The old gods as a live possibility: on the rational feasibility of non-doxastic paganism

Abstract

Pagan revivalism is a growing trend in the contemporary religious landscape. Is it possible to be a neopagan without disregarding the demands of reason? While outright belief in the old gods seems out of the question, I argue that polytheism represents a live epistemic possibility, and that non-doxastic paganism is therefore a viable option. However, the rational, non-doxastic neopagan should only commit to general polytheism, rather than a detailed, specific pagan pantheon (such as the Greek or Old Norse). I also suggest that the ancient pagan conception of the divine as radically immanent must be rejected.

Introduction

After centuries of Christian hegemony in the West, pagan religion started to re-emerge in the twentieth century.¹ Today, growing pagan communities are a common feature of the religious landscape. In some European countries, neopagan organizations are beginning to have substantial cultural impact, like the Old Norse Ásatrúarfélagidh in Iceland (Blain and Wallis (2009)) or the Slavic revivalists Romuva in Lithuania (Senvaitytė (2018)). But what about the rational feasibility of this development? Is it possible to resurrect pre-Christian polytheistic religion while staying true to the demands of epistemic rationality?

The rationality of all religion is standardly questioned in the contemporary West. However, being a pagan seems more questionable than being, for example, a Christian or a Muslim. In our society, already the existence of God is a highly contested issue (for clarity, I use God with an upper-case G exclusively for the monotheist conception of the divine). The pagan revivalist has not one but several gods, gods whose existence no one has believed in for centuries and whose existence even Christians and Muslims sternly deny.

I am somewhat summarily going to assume that literal belief is out of reach here. It seems unrealistic to envision that an Old Norse revivalist could be in an epistemic position which allows for rational, outright belief that the Aesir gods literally exist, that there are extra-mundane realms like Valhalla, that one can work magic with runes, etc.

Fortunately, outright belief is not the only positive, truth-normed attitude one can take towards religious propositions. According to non-doxasticism, a realist religious stance need not be based upon belief, but only on some epistemically weaker attitude like hope or acceptance.² To be rational, a non-doxastic commitment requires that the subject regards the religious worldview as a desirable epistemic possibility in the relevant sense (Palmqvist (2021)). In what follows, I assess whether a non-doxastic commitment to paganism can be epistemically rational by considering whether the central idea of pagan polytheism could represent an epistemic possibility for a philosophically and scientifically informed contemporary subject.

While contemporary paganism has been investigated by sociologists and scholars of religion, it has received little attention from philosophers. This is regrettable since paganism contains ideas and perspectives which differ markedly from those associated with traditional monotheism, providing fresh challenges for philosophers of religion and an opportunity to broaden the field beyond its traditional limits. In this article, I take some first steps towards giving pagan ideas the philosophical attention they deserve.

In pagan studies,³ there is considerable disagreement over the term ‘paganism’. Some, like Michael York (2003), view paganism as a wide-ranging phenomenon, including most historical and present-day religions outside the Abrahamitic and Dharmic traditions. On the other side of the spectrum, Ethan Doyle White suggests a narrow definition according to which only pre-Christian European religion and its modern revival counts as proper paganism (White (2016)). The pagan revivalism I am interested in clearly connects with White's narrow definition, and I leave it to the reader to work out the implications of what I say for any other forms of paganism there might be.⁴

Before continuing, a methodological remark. Real-life religion is messy, with great diversity between groups and individuals concerning both religious practice and intellectual content. To deal with this, philosophers of religion usually concentrate on the central, most fundamental ideas of a religious tradition. Instead of discussing the rational feasibility of being a conservative Lutheran in contemporary Norway, we discuss whether perfect-being theism is a sustainable position. I follow this standard approach and assess the rationality not of real-life revivalist pagans, but of the central idea of pagan polytheism. For religion centred on gods, the question concerning the existence of these gods is surely the most fundamental for assessing its epistemic rationality. Obviously, there are other important pagan ideas as well, such as animism or corpo-spirituality,⁵ but it is unfortunately outside the scoop of this text to address them all.

In the upcoming section ‘Non-doxasticism and the notion of a live possibility’ the non-doxastic approach is presented in detail. In ‘What is a god and how should we understand polytheism?’ I clarify the notion of a god and argue for what I view as the most plausible interpretation of polytheism. It is followed by ‘Polytheism and the arguments concerning the existence of God’, where polytheism is assessed by comparing its epistemic status with that of perfect-being theism. In the last major section ‘Is a world permeated by the divine a live possibility?’ I address the special pagan conception of the divine in terms of radical immanency. The text ends with some concluding remarks.

