All at once, the man writes, God was present, penetrating him with goodness and power, filling his heart with ecstasy. The man describes perceiving God, not with any of his bodily senses, but with his consciousness. God did not appear in any specific form or determinate location. God had no colour, smell or taste. God was invisible. And yet no one, not even Moses on Mt Sinai, could experience a more intimate connection than he had with God in that moment. This man was one of many subjects whose reports appear in Williams James's 1923 book, The Varieties of Religious Experience. Although James's subjects came from all over the world, and from various religious traditions, their accounts were strikingly uniform. The experiences were typically ineffable: subjects struggled to find words that adequately captured what they felt or saw. The episodes were transient – rarely lasting more than a half hour – and passive, as if the subjects were in the grip of a power higher than themselves. Most significantly, the experiences had what James called a noetic quality: they seemed to their subjects to be revelations, full of significance and insight into ‘other orders of truth’.

James boldly proposed to take his subjects’ accounts at face value: that religious experiences may offer ‘windows through which the mind looks out upon a more extensive and inclusive world’. Some contemporary philosophers of religion have taken up James's proposal. They suggest that some religious experiences may be direct perceptions of God and can thereby count as evidence for religious belief in the same way that ordinary sensory experiences can justify beliefs about a perceiver's immediate environment. To see what they have in mind, suppose I am looking at a table in front of me. My eyes are open; the lighting is normal. I see a bowl of fruit. This experience, we think, gives me reason to believe that a bowl of fruit is on the table. I know there is a bowl of fruit on the table on the basis of my experience. In the same way, these philosophers suggest, a religious experience can give people reason to believe that God exists. It provides them with warrant for their religious beliefs.

Others are more sceptical. They object that religious experiences cannot rationally support religious belief unless certain alternatives are first ruled out. What if these episodes are the result of a seizure, schizophrenia or some other neurological disorder? What if they are not confrontations with the supernatural, but mere hallucinations? Hallucinations do not give us reason to believe in much of anything, much less a divine being. It seems rash to allow religious experiences to serve as a believer's evidence for the otherworldly without first eliminating the more worldly possibilities. At the very least, we must show that the experience is more likely to have a supernatural cause than a natural one. But that is a difficult case to make. When faced with different possible causes of a phenomenon, we ordinarily rely on past correlations between cause and effect. For example, if the street is wet, I typically infer that it has rained recently, not that someone has watered the street with a garden hose. This is because I have observed a long correlation between wet streets and rain, and I have rarely if ever seen people water the streets. But in the case of religious experience, we cannot rely on past correlations between religious experience and the supernatural, since it is the very existence of the supernatural that is in question in the first place.

Some defenders of James point out that we face similar problems in the case of ordinary sensory experience. Consider again my sensory experience of the bowl of fruit on the table in front of me. As Descartes famously argues in the opening pages of the Meditations, it is possible that I am dreaming, asleep in bed. Or that my experience is caused, not by a bowl of fruit, but by an evil demon intent on deceiving me: the entire material world is the product of a lifetime of hallucinations. Contemporary external world sceptics suggest that we may be disembodied brains floating in a vat of nutrients, our experiences the result of a network of sophisticated computers stimulating our brains. Unless we can rule out such scenarios, these sceptics argue, our sensory experiences cannot justify any beliefs about the world around us. We must show that the more likely cause of our sensory experience is not an evil demon or a brain stimulating computer, but ordinary objects in the mind-independent world. But to show this, we cannot simply appeal to correlations we have observed between the mind-independent world and sensory experience. For that presupposes that we already have knowledge of the mind-independent world, and this is exactly what the sceptic is questioning. We seem to be in a bind.

One way out of this bind is to adopt a view about perceptual justification called ‘Dogmatism’, a term first coined by Jim Pryor in his paper ‘The Skeptic and the Dogmatist’. The Dogmatist claims that perceptual experience gives us immediate prima facie justification for belief. On this view, I do not need first to rule out the brain in a vat hypothesis before my experience of the bowl of fruit can justify my belief that there is a bowl of fruit in front of me. The experience immediately justifies the belief, unless I have some positive reason to suspect that something is awry. That is, experiences give us defeasible justification of our beliefs: they justify our beliefs as long as there are no defeaters. The advantage of Dogmatism is that it shifts the burden of proof back to the sceptic. There is no need to prove to the sceptic that we are not brains in a vat before we claim to have knowledge about the world around us. Instead, the burden of proof is on the sceptic to give us evidence to suggest that things are not as they seem.

