Imagine sweeping the floor of a very large archive that nobody has dusted in centuries, and finding, scattered under the shelving, a coin here, a buckle there, a torn page from a court diary, a temple's foundation stone with a name still legible. None of it was filed together. None of it was meant to be a story. But you lay the pieces out on a table in date order and a shape starts to appear: bare-handed fighting men, century after century, showing up in places that have nothing to do with each other — a Han dynasty almanac, a Japanese sumo legend, an emperor's literary preface, a Shaolin courtyard plaque, a succession contest decided by a fistfight between two princes. That table of scattered finds is what this page is. Dukes assembles the documentary record of empty-hand combat from sources that were never trying to record the history of martial arts, and the very fact that they mention it in passing — as an aside, a date-stamp, a footnote to something else — is what makes the record worth trusting.1
This page is the documentary spine for the claim that organized empty-hand fighting (Chuan Fa, "fist method," Japanese kempo) existed as a named, taught, datable practice long before the popular Bodhidharma-invented-it story would allow.2 It is built almost entirely from Dukes's chronological appendix — a year-by-year table of Buddhist, medical, and political events across China, Japan, India, and Korea — and from the scattered combative entries inside it.3 The entries are tiny. Most are one sentence. But they are the kind of sentence a forger never bothers to write, because the combat is incidental to whatever the chronicle actually cared about.
The load-bearing finds, in order:
Nine finds, a thousand years apart at the extremes, on two land masses, in genres that have nothing in common — an almanac, a sumo origin tale, a literary preface, a foundation plaque, a festival edict, a succession ruling, a court appointment. That spread is the whole argument.
A single source proving anything is weak; everyone knows the red flag. But a claim that survives because no single source was trying to make it is a different animal. Call it the incidental-corroboration principle. When a court diarist date-stamps a fistfight only because it happened to decide who became emperor, the fistfight is not the diarist's agenda — it is a fact he had to step over to reach his actual subject. Forgers polish the thing they want you to believe. They do not litter unrelated genres with throwaway mentions of it across a millennium.
So the mechanism here is triangulation by indifference. Pan Kuo, writing an almanac around A.D. 32, files "practicing with one's hands" right next to exorcism — for him it is a folk practice worth noting, nothing more.5 Three centuries later an emperor praising literature pauses to record that a man named Tung Chuan could beat people empty-handed.7 Another two centuries on, a succession crisis is settled by putting two princes in a ring.11 Each writer cared about something else. Each one, on the way to that something else, confirmed that organized empty-hand fighting was a normal, named, teachable feature of the world they lived in. The combat is the constant that keeps falling out of unrelated stories.
The named schools do extra work. "Practicing with one's hands" could be brawling. But the Chi Chi Hsiu school at A.D. 50 has a name in two languages, the Jotokashaku method at 706 is a named method with a named creator, and the Chi Chuan school at 950 is institutional enough that an emperor can be formally nominated its master.6812 Names and masters and methods are the fingerprints of transmission — of a thing being handed down on purpose, not reinvented by every generation that needed to hit someone.
And the Shaolin plaque of 728 sits exactly where the popular legend wants the story to begin.9 Dukes does not use it to prove the legend; he uses it the opposite way. The plaque is a physical, datable object at the temple — but the bare-handed entries on either side of it in the chronology (535 before, 706 just before, 748 just after) show empty-hand combat already in circulation across the whole region. The plaque is a node in a network, not the source of the river.
This page is the documentary floor that several big claims in the Dukes corpus stand on. The biggest is Bodhidharma: The Revisionist Biography. That page argues Bodhidharma brought a method, not a religion, into a China that already had centuries of Buddhism. This page supplies the parallel for combat: empty-hand fighting was already here too. The A.D. 32, A.D. 50, and A.D. 535 entries all predate Bodhidharma's traditional arrival around 520. So whatever the wall-facing monk contributed, he did not invent fist-fighting in China; he stepped into a culture that had named schools of it for half a millennium. The two pages together let the vault hold a clean, non-mystical position: Bodhidharma's role was to fuse a meditative philosophy onto a combat substrate that pre-existed him.
It also feeds Chuan Fa Transmission Lineage by giving the lineage external anchor points. A lineage that rests only on a temple's internal memory is fragile; a lineage that can dock onto a court-recorded school in A.D. 50 and an imperially-recognized master in A.D. 950 has handholds outside the temple wall. The named schools here are exactly the docking points.
And it hands Vajramukti as Buddhist Martial Philosophy a crucial concession: Vajramukti is the Buddhist philosophy of combat, but the combat it philosophizes about was a going concern in China long before any Buddhist framework wrapped it. The chronicle entries are the raw clay; Vajramukti is the doctrine that later shaped it.
Take the single strangest entry and read it the way Dukes reads legend — as compressed information. In A.D. 858, the chronicle records, "Emperor Buntoku orders his two sons to engage in bare-handed fighting in order to decide who was to succeed him. The victorious son, Koreshito, became the new Emperor Seiwa."11
Notice what is not in the sentence. There is no explanation of the rules, no description of the technique, no marvel that two princes could fight empty-handed at all. The chronicler treats it as a procedure — a known, usable instrument of state, like an ordeal or a council vote. You do not settle the imperial succession with a method nobody has heard of. The casualness is the evidence: bare-handed combat was institutionally legible enough, in mid-ninth-century Japan, to carry the legitimacy of the throne. A practice has to be deeply normalized before a court will hang a dynasty on it.
Place it next to the entry from 535, three centuries and a sea away, where Emperor Ni Wan-Ti slips a note about Tung Chuan's bare-handed skill into a preface about literature.7 Different country, different genre, different century — and the same unspoken assumption: empty-hand fighting is a real, recognized thing that an emperor can reference without explaining. Two emperors, three hundred years apart, both treating it as common knowledge. That is what the scattered record looks like when it is telling the truth: not one loud source, but many quiet ones that all forgot to be surprised.
You are at the table with the finds laid out, and a skeptic is reading over your shoulder. "One source," he says, tapping the A.D. 50 line. "Could be made up." You don't argue with that line. Instead you slide the other eight toward him. You put your finger on 23 B.C. — a sumo legend, Japan, a patron saint — and then on A.D. 32, a Han almanac in China that has never heard of the sumo legend. You walk him forward: 535, an emperor's book preface. 706, a named method with a named maker. 728, a stone you could in principle still touch at Shaolin. 858, a throne changing hands on the outcome of a fight. 950, a master formally appointed by imperial nod. "Now," you say, "show me the forger who planted bare-handed combat in a sumo myth, a dynastic almanac, a literary preface, a foundation plaque, a succession ruling, and a court appointment — across two countries and nine hundred years — and got them all to agree without any of them mentioning the others." The skeptic goes quiet, because that forger doesn't exist. The argument was never any single line. It was the table.
Evidence base: Dukes's chronological appendix (Appendix 7 in the source's running structure), a dense year-by-year table that is itself a synthesis of Buddhist, medical, and political records across four regions.3 The combative entries are extracted from inside it. [PLAUSIBLE — needs corroboration] applies to the dates themselves: the appendix routinely carries "alt. date" variants for major figures, the chronology is compiled rather than archival, and several entries (notably the "Five Grains" edict, which the source files under a mislabeled "A.D. 589–618" block sitting physically between the 747 and 751 entries, hence dated here ~748) have internal inconsistencies Dukes does not reconcile.10 The pattern survives even if individual dates wobble — but no single date should be quoted as firm without independent confirmation.
