Does the Atheist have a Theory of Mind?

theory of mindby Thomas Coleman III


Planet earth appears to be filled with unseen forces that control the behavior of its inhabitants. No, this isn’t the beginning to a cheesy B-movie science fiction film script. This is reality and even the staunchest of skeptics act as if they believe in these invisible forces. That is, we live in a material world ruled by minds with no physical locality and it is here that we think beliefs, desires, intentions, and other mental states are both responsible for, and explain our behavior [1]. There is nothing particularly magical or surprising about this fact, at least not until we consider particular theories in the cognitive science of religion (CSR) that, for example, suggest atheists may be “socially disabled” [2], have a “malfunction” in their ability to reason about these mental states, or perhaps that there is no such thing as atheism at the level of cognition [3]. Thus, and I ask jokingly, does the atheist have a theory of mind? But, more on this in a moment.

Attributing mental states is something we do to others and ourselves on a daily basis, such that it appears to be commonsense — and it is! In fact, this ability has even been called “commonsense psychology,” among various others terms: social cognition, folk psychology, mind reading, and mentalizing [4]. Our folk psychology may not deliver adequate scientific causal descriptions, but we are nonetheless bound to it in everyday reasoning. While some of these terms have very specific technical meanings within a given discipline or theory, for the purposes of the present essay I will primarily use the term theory of mind (ToM). This essay will present a brief overview of ToM, its relationship to autism spectrum disorders, how this relationship is utilized in CSR, and critically evaluate the suggested links between poor ToM skills and atheism.

Theory of Mind and the Autism Spectrum

Some thirty years ago, Premack and Woodruff [5] shoved the topic of ToM under the empirical eye of the sciences in their paper titled “Does the chimpanzee have a theory of mind?.” Granted that human mental states are unobservable, they were interested in the possibility that chimpanzees might make goal oriented mental state inferences in a similar fashion (e.g., Does Sally the chimp know that the kids at the zoo want to see her and that they become amused when she swings from a branch?). Interestingly, these types of questions focused on animal cognition actually prompted much more research into how humans come to attribute beliefs and desires. That is, all-day everyday, we use our ToM to navigate our world by making sense of past actions and generating predictions for the future (e.g., I became overweight in the past 5 years because I enjoyed junk food, but I am on a diet now so keep those Butterfinger bars away from me). In essence, the use of our ToM is so pervasive in our social life that the sciences had largely taken for granted how humans navigate an intentionless world up until that point.

Later, research progressed on ToM and in humans specifically. In the 1980’s, Baron-Cohen and colleagues [6] asked, “Does the autistic child have a theory of mind?” and provided evidence suggesting that the central explanation for the crippling effects of autism disorders may lay in the inability to properly attribute and understand others mental states. In short, if an individual is unable to understand that a particular piece of information is intended for them (e.g., Ralph this is how you tie your shoe laces) then they may be unable to cognitively register directions as anything other than a random action in one’s environment (e.g., Ralph only hears speech and sees someone moving). With a severe impediment to learning in place, these individuals typically have well below average IQ’s and despite this rather negative term, are considered “low functioning.” Important to note, however, is that while all theories of autism generally acknowledge some degree of trouble in reading the intentions and emotional states of others, they vary on the exact reason for this.

Nonetheless, not every person with autism is incapable of functioning within the rules of folk psychology and many possess a ToM that not only allows him or her to perform simple tasks, but to navigate the wider social world on their own. However, this is not to say that engaging in social situations comes as easily as it would for someone not on the autism spectrum. Reading intentions and picking up on subtle social cues can still be difficult and even exhausting (e.g., If they happen to be at a social gathering, they may not easily read the host hinting that it’s time to leave). These individuals are commonly referred to as “high functioning” and often have accelerated abilities in domains that do not involve heavy reasoning about mental states. Instead, they may maneuver through the social arena by systemizing, analyzing, and constructing rule-based schemata that serve as guides to understanding the world. They may have special interests, for example, in mathematics and the sciences and are perhaps far better at these subjects than most. Furthermore, they usually have average or even above average IQ’s [7].

As research on autism progressed, it became clear that it resisted any strict clinical criterion for diagnoses and its cluster of associated traits became known as the “autism spectrum disorder” (ASD). Interestingly, these traits, such as variations in ToM proficiency and systemizing tendencies, were also found present in the wider public, within populations that did not meet the clinical criterion for diagnoses of ASD. Researchers now use the term broader autism phenotype (BAP) to understand individual differences in ToM and systematizing present in neurotypicals. Now what does all of this have to do with belief and nonbelief in gods?

Cognitive Sciences of Religion

The prevailing view from the cognitive science of religion conceives of “religion” as centered on belief in and behavior stemming from a minimally counter intuitive agent (definition to come). Religious beliefs both stem from and are processed by our average, run of the mill, “garden variety” cognitive capabilities [8]. In this view, rather than religion having supernatural or sui generis explanations, quite natural ones exist and religion is considered a by-product of the mind. While some argue for the possible adaptive benefits of religion, its conceptualization as a by-product is a prerequisite for this view and will not be discussed further here.

Counter Intuitive Agent Representations

Religion, as a by-product, simply plays on our evolved intuitions about physical objects, biological expectations, and psychological capabilities [9]. From an early age, for example, young children possess enough knowledge in these domains to infer that two objects cannot occupy the same space and that unsupported objects will fall: physics domain. They know that crocodiles do not give birth to baby chickens: biological domain. They know that where there is movement, there is an intentional cause responsible: psychological domain.

These domains constitute core knowledge that is shared across our species and religious representations, such as a god or ghost, are considered “counter-intuitive” for violating the properties of these recurring categories. Representations can either “breach” one of these domains (e.g., “a man who walks through walls” is a breach of our intuitive physics by expectations informing things a human can do), or they “transfer” the properties of one domain to that of another (e.g., “a statue that listens to your prayers” transfers psychological information onto our expectations of physical objects) [10]. Representations of this nature are considered “minimally counter-intuitive” (MCI), in that there are just enough violations (one or two at most) to be flagged as unusual and henceforth catchy, yet not so many as to be mentally taxing and non-transmittable (for example, think of how much trouble a game of “Chinese whispers” would be using a long sentence with 30 people). In virtue of how the human mind develops, CI representations stem from our ordinary cognitive faculties and are a universal occurrence. However, many CI representations have nothing to do with a god or ghosts (e.g., the cartoon series Veggie Tales that has talking tomatoes and asparagus, or even Super Man).

Theory of Mind System in CSR

ToM is perhaps the most critical ability to underlie religious beliefs, but it is also a system that facilitates our understanding of the world in general (no ToM, no social interaction whatsoever). ToM is the evolved propensity for systematically treating agents in the environment as capable of self-generated autonomous actions that are caused by mental states in humans [11]. Such mental states (beliefs, desires, wants, emotions), of course, are not observable — we do not have direct access to the contents of another’s mind. However, upon observing a given behavior, the mind automatically begins inferring the mental state (or states) that the individual thinks best explains said behavior (e.g., upon seeing Sally attend church each Sunday, I might infer that she wants to be there because she believes in God).

However, ToM is not necessarily a bearer of epistemic truth in the sense that it relays accurate information about the way the world really works or what another’s mental state really is. Our beliefs are always at risk of being wrong and often are (e.g., I asked Sally why she attends church each Sunday and she said that even as an atheist, she just enjoys the music and company). Importantly, ToM can handle increasingly complex metarepresentations — that is, beliefs about beliefs about beliefs…(e.g., I know that Sally believes that her Pastor thinks she is a believer in God).

When ToM is applied to thinking about CI agents, it does this as effortlessly as it would thinking about Sally’s particular situation or any other, even though a CI agent isn’t a person per se [8]. For example, although a CI agent, such as the angel Gabriel, can pass through walls, it is still characterized as having more or less the same belief-desire reasoning that humans have. The angel can speak, it may have a message to relay, things it wants you to know and so on and so forth. Now, here is where the ToM system links up with god beliefs, as Boyer and Bergstrom [9] suggest:

“…people’s concepts of gods or ancestors may recruit or exploit any of the diverse psychological systems that govern social relations, simply by virtue of the gods’ representations as social agents.”

Although gods and other CI agents (e.g., Santa Claus, Teenage Mutant Ninja Turtles) can be represented in our ToM system, this system is not a “religious specific system” [8], let alone one specifically meant for handling the mental states of a well-fed charitable man in a red suit, anthropomorphic crime fighting turtles, or the editor of this magazine, for that matter.

Beliefs, desires, and wants are how humans in general make sense of the world around us. They are regularly overextended as a heuristic to objects that clearly do not have minds [12] or applied to reason about situations outside its evolutionary selected foci (e.g., our minds did not evolve to perform calculus, although some of us can do so). Perceiving the world in terms of a “mind like ours” makes it easier for a “mind like ours” to operate in an environment ruled by opaque physical causes void of the very intentionality we are presumed to possess [3, 13]. Once a cultural representation of any social agent appears it will be handled by our ToM system with no regard for the truthiness of a particular representation.

Atheism, Autism, and the Deficiency Hypothesis in CSR

Over the course of history, atheists have often gotten a bad rap. Locke famously asserted that common contracts, such as oaths or promises, had no binding effect on atheists and that they were not to be trusted, as they believed in no god with which to enforce these obligations. For a recent empirical example, in one study by Gervais [1] his participants intuited that a man having sexual relations with the carcass of a deceased chicken and subsequently preparing and eating it for dinner (yes the man in the Gervais’s story cleaned the chicken first and even used a condom…) was behavior indicative only of atheists, not any other ethnic minority. But setting the joys of Shake’ N Bake chicken aside for a moment, psychological explanations for atheism began almost a century ago with “the defective father hypothesis” (now discredited). In this view, atheism was the result of a cold and harsh father, or having no father at all. However, attempting to explain atheism as a deficiency continues in some CSR literature today.

As the previous sections detailed, since religious belief is so “easy,” in that it is parasitic to our normal cognitive faculties, then these deficiency arguments typically go something like this: someone who does not do something as easy as religious belief, must have something wrong with them. To this point, Barrett [2] even suggests that not believing in any gods may be comparable to being handicapped — unable to walk he says! Further, if ToM is used to represent the desires, wishes, and beliefs of unseen supernatural agents, then those humans disinterested in communicating with the gods must have a “malfunction” in their ToM system. Or so the story goes.

Now, let us say that an atheist happens to have a normal ToM. Surely they can’t be considered deficient. However, they are often suggested to be “implicitly religious” [2,3]. In Barrett’s [2] argument, he suggests that even if an atheist were to appeal to explaining the occurrence of any particular event, as happening just by “chance,” that this is a “pseudoagent,” which is merely a placeholder for what in another day and age would have been a god explanation (e.g., running into an old friend from high school 30 years later in unusual circumstances and saying “hey, what are the chances?” versus “God brought us together”). On Barrett’s account, it would seem that in lieu of giving a protracted casual-mechanistic explanation for any event (I won’t even attempt to develop one for the chance meeting of an old friend example, as it would be potentially massive) that there is literally nothing one could say except “God.” This argument can be rejected flat out. Just because humans use shortcuts in their reasoning during casual conversations (i.e., “it was chance”) in place of a dry scientific explanation is no reason to believe that they are implicitly appealing to a quasi god.

Bering’s account, however similar to Barrett’s, is situated on a slightly different conceptualization of ToM. Here, Bering [3] suggests that humans have evolved what he terms an “existential theory of mind” (EToM). This EToM causes us to see purpose and intentionality where in fact, there is none. Under this view, in order to deal with the overwhelming burden a meaningless existence might be to humans, we simply project meaning onto the world. For example, say a teenage atheist boy gets into a car wreck and then thinks to himself, if only for a second, “perhaps this happened so I could learn to slow down.” From a scientific point of view there exists no guiding life force (i.e., God) and hence the reason behind this particular car crash is simply a useful illusion. Bering suggests that God will never go away, because the purposeful intentional force that God is so often conceived to be is simply a part of our EToM system. While I am strongly sympathetic to this view, the EToM isn’t just behind religious reasoning, but also philosophical reasoning, as Bering minimally notes [13]. Meaning and purpose can be created and not always God given, as a recent study by Banerjee and Bloom [14], which found these tendencies rooted in more secular propensities, has suggested. However, Bering has chosen not to further develop the EToM in the spirit of philosophical reasoning, only religious reasoning.

While Bering’s and Barrett’s accounts of ToM differ, they are united by the fact that both scientists posit that ToM can be impaired by deficits associated with autism spectrum disorders. This is, at least on its surface, uncontroversial. Few would argue that a person with low functioning autism is going to be communicating with the gods if they are unable to adequately communicate with their caregivers or inferring that their car accident occurred to teach them a life lesson (at any rate, it’s unlikely they would be driving). However, individuals considered high functioning on the autism spectrum are quite another story. Indeed, studies show that significantly lower levels of belief are associated with clinically diagnosed high functioning ASD  [15,16].

Considering that the traits (typically only measured by self-report) that characterize ASD (systemizing and variations in ToM ability) can be found in the wider population, some explanations in CSR have suggested that atheists are unable, or at least find it very difficult to, believe in gods. However, this is allegedly not because they have found God unnecessary, are disinterested in religion, or found God to be implausible, but rather in virtue of their decreased ability to reason about the mental states of others, they cannot reason about God. This, combined with an increase in systemizing (think of a mechanistic view of the world) tips the scale in favor of atheism, or at least very low levels of religiosity.

While this view is certainly tempting, I currently know of no published study that has directly compared atheists with theists. Typically, these conclusions are supported by studies that, while informative, only compare high vs. low religiosity. We wouldn’t use a study looking at frequent vs. infrequent meat eaters to generalize to Vegans, thus studies comparing high vs. low belief doesn’t necessarily speak to individuals with no belief. Perhaps rather ironically, it may prove to be that weak believers in God are the ones with lower ToM abilities and not atheists!

Most unfortunately, some scholars often make the mistake of conflating low belief with no belief, and this confusion bleeds over into citations in published articles and book chapters further reinforcing false assumptions about the psychology of atheists. Similar problems have been noted by psychologist Luke Galen [17] in the domain of research on health, wellbeing, and religion.

Nonetheless, even if studies were conducted to properly test what I will term as the “deficient atheist hypothesis,” methodological rigor would be critically important. For, as any undergraduate psychology student should be able to tell you, correlation does not equal causation. For example, just because the murder rates rise in the summer along with ice cream sales, this doesn’t mean that Ben and Jerry’s causes murder. Furthermore, preliminary results from an ongoing study with my colleagues and mentors, Catherine Caldwell-Harris, Ralph W. Hood Jr. and Robert B. Arrowood have been unable to find differences between atheists and theists in an advanced measure of ToM or in attributing intentionality in social situations.

Perhaps most interestingly, where it is found that the less religious score slightly higher on the broader autism phenotype than stronger theists, the variance explained is almost nil [18,19]. In other words, just as the difference between a football player who weighs 260 pounds and another who weighs 260.1 pounds is very clear — the two weights are absolutely different — it seems comical to say that this is a difference that makes a difference. Where does this leave us in answering the question of whether or not the atheist has a ToM? Or, perhaps more appropriate at this point, a diminished ToM?


The same processes that underlie religious belief underlie atheism [20]. Religious belief can’t claim primacy to our cognitive faculties anymore than religious nonbelief can. Moreover, as the human species carries on into the future, some, but not all, will find little use in the label of “atheist.” Just as the chemists or astronomers of today do not identify as “a-alchemists” or “a-astrologers,” individuals will find little utility in identifying as “a-theists.” Instead, and perhaps only when required to do so, today’s atheists will become tomorrow’s nonbelievers and be identified only secondarily as at some distance from theism [21].

While ToM is certainly necessary to believe in gods, the average human being — religious or otherwise — isn’t using it to communicate with unseen supernatural forces. Instead, our ToM is busy figuring out much more mundane aspects of our earthly existence, such as “why did he or she do this or that?” In the end, ToM underlies the totality of our mental life and applying it to reason about supernatural agency only constitutes a small portion of this. Does the atheist have a theory mind? Of course, just as much or as less as any theist.


Thomas Coleman III is a graduate student in Psychology at the University of Tennessee—Chattanooga. He is interested in exploring multiple methodologies in the domains of qualitative and quantitative research, focusing in the psychology and cognitive science of religion and atheism/nonbelief. Recently, Coleman co-edited a special issue of the journal Science, Religion & Culture titled “Atheism, Secularity, and Science” with John R. Shook and Ralph W. Hood Jr.”

[1] Gervais, W. (2013). Perceiving Minds and Gods: How Mind Perception Enables, Constrains, and Is Triggered by Belief in Gods. Perspectives On Psychological Science, 8(4), 380-394. doi:10.1177/1745691613489836.

[2] Barrett, J. (2012). Born believers. New York: Free Press.

[3] Bering, J. (2002). The existential theory of mind. Review Of General Psychology, 6(1), 3-24. doi:10.1037/1089-2680.6.1.3.

[4] Goldman, A. I. (2012). Theory of mind. In E. Margolis, R. Samuels, & S. Stich (Ed.), The Oxford handbook of philosophy of cognitive science, (pp. 402-424). Oxford: Oxford University Press.

[5] Premack, D., & Woodruff, G. (1978). Does the chimpanzee have a theory of mind?. Behavioral And Brain Sciences, 1(04), 515. doi:10.1017/s0140525x00076512.

[6] Baron-Cohen, S., Leslie, A., & Frith, U. (1985). Does the autistic child have a “theory of mind” ?. Cognition, 21(1), 37-46. doi:10.1016/0010-0277(85)90022-8.

[7] Baron-Cohen, S. (2009). Autism: The Empathizing-Systemizing (E-S) Theory. Annals Of The New York Academy Of Sciences, 1156(1), 68-80. doi:10.1111/j.1749-6632.2009.04467.x.

[8] McCauley, R. (2011). Why religion is natural and science is not. New York: Oxford University Press.

[9] Boyer, P., & Bergstrom, B. (2008). Evolutionary Perspectives on Religion. Annual Review of Anthropology, 37(1), 111-130. doi:10.1146/annurev.anthro.37.081407.085201.

[10] Purzycki, B., & Willard, A. (2015). MCI theory: a critical discussion. Religion, Brain & Behavior, 1-42. doi:10.1080/2153599x.2015.1024915.

[11] Zimmer, C. (2003). Cognition: How the Mind Reads Other Minds. Science, 300(5622), 1079-1080. doi:10.1126/science.300.5622.1079.

[12] Heider, F., & Simmel, M. (1944). An Experimental Study of Apparent Behavior. The American Journal Of Psychology, 57(2), 243. doi:10.2307/1416950.

[13] Coleman, T. J. III, & Hood, R. W. Jr., (2015). Reconsidering everything: From folk categories to existential theory of mind. [Peer commentary on the paper “From Weird Experiences to Revelatory Events” by A. Taves]. Religion and Society: Advances in Research, 6(1), 18–22.

[14] Banerjee, K., & Bloom, P. (2014). Why did this happen to me? Religious believers’ and non-believers’ teleological reasoning about life events. Cognition, 133(1), 277-303. doi:10.1016/j.cognition.2014.06.017.

[15] Caldwell-Harris, C., Murphy, C. F., Velazquez, T., & McNamara, P. (2011). Religious belief systems of persons with high functioning autism. In Annual Meeting of the Cognitive Science Society, Boston, MA.

[16] Norenzayan, A., Gervais, W., & Trzesniewski, K. (2012). Mentalizing Deficits Constrain Belief in a Personal God. Plos ONE, 7(5), e36880. doi:10.1371/journal.pone.0036880.

[17] Galen, L. (2012). Does religious belief promote prosociality? A critical examination. Psychological Bulletin, 138(5), 876-906. doi:10.1037/a0028251.

[18] Lindeman, M., Svedholm-Häkkinen, A., & Lipsanen, J. (2015). Ontological confusions but not mentalizing abilities predict religious belief, paranormal belief, and belief in supernatural purpose. Cognition, 134, 63-76. doi:10.1016/j.cognition.2014.09.008.

[19] Van Eyghen, H. (2015) Workshop: ‘What is religious belief?’ report. Conference report for The Religious Studies Project.

[20] Geertz, A. (2013). Origins of religion, cognition and culture. Durham: Acumen.

[21] Silver, C., Coleman, T., Hood, R., & Holcombe, J. (2014). The six types of nonbelief: a qualitative and quantitative study of type and narrative. Mental Health, Religion & Culture, 17(10), 990-1001. doi:10.1080/13674676.2014.987743.

Gervais, W. (2014). Everything Is Permitted? People Intuitively Judge Immorality as Representative of Atheists. Plos ONE, 9(4), e92302. doi:10.1371/journal.pone.0092302


53 thoughts on “Does the Atheist have a Theory of Mind?

  1. We live in concrete canyons and travel on electronic highways, proud of our accomplishments, satisfied with our knowledge and obese with abundance. Theory of mind has become an abstract phrase, stripped of meaning. We are like mary in the monochrome room, knowing everything and not knowing what we cannot know.

    I toil up a high mountain ridge and hear a distant call, that of a black eagle. I look up into a high domed heaven of impossible blue, startling in its clarity. Eagles soar and swoop on their prey. Below on the rocks, frightened dassies scurry frantically for cover, diving from the boulders into the fynbos. On the mountain slope before me I hear the bleat of kudu and gemsbok, catching a glimpse as they bound out of sight on strong, nimble legs. I quickly step to one side to bypass the arching Cape cobra at my feet. A little further on I see the spoor of that nocturnal predator, the leopard, and a frisson of fear runs down my back.

    I experience the aching thrill of glorious, beautiful life with its fearsome intensity of threat and promise. I turn to my companions and they are hushed with awe. That night, in a high cave, crouched around a warm fire to keep out the mountain chill, their faces are lit with animation. The firelight dances in their eyes and I see another kind of life, an inner life reflecting the outer life we had marvelled at. But it was a deeper, richer life, that of joy, humour, regret, sadness, hope, companionship and love.

    Through the mouth of the cave I see the sparkling immensity of our galaxy in the night sky. I step out and clamber down the slope to quietly enjoy the spectacle. My mind expands as it travels through galaxy after galaxy. My absorption is punctuated by the discrete rustles of nocturnal life in the bushes around me. And then the sounds stop. My skin tingles and a nameless dread grips me. It is a heart stopping moment, beyond fear and dread, tinged with joyous awe.

    I have felt the numinous.

    Days later I return to the dull concrete canyons. I describe my experiences to my wife as clearly as words can allow. She looks at me sympathetically, with interest, and says, ‘that’s nice, dear’. Frustration grips me as I realise that she can never understand the vital, rich experiences of climbing in the mountains. She is mary and she lives in a monochrome room.

    Last night I was reminded of these experiences as I knelt before the altar at Mass. The beautiful sounds of Ave Maria swirled around me. The congregation thrilled to the joy of shared love in the presence of immensity. We celebrated the gift of life that has animated this universe with love.

    Today these good, kind, decent people are back in the concrete canyons. They are sustained and motivated by that encounter with immensity. When some angry, resentful skeptick questions us, we try to explain, but here, just as in the mountains, words have no power to convey the truth of our experiences. We are talking to mary in the monochrome room.

    Liked by 1 person

  2. Philip Thrift, that bm-science paper makes the same error with levels as most if not all who lose track of the usual levels of 1. physics founding 2. chemistry founding 3. cellular biology, founding 4. neural networks. They should expand their understanding of level 3, cells, and work up to 4. neural networks from there, and also more securely down to chemistry and physics. There are lots of network theories grounded by vaguely defined electromagnetic behaviour between neurons, but neurons are cells and cells have limits. The limits to neurons include hormonal and nervous limits, as neurons are equally comprised of sediment moving neurotransmitters (hormones) along tracks laid by ionic current shifting within axons to cross synapses.

    My advice is to expand knowledge of hormones and how they relate to ionic current to create experiences of awareness. I wouldn’t bother with computers and engineering until cellular chemistry is better explored.
    Nor would I bother with using computers as models of neural networks, because the experience is likely to be nothing more than a series of chemical analogies, or commonalities, without any symbols as such. Neurons are likely to be anatomical site extensions to a brain for automatic central coordination, and chemical analogy in all site functions would be integrated directly along neural networks to create sites’ experiences. This is more fact than theory, just extending known elements logically to enable progress. It is where to look and how to understand it, not “computers & engineering”.

    Liked by 3 people

Comments are closed.