Why religion is nothing special but is central

Abstract

It is proposed that explaining religion in evolutionary terms is a misleading enterprise because religion is an indissoluble part of a unique aspect of human social organization. Theoretical and empirical research should focus on what differentiates human sociality from that of other primates, i.e. the fact that members of society often act towards each other in terms of essentialized roles and groups. These have a phenomenological existence that is not based on everyday empirical monitoring but on imagined statuses and communities, such as clans or nations. The neurological basis for this type of social, which includes religion, will therefore depend on the development of imagination. It is suggested that such a development of imagination occurred at about the time of the Upper Palaeolithic ‘revolution’.

1. Introduction

This paper reconsiders how we should approach the study of the evolution of religion. The discussion leads me, however, to a more general consideration of the way social cognition has been approached in recent literature. This reconsideration bears in mind the kind of problems that Colin Renfrew has called the ‘sapient paradox’. The paper proposes a cognitively and neurologically more probable scenario for the development of religion than certain recent theories that are questioned by the problems he highlights.

The problems I am referring to are particularly thrown into focus by a series of theories that originate in Sperber's suggestion that religious-like beliefs are to be accounted for by a subtle mix of intuitive human capacities based on evolved neurological modules, and certain, very limited, representations that, because they go against the core knowledge that the modules suggest, are therefore ‘counter-intuitive’ and ‘intriguing’ (Sperber 1985). The motivation for these theories is to seek an answer to a question. How could a sensible animal like modern Homo sapiens, equipped by natural selection with efficient core knowledge (or modular predispositions), i.e. knowledge well suited for dealing with the world as it is, hold such ridiculous ideas as: there are ghosts that go through walls; there exist omniscients; and there are deceased people active after death? The authors who hold such a theory of religion give the following answers to this question. First, our core knowledge ensures that, however bizarre such ideas might seem at first, when they are more closely examined, they, in fact, turn out to be mainly disappointingly intuitive. Second, even though beliefs in supernatural things nevertheless do involve a few counter-intuitive aspects, if only by definition, these are possible owing to accidental misapplications of core knowledge to domains for which it is not designed. These limited misapplications are, however, so alluring that they make these minimally counter-intuitive beliefs spread like wildfire. They thus become key elements in religions (e.g. Boyer 1994, 2001; Pyysiainen 2001).

The problems with these theories that I shall discuss here do not necessarily imply outright rejection. They are what might be called ‘upstream’ objections since they occur even before we consider the main proposals. The first objection echoes a similar one long ago made by Durkheim, but it has been reformulated more recently by Barrett (2004) when he points out that it is odd to account for such a central phenomenon in the history of mankind as religion in terms of minor cognitive malfunctions. My second objection is that those who propose such theories forget the fact that anthropologists have, after countless fruitless attempts, found it impossible to usefully and convincingly cross-culturally isolate or define a distinct phenomenon that can analytically be labelled ‘religion’.1 The third problem with such theories is that they explain religion as a product of core knowledge or modular capacities, such as naive physics, number, naive biology and naive psychology, all of which, with the possible exception of the last, we share with all our anthropoid relatives. Such a proposal is therefore unconvincing simply because no other animal than humans manifests any behaviour that is remotely like what is usually called religion. This lack also seems to be the case for all hominids or hominims, apart from post-Upper Palaeolithic modern Sapiens. In other words, the explanations that I am challenging account for a highly specific and general characteristic of modern Humans, what they call religion, by general factors that have existed for millions of years before the Upper Palaeolithic revolution when the phenomenon first manifested itself.

The alternative story I propose here avoids these problems. It argues that religious-like phenomena in general are an inseparable part of a key adaptation unique to modern humans. This is the capacity to imagine other worlds, an adaptation that I shall argue is the very foundation of the sociality of modern human society. This neurological adaptation occurred most probably fully developed only around the time of the Upper Palaeolithic revolution.

