Tanya M. Luhrmann

An explanation is an account of one thing in terms of another thing that makes sense of the first thing for an audience familiar with the second thing. We explain the news differently to a child than to a friend. A doctor explains test results differently to a patient than in an academic talk given to her peers. Explanation depends on audience, and in anthropology, the expectations of the audience have shifted substantially since the discipline was founded. We used to see cultural differences, and those comparisons became the crux of what we wanted to explain. No longer.

We know the story well. Margaret Mead and Bronislaw Malinowski set out to use the comparison of people living differently elsewhere to show that the assumptions Americans and Europeans often made about human nature were sometimes wrong. Then came the tumult of the 1960s. Chastened by critiques about power and exploitation, anthropologists began to think about anthropology not so much as a field about explanation through comparison between societies but more as a means to witness the injustices inflicted by some societies on others (Robbins 2013). They became more hesitant to talk about direct comparison, because they became so aware of the many ways comparison can go astray. In one paper, a group of young anthropologists write, “Of all the social and historical sciences, anthropology is perhaps that which is most formally aligned with the very idea of the comparative.… Yet in practice, social and cultural anthropology may be one of the least comparative disciplines” (Miller et al. 2019, 284). “Where have all the comparisons gone?” bemoans a more senior group (Borofsky et al. 2019). “Where did anthropology go?” asks Maurice Bloch (2005, 1).

The time has come to bring back comparison. Elsewhere (Weisman and Luhrmann 2020), I give a reasoned philosophical and methodological account of this argument. Here I want to explain why comparison can feel liberating for an anthropologist, and how anthropological comparison can open up new vistas in which to think and to explain.

I arrived in Cambridge to study anthropology in 1981. For me, Cambridge was a lot like Hogwarts. By late October, the streets went dark before late afternoon, and the leaves skittered across the flagstones in the damp wind. It was damp; it was always damp in Cambridge, and it was easy to believe that there were ancient secrets in the old stone walls. In their shadows, anthropology seemed like something very new. In fact, early founders of the discipline still lunched in college. Meyer Fortes, Edmund Leach, and Audrey Richards sometimes came to the department’s Friday seminar. E. E. Evans-Pritchard’s niece sold fabric at the end of the street. Where anthropology had seemed like one major among others at my undergraduate American university, at Cambridge anthropology seemed like a young, brash discipline that challenged tradition with mud-splashed truths.

In that heady world, the queen of the social sciences was philosophy, not economics (as people sometimes said back home). Down the corridors where John Maynard Keynes had walked, the ghosts that seemed to matter to the anthropology department were philosophical ones. I read through Ludwig Wittgenstein my second year in graduate school the way young American anthropologists now read Giorgio Agamben, with the sense that these were the texts that serious students mastered (see Heywood, this volume). I attended Elizabeth Anscombe’s last lectures as if they were glimpses of the grail. She was perhaps Wittgenstein’s best student, and she had translated the version of On Certainty that I read. One day in class, she was struggling to ferret out the meaning of a sentence in the text. Why, she asked us, had Wittgenstein used the direct article for this noun? There was a long pause in which she looked thoughtfully at her notes. Then she said suddenly, “Ah! I mistranslated that sentence.” Another day, a timid young scholar poked his head into the room and asked us whether this was a class on the philosophy of knowledge. Anscombe gave him a long, measured look. “You could say that,” she responded. “But that is not what you mean.”

There was a clear sense that the philosophers were the smart, respected cultural insiders and that anthropologists were slightly scrubby outsiders. The sharpest anthropology students went to talks by Bernard Williams and Quentin Skinner and they read Willard Quine and Saul Kripke. The department turned out almost to a person when Richard Rorty came to town. When Jack Goody retired, the department hired a man trained as a philosopher, Ernest Gellner, to take his place. He became my adviser. My generation—Pascal Boyer, James Laidlaw, Simon Coleman, Henrietta Moore—all started out writing as if our ultimate goal was to persuade analytic philosophers to think differently about belief, just as Evans-Pritchard had done.

The problem, of course, was that while the philosophers loved having a direct source of sublunary Martians, as Clifford Geertz so splendidly observed, they weren’t very interested in anthropology. They no longer read books by anthropologists. And the anthropologists were not, in general, much good at philosophical argument. That was not just because philosophers use language in very special ways, honed by years of training. (Here a comment by Wittgenstein comes to mind. A group of men are seated under a tree, he wrote, discussing whether the tree exists. They are not mad, he explained. They are doing philosophy [1969, 120, n.467].) It was because the kind of thing that the philosophers were doing—their basic philosophical project—was not anthropological. The philosophers were ultimately interested in the language game of coming up with a compelling description of words like belief. The anthropologists were trying to explain what was going on when their field subjects sacrificed a cow. As a result, the anthropological work on belief was becoming increasingly frustrating, culminating in Rodney Needham’s (1973) book on the subject, which managed to argue that no one believed anything at all.

Meanwhile the problem with Wittgenstein—and with Michel Foucault, Agamben, and the other philosophers young anthropologists read today—was that their questions were so big that no empirical research could answer them (see Heywood, this volume). These authors invite us to reflect on what we know already about human life and to think about it differently, rather than setting us empirical puzzles we can solve. They ask, What is the nature of human understanding? rather than, Why do the English bury their dead and the Zoroastrians leave their bodies to be picked clean by vultures? This was also the era when Clifford Geertz was arguing that anthropology was more like literary criticism than it was like science. That suggested that the details of what fieldworkers discovered weren’t particularly important as discoveries (although one could say that this was not quite what Geertz had meant). After all, the task of a literary critic is to interpret novels readers have already read, and to help readers into a different understanding of them. Literary critics deliver new understandings, not new data. Then Geertz’s postmodern critics took him to task for his arrogance in claiming to represent the Other. The new intellectual environment—the crisis in representation, the postcolonial critique, the comparison to literary criticism—led many of us to think of ethnography more as interpretation than as discovery research. To be sure, we all understood that our goal was to have a question that fieldwork could answer. But the climate invited us to imagine that our goal was to evoke and to challenge rather than to do what other social scientists did, which was to find out something new and explain why it mattered.

In fact, back in the day, as a young graduate student emerging into an intellectual world shaped by the postmodern and postcolonial critique and by Geertz’s lush prose, I imagined that the goal of the book of the research—the ethnography as published—was to provide an account of another world that would capture that world perfectly in a literary way and as elegantly as Geertz had done. One’s goal was surely to write a book that would live forever, just as the books of Evans-Pritchard and Fortes had done. It would have to be a humble book, acute in its understanding of the limits of observation, because that was how the critique had schooled us. And because it was humble, it would explain something in particular—that other world—but nothing in general. This was the chastening effect of critique. The effect was to shift one’s ambitions away from empirical argument (what was the social effect of literacy?) to imagining the ideal book as a gem: compelling, precise, complete. Such a book would stand on its own to challenge the reader’s assumptions. It would make the strange familiar and the familiar change; it would present a complete picture of its ethnographic topic; and nothing more needed to be said.

And of course, this way of imagining the ethnographic book sets the bar impossibly high. If the goal was to astonish, the job had already been done. Once we have the Azande, how many more sublunary Martians do we need? If the goal was to write a perfect gem, it was always easy to see the need for a sharper cut. In many ways, the intellectual climate of the day set up young ethnographers to feel like failures.

It wasn’t until I began to spend time with an interdisciplinary group at the University of Chicago that I began to think of my own intellectual task differently. The Chicago Templeton Network mostly comprised academic psychologists and biomedical researchers—scientists whose papers were often very short and centered on tables and graphs. They simply didn’t believe me when I told them what I had seen from my ethnographic research—that some people were better at prayer practice than others, and that prayer shaped the way that they experienced their world. The group wanted different kinds of evidence to support the claim. This annoyed me. After all, I had spent many years collecting those ethnographic observations. It made me so annoyed that I set out to prove my point. I did some more structured research and found that the outcome supported my ethnographic observations. When I gave my presentation, beginning with my years of ethnographic research and leading up to my first quantitative finding, they listened patiently until the first scatterplot went up on the screen. Then someone smiled at me and said, “Data!”

This also annoyed me.

Yet the work liberated me because it shifted my focus from perfection to puzzles. The specific attempt to compare people with one another systematically raised as many questions as it answered. I had demonstrated that people who differed in ways measured by a standardized scale were more likely to enjoy praying and more likely to experience what were supposed to be the fruits of prayer—a real relationship with God, a sense of a back-and-forth conversation with him, even a sense that God was sensorially present. I had been able to show that some people were better at prayer than others, and that these differences seemed to change how they experienced their world in the most concrete way.

I had done this by using a scale that measures something called absorption, which asks people whether certain statements are true for them—whether they like watching clouds change shape in the sky, whether they sometimes experience things the way they did as a child, whether they can change noise into music just by the way they think about it (Luhrmann, Nusbaum, and Thisten 2010; Tellegen and Atkinson 1974). The scale seemed to probe the way people experienced their inner and outer senses. It seemed to ask, Do you enjoy getting caught up in your imagination? Do you like to pause and drink in the morning? The work suggested that the way people were oriented toward their mental experience was quite important to their experience of prayer.

Let me pause here. My decision to use a scale might seem peculiar to many anthropologists. Scales probably seem like dead tools that reduce the complexity of human experience. And of course, they do reduce the complexity. But in an anthropologist’s hands, they also help to open up the work. Scales, after all, are really just tools to help us to compare: to say that this group, compared with that group, owns more land, or respects their teachers more, or spends more time in meditation. Anthropologists create scales—structured interview protocols—all the time without using the term. A structured list of questions also helps you to see differences between experts and nonexperts. These people seem more interested in narrative, more focused on detail, than those people, and maybe that difference is important. When we ask systematically about the differences, it helps to build an argument. It gives us confidence in what we see. Standardized scales are created by other people who have used them many times and have confirmed that they work in similar ways in at least some different settings. In this case, I had an intuition; I used a standardized scale, and it supported the intuition; I felt bolder when I gave a talk, because I was more confident in what I had to say. In anthropology, these tools are usually called mixed methods, and they were far more part of the early ethnographies than many of us now remember. Margaret Mead collected an extraordinary amount of data of many different kinds, using all sorts of methods. We might remember her for the Samoan girls she wrote about, but her work on the Arapesh is a dense compendium of mixed-methods data.

Why was this liberating? Because I was no longer aiming for the definitive account, the perfect gem. I had identified a puzzle, and it was clear that many people would chew over it and shake it back and forth like a dog with a rope toy. What, after all, was absorption? Was it an individual trait, somehow encoded in a body? Or a cultural invitation, encouraged by different social practices? I knew what the scale’s authors thought—but I wasn’t sure I agreed with them. Now what I was doing was more like detective work, not like literary criticism. Now the details mattered, because it was suddenly clear that no one actually knew them. No one knew whether the absorption scale would pick up similar religious phenomena in other countries, and what it would mean if it did or it did not. No one knew what it meant that people high in absorption had funny little hallucination-like experiences, nor, for that matter, whether it followed that they were like people with schizophrenia. This was fun. It was not because I had added a quantitative dimension to my work. It was because I had a clear sense that there are real puzzles, and empirical research can help to solve them.

To be clear, I still believe that the basic research method of anthropology is ethnography. I still spend long hours with people. I sit in the park with people who hear voices that taunt them. I go on weekends with people who talk with the dead. But I have now begun to think about the goal of ethnography differently from the way I did in the days before my interdisciplinary encounter. I think differently about how and what I want to explain. These days, I think of myself as having findings.

Findings are empirical observations that call out for explanation by generalization. The generalization is a hypothesis, which later findings will support or challenge. Findings are news in the way that the general idea that Moroccans, say, are Moroccan is not. They offer puzzles that need to be solved.

Here is an example. I have begun to spend time in psychiatric settings outside the United States, chatting with patients and doctors, learning about how people find their ways into an inpatient ward. What I do these days that I did not do when I was younger is to ask those patients systematic questions about what they hear. That means that I can compare them with what other patients say elsewhere.

In the Accra General Psychiatric Hospital, I found that people who meet criteria for schizophrenia often reported that the voices who speak to them come from God. On average, they reported that their voices are more positive than those reported by a comparable group of Americans. In Chennai, similar subjects more often said that they heard their kin. They more often said that they heard the disembodied voices of people they already know in the flesh. Not one person in my American sample told me that their dominant voice-hearing experience was positive, and only three Americans out of twenty told me that they heard the voice of someone they actually know (Luhrmann et al. 2015a, 2015b). Why? I think it might have something to do with the social worlds in which they begin to hear those voices. That’s important, because how harsh the voices are seems to have something to do with how well people recover—and if the content of the voices can be shaped by culture, it suggests that medication ought not to be the only way our clinicians intervene.

I think if anthropologists went back to empirical comparison, the field would feel liberated, just as I did. If we as a field did empirical comparison with the aim of observing different patterns of behavior and developing theories that explain specific differences—findings—in such a way that we could say whether those theories are better or worse, we would have so much to say, and to so many people. Empirical comparison does not need to be an assertion of arrogance, as the postmodern critique sometimes assumed, but a concession of humility. Claims based on findings are necessarily limited. They are partial attempts to explain a puzzle. Once presented as such, the ensuing debate keeps one more humble still. They are contributions to a conversation in which one is one of many players.

And findings matter. If I can show that the voices heard by people with psychosis are different in different countries, I am able to argue that voice-hearing responds to learning. That paves the way to think differently about what we should do for those who want to experience their voices differently. Explicit comparison enables us to make claims about the way specific cultural features may have specific consequences. And that, as Mead urged long ago, is how anthropology could change the world.

Now I want to say more about my most recent foray into comparison.

Some years ago I wrote a book, When God Talks Back (Luhrmann 2012), about the evangelical Christianity practiced by maybe a quarter of Americans. In these faiths, people seek out a personal relationship with an interactive God. In these faiths, God talks back. Those who worship imagine God as actively, supernaturally powerful, and as intimate. Such a God cares about humdrum worries; he wants to hear about them and to talk about them. He can and may step in to improve the quality of a haircut. The challenge of the research was to understand how a human was able to experience an invisible other in dialogue. One cannot foreclose the possibility that the supernatural is present (see Bialecki, this volume). But the supernatural is, by definition, nonnatural: nonordinary, not materially available to the senses. In this kind of Christianity, people talk about hearing God “speak.” I wanted to know what they meant in uttering that sentence and what that experience was like.

What I saw was that knowing God in this way was not so much a matter of belief but a matter of skill—something that you learned to do—and that the main vehicle for the learning was prayer. People usually prayed informally, and in that prayer, they conducted daydream-like conversations with God in their minds. The church invited people to develop these daydream-like dialogues in particular ways.

First, the church invited them to think about their minds not as private but as containing thoughts and images and sensations they might once have understood to be internally generated but were in fact communications from God. Not all thoughts counted: people were encouraged to attend to the “texture” of thought, or the “topography” of mind, and pick out thoughts that were “stronger” than others, or felt more spontaneous.

Second, the church invited them in to practice their dialogue with God by in effect pretending that God was present: by going for a walk with God, or by asking God what shirt he wanted them to wear. They did not consider God to be imaginary, nor did they think that they were doing “mere” pretend. C. S. Lewis entitled one of the chapters of Mere Christianity “Let’s Pretend” to suggest that we pretend in order to experience as real. That is what the church intended.

Third, the church invited them to practice being loved unconditionally by God in a variety of ways. People learned to talk to God as they might to a therapist, and waited to hear what this wise, sensible person might say. They learned to remind each other to see themselves from God’s point of view—not as the weak, inadequate person they felt themselves to be, but as a loving and empathic observer would see them. And they stood in for God within the prayer circle, praying out loud to someone the words they felt that God was giving them to say.

Newcomers to this kind of church would begin by saying that God didn’t talk to them. Yet after some months they would sometimes report that they could recognize God’s voice the way they recognized their mother’s voice on the phone. Sometimes they reported that they heard God speak in a way they could hear with their ears. This experience of an audible voice was not a common occurrence—these were not psychotic voices—but they were important to those who experienced them.

One could see, then, that there was learning, and that the learning was not just about learning to say certain things—to adopt a certain discourse—but that it changed something about the experience of mind. My next methodological move came out of my new respect for findings and my burning desire to answer my critics. I added experimental methods to see if I could replicate these ethnographic observations by other means. We brought in over a hundred people and randomized them to prayer practice or to lectures on the gospels. Those in the prayer group were more likely to report more vivid mental images; they seemed to use mental images more; they reported more unusual sensory experiences and more intense spiritual experiences; they experienced God more as a person, felt his presence more, and interacted with him more often (Luhrmann and Morgain 2012; Luhrmann, Nusbaum, and Thisted 2013).

Why bother with the experimental work? Because fundamentally this was a claim that depended on individual differences: on differences between before and after, between trained and untrained. When God Talks Back argued that paying attention to what one imagines makes the world of the mind more vivid, and that this was central to understanding imagination-rich prayer. The big claim is that paying attention to inner events alters someone’s awareness of those events to the extent that the attention can alter their decision about what they perceive to be true—a human experience that explains something about how prayer works in different cultural settings. The ethnographic data presented phenomenological findings that demonstrated general patterns of experience; the experimental data demonstrated that proclivity and practice shifted those patterns in predictable ways. Ethnography is good at capturing general patterns, the underlying grammar of experience. But the most persuasive way to identify individual differences between those who pray and those who do not involves comparison within a group of subjects, and that is most easily achieved by experimental design.

Yet it also seemed that there was something about the experience of hearing God that was hard for my Christian subjects—and this was an ethnographic observation about a specific cultural world. These Christians would talk about struggling to experience God as real in their thoughts. They would say things like, “You’ll think I am crazy,” or “You don’t need to send the white coats for me.” The fear of being crazy was even more pronounced when they reported hearing God speak with their ears. These hesitations seemed to have to do with the secular context and the psychological model of mind among middle-class Americans.

That suggested that there was a scholarly and scientific story not only about the consequence of giving inner experience increased significance through prayer but also through the way that inner experience was understood in a specific social world. In other words, it seemed that there was a story not only about attention in general but about the particular way people learned to pay attention in particular social worlds. It seemed likely that different social groups encourage people to attend to their mental activities in different ways—that they imagine mind differently, that they imagine thought differently, that they have different assumptions about what happens when you think and whether that matters and how people learn. That is a hypothesis about cultural variation in theory of mind.

Psychologists have demonstrated that the basic presumption that the mind is separate from the world is one of the most definitive achievements of childhood development. By the age of three, more or less, toddlers understand that mental states can explain behavior. The classic experiment is the “false belief” task. A child and the child’s mother watch an experimenter hide a toy. Then the mother leaves the room, and the experimenter moves the toy to a new hiding place. When the mother returns, the experimenter asks the child where the mother thinks the toy will be. (This is often represented by dolls, as in the “Sally Ann” tasks.) Very young children point to the second hiding place, because that is where the toy actually is. They presume that all people know what they take to be true about the world. Older children understand that the mother does not know that the toy has been moved, and so they point to the first hiding place. They have developed what psychologists call a theory of mind. They understand that what people think they know may be different from what has happened in the world (Wellman 2013).

Yet what the observer really observes is that the child draws inferences about knowing, believing, wanting, and intending. “Mind” is the name for the container we imagine for those acts. And anthropologists know that different social worlds represent that container differently, although we as a tribe have not done much with that observation. In 1988, the psychologist Angela Lillard pulled together these anthropological findings (Lillard 1998). Euro-Americans, she pointed out, imagine the mind as the seat of mental processes; as located in and identified with the brain; as private but knowable; and, judging by the size of the vocabulary used to describe it, as extremely important. In Euro-America, the mind gives an individual his or her identity. Your feelings, your beliefs, your ideas—in Euro-America, they make you you.

Many other cultures locate the mind not in the brain but in the heart, as the Ilongot (a tribe in the Philippines) do (Rosaldo 1980). Among the Ilongot, many of the mind’s features are attributed to social interaction. “What is important for Ilongots is not what goes on in the rinawa [the closest word for their conception of mind] but between people” (quoted in Lillard 1998, 12). The clear distinction Euro-Americans make between mind and body is more muted in other social worlds. The Japanese, for example, make no simple distinction between the two and do not identify self with mind (Lebra 1993). And mind is often less elaborated outside the Euro-American context. Signe Howell (1981) made determined efforts to describe all mental process terms among the Chewong, a tribe on the Malaysian peninsula. She found five (want, want very much, know, forget, and miss or remember). There was, she reported, no word for think.

Indeed, in many parts of the world, people resist interpreting what is in other people’s minds. This tendency is so pronounced in the South Pacific and in Melanesia that anthropologists have come to speak of “opacity of mind” (Robbins and Rumsey 2008). Mayans also privilege the description of behavior over mental states but seem to be more concerned with the accuracy of the world-word relationship. Eve Danziger has found that Mopan Mayans explicitly disregard mental states and de-emphasize fantasy and pretend play (as Suzanne Gaskins also demonstrates for Yucatec Mayans) because what counts is the direct correspondence between speech and fact—not whether the speaker, at the time of uttering, believed that the spoken claim was accurate (Danziger 2006; Gaskins 2016). Danziger, like Lillard, uses this example to demonstrate that the standard psychological theory of mind is culturally Euro-American. The Mayan child is not behaving in the way that many theorists have suggested, explicitly formulating models of mental states. The child does draw the correct inferences, but in a way different from the Euro-American expectations of the mind.

In 2011, we organized a conference that brought together anthropologists and psychologists to see whether we had sufficient evidence to argue for local theories of mind.¹ We identified at least six such theories:

The Euro-American Modern Secular Theory of Mind: In this theory of mind, people treat the mind as if there is in effect a clear boundary between what is in the mind, and what is in the world. Entities in the world, supernatural or otherwise, do not enter the mind, and thoughts do not leave the mind to act upon the world. The assertion that they do is seen as a symptom of mental illness (thought insertion and thought withdrawal). What is in the mind is not real in the way that tables and chairs are real; one can speak of “mere” imagination. At the same time, what is held in the interior of the mind is causally important. Intentions and emotions are powerful and can even make someone ill.

The Euro-American Modern Supernaturalist Theory of Mind: This theory can be found undergirding charismatic Christianity, contemporary spiritual healing, alternative bereavement practices, paganism, and other practices that are sometimes identified as “new age.” Here people treat the mind as if it conformed to the modern secular theory, except in specific respects. The mind-world boundary becomes permeable for God, or for the dead person, or for specific “energies” that are treated as having causal power and, usually, their own agency. The individual learns to identify these supernatural presences, often through implicit or even explicit training. Other features of the secular theory apply. The training becomes important because the secular model of mind is the default model with which these individuals work.

The Opacity of Mind Theory: This theory is found in varying forms throughout the South Pacific and Melanesia. Its most striking feature is the insistent refusal to infer what other people are thinking unless they verbalize their intentions. In these societies, asserted intention is taken to be causally powerful in a way that felt intention is not. That is, whatever one’s actual intention may be, the intention one asserts (or is taken to have asserted) is taken to be causally powerful. The impropriety of inferring privately held intention is so great that it can be impolite to look directly into another’s eyes. At the same time, the boundary between the mind and the world is often porous, so that spirits (or the Holy Spirit) pass back and forth across it.

The Transparency of Language Theory: In these societies, for which our best examples come from Central America, language is understood to align with the world rather than to express interior states. Fiction may be frowned upon; play by children may be tolerated but not encouraged. The vocabulary to describe mental states may be thin or near-non-existent. Most notably (to ethnographers) beliefs that happen to be false but are not understood to be false by the speaker may be identified as “lying.” When an utterance is assessed, what matters is its truth-relationship to the world rather than its relationship to the intention of the speaker.

The Mind Self-Control Theory: Our best example of this theory locates it in Thailand, but it can be found in different versions throughout Asia. In Thailand, the most important concern around the mind is how well it is controlled. One can have a well-controlled mind, or one that is less well-controlled. When the mind is poorly controlled, emotions and intentions become powerful and can enter other poorly controlled minds as ghosts or spirits. Thoughts thus are real in a way that is quite different from the Euro-American model. One’s mind can be unbunched, and can wander. Thoughts can act in the world and on other minds in ways that are only partially related to those that first thought them.

Perspectivism: This theory suggests that the world is dependent on the perspective one takes on it. Many Amazonian peoples are held to conceptualize theoretically the world as if it is seen from a particular perspective: a human’s or a jaguar’s, for example. Here there is a great deal of interest in the idea that what appears to be blood to a human may seem to be beer to a jaguar (for instance). There is an expectation that a human can become a jaguar and vice versa. The most important feature of a mind seems to be that it can migrate from body to body. People sometimes fear ending up in a non-human form permanently. People can make claims that they have seen other people becoming non-human, or been with humans who become non-human. (Luhrmann et al. 2011, 6–7)

This list is not exhaustive. It may not be accurate. It does not imply that everyone in these social worlds thinks about the mind in the same way. The sheer existence of the list, however, supports the general claim that there are striking cultural differences in the way distinctive social groups imagine the map of this human terrain of knowing, intending, desiring, and feeling—the hilly, pockmarked terrain of human awareness.

Do these differences matters? That is what I set out to answer five years later, drawing on my initial comparative work in the San Francisco South Bay; Accra, Ghana; and Chennai, India (Luhrmann 2017). The Mind and Spirit Project (funded by the John Templeton Foundation) set out to ask, Is there a relationship between the way people think about thinking (their local theory of mind) and the way they experience the supernatural?

To be clear, this is a question about the relationship between the way a social group represents the domain of what we Western English speakers call the mind—what we call their local theory of mind—and the way people actually experience what they identify as the divine or supernatural. There is of course no strict distinction between experience and representation. Decades of philosophical and ethnographic analysis have taught us that. But there is a difference in the target of the research. We set out to identify models of mind across different social worlds and experiences of what William James called the “more”—that which exceeds a materialist understanding of the world as we know it (1935, 501). Our aim is to ask whether there is a systematic relationship between the two.

Why should there be? Much of what humans deem spiritual, to use Ann Taves’s (2009) useful phrase, involves the identification of what a skeptical Anglophone observer would call mental events: an impression or thought, a quasi-sensory voice, an awareness. American charismatic Christians seek for the voice of God within their minds. They expect to hear God speaking back through their thoughts. They must, then, be able to identify a thought—and that is a culturally shaped act. Bruno Snell (1960) and Maurice Leenhardt (1979) each describe a social world (archaic Greece, early twentieth-century New Caledonia) in which people do not have the vocabulary to describe inner experience and in which the gods act in concrete ways; Julian Jaynes ([1976] 2000) argued that the one was responsible for the other.

How do we compare? We chose five countries that, because of existing historical and ethnographic research, we thought encouraged different representations of the relationship between mind and world: China, Ghana, Thailand, the United States, and Vanuatu. (My primary coconspirators here were Christine Legare and Kara Weisman.) We appointed ethnographers competent in the local language and experienced in local research to spend nine months in the field and two years (and more) on the project. (They were Felicity Aulino, Josh Brahinsky, John Dulin, Vivian Dzokoto, Emily Ng, and Rachel Smith.) We found an “apple” for an apple-to-apple comparison: an urban charismatic church with specific features—charismatic experience, aspirationally middle class, with a theological expectation that God would speak back to each of his worshipers. All churches are different, but Pentecostal and neo-Pentecostal congregations are at least presumed to have more similar—more portable—cultures than many (Freston 2013; Robbins 2004).

In each setting, we did a set of intensive, semistructured interviews—one on the mind, another on spiritual experience, each usually one to two hours in length. Each fieldworker did twenty such interviews in each site. We did similar interviews with related groups—a rural variant of the urban church, and an urban and rural faith of local importance (Methodism in the United States, Buddhism in Thailand). To minimize the problems caused by having different interviewers, our team spent four months together reading, talking, developing, and piloting the interviews before the work began. To see if our observations would be consonant with other groups in the same country, we also collected other kinds of data: one hundred shorter interviews in each setting in a location like the local department of motor vehicles; packets of pen-and-paper surveys on mind and spirit with local undergraduates; developmental experiments with children and adults. We have more than one way to compare.

The hard part was deciding what to ask. What serves as an object that can be compared like to like? This is a matter that involves choice, compromise, and limitation. There were three issues. First, time is finite. One cannot talk to everyone about everything, or even to one person about very much. Second, not all questions are worth asking. Some events are so rare or so difficult to describe that asking about them generates mush. I am willing to grant that there is something like a mystical experience with the characteristics William James ([1902] 1935) thought were so important. But in my long experience of asking people about spiritual events, I have found that when I ask people whether they have experienced a remarkable event with qualities of ineffability, transience, noesis, and passivity, people often say, “every afternoon.” That answer is undoubtedly wrong for the rare events James described. We left mystical experiences off our list, although we did ask people to describe their most remarkable spiritual experience and so left room for them to describe such an event. Third, there are unknown unknowns. There are events not thought of in our Euro-American world that do exist in others. In our four-month preparation, one fieldworker—Rachel Smith, who went to Vanuatu—kept asking how we would ask about small blond dwarves. At first I took this as a metaphor for all we did not know to ask. Then it turned out that in Vanuatu, people do indeed report small blond dwarves. So we created a question we could use everywhere that we hoped would lead people to talk about the dwarves in Vanuatu.

That introduces a basic rule: conversation, conversation, conversation. Our interview protocol was built on my own experience in talking with charismatic Christians in Chicago and then on the San Francisco South Bay peninsula. Then we talked and talked about the way those questions might be received by Christians and non-Christians in our different settings. We piloted in the United States, and when the fieldworkers arrived in their sites, we piloted again.

We chose to ask about two kinds of spiritual events: general events deemed religious, particularly about the way that spirit spoke or communicated, and what the folklorist David Hufford (1982) calls “core” experiences. These are events that appear to occur in many cultures; that occur in some people independently of prior expectation; and that appear to form distinct patterns with stable traits. These are events such as voices, visions, a sense of presence, out-of-body events, and sleep paralysis. They are often described in the medical and historical literature. I think of core experiences the way psychiatric researchers these days think about psychiatric diagnoses—that they are somewhat fictional but useful categories that more or less pick out differences in human experience.

Then there is the mind. What on earth is the mind, “the thing that thinks,” in René Descartes’s phrase? There is nothing about which we have more privileged knowledge and about which at the same time we know so little. No one has access to what Augustine called the vast choirs of our memory, and yet when we search for thoughts we sometimes find ourselves grasping shadows in a mist. James ([1890] 1950, 179) described the resistance to describe the mind on the grounds that the concept of mind is hard to describe precisely as “spiritual chloroform.” So we proceeded.

We began with a heuristic: an analytic distinction between the human experience of thinking, intending, believing, wanting, feeling, and the awareness of those mental acts, and the culturally specific representations of those acts. We distinguished them as terrain and map: the first as the phenomenological experience common to humans, and the second as the way those phenomena are represented in a local social world. We presume that not all the human terrain is locally mapped as mind—that specific kinds of mental events, such as anger, might be mapped more as part of body or of spirit. We also presume that the local mapping changes the experience of the event. But we do not presume that the map is a cookie cutter that completely determines human experience.

We set out to identify the way our local communities mapped this terrain by asking our interview participants about a series of stories we hoped would capture the way they imagined what thoughts and feelings could do and how and why; whether thoughts and feelings should be shared, and with whom and when and why; and whether what the mind invents out of whole cloth is real, and should be encouraged, and why.

We have a long way to go, but we appear to have a finding: that when the terrain of the mind is mapped so that it is imaged as more porous, more permeable, so that thoughts cross back and forth across the mind-world boundary to affect the world directly, people report more bodily spiritual events. When the terrain is mapped in a more bounded way, as Charles Taylor (2007) described—when the mind is imagined as an epiphenomenon of the brain, when thought is supernaturally inert, when feelings like anger affect one’s own body but not the material world beyond—spiritual events are less sensory and more thought-like. God speaks into the mind, rather than with a voice one can hear with the ears. When people represent thoughts and feelings as potentially potent, as when ideas about witchcraft and sorcery are salient, or when people imagine the mind as vulnerable to thoughts and feelings from elsewhere, as in divination, they are more likely to describe moments in which they see spirits with their eyes or hear them with their ears.

What constitutes an explanation here? The paradoxical point is that the more we emphasize the finding as something that needs to be explained, the less important, in some ways, the explanation becomes. In some sense, an explanation is always emergent. Our current explanation (Luhrmann 2020) is that these ideas about thought affect the way thought is experienced: as more substantial, as more viscous, as more thing-like. We think that these expectations may alter experience so that sometimes, thought-like events pop out into the world and are experienced as more external. As we learn more, as we do more work, no doubt our explanation will change. What we feel we have achieved so far is the observation of something. There are different models of mind, and something about them seems associated with more vividly sensory spiritual experience. We have an empirical finding that calls out for explanation. We do not have a polished gem that describes something in particular deeply but generalizes nowhere. We have a puzzle, and we hope that it provokes debate.