A Sanskrit scholar will tell you the letter ya is pronounced "ya." But in the Bengali tradition the teacher comes from, the same letter is sounded "ja" — and when he chants the Caṇḍī, he says "ja," because that's how his guru said it, and his guru's guru before that.1 A textbook would call it a mistake. The tradition calls it fidelity. The teacher's instruction is blunt and a little startling: say it the way your guru says it — not the way the grammar book says it, not the way "correct" Sanskrit says it, but the way it came down to you through the living chain of voices. The pronunciation you inherit is the mantra, because the mantra is what was transmitted, not what's printed.2
This is the principle that, in mantra and sacred recitation, lineage transmission overrides textbook correctness. When a regional or lineage pronunciation diverges from "standard" Sanskrit (the Bengali ya→ja shift is the teacher's example), the practitioner follows the lineage's way — "say it the way my guru says it" — because the authority of the mantra runs through the chain of transmission, not through a grammar.3 The living voice that handed you the sound is the authority, not the printed page.
This sits against the modern instinct that there's a single "correct" pronunciation recoverable from the text. The tradition's instinct is the opposite: the transmitted form is the real one, even where it has drifted from the "original," because what you received through an unbroken chain of voices carries the lineage's power, and a "corrected" version — however textually accurate — breaks the chain.4
Walk the logic, because it inverts how we usually think about correctness. We assume the original is the authoritative form and any drift is corruption — so the closer you get to the "true" ancient pronunciation, the better. The lineage principle says: the authority isn't the original, it's the transmission — the unbroken chain of guru-to-disciple voices that carried the sound to you. Each link in that chain blessed and empowered the form it passed on, drift and all. So the pronunciation your guru gave you is dense with the accumulated transmission of everyone who said it before — and a textbook "correction" strips all of that out in exchange for a sterile accuracy that no living chain ever blessed.5
This is why "say it the way your guru says it" isn't anti-intellectual stubbornness. It's a claim about where the power lives: in the living transmission, not the dead text. A mantra is not information you decode from a page; it's a charge passed hand to hand, voice to voice, and the way it was passed to you — including its lineage-specific shifts — is the charged form. To "correct" it toward the textbook is to swap the charged, transmitted sound for an uncharged, reconstructed one. The ya→ja shift isn't a flaw the Bengali tradition tolerates; it's the fingerprint of the lineage, the mark that this sound came down this particular living channel — and chanting it that way is what keeps you inside the channel.6
And there's a humility in it. Following your guru's pronunciation over your own scholarly judgment is a deliberate subordination — you're saying the chain knows better than my analysis, I receive rather than correct. That stance is itself part of the transmission: the disciple who "corrects" the guru's sound has already stepped outside the relationship that makes the mantra work.
This page hands the vault the transmission-over-text principle — the claim that in living traditions, the authority is the unbroken chain of voices, not the recoverable original. It sharpens the vault's many transmission pages (guru-student transmission) with a concrete, almost mundane instance (a single shifted consonant). And it offers a portable tension for any tradition: the conflict between fidelity to the source (textual accuracy) and fidelity to the lineage (transmitted form) — two kinds of faithfulness that can point in opposite directions.
Stay with the smallness of the example, because that's its power. It's not a doctrine or a practice at stake — it's one consonant, ya versus ja. A scholar could prove, with solid linguistics, that ya is the "correct" old Sanskrit value and the Bengali ja is a regional drift. The evidence is on the scholar's side. And the tradition still says: chant ja, because that's how your guru chants it.
This tiny example tests the whole question of where authority lives, precisely because it's so small that no doctrine rides on it — so the only thing at stake is the principle of fidelity itself. If the tradition would "correct" even a trivial regional consonant toward the textbook, it would be admitting that the text outranks the transmission. By refusing to correct even this — by holding the lineage pronunciation even where the scholar is demonstrably right about the "original" — the tradition declares its entire epistemology: the living chain outranks the recovered text, all the way down to a single consonant. The case study's payoff is that the smallness is the proof. A tradition that follows its lineage even where following the text would be easy and defensible is a tradition that has genuinely located authority in the transmission rather than the source — and the ya→ja shift, trivial in itself, is the load-bearing demonstration of that.7
This transfers to any domain where you're a recipient of a living transmission — a craft, a practice, a tradition handed down through people rather than only through texts. The instruction: learn to tell when to receive and when to correct, and default to receiving inside a living lineage.
