Religious Experience

The four arguments so far have been public. Their premises are available to anyone: that a universe exists, that its constants fall in a narrow band, that a concept can be formed, that obligations bind. Anyone can inspect the premises and judge the inference.

This one is not like that, and it is also, by a wide margin, the reason most religious people actually give when asked. Very few believers arrived at their position by weighing cosmological arguments. A great many report something they take to have happened to them.

Whether that has any evidential standing for the person it happened to, and whether it has any for anyone else, is the subject of this article. Both questions have answers that are more interesting than the confident positions on either side suggest.

What counts

Two components are doing work in that definition, and both exclude things that are often confused with the subject.

The experience has to seem to be of something. Being moved by a hymn is a religious feeling and it is not a religious experience in this sense, because nothing presents itself as encountered. The same goes for a general sense of wonder at the scale of the universe, unless it comes with the impression of something's presence.

And the seeming has to be of something objective, meaning something that the subject takes to exist independently of the experiencing. A person who reports a vivid feeling and describes it as their own emotional state has not reported a religious experience in this sense either.

Within those bounds the range is very wide. At one end are the dramatic cases: visions, voices, conversions that reverse a life in an afternoon. At the other, and far more common, is a quiet and persistent sense of being accompanied, reported by people who have nothing dramatic to describe and who would not have thought to mention it if asked about anything else.

For the argument, the quiet end matters more than the dramatic end. Dramatic cases are the easiest to explain away, because they cluster in conditions that have obvious alternative explanations, and there are few of them. The undramatic cases are numerous, distributed across every society anyone has studied, and reported by people who are in no way unusual.

The argument built on this material is not a demonstration and nobody presents it as one. It has the probabilistic shape set out in the beginner tier: the question is whether the existence and character of these experiences is more to be expected if theism is true than if it is false, and by how much.

Two kinds of report

The literature separates two things that popular discussion runs together, and the distinction turns out to matter for the hardest objection later on.

William James gave the description that is still the starting point, in The Varieties of Religious Experience, published in 1902 and delivered as the Gifford Lectures at Edinburgh in the preceding two years. He identified four marks.

Ineffability: the subject reports that no adequate account of the content can be given in words, which is not a report of difficulty but a claim about the kind of thing it was.

Noetic quality: despite resembling states of feeling, these states present themselves to the subject as states of knowledge, carrying insight and an authority that persists afterwards.

Transiency: they do not last, and cannot be sustained by trying.

Passivity: the subject feels acted upon rather than acting, as though the state arrived rather than was achieved, even when practices were used to prepare for it.

The first two marks are an awkward pair, and the awkwardness is the whole epistemological problem in miniature. The experience presents itself as delivering knowledge while resisting any statement of what is known. Ordinary knowledge can be handed on, checked, and corrected. Something that arrives with conviction and cannot be put into a form anyone else can assess is hard to place in an evidential scheme, and this is true whether or not it is veridical.

Walter Stace, in Mysticism and Philosophy of 1960, divided these experiences into two types. In extrovertive mysticism the unity is perceived through the senses, in and behind the ordinary world of objects. In introvertive mysticism it is reached by emptying consciousness of content altogether. Stace held that both culminate in an apprehension of unity, and that claim about a shared core is precisely what the objections below attack.

Rudolf Otto coined the word from the Latin numen in Das Heilige, published in 1917 and translated into English as The Idea of the Holy in 1923. His formula for the object is mysterium tremendum et fascinans: a mystery that is overwhelming and awful in the older sense of the word, and simultaneously fascinating in a way that draws the subject toward what is terrifying them.

Otto's claim was that this response is not a form of belief, not a moral response, and not a rational judgement, and that attempts to analyse it into those things lose what is characteristic about it. Whatever one makes of that, the category earns its place here for a structural reason.

Numinous experience and mystical union are not the same thing and they pull in opposite directions. The numinous is confrontation with something radically other, and the distance between subject and object is the heart of it. Mystical union is the dissolving of exactly that distance. Traditions differ over which they treat as the higher state, and any argument that pools religious experiences into a single body of evidence has to reckon with the fact that its data include reports of a personal other and reports of the disappearance of the personal altogether.

Swinburne's two principles

The argument that religious experience is evidence was given its most careful statement by Richard Swinburne in The Existence of God, published by Oxford University Press in 1979 with a second edition in 2004. It runs on two principles, neither of which is invented for the occasion.

The case for the first principle does not depend on anything religious. Without something like it, ordinary perception collapses, because there is no non-circular way to establish that perception is reliable. You cannot check whether your eyes work without using something, and whatever you use is itself a faculty whose reliability was in question. Every workable epistemology grants perception a default entitlement and then allows it to be overridden case by case, and the principle of credulity is a statement of that policy.

The second is in the same position. Nearly everything anyone knows about history, geography, other countries, or science they have not personally done rests on taking other people's reports as probably accurate. A policy of believing testimony only when independently confirmed would leave almost nothing standing.

Swinburne's argument is then short. Religious experiences are experiences. The principles apply to experiences in general. So they apply to these unless a principled reason is produced for excluding them, and the burden of producing it falls on whoever wants the exclusion. What the burden cannot be discharged by is a general reluctance about the conclusion, since that is not a criterion.

The principles are defeasible by design, and Swinburne allows several kinds of consideration to override them: doubts about the subject's reliability or state, doubts about the conditions, background evidence making the object's existence improbable, and grounds for thinking the putative object was not what produced the experience. The dispute is over whether religious cases are systematically vulnerable to these in a way ordinary perception is not.

The objections

Nothing checks it

The first objection is that religious experience lacks everything that makes perception trustworthy. Other people can look at the tree and report the same tree. Perceptual claims generate expectations that can be disappointed. Errors are detectable and get detected. None of this holds for a reported encounter with God, where there is no second observer, no prediction, and no procedure that would reveal a mistake.

William Alston's reply, in Perceiving God, published by Cornell University Press in 1991, is that the demand smuggles in the standards of one practice and applies them to another. His framework treats sense perception, memory, introspection, and reasoning as doxastic practices: settled ways of forming beliefs, each with its own characteristic checks, none of which can be justified without relying on itself. Sense perception cannot pass the test the objection sets, because any attempt to validate it uses it. If circular support and internal checks are enough for those practices, then a religious practice conducted within a tradition, which does have internal checks in the form of doctrinal constraints and the accumulated judgements of that tradition about which experiences to trust, is in a comparable position.

The result should be stated with the modesty Alston himself gave it. This is a defensive conclusion. It shows that the practice has not been convicted of irrationality by a standard that its critics can consistently apply, and it does not show that the practice delivers truth. Alston knew the difference and said so.

It has a natural cause

The second objection observes that these experiences have neural correlates, that some of them can be induced, and that certain physiological states produce them with some reliability.

The general form of this objection is weak, and the weakness should be stated before the strong version, because the general form is what usually gets asserted. Every veridical perception has a neural correlate. Finding the mechanism by which an experience occurs does not show that the mechanism is not a means of detecting something, any more than describing the optics of the eye shows that nobody sees trees. Identifying a cause and identifying a debunking cause are different achievements.

The strong version is a real difficulty. What undermines an experience is a cause that would have produced it whether or not the object was there. If a particular kind of experience arises reliably from sleep deprivation, or from a specific temporal lobe event, and would have arisen in exactly that form in a world with no God, then that particular experience carries no weight. For some reported cases this is plainly the situation, and the argument's defenders concede those cases and rest the argument on the rest.

The concepts were brought to it

Steven Katz argued in "Language, Epistemology, and Mysticism", published in the 1978 collection Mysticism and Philosophical Analysis by Oxford University Press, that there are no unmediated experiences of any kind, and that mystical experience is shaped throughout by the concepts, expectations, and training the mystic brings. Christians have Christian experiences and Buddhists have Buddhist ones, and this is not a coincidence to be explained by a shared object described differently. It is what the constructive role of the subject's framework predicts.

The reply available to Swinburne and Alston is that mediation is universal and does not by itself debunk. An ornithologist and a novice looking at the same bird have different experiences, structured by different concepts, and the ornithologist's experience is more informative rather than less veridical. Concept-laden perception is still perception.

Where Katz's argument does damage is to a particular use of the evidence. The older versions of the argument leaned on cross-cultural convergence: the claim that mystics everywhere report the same thing, which would be difficult to explain without a common object. If the convergence is an artefact of describing the reports at a level of abstraction that removes everything the traditions actually say, then the convergence is produced by the analyst rather than found in the data. That was the strongest evidential feature of the argument as Stace presented it, and it has not recovered.

The diversity problem

This is a sharper objection than the observation that religious people disagree, and the sharpening is worth setting out.

The disagreement is not confined to the doctrine built on top of the experiences. It reaches into the reported content. Some experiences are of a personal being who addresses the subject. Some are of an impersonal absolute in which distinctions dissolve. Some are of the absence of any enduring self at all, which is what a Buddhist practitioner may report and take to confirm a doctrine incompatible with theism. Some are of many divine beings. These reports cannot all be accurate about their object, and the argument from religious experience proposes to treat all of them as evidence under the same principle.

Three responses are on offer and each has an identifiable cost.

John Hick proposed that the traditions are responses to a single reality encountered under the categories each brings, so that no tradition describes that reality as it is in itself. The experiences are then all veridical in the way perception of a single object through different instruments is veridical. The cost is that the reality as it is in itself becomes something none of the traditions is actually about, and most practitioners reject that as a description of what they take themselves to be doing. The advanced tier takes up this position and its rivals directly.

Alston and Alvin Plantinga argued instead that a person inside a tradition has epistemic resources unavailable from outside it, and remains rational in continuing to rely on that tradition's practice even while recognising that someone in another tradition is symmetrically placed. The cost is that the symmetry is conceded and nothing adjudicates it. Alston regarded this as an uncomfortable result rather than a solution, and said so, which is more than the position's later users have generally done.

The third response is deflationary. The conflicts are sharpest at the level of interpretation and doctrine, and thinnest at the level of what is experienced, so what survives as common evidence is something like this: something transcendent, of overwhelming value, encountered rather than constructed. That is defensible, and after Katz it is the only version of the perennialist claim still standing. Its cost is that what survives may be too thin to be worth the argument. A transcendent something of overwhelming value is compatible with an impersonal absolute, and the argument was supposed to support a personal God.

Next

The arguments in this tier so far have all asked whether theism is true. The next asks a different question, which is whether it might be rational to believe something on grounds that have nothing to do with its being true: because of what believing would gain you, or cost you, or make possible.