You have been somewhere and not been in it.
Walking a street in one city while entirely occupied by a room in another. Sitting through a meal with your attention four years back. Everybody has done this, and the ordinary version has a hard floor: you never stop knowing which one is real. The room is vivid and the street is where your feet are, and if somebody asked you would answer correctly without hesitating.
The question this page is about is what happens when that floor is not there.
Not a stronger imagination. Not a more absorbing memory. A condition in which the remembered place is what the eyes report, the present place is not available for comparison, and the person acts — walks into it, runs at it, jumps in.
The corpus states this as a standing condition of one man, in the narrator's own voice, and then supplies three verified instances of him acting on it.1
The generalisation is not assembled by a reader. It is in the text, flat, as a description of how he travelled:
*"When Śrī Caitanya Mahāprabhu passed through the Jhārikhaṇḍa forest, He took it for granted that it was Vṛndāvana. When He passed over the hills, He took it for granted that they were Govardhana. Similarly, whenever He saw a river, He immediately accepted it as the River Yamunā."*2
Three clauses, three terrain types, one operation.
Note the phrase the translation reaches for twice: took it for granted. Not imagined, not was reminded of, not saw as. Taking something for granted is what you do with the facts you are not currently examining — the category of things so settled that they are not propositions at all.
That is the strongest available description of the condition and the corpus supplies it in passing, as narration, in a travel chapter. The madness chapters then show it happening.
The dune. Antya-līlā 14. He sees Caṭaka-parvata — a sand dune — and takes it for Govardhana, and runs at it at the speed of the wind.3 What follows is the corpus's most extreme physical report: the flesh erupting at every pore "like pimples," hair standing like kadamba flowers, blood and perspiration flowing from every pore at once, the whole body fading to conchshell white and quivering like ocean waves.
The garden. Antya-līlā 15. A flower garden by the beach is Vṛndāvana, and he does not admire it — he searches it, and interrogates its trees about where Kṛṣṇa has gone.4
The sea. Antya-līlā 18. The moonlit sea is the Yamunā, and he jumps in.5 Which is how the fisherman comes to net him.
Three instances, three chapters, verified at full text.6 And in each case the same thing separates this from metaphor: he acts on it. A man comparing a dune to Govardhana stands and compares. A man who runs at it has not made a comparison.
(A fourth candidate — Sundarācala read as Vṛndāvana at Madhya 14 — is known only from the chapter map and was not read at full text. It is not counted toward the three and nothing here rests on it.)
The sentence that reorganises the whole page comes after the dune, when he is brought back:
"Who has brought Me here from Govardhana Hill? I was seeing Lord Kṛṣṇa's pastimes, but now I cannot see them."7
Read what that assumes. He was at Govardhana. Somebody moved him. The complaint is not that he was mistaken and has been corrected; it is that he was somewhere and has been carried away from it, by people, against his interest.
Purī is the displacement. Vṛndāvana is the location. That is the inversion the page is named for and it is the man's own account, not an interpretation of it.
Which changes what the community's care looks like from inside. Every retrieval in these chapters — off the dune, out of the sea, back from the cow shed — is, from his side, an abduction. The devotees think they are bringing him home; on the sentence he actually says, they are the agents of the removal.
And the corpus reports the complaint without softening it and without answering it. Nobody says you were never at Govardhana. They chant into his ear until the joints contract and then they take him back to the room, and the next chapter proceeds.
Which is the most honest thing in the material. A community that had resolved the question would have recorded the resolution. This one recorded the accusation.
And there is a second clause in the complaint that is easy to lose behind the first, and it is the one that names the loss.
I was seeing Lord Kṛṣṇa's pastimes, but now I cannot see them.
The grievance is not about geography. It is about access. What the retrieval took was not a hill; it was a view — an ongoing perception that was in progress and has been interrupted, by hands, mid-way.
Which puts the household in a position nobody in the corpus ever states. Every time they bring him back, they end something he was in the middle of, and they have no way of knowing where in it he was. A rescue from a sand dune and an interruption of the tradition's highest state are, on these facts, the same act — and the corpus supplies no principle for telling which one anybody was performing on any given night.
The three instances are not random terrain and that is worth extracting.
A dune stands in for a hill. A river stands in for a river. A garden stands in for a grove. In every case the substitution is between things of the same kind — the overwrite is not producing Vṛndāvana out of nothing, it is finding the local object in the right category and assigning it.
Which tells you something specific about the mechanism. This is not hallucination in the sense of perceiving what is not there. The dune is there. The sea is there. What has failed is not the perception but the identification — the step where a perceived thing gets matched to a particular, and the particular here is always drawn from one place.
And it means the condition needs the world. He does not lie in a dark room seeing Vṛndāvana; he walks into a landscape and the landscape supplies the raw material. Purī, with its beach, its dunes, its gardens and its water, is unusually well stocked for this — and the corpus never remarks that the town he could not leave was also the one place that could furnish the substitution constantly.8
The technical vocabulary makes this unsurprising, and that is either its strength or its problem.
The taxonomy's claim about the state of separation is that in separation from Kṛṣṇa one experiences the manifestation of Kṛṣṇa, and that the boundary between the one longing and the one longed for fails.9
A doctrine that says absence produces manifestation has already predicted a man who finds Vṛndāvana in Orissa. The overwrite is not an anomaly the theology has to accommodate; it is the theology's own claim, instantiated, in terrain.
Which is the recurring difficulty with this corpus and it should be stated rather than resolved. The prediction is genuinely prior — Madhya 23's taxonomy is set out in the instructional chapters, before the Antya-līlā's events are narrated. That is the shape of a framework doing real explanatory work.
And the record was assembled by people holding the framework. The same property that makes the prediction impressive makes it unfalsifiable: there is no observation of a man in that state, made by those observers, that would have come out any other way. Both facts are true and neither cancels the other, and the corpus is not in a position to help.
Layer A supplies the narrator's generalisation, the three episodes with their physical detail, and the complaint about having been brought from Govardhana.10
Layer B expands the material into the doctrine of the gopīs' separation and the identity of the holy dhāma — that Vṛndāvana is not a location on a map but a spiritual reality available anywhere to a qualified perceiver.11
Worth marking, because Layer B's move is elegant and it dissolves the page. On the commentary's account there is no misrecognition at all. If Vṛndāvana is not primarily geographic, then a man who perceives it at Purī is perceiving accurately, and the devotees who brought him back were the ones with the impoverished view.
That reading is internally consistent and it is not the verse's. Layer A has a narrator saying He took it for granted that they were Govardhana — a sentence written by somebody who knows they were not.
You are somewhere and not in it. Notice whether you can still answer where you are. The ordinary version of absorption keeps the floor; the thing this page describes is what it looks like when the floor goes, and the diagnostic is not vividness but whether you would act.
You are retrieving somebody from a state. Ask what the retrieval looks like from their side. This corpus records a man calling his rescue an abduction, and records it without answering — which is more than most caregiving accounts manage, and the complaint is usually there whether or not anybody writes it down.
You are assessing a framework that predicted what you observed. Check whether the prediction was prior and whether the observers held it. Both can be true, and they pull in opposite directions: priority is evidence for the framework and shared possession is evidence about the observation. Most disputes about this collapse one into the other.
Somebody's condition needs specific raw material. Look at what their environment is furnishing. A man who substitutes landscape for landscape requires landscape, and the corpus's protagonist spent his last years in the one town whose beach, dunes, gardens and sea could supply the substitution continuously.
The evidence is a narrator's explicit generalisation plus three episodes verified at full text, each involving action rather than comparison, and one first-person complaint about the direction of displacement.12
The tension the source doesn't close: the commentary's reading dissolves the misrecognition and the verse's language asserts it. Took it for granted that they were Govardhana is written by somebody who knows the difference; Vṛndāvana is not a place on a map denies there is a difference to know.
A second one. The doctrine predicted this before the events were narrated, and the people who recorded the events held the doctrine. Priority and shared possession are both real here, and the corpus offers no way to weigh them against each other.
Open: what he would have said if asked, at the moment, where he was. Every report is of what he did and what he said afterwards, and nobody in six chapters is recorded putting the question to him while it was happening.
This page and Mādana and Mohana are the prediction and the instance. That page has the claim that separation produces manifestation and the boundary fails; this one has a man running at a sand dune — and the fit is exact enough to be either the taxonomy's vindication or its circularity.
Against The Body That Changes Shape it is the same episodes read for a different variable. That page follows the joints; this one follows the terrain — and Antya 14 supplies both, which is why the corpus's own chapter divisions are a poor guide to what is in it.
And it belongs beside Care as Containment, which holds the community's side of every retrieval here. There the pattern is a household managing its founder against his own impulses and calling each instance service; here is the one place the founder says what he thinks of it — who has brought Me here — and the two accounts of the same acts have never been set against each other.
Expectation Assimilation — that page's mechanism is that the brain does not evaluate an experience in a vacuum: it compares what happens against what was predicted, and the prediction sets the reference point that determines how the experience feels. Show the cheese melting perfectly before the taste arrives, and the taste arrives as exactly what was promised.
The overwrite is that mechanism past the point where it is doing evaluation, and the pairing locates the boundary.
In the commercial case, expectation shapes the verdict on a percept that is itself stable. Everybody tastes the cheese. What the prior controls is whether it registers as disappointing or as promised — an evaluative layer sitting on top of an identification nobody disputes.
Here the prior has reached the identification itself. There is no stage at which the dune is registered as a dune and then favourably assessed. The prediction has consumed the step where the object gets named, and once that happens, evaluation is not occurring at all — the man is not judging the dune well, he is at Govardhana.
Which gives the finding neither page reaches. Expectation operates on a stack, and the deeper it reaches the less the person can do about it. A prior shaping evaluation is arguable — you can be told the cheese is ordinary and taste it again. A prior shaping identification is not, because there is no independent percept left to appeal to, and the corrective everybody reaches for — look again — is exactly the operation that is compromised.
And it explains why nobody in the corpus argues with him. Argument is available against a verdict and useless against a percept, and the devotees behave accordingly: they never once tell him it is a dune. They chant into his ear and carry him home.
Sharpest implication. Expectation operates at different depths, and the depth determines whether anything can be done. A prior that shapes evaluation can be argued with; a prior that has consumed the identification step cannot, because the corrective is look again and looking is the compromised operation. This corpus's community worked that out in practice and never stated it: in six chapters nobody tells him it is a sand dune.
Generative questions.