A census counts houses, not travellers. Send a census-taker down a road and he records every standing dwelling, every registered household — and walks straight past the pilgrim sleeping in the ditch, the tinker, the wandering teacher with no fixed address. Years later a historian reads the census and concludes nobody lived on that road but the householders. He is not lying. He is reading the record correctly. The record was simply built to see one kind of life and blind to another. Dukes argues that the famous claim "Chinese Mikkyo died out after Kobo Daishi" is exactly this mistake — a fact about what the records could see, mistaken for a fact about what existed.1
This page records Dukes's debunking of a standard historiographical claim and the structural argument underneath it. The claim: Chinese esoteric Buddhism — the Mi Tsung (Japanese: Mi-Shu) — "died out" after Kobo Daishi (Kukai), the early-ninth-century figure who carried the esoteric teaching to Japan.2 Dukes's rebuttal is precise. What "died out" actually means, he says, is something much narrower: there appeared to be no "organized, centralized administrative group of temples" propagating the esoteric path after the Japanese patriarchs went home.3 The centralised institution was absent. The teaching was not.
The reason the two get confused is a difference in how Chinese and Japanese esoteric Buddhism organised themselves. In Japan — particularly in Kobo Daishi's school — the tradition built a highly hierarchical structure around "a central and dynamic teacher figure."4 The Chinese esoteric sects set no such precedent; they "never felt a need" for that central figure.5 Instead they produced wandering Acaryas — solitary itinerant teachers who could clarify the inherited iconographic and esoteric tradition to individuals face to face, showing what was essential and what was not.6 The essential dynamic of Chinese Mikkyo, Dukes says, "always remained with the solitary and wandering practitioner," modelled on the questing of the wandering Shakyamuni himself.7 A tradition carried by travellers, not housed in an administration.
Run the logic and the "died out" verdict collapses into a category error. Institutional record-keeping — especially the foreign accounts written by outsiders — sees institutions. It registers temples with administrations, central teachers, hierarchies, the things that file paperwork and leave a paper trail. A decentralised tradition whose whole transmission mechanism is a man walking from student to student leaves almost no institutional footprint by design. There is no head office to record, no central teacher to name, no administrative continuity to chart. So when the outside historian looks for the tradition and finds no central temple group, he writes "died out" — when what he has actually established is "left no centralised record."
Dukes then supplies the positive evidence the institutional record missed, and it comes from an unexpected angle: the Japanese pilgrims themselves. Examine the records of the Japanese monks who travelled to China to study the esoteric teachings after the time of both Dengyo Daishi (Saicho) and Kobo Daishi, and "there are plenty of teachers and temples fully functioning."8 The very tradition said to have died is documented alive and teaching — by the visitors who went looking for it and found it. He stacks on the wider picture: forms of Mikkyo had appeared in Tibet and were growing fast, spreading to Mongolia and re-entering China from the north, so the esoteric current was thriving across the whole region even as the centralised Chinese institution stayed invisible.9 His conclusion is a methodological one: a continuity of this kind, carried by wandering practitioners, "was difficult to discern from the 'outside' and was not amenable to being included in the later historical accounts of foreign countries."10 The Chinese sect continuity "existed long after the time of the Kobo and Dengyo Daishis" — despite some Japanese accounts saying otherwise.11
This page hands the corpus a general historiographical tool: the survival-versus-record distinction. It directly extends Wandering Acarya Transmission Mobility, giving the mobility thesis its sharpest consequence — mobility is not just how the teaching moved but why historians lost track of it. It reinforces Bodhidharma: The Revisionist Biography, where Dukes makes the same move at the level of one man: "the sources are late and quiet" is a fact about archives, not about transmission. And it offers a counter-frame to Guru-Student Transmission as Consciousness Transfer — a transmission that runs person-to-person rather than institution-to-institution is exactly the kind that record-keeping cannot see.
The decisive move is that Dukes refutes a Japanese claim with Japanese evidence. The "died out after Kobo Daishi" line comes, he notes, partly from Japanese accounts.12 His counter-evidence is the travel records of Japanese monks who went to China after Saicho and Kukai — and found teachers and temples fully functioning.13 Watch the structure. The obituary and the rebuttal come from the same culture, even the same kind of source, but they record different things. The obituary records an absence — no centralised institution like the one Japan built at home, so by Japanese organisational expectations the tradition looks dead. The pilgrim diaries record a presence — living teachers met in person, exactly the wandering-Acarya transmission the institutional frame couldn't register. The pilgrims went looking for the teaching and received it; the institutional historians looked for the organisation and didn't find it; both wrote down what they saw. The case study's lesson is that "died out" was never a finding about the teaching. It was a finding about the shape of the container — and a tradition that lives in travellers rather than temples will read as dead to anyone counting temples.
You are reading a survey history and you hit a flat sentence: "the tradition died out in the ninth century." Your pen wants to underline it as a date to remember. Instead you stop and ask the question this page drills: died out, or dropped out of the record? You look at what kind of thing the tradition was. Centralised, with head temples and a named patriarch — then "died out" might mean what it says. But this one transmitted through wandering teachers with no fixed address, no administration, no central figure. So you reach for a different source — not the institutional chronicle, but the traveller's diary, the pilgrim's notebook, the outsider who went looking for the teaching rather than the bureaucrat who catalogued temples. And there it is: teachers met, lessons received, decades after the official obituary. You don't cross out the survey's sentence. You annotate it: "absence of institution, not absence of teaching — see pilgrim records." The history hasn't changed. Your reading of the silence has.
Evidence base: a passage in Dukes's discussion of mandala transmission and the wandering Acarya, where he addresses the "died out" claim directly and cites the post-Saicho/Kukai Japanese pilgrim records as counter-evidence.14 The historiographical argument — that decentralised transmission is structurally invisible to institutional record-keeping — is sound and generalises well; it is one of the more transferable tools in the corpus.
[PLAUSIBLE — needs corroboration — single source; author has lineage interest] on the survival thesis itself: that an unbroken Chinese esoteric continuity persisted "long after" the ninth century through wandering Acaryas. Flag held, not resolved. Dukes has a direct stake here — his own lineage is a wandering-Acarya tradition, so a thesis establishing that such traditions survive invisibly is also a thesis establishing that his could have survived unrecorded. That doesn't make it wrong; the pilgrim-record evidence is real and the structural argument is genuinely strong. But the conclusion he most needs to be true and the conclusion he reaches are the same conclusion, and the page preserves that tension rather than smoothing it.
Open seam: Dukes shows the tradition was alive somewhat later via the pilgrim records, and that Mikkyo thrived regionally (Tibet, Mongolia, re-entering China from the north). But regional thriving in Tibet is not the same as Chinese sect continuity. The pilgrim records establish survival into the generations just after Kukai; the leap to "long after" leans partly on the Tibetan/Mongolian spread, which is a different lineage geography. How far the specifically Chinese wandering continuity actually ran is left open.
This page is where Dukes's greatest methodological strength and his deepest self-interest sit in the same argument, and the honest reading holds both. The strength is real: the survival-versus-record distinction is a genuine contribution, the kind of point a careful critical historian like Geoffrey Samuel would recognise and use — Samuel spends his career insisting that the absence of early textual evidence for a practice is not evidence of the practice's absence, which is this page's claim in a different key. On the mechanism, they converge cleanly. The split is the stake. Samuel deploys the silence-isn't-absence argument to stay agnostic — to refuse premature conclusions in either direction. Dukes deploys it to reach a positive conclusion: the tradition survived, and (elsewhere in his work) it survived into the lineage he himself holds. The same tool that keeps Samuel suspended lets Dukes land. So the wandering-Acarya itinerant — the figure who explains why the record is silent — is also, conveniently, the ancestor-type of Dukes's own school. Reading him here means granting the argument its real force while noticing that it does double duty: it corrects a genuine historiographical error and clears space for a lineage claim its author has every reason to want true.
Wandering Acarya Transmission Mobility — the mobility page and this one are cause and consequence, and joined they produce a principle neither states alone. The mobility page explains how the esoteric teaching moved: through itinerant teachers carrying iconographic and inner tradition person to person rather than through fixed institutions. This page draws the historiographical consequence of that very fact: a transmission mechanism built on movement is, by construction, a transmission mechanism that leaves no institutional trail. So the same feature that made the teaching mobile made it invisible — the wandering Acarya could reach anyone precisely because he was tied to no place, and he was untraceable in the record for exactly the same reason. The insight the pairing yields: decentralisation buys reach at the price of legibility. A tradition gains the ability to survive anywhere and loses the ability to be seen surviving — and historians who equate "documented" with "existing" will systematically under-count exactly the traditions that travelled best. The most mobile lineages are the most likely to be wrongly pronounced dead.
Taoist Forgeries and the Pseudo-Schools — set this page beside the forgery page and a sharp tension surfaces that the history domain has to hold open rather than resolve. Both pages are about what the record can and can't establish, but they push Dukes's invisibility argument in opposite directions. Here, invisibility is a shield for authenticity: the real tradition survived unrecorded, so absence of evidence doesn't disprove it. There, the same logic could shield a fake: a counterfeit school could likewise claim its lineage "survived invisibly through wandering teachers" and point to the very same silence as proof. The structural parallel is uncomfortable — "you can't see it in the record because real transmission is invisible" is an argument available to the genuine itinerant and to the fraud, and the record alone can't tell them apart. The insight the collision produces: invisibility cuts both ways. The same historiographical move that rescues a true wandering lineage from a false obituary also hands every false lineage an unfalsifiable alibi. Dukes wields it for the first purpose; nothing in the move itself prevents its use for the second — which is exactly why his other test (generativity over time, from the forgery page) has to do the work the record can't.
Sharpest implication: "died out" in a history book is often a claim about the shape of the surviving evidence, not about the survival of the thing — and decentralised, mobile, person-to-person traditions are the ones most likely to be falsely buried by institutional record-keeping.
Generative questions: