There is a difference between someone who has the facts wrong and someone who has them backwards, and you can hear it in about two sentences.
The first person is missing something. Give them the information and they update, sometimes gratefully, sometimes not.
The second is harder. All the pieces are there. But the arrangement is reversed — the person who helped is remembered as the person who obstructed, the favour as an imposition — and every new fact you supply gets slotted into the reversed frame and comes out confirming it.
You cannot fix that with information, because information is not what is wrong. The parts are all present and the orientation is inverted, and no fact you add to a picture will turn it the right way up.
Madhya 20 has a two-word diagnosis for the human condition, and this is what it says.
Viparyayo 'smṛtiḥ — reversed memory.1
The claim is precise. A person's constitutional position is servant; what they take themselves to be is competitor. Not stranger, not enemy, not ignorant of the whole business — competitor, which is the servant's relation reversed rather than absent.
And the corpus makes it a memory failure specifically, not an error of judgement. The living entity is not mistaken about a fact it never had. It has forgotten one it did have, and the forgetting has produced a mirror image.
That is a much more particular claim than people are confused about God, and it has consequences the chapter does not draw.
Choosing memory as the category does three things at once.
It locates the content. A forgotten thing was once known and is still, in some sense, in there. This is why the corpus's remedy everywhere is awakening rather than instruction — the material is present and unavailable rather than absent.2
It removes the blame without removing the fault. Nobody is culpable for having forgotten. The condition is a misfortune, and the corpus's tone about ordinary people is correspondingly gentler than its tone about opponents, who are held to be actively resisting rather than merely turned around.
And it makes the reversal systematic rather than random. Errors scatter; a reversal is one operation applied to everything. Which explains why the corpus does not treat the human predicament as a list of mistakes to be corrected individually. There is one thing wrong, applied everywhere.
That is genuinely elegant: the diagnosis predicts the shape of the problem rather than merely naming it.
It also does something less comfortable, and the corpus uses it constantly without marking it.
If the condition is forgetting, then persuasion is reminding. You do not have to argue with somebody who has forgotten; there is no case to answer, because they are not holding a position — they are missing an access. The whole category of disagreement dissolves into not yet reminded.
Which is why this corpus can be simultaneously so confident and so unaggressive toward outsiders. It is not trying to defeat anybody's view. On its own account there are no views out there, only reversals.
And it makes objection unavailable as evidence. Somebody who hears the doctrine and rejects it has not produced a counter-argument; they have produced a sample of the condition. The rejection is what reversed memory looks like from outside.
This build has now met that structure at every one of the corpus's decisive points — failure proves offence, convincing symptoms prove counterfeiting, and now disagreement proves the diagnosis.3 It is not a lapse in one place. It is the shape of this corpus's principal claims, and this is the version applied to the reader before they have said anything.
The chapter attaches an image to the condition, and it is not the image you would expect.4
A criminal is held under water and raised. Held under, and raised.
Note what is not in it. There is no sentence, no term of years, no crime specified and no release. The comparison is to a punishment that consists entirely of repetition — and the repetition is the point of the analogy, because it is being offered as a picture of transmigration.
Three things follow that the corpus does not say.
The punishment is not corrective. Nothing about being held under water teaches the man anything. There is no lesson embedded in it and no behaviour it is training. It is duration without instruction.
And it is not proportionate. A sentence ends when the debt is paid; this does not end at all. Whatever the man did, the arrangement has no accounting in it.
And the man is passive throughout. He does not struggle, escape, or repent in the analogy. He is held and raised by somebody else, which puts the agency entirely outside him — in a chapter otherwise devoted to what a person should do.
Set beside the reversed-memory diagnosis, the two do not quite fit. A memory problem calls for a reminder. A man held under water needs to be let go. The corpus supplies both pictures within a few lines and they imply different interventions.
The corpus does have a remedy consistent with the first picture, and it is worth naming because it is what makes the diagnosis more than a description.
The name is the cue. The whole doctrine of the holy name's efficacy — that it works on somebody in pain, somebody calling their son, somebody mocking, somebody who does not know what they are saying — reads exactly as a retrieval claim.5 The content is intact; the right prompt restores it; and the prompt does not have to be understood, only sounded.
That is internally consistent and it is the strongest form of the corpus's position. A forgetting that can be undone by a cue explains why the remedy is so cheap, why understanding is not required, and why the tradition's central practice is a sound rather than an argument.
It also makes a prediction the corpus never tests: restoration should be immediate and complete, because that is what retrieval is. And the corpus's own material shows the opposite — practitioners chanting for years without the transformation arriving, a case the tradition handles by ruling that the practitioner has committed offences.6
Layer A gives the term, the servant-competitor reversal and the water analogy, compactly, without elaboration.7
Layer B expands the condition into an account of contemporary life and presses the competitor framing hard — the desire to be the enjoyer, the lord, the proprietor — as the diagnosis of a whole civilisation.8
That generalisation is not in the verse. The verse states a condition of the living entity as such. Reading it as a critique of a particular society at a particular date is an application, and a natural one for a commentator writing in 1974 to an audience he was trying to reach.
Worth marking because the two versions have different targets. Layer A's reversed memory is a universal condition nobody chose. Layer B's is something people around him were doing.
Somebody has the facts and has them backwards. Stop supplying facts. Everything you add will be absorbed into the reversed arrangement and come out supporting it, and the effort will convince you they are being stubborn when they are being consistent.
You have diagnosed somebody as forgetting rather than disagreeing. Notice what you have just done to their objections. It may be right. It also means nothing they say can now count against you, and you should be able to state what would.
You are being told your resistance is a symptom of the thing being described. This is the one move worth naming out loud, because it is not detectable from inside the conversation and it is the point at which a description of the world becomes unarguable.
You are stuck in a loop with no lesson in it. The corpus's own image is honest about this — the man in the water is not learning anything, and treating a repetition as though it were instruction is how people spend years extracting meaning from something that has none.
The evidence is a compact term with an attached analogy inside the Sanātana instruction, stated as settled doctrine.9
The tension the source doesn't close: the two pictures imply different remedies. Reversed memory calls for a cue; a man held under water calls for release. One puts the problem in the sufferer and one puts it entirely outside him.
A second one. A retrieval account predicts that the right cue restores everything at once, and the corpus's own practitioners chant for years without it — a case the tradition resolves by finding fault in the practitioner rather than in the account.
Open: what distinguishes reversed memory from a position. The corpus supplies no test, and without one the diagnosis applies to any disagreement whatever.
This is the negative image of Jīvera Svarūpa: Kṛṣṇera Nitya-dāsa and neither works alone. That page states what a person is; this one states what they take themselves to be, and specifies that the second is the first inverted. The pair is what lets the corpus describe the human situation as a single error rather than a condition.
It supplies the mechanism Sādhana Awakens What Is Already There presupposes. Dormancy needs something to have gone dormant; reversed memory names the operation that put it there.
And it sits against The Astrologer and the Buried Treasure as the same claim with the difficulty moved. The parable says knowing is not the problem and procedure is. This term says the knowing is precisely what is broken. Both are in the same chapter's instruction, and a practitioner cannot act on both.
The File Clerk in Your Head: How Memory Actually Gets Made — that page's central correction is that memory does not work like a video recorder. It is reconstruction, not retrieval: what comes back is assembled at the moment of recall, shaped by the frame you are in when you assemble it, and a substantial fraction of what feels like remembering is manufactured.
The corpus is making a memory claim in the strict sense, and setting it against the reconstruction model shows exactly what it has committed to.
A retrieval account says the content is intact and the access is broken. That is what viparyayo 'smṛtiḥ asserts — the constitutional position is unchanged, permanently, and only awareness of it has been lost. It follows that the right cue restores the whole thing, unaltered, and that is precisely the corpus's theory of the holy name: a sound that does not need to be understood, working on the mocker and the man in pain alike.
A reconstruction account has no such prediction available. If recall assembles rather than retrieves, then what comes back is built by the frame doing the assembling — and a person in the reversed frame does not recover the original. They rebuild it reversed, again, and the rebuild feels exactly like remembering.
Which is the finding neither page reaches. The two models differ on whether a cue is sufficient, and every tradition with a cheap universal remedy is committed to the retrieval model whether or not it knows. If restoration is retrieval, a sound is enough and the tradition's generosity is well-founded. If it is reconstruction, a cue delivered into an unchanged frame produces a reversed result — which would predict exactly the outcome the corpus reports and blames on the practitioner: years of the cue, faithfully applied, and no change.
The corpus's own evidence is more consistent with the second model than with the first, and its doctrine requires the first.
Sharpest implication. Diagnosing somebody as having forgotten rather than as disagreeing is the most humane framing available and the least answerable one, because it converts their objection into a specimen. The two properties are inseparable, which is why the gentlest traditions and the most unfalsifiable ones are so often the same traditions.
Generative questions.