Livestock Revolution and Ghostly Apparitions: South China as a Sentinel Territory for Influenza Pandemics

Frédéric Keck. Current Anthropology. Volume 60, 2019.

The chair of the Working Group on the Anthropocene, paleobiologist Jan Zalasiewicz, recently declared that fossil bones of broiler chickens accumulated after 1945 could be used as a marker of a new geological epoch in which humans have changed their environment (Bennett et al; Carrington). These chickens have been genetically selected in such a way that they are twice the size of their wild ancestor, the red jungle fowl (Gallus gallus), domesticated in South China between 10,000 and 7,000 years ago for divinatory and culinary practices (Simmons:298). Indeed, 1945 marks a new threshold in poultry production with the “livestock revolution”: the transformation of small poultry farms into big factories through confinement, concentration, and integration. Such a process has been so successful in North American poultry farms that it was extended to the rest of the world and the food industry. When it was applied in China in the context of the reform policy in the 1980s, this industrial process, also called “chickenization” (Silbergeld:61), dramatically increased the number of poultry raised for meat consumption. Contemporary broiler chickens grow five times faster than meat chickens raised in the 1950s, and their bones are less dense and more deformed. Thus if we look for quantitative indicators of the Anthropocene by considering chicken remains, we could identify three layers with different sizes and densities around three thresholds: the domestication of chickens in China, the industrialization of chicken production in North America, and the expansion of this mode of production in China.

There is another way to tell the same story by including microbes. At the end of the 1970s, virologists warned about the potential of South China as a site of emergence for influenza viruses due to a traditional system of agriculture enhancing the proximity among humans, pigs, and ducks. This warning made sense in the framework of the ecology of infectious diseases relating microbial mutations to ecological changes (Anderson): influenza viruses coevolved with birds, pigs, and humans since the threshold of domestication, and the Industrial Revolution disrupted this ecosystem and amplified lethal viral mutations. The emergence of pandemic influenza viruses was thus described by experts of the Food and Agriculture Organization) as an unintended effect of the livestock revolution (Delgado et al). It can also be described as a feral reversal of the Anthropocene, every threshold in the history of domestication being related to a change in the coevolution between humans, animals, and microbes. Virologists tell us that viruses emerging in China with the domestication of the chicken have been amplified by the division of labor in North America and returned to China in a monstrous form (Davis; Wallace et al).

The same story can be told from a third perspective, not from the models of global experts but from the view of South China by its inhabitants. In Han China, “bird spirits” (shen niao 紳鳥), the most prominent being the phoenix, were considered by scholars as signs of changes and often were associated with tombs (Sterckx:181). In the eighteenth and nineteenth centuries, civil servants sent to South China described it as a place full of miasma (zhang 瘴) and associated it with “cold diseases” (wenbing 温病) (Bretelle-Establet; Hanson). Mass killings such as the Taiping Revolution or the Great Leap Forward led to a proliferation of experts in the communication with ghosts, considered as spirits of the dead roaming without a place to rest (Mueggler; Weller). These complex beliefs and practices were associated by Chinese citizens with the perception of flu-stricken chickens with influenza not as commodities infected with a new virus but as ghosts of epidemics (yigui 以鬼) filling the environment and claiming debts from the living (Benedict:111).

In this article, I combine these three stories ethnographically to describe different perceptions of South China as a sentinel for avian influenza. I define “sentinel,” in contrast with the indicator, not as quantitative but as a qualitative logic of early-warning signals (Keck and Lakoff). A sentinel is not only a technical device of prediction or a military post of surveillance and preparedness: it is also a territory where different actors, human and more-than-human (fowl, pigs, viruses, ghosts), interact in the anticipation of future threats. I thus offer materials for a “patchy” description of the Anthropocene. What if, in the sedimented layers of chicken bones taken by geologists as indicators of the Anthropocene, there were not only viruses but also ghosts? A bird can be perceived either as an indicator of a geological age technologically defined, or as an industrial commodity whose risks must be managed, or as a living entity endowed with intentions. The shift from the logic of indicators to the logic of sentinels leads me to study how beliefs and models are embedded in practices when humans interact with a bird potentially infected with influenza. Since models of influenza pandemics are tainted with colonial and secular histories of public health and surveillance, I want to show that the modes of perception of sentinels capture alternative ways to think and live through the Anthropocene, thus opening paths for postcolonial and nonsecular histories of the livestock revolution.

An ethnography of avian influenza moves from the global models of the livestock revolution to a local territory such as Hong Kong, where viruses emerge and ghosts appear in the interactions between humans and birds. Analyzing the different meanings of the term “release” as it is applied to potentially infected birds, I describe how these different modes of emergence grow out of an uncertain interaction. If the concentration of chickens in industrialized poultry farms has led, following experts, to the “release” of emerging viruses as well as bird spirits, we can ethnographically describe the gesture of releasing a bird from a farm or a market. To what extent does releasing a bird replay the threshold of domestication in an ethnographic setting? What kind of hope does it leave in the ruined landscapes of farms eradicated from live poultry infected with avian influenza? How do sentinel birds appear as phoenixes emerging from the ashes of poultry slaughtered as a preemptive measure against avian flu?

This article thus develops three ethnographic scenes in Hong Kong, reflecting the three narratives of the Anthropocene presented in this introduction: an expert’s view of Hong Kong as a sentinel post on the edge of the epicenter for pandemic flu; a farmer’s view of Hong Kong as a colonial experimentation on ways to raise chickens industrially, in which unvaccinated chickens are used as sentinels of avian influenza; and a bird-watcher’s view of Hong Kong as a sentinel territory full of bird spirits. In these different forms of the sentinel, the military post, the industrial device, and the environmental territory, I ask what it means to release a bird on the threshold of domestication, considered not as an evolutionary step but as a space of friction and trouble (Haraway; Tsing :176) where humans and birds enter into an uncertain interaction.

Hong Kong as a Sentinel Post for Influenza Viruses

My fieldwork in South China, intended to understand the different meanings of influenza pandemic preparedness between 2007 and 2014, started with an ethnographic study of the department of microbiology of the University of Hong Kong created by Kennedy Shortridge in 1972. He had been trained at the school of microbiology founded by Frank Macfarlane Burnet in Canberra with Robert Webster, considered “the pope of influenza” for his prophetic claims about the next pandemic virus (Caduff). With his colleague Graeme Laver, Webster had observed that the antibodies to flu viruses he was modeling in the lab could be sampled from wild birds found dead in the Australian seashore. He consequently hypothesized that flu viruses circulating among humans first mutate among birds, and that waterfowl, diversified in many species living in the same environment, constitute a perfect “reservoir” for the emergence of new viruses that can cause pandemics among humans. Webster collected viral samples from wild birds all over the world to compare their antigenic information and trace their evolution. His lab in St. Jude Children’s Research Hospital in Memphis, Tennessee, concentrates the diversity of flu viruses in the world and has become a reference center for the World Health Organization (WHO) when it must identify a new influenza virus.

If Webster was the pope of influenza, Shortridge became its military officer when he moved to Hong Kong to create the department of microbiology as an outpost to perceive the signals of the enemy on the front line. An influenza virus called H3N2 had emerged in Hong Kong in 1968 and caused a pandemic that killed 1 million people. But because the People’s Republic of China was not a member of the WHO, there was a gap in the surveillance of the mutations of flu in China. Shortridge built networks of personal relationships (guanxi) with veterinarians in Guangdong and collected samples of flu viruses among ducks and pigs in the area. He had observed that rice paddies of South China used wild ducks (Anas platyrhynchos) as biological pest controls, a system known as daotian yangya (Zhang et al.), thus bringing them in close proximity with humans and pigs. Pigs are genetically considered as a “mixing vessel” where flu viruses “reassort” between humans and birds, because they have receptors in their respiratory tracts for bird and human viruses (Webster and Campbell), while ducks are described as “sane carriers” because they shed flu viruses through their digestive tracts without being infected by them. Shortridge asserted that this traditional ecology was an “influenza epicenter” for the rest of the world. “The densely populated intensively farmed area of Southern China adjacent to Hong Kong,” he wrote with the renowned British influenza expert Charles Stuart-Harris, “is an ideal place for events such as interchange of viruses between host species” (Shortridge and Stuart-Harris:812). To support this hypothesis, he remarked that the Chinese character for house (jia 家) contains a pig under a roof—as if one could see the mutations of viruses coming from animals when looking at the various traits of this traditional character.