Non-doxasticism and the notion of live possibility

Non-doxasticism is a truth-oriented, realist approach towards religion, which rejects the traditional assumption that religion must be based on belief to be rational. The central idea is that belief can be substituted by an epistemically weaker cognitive pro-attitude, such as acceptance (Alston (1996)), assumption (Howard-Snyder (2017)), or hope (Muyskens (1979)). The traditionalist Christian believes that ‘God exists’, whereas the non-doxasticist hopes or accepts this proposition. Can the pagan make a similar manoeuvre to secure a rational basis for her religious life?

A preliminary remark. Most accounts of non-doxasticism come in terms of faith. Non-doxastic faith is analysed as a composite attitude, where a cognitive pro-attitude takes the place traditionally held by belief. This way of formulating non-doxasticism might bother the pagan since the term ‘faith’ is heavily associated with Abrahamic religion. Therefore, it is important to see that non-doxasticism without faith is a real possibility. In James Muyskens's (1979) early account, hope is viewed as an alternative to faith rather than a constituent in its analysis. This suggests that a cognitive pro-attitude can function as the cognitive base of religion on its own. Also, I have previously argued that non-doxastic faith differs from what we ordinarily mean by ‘faith’, and that it must be considered a technical notion (Palmqvist (2022)). If so, it would be unobjectionable if the pagan would prefer to call this technical notion something else.

While details vary between accounts, the basic idea of non-doxasticism can be formally stated as two necessary conditions; for S to have an epistemically rational non-doxastic attitude towards some proposition p, the following two conditions must be met:

ND1: S desires the truth of p or judges the truth of p to be an overall good thing.

ND2: S rationally believes the truth of p to be epistemically possible and neither believes nor disbelieves p.⁶

I take for granted that pagan polytheism can satisfy ND1. Paganism holds many attractive features, like a strong reverence for nature and a down-to-earth religiosity. For people dissatisfied with Abrahamic religion, the pagan gods offer a less hierarchical, less moralistic, and less patriarchal view of the divine (York (2009)). Paganism might not be for everyone, but it seems reasonable to suppose that these features make it a desirable option for a considerable number of people.

ND2 is where the challenge lies for paganism. Can pagan polytheism be considered an epistemic possibility in the relevant sense, namely the sense which makes a religious life possible?

Some proponents of non-doxasticism prefer a very wide definition of epistemic possibility, and only demand that p is neither known nor justifiably believed to be false (Schellenberg (2009), 8). Others opt for a narrower sense, requiring p to have a ‘non-negligible probability of being true’ (McKaughan (2013), 13). Since a religious life requires a notion of epistemic possibility which can serve as a rational ground for action, we have good reasons to prefer the narrow sense. Consider the following example. My sister moves around a lot, and we have lost contact. I do not have any outright beliefs concerning her whereabouts. Last time we spoke, two years ago, she lived in Chicago, but she was dissatisfied with her current work-situation and planned to move to New York if things did not soon improve. From my perspective, it seems reasonable to assume that she is now in either Chicago or New York. These are epistemic possibilities in the narrow sense. Of course, much can happen in two years and my sister can be literally anywhere in the world. Therefore, any proposition saying that my sister is in a geographical location represents an epistemic possibility in the wide sense. I do not really know that she has not moved to, say, the countryside of Bulgaria, do I?

Only epistemic possibilities in the narrow sense allow for rational action. If I cannot get hold of my sister, I could search for her in either Chicago or New York but not in Bulgaria. It would not be rationally justifiable to go looking for her in random places, even if her being in any random place represents an epistemic possibility in the wide sense. In what follows, I concentrate on epistemic possibility in the narrow sense, and I henceforth use the term ‘live possibility’ to avoid conflation with the wide sense.

Live possibility: P represents a live possibility for S iff S rationally believes that p has a non-negligible probability of being true.⁷

Given these considerations and my focus on polytheism, the question regarding the rationality of being a contemporary pagan can be restated as follows: can pagan polytheism represent a live possibility for scientifically informed members of contemporary Western society?