When it comes to defeaters, there are two types to look out for: undercutting defeaters and rebutting defeaters. As the philosopher John Pollock characterizes the distinction, an undercutting defeater is evidence that the experience in question is untrustworthy. Suppose I see what looks like a pink piece of paper in front of me. If I also had evidence that someone has replaced all the light fixtures in the room with red-tinted light bulbs, this would be an undercutting defeater. A rebutting defeater, on the other hand, is evidence that the belief in question is false. For example, suppose I have an experience in which it appears that my grandfather is standing in the doorway. The knowledge that my grandfather passed away twenty years ago would be a rebutting defeater. We occasionally come across defeaters of both types that will keep us from taking specific experiences at face value. But we do not have any reason to think we are brains in a vat, or at the mercy of a Cartesian evil demon. And so the Dogmatist would say that our experiences, by and large, justify our everyday beliefs about the world around us.

How would this picture of perceptual justification apply to the case of religious experience? We have been considering the religious sceptic's demand that we first rule out alternative scenarios before we accept the evidential force of a religious experience. In response, we might say that, if Dogmatism is true, a religious experience can supply immediate justification for a person's religious belief in the absence of defeaters. The burden of proof shifts to the (religious) sceptic to give us a positive reason for thinking that religious experiences are untrustworthy or that the beliefs that result from such experiences are false. Has the subject taken any hallucinogenic drugs prior to the episode? Do we have reason to think that the subject is epileptic? Are there any signs of schizophrenia? These would be undercutting defeaters, leading us to doubt that the experience is trustworthy. We would have a rebutting defeater if we had evidence that suggests that the belief based on the experience is, in fact, false. For example, is there a conflict between science and religion? Does the existence of religious disagreement undermine the justification for belief in any specific religion? Is the conception of God – understood as a maximally perfect being – even coherent? Are the perfections commonly attributed to such a God – such as omnipotence, omniscience and benevolence – compatible with one another? Is the existence of a benevolent and all-powerful God compatible with the vast amount of suffering in the world?

Now we can argue about whether these actually are defeaters. After all, some religious traditions believe hallucinogenic drugs offer a gateway to the spiritual realm. And religious believers, unconvinced by the standard arguments against the existence of God, will insist that there are no rebutting defeaters for their beliefs. But the fact that these are plausibly defeaters makes the case of religious experience quite different from the case of ordinary perceptual experience. We have no reason to suspect that our ordinary experience is wholly untrustworthy or that our perceptual beliefs are largely false. We have no evidence for the hypothesis that we are brains in a vat or victims of Descartes's evil demon. As such, Dogmatism is a plausible model for the justification of our everyday perceptual beliefs. The existence of the supernatural is more controversial, and so, arguably, experiences that purportedly support religious beliefs deserve greater scrutiny. For this reason, James's case would be better served by a model of perceptual justification that did not require the absence of defeaters.

One such model is Disjunctivism, a theory which perhaps has the best chance of doing justice to the claim that religious experience can provide warrant for religious belief. Disjunctivism, like Dogmatism, was developed as a response to scepticism about the external world. The Disjunctivist argues that the sceptical argument gets off the ground because it takes for granted a certain traditional conception of perceptual experience – a conception that is tempting, but ultimately misguided. Consider yet again my experience of the bowl of fruit. We can distinguish between the good case and the bad case. In the good case, my senses are functioning normally and the bowl of fruit is actually before my eyes. In the bad case, I am hallucinating and it falsely appears to me as if there were a bowl of fruit in front of me. Now it is at least possible for the experience I have in the bad case to be indistinguishable from the experience I have in the good case. While actual hallucinations are not typically indistinguishable from veridical experiences, we can at least imagine a hallucination that is so realistic that one could easily be taken in. The traditional view concludes from this that all we are ever directly aware of in experience is what is common to the good case and the (indistinguishable) bad case. And since, in the bad case, there is no bowl of fruit in front of me, we have to say that, even in the good case, I am not directly experiencing a bowl of fruit. Rather, what I directly experience is a mental image – or ‘sense datum’ – from which I infer that there is a bowl of fruit in front of me. The external world sceptic gets a foothold by questioning this inference. To borrow an example from Barry Stroud's 1984 book, The Significance of Philosophical Scepticism, suppose you find yourself locked in a windowless room with nothing else but a bank of monitors along one wall. How would you know whether the images on the screens accurately represent what is happening outside the room? The sceptic argues that we are in the same position with sensory experience.

Disjunctivists reject this traditional picture. They propose, instead, that the experiences we have in the good case are fundamentally different from the experiences we have in the bad case. In the good case, we are immediately aware of our environment. When I am in the good case, looking at the bowl in good light, the fact that a bowl of fruit is in front of me is directly present to me. I do not infer that the bowl of fruit is there, as the traditional view would have it. I simply see that it is there. Some Disjunctivists spell this out by saying that perception is a relational state – the object perceived is an essential constituent of the perceptual experience. So you cannot have that perceptual experience unless what you are perceiving is really there. Because of this, Disjunctivists will typically say that if I am in the good case, I am thereby justified in holding the corresponding belief. And, unlike the Dogmatist, the Disjunctivist thinks my warrant for the belief is indefeasible. That is to say, I am not merely justified in believing that there's a bowl of fruit in front of me in the absence of defeaters. Rather, if I am in the good case, it entails that I know that there's a bowl of fruit there before me.

Disjunctivists do not have much to say about the bad case. In the bad case (say, when I am hallucinating) something entirely different is going on, even if it is indistinguishable from the good case. The point is that we should not conclude from the fact that it might be indistinguishable that all we get in the good case is what is common to both the good case and the bad case.

Let us return to the case of religious experience. As with sensory experience, we can distinguish between the good case and the bad case. In the bad case, the subject is not perceiving God and has no justification for her religious belief. But in the good case, God exists and is directly present to the subject of the religious experience. The subject is immediately aware of God's presence. We might even say that God is an essential constituent of the experience. The subject thereby has indefeasible justification for the relevant religious belief: just having the experience entails that she has knowledge of God. In the bad case, the subject is merely hallucinating, having a seizure, etc. The hallucination might be indistinguishable from the veridical religious experience one has in the good case, but we should not conclude from this that one fails to have religious knowledge in the good case.

Of course, none of this establishes that anyone is ever in the good case. The point is just that if someone is in the good case, then her belief in God is justified – and indeed would count as knowledge – just by virtue of having the experience. She does not first need to rule out other alternatives. And no other potential evidence could end up defeating that justification. So what this application of Disjunctivism allows for is the possibility that a religious experience could ground religious knowledge. It does not prove that anyone actually has such knowledge.

It might be helpful in this context to connect the distinction between the traditional conception of experience and the Disjunctive conception with Charles Taylor's distinction between the ‘buffered self’ and the ‘porous self’ in his 2007 magnum opus, A Secular Age. According to the pre-modern worldview, as Taylor describes it, the boundary between mind and world was more porous than it is today. The world back then was seen as enchanted, not only in the sense of containing spirits and magical powers, but also in the sense of being full of meaning and significance. The pre-modern self was penetrable by the world, and people derived meaning and purpose from their place within it. When the enchanted picture of the world was eventually supplanted by a more mechanistic and ‘disenchanted’ picture, the porous self was replaced with the modern conception of the ‘buffered self’. Instead of penetrating the mind, the disenchanted world merely causally interacts with it. Things are meaningful only in virtue of how we respond to them, and any ultimate purpose we have must come from within.

You can see how the picture of the buffered self goes along with the traditional picture of experience. Perceptual experiences, on this picture, occur entirely within the mind, buffered from the external world. Experiences are merely caused by things in the outside world, and anything we know about the world has to be inferred from the character of our experiences. (Think back to the locked room full of monitors.) The Disjunctivist is effectively recommending that we return to a more porous conception of the boundary between mind and world: we should not think of the mind as cut off from the outside world. Rather, in having experience (in the good case) we are immediately aware of how things are; the things in the world are directly present to us. It is therefore fitting that John McDowell, a leading proponent of Disjunctivism, suggests that we reject the ‘bald naturalism’ of our modern age and return to a more enchanted picture of the world.

For those who embrace the Disjunctive model of perception, and the conception of the enchanted world that accompanies it, James's proposal may begin to make sense. In a 2014 discussion of Taylor's work, the New York Times opinion writer Ross Douthat observes that how you interpret a religious experience may depend on which worldview you inhabit. The buffered selves of the disenchanted world will tend to view a religious experience as the ‘internal’ product of a disordered mind. If, on the other hand, the self is porous to the enchanted world, it seems perfectly natural that religious experiences could be direct confrontations with God, and it is unsurprising that those who have such experiences would describe being intimately connected with, or even penetrated by, the divine. In this context, a veridical religious experience could give someone immediate, indefeasible knowledge of the supernatural. However, it is also in this context that such knowledge is taken for granted. As Taylor points out, the enchanted world is a world in which it is hardly possible not to believe in the supernatural. Religious belief is the default. As a result, while reviving the enchanted picture may allow us to make sense of the justificatory power of religious experience, it would also render such justification entirely superfluous.