Two men tie deer antlers to their heads and lower them at each other. The horns clash, scrape, lock; the crowd leans in. This is Go Ti Ta — "antler striking" — and tradition dates its invention to Chi Yu, a figure said to have lived around 2000 BC. Four thousand years later you can still find its bones inside Mongolian folk dances and wrestling, the antler-lunge fossilized into a step.1 Dukes opens his survey of Chinese combat with images like this for a reason. Before a single Buddhist missionary walked into China carrying the Indian empty-hand system, China was already drenched in martial culture — antler-duels, man-against-bull spectacle, "skillful striking" written up in the imperial histories. The country the monks arrived in was not a blank slate. It was a sophisticated military society "in which might counted more than right," and understanding what was already there is the only way to see what the Buddhists actually added.2
This page catalogues the indigenous Chinese fighting arts and combat spectacles that predate the Buddhist transmission, the family Dukes groups under the non-specific label Wu Shu — "military arts."3 Most of these, he notes, were the prerogative of the ruling classes and centred on weaponry; China's enormous range of peoples and cultures meant many groups carried their own distinct, localized forms.4 Four named items anchor the survey:
The page is a baseline. It is what was on the ground before the Indian system arrived.
Dukes is doing two jobs at once on these pages, and they pull against each other. The first job is honest cataloguing: here is the deep, indigenous, pre-Buddhist martial culture of China, real and ancient and not derived from India. The antler-sparring is dated to 2000 BC. The man-and-bull sport runs from 473 BC to its Han-dynasty staging. The Han Shu writes up "skillful striking" with outline drawings. This is China's own thing, and Dukes grants it.
The second job is the loaded detail, and it cuts the other way. Twice on these pages he reaches across to plant the Indian flag in Chinese soil. The man-and-bull carvings, he says, show combatants in "classical Vajramukti Chuan Fa sparring positions quite unlike any Taoist equivalent." The Han Shu's strategy and tactics sections "all seem drawn from the Indian methods." And he names Mei Ching's tiger-and-bull fighter a "Chuan Fa proponent" — applying the Buddhist-Indian term to a Han-dynasty figure who lived before the transmission Dukes himself dates.
Watch the move. The survival mechanism is the genuinely interesting part: a technique can outlast the culture that made it by hiding inside something else. The antler-lunge survives as a dance step in Mongolia; the fighting stance survives as a carving. Forms are durable even when their function dies. But Dukes uses the same durability to run his thesis backward — if a 250 BC carving shows a "Vajramukti position," then either the position is older than the transmission, or Dukes is reading the later system into the earlier image. He wants the first. The honest reader has to hold the second as live.
This page is the negative space that makes Vajramukti Adopted by the Chinese Military legible. That page argues the army went abroad for an empty-hand disarming art because it had "no form of subtle combat training"; this page shows exactly what it did have — weapon arts, sport spectacle, striking forms — so the claimed gap is specific and not a claim that China was unmartial. It supplies Vajramukti as Buddhist Martial Philosophy with its contrast class: a Chinese martial culture organized around might and skill-in-battle, against which the Bodhisattva ethic defines itself. It feeds Seng-Ping: The Dharmapala Warrior-Monks the secular substrate — the indigenous Taoist and folk arts that Dukes elsewhere accuses of plagiarizing Chuan Fa, here shown in their pre-Buddhist depth. And the survival-in-dance mechanism connects to Nata: Kinetic Practice and Movement Meditation: the same fossilization that preserved antler-sparring in Mongolian dance is how Dukes claims the Indian Nata survives in Japanese Gagaku.
Take the man-and-bull carvings, the most loaded item in the survey. Dukes reports Han-dynasty rock carvings, with extant stone examples dated to around 250 BC, depicting combatants who "are taking up classical Vajramukti Chuan Fa sparring positions quite unlike any Taoist equivalent."9 Read it as two claims fused into one.
The first claim is descriptive and checkable in principle: there are old carvings of unarmed man-and-bull combat, and the figures stand a certain way. Fine. The second claim is interpretive and load-bearing: the way they stand is Vajramukti's way, the Indian-Buddhist signature, not a Taoist one. But the carving predates the Buddhist transmission Dukes dates to the second century AD and after. So the interpretation forces a choice he does not flag: either the "Vajramukti position" is a universal of human unarmed fighting (in which case finding it in a 250 BC carving proves nothing about Indian influence), or Dukes is recognizing the later system in the earlier image because he already knows how the story ends (in which case the carving is being recruited as evidence for a thesis it cannot actually carry). The "quite unlike any Taoist equivalent" tells you which way he is leaning — he wants the stance to be foreign even when the date is native. This is the revisionist's risk run in reverse: instead of dissolving a legend, he is solidifying an influence by reading a body-position as a passport stamp. The case study is valuable precisely because the seam shows. Dukes's strongest pre-Buddhist evidence is also where his thesis pulls hardest on the data.
You are in a northern Chinese town, sometime under the Han, and the dynastic festival is on. The arena is dirt. A man walks out with no weapon to face a bull, and the crowd that has seen this a hundred times still goes quiet at the lowering of the head. You watch how he stands — weight back, hands open, hips loaded to pivot off the line of the charge. Years later you are looking at a worn stone carving of that same scene, and a foreign monk beside you says, that stance, that is ours, that is Vajramukti. You look again. Maybe he is right. Maybe a body that has to survive a charging animal arrives at the same stance everywhere, monk or no monk, India or not. You hold both. You write down what you can see — the open hands, the loaded hips, the date on the stone — and you write down separately what the monk says it means, and you do not let the second sentence eat the first.
Evidence base: Dukes's survey of pre-Buddhist Chinese martial arts — Go Ti Ta credited to Chi Yu c. 2000 BC and surviving in Mongolian dance/wrestling; Chiao Ti Tsi man-and-bull sport (473 BC origin, c. 250 BC carvings, 108 BC staging by Emperor Wu); Chi chi hsiu in Pan Kuo's Han Shu (AD 32–92); Mei Ching's tiger-and-bull account.5678
[PLAUSIBLE — needs corroboration] on the descriptive catalogue (the named arts and their rough dates track standard accounts of early Chinese martial culture), but [CONTESTED] on the interpretive overlay — that the carved sparring positions are specifically Vajramukti's, that the Han Shu strategy material is "drawn from the Indian methods," and that a Han-dynasty fighter is a "Chuan Fa proponent." These are Dukes's readings across his own transmission date, and they assume the conclusion (Indian priority) in the evidence (Chinese artifacts).
Open seam, preserved: the dating of Go Ti Ta to Chi Yu "around 2000 BC" rests on a legendary culture-hero attribution, not a datable record; Dukes reports it without flagging the legend, then builds the survival-in-Mongolian-dance claim on top. Whether the antler-sport is genuinely that old or back-projected onto a mythic inventor is left unresolved.
Dukes is most trustworthy on this page exactly where he is least invested. When he simply reports the indigenous arts — the antler-duel, the bull-sport, the Han Shu's striking drawings — he is a careful surveyor, and the material lines up with what other sources say about early Chinese martial culture. The trouble starts where his devotional and lineage commitments enter: every time he reaches across to call a pre-Buddhist Chinese form "Vajramukti" or "Chuan Fa," he is doing the work his temple tradition needs done — establishing Indian-Buddhist priority — and the data starts bending toward the conclusion. Put him beside Geoffrey Samuel and the difference in posture is total. Samuel, facing the same carvings, would refuse to read a religious-system label off a body-position and would treat "seems drawn from the Indian methods" as precisely the kind of unfalsifiable resemblance-claim that lets a tradition annex anything it likes. Dukes treats the resemblance as transmission. The convergence is that both men agree China's martial culture is ancient and pre-Buddhist; the split is over whether its forms are native or imported, and on that question Dukes's warrant is his lineage's claim and Samuel's is the absence of evidence. Neither can close it from where they stand, and the carving stays ambiguous.
Convergent Cultural Evolution — that page argues that similar forms can arise independently from similar pressures, without contact or borrowing, and this page is a live test case that Dukes fails to apply. He sees a "Vajramukti sparring position" in a 250 BC Chinese carving and reads borrowing; the convergence frame offers the rival explanation he never weighs — that a human body learning to survive an armed or charging opponent will independently arrive at a small set of stable stances, the same way unrelated cultures independently arrive at the bow, the spear-thrower, or the wrestling clinch. The cross-domain page supplies the history with a null hypothesis: before attributing a shared form to transmission, ask whether the form is overdetermined by its function. The history supplies the cross-domain page with a sharp instance of the failure mode — a scholar with a transmission thesis treating a functionally-overdetermined posture as a cultural fingerprint. Together they yield a working test the history domain can reuse: a form is evidence of contact only to the degree that it is arbitrary — the more a posture is dictated by the physics of the fight, the less it can prove anyone taught it to anyone. A loaded-hip stance against a bull proves almost nothing; an idiosyncratic ritual salutation would prove much more.
Nata: Kinetic Practice and Movement Meditation — the survival mechanism here (Go Ti Ta's antler-lunge fossilized inside Mongolian folk dance) is the same mechanism Dukes uses for Nata surviving in Japanese Gagaku, and putting them side by side exposes a double standard worth naming. When a Chinese fighting form survives as dance, Dukes treats it as the form persisting after its martial meaning faded — a neutral fossil. When an Indian form survives as dance, he treats it as evidence of living transmission carrying the spiritual content forward. The Nata page supplies the mechanism (movement-meditation as a vehicle that preserves choreography across centuries); this page supplies the control case (a fight preserved as dance with its meaning lost), which suggests the mechanism is meaning-blind — it preserves the moves whether or not anyone remembers what they were for. The insight neither produces alone: choreographic survival is strong evidence of formal continuity and weak evidence of semantic continuity, so a surviving dance-form can never by itself prove that the teaching came with it. Applied back to Dukes, this caps what Gagaku can prove about Nata's spiritual transmission.
Sharpest implication: the more a martial form is dictated by the physics of the fight, the less it can ever prove about who taught it to whom — so Dukes's strongest "Indian influence" evidence (recognizable sparring positions) is precisely his weakest, because those positions are the most overdetermined by function.
Generative questions: