Imagination is often seen as a faculty of free invention, bewildering in its content and arbitrary in its combinations. This received view assumes that imagination is a relatively isolated faculty of human reasoning. This is problematic, however, if one cannot ‘imagine’ human rationality apart from epistemic activities such as envisaging, associating, conjecturing, or hypothesizing. Cognitive scientists George Lakoff and Mark Johnson have thus made a strong case for seeing imagination as ubiquitous in all sorts of human reasoning. Moreover, they have convincingly shown that metaphors and imaginations are neither free inventions nor arbitrary, since they are rooted in bodily experiences with some universal characteristics. A spatial orientation (up/down, here/there, too close/too far, etc.) is thus always combined with qualitative judgments arising from the natural senses (high/low, light/dark, warm/cold, and so on). Also higher-order rational judgments seem to proceed by coordinating and pruning imaginative concepts, rather than replacing ‘images’ with ‘concepts’, ‘metaphorical’ meanings with ‘literal’ ones. The rational thrust towards ordering, combining and patterning are undergirded by networks of spatial images and bodily metaphors. The first part of this essay will be devoted to an analysis of imagination from the perspective of embodied cognitive science.

In the second part I want to argue that the same naturalness of imagination also applies to religious imagination, and that also religious imaginations are governed by rules that to a certain extent constrain the wildness of religious imagination. This point is probably a bit more controversial, and from the outset I must concede a difficulty. I will thus not be able to provide a clear cut essentialist definition of what constitutes a ‘religious’ imagination as opposed to other sorts of imagination. Nonetheless I want to make the case that religious imagination is not something esoteric that can be added (and later subtracted) from other mental images. Rather, processes of imagination open themselves up to still more generalized images, some of which will be deemed ‘religious’ according to standard linguistic usage. It indeed requires a hard work of imagination to sing with John Lennon: ‘Imagine, there is no heaven, it is easy if you try, there is only sky above’. For the gaze of the natural properties of the ever-receding horizon elicits an intuition of endlessness, just as the experience of swimming in the ocean may produce an ‘oceanic’ feeling of embeddedness. Without postulating a specific religious capacity for ‘religious imagination’, I take my point of departure in the mundane observation that religious imageries – as a matter of fact – are triggered by these (and many other) first-order experiences. The attempt to find a clear demarcation line between ‘imagination in general’ and ‘religious imagination in particular’ appears as artificial.

The question is how to explain this fact. I shall here discuss the thesis of evolutionary psychology that religious imaginations are almost hard-wired into the cognitive structure of our evolved brains when triggered by the appropriate circumstances. Almost, I said, because the thesis of evolutionary psychology is not that our brains produce religious imaginations as a result of some built-in and ready-made concepts of God, in terms of a cognitio innata dei. Nor is the argument that the idea of a personal God comes up through a wholly rational inference from the empirical realities of skies and oceans in the sense of a cognitio aquisita dei. Rather the point of scholars such as Steven Mithen and Pascal Boyer is that the emergence of religious imagination is part and parcel of the general human development of cognitive systems. More precisely, religious imaginations and concepts emerge in the creative zones of interaction and tension between pre-conceived templates of human understanding (such as ‘person’, ‘animal’, ‘tool’, ‘being-in’, ‘effect’) and the manifold experiences that challenge our minds to produce new and often contra-intuitive imaginations. Religious imagination, in this perspective, comes about by blending and combining templates of understanding used for other purposes as well. God, for instance, is both imagined as a person with rationality and will, and as light, fire, rock, and sky. In this sense religious imaginations are natural phenomena, deep-seated as they are in the cognitive capacities of the human brain. How far they are also non-arbitrary remains to be discussed.

These findings of evolutionary psychology and cognitive science, tentative as they are, may be taken as a good message for those who wish to see religion as belonging to the human condition, not likely ever to die out. However, the same findings may also be taken to suggest that religion is substantially a pre-rational phenomenon which is fobbed upon us by our ancestors simply because religion has worked so well in the history of homo sapiens, and because it continues to enhance our personal well-being, our social commitments and our cognitive orientation—even though religious imaginations, regarded in themselves, are nothing but useful fictions. My own position on this matter, however, is that evolutionary explanations of the historical emergence of religions, and of their seemingly eradicable persistence, are neutral to the normative question as to whether religious beliefs exist for reasons other than their unreasonable effectiveness. But, as I shall argue in the third part, the constraints on religious imagination are not without relevance for a philosophical or theological understanding of religion. In fact, evolutionary psychology may even illuminate, from an external perspective, what Christians and other religionists mean when they, from an internal perspective, refer to a divine revelation in particular persons and words.

In philosophical tradition imagination has usually been placed in an ambiguous twilight between sensation and thinking. According to Aristotle, phantasía always emerges with sensation or direct perception (aisthæsis) which is nonetheless transcended by imagination. Unlike sensation, imagination is the ‘placing before our eyes’ of absent phenomena; thus imagination can err, since we imagine things not given by perception, as when we form images with our eyes shut. Accordingly, ‘all sensations are true, but most imaginations are false’. Like thinking (diánoia) imagination forms judgments, but unlike thinking the phantasmata are free elaborations of the human cognitive faculties, unconstrained by logic and examination. According to Aristotle, imagination is therefore to be treated with care, if not suspicion. [1] And yet, at the end of his analysis Aristotle casually remarks that ‘the soul never thinks without a mental image (aneu phantásmatos)’. As a matter of fact, images take the place of direct perception in the process of thinking; accordingly also Aristotle is compelled to state that no thinking can proceed without images. [2]

This view is maintained also in theological tradition. At a prominent place Thomas Aquinas quotes Aristotle approvingly: nihil sine phantasmata intelligit anima. But whereas Aristotle was concerned about the overflow of imagination in relation to sensation, Thomas wrestles with the fact that corporeal imaginations flow into religious concepts in uncontrolled ways. Imagination, says Thomas, should be used with utmost care in spiritual matters, since imaginations are derived from sensual realm and are only analogically transferred from here to the spiritual realm. In a theological context, the danger of idolatry requires some hesitancy towards religious imaginations. While concepts such as unity and simplicity adequately reflect divine nature, corporeal imaginations do not. Angels don’t have them, and angels don’t need them, and neither will human persons need imagination in the state of blessedness, when faith is transformed into vision. [3]

A similar ambivalence of distrust and yet recognition of the irreplaceability of imagination is found also in Kant’s rehabilitation of the Einbildungskraft in Critique of Pure Reason. Imagination is here called a ‘blind but indispensable function of the soul, without which we should have no knowledge whatsoever, but of which we are scarcely ever conscious’. More precisely Kant distinguished between reproductive and productive imagination. Reproductive imagination supplements the fragmentary sense inputs, so that when we see a rounded green apple, we assume that it has some continuous features of location and quality, and that its identity persists in a continuous temporal series. Whereas the reproductive or associative imagination (empirische Einbildungskraft) is operative in the apprehension of particular empirical phenomena, the productive imagination (die produktive Synthesis der Einbildungskraft) provides a global orientation by placing any perception in the a priori context of one single, unified experience of all possible states of consciousness. Thus without the productive imagination the reproductive imagination could not work at all. [4] In this manner, Kant actually accords imagination a far more constitutive role for human cognition than did Aristotle; but it is a formal role, and the analysis of the Einbildungskraft is not coupled with reflections on the role of particular imaginations.

Both Aristotle and Kant thus construe the role of imagination in terms of a faculty psychology. The faculty of imagination is one ingredient, as it were, in cognitive processes, while the thinking in categories and concepts retains a realm of its own, untouched by particular imaginations. Compare these classic views of human knowledge with the understanding of knowledge in the empirically oriented cognitive science. In Philosophy in the Flesh: The Embodied Mind and Its Challenge to Western Thought, the cognitive scientists George Lakoff and Mark Johnson argue that ordinary language as well as highly theoretical concepts build on primary metaphors, which are learned through sensorimotor practices from childhood and onwards (‘up/down’, ‘cold/warm’, ‘close/distant’). [5] Concepts and metaphors thus grow out of bodily experiences, which, in turn, are always accompanied by feelings and sensations. Far below the threshold of consciousness, metaphors from one source of experience (e.g. bodily location) are blended and conflated with metaphors from another source (e.g. visual sensation), and they end up making complex networks of imageries. There is a constant flow back and forth between spatial-bodily ‘source areas’ and the ‘target areas’ that one wants to address. With more utterances, metaphors and images expand and are blended and combined in ever-new configurations. Consequently, we may end up understanding our inner life as a ‘journey’, and talk about ‘reaching our goals’ or ‘losing ourselves’. [6] We are here already approaching the topics of a religious language. For networks of metaphors and images function as fundamental thought schemes that guide our very pedestrian activities (usually taken to be the ‘literal’ ones), our most intimate self-reflections (often taken to build on ‘metaphorical imagination’) and our theoretical concepts (often assumed to ‘transcend’ imagination).

Think, for example, on the theological idea of ‘panentheism’ which has received a renewed attention in the last 30–40 years of philosophy of religion. There are several panentheisms, [7] but they all rely on a set of first order imaginations: ‘pan’ reflects the intuition of spaciousness and horizontal infinity, ‘en’ makes use of what Lakoff and Johnson call the ‘container scheme’ (Lakoff/Johnson), and ‘theism’ is in itself a second-order metaphor which, as I’m going to argue in a while, build on a generalized synthesis between templates of personal knowledge (‘knowing’, ‘willing’, ‘revealing’, ‘acting’ and so on) with the cognitive templates of nature and technical knowledge (‘causing’ ‘effecting’ ‘producing’, and so on).

Not only are concepts nourished from metaphors, but because metaphors are rooted in sensorimotor experiences, they are cross-culturally associated with certain evaluative schemes: ‘Up’ is good, ‘down’ is bad, a ‘warm’ smile is better than ‘cold’ one, and a ‘close’ friend is more important than a ‘distant’ relative. As Lakoff and Johnson put it, ‘Reason is imaginative in that bodily inference forms are mapped onto abstract modes of inference by metaphor’ (1999, 77, my emphasis). [8]

The second-generation cognitive science that Lakoff and Johnson exemplify no longer assumes a body-free intelligence (in the mode of earlier AI) but an embodied mind. Accordingly the task of cognitive science is to reconstruct the emergence of mind from the natural history of evolving sensory systems. Embodied cognitive science seems to me to have at least two interesting perspectives for our theme, one more general and one more specific.

The first point is that metaphors and concepts should not be taken as representing objective realities (or phenomena) ‘out there’; but neither should our cognitive capacities be seen as producing wildly ‘subjective’ imaginations. Rather, imaginative concepts reflect ways of coping with the world, a world with which we are already interacting as participatory cognizers who are bound to understand the world from the perspective of bodily-mental metaphors and concepts. Aristotle’s fear that imaginations are polluting the unspoiled direct sensations presupposes an ill-founded dichotomy between ‘inner’ and ‘outer’ as well as a dubious assumption of the absolute reliability of the senses. But note that empirical cognitive science does not commend relativism either. We could thus put the point of embodied cognitive science as follows: What imagination does ‘mirror’ are the continuous and creative interactions between embodied minds and their environments. Information is not to be transported from the ‘external, objective’ world to an ‘internal subjective world’, but information emerges out of the way in which we learn to take up the world in the process of knowing how to ‘receive’ and ‘take up’ our environments. Lakoff and Johnson calls this an ‘embodied scientific realism’, [9] and they point out how this understanding of embodied knowledge is congenial both with the pragmatism of John Dewey and the phenomenology of Maurice Merleau-Ponty. Cognition is rooted in bodily experiences, but bodies are themselves embedded in more encompassing biological, psychological, and cultural contexts. Bodies are never ‘pure’ bodies, but always interpreting bodies.

The other point relating more specifically to imagination is that imagination should not to be seen as a separate mental faculty isolated from (and perhaps overruled by) other mental capacities. The common sense ‘faculty psychology’ that separates the capacities of perception, imagination, reason, memory, will, and so on, is hardly tenable. What is characteristic for the operation of the human mind, rather, is the constitutive conflation of metaphors and cognitive processes, from one end of the spectrum, say the bodily experience of ‘sun light’, to the other end of the spectrum, say the experience of ‘bright intelligence’ or ‘divine light’, and vice versa. In this sense, also what Kant called the a priori productive power of imagination is something which has been learned a posteriori. The distinction between empirical and transcendental evaporates, for the ‘transcendentals’ are themselves the accumulated results of childhood sensorimotor learning and the subsequent learning of linguistic skills.

Let us now place this view of embodied knowledge in a wider evolutionary perspective, and see what this might tell us about the epistemic status of religious imaginations. No child develops as a separate individual, but is raised in human societies with specific needs and imaginations. Paleoanthropologists tell us that the anatomically modern humans (homo sapiens sapiens) have existed in around 100.000 years (kyr), but only in the ‘cultural explosion’ between 60 and 30 kyr ago do we begin to find evidence of technological innovation, art, religion and rituals. Among the Cro-Magnons (ca. 30 kyr) we find elaborate burials with ochre and extensive grave goods, both of which indicate a belief in an afterlife. [10]

Now evolutionary psychology assumes that beneath the surface of cultural variability, human minds have developed some well-winnowed cognitive strategies that are likely to be relatively constant cross-culturally. Just as there is no dichotomy between external objective reality and inner subjective consciousness, there is no absolute distinction between nature and culture. Also cultures have to adapt to the particular problems of survival, reproduction, group cooperation, and world-orientation. Accordingly natural selection applies no less to human cultures than to animal behaviors, though the means and forms of selection may be different. However, unlike sociobiology evolutionary psychologists do not normally believe that the genes are particularly responsible for producing specific behaviors, nor are they concerned about behavior apart from the cognitive mechanisms implied in such behavior. Reproductive fitness as well as cultural fitness is conditioned by the operation of specialized ‘mental modules’, which have proven themselves to be efficient in the past of our hunter-gatherer ancestors, and which, due to their hard-wired neural correlates, persist in ancient agricultures up to modern cultures.

Evolutionary psychologists tend to follow Jerry Fodor’s modular view of human cognition, the so-called Swiss-Army model, rather than Noam Chomsky’s idea of a rule-based, but content-free general-purpose brain. The mind is thus constituted by multiple, specialized and content-rich mental modules that have developed in our ancestors, but are still with us: [11]

(Figure 1)

Figure 1: Two Views of the Human Mind: The content-free, general-purpose ‘Standard Social Science’ model of the human mind, and the content-rich, modular view of Evolutionary Psychology.

Mental modules are operative in face recognition, spatial relations, tool-use, social exchange, perception and emotion, child care, friendship, face-to-face communication, and so on, and they are learnt from childhood experiences, and later enhanced, restrained, or refined by social learning. Thus mental schemes function to a wide extent automatically beneath the level of conscious reflection. [12] They have, over time, become hard-wired in our neural circuits.

What is characteristic for the human mind, however, is that this automaticity is always modulated by an interaction between the different capacities, so that human beings are not simply determined by stimulus-response behaviors. The question is, then, how to account both for the immediacy of human cognition-and-action, and for the highly modulated response acquired through social learning, and facilitated by the access to symbolic networks. There is, as far as I am aware, no really convincing answer to this question; nor does there exist any consensus on the origin of human language. In The Symbolic Species: The Co-evolution of Language and the Brain, the neuroscientist and evolutionary anthropologist Terrence Deacon has argued there is a difference in kind between human communication and animal communication (as evidenced in say, the waggle dance of honeybees, whale songs, or monkey alarm calls). In animal communication, we have a token-token relation between the sign and the reference, whereas the human languages have transcended what Deacon calls ‘the symbolic threshold’ by not only associating signs with ‘realities’ but also forming higher-order symbol systems by which we can establish logical relationships between the symbolic elements (words, signs, etc). This allows us to refer not only to external objects but to communicate about these referents as well as about our own possible interactions with these referents. By way of symbolic systems we interpret something as something in different contexts, and can thus evaluate different possibilities. For example, we learn to discern whether a strained face means ‘aggression’ or ‘pain’, whether a story is an offense or a joke, and in addition we learn to imagine how we could possibly react to possible states. Most importantly, Deacon emphasizes that humans are able to unlearn dispositions by re-interpreting what we see in the wider horizon of possibilities. The ability to say ‘No’ is constitutive for humanity, as is the ability to live in a world of imagined counterfactuals. In both cases combinatorial capacities are crucial: ‘Symbolic reference derives from combinatorial possibilities and impossibilities, and we therefore depend on combinations both to discover it (during learning) and to make use of it (during communication)’. [13] According to Deacon, no human communication exists that did not possess the open-ended structure of second-order, symbolic reflection.