2. The transactional and the transcendental

For heuristic reasons, a consideration of chimpanzee society can serve as a starting point. I turn towards our nearest surviving relatives in order to stress, as is so often the case in the evolutionary literature, a major difference between them and us. Of course, we cannot assume that contemporary chimpanzee social organization is necessarily like that of early Sapiens. There is no way to know; especially since the social organizations of the two extant species of chimpanzees are radically different though both are equally closely, or equally remotely, related to us. In this case, it is not the similarity but the difference that is revealing and this difference provides us with something like a thought experiment that enables us to reflect on certain characteristics of human society.

Chimpanzees do not have anything which remotely resembles the many and varied phenomena that have been labelled religion in anthropology. Indeed, this was probably also true of early Sapiens. But, more importantly, there is also something else that chimpanzees, and probably early Sapiens, do not have. This is social roles or social groups, understood in one particular sense of the word social.

Of course, chimpanzee social organization is highly complex. For example, the dominant animal is not necessarily the biggest or the one who can hit the hardest. Dominance seems to be achieved as much by machiavellian politicking as it is by biting. Also, chimpanzees do create long-lasting coalitions, often of females, and these may well dominate the social organization of the group (De Waal 2000). Such roles and groupings are of a type that I call here the transactional social. This is because such roles and groups are the product of a process of continual manipulation, assertions and defeats. This type of social is also found in modern humans.

However, what chimps do not have is the kind of phenomenon that used to be referred to as ‘social structure’ in the heyday of British social anthropology (Radcliffe-Brown 1952). This I shall label here as the transcendental social. The transcendental social consists of essentialized roles and groups.

Essentialized roles exist separately from the individual who holds them. Rights and duties apply to the role and not to the individual. Thus, a person who is a professor should act ‘as a professor’ irrespective of the kind of person he/she is at any particular stage in the transactional social game. Similarly, in central Madagascar, as a younger brother, I should walk behind my older brother; as a woman, I should not raise my voice in a mixed gathering. All this applies, however powerful I have actually become, even if my prestige is greater than that of my older brother or of a man.

Essentialized groups exist in the sense that a descent group or a nation exists. These groups have phenomenal existence not because the members of the descent group or the nation are doing certain kinds of thing together at particular moments, or because they have been together doing certain kind of things at particular moments in the sufficiently recent past so that it is reasonable to assume that they retain the capacity to behave now in similar ways. One can be a member of an essentialized transcendental group, or a nation, even though one never comes in contact with the other members of the descent group or the nation. One can accept that others are members of such groups irrespective of the kind of relationship one has had with them or that one can suppose one is likely to have with them. Such groups are, to use Benedict Anderson's phrase, ‘imagined communities’ (Andersen 1983).

As noted above, in stressing the system of essentialized roles and groups, I am emphasizing what British social anthropologists, such as Radcliffe-Brown, were referring to when they spoke of social structure. However, my position is theoretically very different from theirs. For them, the human social was equated with the network of such roles and groups. For me, these phenomena are only a part of the social: they are the transcendental social.

The transcendental social is not all there is to human sociality. There is plenty of transactional social in human sociality that occurs side by side or in combination with the transcendental social. The transactional social exists irrespective of the role-like essentialized statuses and the essentialized groups of the transcendental social, though it may use the existence of the transcendental social as one of the many counters used in the transactional game. Human sociality is thus, as Durkheim stressed, double. It has its transactional elements and transcendental element. Chimpanzee sociality, by contrast, is single because the transcendental social does not exist among the chimpanzees.