Much like traditional theism, a polytheistic worldview concerns literally everything and assessing its probability in relation to all available evidence is a huge endeavour. Trying to keep things manageable I only aim at a first approximation. I use two routes to assess whether pagan polytheism could be considered a live possibility. First, I compare the epistemic status of polytheism with that of perfect-being theism. Taking for granted that perfect-being theism represents a live possibility, I assess how polytheism compares when we consider some important evidence concerning the existence of God. If the evidence supports polytheism at least equally well, it too should be considered a live possibility. Second, I assess whether the distinctively pagan notion of the divine constitutes an obstacle for considering pagan polytheism a live possibility.

An objector might be concerned that I move too fast in claiming that perfect-being theism represents a live possibility. However, the very fact that a philosophical view is taken up for serious debate strongly indicates that it is an epistemic possibility to consider. Presenting an argument for p or ¬p makes little sense if we do not regard p as a live possibility. Since philosophy of religion is full of arguments for and against the existence of God, it seems safe to suppose that perfect-being theism represents a live possibility or at least that it is viewed as such by a majority of philosophers of religion.⁸

Before attempting to assess whether pagan polytheism constitutes a live possibility, it is important to get clear on exactly what we mean by ‘polytheism’. That is the topic of the upcoming section.

What is a god and how should we understand polytheism?

In itself, the term ‘polytheism’ only implies the denial of atheism and monotheism: there are gods, and they are more than one. The further meaning of ‘polytheism’ is dependent on what we mean by a ‘god’. It seems that polytheists cannot mean the same thing with ‘god’ as monotheists do, since arguably there is no conceptual space for more than one omnipotent perfect being.

Different proponents of polytheism tend to mean different things with the word ‘god’. There is no ‘orthodox view’ comparable to the perfect-being theism of Abrahamic religion. This should come as no surprise given the cultural and historical diversity of polytheistic religion. In order not to get stuck in the philosophical question of how to define a god, I employ an abstract concept of godhood wide enough to accommodate most existing notions:

Polytheist notion of a god: a powerful non-human agent with the ability to influence our world in significant ways.⁹

This concept might seem overly broad since it includes angels and demons, but polytheism does not always come with a strict boundary between gods and other powerful entities. It might also be objected that we should think of gods in explicitly supernatural terms, to exclude science fiction options such as powerful AIs or hyper-advanced aliens. However, many pagans staunchly deny that their gods are supernatural, as they view supernaturalism as incompatible with their radically immanent understanding of godhood (York (2009)). I find it hard to formulate a more precise notion which does not exclude important views or implement artificial boundaries. As pointed out by Timothy Williamson (2000, 211), ‘sometimes the best policy is to go ahead and theorize with a vague but powerful notion’. That course of action will be pursued here.

In polytheistic religion, the gods are usually pictured as somehow organized. Individual gods are associated with some aspect of reality over which they are considered responsible, even though responsibilities might overlap and change as a religion develops. While not always benevolent, the gods supposedly care for the aspects of the world which they oversee. Together, the gods have dominion over our world (even though there might be other realms outside their control). Let us call such an organized collective of gods a pantheon:

Pantheon: an organized collective of gods in control of our world, where individual gods are associated with and in some sense control specific aspects of it.

Clearly, there might be gods (as previously defined) without there being a pantheon, like gods who do not take part in governing the world, or gods who are not associated with some aspect of reality. However, polytheistic religion almost always comes with their gods organized in a pantheon, and since a pantheon is religiously relevant in a way other alternatives are not, I will leave such possibilities aside.

When we ask whether polytheism is a live possibility, we must carefully consider what it is that we are asking. Are we only asking whether it could be the case that any pantheon exist? Or is the question whether a specific pantheon, like the Old Norse, could exist? The first question asks whether polytheism could be true in a general sense, the second in a specific sense.

General polytheism: a pantheon exists.

Specific polytheism: a specific pantheon P exists.

General polytheism might feel a bit dry and abstract as compared to specific polytheism, but there are strong theoretical reasons for preferring the former. The truth of general polytheism is entailed by the truth of any conceivable specific polytheism so that if any pantheon exists, general polytheism will be true. It is the most probable view, except in the highly unlikely scenario where only one specific pantheon represents an epistemic possibility.¹⁰

The specific polytheist also faces what might be called the intra-polytheist problem of religious diversity, which the general polytheist escapes. She needs to handle the existence of other pantheons in other cultures, pantheons which differ in significant ways from her own. While the Indo-European pantheons like the Greek or Old Norse share some similarity, pantheons like the Egyptian or the Aztec differ markedly. If the gods are real, how come other cultures worship other gods? What reasons could we have for accepting the Norse pantheon while rejecting the Aztec? I can think of two possible ways in which the specific polytheist might handle this objection, but both seem unsatisfactory.

First, the polytheist might claim that her own pantheon is the only pantheon which represents a live possibility. If so, she needs to explain how there can be such a major epistemic asymmetry between different polytheistic traditions that only one deserves to be taken seriously. Monotheists claiming to be in a superior epistemic position as compared to other religions standardly explain the asymmetry with the help of revelation, but revelation is usually absent from polytheistic religion, and it is hard to see what could take its place.

The second way to handle the problem is to suggest that all pantheons exist simultaneously, a view advanced by some contemporary pagan thinkers (Dillon (2015)). There are at least two major problems with this ecumenical solution. First, the worship of a pantheon tends to rise and fall with the culture it is part of, and how can the polytheist account for that? If Zeus literally exists, how come only the ancient Greeks knew about him, and what has Zeus been doing since the Greeks converted to Christianity? I think that the only way to make the idea that pantheons rise and fall with their civilizations plausible is to suggest that gods only become powerful if they are worshipped.¹¹ While this view of the gods is surely possible, it seems highly unorthodox, and I suggest we choose general polytheism to avoid having to rely on it.

A further problem. The gods personify universal forces, but how can gods personifying the same force be distinct? If Zeus, Jupiter, and Thor all represent the universal force of thunder and lightning, what reasons could there be for not identifying these gods with one another? Is there even conceptual space for more than one sky-god? The ecumenical polytheist could perhaps avoid this problem by claiming that the gods control the forces they are associated with rather than personify them. However, this seems hard to reconcile with the immanent conception of divinity common to paganism. Thor not only controls the thunderstorm, but he is also in the thunderstorm. This problem too can be sidestepped by choosing general polytheism.

My focus on general polytheism raises the question about how we should treat the mythological descriptions of the gods found in specific forms of polytheism. I suggest that rather than taking these pantheons literally, we should understand them as culturally specific ways to apprehend and relate to the same gods. The incoherence between different polytheistic outlooks means that they cannot all be correct descriptions of the gods, and since we do not know which (if any) specific pantheon has got it right, we should remain uncommitted concerning the details.

This does not mean that we must restrict ourselves to general polytheism in worship. Religion cannot be practised without its juicy details. What is required is rather an awareness that the details are much less likely to be true than the core ideas. The non-doxastic general polytheist will therefore make use of a specific pantheon and its details in worship, perhaps treating these details as useful fictions.¹² This is very much like how the historically and scientifically informed Christian celebrates the Incarnation at Christmas even though many of the details of the Christmas Gospel are presumably literally false.

General polytheism fits well with how the gods were usually apprehended in antiquity. Polytheistic religion tends to be inclusivist, incorporating foreign gods rather than rejecting them. The Romans are famous for identifying gods of other cultures with their own. They identified Jupiter with Zeus and considered Odin ‘the German Mercury’. Also worth mentioning is the ‘syncretistic’ nature of Hellenistic religion in late antiquity. A prime example is the cult of Cybele, the Magna Mater, who inherited the traits of many different mother-goddesses. All this speaks of a willingness to look beyond concrete details to the larger picture.

General polytheism is the most plausible way to understand polytheism and the best candidate when considering whether polytheism can be a live possibility. I will henceforth use the term ‘polytheism’ to refer only to the general kind. However, concentrating on general polytheism does not imply that the details of concrete myths are irrelevant for present purposes. They provide good examples of how polytheism can be spelled out and I frequently use concrete polytheism in my examples.

Polytheism and the arguments concerning the existence of God

Evidential arguments concerning the existence of God centre on some specific features of the world. Pro-theistic arguments focus on features we would expect in a world created and sustained by a perfect being, and anti-theistic arguments on features which seem incoherent with the theistic picture. What can these arguments tell us regarding the epistemic status of polytheism, as compared to that of perfect-being theism? While I focus on the teleological argument in what follows, I broaden the picture by also reflecting on the argument from evil and an argument about agency detection relying on results from cognitive science of religion.

The teleological argument

The teleological argument tries to reach the conclusion that God exists from the fact that the natural world is ordered. The argument is equally important for polytheism and monotheism. All religions centred on gods tend to agree that one or several deities are responsible for ordering the world. While polytheism standardly lacks the idea of creation ex nihilo, the gods are ‘designers’ of the current cosmic order, and the teleological argument is about exactly that: cosmic design.¹³

The idea of assessing whether polytheism is a live possibility by considering how well it is supported by an argument for perfect-being theism might seem misguided. How could an argument for the existence of a God who is omnipotent, omniscient, and perfectly good be equally relevant for polytheism? The catch is that teleological arguments do not directly support the claim that God exists, but only the weaker claim that the universe was designed. Following Alexander R. Pruss (2009, 91), I will refer to this as the ‘gap-problem’.¹⁴

Because of the gap-problem, the teleological argument supports perfect-being theism and polytheism equally well. As pointed out by Plantinga:

In arguing that God exists, the theist typically means to argue for a proposition equivalent to a conjunction of which the following are conjuncts:

1. (a) The universe is designed.

2. (b) The universe was designed by exactly one person.

3. (c) The universe was created ex nihilo.

4. (d) The universe was created by the person who designed it.

5. (e) The creator of the universe is omniscient, omnipotent, and perfectly good.

6. (f) The creator of the universe is an eternal spirit, without body, and in no way dependent upon physical objects.

Now we can put the objection as follows: the teleological argument may have some evidence (not very strong, perhaps, but not completely negligible) for (a); but with respect to (b)–(f) our evidence is completely ambiguous. (Plantinga (1990), 109)

Polytheism generally denies (c)–(f) and often (b) as well, while perfect-being theists tend to affirm (a)–(f). Since the argument only supports (a), it gives us no ground to prefer perfect-being theism over polytheism.

According to the Old Norse myth of creation (as it is being retold in the Prose Edda), the current world was created by Odin and his brothers from the body of the slain giant Ymir. What exist from eternity are not the gods, but the realms of fire (Muspelheim) and ice (Jotunheim). Odin and his brothers are not even the first gods, but grandchildren of the first god Buri. Comparable stories are found in most polytheistic religion. If one were to strip them of mythological content and rephrase them into a more philosophically informed view, this view would be quite different from the account championed by perfect-being theists. The lesson to learn from Plantinga is that such a view would be equally well supported by the teleological argument.

Richard Swinburne has explicitly argued that teleological reasoning promotes monotheism over polytheism. His first argument attempts to handle the gap-problem by inference to the best explanations. He suggests that by invoking the theoretical virtue of simplicity, we can bridge the gap by claiming that God is the best explanation of why the universe seems designed (Swinburne (2004)). Obviously, the God of perfect-being theism is numerically simpler than a pantheistic pantheon. However, there are other theoretical virtues than simplicity. One such virtue is intelligibility. It might be argued that since the divine attributes held by God (like omnipotence) are far beyond our understanding, we should prefer a collective of gods with less extreme attributes, since such gods would be more comprehensible. This was suggested in relation to the teleological argument already by David Hume:

Why may not several deities combine in contriving and framing the world? . . . By sharing the work among several, we may so much further limit the attributes of each, and get rid of that extensive power and knowledge, which must be supposed in one deity. (Hume (2006), 37)

Hume suggests that if we postulate a collective of designers, we do not need to postulate the extreme divine attributes a single designer requires, which makes the polytheistic theory more intelligible. Hume's suggestion demonstrates that different theoretical virtues might point in different directions. If it is an open question whether monotheism or polytheism is the best explanation in terms of theoretical virtue, Swinburne's solution to the gap-problem fails. Without such a solution, it is hard to see how teleological reasoning could support perfect-being theism over polytheism.

Swinburne's second argument is based on the regularity of the universe. Noticing that our universe is uniform, with the same physical laws holding everywhere, he suggests that we should expect otherwise if the universe was the product of a designer collective: ‘If there were more than one deity responsible for the order or the universe, we would expect to see characteristic marks of the handiwork of different deities in different parts of the universe’ (Swinburne (2004), 147). Swinburne's argument fails to convince. There are many examples of when humans work together on great projects where you cannot trace the individual handiwork of the people involved. The creation of virtual worlds in computer games might be an interesting analogy. Big-budget computer games are generally made by studios with hundreds of employees. Anyone who has played games with vast open worlds know that these worlds are usually internally consistent, and that it is simply not the case that their hundreds of creators have left detectable individual marks. If humans can do this on a small and artificial scale, what reason could there be for thinking that a collective of gods could not create an orderly universe (since, if they exist, the gods will have abilities far beyond our own)?¹⁵

The argument from evil

The question why an omnipotent and perfectly good God would permit the evils we see in the world has been an intellectual challenge to perfect-being theism for millennia. Contemporary versions of the argument are almost always evidential, treating the evils in the world (often in terms of meaningless suffering) as evidence against the existence of God.

These arguments usually ask what kind of world we should expect a perfect being to create. A common answer is that we should expect a perfect being to create a perfect world, or at least the best of all possible worlds. While Leibniz famously argued that our world is in fact the best of all possible worlds, few agree with that conclusion. Arguments from evil can therefore be understood to rest on the great discrepancy between the kind of world we would expect from a perfect being and the world we actually live in. This discrepancy is taken as evidence against the truth of perfect-being theism.

What kind of world should we expect if polytheism is true? Consider some key features of traditional polytheistic worldviews. First, the gods are not perfect beings. Not only is their power severely limited as compared to an omnipotent perfect being, but neither are they perfectly good. Trickster gods such as Loki might even seem evil, or at least morally ambiguous. Second, the gods have their own agendas and their own struggles. They do not always agree and might even oppose each other. Third, the gods are not the only cosmic forces in existence. There are forces of chaos, enemies of the gods opposing their work.

What world should we expect if polytheism is true? Presumably one very much like the actual. An ordered world (ordered by the gods) in which there is considerable ambiguity (because the gods have different agendas) and in which there is evil and suffering (because of the forces of chaos and because the gods lack omnipotence and cannot simply stop humans from doing evil things). Unlike perfect-being theism, polytheism does not suggest that we should expect the gods to deliver a better world than the one we got.

Agency detection and polytheism

According to an important theory from cognitive science of religion, our cognitive systems contain a hypersensitive agency detection device (‘HADD’ for short).¹⁶ We are exceedingly good at detecting agents and agency, but sometimes we also detect agency when there is none. The oversensitivity of HADD is explained by evolutionary factors:

Our agency detection system sometimes leads us to perceive false positives, observing an agent where none is present. The evolutionary rationale for this is that a false positive is less costly than a false negative, as the latter can result in a failure to detect a dangerous predator, a prey animal, or a potential mate, and the former only results in a small waste of energy. (De Cruz and De Smedt (2015), 191–192)

As might be expected, the HADD-theory has sparked controversy between theists and atheists.

Philosophers like Wilkins and Griffiths (2013) think that HADD-theory can be used in a debunking argument against theistic belief. The idea is that an oversensitive HADD leads us to detect agency in the natural world, when in fact there is none. When we think we detect the agency of God or gods in the workings of the world, it is just our HADD playing us a trick. Insofar as our beliefs in supernatural agents build on these erroneous observations of agency, these beliefs are false.

As has been pointed out by Helen De Cruz and Johan De Smedt ((2015), 193) the debunking argument begs the question against theism. Only if we presuppose that the HADD is oversensitive when it detects agency in nature do we have any reason to reject theistic beliefs building on such detection. The theist will naturally hold the contrasting view that the HADD is reliable when detecting theistic agency in nature. Since the HADD is both truth-oriented and oversensitive, it seems that we need independent evidence to distinguish between the true and false beliefs it generates. When it comes to agency in nature, such evidence is not easy to find.

But even though the HADD cannot be used to argue for atheism over theism (or vice versa), it can still give us reason to prefer polytheism over perfect being monotheism. If any kind of theism is correct, it seems reasonable to suppose that the HADD is reliable in letting us detect divine agency in the natural world, even though no agent is physically detectable. The perfect-being theist will think that all such instances are detections of the one true God, while the polytheist will think that there are many different gods at work in nature.

Here, the gap-problem re-emerges for the perfect-being theist. There is a huge theoretical gap that has to be filled to go from agency in nature to the God of perfect-being theism. Agency in nature does not imply omnipotence, omniscience, or perfect goodness. Even the idea that all alleged agency in nature points to one single agent seems a substantial addition to what the HADD could tell us. By contrast, the gods of polytheism, which are plural and lack the omni-attributes, require much less in terms of added theory.

To see this, consider the polytheist who sees Poseidon in the tsunami and Venus in the love affair of a young couple. The polytheist does not have to add powers which go far beyond what she observes – the tsunami is a direct expression of Poseidon's might and the love affair an expression of Venus's. By contrast, a perfect-being theist observing the same agency in the tsunami and the love affair will claim that these are expressions of God's endless power. Likewise, the perfect-being theist will claim that the agencies in the tsunami and the love affair are really one and the same. These are both significant theoretical steps to take, steps not warranted by simply observing agency in nature. What we actually ‘perceive’ when we detect agency in nature is therefore more in line with the polytheistic idea that nature is full of powerful agents, than with the idea that it is controlled by a singular perfect being.

It is also of significance that our world is an ambiguous one, as discussed above in relation to the argument of evil. Once again, think of the polytheist who attributes the love affair to Venus and the tsunami to Poseidon, and the perfect-being theist who attributes both phenomena to God. Imagine that the young couple are killed in the tsunami on their first date, bringing the love affair to a tragic and premature end. The polytheist will interpret this as a conflict between the gods: the young couple favoured by Venus was obviously in Poseidon's way. The perfect-being theist has a much harder task: she must explain how a love affair which tragically ended before anything tangible came out of it could be caused by the same agent which caused the tsunami. Since the polytheist needs no such extra theories, it once again seems that the way she interprets what she ‘learns’ from the HADD requires much less in terms of added theory.

If our ability to detect agency in nature is reliable, it seems to directly support some kind of polytheism.¹⁷ Perfect-being theism contains significant theoretical additions not warranted by simple agency-detection. To go from agency detection to monotheism, we need substantial arguments showing that the seemingly conflicting agencies we detect are in fact one singular agent.

I have considered the question of what support polytheism can receive from some important evidential arguments concerning perfect-being theism. My conclusion is that teleological reasoning does not support perfect-being theism over polytheism, and that the arguments from evil and agent-detection even favour polytheism. However, it would be an overly hasty conclusion to suggest that polytheism is in a better position than perfect-being theism all things considered. There are many further evidential arguments to take into account before any final conclusion can be drawn, and there are also theistic arguments which are purely metaphysical. My only suggestion is that the present considerations should be enough to regard general polytheism as a live possibility.

Is a world permeated by the divine a live possibility?

I have argued that general polytheism represents a live possibility. However, paganism comes with a notion of the divine which differs markedly from other religions. In this section I consider whether this specifically pagan notion of godhood also represents a live possibility.

While perfect-being theism comes with a transcendent view of the divine, the opposite is true of pagan polytheism. Here follows a vivid presentation of the pagan view of the divine, formulated by the neopagans in the UK Pagan Council:

Pagans understand divinity to be imminent [sic!], it is woven through every aspect of the living earth. Pagan worship is mainly concerned with the connection to, and the honouring of immanent divinity. We believe that the gods not only live above us but, more importantly, they live all around us and inside us. The divine is in all things, from the smallest pebble to the tallest tree. (https://www.ukpagancouncil.org/about-paganism/)

Terms like pantheism and immanence do not adequately capture this pagan conception of divinity. Pagans hold that the divine resides in the natural world. For lack of better terms, let us call this view ‘radical immanency’.

The conjunction of polytheism and radical immanency seems highly problematic. It is one thing to postulate powerful non-human agents which are transcendent or situated in other planes of existence, like the heavenly angels of Abrahamic religion. It is another thing altogether to postulate that they exist within in our world.

The idea that the gods exist in the world strongly suggests that their existence is somehow physical. Many pagans therefore conceive of their gods in terms of ‘energy complexes’. One of few detailed accounts is due to York (2009). His attempt to articulate how the gods exist builds on complexity theory:

Pagan gods may themselves be spontaneous self-organizations from a feed-back loop between the attentions of devotee or worshipper and the values perceived to be inherent within the physical and cultural environment . . . It is not inconceivable to the pagan that her deities assume an operative reality that surpasses and is independent of the matrix of realities from which any specific theomorphic figure has taken birth. (York (2009), 299)

Spontaneous self-organization is when the sum unpredictably becomes larger than the parts, and a feedback loop is when that which is generated is fed back into the source. The focus of complexity theory is when unprecedented patterns like spontaneous self-organization emerge as result of a feedback process (York (2009), 298). York suggests that the gods are beings of energy who spontaneously emerge as a result of a feedback process between worship and inherent values. This view seems highly unfeasible.

Ignoring issues like how exactly worship and values are supposed to constitute a feedback loop, York's view contains at least two highly controversial ideas: that the gods come into existence because they are worshipped, and that they are beings of energy.