Other evolutionary psychologists tend to think that human intelligence has eventually evolved through combining a general intelligence with more specialized or modular intelligences which again can be blended in higher-order cognitive processes such artistic creativity and religious thought. In The Prehistory of the Mind, archaeologist Steven Mithen has argued that apart from the general intelligence characteristic for infants (phase 1), children early on develop specialized cognitive domains (phase 2). Children learn to behave differently to human persons, to animals, and to tools, and they thus acquire a set of distinct, ‘intuitive intelligences’. Mithen mentions four such areas of intelligence: (1) social and psychological, (2) biological, (3) technical, and (4) linguistic. What is characteristic for humanity, however, is the extent to which ‘the combining of thoughts and knowledge of the different specialized intelligences is possible’ (phase 3). As soon as the specialized cognitive domains begin to engage one another—and this is what happened in the cultural explosion in the Cro Magnon Age, ‘the result is an almost limitless capacity for imagination’. Mithen refers to this phase 3 of the typically modern human mind as having a ‘cognitive fluidity’. [14] In a schematic form, Mithen presents his view in the form of a gothic cathedral.

(figure 2).

Figure 2. Steven Mithen’s Gothic Cathedral view of the origins of human cognitive systems.

In this perspective religious imagination is a result of the cognitive fluidity attained at phase 3, when intelligences related to social persons and to natural history were combined into the idea of a non-physical after-life (having the persistence of mountains but the features of a personal mental life). These complex notions can be further combined into the generalized notion of non-physical spirits and gods. These, in turn, are thought to be as causally effective as natural events, in so far as they can cause harm or well-being, and yet as communicative as human persons. [15] In this sense, religious imagination uses the same mental modules as are at work in other human activities, such as hunting, cooking, reproducing, nurturing, and so on. Accordingly, there exists no specialized religious module, no natural borderlines between religious and non-religious imagination. Rather the emergence of religion is part and parcel of the liberation of human rationality from the constrained navel structure of the Roman chapels to the open and fluid structures of the Gothic cathedrals.

Apart from its schematism, which Mithen fully acknowledges, his approach does not seem quite satisfying to me for at least two reasons. What he says about religion is after all relatively meager, even though religion is part of the book title. Secondly, and more importantly, Mithen does not discuss the inner constraints on religious (or any other) imaginations. If cognitive fluidity were the one and only explanation of religion, religious imaginations would just teem in all sorts of directions. But, as argued by Pascal Boyer, a cognitive scientist specializing in the emergence of religious concepts, this is not quite the case. Religious imaginations are, after all, highly limited. [16]

In his book with the somewhat exaggerated title, Religion Explained, Boyer offers a model for understanding how religious concepts, often amazingly counterintuitive and sometimes even baroquely exotic, have their natural place in the context of the ordinary workings of human brain. Fundamental to Boyer’s explanatory model is the distinction between concepts and templates. [17] ‘Concepts’ are general ideas referring to particular beings such as a walrus or a giraffe, whereas ‘templates’ are more general schemes. The point is that children as well as adults learn by subconsciously inferring that giraffes, even though they look very different from walruses, have a variety of common characteristics with them, because they are part of the same ontological category: the ANIMAL template. Animals have a body-plan, a living place, eat food, reproduce, and so on. Of course a giraffe and the walrus look differently (long legs and neck versus a trunk), have different habitats (the savannah versus the sea), eat different things (leaves versus fish), and copulate in different ways. The information about these differences will have to be provided by the concepts, derived from particular sets of experience. But the template ANIMAL tells us that all animals – from mosquitos to elephants – live, eat and reproduce. Templates are aggregates of memory. By subconsciously using templates any child is inferring many things that we don’t observe but simply take for granted. The result is a kind of inverted Kantianism, in which the quasi a priori templates have accumulated through the evolution of cognitive systems:

(figure 3)

Figure 3. Chart on the imaginative animal race ‘Zygoons’ which are told to feed on hyenas. Observe the combination between the ontological entry, provided by the templates, and the novel information offered by the empirical concepts.

Now Boyer furthermore contends that there are not many templates of understanding. How many one wants to enumerate is of course a question of definition, but the following may suffice: PERSON, ANIMAL, PLANT, TOOL, ARTEFACT, NATURAL OBJECT. For each of these templates all human beings will have a long-term acquaintance. Tools and natural objects don’t talk and don’t eat; persons do. Persons have memory and act according their past experiences, plants don’t. And yet, since the categories can be blended, new combinations can take place, and the world of religion is full of such cases. Boyer mentions the examples of praying to statutes, of feeding mountains for an exchange of prey, and of special (potentially dangerous) ebony trees who are able to recall the conversations of past generations. In effect, Boyer thinks that religious concepts came about by blending information coming from separate ontological categories, such as the templates of PLANT and PERSON:

(figure 4)

Figure 4. Chart on the idea of special ebony tress able to recall conversations.

The point is here that the information contained by the new information tags contradicts the information presupposed by the ontological categories. The hard core of Boyer’s theory is now that religious concepts and imaginations are always marked by being counterintuitive in the precise sense of counteracting expectations raised by our template categories: ‘[R]eligious concepts invariably include information that is counterintuitive relative to the category activated’ (Boyer 2001, 65). Observe that even though the world of religions do involve oddities of many sorts (at least to outside observers), the mere fact that something is unexpected is not counterintuitive in this technical sense. For example, to imagine ‘a table made out of chocolate’ or ‘giraffe with six legs’ destroys our ordinary expectations, but does not violate an ontological category. The breach of expectations is here still at the level of concepts and natural-kinds However, to say that ‘the table felt sad, when the people left the room’ does break with our assumed information about what ARTEFACTS can do (Boyer 2001, 80–82). This example, by the way, also shows that religion cannot be sufficiently defined by being contra-intuitive (or ‘counter-ontological’, to be more precise). For so are also fairytales and science fiction stories. Boyer is aware that religious concepts have an existentially importance not always present in stories, but he does in fact not provide a very useful understanding of the nature of religious concepts. By religious concepts he mostly seems to understand simply the idea of supernatural beings such as gods, ghosts and zombies, or supernatural events such as miracles. I am not convinced that this is the right angle to approach religious imagination, but I must leave this aside for now.

Let me instead focus on the implications of Boyer’s cognitive approach for the concept of God. His first point is the empirical observation that there are many imaginable, yet non-viable ways of thinking about God. Not all religious concepts can be equally successfully transmitted. Boyer mentions the example of a god watching us in every detail—but instantaneously forgetting about us. This is a notion nowhere found in the history of religion (2001, 51). Or think an omnipotent God existing only on Wednesdays (2001, 56). There are here barriers to the wildness of religious imagination, and cognitive fluidity does not flow in all directions. The reason is that any violation of templates will need to be specific, while at the same time preserving other features of the template (2001, 62).

So, when one imagines God as a PERSON, most believers would violate this category by qualifying the template so that God is a person who is not limited to space and time, because God has no body. By contrast, it would be hard to come through with a religious message saying that God is an infinite person who has a body but no thoughts. The ontological category of personhood would here not only be breached, but simply eradicated. One the other hand, one can also enrich the concept of God by transferring specific knowledge claims gained from other source areas and applying them to God. One can, for example, use the TOOL template and praise God by saying, ‘Thou are the lamp that shines for my foot’, or the NATURAL OBJECT template and say, ‘God is your rock’.

Furthermore Boyer argues that there is an explanation of the pervasive role of PERSON templates in religious thought. Throughout evolution, humans have become used to pay attention to differences, especially to salient features that can be treated as signs standing for something, having some hidden meaning. To put it a bit robustly, we are evolutionarily designed to look for signs, and we are bound to read events as signs indicating the activity of somebody else. Boyer here refers to the evolutionary psychologist Justin Barrett, who has argued that human cognitive systems are marked by a ‘hyperactive agent detection’: ‘Our evolutionary heritage is that of organisms that must deal with both predators and prey. In either situation, it is far more advantageous to overdetect agency than to underdetect it. The expense of false positives (seeing agents where there are none) is minimal, if we can abandon these misguided intuitions quickly. In contrast, the cost of not detecting agents when they are actually around (either predator or prey), could be very high’ (2001, 144).

