You walk into a place you have not been for twenty years and for a few seconds you cannot see any of it.
The chairs are there. Your eyes are working. But nothing is arriving as a chair — the whole room is arriving as one thing, a single atmosphere with a name, and you are inside it before you have registered a single object in it.
Then it resolves, and there is a table, and a window, and the carpet is different from how you remembered.
Almost everybody has had a version of that. What the corpus describes is that first state, not resolving.
The mahā-bhāgavata — the most advanced devotee — sees the mobile and immobile things of the world but does not see their forms.1
Two clauses and both are needed.
Sees — this is not blindness, absorption, or a trance. Perception is happening and the world is present.
Does not see their forms — what is absent is the individuation. The distinguishing outline, the this-not-that, the property by which a tree is a tree and not a wall.
So the claim is not that the advanced devotee sees something else in place of the world. It is that they see the world with one specific operation switched off.
The corpus's own gloss is that what such a devotee perceives, everywhere, is Kṛṣṇa — that the forms drop out and the relation remains.
Which is easy to file as devotional excess and is more interesting than that, because it is a claim about where forms come from.
Ordinary perception is not a passive reception of shapes. It is a sorting: this is a tree, that is a person, this belongs to the category of things I step over. The forms are produced by a categorising apparatus running continuously and invisibly, and the categories are supplied by training, need and language.
The corpus is describing a state in which the sorting has stopped without the seeing stopping, which is not incoherent and is reported in some form by a great many contemplative traditions.
What is distinctive here is what fills the gap. Not emptiness, not undifferentiated light — the corpus is explicit that the impersonal option is the wrong one.2 The residue is a someone.
Notice that the corpus tells you what the state lacks rather than what it contains.
Does not see their forms. That is a subtraction. And a subtraction is a very odd way to describe an attainment — the tradition's other summits are described by addition: more love, more intensity, an increase by ten million times.3
Three reasons a description would go negative, and they are not exclusive.
There is no vocabulary for the positive content. The state is not composed of anything nameable, so the only available handle is what has been removed.
The negative version is checkable and the positive is not. Does not see forms is a claim a person could recognise in their own experience; sees Kṛṣṇa everywhere is not a report anybody could confirm or fail to confirm.
And the subtraction is the argument. If forms are produced by a sorting apparatus, then removing the apparatus is not a loss of information — it is the removal of an addition. On that reading the mahā-bhāgavata is not seeing less. They are seeing what was there before the categories arrived, and everybody else is seeing the world plus their own filing system.
The third reading is the one that makes the claim worth having, and the corpus does not state it.
The passage is brief and its context is doing work.
It appears in the chapter that also produces the eleven-rung ladder, the ornament allegory and the sakhī multiplier — the corpus's most technical and most enumerative chapter.4 And this claim is the opposite of enumeration: a state in which the distinctions the rest of the chapter is built from have dissolved.
The corpus does not remark on the juxtaposition. Which is a shame, because it is the interesting one: a tradition whose contemplative method is exhaustively itemised — twenty ornaments, eleven rungs, sixty-one readings — locates its highest perceptual state in the absence of items.
The available reconciliation is that the itemisation is the route and the dissolution is the destination, which is coherent and is how most traditions with both would answer. The corpus does not say it.
The extraction record flagged this as a probable enrichment target rather than a page, on the grounds that the vault might already hold the material.
A search of the eastern-spirituality domain finds no prior page on this state, so this is first coverage rather than a duplicate.5 It is built at MEDIUM because the corpus gives it two clauses and no development — a thin page honestly sized is better than a thick one padded, and the corpus's brevity here is a fact about the corpus rather than a gap to be filled.
Layer A gives the claim in one sentence, inside the dialogue.
Layer B expands it into the standard three-grade taxonomy of devotees — kaniṣṭha, madhyama, uttama — and the associated behavioural criteria.6 That is real systematic context the verse assumes and does not supply.
Worth marking the shift, because it is the same one this build keeps finding. The verse describes a perceptual state; the commentary supplies a rank. Does not see forms becomes is a first-class devotee, and the second is a classification while the first was a report.
You walk into a place and for a moment nothing resolves into objects. Stay there as long as it lasts. That interval is your perceptual system before the filing has started, and it is available a few times a year at most.
You are certain you are seeing a situation clearly and the situation has arrived pre-sorted — this is a problem, that is a person who is difficult, this belongs in the category I already have. The categories are yours and they arrived before the seeing did.
Somebody describes an attainment entirely by what it lacks. That is often the only honest description available, and it is also the form that cannot be checked from outside. Ask what they would expect to find themselves doing differently, which converts a subtraction into something with consequences.
You have built an exhaustive method — a checklist, a taxonomy, a practice with twenty steps — and you notice the people who are best at it seem not to be using it. That is the itemisation-as-route pattern, and it is worth saying out loud to whoever is still on step three.
The evidence is a single sentence with two clauses, inside the corpus's densest dialogue, with no development.7 The page is sized to that and says so.
The tension the source doesn't close: the corpus's method is exhaustively itemised and its highest perceptual state is the absence of items. The route-and-destination reading reconciles them and the corpus never offers it.
A second one. The claim is stated negatively and the corpus's other summits are stated positively — by increase, by multiplication, by accretion. The one attainment described by subtraction sits in a tradition whose entire vocabulary of progress is additive, and the corpus does not register the exception.
Open: whether the state is continuous or intermittent. The corpus says the mahā-bhāgavata sees this way and does not say whether they can also, when required, see a chair.
This page sits oddly and productively against Rādhā's Ornament Allegory, which is in the same chapter. That passage builds a contemplative object out of twenty distinct items with an order; this one describes a state in which items have gone. One chapter, a maximally itemised practice and a formless summit, and the corpus does not say the first leads to the second.
Against Advaita Vedānta: Śaṅkara the passage is unexpectedly close to the position the corpus spends chapters attacking — the dissolution of distinctions in perception — and the difference is entirely in the residue. Forms dissolve and a person remains, which is the corpus's standing amendment to every impersonalist claim, applied here to perception rather than to metaphysics.
And it belongs with The Eho Bāhya Ladder as the chapter's other statement about what advancement looks like from outside. The ladder is content; this is perception; and the corpus supplies no bridge between the two.
Looking vs. Seeing: As a Man Is, So He Sees — Blake's line, which that page turns on: a tree that moves some to tears is to others a green thing that stands in the way. As a man is, so he sees.
Both passages claim that perception varies with the perceiver rather than with the object, and they disagree about what the variation consists of — which is the useful part.
Blake's variation is in what the object means to you. The tree is fully present as a tree in both cases; one person is moved and the other is inconvenienced. The forms are stable and the response is not.
The mahā-bhāgavata claim is stranger and more radical: the forms themselves are not stable. It is not that the tree means something different; it is that the tree-shape has stopped being produced, while the seeing continues.
Which produces a distinction neither page draws, and it is a real one. There are two places perception can differ between people — in the sorting and in the valuing — and they are usually collapsed. Blake's contrast is about valuing, and almost everything written about how differently people see the world is about valuing, because valuing is the part we can discuss.
The sorting is much harder to reach, because it runs below report: by the time you can say what you saw, it has been sorted. You cannot describe your own categories from inside them, which is why the corpus's claim is stated as a subtraction and why Blake's is stated as a response.
And the practical upshot is uncomfortable. When two people disagree about what is in front of them, the disagreement is usually about valuing and occasionally about sorting, and only the first is negotiable by conversation. If somebody's categories are different from yours, no amount of pointing will help — they are not seeing your object and disliking it. They are not seeing your object.
Sharpest implication. The forms in your visual field are not given; they are produced, continuously, by a filing system you did not choose and cannot inspect from inside. This corpus's highest perceptual claim is not that its adepts see something extra but that they have stopped adding — which is a subtraction described as an attainment, in a tradition that otherwise measures everything by increase.
Generative questions.