The idea that gods come into existence by being worshipped seems questionable because we know of no other examples in which conscious agency spontaneously arises as the result of collective human action. When collective action leads to the emergence of entities larger than the sum of their parts, like an economy or a social culture (examples mentioned by York (ibid., 299)), these are social realities without agency or consciousness. Granted, we do not know exactly how consciousness and agency arise in humans and animals, but collective action just seems like the wrong kind of process.

Perhaps anticipating this objection, York claims that the process is too complex for prediction, and that only in hindsight, by ‘retrodiction’, can one understand what has happened (ibid., 298). I find the idea of an explanation that is available only in retrospect unintelligible. If we lack any reason to think that a certain process could bring about a god, so that we cannot assign any meaningful positive probability to such an outcome, we also have no reason to think that we could ‘retrodictionally’ explain the existence of a god by this process. It is because process a might result in b that we can explain the occurrence of b with the help of a. If there is no reason to think that a might result in b, we will only see covariation and not causation when b appears in connection to a.

The idea that the gods are beings of energy is an attempt to fit radically immanent gods into our current worldview. However, if gods are beings of energy, it seems an inescapable conclusion that they are empirical phenomena. That is problematic, because there is little room in our current worldview for empirical gods. Had there been powerful beings of energy influencing our world, scientists would surely have found them by now. If we accept radical immanency of the divine, I cannot see how we could escape the conclusion that the gods are empirical phenomena, and the existence of such gods is not a live possibility according to our current worldview.

To see why the original pagan view of the divine is out of reach for us, it might be helpful to consider Charles Taylor's (2007) analysis of the secularization. While the secularization itself mostly means that religion has become a personal matter and religious belief optional (Taylor (2007), 3), it is the accompanying process of disenchantment which has made the pagan view implausible.

The term ‘disenchantment’ was originally coined by Weber. It denotes the process in which the old view of the world as infused with religious meaning and sacredness has given way to the rationalized, mechanical picture promoted by science. According to Taylor, our view of the world we live in, which he calls the ‘immanent frame’, is now completely disenchanted. That does not mean that this world is all there is – the frame is open to the possibility that there might also be a transcendent, divine reality (ibid., 539–549). However, it is not open for the possibility of the divine existing within the immanent frame. Our world has been irreversibly disenchanted. Just as we cannot go back in time scientifically to a pre-Copernican era where the Ptolemaic view of cosmos was a live possibility, we cannot go back religiously and treat radical immanency as a live possibility.

The pagan view of the divine as radically immanent cannot be reconciled with the current worldview. What remains to be considered is therefore a reformed kind of paganism which rejects radical immanency and combines general polytheism with a view of the divine which locates the gods outside the immanent frame. The Christian God is often considered both transcendent and immanent, and perhaps the Christian idea of the Holy Spirit as working within the world while not being part of it could be a suitable model for thinking about the pagan gods.

Concluding remarks

Is it possible to be a neopagan while staying true to the demands of reason? Assuming that non-doxasticism is the best truth-normed approach available, I recast the original question as the question whether pagan polytheism constitutes a live possibility.

I argue that neopagans should adopt general polytheism rather than committing to the existence of some specific pantheon, though the mythological picture of the gods might be treated as a useful fiction facilitating pagan worship. If perfect-being theism constitutes a live possibility, so does general polytheism. Polytheism can explain the design in the world equally well as perfect-being theism. It is more consistent with the existence of evil and the results from cognitive science of religion concerning agency-detection.

Pagans traditionally view their gods as existing within the world, but this view of the divine as radically immanent is not a live possibility in the disenchanted contemporary world. The neopagan therefore needs to combine general polytheism with a more transcendent view of the divine. Such a ‘reformed’ pagan polytheism constitutes a live possibility to which one might commit non-doxastically without flaunting the demands of reasons.

However, the fact that it is rationally permissible to commit non-doxastically to paganism only means that we may do so. It does not necessarily mean that we should adopt non-doxastic paganism all things considered. On non-doxasticism, religion is voluntary, and there is an ongoing debate concerning which religious view a pro-religious agnostic should prefer.¹⁸ I have suggested that pagan polytheism constitutes a desirable religious view, but is its truth desirable enough when compared to the relevant alternatives? Can it contend with views such as perfect-being theism or ultimism in this regard? That is the next major question which needs to be addressed to determine the rationality of being a non-doxastic pagan.