Seeing Like a ‘Sea Gypsy’: An Anarchist History of Hong Kong

Tai O

I’ve mostly finished reading through James Scott’s The Art of Not Being Governed. Like so much of his work, I simply find it brilliant. He animates the worlds he writes about more than any other author I know. There is no linear progress for Scott, but instead endless actions and reactions by groups with varying degrees and types of power over each other. Rather than the world being composed of the powerful and the powerless, there are the various forms of states and people with differing degrees of autonomy from the state.

I think much of what Scott writes about applies to Hong Kong. Let’s start way back, with the “Hundred Yue.” Originally “Yue” was a description of the people living near the Yangtze when that was still the periphery of Han civilization. It evolved into a term for people of south and southeast China, pushing further and further south until ‘Yue language’ became another term for Cantonese. Vietnam translates directly into ‘South Yue.’

It looks a lot like the dynamics Scott describes, with “Yue” being a state term for a very diverse glob of stateless people that don’t really constitute a ‘group’ in any meaningful sense. The term was eventually adopted by different groups in very different contexts to its original meanings. It was an exonym-cum-endonym, an outside term eventually being appropriated as local by those still on the periphery of Chinese state power.

It also serves as a useful point of departure in thinking that this area was ‘always’ Chinese. Instead, there’s good reason to think that everyone from the Austronesians (Filipinos, Indonesians, Malaysians, Malagasy in Madagascar[!]) to the ethnic minority groups of Guizhou, Guangxi, and Yunnan are likely refugees of the Han push southward. What are today Filipinos and the Hmong were probably then the ‘Yue.’ As Scott points out, it’s not that group X left as Group Y came in. Some stayed, some left, and everyone reinvented their identities and histories.

This is to say there is no biological or foundation to what it means to be “Han Chinese,” nor are their essentialist characteristics of what it means to not be Han. The Han are simply those people who, for historical and political reasons, were either captured by Han empires or chose to integrate. This Han civilization was further broken up by a cellular structure (Skinner, 1971),

 Apart from certain remote and sparsely settled areas, the landscape of rural China was occupied by cellular systems of roughly hexagonal shape. The nucleus of each cell was one of approximately 45,000 market towns (as of the mid-nineteenth century), and its cytoplasm may be seen in the first instance as the trading area of the town’s market. The body of the cell – which is to say the immediately dependent area of the town- typically included fifteen to twenty-five villages, usually but not necessarily nucleated.

From within these cells shards of culture were taken from the larger whole during times of dynastic heights. When dynasties fell, they would close off. The first stage of closure is what Skinner calls ‘normative closure,’ which was a “reaffirmation of local norms, a renewed emphasis on the particularized subculture of the local system, and an increased resistance to the influence of exogenous cultural influence.”

How does southern China, and Hong Kong in particular, fit into this ‘cellular’ picture? One of Scott’s arguments is to not look at states as being defined by boundaries on a map. Especially in historical terms, it’s best to see them as a allied urban cores, surrounded by fixed-field agriculture, which in turn is surrounded by peripheries with distance-making geographies – hills and wetlands that are difficult to conquer and control.

One way of visualizing how the friction of distance might work is to imagine yourself holding a rigid map on which altitudes were represented by the physical relief of the map itself. Further, let’s imagine that the location of each rice-growing core is marked by a reservoir of red paint filled to the very brim. The size of the reservoir of paint would be proportional to the size of the wet-rice core and hence the population it might accommodate. Now visualize tilting this map, now in one direction, now in another, successively. The paint as it spilled from each reservoir would flow first along level ground and along the lowland water courses. As you increased the angle at which the map was tilted, the red paint would flow slowly or abruptly, depending on the steepness of the terrain, to somewhat higher elevations. The angle at which you had to tilt the map to reach particular areas would represent, very roughly, the degree of difficulty the state would face in trying to extend its control that far… Our metaphorical map, like any map, though it serves to foreground the relationships we wish to highlight, obscures others. It cannot be easily account, in these terms, for the friction of distance represented, say, by swamps, marshes, malarial zones, mangrove coasts, and thick vegetation.

This fits my observation that Hong Kong is, generally speaking, a terrible place to build a city. It is precisely the sort of place where the paint stops in Scott’s metaphor. Most of the land in Hong Kong today that we’ve built on has been transformed – whether from sea to land or from mountain to mini-plateau (i.e., the earthwork needed to build on The Peak). The population, when the British arrived, was roughly seven thousand, which is the population of Lamma Island today.

Screenshot 2014-01-16 00.38.45

And who were the people? There were four main groupings. The vast majority were Tanka, who fit Scott’s description of a Zomian people nearly perfectly. They seemingly have no history and are defined mostly by how much they escaped the state, for instance being pirates, fifth columns for outsiders [British and the Japanese], and ignorers of state monopolies [like salt]. They were pariahs and outcasts (even today it is Tankan women who clean roads and beaches in my neighborhood) whose name, which translates to “egg people,” was only what others called them.

When they call themselves anything, they prefer 水上人 – the People of the Water. The British just called them “Sea Gypsies”, and in Guangzhou they weren’t allowed off their boats and many of their women were prostitutes, using the boat as a brothel. In Hong Kong, they were the only group to interact with – and socialize with – the new foreigners. They are what remained of the groups that weren’t Han-ified.  As an early Christian missionary described them, they were the people who

… live in the boats, are full of life. They are an aboriginal tribe, speaking an altogether different language from the Chinese. On the land they are like fish out of water. They are said never to intermarry with landlubbers, but somehow or other their tongue has crept into many villages in the Chiklung section. The Chinese say the Tanka speech sounds like that of the Americans. It seems to have no tones. A hardy race, the Tanka are untouched by the epidemics that visit our coast, perhaps because they live so much off land. Each family has a boat, its own little kingdom, and, there being plenty of fish, all look better fed than most of our land neighbors.

Of the three remaining groups, all were self-identified as Han. There was the curious misnomer Punti , or “local”, who began arriving between 1500 to 500 years ago. They can be seen as the forefront of Han state formation. There are the somewhat mysterious Hakka, who have an even longer (and earlier) history, who can be primarily seen as marginalized Han. They were both refugees from the north, the difference being their closeness to the core of Han civilization. The Punti lived in the cities, valleys, and river junctions; the Hakka mostly took to the hills.

There were also the Minnan (or Hokkien / Fujianese / Hokla), who seem perpetually on the move and are the Chinese we run into throughout Southeast Asia and the United States. The Han-identified groups were primarily distributed in the New Territories, as nearly everything else was mountain, sea, and mangrove swamp populated by ungoverned and ungovernable people.

This means that Hong Kong, historically, has been in what Scott calls a non-state space or a ‘shatter zone.’, which is “an effect of state-making and state expansion. Shatter zones and regions of refuge are, then, the inescapable ‘dark twin’ of state-making projects in the valleys” (pg 347). Hong Kong was a swampy, mountainous outcropping populated mostly by Tanka ‘boat people’ with identities defined by and against the larger Han core.

In fact, one could argue that most of Guangdong (outside of Guangzhou), for a long time, was a fracture zone. One of the most interesting historical themes of this region were evacuations to cut local people from the sea. First, there was the Haijin (海禁) during the Ming Dynasty, in force between 1368 and 1567, that demanded:

… all coastal inhabitants should be living less than 20 Li away from the city. Beyond 20 Li, an earthen wall shall be built to serve as a border line; not a single sampan would be allowed to go into the water, no one shall be allowed beyond the border line, any person found shall be executed on the spot. Armed soldiers patrolled the border constantly, would behead anyone caught over the border line.

The Qing later tried the same thing with  Great Clearance / Frontier Shift (遷界令), from 1661 to 1669, but made the distance even from the sea even further at fifty miles! Even settled ‘cores’ like Dongguan and Zhongshan – and even Yuan Long – were a fracture zone with suspect and changing allegiance to the Han state. Even what is called “Punti” in Hong Kong refers to Northern refugees during “the Southern and Northern Dynasties, the Jurchen-Khitan Wars, the fall of the Northern Song Dynasty, the Mongol conquest of the Jurchen Jin Dynasty and finally the Southern Song Dynasty.”

kaiping watchtowers

Aside from the people of the coast, one notes how the Hakka, throughout Guangdong, found remote places to call home and often fortified them. We see traces of this even with the Punti through the 20th century, with the Kaiping Diaolou towers that dot the landscape on the western side of the Pearl River Delta. Our histories are almost entirely that of the “state spaces” and:

…neglect or ignore altogether both “nonstate spaces” beyond their reach and the long periods of dynastic decline or collapse when there is hardly a state at all. In a truly evenhanded, year-by-year, chronology of precolonial, mainland Southeast Asian states, most of the pages would be blank. (pg 55)

Scott is asking us to see most these original Hong Kongers – including, perhaps, the ‘original’ Punti – as people who intentionally lived with a distance Han civilization in an isolation that allowed them to define their own relationship with the state. Scott would  have us question how ‘ancient’ and discrete the original groups were – did they accept new refugees? Were a new wave of Han migrants not accepted into Punti culture? How much did the poorest and most desperate refugees simply emulate the lifestyles of one of the marginal groups, either a Hakka in the hills or a Tankan in the water eventually adopting the language and fluid culture? Or, as we’ll get to, did Tankan take advantage of political openings to redefine Punti?

Hong Kong’s socio-cultural origins are a fracture zone with multiple, and often simultaneous, identities. How ‘Han’ or ‘Chinese’ the people who lived here were probably waxed and waned with the rise and fall of dynasties. When the reach of the state was great, and when the terms of trade were good, Tanka, Punti, and Hakka people likely traded and interacted a great deal with the ‘core’ Chinese culture and economy. When dynasties declined, merchants would probably have become pirates on the move and cultural clusters with changing alliances and allegiances.

Whatever we think of as ‘Hong Kong culture’ seems to have been a co-creation between the British and the Tanka. The British created the geopolitical space of Hong Kong; the Tanka filled in a relatively empty place to leverage the political and economic opportunities of working alongside the new rulers:

Tanka people, forbidden by Chinese law (since A.D. 1730) to settle on shore or to compete at literary examinations, and prohibited by custom from intermarrying with the rest of the people, were from the earliest days of the East India Company always the trusty allies of foreigners. They furnished pilots and supplies of provisions to British men-of war, troopships and mercantile vessels, at times when doing so was declared by the Chinese Government to be rank treason, unsparingly visited with capital punishment. They were the hangers-on of the foreign factories of Canton and of the British shipping…

Perhaps most interesting is how the Tanka then became Chinese – and then forgot about it. The original Tanka and Hakka inhabitants of Hong Kong did not speak Cantonese as a mother tongue, neither did most of the migrants that came after. What seems to have happened instead was that marginalized Tanka and Hakka began a process of transforming into local/’punti’ people as they came off of their boats and down the hills. The Tanka:

…invaded Hong Kong the moment the settlement was started, living at first on boats in the harbor with their numerous families, and gradually settling on shore. They have maintained ever since almost a monopoly of the supply of pilots and ships’ crews, of the fish trade and the cattle trade, but unfortunately also of the trade in girls and women. Strange to say, when the settlement was first started, it was estimated that some 2,000 of these Tanka people had flocked to Hong Kong, but at the present time they are about the same number, a tendency having set in among them to settle on shore rather than on the water and to disavow their Tanka extraction in order to mix on equal terms with the mass of the Chinese community. (Eitel, p. 169)

This is to say an ostracized group in China, when presented an opportunity, transformed themselves into a Han Chinese group. Hong Kong opened a space for the Tankan to play both roles, to appear Han while remaining far removed from the Chinese state. They adopted Cantonese, a trade language they had long known. But they have, mostly, chosen to leave behind cultural traits that other Chinese identified as ‘backward.’ It’s curious that there is no study of Tanka language and culture in Hong Kong. This seems to be their collective choice.

The arrival of the British seems in equal measures the cause of, and because of, the decline of the Qing Dynasty. There was a violent process of state formation, reformation, and resistance both before and long after they set up shop in Hong Kong. There were civil wars, epidemics, revolutions, invasions, famines, and purges. Hong Kong, a new core protected from these violent episodes, absorbed refugees from throughout China – particularly from the Chaozhou-Shantou (Chaoshan) area and Shanghai.

Old Sampan new Shipliner

I think it’s fair to say that the Tanka radically opened a new space for the social production of what we would now consider popular Hong Kong culture. The British enclosed a relatively empty “fracture zone,” the Tanka colonized what they could and took on a new ‘local’ identity, which in turn paved the way for a flood of new migrants to become ‘local’ as well. A hundred years ago Hong Kong was a far more diverse place, but it has coalesced around (more-or-less) a single, relatively homogenous culture that poses as local/’Punti’ culture. The Hakka, Tanka, Chaoshan, and Shanghai families all seemed to try to forget there own histories and languages to create and sustain this. In fact, it looks a lot like the historical waxing and waning of Chinese “openness” that Skinner describes.

Let’s take stock of where we are: my argument right now is that demographic bulk of Hong Kong, at the outset, was not ethnically Chinese. The disruption that the British caused allowed them to come off their boats and take advantage of their relations with the British, displacing and becoming the new ‘Punti.’ This is to say sometime in the early 20th century a ‘Hong Kong culture’ different and distinct from the Punti and other ‘ethnic’ cultures, began to emerge with Cantonese language as an essential element but spoken primarily by non-native Cantonese speakers.

It would seem fair to say they absorbed the Punti rather than the other way around. Sampans in Hong KongThey were able to do this because, even for the ‘real’ Punti, culture might be described as “a bandwidth of traits or identities that could be deployed or performed as the situation required [with] A person’s… identity [being] the repertoire of possible performances and the contexts in which they are exhibited” (pg 275). The Tanka didn’t disappear, as one sometimes reads, they merely changed their ‘cultural bandwidth’ to blend-in and co-create a new socio-political space with the British. This, in turn, opened a space for new migrants.

These new Hong Kongers, much like their predecessors, were “cultural amphibians.” As Cheng Kai-Ming observed, “we are British at work and Chinese at home.” I would interrogate that further and say, that for a long time, there were multiple meanings of “Chinese” that began converging into something with a more local context, with various elements of divergence from the main Chinese and English core. The Tankan insist they’re Chinese, even if everyone else says they’re not. But in Hong Kong, they could be “Chinese enough” to other Chinese, “Western enough” to the British, and the bandwidth in-between that we would call “local Hong Kong culture” today.

One could make much work of Scott’s work on “the art of not being governed” in context to the discontent with the handover of Hong Kong to China. While Hong Kong certainly isn’t a non-state space, we see that the political and economic rise of Mainland culture has seemed to entrench an identity that was created in the 20th century. How much of the pessimism in Hong Kong about reintegration with the Chinese polity stems from the fact that many of the “local” people were, in fact, people who had spent more than two thousand years on the run from the Han Chinese state?

One might ask, then, whether ‘essential’ Hong Kong cultural traits, like the use of Cantonese, were used because they were mostly illegible to the wider Chinese state? How much more autonomy did the new Hong Kongers have by not using simplified script or Mandarin in the classroom and media? How much did the other Han, like those from Chaoshan and Shanghai, encourage their children to speak Cantonese to build a socio-linguistic wall with the polity that they had fled? How much do the idiosyncrasies of our Special Administrative Region, ones that Hong Konger’s fight so hard to preserve, serve to make sure there are “no handles” to easily grasp and manage the people (pg 351)? Returning to Skinner’s model, doesn’t this also look a lot like the ‘nominal closing’ that he describes – the retreat into manufactured localism?

It wouldn’t necessarily be planned, but an evolution of one choice being preferred over other options. There’s no reason Hong Kong couldn’t have made a collective shift towards Mandarin like Singapore and Taiwan did – but neither Singapore or Taiwan were incorporated into the Beijing state. In so much as language was selected to distance Hong Kong from the core, one might wonder whether if Cantonese had become the official language of China – as almost happened – would Tankan, Chaoshan, Hakka or even a pidgin have come to dominate in the territory? This is Scott’s most compelling argument – social groups are both repelled from and attracted to the state core. Never doubt the social fluidity that would have allowed Hong Kong’s de facto language and culture to turn on a dime if circumstances warranted.

This all to say, how much of what we now call Hong Kong culture is the accumulation of adaptations to distance itself from the problems that came with both new and “old” China? I think the most we can say for sure is this: the Hong Kong culture we see today is not merely a colonized offshoot of Han society that was here before the British arrived. Nor was it a mere shelter from the violence of state creation further north, and thus an incubated and more ‘authentic’ and preserved Chinese culture than we see in the north. It is instead a constantly evolving cultural shard, geographically situated in a historically non-state space – a cultural ‘shatter zone’ enclosed by the British – with an evolved (and evolving!) bandwidth of cultural adaptations that have allowed it to approach mainland China, and the world, on its own terms.


  1. Sadly I lack the staying power to take all this in in one go, particularily since I am not really interested in China, but it seems to me that much of European history could really do with this approach, Turkey in Europe and its marches, i.e. my world, possibly right up to the present.

    December 2, 2013
  2. […] There is a demographic edge to this, as I wrote about in the introduction: young people see themselves as far more as Hong Konger more than ‘Chinese’ in the way Beijing would see it. Hong Kong, for many different reasons, is culturally and socially heterogenous with the mainland. Hong Kongers see difference everywhere: in politics, in education, in language, in taste, in humor. In many ways, the British project to make Hong Kong more ‘Chinese’ was, in fact, to create a peculiarly stamped Hong Kong nation. […]

    February 1, 2015

Leave a Reply