You've been taught something by a living teacher in a way that diverges from the "book-correct" version — a technique, a pronunciation, a method that the textbook would mark wrong. Your educated instinct is to correct toward the source. The lineage principle counsels restraint: ask first whether you're inside a living transmission whose authority is the chain itself (in which case the transmitted form is the right one, drift and all) or merely reconstructing from texts (in which case accuracy matters more). Inside a living lineage, receive the form as given — including its idiosyncrasies — because those carry the transmission's charge, and "correcting" them steps you outside the chain. The discipline is a deliberate subordination: the way it was handed to me is the way I'll hand it on, holding fidelity to the living channel above fidelity to the reconstructed original. And the self-honesty: notice when your urge to "correct" is really an urge to step outside the relationship and be your own authority — which is exactly the move that breaks the transmission.
Two failures, opposite. Fail toward textualism and you get the pedant: "correcting" the living transmission toward the textbook at every turn — chanting the "right" Sanskrit your guru never used — and in the name of accuracy breaking the chain, swapping the charged transmitted sound for a sterile reconstructed one no living lineage ever blessed. The tell is fidelity-to-the-source that overrides fidelity-to-the-teacher, scholarship used to step outside the relationship. Fail toward inertia and you get the frozen error: elevating every drift and mistake to sacred status, so that genuine corruptions, errors, or even harmful distortions get preserved forever under "that's how my guru said it" — fidelity weaponized against any correction, even necessary ones. The tell is a tradition unable to distinguish a lineage fingerprint (the meaningful transmitted form) from a genuine error (a mistake that crept in), treating both as untouchable. The form needs discernment: inside a living transmission, receive the form as given (don't pedantically correct the lineage's charged idiosyncrasies), and retain the wisdom to tell a transmission's fingerprint from a corruption that should be addressed through the lineage itself, not by a disciple's unilateral "fix."
The principle that lineage pronunciation governs mantra-recitation (transmission over textbook) is well-attested across living mantra traditions; the Bengali ya→ja shift is a real, well-known regional feature.8 The teacher's "say it the way your guru says it" is a faithful articulation. Open question: where is the line between a lineage fingerprint (a meaningful transmitted form to be preserved) and a genuine corruption (an error that crept in and should be corrected) — and who gets to decide, given that the principle itself forbids the individual disciple from unilaterally "correcting"? The principle that protects the transmission also disables the mechanism for fixing real errors in it.
One teacher, so the tension is between this transmission-over-text principle and the vault's critical-historical, text-recovering scholarship (Samuel/White and the broader philological tradition), whose whole project is to recover the original behind centuries of drift. The lineage principle says the transmitted form is authoritative because transmitted; the philological project says the transmitted form is evidence to be corrected toward the reconstructed original. The convergence: both take the transmitted forms with deep seriousness (the philologist studies the drift the practitioner reveres). The split: is drift corruption (to be corrected, philology) or charge (to be preserved, lineage)? The split reveals what neither states alone — that "fidelity" splits into two incompatible faithfulnesses: fidelity to the source (recover the original) and fidelity to the channel (honor what was handed down), and that a living tradition and a critical scholarship can study the same pronunciation while wanting opposite things from it. The practitioner chants ja to stay in the channel; the scholar notes ja as the drift that proves the channel's history — and both are being faithful, to different masters.
These domains connect because they all confront the difference between fidelity to an original and fidelity to a transmission — and the question of when a deviation is a flaw versus a meaningful inheritance.
The Sharpest Implication If the authority of a mantra lives in the unbroken chain of voices rather than the recoverable original — if "say it the way your guru says it" outranks the grammar book even when the grammar book is demonstrably right — then our deep modern assumption that the source is the authority and transmission is just lossy copying is, in living traditions, exactly inverted. The disturbing implication is that fidelity is not one thing but two, and they can point opposite ways: you can be faithful to the origin (recover the pure original) or faithful to the channel (honor what was handed to you), and a single shifted consonant can force you to choose. The further edge: your educated reflex to "correct toward the source" may be, inside a living transmission, an act of defection — the move that steps you outside the relationship and breaks the very chain whose power you were trying to access. Sometimes the faithful thing is to say it "wrong," on purpose, because that's how it reached you.
Generative Questions