Skip to the content | Change text size

Eras Journal - Bulbulia, J: Review of "Religion Explained: The Evolutionary Origins of Religious Thought", Pascal Boyer

'Unweaving the Religious Mind': A Review of Pascal Boyer,Religion Explained: The Evolutionary Origins of Religious Thought

Basic Books, New York, 2001.

Isbn 0 465 00696 5

Pascal Boyer is one of the leading figures in the cognitive psychology of religion - a field of inquiry less than fifteen years old, but one which has already shed promising new light on the psychological processes that underlie religious thought and practice. Religion Explained is an accessible book that gathers the latest research in the field, outlines its achievements and looks to intriguing new prospects for discovery on the horizon of inquiry.

In spite of the book's ambitious title, the moral of Boyer's story is that there is no simple explanation for religious consciousness. Some aspects relate to inference systems in the brain (how we think about causality and ontology), others relate to features of our social awareness (how we represent and interact with persons), and still others relate to how memories are processed and stored. Because 'religious thought' does not describe a single cognitive dimension there can be no straightforward psychological account of what non-specialists refer to with that expression. Given this approach, 'Religion Explained' is misleading advertising. Clearly, if there is no such thing as religion then religion cannot be explained. Rather, aspects of what we in ordinary language call 'religion' can be explained. The point is subtle but illustrates a crucial turn in the psychology of religion where the most fruitful and promising research programs have abandoned the search for a global explanation of religion, focusing instead on more narrowly circumscribed cognitive domains. Marketing aside, however, there is enough interesting psychological detail in Boyer's account of some of the cognitive systems that give rise to aspects of religious consciousness to suggest that this is an important book that deserves careful reading.

Boyer is mainly concerned with explaining the various psychological processes involved in the acquisition and transmission of ideas concerning the gods, including ideas regarding what is owed to them and regulations governing how we are to behave toward each other in light of their existence. Clearly many religious persons acquire religious ideas and practices through social exposure. The child of a Zen Buddhist will not become an evangelical Christian or a Zulu warrior without the relevant cultural experience. While mere exposure does not cause a particular religious outlook - Jane may have been raised a Roman Catholic but leave the church - nevertheless some exposure seems required - Jane will never invent Roman Catholicism out of thin air. Boyer claims that cognitive science can help us to understand the psychological mechanisms that account for these manifest correlations and in so doing enable us to better understand the nature of religious belief and practice. But what motivates a cognitive approach to religion in the first place?

Until the 1950s the best psychological account of learning was behaviourism, the leading proponent of which was B.F. Skinner. Skinner held that children acquire beliefs and habits through positive and negative stimulus from their environment. Reward or punish a child in the right way and you can create any kind of adult you like - a doctor, a historian, or a criminal.

Behaviourism seemed to provide a scientifically respectable story about why persons vary according to their circumstances. Its message was also politically attractive. Given the right stimulus/response conditions, we can create Utopia. Comforting idea but, alas, hopeless as a psychological program. With regard to language, Noam Chomsky demonstrated that the window for learning is far too narrow for children to acquire linguistic competence by conditioned learning. There is 'a poverty of stimulus' for the relevant linguistic information. Generative linguistics was developed in response to the need for a plausible theory of linguistic development. The approach was based on the idea that 1) there is a human language faculty; 2) the language faculty generates rules for the interpretation and production of meaningful sentences; and 3) the language faculty is largely innate and invariant among different communities of speakers. All children are born with the same language faculty, whose growth and development resembles organ growth: children do not 'learn' a language any more than they learn to grow teeth.

Generative linguistics is only one of the more promising areas of research within a broader computational approach to mind. In the late 1930s, Alan Turing showed that for any information-processing task, there is a computational system that can model it ('On Computable Numbers, With an Application to the Entscheidungs Problem' [1938]). While the brain is not a computer, Turing showed that the activity of both brains and computers can be accounted for by appeal to specific computational principles that describe their operation. (Just as both airplanes and birds fly in virtue of the common principles of aerodynamics). The brain, like a computer, processes information according to rules. By conceiving of the mind as a computational device, and analysing cognitive tasks as information processing problems, researchers have achieved new depth in their understanding of a large number of cognitive domains, including vision, memory, mental imagery, spatial representation, and social exchange.

Given this background we can begin to see what compels Boyer's approach to religion. To the extent that the mechanisms controlling the acquisitions and transmission of religious concepts rely on human brains, the mechanisms are open to computational analysis. All thought is computationally structured, including religious thought. So presumably, computational approaches can shed light on the nature and scope of religious cognition. How then can we begin to uncover the computational natural of religious thought?

Boyer moves outside the leading currents in mainstream cognitive psychology and suggests that we can use evolutionary biology to unravel the relevant mental architecture. Our brains are, after all, biological objects and the best naturalistic account of design in nature is Darwin's theory of evolution. To the extent that mental architecture exhibits intricate design, it is plausible to think that the design is the result of evolutionary processes working over vast periods of time. Like all biological systems, the mind is optimised to promote survival and reproduction in the evolutionary environment. On this view, all specialized cognitive functions broadly serve those reproductive ends.

In adopting this perspective, Boyer takes inspiration from a newly emerging branch of cognitive science called 'evolutionary psychology.' Evolutionary psychologists are ordinary cognitive psychologists who use evolutionary reasoning to generate hypotheses for how the mind works. They argue that by reflecting on the cognitive demands on survival and reproduction in the ancestral environment (the setting in which human evolution took place) and assuming that our cognitive features provide (approximately) optimal solutions to the various reproductive tasks that the ancestral environment presented, we can generate an ideal specification for our mental architecture. With this blueprint in hand we can search for actual design features matching the idealization. This ideal blueprint enables us to formulate more precise hypotheses for actual mental organization.

A key problem in accounting for religious belief and practice from an evolutionary standpoint is that ostensibly religion does not further an organism's reproductive interests. Given the exacting specifications imposed on mind design by natural selection, and given the massive resources expended through religious activity, we would predict evolution to have predisposed human beings to intense religious scepticism. It is difficult to discern reproductive ends furthered by dispositions to believe in invisible beings and to devote enormous resources to communicating with and pleasing them. It seems we would have been better off with more parsimonious dispositions, for example, with inclinations to lavish spare resources on our kin, to believe in and respond to predators, prey, and potential mates only. It is hard to imagine why natural selection would have produced cognitive mechanisms that led us to sacrifice livestock to gods or support priests. Call this theoretical problem 'the design problem'.

Given his advocacy for the evolutionary psychology of religion, I was surprised that Boyer did not pursue an adaptivist strategy of explanation for religious dispositions - resolving the design problem by looking for indirect benefits accruing to manifestly wasteful activity - and then use this analysis to generate hypotheses about the cognitive systems underlying religious thought and practice. I expected Boyer to argue with E.O. Wilson (Consilience [New York, 1998]) for the view that such dispositions build social solidarity, which in turn promotes survival in some approximately evolutionarily optimal fashion.

Boyer's solution to the design problem is to claim that religious thoughts and practices are the epiphenomenal outcomes of other cognitive faculties that do serve adaptive functions (p. 202). In the language of biology, religion is a 'spandrel': religious thoughts and practices are best explained as cognitive noise the brain produces when fully optimised mental machinery runs as it has been designed. Given the way our minds are organized, we find ourselves attracted to religious thoughts and practices, but being religious is by no means a forgone conclusion. According to Boyer, religion itself is not a dimension of human nature. Unlike language, it is not an adaptive feature of our biological design.

To support this conclusion Boyer points out that not everyone develops beliefs in gods, and even among those that do belief is not always acquired and represented in similar ways. Take belief in unseen spirits. Elaborate theological constructions aside, cognitive research into ordinary theological thought reveals that most people who believe in the gods represent them as similar to human beings. The gods are always subject to human-like emotions: anger, jealousy, concern and the like, speak (native) languages, have memories, representations of themselves as individuals, entertain social relations and so on. They always differ from human beings in only a few respects. A god may have infinite power to create and destroy, or may live forever, or may entertain social relations with all beings. In Boyer's view, we acquire god-concepts when exposed to them because 1) our minds are already designed to represent attunement to other persons (especially superiors who can make or break us) and 2) we easily remember persons who are extraordinary in a few ways. Importantly, elaborate theological constructs that depict the gods as different from us in many ways are difficult to remember, and are not representative of the ways in which ordinary non-specialists think about deities.

Readers friendly to Richard Dawkins's 'meme-theory' of religion will find some comfort in Boyer's explanation. For Boyer, as for Dawkins, religious representations adhere to our cognitive systems because the representations are well adapted to do so. Crucially, those representations do not necessarily promote the biological fitness of the organisms that host them. Boyer's view enriches and extends meme-theory by exploring the cognitive systems that support religious ideas and constrain their formal expression. Without that account of those systems, meme-theory remains abstract and metaphorical.

Some students of religion may want to object that religion has more to do with practice than with 'ideas', doxa remaining a central concern of Christian, but few other, religious traditions. Boyer devotes a chapter to ritual that seeks to broaden the perspective of cognitive and evolutionary psychology beyond 'religious concepts' to the arena of practice. Here, again, he shies away from providing any directly adaptivist rationale, instead arguing that ritual practice arises from the proper function of other computational features of our brains.

Boyer's reconstruction begins with an analysis of the inference-systems that govern our perception of cause-effect relations, systems organized by natural selection that produce concepts of unseen powers and agencies. Consider the concept of momentum. We all know that if you throw a baseball it will continue (roughly) in a straight path forward before falling to the ground (rather than flying at the speed of light towards Jupiter, etc.) No one teaches us to think that way. Instead this pattern of inference forms part of our intuitive psychology of physical objects: human beings grasped the concept of momentum long before Newton gave it a theoretical explanation. Moreover we all know that a seedling may grow into a tree, but not into a snake or zebra. We know this because we come equipped with an intuitive folk-biology that provides us with the relevant inference structures to predict the behaviour of organic life. These inferential structures, which suit us well to the circumstances of biological existence, employ concepts of unseen powers. Next Boyer looks to the systems that generate internally structured beliefs regarding the causal effects of various human activities. When I throw a ball it flies, when I plant a seedling it grows, when I kick Paul, he groans, and so on. Generalizing, humans can engage in activities that produce reliable outcomes.

Boyer explains ritual action in virtue of the interaction between these two systems. Given that our cognitive systems predispose us to think both that hidden causal forces operate in nature, and that our actions have effects, we are naturally disposed to engage in practices in which we act to influence hidden powers in favourable ways. We seek to influence the hidden world of causation to further our interests by relating our activities to it.

Boyer furthermore observes that the gods are not indispensable to these ritual practices because the causal understandings may be impersonal. Boyer suggests that secular rituals having to do with 'our tradition' and 'society' typically assume the relevant causal roles. Nevertheless, given the salience of our social models of inference, again largely innate, we tend to ascribe personal agency to the hidden causal powers. This is why, in Boyer's view, the gods are often, but not always, objects of ritual concern (p. 296).(This explanation is not original to Boyer. It is essentially the position of Thomas Lawson and Robert McCauley, Rethinking Religion: Connecting Cognition and Culture, Cambridge University Press, Cambridge, 1990.)

Though I think computational considerations strongly motivate the broader cognitive approach to religion that Boyer promotes, I have a few small reservations about Boyer's explanatory account.

First, beyond exploiting the idea that human beings have intuitive inferential systems relating to their environments, Boyer does not attempt to use evolutionary biology to generate hypotheses about religious cognition. He assumes that because not all persons express religious commitments, and because religious commitments are acquired and represented in different ways and are (apparently) sub-optimal, there can be no dedicated religion faculty. But this conclusion does not follow. Boyer's own research on inference systems suggests that 1) we tend to posit invisible causal forces and 2) these forces tend to be personified. It may turn out that what we describe as 'religious thought' is an internally guided expression of the normal operation of biologically adaptive faculties, systems which have evolved to produce close to optimal reproductive outcomes, at least in the ancestral environment. In spite of apparent inefficiencies, there may be some direct adaptive benefits accruing to dispositions to believe in and rally around unseen personalities or sacred powers. The systems may exploit the features of intuitive psychology that Boyer examines, but in ways designed to promote the religious organism's reproductive ends. Given nature's efficient economy, I think the adaptivist approach to the design problem deserves a more careful consideration than Boyer accords it.

Moreover, it may be the case that 'religious representations' are not acquired in quite the way Boyer claims. Research in linguistics suggests that the semantic component of language is largely innate. Words offer richly structured perspectives for viewing the world and interpreting experience. Sometimes these perspectives conflict. I can paint a door brown and then walk through it, but I cannot do the same with a wall. I can finish a bottle and then break it, but not a sandwich. Linguists are now extending 'poverty of stimulus' arguments to show that these meanings and outlooks are not acquired through learning. It may turn out that the religious concepts that Boyer describes as 'acquired' are essentially innate, even if social conditions set certain parameters to the system and prompt development. This would be a surprising but far from implausible result. I think the verdict is still out on this question.

Even if Religion Explained provides less than its title claims, I do not want to understate the importance of Boyer's work to the psychology of religion. Empirically weak accounts of religion have dominated psychological inquiry in Religious Studies for most of its history. Boyer's book outlines a research program drawing on the most powerful areas of naturalistic inquiry into the mind. In my view, the cognitive psychology of religion has set a new standard of rigor and precision for our field. This is a tremendously important book, and anyone interested in promising new currents in the psychology of religion should read it.

Dr Joseph Bulbulia

Religious Studies, Victoria University of Wellington, New Zealand