The deeper tension: this page reads the scattered record as strong because indifferent. A critic from the vault's own Samuel/White corpus would push back — scattered late mentions can also be a tradition retro-projecting its antiquity, salting old chronicles with the practice it wants to look ancient. Dukes does not address that possibility, and this page does not resolve it. The incidental-corroboration principle is powerful, but it assumes the chronicles are clean witnesses rather than partisan ones. That assumption is doing quiet work throughout.
Open: the nine entries span China and Japan but the thread between them is never documented. Did the Chi Chi Hsiu school of A.D. 50 have any real relation to the Chi Chuan school of A.D. 950, or are they independent flowerings of a recurring human impulse? The chronicle gives dates, not descent. The honest reading is recurrence; the tempting reading is lineage; the gap between them is the open seam.
Reading Dukes here feels like watching a man build a very strong case out of bricks he did not bake. The chronology is his compilation — and that is the catch that never fully goes away. The same Kongoryuji warrant that haunts the Bodhidharma biography haunts this table: the decisive framing (these scattered finds add up to a continuous combat tradition Buddhism later absorbed) rests on Dukes's own temple's reading of the record, the privately-held Kempo Hishu chain standing behind his confidence. An outsider can check that an emperor's preface exists; an outsider cannot check the temple memory that tells Dukes how to string the prefaces together.
Set him against Geoffrey Samuel and the half-alliance from the Bodhidharma page repeats with a twist. Samuel would love the move of reading incidental sources against each other — that is his own method on yoga's history. But Samuel would immediately ask the question Dukes skips: are these chronicles witnesses or advocates? Samuel demythologizes toward "we cannot be sure"; Dukes demythologizes toward "and therefore the combat tradition is older and realer than the legend admits." Both use the scattered record. One uses it to dissolve confidence, the other to build it. The vault holds both readings of the same nine entries and resolves neither — that is the point of keeping both authors in the room. White, for his part, would seize on 858: a throne decided by single combat is exactly his "sinister yogi," the holy realm and the violent one collapsing into one ruling, and he would read it as a pattern rather than a curiosity.
Samuel's Critical Anthropological Method — Samuel's method and the incidental-corroboration principle on this page are two settings of the same instrument, and comparing them exposes a blind spot each one has. Both read sources that were not trying to make the claim the historian wants — Samuel reading grammar and ritual manuals for what they assume rather than what they assert; Dukes reading almanacs and prefaces for the combat they step over on the way to something else. The convergence: a fact mentioned in passing, by a writer with no stake in it, is worth more than a fact a partisan source insists on. The split, and the insight: Samuel turns that principle against the tradition's antiquity claims, asking whether the passing mentions are themselves seeded. Dukes turns it toward the tradition, treating the passing mentions as innocent witnesses. Put them together and you get a rule the history domain needs: incidental corroboration is only as strong as the chronicle's innocence, and you cannot read the combat out of the almanac without first asking whether someone put it there. Neither author asks that of his own evidence; the vault has to ask it for them.
Bodhidharma Dating Controversy, Source by Source — that page is a single-figure version of the problem this page solves at scale. The dating-controversy page argues over when one man arrived by weighing source against source. This page does the same operation but with no single figure at the center — just the practice itself, surfacing across genres and centuries. The structural parallel: in both cases the method is to refuse any one source's authority and let the agreement-by-accident across many sources carry the weight. The insight neither produces alone: when you have one famous figure, source-criticism becomes a fight over his biography (did he exist, when did he arrive); when you drop the figure and track the practice, source-criticism becomes triangulation, and the absence of a hero is a strength — there is no biography to inflate, only a recurring fact to confirm. The Bodhidharma debates are loud because a person is at stake. The Chuan Fa chronicle is quiet because nothing but the practice is. The quiet version is, paradoxically, the more defensible history — and it is the one nobody fights about, precisely because there is no hero to claim.
Sharpest implication: if a practice survives in the record because no source was trying to preserve it, then the strongest evidence for any old tradition is not its famous founding text but the offhand mentions in documents about something else entirely. The history domain should hunt for the asides, not the manifestos.
Generative questions: