Imagine a long library table with twenty Chinese books laid out in a row, none of them agreeing with the one beside it. One book, written in 547, has a Persian monk standing in front of a particular temple admiring its beauty — and that temple was wrecked by a storm in 526, so the monk had to have been alive and in China before that. Another book, compiled five centuries later, says the same monk first set foot in Canton in 526. Both cannot be true. A third book never mentions the monk at all and names a completely different man as the founder of his school. This page is a walk down that table. Dukes does not tell you a single tidy birthdate for Bodhidharma; he hands you the receipts and shows you how the dates are forced earlier and earlier by texts that have nothing to gain from the conclusion. The argument is built from a pile of ledgers, and the ledgers fight.1
This is the source-criticism apparatus underneath the Bodhidharma Revisionist Biography — the actual list of named Chinese texts, with their dates, their claims, and the arithmetic Dukes runs on them. The biography page gives you the man who steps out of the fog; this page shows the fog being measured. It collects roughly twenty source documents spanning from circa 472 to 1237, sorts them by what they assert, and uses the disagreements among them to triangulate when Bodhidharma was actually in China.2
The headline finding is a dating war. The later, famous, partisan account — the Ching Te Chuan Teng Lu of about 1044, which is where almost everyone gets the Wu Ti meeting and the nine years of wall-gazing — puts Bodhidharma's arrival at 526.3 Three earlier, less partisan texts push that date back by decades. The Loyang Chia Lan Chi of 547 — the earliest record of his presence in China — has him admiring the Yung Ning Temple before 526.4 The Hsu Kao Seng Chuan of about 645 records a disciple of his, Seng Fu, who left the north in 494, which only works if the master was already teaching there well before then.5 Underneath the dates sits a second, deeper fight: several texts deny Bodhidharma was the founder of the school at all, naming Gunabhadra instead.6 So the ledger does not just argue about a year. It argues about whether the man at the head of the table belongs there.
The method is subtraction by witnesses. Dukes never asserts a date; he lets the texts box each other in.
Take the temple-date proof. The Loyang Chia Lan Chi was written down by Yang Hsuan Chih and finished in 547. In it, a Persian monk named Bodhidharma comments on the beauty of the Yung Ning Temple. That temple has a paper trail of its own: built in 516, damaged in a storm in 526, turned into an army barracks in 528, burned down in 534. A man cannot admire the beauty of a building that is a barracks or a ruin. So Bodhidharma must have seen it intact — before 526. And that single inference detonates the later Dentoroku's claim that he first reached Canton in 526. He was already up north admiring temples by then.4
Now stack the ordination math on top. The Hsu Kao Seng Chuan, composed around 645 by a Vinaya-school monk named Tao Hsuan, records a northerner called Seng Fu as a disciple and ordinee of Bodhidharma. Seng Fu left the north in 494 and died in the south in 524 at age 61. Run it backward: if he was at least a teenager when ordained, his master was active in northern China before 494 — pushing the arrival date earlier still, or forcing an earlier visit.5 Dukes prizes this witness for a specific reason: Tao Hsuan was not a Chaan partisan, he wrote only about a century after Bodhidharma's death, and he treats the man "as just one of many other monks of the period."7 A hostile or indifferent witness who still records you is worth more than a devoted one who inflates you.
The same engine runs on the founder question. The Li Tai Fa Pao Chi, a Tibetan-linked text found in the Tun Huang caves, asserts that Gunabhadra — not Bodhidharma — was the first Chinese Chaan patriarch.8 The Leng Yen Chia Jen Fa Chih of about 700 agrees, calling Gunabhadra the founder, and pointedly carries no story of Hui Neng succeeding Hung Jen, no robe, no poems.9 The chronology even gives Gunabhadra the Lankavatara translation Bodhidharma is supposed to have transmitted: Gunabhadra rendered a version of it in 435–436, and a 443 "Lankavatara Jewel" translation is the one "probably used by Bodhidharma."10 So the very scripture that seals Bodhidharma's lineage may have reached him through the man some texts call the real founder.
This page is the load-bearing footing under the whole Dukes biographical structure. It supplies the Bodhidharma Revisionist Biography with the hostile-witness evidence that keeps the man from dissolving into legend — the biography asserts a real, datable person; this page shows the arithmetic that earns the assertion. It feeds Tibetan Accounts of Bodhidharma directly: the Jen Tien Yen Mu here names the Bodhiruci-translator as the poisoner, and the Tibetan record independently carries the six-poisonings motif, so the two pages corroborate across a language barrier. It underwrites the patriarch-count problem that the biography only gestures at — 23 versus 27 versus 28 — with the specific texts on each side. And it gives Chuan Fa Transmission Lineage its evidentiary spine: a lineage is only as solid as the documents that record it, and here the documents are caught contradicting one another.
One entry on the table shows the whole method working in a single paragraph. Dukes notes that the Japanese Zen sect names 28 patriarchs up to and including Bodhidharma, while the Tendai sect names 24, using a different list — and both lists borrow names from common Chinese sources.11 A casual reader files this as trivia. Dukes reads it as a fossil record. Here is the tell: most Chinese copies of the Hsieh Mai Lun surviving in Japan date only from the 14th century and claim to be copies of Tang-dynasty originals, and in those late copies, 28 patriarchs are given. But the Tun Huang cave excavations turned up much earlier copies — and in most of those, only 27 patriarchs are common.11 The 28th name was added later. You can date the inflation of the lineage by counting heads in old manuscripts versus new ones. The earlier the copy, the shorter the list. That is exactly the texture the revisionist biography promises: a legend you can watch assemble, layer by layer, because the early strata physically survive in a desert cave and the late strata sit in a 14th-century library. The patriarch count is not a doctrine. It is a tree-ring.
A second tell in the same case: the Fu Fa Tsang Yin Yuan Chih of about 472 — the earliest literal account of the patriarchs — lists 23, from Mahakasyapa to Simhanada, and Bodhidharma is not in it.12 The man at the head of the Zen table is absent from the oldest copy of the guest list. Later texts add him; the Chuan Fa Chen Tsung Lun of about 1061 has to argue, against a Tang-dynasty critic named Shan-Ki, that Bodhidharma "was an orthodox patriarch" at all.13 You do not have to defend something nobody is doubting. The defense is the evidence of the doubt.
You are at the library table, late, with a yellow legal pad. You write the disputed claim at the top: Bodhidharma arrived in China in 526. Then you go down the row of books, and for each one you write two columns — what it says, and what it has to gain. The Dentoroku, 1044, Chaan monk, says 526; gain: it is the sect's own founding story, every incentive to tidy it. You put a question mark beside it. The Loyang Chia Lan Chi, 547, secular city-gazetteer, says he admired a temple that was wrecked in 526; gain: none, it is just describing buildings. You underline it twice. The Hsu Kao Seng Chuan, 645, Vinaya monk who thought Chaan people were "lazy and improper," records a disciple ordained before 494; gain: negative, the author has no love for the school. You underline it three times. Now you look at your underlines. The dates with no axe to grind all point earlier than 526. The only source insisting on 526 is the one that profits from it. You cross out 526 at the top of the pad and write before 494, on the weight of the hostile witnesses. The pad is uglier than the legend. It is also the only version you can defend out loud.
Evidence base: Note 39's appendix of "Significant Chinese Accounts relating to Bodhidharma's History," a roughly twenty-text annotated catalogue running from the Fu Fa Tsang Yin Yuan Chih (c. 472) to the Shih Men Ch'eng T'ung (1237), cross-checked against the dated chronology entries for 472–547.2 Dukes also names his outside scholarly backstops — Blyth's Zen and Zen Classics, Takakusu, Nanjio's catalogue of the canon.14
The hardest tension is the poisoning thread, and Dukes does not resolve it. The Jen Tien Yen Mu (c. 985–1061) asserts Bodhidharma did not use the Lankavatara Sutra at all, and that the man who translated Bodhiruci's version was the same man who later poisoned him.15 Yet other texts on the same table make the Lankavatara central to his transmission and hand it to Hui K'o.16 Did he use the sutra or not? Was he poisoned by a rival translator, or is the "poisoning" the same Three-Evils allegory the biography reads in the severed-arm story? The table holds both and settles neither — preserved here, not collapsed.
The founder dispute is equally live. [CONTESTED] on the identity of the first Chinese Chaan patriarch: the Li Tai Fa Pao Chi and the Leng Yen Chia Jen Fa Chih both say Gunabhadra; the orthodox lineage says Bodhidharma. Dukes records the dissent without erasing it.6
A naming caution worth flagging: the Tibetan-linked Ta Mo Ta Lo Chan Ching ("Dharmottara Chaan kyo") was translated by Buddhabhadra (c. 308–422), well before Bodhidharma's own floruit — a reminder that "Dharmottara"/"Tamo" name-forms attach to more than one person and that some apparent Bodhidharma references may be homonyms.17
Reading this page, you can feel Dukes doing two things at once that pull against each other. As a critic he is superb and genuinely even-handed: he openly prefers the secular gazetteer and the hostile Vinaya monk over his own sect's founding legend, and he warns the reader off the very source Western scholars lean on. That is real source criticism, the kind Geoffrey Samuel would recognize and respect. But the warrant flips at one specific spot. The single text that anchors the orthodox patriarch line — the Fu Fa Tsang Yin Yuan Chih of 472 — Dukes cross-references to a manuscript held at Kongoryuji, his own temple.18 So the public, checkable witnesses all push the date earlier and stay agnostic about the founder; the private, in-house witness is the one that secures the lineage. This is not a contradiction to resolve — it is the structural signature of an insider doing outsider-grade criticism. He can be ruthless with the documents anyone can read, precisely because the document that matters most to him sits in a cupboard the rest of us cannot open. The Hsieh Mai Lun attribution and the Dogen-forgery thread that run through the biography page live here too as unresolved flags. [PLAUSIBLE — needs corroboration] on the whole reconstruction insofar as its keystone is a privately-held temple manuscript.
Samuel's Critical Anthropological Method — this page and Samuel's method are the same instrument turned on different traditions, and laying them side by side exposes a rule neither states alone. Samuel dates yoga's claims by playing texts against each other and trusting the witness with the least to gain; Dukes dates Bodhidharma by playing twenty Chinese books against each other and trusting the secular gazetteer over the sect's own legend. Both refuse the tradition's self-flattering chronology. The difference is the cupboard. Samuel's whole evidence chain is public — every citation checkable, every conclusion deliberately thin where the record thins. Dukes's chain is public right up to the keystone, which is his temple's private manuscript, and there the conclusion thickens exactly where outsiders lose the trail. Put the two together and you get the insight: source criticism splits into two species depending on whether the critic stands inside or outside the lineage. The outsider's version is fully checkable and structurally blind to what the temples hold; the insider's version can reach the held material and is structurally unable to prove it. A vault holding both has to grade every dating claim twice — once for the quality of the public witnesses, once for whether the decisive witness is one anybody can read.
Source Laundering and Provenance Inflation — the patriarch-count fossil is a provenance-inflation case study with the dates attached. The behavioral-mechanics page describes how a claim's authority gets manufactured by attaching it to a more impressive origin than it actually has — backdating, borrowing a famous name, claiming descent. Here you can watch the mechanism happen in slow motion across real manuscripts: the 472 list has 23 patriarchs and no Bodhidharma; the 14th-century copies have 28 with Bodhidharma enthroned; the Tun Huang caves preserve the 27-name intermediate stage. The skin-flesh-bones succession verse is itself borrowed wholesale from Nagarjuna and re-attributed to Bodhidharma.19 The history page gives the behavioral page something it usually lacks — a dated, physical time series of provenance inflation, where you can literally count the added name in older versus newer copies. The behavioral page gives the history page the mechanism behind the motive: lineages inflate because authority flows backward from impressive ancestry, so a school under attack (as the Chuan Fa Chen Tsung Lun shows Chaan was) has every incentive to lengthen and ennoble its founder list. Together they yield a working tool: when a lineage claim's prestige rises over time while its oldest copies say less, you are watching provenance inflation, and the early manuscript is the audit trail.
Sharpest implication: the most reliable date for a contested holy figure is the one supplied by a witness who disliked him, dated his disciples instead of him, and never meant to write a founding legend — and that date almost always sits earlier and plainer than the famous one.
Generative questions: