Why “belief is hard work

This work is licensed under the Creative Commons | © Pascal Boyer. Attribution-NonCommercial-NoDerivs 3.0 Unported. ISSN 2049-1115 (Online). DOI: http://dx.doi.org/10.14318/hau3.3.014


Why “belief is hard work

Implications of Tanya Luhrmann’s When God talks back

Pascal BOYER, Washington University in St. Louis

Comment on LUHRMANN, Tanya. 2012. When God talks back: Understanding the AAneriean Evangelical relationship with God. New York: Alfred E. Knopf.

Tanya Luhrmann’s When God talks back (Luhrmann 2012), henceforth GTB, is a fitting companion volume to her first (and equally important) book Persuasions of the witch’s craft (Luhrmann 1989). The two books address a similar issue-briefly, how belief, far from being a simple matter of receiving and accepting information, requires complex cognitive processes, some of which can be illuminated by meticulous ethnographic investigation. The situations are certainly different. The London practitioners of “witchcraft” among whom Tanya Luhrmann did her first fieldwork engaged in practices widely perceived as ridiculous, indeed preposterous. Their stated beliefs were eclectic and generally couched in rather inchoate metaphors. By contrast, American evangelicals practice a respected version of mainstream Christianity. What makes them special is a clearly articulated belief that God can, precisely, talk back.

But the rub is, he does not. Or, to be more specific, the definite intuition that an agent is around, that this agent really is the god, that the god is talking, requires a lot of work, and is rather rare and frustratingly elusive. Even among the most accomplished of believers, a few islands of intuition are surrounded by oceans of doubt and disbelief (GTB: 133). This is splendidly illustrated in Lurhmann’s ethnography, and raises important questions to do with both our understanding of “beliefs” in anthropology and our cognitive models of belief states.

The challenge: Creating the god’s presence

As Luhrmann’s book demonstrates at length, it takes a considerable amount of work to reach some degree of intuitive belief that “God” is around, that “He” is listening and talking back. From the outside, Evangelicals are often perceived as people with certainties: they know there is a god, they know what he is like, they communicate with him. Inside the group, we find more or less the opposite. Christian beliefs are of course held with fervor, but the crucial element, the presence of and communication with a superhuman agent, are described as goals to achieve rather than a starting point. Many Evangelicals readily admit that they have not (or not yet) reached that point.

The cognitive work takes many different forms, which constitute most of the community’s religious practices. People must train their attention (GTB: 41ff). Everyone’s conscious mental life is replete with transient, floating, and inconsequential intrusive thoughts whose origin is obscure and unimportant. For Evangelicals, this is where superhuman communication may sometimes occur, if one can train oneself to accept and ponder elusive thoughts rather than discard them. People must also train their sensory imagination (GTB: 159ff.), and of course auditory imagination in particular. They must seek places and situations where perceptions do not crowd out self-generated imagery. People must train their emotional imagination or simulation (GTB: 11 Iff.). They assume that the presence of God is marked by specific emotions, but openness to such states, for them, requires careful monitoring and conscious appraisal of one’s emotional experience. Most important of course, people must learn to pray (GTB: 31). What most outsiders would consider the most straightforward activity, addressing an agent who you think is listening, is most difficult because the agent’s presence is, precisely, highly problematic.

Practice works—somewhat, sometimes. Many members of the group have experienced the “breakthrough” when inchoate thoughts or images seem to organize themselves into a coherent feeling of presence and a clear message from the imagined agent (GTB: 53). Personality variables clearly help in the process, as Luhr-mann’s data on interpersonal differences demonstrates (GTB: 15ff.), but the main factor remains dedicated practice—one is led to the intuition of a god’s presence through sustained practice.

But why on earth is it that difficult?

The world over, people do not (easily) believe in gods and spirits

A most common feature of ethnographic descriptions is that the author takes people’s beliefs (I will mainly consider beliefs of the “religious” variety) as a starting point in explaining behavior. The Fang believe that dead people can harm them if offended—that is why, in particular ceremonies, they try to placate these imagined agents and compensate them for perceived slights (Fernandez 1982). The Turkana believe that ancestors can protect them against misfortune, and consequently sacrifice an ox to nudge them toward benevolence (Lienard and Boyer 2006). So far, so good—such statements linking behaviors to beliefs explain the former in terms of the latter, an operation that is fundamental to our folk-psychology (Sperber 1985).

But this way of using belief as an explanans also conveys the implication, not always clearly intended, that the belief in question is actually there, that the people concerned, Fang or Turkana, do consider that spirits and ancestors are around and have the powers in question. But, I would argue, that is far from clear—indeed, understanding that it is often not the case explains a lot of otherwise mysterious cultural institutions.

This problem is compounded by the anthropological use of indirect reported speech, writing such things as “among the Fang, only ancestors can make one sick.” Of course, all readers know that the intended meaning is “only [what the Fang would describe as] ancestors can [according to some Fang people] make one sick.” But this usage perpetuates and strengthens the assumption that the belief is actually there.

To see how damaging the assumption of belief may be, consider a classic case, Claude Lévi-Strauss’ description of “symbolic efficacy” in shamanistic rituals surrounding problematic childbirth (Lévi-Strauss 1963). The central argument is that the shaman’s chant presents the patient with a series of analogical mappings among her body, the fetus, the birth-canal, et cetera, on the one hand and a set of mythical villages and landscapes on the other. Lévi-Strauss also suggests that these symbolic mappings actually help with the delivery process. The tacit assumption in this well-known exercise in ethnographic interpretation is that the participants be-lieve, in a straightforward manner, that the spirits actually are there, that the mythic villages in question do exist, that the shaman has access to them, and so on. Because rituals activate a set of preexisting beliefs of this kind, they may be “effective,” have effects on people’s mental states.

But Lévi-Strauss was quite wrong about that. Observing rituals in the flesh, so to speak, one is bound to derive the opposite impression, that beliefs are often an occasional and elusive consequence of ceremonies rather than their foundation. indeed, if beliefs were as straightforward as Lévi-Strauss (and many others) assume, rituals would be quite strikingly inefficient. As my colleague Denis Vidal once put it, if it takes a whole night of scripted ritualized behavior and 10,000 verses of opaque speech to cure a common cold, then calling all this “efficacité” seems a bit of a stretch. (Note that Lévi-Strauss wrote all this in French, an impoverished language in which a single term means “efficient,” “effective,” and “efficacious.”)

But the rituals are not as absurd as that—precisely because the beliefs are problematic. Very few people, it seems, hold as true a simple declarative statement of the form “the shaman can encounter spirits” or “the ancestors protect us.” Such statements are among the many conjectures they do entertain, as the best or at least a plausible interpretation of events and circumstances. They are generally held as “metarepresentational beliefs.” That is, one holds the belief that “[it is probably the case that] ancestors protect us” or “[it is certainly true that] ancestors protect us” (Sperber 1982). This of course explains why people can readily make such statements, and be committed to them, while having little or no intuitions about what the statements actually entail.

In the spirit of Lurhmann’s ethnography, one should generalize the observation. Many religious practices seem exceedingly odd if we see them as based on preexisting unproblematic beliefs. Once we realize that the belief is a conjecture, these activities make more cognitive (and existential) sense. initially, spirits may or may not be around. But after the whole night of ritual and the 10,000 verses, to some people at some junctures this conjectural representation becomes more vivid, more accessible, is associated with actual experience, is given some explanatory power—in other words is potentially turned into what we commonly call a belief. It is highly doubtful that shamanistic songs ever helped deliver infants—but making people think that might be the case is the real “symbolic efficacy.”

Tricks of belief—from reflection to intuition

At this point, it may be of help to emphasize the distinction between two kinds of mental states, that we can call “intuitions” and “reflections” (Baumard and Boyer 2013), or “intuitive” versus “reflective beliefs” (Sperber 1997), or “aliefs” vs. “beliefs” (Gendler 2008). Such a distinction is the basis of many “dual-process” models of cognition, based on a contrast between intuitive processes (often called System 1) that are quick, automatic, and implicit on the one hand, and reflective processes (called System 2 by contrast) that are slow, deliberate, and explicit, on the other (Evans 2008). Human minds comprise vast numbers of intuitive systems that produce specific representations of the environment. These are called “intuitive” because they “pop up” involuntarily—the processes that lead to them are not accessible to conscious inspection. Human minds also support deliberate chains of thought that explain, mention, or comment on intuitions. For instance, we all have physical intuitions, which for instance help us to predict the trajectory of the ball that bounced on the floor. We can also entertain reflective thoughts about the fact that a ball in motion contains momentum or force.

The distinction is relevant to the question at hand, because in most places, at most times, most people’s representations of superhuman agents (gods, spirits, etc.) are of the reflective type. People entertain deliberate thoughts to the effect that, for example, “So-and-so’s illness had to do with the spirits” or “God has a plan for me.” These are the metarepresentational “beliefs” that we anthropologists elicit or infer from people’s statements and behaviors (Baumard and Boyer 2013).

But such reflective thoughts can, sometimes, be associated with specific intuitions. This happens for instance when a magician announces that his “mystical force” will “annihilate” the object placed on the table (such statements create reflective, explicit representations in the minds of the audience), and, indeed, touching the object makes it disappear, or so it seems. The reflective thoughts about “mystical force” are now associated with intuitions (remembered perceptions), which of course makes them vastly more attention-grabbing, and potentially more plausible.

One should not look down on such cheap tricks. They are important, if not essential, in many religious traditions. Getting to see an image of a god in a piece of toast may not seem to us the most profound instance of religious experience, but that is because we are used to highly intellectualized, institutionalized forms of religious activity. In many places the world over, conjuring tricks and manufactured illusions are perfectly respectable adjuncts to more sober myth and ritual. And after all, the gospels are replete with conjuring tricks (e.g., turning water into wine) of a kind very familiar to first century Palestinians (Smith 1998) and indeed to amateurs of magic everywhere else. In some cases, the association of reflective thought with intuitions takes on a more elaborate form, consisting for instance in the recitation of thousands of lines of verse, or in demonstrations of trance and possession. Call these tricks or rituals or disciplines—the cognitive effect is that what used to be mostly reflective though is associated with definite intuitions.

This matters because religious representations are potentially much more compelling, attention-grabbing, and memorable when they are associated with intuitive content, preferably with perceptions, in this way. The process is necessary for the transmission of religious representations, for the constitution of religious traditions (Boyer 1998, 2001). Obviously, this does not imply that anyone is really trying to create traditions and transmit religious information. The point is simply that cultural transmission is a selective process. some mental representations “make it” through many cycles of individual transmission in (roughly) similar form, and others just do not. The former are what we call “cultural” representations, to denote their presence in many different minds in a group (Boyd and Richerson 1985; Sperber and Hirschfeld 2004).

Disowning one’s own thoughts

The Evangelicals in Tanya Luhrmann’s group have set themselves the Herculean task of associating reflective beliefs with intuitions without ever resorting to cheap tricks. Instead, the process requires gradual changes to their conscious appraisal of their own thoughts. Starting with material that most Christians would agree with, for example, that “God is everywhere, can hear all our thoughts and talk to us,” they endeavor to calibrate their own mental systems until this conceptual description fools, so to speak, their perceptual systems.

How can one achieve that? The techniques used are all “empirical,” fashioned though trial and error in the various Evangelical communities, and taught largely through individual testimonials. I cited above the various domains of training—imagination, sensory imagery, emotion. But how does this lead to the intuition of superhuman presence? Given a variety of specific thoughts and experiences, some more coherent or vivid than others, how is one to judge that a particular one is the real thing?

This is indeed the pivotal question in the Evangelical’s progress (GTB: 41). A crucial element here is the ownership of thoughts. To become (somewhat more) convinced that a thought of yours is a direct message from the god, you have to feel that it is not yours. More accurately, once you feel that a particular thought did not come from your own cogitations, the conjectural reflective interpretation, that it came from another agent, is considerably strengthened. This is why believers train themselves to identify and monitor those thoughts, the ownership of which is not certain. They are told about and pay special attention to various diagnostic signs (GTB: 63ff.). First, pay attention to emergent thoughts that seem too odd or unexpected to feel like “yours.” Second, check that the thoughts in question “fit” what you would imagine the god might tell you. Third, most important, others around you should agree that the thought may be of divine origin. Fourth, the thought should trigger a unique feeling of peace, the emotional signature of an experience that supposedly cannot be self-generated.

Ownership of mental contents, the intuition that this or that intrusive thought is mine, usually is a tacit aspect of conscious experience. Questions of ownership only arise when people have a definite intuition that some thought or behavior is not theirs. People may come to feel that parts of their bodies are alien (Feinberg, Haber, and Leeds 1990), that their own image in a mirror is that of an impostor (Edelstyn and Oyebode 1999; Ellis and de Pauw 1994). More directly relevant to the Evangelicals’ disciplines, patients in many schizophrenic presentations experience their own behavior, but more often their thoughts or speech as controlled by other agents (Farrer et al. 2004; Frith 1999).

Such states are most often described in the context of rare pathologies, but they are relevant to an understanding of ordinary cognition (Frith 2005; Gerrans 2001). Experimental studies show that the intuition of ownership, although completely intuitive to most of us in most contexts, is certainly not a given. It depends on the workings of dedicated neuro-cognitive processes that match anticipation (e.g., planning to utter a sentence, planning a gesture) and perception (e.g., hearing the sentence from one’s own voice, seeing one’s limb move) (Blakemore, Frith, and Walpert 1999; Farrer and Frith 2002). Any experimentally induced disruption in these processes, in normal subjects, produces a striking intuition that the thought or gesture was “remote-controlled.” Even the illusion of an alien hand can be easily triggered by inducing a delay between the subject’s decision to make a gesture and their access to visual feedback (Wegner 2003; Wegner, Sparrow, and Winerman, 2004).

So it would seem that Evangelicals, in their quest for intuitively non-self-generated thoughts, follow similar tracks to neuroscientists and neuropsychologists exploring the underpinnings of Scheiderian symptoms like alien thoughts and speech. In some sense the situations are symmetrical. The believers expect and hope that external control occurs and interpret some experiences as validating that expectation, while participants in lab experiments just experience unexpected, non-self-generated content and describe the experience as a feeling of external control. But the latter case precisely shows how automatic and compelling the inference is, from non-self-generated to other-agent-generated content. Because of the constraints of our intuitive psychology, it is exceedingly difficult for us to interpret any thought or behavior in mechanistic, nonintentional terms—an agent must be involved (Leslie, Friedman, and German 2004; Malle and Knobe 1997). Culturally successful notions of superhuman agency “free-ride” on this powerful intuition.

Beyond belief

As I emphasized, a great merit of Luhrmann’s detailed ethnography is to show how difficult it is to achieve an intuitive grasp of something—the presence of a god—that is reflectively accepted as certainly true. At the beginning of the book, Luhrmann comments that evolutionary psychology (so far) does not explain why many people think of their gods as real (GTB: xii). That is quite true. In fact one could go further. The more we know about our evolved psychology, the more we understand why most people, at most times, in most situations will not consider their gods real, in the sense of having a definite intuition that the gods are actually there. Our agency-detection and behavior-interpretation mechanisms were tailored to allow the smooth operation of human communication and coordination. Getting ownership right is part of the design of the system, so we should expect that, barring severe pathology, intuitions of nonownership will remain exceptional and difficult to cultivate.

This of course may seem surprising, as a reflective notion of superhuman agency, and its involvement in human affairs, is so pervasive in human cultures, indeed probably one of the most easily acquired pieces of socially transmitted information. But, as I indicated above, the paradox is mostly an artifact of our folk understanding of “belief,” which gets in the way of a proper understanding of mental states. We cannot really understand why a successful cultural notion describes an exceedingly rare intuition as long as we confuse intuitive mental content with explicit reflections, as is very generally the case in anthropology, but also, sad to report, in many areas of cognitive psychology. Sometimes ethnographers have to step in where cognitive scientists got it all muddled up. We should be grateful to the rare anthropologist who, like Tanya Luhrmann, does just that.


Baumard, Nicolas, and Boyer, Pascal. 2013. “Religious beliefs as reflective elaborations on intuitions: A modified dual-process model.” Current Directions in Psychological Science 22 (4): 295–300.

Blakemore, Sarah J., Chris D. Frith, and Daniel M. Wolpert. 1999. “How do we predict the sensory consequences of our own actions?” Society for Neuroscience Abstracts 25 (1–2): 1560.

Boyd, Robert, and Peter J. Richerson. 1985. Culture and the evolutionary process. Chicago: University of Chicago Press.

Boyer, Pascal 1998. “Cognitive tracks of cultural inheritance: How evolved intuitive ontology governs cultural transmission.” American Anthropologist 100: 876–89.

———. 2001. Religion explained: Evolutionary origns of religous thought. New York: Basic Books.

Edelstyn, N. M. J., and F. Oyebode. 1999. “A review of the phenomenology and cognitive neuropsychological origins of the Capgras syndrome.” International Journal of Geriatric Psychiatry 14 (1): 48–59.

Ellis, Hadyn D., and K. W. de Pauw. 1994. “The cognitive neuropsychiatric origins of the Capgras Delusion.” In The neuropsychology of schizophrenia: Brain damage, behaviour and cognition series, edited by J. Cutting and Anthony S. David, 317–35. Hove, UK: Lawrence Erlbaum Associates, Inc.

Engel, Christoph, and Wolf Singer 2008. Better than conscious?: Decision making, the human mind, and implications for institutions. Cambridge, MA: MIT Press.

Evans, Jonathan S. B. T. 2008. “Dual-processing accounts of reasoning, judgment, and social cognition.” Annual Review of Psychology 59: 255–78.

Farrer, Chloe, N. Franck, Chris D. Frith, J. Decety, N. Georgieff, T. d’Amato, and M. Jeannerod 2004. “Neural correlates of action attribution in schizophrenia.” Psychiatry Research: Neuroimaging 131 (1): 31–44.

Farrer, Chloe, and Chris D. Frith. 2002. “Experiencing oneself vs another person as being the cause of an action: the neural correlates of the experience of agency.” NeuroImage 15 (3): 596–603.

Feinberg, Todd E., Laurence D. Haber, and Norman E. Leeds. 1990. “Verbal asomatognosia.” Neurology 40 (9): 1391–94.

Fernandez, James. 1982. Bwiti: An ethnography of the religous imagnation in Africa. Princeton, NJ: Princeton University Press.

Frith, Chris D. 1999. “How hallucinations make themselves heard.” Neuron 22 (3): 414–15.

———. 2005. “The self in action: Lessons from delusions of control.” Consciousness and Cognition: An International Journal The Brain and Its Self 14 (4): 752–70.

Gendler, Tamar Szabó. 2008. “Alief in action (and reaction).” Mind and Language 23 (5): 552–85.

Gerrans, Philip. 2001. “Authorship and ownership of thoughts.” Philosophy, Psychiatry, and Psychology 8 (2/3): 231–37.

Leslie, Alan M., Ori Friedman, and Tim P. German. 2004. “Core mechanisms in ‘theory of mind.’” Trends in Cognitive Sciences 8 (12): 529–33.

Lévi-Strauss, Claude. 1963. Structural anthropology. New York: Basic Books.

Lienard, Pierre, and Pascal Boyer. 2006. “Whence Collective Rituals? A Cultural Selection Model of Ritualized Behavior.” American Anthropologist 108 (4): 814–27.

Luhrmann, Tanya M. 1989. Persuasions of the witch’s craft. Oxford: Blackwells.

———. 2012. When God talks back: Understanding the American Evangelical relationship with God (1st edition). New York: Alfred A. Knopf.

Malle, Bertram F., and Joshua Knobe. 1997. “The folk concept of intentionality.” Journal of Experimental Social Psychology 33 (2): 101–21. http://www.academicpress.com/jesp.

Schneider, David M. 1984. A critique of the study of kinship. Ann Arbor: University of Michigan Press.

Smith, Morten. 1998. Jesus the magician: Charlatan or son of God? Berkeley, CA: Seastone.

Sperber, Dan. 1982. “Apparently irrational beliefs.” In Rationality and relativism, edited by Martin Hollis and Steven Lukes, 35–63. Cambridge, MA: MIT Press.

———. 1985. On anthropological knowledge: Three essays. New York: Cambridge University Press.

———. 1997. “Intuitive and reflective beliefs.” Mind and Language 12 (1): 17.

Sperber, Dan, and Lawrence Hirschfeld. 2004. “The cognitive foundations of cultural stability and diversity.” Trends in Cognitive Sciences 8 (1): 40–46.

Wegner, Daniel M. 2003. “The mind’s best trick: How we experience conscious will.” Trends in Cognitive Sciences 7 (2): 65–69.

Wegner, Daniel M., Betsy Sparrow, and Lea Winerman. 2004. “Vicarious agency: Experiencing control over the movements of others.” Journal of Personality and Social Psychology 86 (6): 838–48.



Pascal Boyer
Departments of Psychology and Anthropology
Washington University in St. Louis
One Brookings Drive
St. Louis, MO 63130, USA