Common sense holds there are distinctive religions, an intuition that informs most scholarship and teaching in religious studies and the social sciences, but the intuition is somewhat misleading. In spite of apparent religious difference, recent psychological inquiry suggests that religion emerges from a single panhuman psychological design that strongly constrains variation. There is some variation in the religiosity of individuals and groups, but not the variation of “traditions”. This paper uses recent research in the cognitive and evolutionary study of religion to explore some basic properties of the mental architecture that generates human religiosity, including features that enhance the illusion of religious difference.
Key words: behavioural ecology, biology, costly signalling, evolutionary psychology, game theory, healing, linguistics, prisoner’s dilemma, morality, religion.
The purpose of this paper is to examine recent advances in the evolutionary psychology of religion, and to explore how this research revises standard accounts of religion and religious variation.2 For example, whereas ordinary language (and first year world religions courses) carve out particular doctrines and practices as belonging to distinctive religious kinds—Islam, Judaism, Christianity, Hinduism, California New Age, Satanism and so on—it seems these labels do not describe aspects of the natural world (they do not reflect the kinds of nature.) Instead, panhuman psychological architecture generates only limited, but strategically important, variation in the religiosity of individuals and groups. Religion is thus like language—for linguistic inquiry has revealed that variation in human languages is limited to superficial aspects, which conceal a universal biological design. I examine the similarities between language and religion in the first section below. I then describe some important aspects of the psychological design that equips us for religious thought and practice, and show how this design emerges from genotypic resources that produce religiosity within narrow parameters. Though cultural environments play a role in an agent’s particular brand of religiosity, acquisition events do not explain the intricately structured understandings and motivations that emerge through religious cognition. I show how the psychological faculties that subtly prompt and guide religiosity are designed to promote biological success in the ancestral world. I conclude that in spite of striking apparent variation, core aspects of religiosity remain invariant. From the vantage point of cognitive architecture, it appears that there is only one human religion with minor but strategically important variation in its conventional expressions.
Let me first address the objection that my thesis—the idea that there is one human religion with variation at the margins—relies on an ambiguity in the similarity relation. Notice our judgment of similarity is always interest relative. Depending on our purposes, we may correctly judge that oranges and apples are both similar and different. Oranges are similar to apples when compared to black holes, but different when compared to tangerines. Black and white are distinct colours, yet both are colours, so in this respect similar. But this does not entail that there is “only one human colour with variation at the margins”, or “essentially one human fruit”. It seems plausible to say that religions are both similar and different depending on what we take to be the relevant comparison classes, and the practical specifications we use in judgement. Moreover it seems fruitful to discard the idea that “all religions are essentially the same” as too crude and reductive to have explanatory value. For it is unclear how we could ever gain by lumping the products of culture into gigantic categories that way. (Compare: “All games are the same;” “All marriages are the same” “All science is the same”.)
I do not deny the legitimacy of using concepts of distinctive “religions” in making pragmatic sense of religious beliefs, institutions, and practices. One can truthfully (and helpfully) say, “Indonesia is largely Islamic” but not, “Indonesia is largely Hari-Krishna”. The Taliban’s religiosity differs from feminist Earth goddess religiosity; for example, the Taliban aren’t feminists. Generalising: there are undeniable differences among agents in religious labelling and doctrinal, moral, and ritual conventions. And there are local differences in the intensity of religious commitments, and in the uses to which religious ideologies are put. Most scholars of religion labour amid these peculiarities of convention and circumstance, with illuminating results. I see nothing in the evolutionary psychology of religion that challenges this interest.
Yet there is new understanding to be gained from an evolutionary perspective that views religion as a panhuman capacity. We will see how the psychological architecture that supports religious commitment and practice is largely invariant across culture and era. And this architecture reliably produces functionally similar expressions and commitments, irrespective of labelling and convention. Though religiosity is always framed by local settings and histories, we will see that that “religion” is no more a “product of culture” than jealousy or friendship. We can talk meaningfully of the distinctive religiosity of Indonesians, the Taliban, and Earth worshipers while still appreciating that variation is strongly bound and directed by a common psychological design. The trail of this universal design runs over all religious products. The fact that this design is a common human possession that we share through biological endowment is a scientific discovery, not word play. I seek to explore aspects of this discovery below. As a way to understanding the issues, consider language.
On the face of it, there appear to be numerous and different languages, “English,” “Japanese,” “Swahili,” “Greek,” “Hindi,” and others. That they differ is made vivid when conversation partners do not share one: the tourist of China, lacking Chinese, falls on hard times. Even the same language varies over time. A speaker of Chaucer’s English could not communicate with a contemporary speaker of London’s English, though an unbroken chain of only twenty-odd generations of communicating English speakers separates them. 14,000 years from now no one will speak our English. Our distant successors will look on our speech as the Chinese look on the English-only tourist, incomprehensibly.
Pronunciations of words vary. Germans say, “rot” to mean “red,” not intending the “rot” of “rotten wood”. French speakers use “raison” to denote “correct” instead of that black dried fruit; they say “anniversaire” to mean “birthday,” instead of “anniversary” and “blanc” to mean “white” not “black” [see (Harman 1998)].
Grammars also vary. English displays relatively stable word order regularity, with adjectives typically preceding nouns that they modify, and subjects usually preceding transitive verbs, and (though less frequently) intransitive verbs. Yet differences among languages in this respect seem quite extreme. Aboriginal speakers of the Australian language Warlpiri use case markers to convey grammatical relations and noun modification. In certain Native American languages, there are few noun phrases within clauses, and grammatical relations are expressed by attaching strings of agreement affixes onto verbs (Pinker and Bloom 1990). More basic differences include rules governing the designation of subject and object, either by word order, as in English and Japanese, or by case, as in German, Latin and Gzechoslovakian (Pinker 1999)—a bane to those of us who study these languages as adults. Lots of apparent differences here.
We observe that children learn the language of those around them. An African born in Toronto utters English sentences ending with “eh?” instead of some African language. Language seems to be something children pick up from their surroundings.
Call linguistic externalism the view that explains linguistic competence as structured through local acquisition events. Many theoretical versions of linguistic externalism are possible. I’m interested in the version that most closely approximates our common sense view that language has something to do with acquiring an artefact called “language” from a community or culture. Folk linguistic externalism undertakes something like the following commitments:
1. Languages are public tools for communication shared by members of different linguistic communities.3
2. There are different languages.
3. These communal tools are acquired through learning. Roughly, something “external”—a language—is internalised by individuals in the course of development through cultural exposure.4
Each of these assumptions turns out to be false. Noticing why will help us to see where folk externalism about religion goes wrong.
Consider linguistic difference. While pronunciations vary, grammatical variation is tightly constrained. It is not the case that grammatical differences of the kind that distinguish English from French or Cherokee are rigid. English speakers can invert subject-predicate word orders, for example: “The instructor was driven to drink by the student’s passive sentences”. In fact, English speakers do sometimes employ case markers, for example, “’s” for possession as in: “the student’s passive sentences”. We can produce ergative constructions, replacing “the bottle broke” for “I broke the bottle”. Moreover, there are converse orderings of English-like constructs in apparently grammatically distinct languages [For discussion see (Pinker and Bloom 1990)].
Focusing on the grammatical component more closely we find invariant principles of sentence formation children never learn. Take a descriptive generalization called the coordinate structure constraint. This constraint exposes fundamental differences in structure between sentences of the following form:
1.) Mary saw Peter with Paul.
2.) Mary saw Peter and Paul.
In English, questions can be formed by inserting a question word at the head of the sentence, followed by an auxiliary verb. Hence:
1.) Whom did Mary see Peter with?
However this transformation doesn’t work when the question word is conjoined with another noun phrase, as in:
2.) Whom did Mary see Peter and?
Children never explicitly learn the coordinate structure constraint, nor do they ever say, “Whom did Mary see Peter and?” because the constraint is hard-wired.
Internal knowledge extends beyond the construction of sentences to the meanings of words (as opposed to their pronunciations). Word meanings possess extremely complicated relational properties that we never learn. Chomsky illustrates this point through the following examples. Take “house” in the sentence “Peter is near the house”. Notice the implication is that Peter is outside, not standing near the inside wall. So it is with “car,” “airplane,” or even an impossible object like “rectangular sphere”: we assume the same for “Peter is standing near the rectangular sphere” (i.e. near the exterior surface). Similarly, when we say “Peter painted the house red” the default assumption is that he did something to the outside, not the inside. If Peter cleans the house, however, the default assumption is that he rearranges objects on the inside. We conceive of “house” therefore as an exterior surface with internal spaces, both of which have complex properties. Chomsky notes that “home” has different implicit properties. If I have shifted my house from New York to Moscow, I have moved a massive wooden object. Notice I convey a different understanding when I say that I have shifted my “home” from New York to Moscow. “Home” has both a concrete and abstract aspect. Exploring “home” further we note its abstract properties differ from those of “book,” which is also concrete and abstract but in different ways. You and I can simultaneously read the same book even if we live at opposite ends of the planet, but you and I cannot simultaneously live in the same house or home at opposite ends of the planet (Chomsky 2000: 31–37, 62–66).5
Not just grammar, then, but substantial semantic components of the psychological systems that produce language remain invariant. Strikingly, it appears that we all think from the same mental dictionary, with variation limited largely to conventions of pronunciation, how we say “black” “home” “Moscow” and other words.
Linguistic externalism is committed to the view that we learn these intricate meanings and grammatical rules through acquisition events, but this is implausible because competence emerges from severely limited exposure, a “poverty of stimulus”. Acquiring the massive and largely tacit knowledge required for linguistic understanding and generalisation is computationally intractable within the time of childhood development. Children are simply never exposed to all that they know.
Rather than thinking of language as an acquired artefact, linguists think of language as a “mental organ”, which develops along a more or less fixed schedule, in response to environmental inputs but whose intricate structure cannot be explained meaningfully in virtue of those inputs (Pinker 1994).6 We assume as much with ordinary organs (the spleen, liver, eardrums …) whose organic design and development cannot be meaningfully explained in terms of “food”.
Clearly, it would mislead the English-speaking tourist to say: “Don’t worry about China—you speak the same language”. Yet we can understand the differences without appeal to artefacts called “languages”. The differences adhere to the slight variation in language faculties or “language organs” of individual speakers (linguists call these “I-languages”). I say to you, “Bring your translation book”. You understand these words because my I-language specifies the linguistic (phonetic, semantic, structural) properties of the sentences that you hear. The state of your I-language is similar to mine because you have been exposed to a community of speakers of a related linguistic heritage. This enables you to select an appropriate analogue with which to interpret my utterances with sufficient precision for understanding. Individuals unlike us whose I-languages develop in different environments may not be able to select an appropriate analogue if the state of their I-language differs too much (again variation in I-languages is limited pronunciations and a handful of grammatical rules).
Notice that we can account for the tourist’s failure to communicate in China, or for the differences between Italian and German courses entirely on these terms, without appeal to public entities or artefact, which individuals grasp partially through “learning” aspects of them. Social exposure explains some conventional aspects of language, but the skill is structured through biological endowment.
I discuss language at length because it is a well-studied psychological phenomenon, and the results of that study cast serious doubt on the assumptions of linguistic externalism.7 However, it is possible to accept the internalist view of language while remaining an externalist about other aspects of cognition. Take our knowledge of quantum field theory. Here complex understandings are not structured by “mental organs” (at least not straightforwardly). Rather, knowledge hinges on scientific discovery and active transmission.
Clearly externalist projects are relevant to the explanation of many facets of human skill and intelligence. As Sterelny notes, “[c]hildren resemble their parents because of a flow of genes. But they also resemble their parents because of a flow of information” (Sterelny 2003). We are not born with blueprints for all we will know and will become: part of us is made by the culture we are born to. Cultural inheritance plays an important causal role in phenotypic expression. For this reason cultural inheritance may also play an adaptive role: for example it enables successive generations to accumulate adaptive knowledge over time. Cultural know-how—how to make a pot, how to construct a canoe, where to find water in the dessert, how to detoxify seeds—may be vital to success (Henrich and McElreath 2003). Moreover, such knowledge clearly varies between cultural lineages. Consider war technologies. The history of colonialism makes no sense without recognizing some human societies have been better equipped to savage others. Reading and writing ratchet civilization forward, because the technology enables the accurate storage and transmission of cultural know-how. Yet these abilities come too late to have been targeted by selection (Dehaene 2003). Moreover, once concocted, culturally born technologies and artefacts may alter selective landscapes. We do not merely respond to our environments, we transform them as “ecological engineers” (Odling-Smee, Laland et al. 2003). The products of culture frequently have important fitness affects on those who come after. The invention of penicillin makes the world safer for surgery, and both technologies improve the fitness of agents prone to acute appendicitis, and their descendants, altering the frequency of the relevant alleles down stream.
Reflecting on the importance of cultural endowment to all of human life, it may seem as if religion’s explanation demands an approach that take us beyond genetic endowment, for religion seems to be a product of culture par excellence. There was no Buddhism before the Buddha invented it. The religion did not spring to the young Richard Gere’s mind because it was prefigured in his parent’s gonads.8
Call religious externalism the plausible view that religion is an aspect or product of culture. Roughly, the view asserts that:
1) Religion is substantially a cultural invention or technology for living and understanding.
2) There are different religions
3) Children (and adult converts) acquire their religion from others during the course of their education or social experience.
I will show that recent naturalist inquiry casts doubt on each of these three propositions. In contrast to religious externalism, call religious internalism: the view that:
1) Religion is substantially structured by genetic resources (the architecture of religious cognition is not an invented technology).
2) Religious variation is minimal.
3) Children are endowed with religion, which “grows” in certain cultural environments, because it is a feature of panhuman psychological design.
Now I need to describe the sort of evidence that would count in favour of an internalist approach to religion.
Begin with following relatively uncontroversial view about nature and function of cognition.9 Minded organisms are equipped to process information in ways that enable them to flourish. They gather, store, and access information about the external world and their various internal states, which they exploit to respond and behave to enhance biological success. [”Behave” in the widest sense: on this view, an immunological response or dying old (lifespan) is behaviour.] The functional organisation of any cognitive agent is in part specified by species-shared properties of its genotype. Thus, honeybees are equipped to communicate information about the location of nutritional resources through the subtle staging and monitoring of a figure eight dance. They furthermore possess cognitive powers to plot a course from hive to the (dance-referenced) nutritional supply, exploiting environmental cues like sun angle or (on cloudy days) light polarization and other local markers (Gould 2002). By all accounts these capacities are structured by genetic inheritance. You cannot teach these skills to drosophilae. Clearly environment plays a roll, even here. For local informational resources matter to an agent’s thought. The particular angle of the sun during a particular honeybee’s flight enables it to determine its course. Moreover, structural features may emerge through type-specific environmental interactions over time. Male zebra finches placed in song impoverished settings never acquire the capacity to sing. These bird capacities require external stimuli. The linguistic inputs of children raised in Nepal influence their individual (token) I-languages, which do not stabilize until late adolescence. Their I-languages will differ somewhat from those of children raised in Los Angeles. In these cases, however, though structured capacities key to environmental features, structure comes from genetic resources. Zebra finches raised in L.A. will never speak like those valley kids.
The key to assessing the internalist stance on religion lies in understanding the relative importance that environment plays in scaffolding religious cognition. Language and bee-navigation (exemplars of internal design) illustrate the salience of genetic resources to cognitive structuring. We may say that such structuring is strongly canalized. Canalized design features are functionally organized phenotypic expressions whose end-states remain stable under a range of normal environmental conditions, (“normal” here defined in virtue of the environment of evolutionary adaptation, the EEA) (Ariew 1999). Selection targets a strongly canalized psychology because it enables agents to rapidly solve adaptive problems by narrowing their search space (Cosmides and Tooby 1992).10
To the extent that selection has targeted open-ended learning capacities in our species, it has done so to enable storage and transmission of abilities and knowledge relevant to understanding and navigating the world. Culture is a warehouse of good tricks, and the reality checks its stock. Yet we need not appeal to the reality of the gods to explain why sensible and rational agents become so passionately committed to them. Religious thought has not evolved to track and engage a supernatural world. Far from a warehouse, religiosity produces castles in the air. Variation in the configuration of religious doctrines, practices, and moral rules make it appear that religious understandings vary widely. Yet because these illusions are designed to solve central and invariant problems of human life, variation must be limited. Not just any religious understanding or act will clear adaptive hurdles. Religion engages fictions, but fictions with tremendous human weight and adaptive importance.11
We shall see that selection has canalized religious psychology to foster rapid solutions to two recurrent adaptive problems: getting along with others, and getting along with ourselves.
I suggest that the prospects for internalism in the study of religion depends on the degree of religiosity’s canalization. For religiosity to be strongly canalized:
(1) Religious cognition must be structured, such that specific information is routinely acquired, stored, organised, and accessed to produce specific cognitive and behavioural outputs.
(2) This structuring must remain stable over a variety of “normal” environmental conditions, that is, conditions closely approximating the EEA (Environment of Evolutionary Adaptation).
(3) We cannot usefully explain structuring by environmental or cultural resources (for example, because there is a poverty of stimulus, or because the relevant computational problems are too complicated to solve by “learning”.)
In the next section, I review research suggesting that core aspects of religiosity are highly structured, that these structures remains stable over a variety of normal environmental conditions—including a variety of cultural conditions—that structure appears to emerge from a poverty of stimulus and on a developmental schedule. Religiosity, like language, flows from a species general biological design.
Numerous research programmes in cognitive psychology have converged to the view that religious concepts heavily rely on implicit folk psychological understandings [for reviews see (Barrett 2000; Atran 2002; Boyer 2003).] These folk psychological understandings govern how we think about physics and biological kinds and other minds. They are highly structured, constant, and surface from a severe poverty of stimulus. For example, before the age of one, children are aware that objects have certain stable properties (e.g. number, mass, substance) without ever explicitly learning this information. By early childhood, children are aware that human and biological kinds have properties that differ from physical objects (Spelke, Phillips et al. 1995). They know that organisms require nutrition, have life spans, and that species properties hold in spite of appearances—for example children know that a cow in horse costume is still a cow, though a table chopped up and made into a chair becomes a chair (Gelman and Markman 1987). Intuitive folk psychology, like language, is largely invariant across cultures (Petrovich forthcoming). Cognitive psychologists maintain that religious concepts are strongly determined by these folk intuitions, which guide our understandings and inferences about the supernatural (Boyer 1992; Boyer 1994; Boyer 1996; Atran 2004).
Consider a common supernatural concept: “ghost”. Probing this concept, we see how heavily it relies on the tacit folk psychological understandings that organise our interpretations of “person”. For example, a ghost will be interpreted to have psychological states—beliefs, desires, memories, attitudes, intentions and plans. No one ever teaches a child that a “ghost” may possess complicated intentions, beliefs, desires, and plans: we just assume this. Some aspects of folk physics and biology are also retained; it is easy to imagine that a ghost can communicate or can be seen at a distance or can move through physical space—perhaps threateningly towards you with its ghost fangs exposed. We think to these unreflective interpretations without ever noting the oddity of inference to physical properties from entities imaged as immaterial. We do not pause to consider how immaterial beings can think with no brains—though we would pause at the idea that a chunk of dirt or atmosphere could think without brains. We assume without inquiry the relevant causal connections. Nor do we do generally wonder how an immaterial spirit tooth—no matter how long and pointy—may injure (for what exactly is long and pointy?). Moreover some understandings are more salient. We think to the view that a ghost is capable of moving through walls more readily than that it may move through the centre of the sun, or a refrigerator. It knows where you are hiding is more salient that it knows you enjoy burritos. We do not learn much beyond a few encyclopaedic details of ghosts: instead folk psychological understandings do the thinking for us, generating rich inferences that allow us to represent and communicate about ghouls (Boyer 2001). Such is true of all supernatural concepts, which rely on the conceptual understandings of the natural concepts they resemble (Boyer 1998; Boyer and Ramble 2001).
Further, though supernatural concepts vary—some rely on intuitions governing persons, others places, and others, powers and forces—they always explicitly violate a few, but not many of these tacit assumptions. Ghosts may be able to move, but unlike us they can move through walls, which is startling. God may be like a man, but unlike a man he has the power to read minds, or create a universe, or punish and reward you for all you’ve done, also startling. Hence, any individual’s religious concepts will retain much of the structure of intuitive psychology. Yet they will also contain a few violations of that implicit knowledge, and this makes religious concepts memorable (Boyer 1994). Boyer hypothesizes: it is precisely because religious concepts violate implicit assumptions for the relevant kinds that religious concepts are interesting and memorable, and thus easily spread in populations. Critically, the violations, though arresting, must be minimal, because religious concepts that are too complex (violate too many expectations) become too difficult to recall (Boyer 2000; Boyer 2001; Atran 2004). For each religious person grows a religiosity that is unique.
The constraints imposed by intuitive psychology on the articulation and scope of religious concepts are, in turns out, extreme, and lead to inferences that, for example, fly in the face of philosophical theologies. Barrett and Keil observe that in spite of subscribing to anti-anthropomorphic theologies, religious persons in India and the United States imagine deities as ordinary persons, different in minimal (but arresting) respects from ordinary persons. Devotees notionally understood and subscribe to theological positions that depict the gods as unlike persons, yet their interpretations belied a powerful anthropomorphising tendency (Barrett and Keil 1998). This tendency holds irrespective of culture and epoch because the intuitive psychology of persons is largely invariant (Knight, Sousa et al. forthcoming).
Thus, on the cognitivist view, religious cognition materializes from a mind already guided by expectations about the natural world it will encounter. Religious concepts violate these intuitive expectations in minimal ways, eliciting our “passion for surprise and wonder” to adopt Hume’s idiom (Hume 1993).12 The panhuman mental architecture of folk psychology heavily structures and constrains religiosity. And the methods appropriate to the study of religious cognition, are just the methods appropriate to the study of the cognitive systems that underwrite it [for recent reviews see (Barrett 2000; Andresen 2001; Atran 2002; Boyer 2003)].
Most cognitive psychologists view religiosity as an after-effect of structured cognition, as cognitive noise. Here the constraints are strong enough on religious cognition to reconsider the role of cultural inheritance mechanisms or “tradition” in composing religious thought (Boyer 1990). Though we acquire a few aspects of our religion from those around us, a panhuman cognitive design configures and guides religious thought.
Adaptationists maintain that the constraints on religious cognition are even more extreme, for religious thought is the product of dedicated cognitive architecture. Adaptationists accept that cognitive methodologies illuminate aspects of religious cognition. But they further hold that we can deepen that understanding by approaching religiosity as functionally organised to promote biological success. For if religiosity has been targeted by selection, we can reverse engineer its design by contemplating the adaptive problems religious thought is designed to solve. Moreover, if dedicated cognitive architecture prompts religiosity, then it is likely even more internally organised and constrained than most cognitive psychologists suspect. For biological success frequently demands exquisite precision and speed, which open-ended problem solving architectures simply cannot deliver (Marr 1980; Cosmides and Tooby 1992).
Recall that cognition fuels success by tracking and responding to the world. Yet I have suggested that 1) religious cognition involves beliefs and practices relative to supernatural beings and 2) we need not suppose these supernatural beings exist. Thus on my minimal naturalistic assumptions, religiosity need not track and register features of a supernatural world, for cognitive agents may come to believe in such worlds though nature is secular. This presents a problem. If religion is an adaptation, then how is responding to things that do not exist helpful? Given religious costs, it seems selection should instead have culled religious tendencies. However adaptationists have demonstrated that religious errors and costs, if systematic, can help in at least two ways: by forging reliable social alliances and by fostering heath-inducing outlooks and practices. I consider these in turn.
One of the keys to biological success in our lineage has been the ability of individuals to forge large and stable alliances among non-kin (Dunbar 1998). In anchoring individual fates to collective fates, and in co-ordinating activity to achieve common ends, individuals are able to promote their interests more effectively than by taking on the world alone. Yet such alliances are notoriously unstable, subject to invasion by agents willing to exploit these advantages without contributing. Social exchange often pays better than it costs. But it is fragile, because cheating social life often pays better still (Skyrms 1996). A society of cheats is no society but a confederation of individuals. Adaptationists urge that a core function of religiosity is to help to make social life possible. For religion fosters co-operation by enabling individuals to reliably predict and secure social exchange where there are rational incentives to defect from cooperation (Cronk 1994; Irons 1996; Sosis 2000; Irons 2001; Atran 2002; Sosis 2003; Sosis and Alcorta 2003; Bulbulia 2004). Religious cognition leads to inaccurate beliefs in gods, but religious expressions accurately lay bare motivational states in a way that enables us to forecast how others will act, and plan. Through religion we do not track gods, we track each other. And the accuracy the system affords solves co-operations problems. Thus, religion functions as a commitment device—a mechanism that binds us together—as the Latin root of the word “religare” (“to bind”) suggests.
To understand the issues, consider an exchange among a pair of unrelated individuals that is modelled as a prisoner’s dilemma. Assume that P is the fitness payoff of non-co-operative individual action (going at it alone). Gall R, the reward for co-operative action (give and take). Let T = the temptation to cheat co-operation (to take but not give), and let S = the payoff of cooperating when an exchange partner cheats (to give but not take). A large class of interactions will conform to the following payoff matrix:
Natural payoff matrix: T (cheating) > R (reciprocity) > P (solitary action) > S (getting cheated)
Because T > R, it pays better to defect rather than co-operate. And because P > S it pays better to defect if the other agent defects. Thus defection strongly dominates co-operation because it pays better regardless of what the other agent does. Defection is what economists call the “Nash equilibrium” for this interaction (Nash 1951; Schelling 1960). Yet observe that co-operation is “strictly efficient”. If both agents opt for fair exchange, both are better off than by going at the world alone (Skyrms 1996).
Notice that specifically structured religious understandings may solve co-operative problems by altering the expected returns of co-operative action. Whatever the actual payoff matrix, if agents believe that co-operation always pays better than defection, they will be rationally motivated to exchange. Suppose the reward for a cooperative response is eternal bliss and the punishment for defection is reincarnation as a sweat gland. Such belief is at odds with nature, for it inaccurately depicts probable outcomes. The relevant causal processes are supernatural. Yet precisely because cognition errs in just this way, it motivates cooperative exchange.
Perceived supernatural payoff matrix: R > P > S > T
Selection may equip agents with dispositions to form such moralizing illusions, though crucially, only where it is possible to reliably recognize the presence of these illusions in others. The recognition constraint is critical to the evolution of religious cognition. For religion to foster exchange, agents must be able to reliably discern religious commitments in their partners. Yet it is always in a defector’s interest to state a religious commitment, only to defect when conditions allow. Defectors will have incentives to pollute the epistemic environment of religious co-operators with signals of authentic religious commitment, only to defect when the opportunity presents itself. Hence religionists must have a reliable safeguard from pretend co-religionists. Finding a safeguard is difficult, given the advantages to bypassing it.
Behavioural ecologists have shown that selection may outfit organisms with signalling and detection systems capable of producing and assessing co-operative commitment or worth. Frequently, such systems rely on costly expressions or “handicaps” that target the relevant property. An authentic signal will be arrayed so that only a signaller that genuinely possesses the relevant property (in this case, the presence of a motivating religious belief) will be able to afford producing the signal. Such costly signals of worth are superabundant in nature (Zahavi 1980). For example the songs of male bulbuls (Pycnonotidae) are carefully arranged to signal strength and reproductive fitness, and so are capable of warding off potential rivals and of attracting mates. Only fit birds can manage such ornamental song. Similarly, the whistling of skylarks when chased by predatory merlins is a costly signal of aerobic fitness, one that generally forestalls the chase. Skylarks able to sing while avoiding beak and talon are unlikely to be caught, so merlins don’t generally chase them (Zahavi and Zahavi 1997). The vigorous leaping or “stotting” of Springbok gazelle authenticates their ability to outrun predatorial lions, also forestalling a pointless chase (FitzGibbon and Fanshawe 1988). The costumes and ornaments of real estate agents are configured to signal success, etc.
In our lineage, costly religious signalling has evolved as a means to authenticating the relevant religious beliefs. First, emotional displays of religious commitment—the thralls of devotion and piety—accurately convey a religious lifeway. It is notoriously difficult to fake emotions because they are not under conscious control. We can easily lie with words but not with our faces (Frank 1988; Ramachandran and Blakeslee 1998). Religious emotions express specific altruistic motivations and religious beliefs. That is, emotional signals communicate a behavioural trajectory. Weeping in prayer before LORD FOZ suggests commitment to a supernatural scheme in which LORD FOZ relates to our lives. Doing the same before LORD CHOP suggests otherwise. Moreover such displays modulate in ways that enable an audience to discern strength of commitment. Religious emotions measure a commitment’s intensity. We can thus discern the strength of a religious commitment from the power of an emotional display. And this information can be factored into an assessment of future interactions. As defection incentives rise, the strength of an emotion sufficient to secure exchange will also rise. Knowing the bounty, Bin Laden will search the religious expressions of his cohort [For discussion see (Sosis 2003; Sosis and Alcorta 2003; Bulbulia 2004; Sosis 2004)].
Second, religious ritual provides a forum in which religious emotions, and other costly signals, can be produced, evaluated and inculcated. Though private rituals may function to reaffirm religious commitment, such displays are wasted signals.13 Sosis and colleagues have shown that ritual participation is an excellent predictor of altruistic exchange with non-kin (Sosis and Ruffle 2003; Sosis and Ruffle 2004). Religious rituals also correlates positively with communal longevity (Sosis 2000; Sosis and Bressler 2003). As with emotional display, ritual displays communicate concrete intentions and plans, enabling audiences to plot a behavioural trajectory. Sacrificing a lamb in the ritual of FOZ conveys a different commitment in the ritual of CHOP. Further ritual costs modulate to assess the intensity of commitment. As the relative expense of a ritual display rises so too does the relative degree of commitment it assesses (Sosis 2003; Bulbulia 2004). The strongly committed will be willing to spend more on ritual offerings, because they expect future supernatural returns to repay the expense. For religious agents perceive rituals as investments repaid by the gods. And because defectors perceive rituals as resource waste, they have disincentives to opt out of these costs. Moreover because they elicit powerful religious emotions, dramatic religious rituals provide experiential evidence for the supernatural understandings that bind a group together. Thus religious rituals not only afford opportunities to assess religious conviction, they instil them (Sosis 2003; Alcorta and Sosis under review).
The adaptationist approach to costly religious signalling leads to a few surprising outcomes. When facing crisis, religious agents will tend to devote more time and material resources to ritual participation than when secure. This prediction that has recently received empirical support, see (Chen 2004). Such outlays may seem puzzling. With adversity we would ordinarily predict tendencies to conserve. Yet from an adaptationist perspective the importance of group resources to individual well-being escalate during crisis. And so too do defection incentives. Hence for co-operation to be reliable, commitment costs need to be amplified. [For overviews of the adaptationist literature see (Sosis and Alcorta 2003; Bulbulia in press)].
“The space between skins is a no-man’s land, controlled by no-one; designed for no-one; littered with the detritus of biological struggle” (Sterelny forthcoming). Religiosity has evolved to tame aspects of that external world, fostering exchange by restraining intra-group competition and promoting self-sacrifice. Yet religion may also do work within our skins. For though not contested in the same sense, an agent’s internal psychological states need to be integrated and co-ordinated for it to flourish. We must make peace not just with the world, but also with ourselves. And religiosity may mend our minds as well, in its function as a mental harmonizer.
Making only conservative assumptions about ancestral conditions, we can appreciate how supernatural commitments foster psychological and physical well beyond reciprocity. Consider a certain idealisation, which I will call “traumatic world”: agents regularly face situations where natural probabilities, if assessed accurately, paint bleak reproductive outlooks (infection and illness, childhood mortality and other family death, loss of body parts and disfigurement, dental issues, predator culling, shipwreck, famine and war, etc.) It is reasonable to assume that the ancestral world was to a substantial degree a traumatic world. Suppose further, that accurate estimations of the traumatic world have damaging side effects, which come by way of various psychological and physical stressors. For example understanding one’s odds of surviving a shipwreck may support inferences to the view: “I am almost certainly going to drown”. Given the rest of our psychological make-up, such an inference may present an agent with overwhelming grief, anxiety, or fear. And we know that stress is physiologically and psychologically damaging. (It is hard to maintain a state of wellbeing with a knife held to one’s neck) [for and overview see (Kompier and Cooper 1999).] Averting this damage requires a refinement of response to the traumatic world. Optimising our responses, selection will restrain and dampen the effects of bad news. But it is reasonable to assume that there are few direct evolutionary pathways to dampening. If our loves drive us, then grief and anxiety at love’s loss may be fairly inexorable. For a motivation easily overturned is hardly a motivation.
Yet specific varieties of religiosity may facilitate solutions, through the inaccurate conceptions of the world supernatural understandings afford. Directed supernatural distortions and biases may synchronise internal psychological states and physiological responses, providing religious agents with integrated, meaningful, and broadly optimistic understandings of the world and their prospects in it. Atran and colleagues have shown that supernatural understandings become salient to individuals confronted with stories of death. Though not an adaptationist, Atran suggests that “emotionally eruptive existential anxieties motivate supernatural beliefs” (Atran 2004). Supernaturalisms seem to help us to endure the foxholes of life.
Clearly, misconstruing the traumatic world as not hopelessly bad may reduce stress, fear, and malaise. Here salubrious religious illusions shield agents from the slings and arrows of existence by altering damaging (though accurate) assessments of the world. In this religiosity not only binds us to each other, but also to ourselves. [For a comprehensive surveys of the religion and health literature see (Pargament 1997; George, Ellison et al. 2002)]. Though the psycho-causal connections that facilitate health through religion remain somewhat unclear (Shapiro and Shapiro 1997), assuming these connections, we can compare the fitness effects of accurate judgement to inaccurate optimistic judgment, and consider the downstream biological consequences of each.
Accuracy v. supernatural illusion
| Perception | Response | Outcome |
|---|---|---|
| Traumatic World | Stress, Dread, Fear, Malaise, Anxiety, Despair |
Poor health |
| Enchanted World | Optimism | Good health |
Given these benefits, it is easy to see how religious agents may slog up adaptive hills through very specific supernatural understandings of their prospects: belief in gods that make the world a better place.14
However so, religious distortions will need to be highly encapsulated. In most cases, there is nothing more lethal to an organism than misunderstanding its actual circumstances and prospects. Selection may work with these illusions, but in relative isolation from survival. Religion is like a surgeon’s scalpel: effective only when precisely guided, lethal otherwise. Lost at sea, we may believe that the gods will save us—and such inspiring conviction may benefit us—but we must still swim.15 We may believe the gods protect our children, our genetic legacy hinges on restraining our children from the jungle. We may believe that God is just and all-powerful, but to survive we must still visit harm to infidels who threaten us.
Researchers have only recently begun to examine the fitness effects of religious illusions outside of social exchange, so further adaptive functionality beyond co-operative domains remains somewhat conjectural. Yet if supernatural illusions and motivations sustain hope and meaning, provide comfort and solace (Atran 2002; Atran 2004), and even facilitate healing (McClenon 2002), and is encapsulated (Bulbulia 2004) (plausible assumptions) then it seems selection will amplify and entrench supernatural dispositions.
The story of evolutionary research on religiosity over the past decade is a story of increasing awareness of how biological constraint produces exquisite refinement in panhuman psychological design. Religiosity furnishes mental gear for grappling with the world. To facilitate success religious cognition must be delicately structured. Reverse engineering religions has produced the following crude picture of the common architectural features of the religious mind.
1) Conviction Generator. The cognitive architecture that employs religious concepts produces ontological commitment to non-existent agencies and powers. Such commitments distort expected outcomes, to motive predictable responses and behaviours. Local culture environments supply names to these agencies and powers, along with some encyclopaedic information. But the functions these commitments perform remain stable. Notice, the concepts “Dracula” and “Bugs Bunny” violate intuitive expectations governing the relevant domains (Slavic noblemen and rabbits, respectively). And as cognitive psychologists observe, these concepts produce memory effects. Observe however that few adults believe these concepts apply to reality. Yet gods and supernatural powers frequently become the objects of sincere and deep conviction, expressed as certainty. Such convictions differ from a mere memory effects [for discussion see (Atran 2002).] Our minds appear designed to respond to our cultural environments by developing the relevant supernatural commitments. We are predisposed to embrace the spirituality of our social worlds (Bering in press).
2) Experience. Religiosity involves inferences from a purely secular world to a robust supernatural world. To accomplish this task, there must be an active biasing and distorting of information to produce and reinforce these inferences. A psychological design that produces supernatural commitment will contain structures that modulate perception to produce supernatural belief-affirming experiences. Psychologists observe that there is the tendency to experience the world as animated by supernatural beings and powers and to interpret random events as meaningful (Heider and Simmel 1944; Guthrie 1993; Scholl and Tremoulet 2000; Guthrie 2001; Bering 2003). Such tendencies are well explained in light of a dedicated religious experience function: a mechanism that processes the world as containing supernatural elements.
3) Love and condemnation: strategic response plasticity. Supernatural concepts are configured to support co-operation in prisoner’s dilemmas. Such concepts are thus further constrained to possess a moral aspect. The gods and godly powers take an interest in the morality and piety of mortals, to harm or reward agents relative to their behaviour. In this way, social action is related to “just” desserts through supernatural policing. I place “just” in shutter quotes to denote an inherent relativity in religious morality, for the morality of these concepts tends to be localized to the relevant exchange group (and to the gods of the relevant exchange group). For if the gods were to sanction exchange with defectors, then selection would quickly replace religious inclinations with a cold economic calculus. Religious morality is arbitrary but not random (Wilson 2002). As a corollary to intra-communal policing, god concepts may be arrayed to favour inter-group asociality or anti-sociality. With respect to those outside the community, moralizing gods may well enjoin indifference, or where these outsiders compete for resources, condemnation or violence. Notice, an optimal design would enable agents to adjust strategies to local situations. What may benefit agents in one setting may prove disastrous in another. Thinking of Lord CHOP’S absolute justice will not provide the inspiration helpful to the stowaway (unjustly, by her lights) stranded at sea. Such a Caruso needs to focus on CHOP’S loving power to assist the faithful in need. So the expectation is for response plasticity targeted to strategic settings. Notice however that the relevant information to decide how to act is often not available, or obscure. “Do we convert, or ignore or castigate these cannibals?” With respect to intra-group settings, we can predict that gods will generally enforce pro-social morality. However the epistemological waters are muddied with respect to out-groups. Where conversion clearly builds stronger coalitions the expectation is for god concepts enjoining missions to out-groups. Where out-groups are perceived to pose genuine threats, adaptationists predict religiously sanctioned caution or violence.
4) Illusions of difference. The similarity underlying language is not obvious largely because pronunciations vary. With respect to religious similarity, it may be strategically useful for consciousness to actively enhance the illusion of difference with moralistic overtones. Where religious competition is acute, an optimal psychological architecture will minimize and distort religious similarity to favour strong religious branding. for holding that all religions are essentially the same stands in tension with the strategy, central to moralizing religion, that one’s religion aligns to one’s coalition and exchange group. An optimal design will modulate the perception of differences to map strategically to perceived exchange outcomes, with a distortion and biasing of information flow frequently obscuring religious similarity, even obvious similarity, as in the case of Northern Irish or Yugoslavian Christianities [compare (Hardin 1995)].
5) Placebo effects. Religious architecture may bias and distort information flow in ways that do not immediately relate to exchange concerns. For example, I have sketched an adaptive rational for a design that promotes distortions of hope, solace, meaning, and other psychological states that correlate to physical and mental well-being. It is easy to see how selection could forge channels integrating the conceptual materials of religion with life enhancing understandings of reality and forecasts, to fabricate judgements and attitudes that fuel success. For in actively distorting and misunderstanding the outlay of reality—incorrectly assessing the causal powers of a placebo, finding meaning in a random occurrence, or solace in the face of despair or horror—an organism’s prospects may rise over capturing the world as it is, or is likely to be (the effect of which may trigger debilitating despair, or other damage). Religiosity appears to function a mental harmonizer with flow on effects to physical health.
6) Informational encapsulation. Critically, whatever the functions performed by supernatural belief, religiosity must be informationally encapsulated to only those functional domains it serves. Importantly, with regard to maladaptive behaviour, religiosity tends to be practically inert. Religious cognition rarely thwarts life. Though the religious believe as if the gods reward and punish, they do not generally alter their behaviour to reflect that commitment over the basic (non-coalitional) circumstances of life. Religiosity does not generally hinder the devout from building shelters, gathering nourishment, and battling enemies (Bulbulia 2004). Religion is thus similar to other forms of self-deception, where reproductive value comes from a systematically distorted representation of oneself and one’s world [see (Trivers 2001)].
7) Costly display. Where religious costs do appear—the sacrifice of material and opportunity resources on behalf of supernatural agents, strong emotional displays, the undertaking of risky projects, bodily mutilation, scripture study, fasting and so forth—these costs often figure as public signals of authentic commitment to binding religious ideologies. Such costs are authenticating mechanisms that facilitate the solution of prisoner’s dilemmas. And they modulate to circumstances: cost outlays rise as religious groups face crisis (Sosis 2003; Chen 2004). Non-signalling costs are minimal, and generally relate to the education of religious emotion (private rituals)—investments that maintain the overall functionality of the relevant cognitive systems.
8) Developmental patterns. There appears to be a structured, internally driven developmental schedule leading to mature religious cognition. Young children readily produce supernatural understandings of the world (Barrett and Richert 2003), believing the universe around them to be inhabited by god-like creatures (Kelemen in press). Children exhibit a bias for intentional accounts of the origins of nature, a bias that strikingly holds regardless of the religiosity of their parents (Evans 2000; Evans 2001). Thus the children of atheists tend to theism, suggesting their early religiosity comes from within. Distinctively moralizing versions of invisible agents appear by the age of 2 1/2, and children seem to regulate their moral behaviour in virtue of ontological commitment to them (Bering in press). This ambient religiosity of childhood takes on strongly affective contours by late childhood and early adolescence, a time when emotional authentication of religious commitment is critical to forging reliable social exchange [for extensive discussion of the underlying neurological basis of adolescent religion, see (Alcorta and Sosis under review)]. While developmental details remain somewhat obscure, the patterns appear to hold irrespective of culture (Bering (in press); Knight, Sousa et al. forthcoming). From a severe poverty of stimulus, children grow to interpret their world as holding supernatural powers who are interested in them and what they do—in particular how they treat others. By early adolescence they display emotional responses to these beings (Francis and Kay 1995), and organise their activities in virtue of the supernatural commitments these emotions certify [for reviews of the developmental literature see (Bulbulia in press; Kelemen in press)].
In sum, the results of adaptationist inquires appear to vindicate the internalist stance on religion. For it appears that religious cognition is:
(1) is heavily structured so that supernatural conceptual information is acquired, stored, and accessed in ways linked to specific cognitive and behavioural outputs. The psychological scaffolding of religion comes mainly from biological, not cultural endowment.
(2) remains stable over a variety of “normal” environmental conditions (conditions closely approximating the EEA). This structural stability follows from the problems religiosity has been designed to solve: only functionally specific and informationally encapsulated distortions of reality, and functionally specific hard-to-fake emotions and costly behaviours will foster the adaptive understandings and motivations religiosity provides
(3) materializes through an internally directed developmental programme: the “intuitive theism” of childhood grows through adolescence (in suitably religious contexts) to motivate reciprocity and promote physical and mental well-being. From fragmented cultural sources, religious commitment springs to a rich and intricate shape, in ways that equip religious agents to flourish.
It therefore appears that religion is strongly canalized; Religious variation operates within the narrow constraints of a psychological design optimized for biological success in the ancestral world.
Time to summarize the argument. There is undeniable variation in religious thought and expression. This variation flows from cultural circumstances and processes. Whether an individual worships a god named “FOZ” or “CHOP” depends on the exigencies of local history and institutional transmission. So too are the elemental features of ritual: whether to walk fires for FOZ or prostrate before icons of CHOP; what to wear to an inquisition; which melodies to sing before waging war; how to read the future from a sheep’s entrails, and others. And at least some religiously motivated moral strategies appear to vary by convention. Many Aztecs worshipped gods imagined to require human blood sacrifice. California Zen Buddhists perceive different demands. Their religious brief is roughly: “sit and think nothing and do nothing”. The morality of inclusiveness varies along institutional lines: Chinese Buddhists and Chinese Christians differ markedly in how they treat out-groups (Hansen and Norenzayan 2004). More basic moral differences seem to flow directly from religious understandings: some are inspired by their religion to pacifism, others to bellicosity (Atran 2003; Atran 2004). With respect to the conventional character of religion, internalist inquiry has limited predictive value (as it does in the study of language). A complete psychological understanding of panhuman religious architecture could not have predicted a group would call their god “Jehovah”. May the scholars of local detail continue to illuminate such peculiarities.
Nevertheless current evolutionary and cognitive research suggests that peculiarity operates within the constraints of a common religious psychological design. For religiosity appears to flow from dedicated cognitive structures configured to form precise non-natural understandings that motivate intricate adaptive responses to ancestral conditions. Are there any religions? In fact there are innumerably many. For each religious person grows a religiosity that is unique. Yet religious differences among individuals and groups, though striking, are largely conventional, owing to differences in the configuration of supernatural concepts fueling the system, variation in the doctrinal and ritual proscriptions, and the perceived strategic exigencies of local happenstance. Religious development consists largely in fixing labels to pre-existing cognitive structures, and applying these understandings to local settings. Settings vary, so religious responses vary, but such is true of the most canalized capacities. (For example, settings vary, so visual responses vary, but we don’t “learn” to see.)
Of course religious labeling is generally important to religious agents themselves. For it is through these labels (and associated signaling practices) that religious persons make sense of their world, understanding whom to trust and how to organize their lives. Though Hindu and Islamic gods and rituals are functionally similar, recent Indonesian history suggests that only Islamic understandings will coordinate and inspire religious agents on the ground there. As with language, religiosity is an aspect of our sociality. Slight variation in individual religiosities can have profound practical consequences in local strategic settings. Worshipping Mother Earth among the Taliban may prompt a stoning. Nevertheless, although the names of gods and the specific costs and elaborations of rituals obviously vary, these relatively superficial differences obscure vast commonalities in the cognitive undergirding that supports the religious mind. Cultural inheritance explains some conventional attributes of an individual’s religious outlook and practice, but biological endowment accounts for the massive cognitive structuring necessary for religion to solve adaptive problems. Without this psychological architecture, the conventions that regulate and control religiously motivated thought and exchange would not be possible: without the support of internally directed cognition, all sacred canopies fall. To comprehend this panhuman psychological architecture, we must look past a thicket of striking cultural variation to universal structures that buttress and facilitate religiosity in our species. To be sure, use of label “Jehovah” was never predictable. Yet to understand why religious perspectives employing this concept make sense to so many otherwise sensible people (instead of infallibly triggering incredulous stares and slammed doors) we need to explore the elaborate design of the Pleistocene adapted mind that produces religiosity. Through current naturalistic inquiry, this fascinating and powerful dimension of human nature is becoming less mysterious.
Religious Studies
Victoria University
P.O. Box 600
Wellington
New Zealand
Alcorta, G. and R. Sosis (under review). Religion, emotion, and symbolic ritual: The evolution of an adaptive complex.
Andresen, J., Ed. (2001). Religion in Mind. Cambridge: Cambridge University Press.
Ariew, A. (1999). Innateness is Canalization: In Defence of a Developmental Account of Innateness. Where Biology Meets Psychology: Philosophical Essays. V. G. Hardcastle (ed.). Cambridge: MIT Press.
Atran (2002). In Gods We Trust: the evolutionary landscape of religion. New York: Oxford University Press.
Atran, S. (2003). Genesis of suicide terrorism. Science 299: 1534–1539.
— (2004). Combating Al Qaeda’s splinters: Mishandling suicide terrorism. The Washington Quarterly 21, 67–90.
— (2004, in press). Religion’s evolutionary landscape: Counterintuition, commitment, compassion, communion. Behavioral and Brain Sciences.
Barrett, J. L. (2000). Exploring the natural foundations of religion. Trends in Cognitive Sciences 4 (1): January.
Barrett, J. L. and F. C. Keil (1998). Cognitive constraints on Hindu concepts of the divine. Journal for the Scientific Study of Religion 37: 608–619.
Barrett, J. L. and R. A. Richert (2003). Anthropomorphism or preparedness? Exploring children’s god concepts. Review of Religious Research 44 (3): 300–312.
Bering, J. (2003). Towards a cognitive theory of existential meaning. New Ideas in Psychology 21: 101–120.
— (in press). The evolutionary history of an illusion: Religious causal beliefs in children and adults. In B. Elllis and D. F. Bjorklund (eds). Origins of the social mind: Evolutionary psychology and child development. New York: Guilford Press.
Bigrami, A. (1992). Belief and meaning. Oxford: Blackwell.
Boyer, P. (1990). Tradition as Truth and Communication: a cognitive description of traditional discourse. Cambridge: Cambridge University Press.
— (1992). Explaining religious ideas: Elements of a cognitive approach. Numen XXXIX (1): 27–57.
— (1994). The naturalness of religious ideas: a cognitive theory of religion. Berkeley: University of California Press.
— (1996). What makes anthropomorphism natural: intuitive ontology and cultural representations. Journal of the Royal Anthropological Institute (2): 1–15.
— (1998). Cognitive aspects of religious ontologies: how brain processes constrain religious concepts. Theories and Method in the Study of Religion. T. Alhback, Donner Institute.
— (2000). Functional origins of religious concepts: Ontological and strategic selection in evolved minds. Journal of the Royal Anthropological Institute 6: 195–214.
— (2001). Religion Explained: the evolutionary origins of religious thought. New York: Basic Books.
— (2003). Religious thought and behaviour as by-products of brain function. Trends in Cognitive Sciences 1 (3): 119–124.
— and J. Ramble (2001). Cognitive templates for religious concepts: cross-cultural evidence for recall of counter-intuitive representations. Cognitive Science 25: 535–564.
Bulbulia, J. (2004). Religious costs as adaptations that signal altruistic intention. Evolution and Cognition 10 (1).
— (in press). Area review: The cognitive and evolutionary psychology of religion. Biology and Philosophy.
Chen, D. (2004). Economic distress and religious intensity: Evidence from Islamic resurgence during the Indonesian financial crisis. Under review.
Chomsky, N. (2000). New horizons in the study of language and mind. New York: Cambridge University Press.
Cosmides, L. and J. Tooby (1992). The psychological foundations of culture. In J. H. Barkow, L. Cosmides and J. Tooby (eds). The adapted mind: Evolutionary psychology and the generation of culture. New York: Oxford University Press.
Cronk, L. (1994). Evolutionary theories of morality and the manipulative use of signals, Zygon 29 (1): 81–101.
Dehaene, S. (2003). Natural born readers. New Scientist 179 (2402): 30.
Dummett, M. (1986). A nice derangement of epitaphs: Some comments on Davidson and Hacking. In E. Lepore (ed.). Truth and Interpretation. Oxford: Blackwell: 459–476.
Dunbar, R. I. (1998). The social brain hypothesis. Evolutionary Anthropology 6: 178–190.
Evans, E. M. (2000). The emergence of beliefs about the origin of species in school-age children. Merrill Palmer Quarterly 46: 221–254.
— (2001). Cognitive and contextual factors in the emergence of diverse belief systems: Creation versus evolution. Cognitive Psychology 42: 217–266.
FitzGibbon, C. D. and J. H. Fanshawe (1988). Stottmg in Thompson’s gazelle: An honest signal of condition. Behavioral Ecology and Sociobiology 23: 69–74.
Francis, L. J. and W. K. Kay (1995). Teenage Religion and Values. Leominster: Gracewing.
Frank, R. (1988). Passions Within Reason: The Strategic Role of The Emotions. New York, Norton and Company.
Gelman, S. A. and E. Markman (1987). Young children’s inductions from natural kinds: The role of categories and appearances. Child Development 58: 1532–1540.
George, L. K., C. G. Ellison, et al. (2002). Explaining the relationships between religious involvement and health. Psychological Inquiry 13 (3): 190–200.
Gould, J. L. (2002). Can honey bees create cognitive maps? In M. Bekoli, C. Allen and G. Burghardt (eds). The Cognitive Animal: Empirical and Theoretical Perspectives on Animal Cognition. Cambridge: MIT Press.
Guthrie, S. (1993). Faces in the clouds: a new theory of religion. New York: Oxford University Press.
— (2001). Why gods? A cognitive theory. In J. Andresen (ed.). Religion in Mind. Cambridge, Cambridge University Press.
Hansen, I. and A. Norenzayan (forthcoming). Religious devotion, religious exclusivity, and tolerance for religious outsiders.
Hardin, R. (1995). One for all: The logic of group conflict. Princeton: Princeton University Press.
Harman, G. (1998). Moral philosophy and linguistics. In K. Brinkmann (ed.). Proceedings of the twentieth world congress: Volume I: Ethics. Bowling Green: Philosophical Documentation Center: 107–115.
Heider, F. and M. Simmel (1944). An experimental study of apparent behavior. American Journal of Psychology 57: 243–249.
Henrich, J. and R. McElreath (2003). The evolution of cultural evolution. Evolutionary Anthropology 12: 123–135.
Hume, D. (1993). Of miracles. David Hume: Writings on Religion. A. Flew. La Salle: Open Court: 63–88.
Irons, W. (1996). Morality, religion, and evolution. In W. W. W.M. Richardson (ed). Religion and science: history, method, and dialogue. New York: Routledge: 375–399.
— (2001). Religion as hard-to-fake sign of commitment. In R. Nesse (ed). Evolution and the Capacity for Commitment. New York: Russell Sage Foundation.
Kelemen, D. (in press). Are children “Intuitive Theists”? Reasoning about purpose and design in nature. Psychological Science.
Knight, N., P. Sousa, et al. (forthcoming). Children’s attributions of beliefs to humans and god: Cross-cultural evidence.
Kompier, M. and C. Cooper (1999). Preventing Stress, Improving Productivity. New York: Routledge.
Lawson, E. T. and R. N. McCauley (1990). Rethinking religion: Connecting cognition and culture. Cambridge: Cambridge University Press.
Marr, D. (1980). Vision. New York: Freeman and Company.
McCauley, R. N. and E. T. Lawson (2002). Bringing Ritual to Mind. New York: Cambridge University Press.
McClenon, J. (2002). Wondrous Healing: Shamanism, Human Evolution, and the Origin of Religion. DeKalb, Illinois: Northern Illinois University Press.
Nash, J. (1951). Noncooperative games. Annals of Mathematics 54: 289–295.
Odling-Smee, J., K. Laland, et al. (2003). Niche Construction: The Neglected Process in Evolution. Princeton: Princeton University Press.
Pargament, K. I. (1997). The Psychology of Religion and Coping: Theory, Research, Practice. New York: Guilford.
Petrovich, O. (forthcoming). The child’s theory of the world.
Pinker, S. (1994). The language Instinct. New York: HarperPerennial.
— (1999). Words and rules: The ingredients of language. New York: Basic Books.
Pinker, S. and P. Bloom (1990). Natural language and natural selection. Behavioral and Brain Sciences 13: 707–784.
Ramachandran, V. S. and S. Blakeslee (1998). Phantoms in the brain: Probing the mysteries of the human mind. New York: Quill William Morrow.
Schelling, T. (1960). The Strategy of Conflict. New York: Oxford University Press.
Scholl, B. J. and P. D. Tremoulet (2000). Perceptual causality and animacy. Trends in Cognitive Sciences 4: 299–309.
Shapiro, A. and E. Shapiro (1997). The placebo: Is it much ado about nothing? In A. Harrington (ed.). The Placebo Effect. Cambridge: Harvard University Press: 12–36.
Skyrms, B. (1996). Evolution of the social contract. New York, Cambridge University Press.
Sosis, R. (2000). Religion and intragroup cooperation: Preliminary results of a comparative analysis of Utopian communities. Cross-Cultural Research 34 (1): 77–88.
— (2003). Why arent we all Hutterites?’ Human Nature 14 (2): 91–127.
— (2004). The adaptive value of religious ritual. The American Scientist 92: 166–172.
Sosis, R. and G. Alcorta (2003). Signalling, solidarity, and the sacred: The evolution of religious behavior. Evolutionary Anthropology 12: 264–274.
Sosis, R. and E. Bressler (2003). Co-operation and commune longevity: a test of the costly signalling theory of religion. Cross-Cultural Research 37 (2): 11–39.
Sosis, R. and B. Ruffle (2003). Religious ritual and cooperation: Testing for a relationship on Israeli religious and secular Kibbutzim. Current Anthropology 44(5): 713–722.
— (2004). Ideology, religion, and the evolution of cooperation: Field experiments on Israeli Kibbutzim. Research in Economic Anthropology.
Spelke, E., A. Phillips, et al. (1995). Infants knowledge of object motion and human action. In D. Sperber, D. Premack and A. Premack (eds). Causal Cognition. Oxford: Clarendon Press.
Sterelny, K. (2003). Though in a Hostile World: The Evolution of Human Cognition. Oxford: Blackwell.
— (forthcoming). Made by each other: organisms and their environment. In J. Odling-Smee, K. Laland and M. Feldman (eds). Niche Construction: The Neglected Process in Evolution in Biology and Philosophy.
Trivers, R. (2001). Self-deception in service of deceit. In R. Trivers (ed.). Natural Selection and Social Theory. New York: Oxford University Press.
Whitehouse, H. (2000). Arguments and icons. Oxford: Oxford University Press.
Wilson, D. S. (2002). Darwin’s Cathedral: Evolution, Religion, and the Nature of Society. Chicago: University of Chicago Press.
Zahavi, A. (1980). Ritualization and the evolution of movement signals. Behaviour 72: 77–81.
Zahavi, A. and A. Zahavi (1997). The Handicap Principle: A missing piece of Darwin’s Puzzle. New York: Oxford University Press.
_______________
1Acknowledgements: Many thanks to Gandace Alcorta and Rich Sosis for excellent comments on an earlier draft.
2By “religion” and its cognates I will mean, beliefs and practices relative to supernatural beings, places, and powers: “gods” “spirits” and “heavens” and the like. The theoretical motivation for this use of “supernatural” as a distinctive psychological kind comes from Boyer’s work, discussed below (Boyer and Ramble 2001). Often I use “religiosity” to capture this intention. I am therefore interested in religion as a psychological phenomenon, a dimension of how many think and act.
3For example “French” is the tool used by communities in France, Corsica, Quebec, parts of Switzerland, West Africa and elsewhere (expensive restaurants, certain Gajun precincts, among poseurs, and so forth).
4For an example of this view, see (Dummett 1986).
5See also (Bigrami 1992).
6By “mental organs” cognitive psychologists do not mean bits of brain matter. Rather, they look at the brain as an integrated architecture of information processing modules—roughly a set of programs that supervene on the material brain, much as computer programs supervene electrical hardware. “Mental organ” describes these programs, not the underlying neurology that enables them. The classic “levels” discussion is set forth in (Marr 1980).
7It would be interesting to study the interface between the language faculty and the systems that control supernatural cognition. For at a minimum, we use language to express and understand supernatural commitments and practices. Thus, there must be legibility conditions at the interface between the systems that control each domain. I do not seek to explore these conditions in any detail here. Nor do I maintain that the faculties governing religious thought and development resemble those of the language faculty in anything but superficial ways. Such is true of ordinary organs: the liver’s development and design differs in critical respects from that of teeth or sexual organs. Hence, it would be striking were the developmental parallels between language and religion to prove substantial. [For a contrasting stance see (Lawson and McCauley 1990)].
8Notice, neither did Gere’s English, so we must take care in externalist generalizations.
9There are controversies in the details of this view, but not about its basic outlines.
10Two points to bear in mind. First, structure does not imply determinism: canalized structure may specify a range of behavior outputs, given one or several informational cues and triggers, none of which is exclusively determined in advance by genotype. Detecting a predator an organism may fight, flee, monitor, go about business as usual but under a higher state of alert, none of which is automatically determined by the predator cue (Sterelny 2003). Second, canalization clearly admits of degree.
11Nothing I say here should be taken as a wholesale rejection of the religious dimension of human nature. At the core of most loves (and hatreds) lies the worm of bias and distortion, but one that spins the fabric of our entire existence. Some religious outlooks and practices are morally vicious, but not all (probably not most) and in my view we should only reject the vicious. We should no more discard good religion than we should good parenting or good marriages, which frequently rely on the illusion that our own children or spouses are profoundly special (they are not).
12For reviews of constraints on ritual actions see: (Lawson and McCauley 1990; Whitehouse 2000; McCauley and Lawson 2002).
13Sosis argues that private rituals “appear to be critical in keeping out potential free-riders by raising the overall costs of ritual performance. Since a skeptic will not perform private acts of devotion (because they are not observed) the overall costs of ritual behaviors imposed on a community must be significantly higher to keep out skeptics than if a community did not impose private demands on their members” (per. comm.) See also (Sosis 2003). Maybe so, but if a community has access to information about “private” ritual, then such rituals are not private.
14Here again we need not assume supernatural reality to explain commitment to supernatural reality.
15This idealisation trades on the beneficial effects of optimistic belief. But clearly selection may favor other forms of misconstrual, even pessimistic outlooks, where these on average enhance life and reproduction. Though again, whatever the error vector, it will need to be encapsulated, lest the remedy turn to poison.