The double character of the human social can be illustrated by the example of a Malagasy village elder I have known for a long time. By now, he is old, physically weak and a little bit senile. He has difficulty in recognizing people. He spends most of his days in a foetal position wrapped up in a blanket. Yet he is treated with continual deference, consideration, respect and even fear. Whenever there is a ritual to be performed, he has to be put in charge so that he can bless the participants. When he is treated with great respect he is being behaved to, and he accordingly behaves towards others as a transcendental elder. This does not mean, however, that he is not also within the transactional social system. While as a transcendental elder he is little different to what he was when he was in his prime several years ago, as a transactional player he has lost out completely in the machiavellian game of influence, and nobody takes much note of him anymore or of his opinions since in the continual power play of daily life he has become insignificant.

This kind of duality is impossible in chimpanzee society. There, once you are weak or have lost out in the continual wheeling and dealings of power, you lose previous status. In an instant, a dominant animal is replaced in his role (De Waal 2000). A chimpanzee's rank depends entirely on what those it interacts with believe it can do next. Chimpanzees do pay respect to each other in all sorts of ways… for instance, bowing to a dominant animal, but once this animal has lost out in the power game, this behaviour stops instantly. A social position in chimpanzee society never transcends the predictable achievements of the individual. This absence of transcendental roles is where the fundamental difference between chimpanzee and human sociability lies. The Malagasy in the village where this elder lives bow to him just as much now that he is weak as they ever did, even though he has become obviously without transactional influence. It is important to remember, however, that the respect shown to him does not mean that he is an elder all the time. The people, who interact with him, and probably himself, represent him in two ways. These two ways are not experienced as contradictory, but they are clearly distinguished and made visible by the behaviour of all concerned. Everybody knows that he is a weak old man whose hands shake and whose memory is going, and people sometimes behave towards him in terms of that representation, even with occasional cruelty. They also behave towards him in terms of the respect as described above. Thus, he belongs to two networks and, although the two are different, the transcendental network is taken into account in the transactional network while the transactional network affects the transcendental network only indirectly; for example, when another person is ultimately able to replace an elder in his transcendental role through revolutionary manipulation (for example in a traditional African society, convincing people that he is a witch; Middelton 1960).

In order to fully understand the role of an elder such as the one I have in mind, it is essential also to remember that, as a transcendental being, he is part of something that appears as a system, even though this systematicity may be something of an illusion. The transcendental elder implies the existence of transcendental juniors, of transcendental affines, transcendental grandchildren, etc. The transcendental network involves gender roles, thereby creating transcendental women and men. It is a system of interrelated roles and it is this complexity of interrelations at the transcendental level that most critically distinguishes the human social from the sociality of other species.

This transcendental network also includes what the structural functionalists called ‘corporate groups’, but which I have referred to above as essentialized groups. These are transcendental groups. By this, I mean that, for example, members of a clan are dual. At the transactional level, they differ from each other just as much or as little as they do from people of the next clan. But, in the transcendental social mode, all members of such a group are identical as transcendental members. They are, as is often said, ‘one body’. As one body, they differ absolutely, and all in the same way, from those others in the other clan. The transcendental character of such groups is made all the more evident when we realize that the composition of such groups, whether they are clans or nations, may equally include the living and the dead. Thus, when in the transcendental one-body mode, members can make such bizarre statements as ‘We came to this country two hundred years ago’. The transcendental can thus negate the empirically based transactional in which people do not live for 200 years. Thus, the transactional social can as much ignore the present physical state of an elder as it can ignore death and individuality. The transcendental network can with no problem include the dead, ancestors and gods as well as living role holders and members of essentialized groups. Ancestors and gods are compatible with living elders or members of nations because all are equally mysterious invisible, in other words transcendental.

3. The transcendental social and religion

This social indissoluble unity between the living and the dead and between what is often called the ‘religious’ and the ‘social’ has never been better explained than in a famous article by Igor Kopytoff ‘Ancestors as elders in Africa’ (Kopytoff 1971). Although the article is phrased as a criticism of earlier work by Fortes, it actually follows the latter author closely. Kopytoff points out how in many African languages the same word is used for living elders and for dead ancestors whom, it has often been said in the literature, Africans ‘worship’. This is because in a sense, in the transcendental sense, they are the same kind of beings. Kopytoff stresses how both ancestors and elders have much the same powers of blessing and cursing. This leads him to assert that to talk of ‘ancestor worship’, and thereby to suggest something analogous to an Abrahamic notion of a distinction between material and spiritual beings, is an ethnocentric representation that imposes our categorical opposition between the natural and the supernatural, or between the ‘real’ and the religious, onto people for whom the contrast does not exist.

I accept much of Kopytoff's and Fortes' argument and want to expand it. What matters here is that if they are right, there is no reason why we cannot reverse his argument, something that Kopytoff himself suggests. If dead ancestors in an ‘ancestor-worshipping society’ are the same ontological phenomena as elders, then elders have the same ontological status as ancestors. If there is a type of phenomenon that merits the appellation ancestor worship, which suggests the kind of things that have often been called religion, then there is also elder worship or elder religion. And since elders are part of a system, there is in the traditional sense, junior religion, descent group religion, man religion, woman religion, etc.

Although to talk in this way may be fun, we have to use our words with the meanings that they have historically acquired. So it might be better to rephrase the point and say that what has been referred to above as the transcendental social and phenomena that we have ethnocentrically called religion are part and parcel of a single unity. This implies that the English word religion, inevitably carrying with it the history of Christianity, is misleading for understanding such phenomena as ancestor worship since, in such cases, there is not the same boundary between the ‘supernatural’ and the ‘natural’ as that perceived to occur in societies caught in the history of the Abrahamic religions. The boundary exists also in these cases, however, and it occurs between one type of social (the transcendental social including the phenomena that have been called religion) and the transactional social. This boundary is clear in the kind of society I am referring to and explains the two different ways of acting towards the Malagasy elder noted above.

The inseparability of the transcendental social and the religious is not only manifested in cases of so-called ancestor worship. Hinduism is a phenomenon that is often assumed to be comparable with the Abrahamic religions, but such an equation is misleading for the same reasons as apply to the African examples discussed previously. For example, Fuller begins his study of popular Hinduism by pointing out that a wife should, and indeed does, at some moments, treat her husband in the same way as she treats the gods. The same gestures and bodily positions are used in both cases in performing puja and the husband can thus be said to be a ‘god’ to his wife in the Hindu sense of god. The point is that here also the transcendental social husband and wife role is part of one single overarching transcendental hierarchical social system that includes the gods (Fuller 1992).

The societies I have discussed above clearly present a challenge for the kind of theories referred to at the beginning of this article, i.e. the theories advocated by, among others, Boyer. This is because they explain a phenomenon that can only be distinguished from a greater whole: the transcendental social, by using a contrast between the religious and the secular that is borrowed from a relatively modern system of representations that simply does not apply in their cases. Consequently, I shall argue that it is the greater whole in its totality, i.e. the transcendental, that needs to be explained. However, such a redefinition of the project presents an obvious difficulty. If Boyer is wrong to take a specific type of society, those with religion, to represent the human condition in general, is it not equally wrong to take specific other societies, those discussed in this paper so far, as representing human nature?

4. Historical excursions

In what follows I argue that this is not so because societies with religion are the subsequent product of an inessential and superficial modification of the societies discussed above. A full demonstration of this point would require much more space than is available in this short paper. What follows is therefore nothing more than a tentative sketch of what such a proposal would look like. So, in order to explain how a certain state of affairs occurred for some, and only some, human groups, I move to a historical argument to argue that it is in certain specific historical circumstances (admittedly of great importance for the majority of mankind though not for all) that the kind of phenomena we call religious take on a separate appearance that seems to distinguish it from the more inclusive transcendental social.

