Buddhist cosmology puts six heavens immediately above the human world, all still inside the Realm of Desire — meaning their inhabitants still want. And it specifies, with unusual concreteness, how they want.
In the Yāma heaven, males and females embrace. In the Tuṣita heaven, they clasp hands.
In the Nirmāṇarati heaven, they smile at one another. In the Paranirmita-vaśavartin heaven — the highest of the six — "the beings satisfy their sexual desires by glancing into one another's eyes."1
Read that ladder in order and the pattern is unmistakable: as you go up, the contact gets less. Embrace, hands, smile, glance. Four rungs, and the physical component thins at each one until nothing is left but looking.
A line of human recognition gets added that does more work than it looks like: "When we talk of love at first glance, this is a faint reflection of this type of sexual contact (as well as karmic memory) and since it involves the eyes, we can guess that it involves shen."1 Nine pages later two more graded variables arrive: longevity, measured in earth years, and — for males — the number of wives, both varying "according to one's rank of spiritual attainments in these heavens."3 This is used to rescue a claim most Western readers file as folklore: Mohammed's seventy-two wives in paradise is "actually a perfectly valid statement regarding certain Desire Realm heavens," though "you need not hold onto the number seventy-two as being the rule."3 The mechanism offered is a balance requirement — the yang nature of those realms "must be balanced by an equivalent existence of refined heavenly yin."3
And then comes the question that undoes the paragraph: "with the large ratio of wives to husbands… you have to ask, 'What makes an individual think they'll be reborn in that heaven as a male rather than as a female?'"3 That's the sharpest sentence in the chapter and it's aimed at his own reader. The reward being described is only a reward from one of two positions, and the ratio that makes it impressive is exactly what makes the position unlikely.
The scheme isn't arbitrary and the axis it runs on gets named. The three essences — jing, chi, shen — are ranked, coarse to refined, and they have locations. Jing is the reproductive and endocrine substrate. Shen is what you see in the eyes.
So: "The sexual relations for human beings, on the other hand, primarily involve jing, and only rarely do they ever involve the true chi of cultivation. Since jing and chi are much lower than shen on the spiritual scale, it's natural to assume that this form of sexual activity would correspond to lower beings on the spiritual totem pole."1
That converts a quaint cosmological detail into a claim with structure. The ladder of heavens is a ladder of which essence the exchange runs on. Human sex is a jing transaction. Eye-to-eye contact at the top of the Desire Realm is a shen transaction. And the intermediate rungs — hands, smile — sit between.
Which explains why the beings higher up need less: they aren't being ascetic. They're exchanging something finer, and a finer substance needs less machinery to move it. The inference from the mode back to the beings is exactly this: eye contact "already gives you an indication that their 'physical forms' are much more advanced than our own because this type of activity shows that it's easy for them to cultivate their shen."1
Now the thing that makes this page worth building, and the text walks straight past it.
The four modes get mapped onto the four classes of Tibetan tantra:
In Action Tantra the practitioners look at one another or at beautiful deities in order to generate the bliss of the path. In Performance Tantra, practitioners use smiling and laughing to try and generate the spark of samadhi bliss. In Yoga Tantra, the practitioners hold hands in order to generate bliss and in Highest Yoga Tantra, people use the contact of sexual embrace in order to generate bliss.1
Set the two ladders side by side.
The heavens, ascending: embrace (Yāma) → hands (Tuṣita) → smile (Nirmāṇarati) → glance (Paranirmita-vaśavartin).
The tantras, ascending in the Tibetan classification: Action → Performance → Yoga → Highest Yoga. Which by the mapping given is: glance → smile → hands → embrace.
The two hierarchies run in opposite directions. The mode belonging to the highest heaven belongs to the lowest tantra class. The mode belonging to the lowest heaven belongs to the class called Highest.
The text gets within one sentence of saying so — "In the Paranirmita-vaśavartin heaven, which would correspond to Action Tantra activities."1 Then it uses that correspondence to argue the beings are advanced. What goes unnoticed is what the same correspondence does to the practices: the one matching those advanced beings is the entry-level tantra, and the one named Highest matches beings at the bottom of the ladder. This isn't a small inconsistency. Both hierarchies are internal to traditions the book treats as authoritative, both are about the same four modes of contact, and they disagree completely about which mode is the refined one.
Three readings are available and the book takes none of them.
The tantra ranking is about efficacy, not refinement. Highest Yoga Tantra might be highest because sexual embrace generates the strongest energetic event, not because it is the most rarefied — a bigger lever for a coarser body. On this reading there's no contradiction: the heavens rank beings by refinement, the tantras rank methods by power, and a coarse being needs a powerful method. This is the most charitable reading and it's consistent with the line that human sex runs on jing.
The correspondence is spurious. Four items matched to four items is exactly the kind of pattern that appears whether or not it means anything, and nothing forces the Tibetan classes to map onto Desire Realm modes at all. The neatness is the warning sign.
