The Two Windows: How Fujian and Hong Kong Took Kung Fu Abroad

A teenager in Des Moines throws a reverse punch in a Wing Chun class. A construction worker in Kota Kinabalu finishes a Five Ancestors form at dawn. A film student in Berlin pauses a frame of The 36th Chamber of Shaolin to study the footwork.

All three are downstream of the same event — the export of Chinese martial arts — but not of the same mechanism. The usual telling collapses them into a story about individual genius: Bruce Lee crossed the Pacific and taught the world, Ip Man carried Wing Chun out of Foshan, one great master ferried an entire system across an ocean. That story sells tickets. As history it is thin.

What actually happened is that the arts left China through two mechanisms with almost nothing in common except the techniques they carried.

One was a migrant network. Southern styles travelled to Southeast Asia inside clan associations, dialect groups, and sworn brotherhoods, as part of the apparatus that let communities without legal protection survive. The art was communal infrastructure. It went deep and stayed narrow.

The other was a port city’s media industry. After 1949, masters who had lost everything arrived in Hong Kong and taught whoever would pay, while a studio system turned their techniques into films distributed across the world. The art was a commodity. It went wide and stayed shallow.

Neither mechanism produced the global practice that exists now. The first built lineages that outlasted their founding communities; the second built an audience of hundreds of millions, most of whom never met a master. Modern kung fu is what happened when those two outputs met.

The distinction is structural, not geographic

Fujian and Hong Kong name the two cases, but they are not the explanation. The relevant difference is not which province a master came from. It is what kind of migration he was part of, and what he had to do to eat.

	Southeast Asia	Hong Kong
Migration type	Labour migration, chain migration through clan networks	Political refugees, urban drift
Social structure	Clan halls and dialect associations governing community life	Colonial law, commercial society, no clan substitute
How a master earned	Embedded in a community that needed him	Teaching fee-paying strangers
Legal protection	Almost none from the colonial state	Limited but real
Result	Deep, narrow, ethnically bounded	Wide, shallow, cross-ethnic

Ip Man is the case that makes this concrete, and the one that breaks a purely geographic reading. He was from Foshan — Cantonese, not Fujianese. He left in 1949 as a refugee, not as a labour migrant. And when he arrived he did not teach inside a clan association, because he did not have one. He taught in a restaurant workers’ union hall in Sham Shui Po, introduced by Leung Sheung, and his early students were bus workers, civil servants, and policemen.

Nothing about that is Foshan tradition. It is what a master does when the structure that used to hold him is gone.

Window one: the migrant network

Why the coast emptied

Fujian’s interior is mountain and thin soil; for centuries the sea was the only road out. Quanzhou and Zhangzhou had trade routes across the South China Sea long before anyone thought of emigrating along them.

The migration that carried martial arts was part of a much larger movement, and it came in several distinct streams that are often merged into one. The indentured labour trade of the mid-nineteenth century — the coolie traffic — ran mainly to Latin America: roughly 140,000 Chinese were shipped to Cuba between 1847 and 1874, of whom about 125,000 arrived alive, with a comparable number sent to Peru.[1] That trade was a different thing from the migration that built Southeast Asian Chinese communities, which was larger, longer, and largely made up of free migrants and chain migration financed through clan networks rather than contract labour.[2]

The distinction matters for this argument. Martial arts did not travel to Nanyang in the holds of the coolie ships. They travelled with the free migrant who had a clan hall to report to when he landed.

Fujian’s southern counties were also poor, over-populated, and violent. Village feuds over water and boundaries were endemic, banditry was ordinary, and the state was distant. The styles that developed there — Fujian White Crane, Five Ancestors Fist, Ground Method Boxing, the various southern Shaolin systems — were built for close quarters, uneven ground, and ending an encounter fast against someone stronger.

What the clan hall was for

Migrants arrived as members of surname clans and dialect groups, not as individuals. The clan association provided shelter, credit, burial funds, dispute mediation, and physical protection. That last item is the one that matters here.

Colonial administrations in Southeast Asia offered Chinese migrants almost no legal remedy. Anti-Chinese violence was periodic. Chinese communities competed with each other over tin, opium, transport, and trade. A clan association without the capacity to defend itself was not a mutual-aid society; it was an opportunity.

This is where the scholarship is more careful than the popular telling. Judkins and Nielson’s social history of the southern arts documents the function of martial practice within overseas Chinese community institutions — as a practical resource for self-defence and collective protection in communities that could not expect impartial policing — without claiming that secret societies conscripted members into fighting.[3] The stronger claim, that a mine’s brotherhood required fighting men, is a plausible inference and not a documented fact. This page states the weaker version, because it is the one the evidence carries.

Five Ancestors Fist as the case

Wuzu Quan is the clearest example because its spread is documented on both ends. The system was organised in Jinjiang, near Quanzhou, by Cai Yuming in the second half of the nineteenth century — the same decades in which migration from that district to Southeast Asia was at its height.

Its transmission outward is traceable through one man. Shen Yangde, Cai Yuming’s last disciple, established the style in Xin’an village in Xiamen, bringing six Qiu clan halls under one organisation. When Xiamen fell to Japanese forces, he left — first to the Longshan Hall in Myanmar, where he taught at Jiande Temple, then to Singapore, where he founded the Yuming Martial Arts Research Society and the Lanqing Martial Arts Hall.[4] A local newspaper account and the county government record differ on whether the departure was 1937 or 1938; both agree he left, and both agree where he went.[4][5]

The point is not that Shen was an evangelist. He was a refugee taking the only route available to him, which happened to be the route his district’s migrants had already paved. The network did not carry ideas. It carried people, and the ideas travelled with them because the people needed them.

Why it stayed Chinese

Three features distinguish this mechanism from what came later in Hong Kong.

It was bounded. Teaching happened inside dialect and clan lines — a Hokkien association taught Hokkien styles, a Hakka community taught Hakka ones. This was not snobbery; it followed from the fact that the training was tied to communal defence, and you do not hand the keys of the clan hall to strangers.

It was deep. A style might have a few hundred practitioners in one city, but those practitioners trained hard, kept lineage ties to home villages, and preserved forms with unusual fidelity, because the art functioned as inherited property rather than as a product.

It was invisible to the outside. These arts did not appear on film, were not written up in the English-language press, and did not recruit foreign students. They existed in the interior of Chinese community life. Wang Gungwu’s account of overseas Chinese society describes exactly this pattern — institutions organised around community self-government that operated largely beneath the notice of the colonial state rather than seeking its recognition.[2]

The limitation followed from the strength. By 1950, Five Ancestors Fist and White Crane were thoroughly rooted across Southeast Asia and had still not crossed an ethnic line. A Malay, Indian, or Eurasian in Singapore in 1950 had essentially no route into either.

Breaking that seal required a different kind of city.

Window two: the port and its films

The collision

Hong Kong’s role in this history was not designed. It was a consequence of collapse.

The civil war and the establishment of the People’s Republic sent a wave of refugees south. The population had fallen to about 600,000 by 1945 under occupation; by 1950 it stood at about 1.97 million, and by 1955 at about 2.49 million — a four-fold increase in a decade, compounded of returning residents, refugees, and a high birth rate.[6] Among those arrivals were martial arts teachers from across the southern systems.

What made the colony different from Singapore or Manila was not the concentration of talent but the institutional configuration. Hong Kong was a British free port with minimal censorship and no exchange controls. Shanghai’s film and publishing industries had largely relocated there as the war advanced. The Shaw brothers had moved their operation from Shanghai to Singapore and then to Hong Kong, and by the late 1950s were building Movietown.

The refugee masters arrived into that environment with nothing. Unlike their counterparts in Southeast Asia, they had no clan structure to teach inside. They taught whoever paid.

The studio as factory

The Shaw studio was an industrial operation in the full sense. By the late 1960s it was producing between 40 and 48 features a year, on a lot where directors, actors, screenwriters, choreographers, and set builders were all under contract.[7] That rate alone is the argument: no lineage-based system could produce at that tempo, because a lineage transmits one student at a time with a master’s attention attached.

The output was not aimed primarily at Hong Kong. Shaw built distribution across Southeast Asia, Taiwan, Korea, Japan, and the Chinatown circuits of North America, Europe, and Australia. Films were dubbed, subtitled, and recut per market. A film shot in Hong Kong could be playing in San Francisco, Lagos, and rural Thailand within months.

What this did to the art is the point. A martial art transmitted by migration needs a master, a student, a room, and a community. A martial art transmitted by film needs a projector. It moves at the speed of commerce rather than the speed of migration, and it crosses ethnic boundaries because anyone who buys a ticket can watch.

Call this the cinematic air-drop. The arts did not spread outward from Hong Kong in rings, carried by departing students. They appeared, fully formed, in places that had no Chinese martial arts community at all, because a print arrived before any teacher did.

The evidence for the sequence is in the box office. Five Fingers of Death was released in the United States by Warner Bros in March 1973 and is generally credited with setting off the American kung fu cycle — more than thirty similar films reached US screens that same year.[8] Its North American rentals came to about $4 million, which made it the second-highest-grossing martial arts film of 1973 in that market, behind Enter the Dragon at roughly $4.25 million.[8] Both figures are rentals, not grosses, and neither film topped the all-genre annual chart.

The second-place figure is the interesting one. A Hong Kong film with no internationally known star outdrew everything in its category except the biggest martial arts film ever made — and it had arrived first. The demand existed before the man who is supposed to have created it.

Bruce Lee, correctly placed

Lee’s importance is real but narrower than the myth. He was not a traditional master who happened to appear in films; he was a performer who understood screen logic, and his three Golden Harvest pictures gave the genre a visual language that needed no cultural translation.

But he did not open the door. Warner had already released a Hong Kong martial arts film in the American market before Enter the Dragon, and the audience that made his films profitable was assembled by the films that preceded them. Lee made the genre a mass phenomenon. He did not create its market.

The pulp layer

Alongside the studios ran a cheaper distribution system: serialised fiction. Jin Yong and Liang Yusheng published wuxia novels in Hong Kong newspapers from the 1950s, and the form gave cinematic kung fu a literary vocabulary — a stock of images, ethics, and historical resonance that a fight scene could draw on without explaining itself.

Those novels were not transmitted through lineages. They were sold at newsstands and read on public transport. They converted martial arts from a skill you learned from an uncle into an imaginative space any literate person could enter.

Why it went wide and shallow

The inverse of the migrant mechanism, on all three counts.

Cross-ethnic from the start. Hong Kong films were made for export. By the mid-1970s kung fu was being taught in Birmingham, Detroit, and Durban to students who had never been to a Chinatown and knew nothing about Chinese clan structure.

Wide rather than deep. A style could acquire tens of thousands of practitioners worldwide within a decade of appearing on screen, with transmission fidelity varying enormously. A student who learned from a student who had learned from a film could produce something barely recognisable.

Spectacularly visible. Which meant that for most of the world, “kung fu” was defined by what happened on a screen rather than in a school.

The cost was that much of what spread was a cinematic simulation of martial arts, stripped of training method, combative pressure, lineage discipline, and communal context. It was possible in 1985 to be a devoted kung fu student in Ohio who had never sparred and never learned a form in full.

Where the two met

By the late 1970s the mechanisms had begun to interlock, and modern global practice was formed in that interlocking.

Diaspora audiences met the films. A third-generation Chinese-Malaysian teenager who had grown up watching his grandfather practise Five Ancestors Fist in the back of a clan hall — and thinking of it as something old men did — saw a Shaw Brothers film and encountered the same art as heroic, modern, and worth caring about. The film did not replace the family transmission. It gave it a new vocabulary of value.

Film audiences traced their way back to lineages. The first non-Chinese students in the West often learned from whoever was available, frequently other Westerners with a few years’ experience. As they matured, many went looking for sources, and found them in two places: Hong Kong, where Ip Man’s students were teaching, and Southeast Asia, where the migrant systems had been preserved in relative obscurity. The media window created the market; the diaspora window supplied the content that could sustain it past the initial enthusiasm.

Hong Kong became hybrid space. Choreographers and performers moved between Singapore, Kuala Lumpur, Taipei, and Hong Kong, carrying methods across networks that had previously been separate. Several forms now taught worldwide as “traditional” were standardised in Hong Kong in the 1950s and 1960s, sometimes by masters who had themselves trained in Southeast Asia.

A counter-case, to keep the model honest

The two-window frame is a useful abstraction and it is not a law. Some arts did not travel through either window.

Wing Chun in Foshan stayed in Foshan. It was transmitted locally, from Chan Wah-shun to Ip Man and his generation, without a migration story attached, until Ip Man himself moved to Hong Kong in 1949 — and even then he was one master making one journey, not a network in motion. Ip Man’s record covers that transmission on its own terms.

Wing Chun also shows why the geography can mislead. Fujian has its own Wing Chun line, listed as a separate national heritage item declared by Fuzhou, with its own transmission and its own claims.[9] The Hong Kong line runs through Foshan. The two share a name and a reputation and are not the same documented lineage. Anyone using “Wing Chun” as evidence for either window needs to say which line they mean.

What the model is good for

It explains a specific puzzle: why the same body of technique produced two completely different kinds of global presence. Deep and narrow in one place, shallow and wide in another — not because the arts were different, but because the migration and the media environment were.

And it predicts something checkable. If the mechanism is what matters, then a style that entered Southeast Asia through clan networks should show high lineage fidelity and low cross-ethnic reach, while the same style entering through Hong Kong’s teaching market should show the reverse. That is roughly what the record shows, and it is the kind of claim that can be tested against more cases than this page has room for.

Both engines are slowing

The studio system that created global kung fu is gone. Shaw Brothers ceased production in the mid-1980s; Golden Harvest was absorbed into a larger group. Hong Kong cinema’s industrial peak is a historical period now.

The clan associations of Southeast Asia are contracting. The grandchildren of coolies and miners are middle-class professionals with no need for martial arts as communal defence, and the halls that once required fighting men now struggle to fill a membership list.

So the two mechanisms that built the global practice are both weakening, while the appetite for the practice keeps growing. That is the tension worth watching: the systems that produced depth are in decline, and the systems that produce reach are getting cheaper and faster. More visibility, less transmission.

What replaces them is not yet clear. But the shape of the answer is visible in places where non-Chinese students train seriously in migrant-derived systems, where Southeast Asian masters teach seminars in Europe, and in archives that put the documentary record where anyone can check it — including, one hopes, this one.

Sources

1. [1] Meagher, Arnold J. The Coolie Trade: The Traffic in Chinese Laborers to Latin America 1847–1874. Xlibris, 2008 — the title itself fixes 1847–1874 as the Latin America period. The Cuba figure of roughly 140,000 shipped and 125,000 landed is given in Zeuske, “Coolies – Asiáticos and Chinos: Global Dimensions of Second Slavery,” in Bonded Labour, De Gruyter, 2016, drawing on Meagher.
2. [2] Wang Gungwu, The Chinese Overseas: From Earthbound China to the Quest for Autonomy, Harvard University Press, 2000; and Kuhn, Philip A., Chinese Among Others: Emigration in Modern Times, Rowman & Littlefield, 2008 — the structure of migrant communities, clan and dialect organisation, and their relation to colonial authority.
3. [3] Judkins, Benjamin N., and Jon Nielson, The Creation of Wing Chun: A Social History of the Southern Chinese Martial Arts, SUNY Press, 2015 — the social function of martial practice in overseas Chinese communities. Used here for the weaker claim, not for the conscription claim.
4. [4] 台海网 (Taihainet), 新垵：厦门尚武村落, 18 July 2007 — Shen Yangde’s departure, the Longshan Hall in Myanmar, Jiande Temple, Singapore, and the two organisations he founded there.
5. [5] 厦门市海沧区人民政府, 新垵五祖拳（海沧非遗）, 17 February 2022 — the six Qiu clan halls, the Heyang Tang, and the 1937 date for the Japanese occupation of Xiamen. The 1937/1938 discrepancy between this record and [4] is noted in the text rather than resolved.
6. [6] Hong Kong population series: 1945 ≈ 600,000; 1950 ≈ 1,974,000; 1955 ≈ 2,490,000. The decade growth is reported here as roughly four-fold, and the components — return migration, refugee influx, natural increase — are stated rather than attributed to a single cause. Figures vary between series; the archive uses the ones consistent across the Hong Kong government series and the standard demographic compilations.
7. [7] “Crafting a World-Class Brand: Shaw Brothers’ Appropriation of Foreign Models,” Media Industries Journal 7, no. 1 (University of Michigan) — Table 1, annual production 1964–1971: 23, 26, 41, 44, 45, 48, 40, 40 features produced per year.
8. [8] Variety, “Big Rental Films of 1973,” 9 January 1974, p. 19 — North American rentals for Five Fingers of Death (US$4m) and Enter the Dragon (US$4.25m); and Variety, 9 January 1974, p. 72, on the 1973 cycle of more than thirty chop-socky releases in the US.
9. [9] China Intangible Cultural Heritage Network, item Ⅵ-77, 咏春拳, fourth batch 2014, declared by Fuzhou, Fujian — the separate Fujian Wing Chun line. Protection unit currently recorded as 暂无合适保护单位.

Two related records: Ip Man, the master whose move from Foshan to Hong Kong is the hinge of the second window, and Wuzu Quan, the clearest documented case of the first. The companion essays are the fact-check of Bruce Lee’s teachers and Wong Shun-leung and Wing Chun’s combat science.

Frequently Asked Questions

How did Chinese martial arts spread outside China?

Through two different mechanisms, a century apart. Migrant networks carried southern styles to Southeast Asia inside clan associations and dialect communities, where the art functioned as communal protection and stayed largely Chinese. Later, Hong Kong's film industry turned regional fighting styles into a global visual commodity, reaching audiences with no Chinese ancestry at all. The first was carried by people; the second was distributed by studios.

Did Bruce Lee introduce kung fu to the West?

No, though he is the reason most Westerners noticed it. Hong Kong martial arts films had been entering the United States before Lee's American releases, and Five Fingers of Death was released in the US by Warner Bros in March 1973, the same year as Enter the Dragon. Lee's importance is that he made the genre a mass phenomenon — not that he opened the door.

Why did Fujian martial arts stay within Chinese communities while Hong Kong's went global?

Because the two were transmitted for different purposes. In Southeast Asia the arts were bound up with clan defence in communities that had no access to colonial law, so they were taught inside kinship and dialect lines. In post-1949 Hong Kong, refugee masters had no clan structure to fall back on and taught whoever paid. One was communal infrastructure; the other was a market.

Is Wing Chun a Fujian style or a Hong Kong style?

Both claims miss the point. Wing Chun's documented transmission runs through Foshan and, from 1949, Hong Kong. Fujian has its own separate Wing Chun line, recognised as a distinct national heritage item declared by Fuzhou. They share a name and a reputation, not a single documented lineage.