The creation of an apparently separate religion is closely tied to the history of the state. It has long been noted that in early states such as Mesopotamia, Egypt, China, the early Andean states and many other places that the religious and the ‘political’ were inseparable. Frankfort long ago argued that in ancient Egypt, pharaoh was a visible god interacting on a compatible footing with the invisible gods. The organization of the state was part of the divine order (Frankfort 1948). The ancient Egyptian kingdom was part of an explicit cosmic ordering of space and time. The recurrence of the flooding of the Nile was represented as the consequence of the repetitive cyclic action of the gods, including the pharaoh. The world was centred on the capital with distant uncivilized, barely human, peripheral peoples far from its centre. Egypt was, to borrow the Chinese phrase, the empire of the centre. All this is the familiar attribute of what has been called divine kingship, whether it is that of the Swazi, Indic states or the Mesopotamian city-states.

The transcendental representation of such states was not all there was to political organization. There were also other available transactional representations of the state, pharaoh, time and space. In much the same way as the Malagasy elder is dual, it was also possible to see the pharaoh in more straightforward terms, and that was in spite of the prodigious efforts that were made to transform him through his palace and his tomb into an empirical manifestation of his transcendental side.

The transcendental construction of such states is also accompanied by another corollary process. The development of the Merina state in Madagascar in the eighteenth and nineteenth century shows how the construction of the symbolic state is accompanied by a partial destruction and reformulation of the symbolism of the subjects. Thus, certain key attributes of elders/ancestors were forcibly transferred from local descent groups to the king and his palace (Bloch 1986). Interestingly, a similar process involving the diminution of the transcendental social of subjects for the benefit and construction of the royal transcendental has been examined for early Egypt by Wengrow (2006).

Thus, the royal centralized transcendental construction depends on the partial destruction or at least transformation of the symbolical system of subjects. In Madagascar, the focus of the symbolism of the subjects migrated, thanks to violent encouragement, from the house to the tomb, as the palace became the symbolical house of the kingdom with the ruler as its central ‘post’ (in Malagasy, Andry the root of the word for ruler Andriana: lord; Bloch 1995). Similarly, and in more detail, I described how the descent group ritual of circumcision subsequently became orchestrated by the state and how certain aspects were taken away from the elders to become constitutive elements of grand-state occasions. The descent groups lost key elements to the representative of the state and were punished if they attempted to perform the full ritual independently (Bloch 1986).

Since in such systems the transcendental social and the religious are identical, it is not just the religious that is being reorganized in a centralized state and sucked up to a point into a centralized, organized, organic-seeming system, it is the whole transcendental social. The creation of this transcendental holistic image of the complete kingdom, including gods and men, thus requires the creation of the incompleteness and disorganization of the subjects' transcendental social, which can only be made complete in the kingdom.

After such a process a change that is different to the symbolic centralization of the state happens. States are unstable and political systems continually collapse. That causes a new problem. When the royal state collapses at the hand of its enemies, the subjects find themselves bereft because the construction of the state had previously made them transcendentally incomplete and the state, after its collapse, is not there anymore to complete them.

The same Malagasy example can again illustrate this point. The growth of the Merina kingdom in the nineteenth century had led to the circumcision ritual being partly taken out of the transcendental construction of descent groups and being placed in the realm of the symbolical construction of the kingdom. However, in 1868, when the Merina kingdom became disorganized, in part owing to the influence of Christianity, the ruler failed to perform the royal circumcision ritual. At that point, a popular movement arose which sought to force him to perform it.

Why did the subjects feel bereft by the royal non-performance when originally the ritual had been their privilege? Why should they seek the state that in many ways exploited them? Because, given the previous process, when the state collapsed, they were left with nothing but their incomplete transcendental social and, for reasons that I cannot explain, it seems as if the deprivation process is irreversible. Thus, when the state, having confiscated a large part of the transcendental social so as to create its own ordered pseudo totality of cosmic order, then collapsed, a totalizing transcendental representation without its political foundation remained, floating in mid air, so to speak. This begins to look like what we call religion. For example, the collapse of the political base of the transcendental social may lead to the occurrence of these ritual, sacred, pseudo-royal systems of Africa that so fascinated Frazer, where as Evans-Pritchard said, the king ‘reigns but does not rule’ (Evans-Pritchard 1948). It is what leads to shadow ‘states’ that only exist in mystical form as spirits that possess mediums. Examples of these are found among the Shona or in western Madagascar where they were caused, Feeley-Harnik argues, by the collapse of the political as a consequence of colonial rule (Feeley-Harnik 1991). This is also what explains the bizarre institutions of contemporary European monarchies. These post-state states are ‘religions’, i.e. phenomena apparently distinct from the rest of the transcendental social.

The Abrahamic religions offer another example of the process. The historian of Judaism and of early Christianity, J. Z. Smith, argues that Jewish monotheism must be understood as the product of a longing for the unified, centralized, holistic transcendental Mesopotamian city-states with Ziggurats at their centre. These were a kind of state that the Jews, as minor peripherals to that system, hardly ever managed to achieve for themselves, or, when they did, did so on a tiny fragile scale. Early Judaism is therefore also a transcendental incomplete residue: religion. This residue was modelled on the Mesopotamian prototype with, at its centre, the Ziggurat in a purely religious form, i.e. the temple in Jerusalem (Smith 1982).

With this sort of situation, we therefore get religions that are only apparently separate from the transcendental social state, but this separation is always uncomfortable and unfinished and it leads to the kind of flirting processes between state and religion that has characterized history in much of the Abrahamic world. At least in Europe and those great sways of Asia and Africa that are still under the ghostly spell of ancient Mesopotamia and Egypt this flirting takes various forms. One form of the process involves new states taking on, ready made, one of these politically detached religions issued from clearly different political entities. Rome was an example of the process. Imperial Rome became one of these centralized systems where political conquest led to the creation of a transcendental social representation of the state through making incomplete the transcendental social of subjects. Yet the transcendental construction never worked very well and when Rome got into even more trouble than usual, the system broke down. This led to the adoption of foreign and abandoned centre religions, therefore, ‘hungry’ for the recovery of their lost politico/transcendental social element, e.g. Judaism, other eastern religions and ultimately, one of the many forms of Christianity. Rome was therefore taking on the religious side of a centralized system from a collapsed tiny city-state as a late attempt at reorganization of a unified transcendental (Beard et al. 1998).

The process repeated itself. When, in the seventh century, the Franks began to develop centralized entities in western Europe, they picked up Christianity and, so to speak, ‘put it on’ with modifications to make it fit. One of the most spectacular moments was when Charlemagne in 800 invented a ritual that made him the Holy Roman Emperor with bits borrowed from the old testament, from Frankish rituals and of course, above all, from Roman rituals (Nelson 1987).

The other form of the relation between religion and the state, made necessary by their previous separation as a result of the collapse of a centralized unit, is for the religious bit to try to grow back its lost political undercarriage. Again and again, the popes tried. The Ayatollah Khomeini was more successful. Most of the movements that have been called millenarians try this sort of thing. Mormon history furnishes a particularly interesting example. Joseph Smith started the Mormon religion in the eastern USA for people who were heirs to a Christian religion that at many removes was heir to a long history of trouble between the religious-like pretensions of the state and the state-like pretensions of religion. However, the Mormons were in a place where the state was weak and, unusually, where the totalizing cosmological pretensions of post-state religion were strikingly incoherent, largely because they were meant to apply to a country not included in the cosmology of the Bible. So the Mormons put that to right by finding a Gospel that did mention the New World and its inhabitants and, in their creative enthusiasm, began to rebuild the political part of the destroyed transcendental entity. Not surprisingly, this annoyed the other state in Washington and they had to try to build it up in the desert, which, amazingly, they just about succeeded in doing. At the centre of this renewed unitary entity, where the transcendental social and religious were again to be an inseparable totality as in ancient Egypt or Mesopotamia, they built their temple: a temple that looks strikingly like a Ziggurat.

5. Conclusion

The point of these historical excursions is to suggest that the separation of religion from the transcendental social in general is, even in the places where it appears at first to exist, superficial and transient. In any case, this superficial phenomenon has occurred in human history only relatively recently.

It is this transcendental social in its totality that should be our focus. It is what distinguishes the human social from that of other closely related animals, such as chimpanzees. It is a unique characteristic and an essential part of human sociality, which, as often suggested, is the fundamental difference between humans and other anthropoids. An explanation of its occurrence cannot thus be in terms of a minor evolutionary adaptation, or misadaptation, as is suggested by Boyer-type theories.

Such a conclusion is negative, but it is possible to propose a more positive and fruitful one.

What the transcendental social requires is the ability to live very largely in imagination. We often act towards elders, kings, mothers, etc., not in terms of how they appear to the senses at any particular moment but as if they were something else: essential transcendental beings. Once we realize this omnipresence of the imaginary in the everyday, nothing special is left to explain concerning religion. What needs to be explained is the much more general question, how it is that we can act so much of the time towards visible people in terms of their invisible halo. The tool for this fundamental operation is the capacity for imagination. It is while searching for neurological evidence for the development of this capacity and of its social implications that we, in passing, will account for religious-like phenomena. Trying to understand how imagination can account for the transcendental social, and incidentally religion, is a quite different enterprise to accounting for the religious for itself in terms of modules, or core knowledge, which, in any case, we share with other primates. Unlike this, imagination does seem to distinguish us from chimpanzees and perhaps also distinguishes post-Upper Palaeolithic humans from their forebears.

A number of recent writers have suggested something along the same lines. In a book by Paul Harris about imagination, the author shows how the ability to engage spontaneously in pretend play begins very young and develops in a multitude of ways such as creating ‘imaginary friends’ and other forms of explicit make believe. Such imagination practice seems essential for normal human development. Nothing like that occurs in other species. Clearly, this capacity is necessary for engaging in the transcendental social as defined above, inevitably including the religious like. The selective advantage this form of sociality procures explains its evolutionary potential. It is central to human life. Harris suggests this centrality in an adventurous introduction when he notes that the first evidence for such a capacity is the cave paintings of Europe dating back to ca 40 000 years ago (Harris 2000). He might have gone a bit further back to what has been called the Upper Palaeolithic revolution, one feature of which was the first suggestion of transcendental roles found in grand burial.

Again, in a parallel argument, also taking empirical data on ontological development as its starting point, Hannes Rakoczy connects the imagination and the transcendental social even more explicitly (Rakoczy 2007). In that work, and that of his co-workers, this is referred to as ‘status functions’ but it is as yet little developed. However, the argument is strikingly similar to that proposed above and totally congruent. It does not however, like Harris, touch on the topic of religion, but according to my argument, this is inevitably subsumed under this type of discussion of the social.

To explain religion is therefore a fundamentally misguided enterprise. It is rather like trying to explain the function of headlights while ignoring what motorcars are like and for. What needs to be explained is the nature of human sociability, and then religion simply appears as an aspect of this that cannot stand alone. Unfortunately, the recent general discussion on social cognition does not succeed in doing the job that is needed to understand the transcendental social either. This is because, for the most part, it has considered the human social as an elaboration and an expansion of the type of social found in other animals, especially other primates (Dunbar 2004). This is useful but it obscures a fundamental difference between humans and others. Such an approach only pays attention to the transactional, or the ‘Machiavellian’ social, since that is what is shared by, for example, baboons and humans. It ignores the uniquely human transcendental social that represents a qualitative difference with other non-human socialities. What is essential to understand is the evolution of this specificity. Concentrating on that equally unique human capacity, imagination seems the most fruitful approach in that enterprise and, in passing, we will also account for religion since it is nothing special.