Or the inversion is a genuine finding about the esoteric schools — that a system ranking sexual embrace as its highest practice has its hierarchy upside down relative to the cosmology it inherited from the same corpus. That would be a real criticism, and it's adjacent to the argument at The Esoteric Schools Run It Backwards — though that page's charge is different, about starting from results rather than causes.
The third reading is available here, and would strengthen a critique made elsewhere at length, and it never gets made. Which suggests the inversion went unseen rather than dismissed.
The wider passage contains a definition that quietly makes the whole ladder harder to climb than it appears:
here we must remember that "desire" really includes all kinds of clinging and attachment including the desire for purity, calmness and the cultivation of Tao as well.2
That reclassifies the practitioner's own motivation as Desire Realm material. The consequence is stated: "You have to be empty of everything to achieve samadhi and leave the Desire Realm with its sufferings, including the desire for the mental states that seem to be stages of purity. You have to cultivate and just let them happen rather than try to grab them or create them. If you go after or cling to these states of mind, you can only achieve a false samadhi."2
So the exit condition from the lowest realm includes not wanting to exit it. That's not a paradox the passage tries to soften, and it's the same structure as Noticing Purity Is What Spoils It — a condition that the wanting of it forecloses.
It also relocates the whole discussion. A reader arriving at the graded-sexuality material expects it to be about sex. The definition says the operative variable is clinging, and sex is simply its most legible instance.
Amid the cosmology there's one sentence that is a practical instruction and it's the most useful thing on the page:
whenever you can freely abandon clinging to even some small measure of your desires so that they don't rule you, you are making definite progress in freeing yourself from the Realm of Desire.2
Freely abandon. Not suppress, not abstain — release, without effort maintained against resistance. Even some small measure. And so that they don't rule you, which locates the test in whether the desire is steering rather than whether it's present.
That's a real criterion, it's incremental, and it doesn't require any cosmological commitments. You can run it on a Tuesday without believing in the Yāma heaven.
Which is worth noting against everything else on the page. The heavens are unverifiable, the tantra correspondence is questionable, and the essences are a framework — and the operative instruction underneath all of it survives all three being wrong.
The passage ends by locating the exit precisely. Getting past sexual desire is what admits you to the first dhyāna. And that desire is defined broadly — "basically attachments to the body and its sensations, attachments to body consciousness."1 Someone who reaches it "has finally surpassed in stature the Desire Realm, and its ordinary (non-cultivating) inhabitants."1
Note what that says about the heavens just described. The beings in Paranirmita-vaśavartin, refined enough to conduct their sexual life through eye contact alone, are still below the first dhyāna. Their extraordinary refinement is within the Desire Realm and doesn't get them out of it.
And there's a wrinkle flagged as rare: "Buddha already told us that you can cultivate four different samadhi before the first dhyana."1 So the Desire Realm has its own internal samādhi ladder — which is why its inhabitants "can do various supernatural things" without being anywhere near the real absorptions. That's the standing deflation of Superpowers Are Not Evidence, given a cosmological address.
You want to test the actual claim underneath this — that the refinement of a want shows up in how much machinery it needs. Take one appetite that isn't sex, because sex is too loaded to observe cleanly the first time. Coffee works. So does checking your phone.
For one week, don't change the behaviour. Change only one thing: notice how much has to happen before the want is satisfied.
With coffee: is it the drink, or the cup in your hands, or the walk to the machine, or the smell, or the ten minutes of not working? Find out by accident, not by experiment — some morning the coffee will be bad, or cold, or you'll get interrupted, and you'll discover which part you actually missed.
With the phone: is it the message, or the unlocking, or the first two seconds of the screen coming on? Write one line each time you notice.
By the end of the week you'll have a rough map of where the want actually lives, and it's almost never where the object is. That's the whole content of the graded-heavens image, stripped of cosmology: the same appetite can run on a great deal of machinery or very little, and the amount of machinery is a fact about you rather than about the thing.
Then the harder half, and it's the one the passage insists on. Notice whether you want this exercise to show that you're doing well. That wanting is on the same list — the desire for purity, calmness and the cultivation of Tao. It doesn't disqualify anything. It's just the same mechanism, one level up, and the point of the week is to see it operating rather than to have got rid of it.
The strongest content is the jing–chi–shen axis, because it makes the ladder mean something structural rather than decorative: the rungs differ in which essence the exchange runs on, and that's consistent with the book's physiology elsewhere.1 The central tension is the inversion. The heavens rank embrace lowest and glance highest; the tantra classification mapped onto them ranks glance lowest (Action) and embrace highest (Highest Yoga). Both hierarchies are internal to the corpus, both concern the same four modes, and the book never notices they disagree.1
A second tension: "desire" is defined to include the desire for purity and for the Tao, which makes the exit condition from the Desire Realm include not wanting to exit it.2 Stated without softening, and never reconciled with a book that spends six hundred pages recommending the wanting.
