You stub your toe and you say something.
It might be a name. In half the languages of the world the thing people shout when a shelf lands on their foot is a god's name, and not one of the people shouting it is praying. It has been worn down by a thousand years of use into pure noise — a sound the body makes under load.
Now imagine a tradition that says: that still counts.
Not as a nice thought. As a claim about mechanism, argued from cases, with a boundary drawn around it and a list of exceptions. That is Antya 3, and getting the boundary right is the single most easily botched point in this entire corpus.
Haridāsa Ṭhākura is asked how people outside the tradition will ever be delivered, and his answer is that they already are being.
The yavanas — Muslims, in this text's usage — habitually say hā rāma, hā rāma, an ordinary exclamation with no devotional content whatever.1 And Rāma is a name of God.
So they are chanting, constantly, without knowing it, without meaning it, and often while doing something the tradition would deplore.
The general principle is stated flatly:
"The letters of the holy name have so much spiritual potency that they act even when uttered improperly."2
That is a claim about syllables. Not about faith, not about understanding, not about the heart. The potency is in the sound, and the sound does not check who is producing it or why.
The supporting cases are chosen to make the point unmissable. A mleccha gored by a boar, crying out in pain. Ajāmila, at the moment of death, calling his son — whose name happened to be Nārāyaṇa.3 Neither man is addressing God. One is in agony and one is calling a child.
Read only that far and you have the corpus saying intention is irrelevant. Which is a startling reversal, because the same corpus rests its central distinction between love and lust entirely on the direction of the actor's intent.4
But the boundary is drawn in the same speech, and it is the reason this page exists.
Chanting with offence — for bodily benefit, for wealth, for followers, out of greed, out of atheism — "will not produce the desired result very soon."5
So the doctrine is not intent doesn't matter. It is considerably more precise than that:
No intent works. Wrong intent does not.
Absence of intention — the cry of pain, the father calling his son, the habitual exclamation — is harmless and effective, because there is nothing in the way. The syllables arrive unobstructed.
Instrumentalising intention is an offence. The person chanting in order to get something has interposed a purpose between themselves and the sound, and the purpose is what blocks it.
That is a coherent position and it is not the position most summaries of this material report. It is also, structurally, the same claim the corpus makes about love: what disqualifies is not the presence of a self but the aiming of the act at that self. The name doctrine and the direction test are the same rule in two registers.
The line is worth walking, because the corpus's own hardest case sits right on it.
Effective: the boar's victim. Ajāmila. The yavana's exclamation. A child imitating a sound. Somebody reading the syllables aloud from a page without belief.
Effective, and this is the sharp one: the meat-eaters of Ādi 17 who chanted "Hari, Hari" to mock the Hindus, and then could not stop — "I have no desire to say it, but still my tongue says it. I do not know what to do."6
Derision falls on the no intent side, not the wrong intent side. The mockers were not instrumentalising the name to obtain anything; they were making a noise at somebody. Nothing was interposed, so the syllables did their work.
Not effective: the man who chants to become wealthy, to acquire followers, to neutralise the consequences of his own conduct. Those are the corpus's named offences, and each of them has a purpose standing between the chanter and the sound.
Which produces the doctrine's most uncomfortable ranking. A hostile mocker is better placed than a sincere devotee with a motive.
That is not a paradox once you see what the criterion is tracking. It is not tracking goodness, or sincerity, or effort. It is tracking whether the sound is being used. Contempt does not use the name; it discards it. A prayer for wealth uses it, precisely and instrumentally, and the using is what interposes.
The same speech makes an argument about loudness that reads as a devotional flourish and is actually a direct consequence of everything above.
Chant loudly, because others will hear it.7
If potency is in the syllables and reception requires nothing of the receiver, then every person within earshot is being acted upon, regardless of what they think they are doing. An audible practice is not a performance. It is a distribution mechanism with a physical range.
Which is why this tradition's central practice is congregational and outdoors rather than silent and interior, and why the corpus ranks vocal chanting above silent remembrance.8 Those are not preferences about temperament. They follow from the phonetic claim: a silent practice reaches one person and an audible one reaches everybody within a hundred yards, and the hundred yards is the whole point.
It also explains why the extraction record folded the loudness material into this page rather than giving it its own. It is not a second doctrine. It is the first doctrine's delivery specification.
And then the chapter does something rare: it lets someone state the obvious objection.
If the name works on anybody who hears it, and everybody hears it, then everybody is liberated — and Caitanya's stated worry is logistical: "If all living entities were liberated, the entire universe would be devoid of living beings."9
That is not the objection a modern reader expects. It is not this makes salvation too cheap or this removes the need for practice. It is a population-balance problem, and it is raised by the figure the book exists to worship.
The answer given is a conservation argument, and it belongs on its own page.10 What matters here is that the corpus saw the consequence of its own phonetic doctrine and took it seriously enough to need an answer — which is a point in favour of the doctrine being meant literally rather than devotionally.
Layer A gives the mechanism, the cases, the boundary and the loudness argument, in a devotee's mouth, as an argument in a conversation.
Layer B is enthusiastic and extends the yavana material considerably, generalising it to anyone using the name in any language and pressing the point that Western readers unfamiliar with the tradition are in the same position.11 That is a legitimate extension of the verse's own logic.
What is worth flagging is the term itself. Yavana in this text means Muslims, and the passage is warm — it is an argument that people outside the tradition are being delivered without knowing it, offered as good news. That is the friendlier of the two things this corpus does with its Muslim neighbours, and it sits ten chapters from the other, where a Muslim judge is made to disparage his own scripture.12 The reader should hold both.
You catch yourself repeating something you don't believe — a slogan, a phrase from a job you left, a formula you learned as a child. Notice that it is still doing something. The tradition's claim is that repetition operates below assent, and whatever you make of the theology, the observation about repetition is not in dispute.
You are trying to install something in a group and you are working on their agreement. Consider working on their exposure instead. The corpus's whole practical architecture rests on the claim that being within earshot is sufficient and consent is not required, and every advertising department in the world agrees with it.
You are doing a practice in order to get an outcome. This doctrine says that is the one configuration that fails — not scepticism, not distraction, not derision, but purpose. Sit with how counterintuitive that is, and with how well it describes what happens to a practice you have started using as a tool.
Somebody mocks the thing you care about, using its own words. The corpus's position is that they are closer than the earnest person with an angle. You do not have to agree, and it is worth noticing that you probably treat it the other way round.
The evidence is a single sustained speech containing the mechanism, three worked cases, the explicit offence-list boundary and the loudness argument.13 It is one of the more tightly argued passages in the corpus.
The tension the source doesn't close — and it must not be flattened, because it is a refinement rather than a contradiction. The corpus holds a phonetic-efficacy theory of the name and an intention theory of devotion. Antya 3's boundary reconciles them by distinguishing absence of intent from instrumentalising intent. That reconciliation is real and it is the corpus's own. What it does not explain is why the same distinction is not applied in the disciplinary chapters, where behaviour is judged without any reference to it.
A second one. If the name acts on anyone within earshot, the corpus's severe rules about qualification, initiation and eligibility govern a practice that is, by its own account, already working on everybody involuntarily. The gate and the flood are in the same tradition and the corpus never asks what the gate is for.
Open: the doctrine's uncomfortable ranking — mocker above motivated devotee — is entailed by the boundary as stated and is never drawn by the corpus. Nothing in the range read either endorses or blocks it.
This page and the Ādi 17 mockery material — the-tongue-that-would-not-stop, not yet built — were assessed for merger during the consolidation pass and the merger was declined, which is upheld here. They are the same doctrine and one is mockery and one is habit, and the corpus's own distinction — no intent works; wrong intent does not — runs precisely through the gap between them. Merging would have produced one page that could not exhibit the boundary.
Against Kāma and Prema: The Direction Test the relation is closer than it looks. That page's criterion is whose satisfaction an act aims at. This page's boundary is whether a purpose has been interposed between the chanter and the sound. Both rules disqualify the same thing — the act bent toward the actor's own account — and the corpus states them in registers so different that it never notices they are one rule.
And it sits directly on Nāma-saṅkīrtana Above Smaraṇa, which it explains. That page holds the ranking of vocal chanting above silent remembrance. The phonetic-efficacy claim is why: a silent practice cannot reach a bystander, and reaching bystanders is the mechanism.
The Magic Power of Words and Formulas — Le Bon's claim, quoted on that page, is that a pyramid taller than Cheops could be built from the bones of men killed by words and formulas: short syllables, vague enough to mean what the hearer wants, charged with image, immune to argument. Deus vult, Lebensraum, shareholder value.
Both pages hold that certain sounds operate independently of the hearer's assent. The disagreement is about what the sound is doing.
Le Bon's formulas work by being semantically empty and emotionally loaded — they carry no checkable content, so nothing in them can be refuted, and they attach to whatever the hearer already wants. The mechanism runs through meaning: the phrase means something, vaguely, and the vagueness is what makes it unstoppable.
The nāmābhāsa doctrine claims something stronger and stranger: the sound works when it means nothing at all to the speaker, and even when it means something else entirely. Ajāmila is calling his son. That is not vagueness; that is a completely different referent, correctly understood by everybody present, and the claim is that the syllables act anyway.
Put the two together and you get a spectrum neither page draws, and it is genuinely useful. At one end, sounds that work because of what they mean. In the middle, Le Bon's formulas, which work because of what they vaguely mean and cannot be pinned to. At the far end, a claim that sounds work irrespective of meaning altogether. The three are not rivals — they are three different mechanisms, and the practical question about any repeated phrase in your environment is which one you're dealing with.
That matters because the defences differ completely. A meaning-dependent phrase is answerable: you can argue with it. Le Bon's formula is not answerable but is unmaskable — you can show that it means nothing, and the showing works. A genuinely meaning-independent sound has no defence at all, because there is nothing to answer and nothing to unmask; it works on people who have seen through it, which is exactly what Ādi 17's mockers discovered about their own tongues.
Whether anything actually occupies that far end is the open question. The corpus is entirely sure something does, and the honest position from outside is that the only well-evidenced part of the claim is the involuntary repetition — the jingle, the slogan you cannot unhear, the phrase from a job you left. That much is not in dispute by anybody. What it produces is.
Sharpest implication. The one configuration this doctrine says fails is the one everybody assumes succeeds: doing the practice on purpose, in order to get the result. Mockery, accident, agony, habit and ignorance all work. Purpose interposes, and a tradition that says so has ruled out most of what its own institutions are built to encourage.
Generative questions.