square (A) moves in a straight line until it reaches the second square (B), at which point A stops moving and B starts moving along the same trajectory.

(2) Two small squares are sitting in a line, separated by several inches. The first square (A) begins moving in a straight line towards the second square (B). As soon as A gets close to B, B begins moving quickly away from A in a random direction, until it is again several inches from A, at which point it stops. A continues all the while to move straight towards B’s position, wherever that is at any given moment. This pattern repeats several times. 

Objectively, all that is happening in such movies is the kinematics described above. Perceptually, however, a striking thing happens: in the first movie, you see A cause the motion of B, and in the second movie, you see A and B as alive, and perhaps as having certain intentional states, such as A wanting to catch B, and B trying to escape . . . . 

I think we can all accept this point. We know how easy it is to imagine unusual sounds in a still house are evidence of an intruder, “or “how light patterns on a television screen are people or animals with beliefs and desires”. Barrett cites an article by Scholl and Tremoulet [link is to pdf] addressing the question. Example in the side box.

An agency detection device is quickly activated to explain a wisp of fog or blowing sheet on a clothesline as a person, even a ghost, for a moment.

An agency detective device, in particular a hyperactive one (as they tend to be), come to the fore to “explain” events around us:

An example of an event that may trigger agent detection might be the following. When walking through a reportedly haunted castle, a decorative sword falls and narrowly misses cutting off your arm, just after you scoffed at the idea of ghosts. (Foolish mortal!) Given that a physical object appeared to have moved in a way that was not readily explained by the non-reflective beliefs of Naïve Physics (because inanimate objects cannot move themselves), and the movement seemed to be goal directed (to answer your skeptical comment), your HADD might detect agency at play. HADD searches for a candidate agent. As the sword is not an agent in its own right, a non-reflective belief that an unseen agent must have moved the sword arises from HADD. Connecting this non-reflective belief in present agency to the reflective proposition that ghosts inhabit the castle makes acquisition of a reflective belief in a ghost perfectly natural given the circumstances. The sword falling is an event that HADD might detect as agency.

Barrett adds “physical traces” to the list of things an agency detection device seeks to explain. The word “traces”, though, in my estimation, begs the question because it assumes a causal agent within the very definition of the word. So we have arguments over whether crop circles are caused by space aliens or human pranksters. I would prefer “physical phenomena” to “traces” — many are easily explained but we do come across even natural phenomena that look strange and we are easily led to imagine (erroneously) intelligent agents as their cause.

Here we get into a very messy realm. Many Bible believers are convinced that they see “traces” or evidence of fulfilled prophecy, for example, in their holy writings. Others see “miraculous” gematria at work. And so forth. We know where this agency toolbox leads in these contexts.

Another context Barrett pinpoints as significant is what he calls “contextual” — meaning that if we walk alone through a tree-thick park as evening descends we are more likely to “detect” false-positive indicators of a mugger waiting to surprise us than we would in the fresh sunlight of a morning.

Another opportunity for the agency detection device to come to the fore is in a religious ceremony. (Though again, are we not here begging the question somewhat? But we do see an opportunity for reinforcement, certainly.)

The salience of explicit, reflective expectations for witnessing agency increases agency detection as well. Told that a particular religious event (e.g., a ritual or petitionary prayer) will cause a god to act in a particular way, I would be primed to find seek confirming evidence of the god’s activity.

McCauley and Lawson observe that many religious rituals in which a god or representative of a god performs an action (e.g., as when a priest marries a couple, or a minister baptizes a person, or an elder initiates a youth), the ritual and surrounding trappings tend to be more elaborate and more emotional. One reason offered for this correlation is that the heightened sensory pageantry and resulting emotional arousal impresses upon the participants and observers that the gods are really acting because of or through the ritual (McCauley & Lawson, 2002). This observation leads me to speculate that the additional sensory pageantry excites HADD (perhaps through general arousal) and, in some cases, provides more potential events and traces that might be attributed to the gods.

Yes, we can quickly override those mental temptations to impute agency to a sound from a tree branch moving in the wind, a sound of our house timber shrinking in the cool of the night, but we are still left with our propensity to quickly imagine an unknown and intentional agent being responsible for these sounds, etc.

Even though a single HADD experience might be overridden and not result in a reflective belief, these experiences will be remembered and cumulatively lend plausibility to reflective belief.

To illustrate this dynamic in another way, compare three people:

1. A person walks through a wood after having been told that forest spirits dwell in it. Something ambiguous happens prompting this person to have HADD experiences. The experiences will be likely to reinforce the belief in forest spirits.

2. A second person walks through a wood and has the same HADD experiences as the first person but does not successfully identify an agent that accounts for them, and so does not form a reflective belief in an agent. But after the fact, the person hears that forest spirits dwell in that stretch of woods. Memories for the HADD experiences might be triggered and evaluated as evidence for reflective belief.

3. A third person walks through the wood, having heard that it is populated by forest spirits but has no HADD experiences at all.

Clearly, the first two people are considerably more likely than the third to form a reflective belief in forest spirits. HADD would play a pivotal role in belief formation or encouragement.

It might be that HADD rarely generates specific beliefs in ghosts, spirits, and gods by itself, and hence does not serve as the origin of these concepts. Nevertheless, HADD is likely to play a critical role in spreading such beliefs and perpetuating them.

Next — how the agency detection tool connects events in our lives, and how we live our lives, with unseen agents.

Barrett, Justin L. 2007. “Gods.” In Religion, Anthropology, and Cognitive Science, edited by Harvey Whitehouse and James Laidlaw, 179–207. Durham, N.C: Carolina Academic Press.

Scholl, Brian J, and Patrice D Tremoulet. 2000. “Perceptual Causality and Animacy.” Trends in Cognitive Sciences 4 (8): 299–309. https://doi.org/10.1016/S1364-6613(00)01506-0.



Gods – 3 (An Anthropology of Religion Perspective — Why People Believe in Gods)

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.

by Neil Godfrey

In brief, people believe in gods because gods gain tremendous support from the natural and ordinary operation of mental tools. Note that because mental tools and their processing biases arise primarily as a consequence of biological endowment plus essentially universal features of human environments, the factors that prompt belief in gods in Melanesia are the same as those that prompt belief in Scandinavia. Below I sketch several ways in which god concepts receive this support. The first way concerns how god concepts are minimally counterintuitive.

a minimally counterintuitive concept pixabay

As per the previous post, by mental tools we mean the way our brains come equipped with agency detection, with a theory of mind, with a basic set of inferences about the physical and biological environment.

Also as per the previous post, the more inferences we can bring into play from our various mental tools and apply to any proposition or idea, the more likely we are to reflexively or intuitively believe that idea. Our mental tools prepare us to expect our environment to behave in certain ways; that is, we intuitively expect objects to fall when they are unsupported, that foods are made of organic matter, that self-propelled and goal-directed agents act with intent.

Concepts that violate in major ways these properties that our mental tools have come to expect are not plausible and not entertained in our thinking for very long. An example Barrett gives is of a dog that

    • experiences time backwards
    • is born or a rhino that mated with a bullfrog
    • sustains itself on graphite
    • speaks Latin
    • changes into cheese on Thursdays

Such a dog violates so many non-refective intuitions or inferences we expect a dog to have that we no longer have a “portable concept of a dog” at all, but rather “a laundry list of features that do not seem to hang together.” That sort of concept won’t be shared or last easily. We can say that highly counterintuitive concepts don’t last.

However, a dog violates just one or two expectations that our mental tools prepare us for is still easily understood, remembered, and shared: e.g. a dog that is as large as a small horse. Or, minimally counterintuitive concepts can last.

Minimally Counterintuitive Concepts

Take a statue that hears prayers. The statue meets all the expectations of our “naive physics” tool; but it has one extra attribute — a mind, or at least what our theory of mind tool leads us to expect of minds. That is a concept that is easily understood. It conforms to what we intuitively expect of both objects and minds.

A statue that hears prayers may only involve a simple transfer of mental properties to an artifact. Except for this one transfer of property, the artifact meets ordinary intuitive expectations for artifacts (that is, non-reflective beliefs), and the mind of the statue meets ordinary non-reflective beliefs about minds. . . . Compared to how massively counterintuitive concepts could be, successful religious concepts tend to be rather intuitive. They conform to non-reflective beliefs governing the sorts of things that they are—their intuitive ontology. Hence, general plausibility is maintained. But being only slightly or minimally counterintuitive provides god concepts with another asset: facilitated transmission. (186)

We avoid a laundry list of oddities. The statue that hears prayers does not, for example, violate our theory of mind tool. The statue does not hear prayers yet completely misunderstands them; it does not hear prayers that are uttered many miles away; and so on.

And the concept is easily communicated. If only one person has a belief then that belief is a mere oddity. It is when groups of people share the same beliefs in supernatural agents that we have religious beliefs.

So what makes a good religious belief? The cognitive approach to religious beliefs proposes that the best candidates are those that violate only one or two intuitive expectations.

Those that violate a small number of intuitive assumptions can actually make for very strong candidates. Concepts that meet most non-reflective beliefs, but violate just a small number (e.g., one or two at a time) have been called minimally counterintuitive (MCI) concepts (Barrett, 2004; Barrett & Nyhof, 2001). A dog that speaks Spanish would be a minimally counterintuitive concept. A dog that gives birth to kittens would be MCI. A dog that can never die would be MCI. Such concepts enjoy good conceptual integrity and as such are easily remembered, recalled, and shared. Further, the counterintuitive feature may help the concepts to stand out against a backdrop of more mundane concepts, hence improving their salience and the attention devoted to remembering them. Experiments show that MCI concepts are transmitted more faithfully than ordinary or simply unusual ones (Barrett & Nyhof, 2001 ; Boyer & Ramble, 2001*). (187 f. * link is to pdf)

But not all minimally counterintuitive concepts are likely religious beliefs. What makes some better than others?

Take, for example, a potato that vanishes whenever you look at it versus a potato that talks. Both potatoes are counterintuitive but the vanishing potato scores poorly in terms of inferential potential. That is, some concepts more readily generate inferences, explanations, and predictions than others do. Some concepts excite a greater range of mental tools and some mental tools more completely. Consequently, they touch on more human concerns and, due to the convergence of many non-reflective beliefs, carry more reflective credibility. Even if I had some evidence of a potato that vanishes whenever someone looks at it, not much follows from its discovery. A potato that talks? Now that sets the imagination running a bit, especially at suppertime. (188)

Religious belief systems are mostly populated with intentional agents, minimally counterintuitive intentional agents — to be taken up in the next post.

Barrett, Justin L. 2007. “Gods.” In Religion, Anthropology, and Cognitive Science, edited by Harvey Whitehouse and James Laidlaw, 179–207. Durham, N.C: Carolina Academic Press.


Gods – 2 (An Anthropology of Religion Perspective)

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.

by Neil Godfrey

Explorations into why we believe and think the way we do should be shared as widely as possible and not restricted to scholarly publications. Hence these posts. (They cover ideas that we have presented before in different ways as they derive from different researchers, but slightly different perspectives on the same fundamental concepts can deepen our understanding of the matter.)

In the previous post we began with the point that we have two types of beliefs: reflective and non-reflective. Here we identify where these different types of beliefs come from. We will see in future posts how this model explains why belief in gods and spirits is in effect universal.

Where Non-Reflective Beliefs Come From

We are not taught everything we know. We are born with a brain that comes pre-packaged with a set of tools that enable us to make reliable inferences about how our world works.

These mental tools automatically and non-reflectively construct perhaps most of our beliefs about the natural and social world. Non-reflective beliefs arise directly from the operation of these mental tools on inputs from environment. The vast majority of these beliefs are never consciously evaluated or systematically verified. They just seem intuitive, and that is usually good enough. (Barrett 182)


We focus on four of these mental tools.

Our Naive Physics Tool

Even as infants we “know” that physical objects:

    • tend to move on inertial paths
    • cannot pass through other solid objects
    • must move through the intermediate space to get from one point to another
    • must be supported or they will fall
Our Agency Detection Tool
    • automatically tells us that self-propelled and goal directed objects are intentional agents
Our Theory of Mind Tool

Theory of mind gives us non-reflective beliefs concerning the internal states of intentional agents and their behaviors: Continue reading “Gods – 2 (An Anthropology of Religion Perspective)”


Gods (An Anthropology of Religion Perspective)

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.

by Neil Godfrey

Justin L. Barrett earned degrees in psychology from Calvin College (B.A.) and Cornell University (Ph.D). He served on the psychology faculties of Calvin College and the University of Michigan (Ann Arbor), and as a research fellow of the Institute for Social Research. Dr. Barrett is an editor of the Journal of Cognition & Culture and is author of numerous articles and chapters concerning cognitive science of religion. His book Why Would Anyone Believe in God? (2004) presents a scientific account for the prevalence of religious beliefs. He is currently Senior Researcher at the University of Oxford’s Centre for Anthropology and Mind. — from “Contributing Authors”, p. xxiii, of Religion, Anthropology, and Cognitive Science.

If I want to ensure a good harvest, I might take care in preparing my field, fertilize, use the best seeds possible, weed, and irrigate. I might also pray or conduct a ritual or in some other way try to get some supernatural help. If I wish to join a community or society, I might register or pay dues or even undergo an initiation ceremony. But I might submit myself to an initiation that appeals to ancestors, spirits, or gods. (Barrett, 179)

Thus begins Justin Barrett’s contribution to Religion, Anthropology, and Cognitive Science. The title of his chapter is Gods. His contribution is an exploration of why it is that people around the world, and for ages past, have made appeals to superhuman or supernatural agencies. It’s not as if the idea of “god” or “spirits” are unique in their ubiquity. Other beliefs are also found in common throughout the human experience: people universally believe in other minds; they also believe in the constancy of physical laws. It’s not only gods and spirits that are some sort of universal.

Barrett begins his discussion by how it is that people come to believe anything at all. And this brings us to the work of psychologists and their experiments on people at different stages of development. One thing has become clear: our minds don’t simply register “the world as it is” through our senses and accordingly “map reality” into our heads like a sponge responding to finger pressures to register this or that “reality point”. No, our minds are a storehouse of modular processing machines. Nothing enters that is not pre-processed in some way:

[The mind’s] normal functioning may better be likened to a workshop equipped with lots of specialized tools for processing particular classes of information. These mental tools arise with built in biases that influence which bits of information will be attended to and how that information will be represented (which might include its being distorted). (Barrett, 180)

There are two types of belief, Barrett explains:

1. Reflective Beliefs

If someone asks you if you believe in something, your answer will draw from a reflective belief. You will know you are not alone in those beliefs. Examples of reflective beliefs:

  • Toyotas are more reliable than Yugos
  • E=mc2
  • pumpkins are orange
  • Michael Johnson holds the world record in the 200 meter dash
  • Harvey Whitehouse is six-feet, five-inches tall

2. Non-Reflective Beliefs

Non-reflective beliefs, in contrast, operate in the background without our conscious awareness. These beliefs may not be consciously accessible and do not arise through deliberation. Rather, our minds produce non-reflective beliefs automatically all the time.


  • People act so as to satisfy their desires.
  • Rainbows exhibit six bands of color.
  • Raccoons and opossums are very similar animals.
  • People from outside my group are more similar to each other than people inside my group.
  • Animals have parents of the same species as themselves.
  • My pants are blue.

Non-reflective beliefs do not depend on verbal reasoning and statements. We can even identify more nonreflective beliefs by studying babies. Babies, we can tell from their eye-gaze, believe non-reflexively that

  • solid objects cannot pass directly through other solid objects
  • unsupported objects fall
  • inanimate objects must be contacted before they may be set in motion whereas people need not be . . . .

So where do these nonreflective beliefs come from? We are not taught them. How do they arise? . . .

Continued in next post in this series . . . . . 

Barrett, Justin L. 2007. “Gods.” In Religion, Anthropology, and Cognitive Science, edited by Harvey Whitehouse and James Laidlaw, 179–207. Durham, N.C: Carolina Academic Press.


Oh my god! People still believe these ancient myths

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.

by Neil Godfrey

By Michael F. Bird:

The World was Made For Us to Reign Over It

In terms of the biblical narrative, I think it helps if we remember that we are consistently given a picture of God’s intention to make humanity his vice-regents who will reign with him and for him over the world. This was the role of Adam in Eden, Israel in Canaan, and the church in Christ’s kingdom . . . .

What is no less interesting is how two very different texts, both post-70 AD, 4 Ezra and the Shepherd of Hermas, treat creation as made explicitly for them to rule over it.

. . . .
In which case, I think it safe to say, that “salvation,” in its eschatological coordinates, has to include the notion of creaturely participation in God’s rule over all things.
Can you imagine the ants, or any other sentient creature, saying that they have concluded that god gave them the responsibility to rule the world as his agents and then to act as if all else exists to be subordinated to them?
Dear Bible Believers, it is time to put away ancient mythology and to understand that we are part of a web of life on a fragile planet. Carrying on as god-ordained supremacists, as if our species is the ultimate “goal” of this universe, is creating the sixth great extinction right now and threatening the very survival of organized human life. Time to face the fact that grass has so far proved itself far more successful in evolutionary terms than intelligent life forms.


Religious Credence is Not Factual Belief: 2

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.

by Neil Godfrey

Continuing from Religious Credence is Not Factual Belief: 1

From http://www.armageddonbooks.com/map.html
From http://www.armageddonbooks.com/map.html

In the previous post we saw the three core features of factual beliefs as understood by Neil Van Leeuwen; we also saw that religious “beliefs” or credence do not share any of these characteristics of factual beliefs.

Van Leeuwen distinguishes factual belief from what he terms secondary cognitive attitudes — that is, factual belief has different characteristics from fictional imagining, hypothesis, assuming for sake of an argument, and so forth. Among these secondary cognitive attitudes he places religious credence. But religious credence is different again from these other secondary attitudes because of three characteristics. In this post we identify those three distinguishing properties.

But before continuing, however, I’ll jump to Van Leeuwen’s conclusion where he explains how religious credence and factual belief relate to one another in the mind of the religious person:

An agent’s religious credences comprise a map she uses for short- and long-term orientation in life. The map is colored with features that are taken to have normative force in virtue of their being part of the map at all; the colors represent sacred, sinful, eternal, righteous, holy, and the like. The map, in more or less detail, is determined by the dictates of individuals who are taken as special authorities in the community of the agent. But the agent herself also freely elaborates on the credence map in ways she finds useful for normative orientation. The map colors the authorities as holy. Other individuals in the community are painted as faithful. This map doesn’t just represent normative properties, however; it also represents objects, people, places, events, and supernatural beings that make the normative properties memorable or salient.

But the agent is always using another map: factual belief. This map comes to be in a different way from the religious credence map. It is generated chiefly by perception and rational expansion thereon. It helps us avoid falling in ditches and eating poisonous berries. The religious credence map lies on top of the factual map like a colored transparency, so that the objects, events, people, and places in the factual map can also appear religiously colored. Thus, only by careful scrutiny do we see the two maps are distinct. (My bolding)

So that’s how Van Leeuwen sees the two types of cognitive attitudes co-existing. Now to those points that make religious credence a stand-alone. Continue readingReligious Credence is Not Factual Belief: 2″