A third, on evidence: nothing in the graded-heavens material is sourced. No sutra, no footnote, no locator. The four modes are among the most specific factual claims in the chapter and they arrive bare. The same is true of the wives, the longevity gradient and the male-birth-from-the-knees detail at p 279 — the reader is referred to "Buddha's various descriptions of the heavens" with no work named.3
A fourth, and it needs stating plainly. The gender doctrine at p 279 holds that male rebirth in these heavens requires transforming yin chi into yang chi, and therefore cultivation attainment. Yin merit alone yields rebirth "as a heavenly female" — with the disclaimer "naturally this isn't a sexist statement, for it's just the way things are."3 A disclaimer that asserts the claim is factual doesn't address the objection to it. Recorded as the source's stated position and not carried as fact, alongside its two internal counterweights. The balance mechanism makes the female position necessary rather than deficient. And the question raised about which side of the ratio a reader expects to land on undercuts the reward the doctrine is being used to explain.
Also recorded and not carried: the same page's claim about jihad — that killing in Allah's name yields "unfortunate retribution rather than a reward," and that the belief is "crazy indoctrination which is fostered entirely for political ends."3 The karmic argument is internal to the framework and unremarkable within it; the characterisation of a living tradition's adherents is the recurring pattern this vault excludes, and it's excluded here.
Open question: is the tantra-class correspondence a construction here or a received one? If it's his, the inversion is an artefact of his mapping and means nothing. If it's traditional, it's a genuine unresolved disagreement inside the Tibetan corpus about whether refinement or power sets the ranking — and that would be worth pursuing.
The reading of the four heavens and the reading of the four tantras are governed by different principles, and he applies them a paragraph apart without noticing. The heavens are read by refinement: less physical contact, finer essence, higher being. The tantras are read by method: what practitioners do to generate bliss. Under the first principle Action Tantra should be the highest class. Under the Tibetan ranking it is the lowest.
He resolves nothing because he doesn't see a conflict — the correspondence is offered as illumination, and it is, but it illuminates a disagreement rather than an agreement. Against The Four Descending and Ascending Blisses, this page shares the structural finding. There, four bliss-stations turn out to be named by their order rather than by their locations, so that the ascending gloss reverses every assignment — the four blisses are a count, not a map. Here, four modes of contact are mapped onto four tantra classes and the resulting order runs backwards. Both are cases of a four-item scheme whose ordering does less work than its items suggest, and in both the author supplies the evidence and not the observation.
Against Jing, Chi, Shen — The Three Essences, this is that framework applied at cosmological scale, and it's one of the cleaner applications: the essences give the heavens their ordering principle, and the ordering principle predicts the modes rather than being read off them.
The vault's page on Dhvani: Suggestion as the Soul of Content carries the Sanskrit poetic doctrine that the highest kind of language is the one that suggests rather than states. What's implied outranks what's expressed, and a poem's power is measured by how much of it never appears on the page. The graded heavens are that doctrine as a cosmology. Embrace states; glance suggests. And the ranking runs the same way in both: the mode carrying the least explicit content sits highest, not because explicitness is coarse in itself, but because a receiver who needs less signal to receive the whole thing is a more capable receiver.
That last clause is the insight the pair produces, and neither page states it alone. Dhvani is usually read as a claim about the maker — the skilled poet implies rather than declares. Read against the heavens, it becomes a claim about the receiver's capacity, which is what actually determines whether suggestion works. A glance transmits nothing to a being who needs an embrace. Suggestion isn't a compression the sender achieves; it's a bandwidth the pair share, and the poet's restraint is a bet on the reader.
Which explains something both traditions handle badly. Dhvani theory has no good account of what happens when the reader misses it, and the graded heavens have no account of a being who tries to operate above their rung. The failure mode is identical and it isn't the sender's: a signal pitched above the receiver's capacity doesn't arrive faint, it doesn't arrive at all — and the receiver experiences nothing, so they can't even register that something was sent.
Run it the other way, and the inversion problem gets a possible resolution the cosmology couldn't supply. Highest Yoga Tantra using embrace isn't a coarser practice; it's a practice designed for a receiver of our bandwidth. Explicit means for a low-bandwidth audience is the same design decision a poet makes writing for the general reader. And calling it the highest class makes sense once you notice what a method is ranked by: what it accomplishes on the practitioner available, not its own refinement.
(One handshake. The physiological framework this page runs on is the source's own and is carried on its own page.)
Sharpest implication: the ladder is not about restraint, it is about bandwidth — how much machinery a want needs in order to complete. Which reframes the whole ascetic tradition's relationship to this material: the practitioner who abstains is not performing the higher mode, they are performing the lower mode with the object removed. The heavens describe beings for whom the embrace would be redundant, not forbidden.
Generative questions: