Skip to content Skip to footer

Chinese in Vietnam: Cholon, the Minh Huong, and the Second Exodus

Street-level frontage of Chợ Bình Tây market in District 6, Cholon, with its clock tower, blue-and-white mosaic gable and shoppers at the entrance

Two displacements, and the paper between them

A community can be displaced twice without ever crossing the same border. The Chinese of the Mekong delta were moved first by a dynasty: a defeated Ming force, posted in the spring of 1679, as the dynastic record has it, to a frontier the Nguyễn did not yet hold.

The scholarship dates the posting to 1681, to late 1682 to 1683, or to 1683 itself, and nothing here settles the difference. A tax register in 1698 wrote them into the Vietnamese state, with a headman, communal land and a right to sit the examinations. Their descendants became 明乡 (Míngxiāng, Minh Hương), the Ming-loyalist community, and built the quarter the Cantonese called 堤岸 (Dī’àn, Chợ Lớn), Cholon. Then commerce changed, and in March 1978 the state abolished private trade in the south: a quarter whose worth was inventory and goodwill became illegal within weeks.

The people on the boats were by then mostly 华族 (Huázú, Hoa), the Cantonese, Teochew, Hokkien, Hakka and Hainanese who had made Cholon the rice mill and the credit line of French Cochinchina, while those who had already become Vietnamese stayed to watch them go. Both were carried out on paper: a register that wrote the community in, and a register that wrote it out.

The case is an anomaly: no indenture contract at departure, no exclusion statute at the far border, no recruiting company working a line. The instrument here was capital, and capital can be legislated out of existence in a month. The boats come late, and as arithmetic: the last consequence of the paper, not its subject.

Street-level frontage of Chợ Bình Tây market in District 6, Cholon, with its clock tower, blue-and-white mosaic gable and shoppers at the entrance
Chợ Bình Tây (Bình Tây Market), District 6, Cholon — the market 郭琰’s estate funded and which opened in 1928, a year after his death. Shops stacked with goods run along both wings.

A defeated army on a border the Nguyễn did not yet hold

Some three thousand men and fifty war ships came into the roadstead between Đà Nẵng and Hội An, the Nguyễn dynastic record says, under two commanders it names: 杨彦迪 (Dương Ngạn Địch) of Long Môn and 陈上川 (Trần Thượng Xuyên) of Gao-Lei-Lian. Two further names, 黄进 and 陈安平, appear in Li Tana’s fuller reading; the record names two, the reading adds two, and the discrepancy is left standing.

They were not absorbed: the council would not take three thousand armed men into the army, would not let them sit near the capital, and, Li Tana argues, feared a Qing demand for the fugitives. They were posted instead to Mỹ Tho and the Biên Hòa country, on a Khmer frontier the Nguyễn did not control. A defeated foreign army was used as a border garrison: nobody here was admitted.

The settlement 陈上川 founded at Cù Lao Phố, in the bend of the Đồng Nai, flourished for a century, traditionally dated 1679 to about 1776. It weathered the 1747 rising of the Fujian trader Lý Văn Quang and did not survive the Tây Sơn wars. Survivors came down to the arroyo, and the quarter they rebuilt is usually dated around 1778, with dates from about 1775 to 1782 in the literature; 1782 is better read as the year of the Tây Sơn massacre and reconstruction than as a founding. No document founds Cholon. What founded it was water.

The arroyo silted, was dredged, silted again; in 1819 it was cut and embanked as the kênh Tàu Hũ, and the quarter’s address moved with the channel each time the dredgers returned.

Colonial-era postcard of the Cholon canal lined with wooden boats and sampans, waterside shop-houses and a footbridge
A printed postcard captioned “68. — CHOLON. — Le Canal” (A.F. Decoly, Saigon), carrying its own engraved caption and publisher’s mark. The arroyo — later cut and embanked as the kênh Tàu Hũ — lined with wooden boats and sampans, waterside shop-houses and a footbridge.

The book older than the army

The community’s clock does not start with the army. It starts with a genealogy, and the genealogy starts in a Fujian village: 陈养纯 (Trần Dưỡng Thuần), born in 1610 at 玉洲上社, in the twenty-eighth du of 龙溪县 in 漳州府 — an address recorded down to the sub-canton, which is the point.

He reached Vietnam around 1650, on 陈荆和’s reconstruction; the register states no arrival year, and the reconstruction is an inference from what the family wrote about itself. The commune, 明香社 (Míngxiāng shè), is dated by the same scholarship to 1645–1653, probably about 1650, and is attested before the army on stone: a 1653 plaque at the 萃先堂 hall in Hội An, Quảng Nam, naming the commune and its officers. That plaque attests a commune with officers a generation before the posting. It does not date a foundation, and it is not a census.

The gap between stone and record measures how thin the archive is: the dynastic record first names the community in 1746, ninety-three years later. Between them sits a family book compiled four times — early eighteenth century, 1739, 1875, 1930.

It is also one elite family’s book. A single lineage’s register is not the Minh Hương experience but the experience of a family that could afford a register, and a genealogy is advocacy as well as record, arguing for the standing of the generation that pays to have it copied.

Whether Hội An seeded the southern communes, as 陈荆和 held, or the Nguyễn rolled out a village form across the south, as 李庆新 argues from steles down to Hà Tiên, remains open between a genealogist and an institutional historian.

The register, not the concession

In 1698 the lord 阮福淍 created the prefecture of 嘉定府 (Jiādìng fǔ) and sent 阮有镜 (Nguyễn Hữu Cảnh) to survey the eastern frontier. The instruction that came back registered the Chinese descendants at Biên Hòa as one commune and those on the Cholon side as another, enrolling both as ordinary taxable households. The Minh Hương commune became a tax unit.

It was not a land concession, and no source consulted here supports the version in which 陈上川 was given land in 1698: he had been posted to the Đồng Nai in 1679 and died in 1720, and the 1698 act belongs to 阮有镜.

What the register created was a village of a recognisably Vietnamese kind. A 社长 (shèzhǎng) headed it; a 丁 (dīng) roll listed its adult males; it held 公田 (gōngtián), communal land owned by the commune rather than a lineage; it carried a fixed quota, corvée, a limited tax holiday while the land was opened, and eligibility for the examinations. The office list survives on a 1926 stele at the 来远桥 in Cholon, transcribed by 李庆新: a village head, a keeper of the three treasures, a guard officer, a hamlet head, a master of rites, a 管谱 (guǎnpǔ) or register-keeper, a 守银 (shǒuyín) or cash-keeper, and duty officers.

The last two offices are the argument. A merchant guild does not employ a man whose whole job is the genealogy and another whose whole job is the strongbox; a commune does. That is what incorporation looked like from inside: not a certificate, but a page with a headman’s name at the top, a quota in the margin, and a corvée figure at the foot.

One character, one dynasty

Two acts, a year apart, routinely run together. In 1826 the Board of Revenue fixed the Minh Hương head tax at two taels of silver per adult male, the old and sick exempt. The rate matters less than what it replaced: assessments ran from two taels of silver in the centre to cloth and coin in the south, and even the silver columns were set province by province — seventy-odd payers assessed at sixty taels in one, a hundred at two hundred in another. Minh Hương status was fiscal before it was ethnic.

In the seventh month of Minh Mạng’s eighth year — 1827 — a second order rewrote the name. Every 明香 in the realm became 明乡: xiāng, incense, replaced by xiāng, township. Same sound, different character, one dynasty deleted from the map. Older literature says only “the 1820s”; the record is more precise, and the two acts should not be merged into a reform.

The rename can be read as one clause in a broader naturalisation programme, which is how 向大有 reads it — one that also touched examinations, dress and the purging of village registers. What is not interpretation is the condition attached: the same reign forbade Minh Hương to join Vietnamese villages, to call themselves 华人, or to wear the Qing queue, and required them to register their names. 明乡 and 清人 remained two legal estates: naturalised land-holding subjects on one side, foreign traders under 帮 (bāng) surety on the other.

What assimilation bought is visible in the careers. 郑怀德 (Trịnh Hoài Đức, 1765–1825), 吴仁静 and 黎光定 — the 嘉定三家, the Gia Định Three — were Minh Hương descendants who reached the ministries, and 郑怀德’s gazetteer of about 1820 wrote the commune’s location, population, customs and produce into the state’s official geography: the governed writing themselves into the map by which they were governed.

Both halves stand. A community manufactured as a middle estate — register, tax rate, route to office — is not a community dissolved.

The association before the French

The standard line is wrong by about a century. In 1790 an order registered 唐人 (Tángrén), Chinese residents, by native place — Guangdong, Fujian, Hainan, Chaozhou and Shanghai, each named as a 省 and given its own officer, though Chaozhou was a prefecture rather than a province. The 帮 was the corporation that followed, and the 帮长 (bāngzhǎng) its head, answerable to the state for a headcount, a tax, policing and the settlement of disputes among his own people.

That is a Nguyễn instrument, issued from the dynastic record and not a clan book. The French inherited and rationalised it: the congrégation was an administrative unit with a registered electorate and a chairman, and the 1911 voters’ roll at Rạch Giá proves a conversion rather than an invention. The 1871 reorganisation fixed the number of congregations at five.

Manila’s Chinese were a converted, ghettoised minority under a Spanish regime that baptised and penned them; Cholon’s were a registered tax unit under a Vietnamese one that counted and taxed them, as the Binondo story shows from the other side. One was a conversion, the other a register.

A clarification, because it is the easiest mistake here: Cholon’s Chinese were registered by province of origin into five 帮, and the founding cohort came from Zhangzhou and Gao-Lei-Lian — not from the five counties of the Pearl River Delta that most English-language diaspora history describes. Vietnam’s Chinese were sorted by the province they left, not the county their grandfather farmed.

The mill, not the shop

The unit of Cholon’s economy was not the shophouse. It was the 米较 (mǐjiào), the rice mill, and the mill is what turned a frontier market into a near-monopoly.

The first Chinese-owned mill went up in 1877/8 on Brocheux’s dating; another summary gives 1876, a divergence worth flagging rather than adjudicating. The concentration was already extreme: seven of the nine mills with capacity above 500 tons of paddy a day by 1910, in Brocheux’s census. Individual mills ran 500 to 1,000 tons a day, and by 1914 Chinese owners held essentially all of Cochinchina’s reported milling capacity — a monograph’s tabulation, not a customs return.

The labour inside that count was not one interest. Teochew and Cantonese gangs carried and bagged for bosses who spoke their dialect and lived on different streets; the class line ran along the boundaries of the 帮.

What did not happen matters as much. Cholon’s Chinese were not shipped under contract and no coolie fleet called for them; labour arrived through 帮 networks and dialect kinship, a different mechanism from the traffic that carried contract labour to the Americas.

The built object is a market. 郭琰 (Quách Đàm, 1863–1927), probably a Shunde native, was the miller whose estate — not whose lifetime, since Bình Tây opened in 1928, a year after his death — funded the market that opens this article. He did not build it.

The same merchant houses that financed the rice also financed a revolution, a thread this site has followed elsewhere.

Credit without a Chinese bank

The chain began in a field. 青苗 (qīngmiáo), green rice, was an advance against a crop still standing, made by a Chinese merchant to a grower before harvest and binding the grower to sell to that merchant at a price agreed in advance. The merchant took the paddy to a Chinese mill; the mill sold to a French export house that shipped it. Purchase, processing and export were three links in one relationship, not three markets.

What makes the arrangement remarkable is a negative. No known or identified ethnic Chinese bank appears in the literature on Cholon; the Banque de Cochinchine of 1908 was short-lived. Shipping, insurance and discount came from Hong Kong houses, trust from 帮 and lineage ties — precisely the collateral a French bank could not underwrite. Cholon was integrated across the South China Sea, not inside Saigon.

The output was measured in tonnage, and the tonnage is a monograph’s figure, not a customs return: by 20 September 1907, Cochinchina had exported 1,023,649 tons of rice worth 178 million francs. In the same series, rice was more than 70 per cent of the colony’s exports up to 1930, and about half of that went to China, through Hong Kong and Singapore into Shantou and Xiamen.

The collapse was a tenure story. When paddy prices fell on the Cholon market in 1930–34 the cost was paid in title deeds: land moved from small owners to large, and the credit that had financed the crop foreclosed on the field.

The mechanisms elsewhere were a contract or a border: Cuba’s Chinese arrived on coolie ships and were counted at 124,813, Peru’s landed 90,000 to 100,000 and were sent to the guano pits, Guyana’s came under indenture and produced a president. Cholon’s were neither shipped nor excluded.

A government financed per head

The 帮 was financed per head, and that is what turns a set of temples into a government. Chinese residents paid an impôt de capitation, levied on each person and collected through the 帮长 rather than the Vietnamese commune; sources describe the rate as set above the rate paid by Vietnamese. Above, not a ratio: the literature gives no percentage.

Five 会馆 (huìguǎn) were the constituencies of the five 帮, and they have addresses. The Cantonese hall is 穗城 (Suìchéng) 天后庙 at 710 Nguyễn Trãi; the Teochew is 义安 (Yì’ān) 关帝庙. The Hokkien hall is 二府会馆 (Èrfǔ huìguǎn) at 264 Hải Thượng Lãn Ông, built jointly by merchants from 泉州府 and 漳州府 — two prefectures, which is what the name says. Hainanese and Hakka halls complete the five, and the same Fujian community also ran 霞漳 (Xiázhāng) and 温陵 (Wēnlíng) as separate buildings: five halls is a statement about the 帮 system, not about the number of roofs in District 5.

Frontage of Hội quán Nghĩa Nhuận, a Chinese assembly hall in Cholon, with a dragon-ridge tile roof, guardian lions and a plaque above the doorway
Hội quán Nghĩa Nhuận (義潤會館), a Chinese assembly hall in Cholon: dragon-ridge tile roof, guardian lions, and the dark horizontal plaque above the doorway carrying the hall’s name in characters, read right to left, above a Vietnamese banner giving the same name. Halls of this kind held each 帮’s temple, school and arbitration court.

What they ran is reported as hospitals, schools, 义山 (yìshān), charity cemeteries and arbitration, each hall keeping its own deity, festival date and register of the dead. A Cholon Chinese could live an entire life — born, cured, married, buried, a dispute settled — without entering a French court.

That is why the 帮’s ledger is the best surviving census of the community. The 义山 and 义祠 (yìcí) graves record village of origin, generation name and burial society. A Guangdong county that no colonial census kept is cut into a stone in a cemetery outside the city.

The hinge: who was left when the boats sailed

The demographic series is short and its base is the argument. In 1921 the Minh Hương of southern Vietnam were counted at 64,500 — 42 per cent of southern Vietnam’s Chinese population. In 1931 they were 73,000. By 1950 they were 75,000, and 10 per cent. Each share is of the Chinese population, never the general population, and 64,500 is the literature’s figure, not a clean Minh Hương-only census.

The count rose while the share fell. That is a sentence about new arrivals outrunning Minh Hương natural increase, not about a population dying out. The arrivals that outnumbered them came with the mass emigration of the 1840s to 1940s.

By 1950 the category had moved underneath the people it described. An 1842 register rule pushed locally born Chinese into the Minh Hương register at eighteen, which is how Minh Hương stopped meaning Ming loyalist and started meaning born here; 蒋为文, transcribing the colonial figures, doubts the label names Minh Hương at all, since the surveys counted sojourners and filed a naturalised people under the same heading.

Then the state re-sorted them. A 1956 nationality measure — its number unresolved in the sources reachable here, so not printed — formalised who was a citizen. After 1975 Minh Hương descendants were reclassified 京族 (Jīngzú), the Kinh majority, and many hold both identities at once; a 2020–24 field study found members describing themselves as Kinh and as Minh Hương in one conversation. In 1978 an ethnicity re-registration pushed families back into the Hoa column at the precise moment the state began expropriating Hoa.

So the second exodus was a 华族 exodus, and the Minh Hương are, broadly, the ones who stayed. Broadly is the operative word: the campaign fell mainly on Hoa categories, and Minh Hương status had shifted for many families in the preceding decade, so stayed describes a demographic tendency, not a clean binary.

March 1978: the capital, deleted

In March 1978 the state abolished private trade in the south. Vietnamese memoir and diaspora accounts give the morning of 23 March 1978 for the order; the legal instrument is not pinned in the sources reachable here, and no decree number is offered.

The reason a commercial minority was hit hardest by a measure about trade is arithmetic, not conspiracy. A Cholon shop’s entire worth was inventory and goodwill, and both became illegal in the same month: households sold within weeks at a fraction of value, by memoir and diaspora accounts, and the currency reform turned đồng savings into near-nothing. In May 1978 came household class classification and the label tư sản mại bản, the comprador bourgeois, which disqualified children from university and state employment. Firms were sealed, deposits frozen, residence and household registration cancelled. New economic zones followed.

The infrastructure went with the firms: Chinese-language schools and dailies closed, and the five 帮 were folded into a single state-managed body. That is dismantled infrastructure, not a community ended.

One testimony, offered as testimony and not as a rule: in the seven families 黃雋慧 interviewed, the shop was sold before the boat was bought.

The routes out are different quantities and are kept apart. By sea, people left from Vũng Tàu, Mỹ Tho and Cần Thơ; on foot, others crossed from the north into Guangxi and Yunnan. The sea figure is not a departure count: almost 800,000 people arrived by sea between 1975 and 1995. The land count is a crossing count: 265,000-plus entered China between April 1978 and June 1979. They are never added.

The ethnicity of the sea route is the finding. Sixty to seventy per cent of those who reached Hong Kong and neighbouring first-asylum countries in 1978–79 were ethnic Chinese (Amer, in Chan), against a Chinese population of about 2.5 per cent of Vietnam’s. Between 200,000 and 250,000 died at sea, a UNHCR-derived estimate of 2000 in which 250,000 is the ceiling of the range, not a count of the dead. China resettled a reported 260,000–280,000 people on state 华侨农场 (Huáqiáo nóngchǎng) in Guangxi, Yunnan, Guangdong and Fujian — a resettlement figure, and a tentative one, not a count of border crossings.

The mechanism was a class measure; the outcome was an ethnic event, because the class it defined was, in Cholon, a minority, and a community whose wealth is inventory cannot survive a law against inventory. That is not to say the measure was aimed at them.

Passage was priced. Places sold at taels of gold per head; gold leaf was sewn into hems or hidden in food tins; grandparents were left behind as caretakers in case anyone returned; in some families one parent travelled with the children while the other stayed to liquidate. Each detail comes from 黃雋慧’s seven families, not from every family.

Every family narrative that ends in arrival is a sample of survivors.

The halls that outlived their function

District 5 is a working quarter with a heritage overlay, and the overlay is what is growing. The 会馆 were restored with state permission from the 1990s; halls that once collected taxes now run 华文 (Huáwén) centres teaching characters to the grandchildren of the people they registered, and the 元宵 (Yuánxiāo) lantern parade is a municipal cultural asset. The quarter is still a market before it is a monument.

The building this article closes on is 明乡嘉盛会馆, the Minh Hương Gia Thạnh hall. The entrance stele gives 1789 and 81 founders, with 1778 the rival year in the Chinese literature and a 1688 date belonging to a different event, the village rather than the hall. The name was granted under Gia Long — the sources divide on the year, and none is printed. The hall is dedicated to 陈上川, with 郑怀德, 阮有镜 and 吴仁静 alongside, and on the wall a couplet: 耻作北朝臣,纲常郑重;宁为南国客,竹帛昭垂 — ashamed to be a subject of the northern court, content to be a guest in the southern kingdom.

The roof is the agenda item, and it is a budget line, not an elegy. A field study of fifty-nine days between 2020 and 2024 counts about 257 Minh Hương people attached to the hall; its largest service is 祈安 (Kỳ Yên) on the sixteenth of the first lunar month, twenty to twenty-five tables of ten, roughly 200 people. Membership is reported as a male-line rule, verified in the 2020–24 fieldwork rather than 2013: men of twenty and over, descended in the male line from the 81 founders, with two members standing surety. The roof is paid for out of three purses — a grant from the city’s department of information and culture, donations from members and worshippers, and rent from two shopfronts beside the gate. No published source gives the grant’s size. The hall was recognised as a national historical-cultural relic on 7 January 1993.

What a descendant can approach is narrower than the guides suggest. The city’s second national archive centre catalogues twenty-nine fonds for the French period, the governor’s papers for Cochinchina alone running to 2,435.5 linear metres, and inside that index a series filed under congrégations diverses: the 帮 dossiers. A reader’s card takes a passport and two forms; what may be seen is decided dossier by dossier. No Minh Hương-specific register appears in the catalogues reachable here; the dossiers themselves were not consulted.

The stones are more certain than the catalogues. Reading a clan genealogy (族谱, zúpǔ) is a method rather than a promise — the technique is set out elsewhere — and this article’s gift is smaller: a stele with a county name on it. The 侨批 (qiáopī) archive that survives in bulk is a Wuyi and Fujian corpus that does not map onto Cholon, and that difference matters.

Most searches stop at a street and a hall, not a village: the records that would go further were dispersed with the shops.

What travelled was the form. Paris’s thirteenth arrondissement was rebuilt between 1975 and 1985 largely by Teochew and Cambodian Chinese refugees into a second Cholon; elements of the 帮 reappeared in Sydney, Toronto and San Jose — dialect-based, temple-centred, cemetery-owning — in cities chosen by resettlement offices, not the associations. A fourth-generation child now learns 华文 in a District 5 hall — in the city its great-grandparents paid gold to leave — while a cousin may no longer be able to say 阿公 (āgōng), grandfather.

The register in the committee room holds about 257 names; the tables at Kỳ Yên seat roughly 200 of them once a year; and the roof over the room is the item the committee actually votes on.

Street frontage of Đình Minh Hương Gia Thạnh in District 5, with an arched plaque in quốc ngữ reading MINH HƯƠNG GIA THÀNH and ceramic dragon figures on the roof ridge
Đình Minh Hương Gia Thạnh, District 5 — the hall this article closes on. The arched plaque over the gate is in quốc ngữ romanisation, reading MINH HƯƠNG GIA THÀNH; ceramic dragon-ridge figures sit above, and city shophouses and overhead wires crowd the frontage.

Continue Reading

Stay Connected

Follow RootsOfChina on Facebook and Instagram for new article updates, behind-the-scenes from Kaiping, and stories from the Chinese diaspora.

Sources

Suggested Reading

Is this your family’s story?

Roots Of China is built from readers’ family histories. If a story here sounds like your own — a surname, a village, a journey — leave a comment below, or share your family’s story with us and we may be able to help you find your ancestral village.

Share your story →

Leave a comment

0/100