Imagine a student monk who has waited weeks for ten minutes. No phone, no books to fall back on, no second chance for a long while — just a single appointment with a teacher who may answer his most pressing question with a remark that seems to refuse the question entirely. He cannot record it. He cannot publish it. So afterward, in the dark, he scratches down a few lines for himself — not a book, not scripture, just a knot tied in a handkerchief so he will not forget what happened in that room. That private scrap, multiplied across many students and many interviews and bound loosely together, is what the Chinese call Yen Li and the Japanese call Goroku — "sayings." Dukes uses the category to explain what kind of document the Hsieh Mai Lun really is, and the answer reframes the whole text: it is not a treatise. It is somebody's memory aid, never meant for our eyes.1
A Goroku (Chinese Yen Li), in Dukes's account, is a record of students' questions and a teacher's answers, "created, most likely secretly, as an 'aide mémoire' for the students' personal use."2 The defining trait is where it sits: these aids "form a unique category of spiritual literature unaccounted for within the earliest canonical divisions of the sacred records."2 They fall outside the canon by design — not because they failed to make the cut, but because they were never candidates. The canon is public, authorised, ordered. A Goroku is private, ad hoc, and personal: the residue of a living oral exchange, written down only so the exchange would not evaporate.
This matters for reading the Hsieh Mai Lun, because it sets the genre expectations correctly. A Goroku "reflects a wide range of students' enquiries" — some basic, some subtle — and its parts do not necessarily belong together.3 Some queries, Dukes notes, "are merely preludes to longer teachings and are probably interpolated accounts from other occasions added later in order to explain a point."3 So the text is a patchwork: this question from one interview, that answer spliced in from another, the seams left where someone later stitched a clarification across a gap. You are not reading an argument. You are reading a notebook.
Here is the counterintuitive part. In the Mikkyo (esoteric) sects, oral teaching transmitted directly from a living master is called Kuden or Denbo, and within Buddhism this oral-based teaching "is held to be more reliable and accurate than literal renderings."4 So the private, unpublished, fragmentary Goroku is closer to the trusted source than a clean published sutra — because it preserves the live transmission rather than a polished literary reconstruction of it. The canon's order is the very thing that makes the canon suspect to this tradition: order implies editing, editing implies distance from the room where the teaching actually passed.
That inverts the usual hierarchy of evidence. A scholar prizes the published, datable, cross-checkable text and distrusts the private scrap. The Mikkyo logic prizes the private scrap precisely because it is closer to Kuden, and treats the published text as a downstream literary artifact. The Goroku's defects by canonical standards — unauthorised, disordered, incomplete — are its credentials by oral-transmission standards. The handkerchief-knot beats the printed book.
And this is also why such a text reads strangely. Dukes warns that in many sections "Bodhidharma is replying supra-logically in order that the student may realize that even asking a serious question is already missing the point."5 The answers refuse the questions on purpose. A Goroku preserving that refusal looks, to an outside reader expecting a treatise, like a broken or evasive document — when in fact it is doing exactly what its genre exists to do.
This page sets the evidential frame for every other page drawn from the Hsieh Mai Lun. Hsieh Mai Lun: Mind Is Buddha and Movement Is the Source of No-Mindedness both quote a text that, on this page's account, is a private notebook of uncertain provenance — so any claim sourced from it inherits the Goroku's status: trusted-as-oral by the tradition, unverifiable by outsiders. It also gives Guru-Student Transmission as Consciousness Transfer its documentary shadow: the Goroku is what Kuden leaves behind on paper, the trace of a transmission that was meant to happen mouth-to-ear. And it sharpens Wandering Acarya Transmission Mobility — a teaching that travels orally needs exactly these portable private scraps to survive the journey when the master is gone.
The most checkable — and least checkable — passage is Dukes's own account of where his Hsieh Mai Lun text comes from, and it is worth deconstructing because it is the whole evidential foundation laid bare. His version, he says, integrates two different Chinese works.6 The first is the Tamo Chaan Ching ("An Account of the Meditation Methods Used by Bodhidharma"), which came to him by a route that is impossible for an outsider to follow: "a Chinese manuscript... presented to me by the Ajari Senda of Okinawa in traditional Chinese Chuan (manuscript rolls)."6 A named individual, a private gift, manuscript rolls — and no public archive, no catalogue number, no second witness. The second source is the Tamo Hsieh Mai Lun, which Dukes describes very differently: "a work preserved in several Chinese and Korean monasteries and recently rediscovered in modern archaeological excavations of the Tun Huang Cave Temples."6
Notice the asymmetry, because it is the case. The Tun Huang source has a public, corroborable pedigree — the Dunhuang cave-temple manuscripts are a real, famous, scholar-accessible archaeological corpus, and "several Chinese and Korean monasteries" is in principle checkable. The Ajari Senda manuscript rolls have a private pedigree that terminates in a single named donor and Dukes's own possession. One source you could, in theory, go and verify; the other exists only on Dukes's word and in his hands. He has "integrated the meanings of all the versions into this one work," which means the public Tun Huang text and the private Okinawan rolls are blended in the final product — and once blended, you cannot tell which claim rests on which.7 The verifiable and the unverifiable are stirred into a single cup. That is the provenance situation in one image, and it is exactly the situation the whole Kongoryuji-warrant flag exists to mark.
You pick up a translated text attributed to an ancient master and the first thing you do is ask what kind of document it is. Not "is it true" — "what is its genre and where did it come from." You find Dukes's answer: it is a Goroku, a private aide-mémoire, fragmentary by nature, parts spliced across occasions, never meant for publication. Good — now you read its disorder as a feature, not a defect, and you stop expecting a clean argument. Then you trace the provenance, and you sort the sources into two piles. One pile has a public address — Tun Huang, named monasteries, an archaeological corpus you could write to. The other pile ends in a person's hands: a named Ajari, manuscript rolls, a private gift. You keep the piles separate in your mind even though Dukes has blended them on the page. When you later cite a striking line from this text, you carry the genre and the provenance with the quote like a label on a specimen jar — Goroku, oral-tradition-trusted, part-public part-private, integrated, unverifiable at the seams. You have not dismissed it. You have shelved it correctly.
The genre claim — that Yen Li / Goroku are private aides-mémoire outside the canon — is plausible and broadly consistent with how the Chan/Zen yulu ("recorded sayings") literature is understood in mainstream scholarship, so this part is not idiosyncratic to Dukes. [PLAUSIBLE — needs corroboration] on the specifics he attaches to the Hsieh Mai Lun.
The load-bearing flag is provenance. The Tamo Chaan Ching reaches Dukes via "the Ajari Senda of Okinawa" as private manuscript rolls, with no public corroboration possible; the Tamo Hsieh Mai Lun he ties to the genuinely public Tun Huang corpus.6 These are two very different evidential statuses blended into one translated text. [DO NOT FILE as fact — document-don't-endorse] the authenticity of the privately-held Okinawan manuscript as a sixth-century Bodhidharma source: it cannot be examined, dated, or cross-checked by anyone outside Dukes's tradition. The vault records that he claims it; the vault does not vouch for it.
Dukes is admirably candid about the translation hazards — no punctuation, no tense, no number-marking in the medieval Chinese originals, characters obsolete in both modern Chinese and Japanese, resolved only by consulting "dictionaries of medieval Chinese in the Kongoryuji Library."8 That candour is real and to his credit, and it also quietly compounds the provenance problem: even the public Tun Huang text passes through a translation whose decisive reference tools sit in his own temple's library. The chain never fully leaves Kongoryuji.
Open: which specific teachings in the vault's Hsieh Mai Lun pages rest on the public Tun Huang manuscript and which on the private Okinawan rolls? Dukes's integration makes the question unanswerable from the published text — and that unanswerability is itself the finding.
This page is where Dukes the insider and the vault's outside critics meet most directly, because the dispute is purely evidential. Samuel and White would have no quarrel with the concept of a Goroku — recorded-sayings literature is well documented. Their question lands on the Ajari Senda rolls: a sixth-century teaching surviving as private manuscript rolls handed to one practitioner-scholar by one named donor, integrated into a published translation, is precisely the kind of claim their method cannot accept and cannot disprove. Dukes's warrant is initiatory and possessive — the decisive manuscript lives in or passed through his hands and his temple's library. The outside scholar's warrant is public and reproducible. Reading them together produces the standing caution this whole re-ingest preserves: where Dukes's evidence is the public Tun Huang corpus, his claims are checkable; where it is the private Okinawan rolls, they are not — and his own integration of the two means the vault must read every Hsieh Mai Lun citation twice, once as possibly-public and once as possibly-private, because the text itself no longer tells them apart. The convergence is on the genre; the unresolved tension is the cupboard.
Unverifiable Provenance as Authority Claim (behavioral-mechanics). A privately-held manuscript that cannot be examined is not only an evidential gap — it can function as a source of authority. The behavioral-mechanics frame names the move: scarcity and inaccessibility manufacture prestige, and "I possess the document you cannot see" is a position of power, not merely a limitation. The structural parallel to Dukes's Ajari Senda rolls is exact — a secret, named-lineage manuscript is simultaneously the warrant for the teaching and a thing no rival can audit. But the handshake produces an insight neither domain reaches alone, and it is a genuinely double-edged one. The behavioral-mechanics reading would treat the private manuscript with flat suspicion — convenient, unfalsifiable, self-serving. Dukes's Goroku frame supplies the counter-case the cynical reading cannot generate: in an oral tradition that genuinely trusts Kuden over published text, the private, unpublished, inaccessible source is not a power-grab but the expected form of the most trusted material — the canon being the suspect document, not the scrap.4 So the same artifact — inaccessible private manuscript — reads as either fraud-structure or authentic-transmission depending entirely on whether the surrounding tradition's epistemology is sincere. The combined insight: inaccessibility cannot, by itself, distinguish a genuine oral lineage from an authority racket, because both produce identical documentary signatures. The vault needs that precisely so it neither dismisses Dukes as a fraud nor credits him as proof — the signature is the same either way, and only corroboration the lineage structurally cannot provide would tell them apart.
Oral Transmission Versus the Written Record (cross-domain). The Mikkyo claim that Kuden (oral, living transmission) is "more reliable and accurate than literal renderings" is a sharp instance of a tension this cross-domain page tracks across many traditions: the written record is checkable but dead, the oral transmission is living but unverifiable, and the two epistemologies rank evidence in opposite orders.4 The structural parallel is the Goroku as the hinge object — a written trace of an oral event, valued because it stays close to the orality and distrusted (by outsiders) for the same reason. The insight neither side reaches alone: a written record's reliability is usually assumed to increase with editing, ordering, and publication — yet on the oral-primacy view those very operations move the text away from the trusted source, so reliability decreases with literary polish. That inverts the historian's instinct entirely. For a tradition that ranks Kuden first, the messy private Goroku is more trustworthy than the clean published sutra, and the cross-domain lesson is that "reliability of a source" is not a property of the document but a function of which transmission-epistemology you are standing inside — the same scrap is best-evidence or worst-evidence depending on the frame, and there is no neutral place to stand that settles it.
Sharpest implication: every teaching the vault draws from the Hsieh Mai Lun is sourced from a document that is, by Dukes's own account, a private memory-aid integrating a public manuscript with one nobody else can see — which means the citation's reliability is undecidable at the seam, and the honest move is to carry that undecidability forward with the quote rather than launder it away.
Generative questions: