The research

words
of testimony.

Across interviews with people, drawn from archives, patterns repeat: in what people experience during an awakening, what sets it in motion, how it changes the body, what is hard in the years that follow, and what helps them settle.

Key findings

What the data shows

The clearest result here is not about any one experience. Every account was read twice, by two independent readers working from the same rulebook, and only of the features tracked were described plainly enough for both readers to draw the same line. Every one of those is concrete: something a person saw, heard or felt. Every feature they could not agree on asks the reader to judge what an experience meant. Concrete experience can be counted from interviews at this scale. Interpretive categories cannot.

The headline figure in each panel below is the share of people both readers marked. The wider figure, where at least one of them did, is given in the text beside it. Where the two run far apart, the number is a shape rather than a measurement, and it says so.

Across the whole taxonomy

features out of were described clearly enough that two independent readers drew the same line. All of them are concrete: light, emptiness, visions, hearing guidance, energy in the body, leaving the body, the body moving on its own, psychedelics. This split is the strongest thing here.

The experience

described unity or oneness, the boundary between self and everything else dissolving. Up to on the wider reading, and that long tail is the point: the most universal experiences are the ones people reach for the same abstract words to describe, which makes them the hardest to count.

The experience

described hearing guidance, a voice, a knowing, or a transmission they couldn't attribute to their own reasoning. Up to if you include what only one reader marked, and the two of them agreed here more closely than almost anywhere else.

The experience

described leaving their body and watching their own physical form from outside it. Up to . Nobody in this collection nearly died, so this is not carried by near-death accounts. It is also the feature the two readers agreed on most closely.

Physical changes

described the body moving on its own, shaking, jerking, rocking, or holding postures nobody chose. Up to . It is the only thing the body does that both readers could agree on, and often the part people find hardest to talk about.

Triggers

could name nothing that set it off. No practice, no ceremony, no preparation. Up to on the wider reading, and that gap is one of the widest on the site: what counts as a cause is exactly the kind of judgment two readers make differently.

Life after

could not explain what happened to the people around them. Up to , the most commonly described difficulty afterwards. Not one feature in this whole section reached the bar where both readers agreed closely, so read the ordering, not the values.

What helped

steadied themselves by reading and study, more than by anything else they named. Up to . As with life after, nothing in this section was agreed closely enough to quote on its own.

Why it matters

What scale makes possible

Most published studies of spiritually transformative experiences work with twenty to a hundred accounts. That is enough to show something exists. It is not enough to say how common it is, because at that size a feature turning up in half the sample and a feature turning up in a quarter of it are hard to tell apart. At interviews with people, they separate. Patterns that look like coincidence in a small sample start to look like structure.
This collection draws on separate interview archives, from meditation communities to kundalini-focused programmes to general awakening interviews, and none of them supplies a majority. That matters because every archive has its own framing, its own vocabulary, and its own reasons for choosing a guest. A pattern that holds across all of them is more likely to be about the experiences than about one community's way of talking.
Every account was read twice, by two independent readers working from the same written rulebook, neither able to see what the other had marked. Where both marked a feature, it is recorded as clearly described. Where only one did, it is possibly described. The gap between those two is left visible on every chart rather than averaged away, because it is the honest measure of how much a figure depended on judgment. Reading twice does not remove error. It shows you where the error lives.

The result is a dataset large enough to compare groups within it: people whose awakening followed a practice against people who could name no cause at all, one interview community against another, what happened during the experience against what was hard for years afterwards. That has been difficult in this field for a plain reason, which is that large, consistently coded collections of first-hand accounts have not existed. This is one attempt to build one.