Technique Is Local, Crisis Is Universal

Spend a few days reading across the meditation literature in this library (Buddhaghosa beside the hesychasts, the Daoist alchemists beside the Sufis, Iamblichus beside the quietists) and something unexpected comes into focus. These texts are not devotional mood pieces. They contain protocol specifications, dose warnings, staging models, and individual-differences prescriptions. Comparative scholars have long read the traditions against each other; what nobody has quite done is read these texts as a pre-registration archive: dated, priority-stamped, falsifiable claims, waiting for their trials.

Here is the fifth-century Visuddhimagga, the great Theravada practice manual, which we hold in a 1,282-page edition, translated nearly in full, specifying the middle stages of breath meditation. It does not say “watch your breath.” It prescribes a sequence that deliberately generates rapture, then regulates it:

The same manual prescribes practices by temperament (breath for the deluded and the discursive, the divine abidings for the angry, foulness for the lustful), a personalized-medicine claim from the fifth century that modern meditation research, which mostly treats meditation as one intervention, has never directly tested. That is the spirit of this note: not “ancient wisdom anticipated science,” but that these sources state specific, falsifiable claims, and someone should falsify them.

Three Places the Sources Contradict the Modern Picture

1. The West was meditating all along

The common story (meditation arrived from Asia in the twentieth century) mistakes a vocabulary change for an import. The Stoics and the Byzantine hesychasts use the same Greek word, prosochē, for vigilant attention, a millennium apart. Medieval Latin used “meditation” for the active, discursive stage of practice and “contemplation” for received stillness, nearly the reverse of current English usage, which is why reading Teresa of Ávila or the Philokalia with today's definitions systematically misleads. And when Indian practice did reach the Victorian West, it was assimilated not to “mindfulness” but to mesmerism. Henry Steel Olcott, co-founder of the Theosophical Society, introducing the 1885 Theosophical edition of The Yoga Philosophy (edited by Tookaram Tatya), explains samadhi through animal magnetism:

2. The rainbow is younger than your grandmother

We wrote in February about recovering the chakra tradition from its primary sources. A quantitative pass over the corpus sharpens the point. Across ten subtle-body systems in 432 books (Hindu tantric, Buddhist tantric, Daoist, Sufi, Kabbalistic, and others), the systems align on almost nothing except body location: heart and throat centres appear in ten out of ten. The modal chakra count in period Hindu tantric sources is six (the crown usually sits above the count, not in it). And only about 12% of pre-modern colour assignments match the familiar rainbow scheme, a rate that is roughly the same on either side of 1875, because the red-to-violet mapping does not come from the Sanskrit sources, nor even from Leadbeater's 1927 Theosophical colours, but from Christopher Hills's Nuclear Evolution (1977). The chakra chart in every yoga studio is younger than the transistor radio.

3. The safety documentation we threw away

Every tradition that uses imagery independently wrote the same warning list. The 1615 Daoist compendium Xingming guizhi tells the adept that deep stillness will conjure “towers, pavilions, jewels, pearls, female musicians”, and that none of it is attainment. Zen calls such productions makyō. The Sufi master al-Hujwīrī condemns taking any mental image for God as “utter anthropomorphism and manifest error”; Teresa subjects visions to discernment; the Persian Kashf al-Maḥjūb keeps beginners away from ecstatic music sessions “for there are great dangers in it.” The traditions disagree about whether to use images, but they agree about the failure mode: being used by them. Modern secular mindfulness shipped the techniques without the teacher, the map, or the warning lists, and the adverse-effects research programme (the “Varieties of Contemplative Experience” project: Lindahl, Britton and colleagues' 2017 mixed-methods study of meditation-related challenges in Western Buddhists) is now empirically rediscovering the territory those lists covered.

A Measurement, and the Hypothesis It Suggests

This summer we ran a probe-based survey of first-person experience across the corpus: retrieval by experiential probe sentences, classification, a verbatim-quote gate, and embedding of the surviving passages. It is a draft working paper: single-model classification of AI-translated text, no human-validated subset yet. The same caveat covers the subtle-body numbers above, which come from the same survey. Treat everything in this section as provisional, including the phrase this post takes as its title, which is offered as the hypothesis we would most like someone to falsify, not as a finding.

Method travels in manuals; vision travels in reports. Contemplative-emptiness material arrives as 57% instruction and 6% first-person testimony. Visionary body-dissolution material inverts it: 7% instruction, 63% testimony. This is a fact about the record, and it has a sharp consequence for contemporary research: a phenomenology questionnaire can only collect testimony, so the traditions modern science studies most are precisely the ones whose literature is structurally worst at supplying first-person reports.

Technique looks local; crisis looks universal. We extracted the Visuddhimagga's stage descriptions from our own copy, in its own vocabulary, and matched them across the whole corpus. Inside the Buddhist tradition, matches are dominated by concentration attainments (roughly half, against a third for the crisis stages). Outside it, the proportions flip: the knowledges of fear, misery, and disgust (the deliberately mapped stretch modern practitioners call the dukkha ñāṇas) take about two-thirds of the matches, a three-fold shift in the odds. Two honesty notes before any interpretation. First, the same survey's own negative result: in embedding space, which book and which language a passage comes from predict its neighbours far better than what experience it describes; text organizes the space more than phenomenology does, which is why no claim here rests on raw embedding similarity. Second, a confound we have not ruled out: concentration vocabulary (jhāna, rapture, one-pointedness) is lexically Buddhist almost by construction, while crisis vocabulary (fear, darkness, dissolution) is generic across translated devotional prose, so part of the asymmetry could be about words, not experiences. With those on the table: if the pattern survives better instruments, the cross-cultural “common core” of deep practice is not bliss or union, as the popular perennialist picture has assumed; it is the crisis. Every staged map we hold, from the vipassanā knowledges to John of the Cross's nights to the Sufi contraction (qabḍ), budgets for it.

Seven Hypotheses the Sources Hand Us

Each of these is stated in, or directly implied by, a primary source we hold. Roughly ranked by feasibility:

1. Temperament matching. Buddhaghosa's trait-by-technique prescription table predicts a measurable interaction: practitioners matched to their temperament should outperform mismatched ones. Nobody has run this as a moderator study.
2. Staged versus flat breath practice. The sixteen-step protocol (induce rapture, then gladden, concentrate, liberate) against undifferentiated breath focus, on affect and absorption outcomes.
3. Breath cessation. Daoist “embryonic breathing” and the fourth jhāna independently claim respiration becomes imperceptible in deep absorption. Jhāna respirometry already exists (a 2024 case series by Chowdhury and colleagues measured significantly slowed breathing in absorption), but nobody has run the same measurement against the Daoist claim in one design. Two unconnected traditions, one protocol.
4. Constructive practice trains imagery. Kozhevnikov and colleagues reported in 2009 that Tibetan deity-yoga practitioners get a large (if temporary) boost in mental rotation and visual memory after practice: fifteen self-selected practitioners, so a lead, not a result. The untested extension: Ignatian exercitants, the West's deity-yoga analogue, should show convergent gains despite zero shared lineage.
5. Warning-list validity. The Varieties of Contemplative Experience project already grounds its adverse-events taxonomy in Buddhist maps. The open questions are cross-tradition (do the Daoist and Sufi warning lists carve the same joints?) and predictive: does the traditional category an event falls into predict whether it resolves benignly or escalates? The lists are specific enough to code against existing case data.
6. The heart-locus prior. Ten-of-ten convergence on a heart centre could be deep structure, or just interoceptive salience, since hearts are feelable. The test is a null model: is cross-tradition convergence greater than interoception alone predicts?
7. Dose-response of repetitive invocation. Iamblichus, defending theurgy around 300 CE, states a training-curve model of ritual prayer: “prayer, when frequented for a long time, nourishes our intellect … little by little, it perfects our infirmities … until it leads us to the summit”. Strip the polytheist frame and this is a duration-dependent plasticity claim about repetitive invocation (the same claim the hesychasts make for the Jesus Prayer and Buddhaghosa for the breath), and it is measurable.

Caveats, Because This Is the Whole Point

The corpus translations quoted here are AI working translations with the originals preserved alongside, with two exceptions worth naming exactly: the “anthropomorphism” quotation is Nicholson's 1911 English, while the samāʿ warning is our own AI translation of Zhukovsky's 1926 Persian critical edition. The experience survey is a single-model draft. The corpus itself is a biased sample: what was written, kept, scanned, and imported. None of these numbers should be quoted without those qualifications, and the hypotheses above are offered precisely so that people with laboratories can do what we cannot.

What the library adds is the thing laboratories have lacked: the primary record, at scale, in the original languages with working translations, linkable to the page. Every quotation in this note resolves to the leaf of the edition it came from. The manuals are open. The protocols are written down. They have been waiting, some of them, since late antiquity.