In this sense, gods and spirits can be (and certainly have been) perceived as predators that provoked fear and anxiety. But gods and spirits can also be seen as invisible partners with whom one can seek refuge, communicate, but also possibly exchange goods (2001, 146–150). What is distinctive for religious communication is here the fact that the communication with invisible partners is decoupled from the external social exchange, and thus offers a space for learning both social and self-reflective skills against a stable background, constituted by the relation to the Invisible Other (2001, 149). In other words, the pervasiveness and persistence of notions of a personal God, despite the criticisms coming from philosophers such as Spinoza or Fichte, can partly be explained by the naturalness of the mental modul of agency detection, which is operative far below the threshold of reflection. And yet, the agency detection also elicits reflection.

Also Boyer thus argues that religious imagination (however we want to define ‘religious’) uses the same inference systems as the human brain and mind in general. Boyer refers to empirical investigations showing that people tend to memorize better violations of expectations than no violations, and that people tend to better recall specific expectations than too many. At the level of concepts, one better remembers a ‘one-armed man’ than a ‘two-armed man’; however, if we began to violate a human being much further (e.g. without perceptible face, or without feeling) imagination gets too strained. Similarly in the world of religious imagination. We are evolutionarily bound to pay attention to salient features, and to seeing events as signs, and signs as traces of personal activity. However, as Boyer also points out, not least the elite representations of religion risk the danger of making too many violations of expectations. He mentions the Christian doctrine of Trinity which is notable for being difficult to transmit culturally, or literate Buddhists who endorse a wholly non-anthropomorphic universe. In both cases, we find that the theological correctness maintained by scholars, priests, or monks is counteracted in popular piety by giving priority to one of the three Trinitarian persons, usually Christ, or by supplementing the non-anthropomorphic universe with a world of very anthropomorphic ghosts and spirits.

Summing up, it should be noted that Boyer’s theory only pretends to explain the adaptive power of religious concepts, not their particular semantics. His aim is to explain the emergence as well as the persistence of religious concepts. But also here Boyer admits that his approach does not explain the particulars of religions, nor the beliefs of individual persons. Evolutionary psychology can explain ‘the likelihood of religious ‘belief” as well as the ‘vast trends in human groups’ (2001, 319), but not the particular shape of particular religion. However, not only is the explanatory range of Boyer’s theory limited (as are all scientific explanations). There are also self-imposed limits to the explanatory enterprise itself. In his own words, Boyer concerns himself with cultural adaptation, with ‘cultural fitness’, but not with the ‘conceivability’ of particular religious imaginations (Boyer 2000, 104). As an evolutionary psychologist, he can discuss the viability and plausibility of religious beliefs, but not their internal rationality. In other terms (borrowed from Wheeler’s expression about the unreasonable effectiveness of mathematics), the reasons that may undergird unreasonable effectiveness of religious thought lie outside the scope of evolutionary psychology. To these philosophical and theological questions we now turn.

Let me begin by saying that I am committed to the view that a philosophy of religion (at least in the context of a Christian tradition) can fully acknowledge the naturalness of religious imagination. Religion is a natural phenomenon, simply because it belongs to the order of God’s creation. Hereby I subscribe to a principle of a continuum of reality, but without wanting to pre-define this continuum of reality as ‘natural’ in terms of a ‘scientific naturalism’, or any other specific world-view. There is a richness to reality that may not be captured by scientific approaches, as known from, say, standard physics and evolutionary theory. However, by subscribing to the idea of a continuum of reality, one does not need to buy into the idea of a continuum of rationality, in the sense of believing that nature – in our case the phenomenon of religious imagination – can be adequately represented and explained within the perspective one sort of scientific theory—in this case by cognitive science and evolutionary psychology. Neither does one need to assume that the scientific explanations can be seen as one unbroken series of arguments—from physics to psychology to religious reflection. Rather, the natural and social sciences build up a pluralistic patchwork of partially overlapping, partially conflicting long-range and short-range explanations of a ‘dappled world’ that remains beyond their grasp. [18]

Accordingly, explaining religion in the sense defined above cannot mean explaining it away in a reductionist manner. Nor does explaining the emergence of religion answer the philosophical question of the possible validity of particular religious truth-claims. Evolutionary psychology rather suggests to me that religions consistently produce truth-candidates which should be treated on par with all other sorts of human thinking. I also hope to make a case for showing how the naturalness of religious imaginations does not preclude a sensible notion of ‘revelation’, as provided from the internal perspective of a given religion, in my case the Christian tradition.

[The remainder of the paper will be presented at the conference.]