In the ecology of infectious diseases, a school of medical thinking initiated by Burnet, microbiologists have borrowed concepts such as “avian reservoir” or “influenza epicenter” from other fields of knowledge, such as nuclear physics and seismology, to describe the antigenic shifts they observe in their viral samples. But these ecological concepts also have practical consequences because they allow microbiologists to intervene as public health experts. Shortridge was both an observer of and an actor in the “livestock revolution”: he recommended biosecurity measures to control the proliferation of microbes by making an opposition between dirty tropical landscapes and clean interventions of the government, but this recommendation privileged big industrial farms who can afford to implement these biosecurity measures in compliance with governmental rules.

Influenza research in Hong Kong was the outcome of a colonial and postcolonial history of science in the Asia-Pacific area, which applied immunological concepts to Asia from Australia, conceived as the vanguard of the Western world (Anderson). Following an outbreak narrative that distributes blame for epidemic events, it stigmatizes groups and places as traditionally unclean (Wald:8). It conceives the colony of Hong Kong as a reduced model of South China, where viral mutations can be observed on a regular basis and an alarm can be sent to the rest of the world. Because Hong Kong is situated at the end of the Pearl River Delta, it has many wild birds coming to roost in the shallow waters and many immigrants crossing the border in the hope of a better life. The mutations of viruses that accompany the movements of birds and humans can thus be simulated in Hong Kong as an experimental live model, contrasting modern Hong Kong and traditional China.

This colonial configuration was confirmed in 1997, when Hong Kong was returned to Chinese sovereignty. A new flu virus named H5N1 appeared in the poultry markets of Hong Kong, killing 5,000 chickens and one child. By contrast with ducks, chickens die massively and rapidly from highly pathogenic influenza, and by contrast with humans, their symptoms appear in the digestive tract rather than in the respiratory tract. Shortridge set up an inquiry in the thousand poultry markets of Hong Kong and found that 36% of chickens tested positive for H5N1. He recalled, “One moment birds happily picked their grains, the next they fell sideways in slow motion, grasping for breath with blood slowly oozing from their guts. I had never seen anything like it. I thought, ‘My God. What if this virus were to get out of this market and spread elsewhere?'” (quoted in Greger:35). Here Shortridge uses imagination to simulate the release of viruses from markets on the mode of the “as if.” He later said that the dead bodies of chickens in the Hong Kong wet markets reminded him of his mother’s description of the human victims of pandemic H1N1 influenza (or “Spanish flu”) in his native Queensland in 1918 (Greger:xi). In Shortridge’s narrative, bird flu viruses connected dead chickens in Hong Kong markets, memories of dead bodies in Queensland, and potential threats in Chinese farms.

Shortridge’s strategy was to make a vaccine for new flu viruses when they mutate among birds, pigs, and humans. But because this virus was very lethal among chickens, it was impossible to make a vaccine. Flu viruses had been circulated and attenuated in chicken eggs since their process of replication had been studied in this animal model by Burnet in the 1950s. There is consequently a whole industry of egg rearing for vaccination that parallels the emergence of flu viruses in avian reservoirs and that may explain the analogies between labs and farms. Still today, despite research in other techniques such as cell-based and recombinant flu vaccines, 90% of flu vaccines are produced in eggs, which is time-consuming, as three eggs are necessary to make one vaccine for three flu strains. It is estimated that 500 million high-quality chicken eggs are used every year for the production of immunization shots (Caduff:89; Manini et al).

In November 1997, Shortridge—in agreement with experts sent from the US Centers for Disease Control and Prevention—recommended that the Hong Kong government kill all the live poultry on the territory in order to eradicate the animal reservoir of H5N1. This measure had some efficacy: the H5N1 disappeared until 2002, and when it reappeared, a similar killing was conducted by the agriculture authorities, who became used to putting chickens in bins with gas and then incinerating them. “We didn’t cull; we conducted a slaughter!” Shortridge told me (interview, Hong Kong, February 2009). When I asked him how the massive killing was acceptable to the Hong Kong citizens, he said that 5 years earlier, he had recommended the closure of the horse races in Hong Kong because there was an outbreak of equine influenza, which is lethal in horses but not transmissible to humans. In a city where horse racing provides the only opportunity to gamble, this closure was, in his opinion, more costly to many Hong Kong citizens than the killing of their backyard poultry. This anecdote tells a lot about the differential value of animal life in the context of zoonotic outbreak, and it shows that the closure of livestock activity was in itself a costly measure in a liberal economy that relies on the circulation of persons and commodities. However, it does not tell much about the meaning of the slaughter for Hong Kong citizens, as we will see later.

Shortridge later justified the repeated killings of poultry infected with bird flu as a preemptive measure to avoid a pandemic:

Poultry were killed market-by-market as signs became evident, leading to the pre-emptive slaughter of all poultry to prevent human infection. Early detection and reaction was the order again in 2002 and 2003. Thus, there now lay the prospect for influenza-pandemic preparedness not only at the human level but, better still, at the baseline avian level with the ideal that if a virus could be stamped out before it infected humans, an influenza incident or pandemic will not result. In 1997, the world was probably one or two mutational events away from a pandemic, while in 2002, with earlier detection, it was probably three or four events away. (Shortridge:10)

The idea that the early detection of viruses in their animal reservoirs and rapid intervention led to stopping a pandemic before it happened was confirmed in 2003 with the SARS crisis (Severe Acute Respiratory Syndrome). A coronavirus circulating among bats and transmitting accidentally to humans through the civet cats consumed in Chinese traditional medicine returned to its animal reservoir when civet cats were killed and their sale forbidden. Shortridge then wrote an article with his two colleagues at the Hong Kong University department of microbiology who had identified the SARS virus in animals and humans, Guan Yi and Malik Peiris, in which he concluded: “The studies on the ecology of influenza led in Hong Kong in the 1970s, in which Hong Kong acted as a sentinel post for influenza, indicated that it was possible, for the first time, to do preparedness for flu on the avian level” (Shortridge, Peiris, and Guan:79).

Biosecurity measures applied in Hong Kong were taken as models by WHO when the H5N1 virus moved from Hong Kong to the rest of Asia and then to Europe in 2005, increasing the fear that this zoonotic virus would become pandemic. They belong to a series of techniques of anticipation of disasters invented under the Cold War for nuclear blast and applied to all-hazards management, such as early-warning signals, worst-case scenarios, and stockpiling of emergency equipment (Lakoff). In August 2005, Michael Osterholm, an epidemiologist who had just been appointed secretary of the Department of Health and Human Services and who was appointed to the newly established National Science Advisory Board on Biosecurity, published an influential paper in the journal Foreign Affairs in which he justified the massive stockpiling of Tamiflu, flu vaccines, masks, and protective equipment launched by the Bush administration. His article ended with these figures:

The population explosion in China and other Asian countries has created an incredible mixing vessel for the virus. Consider this sobering information: the most recent influenza pandemic, of 1968–69, emerged in China, when its population was 790 million; today it is 1.3 billion. In 1968, the number of pigs in China was 5.2 million; today it is 508 million. The number of poultry in China in 1968 was 12.3 million; today it is 13 billion. (Osterholm:35)

These statistics were then often repeated in presentations by public health officials, but they are difficult to confirm. It is ethnographically more meaningful to ask how the livestock revolution was experienced by actors who faced the growing risk of avian influenza in the poultry farms of South China.

Model Farms in the Hong Kong New Territories

The industrialization of chicken production started in the United States after 1945 with companies such as Tyson, Holly Farms, and Perdue, who reduced the time span to turn a day-old chick into a broiler through intensification and diversified poultry products through advertising. With confinement, animal numbers are not limited by the supply of natural resources or the carrying capacity of a territory (Silbergeld:39). With transportation, they do not need to be raised close to the place of hatching. While in 1929 300 million chickens were widely dispersed across the United States, by 1992 US poultry production was largely concentrated in a few states in the South, hosting 6 billion broilers at an average flock size of 30,000 birds, with increasingly dangerous conditions of life for animals as well as for humans (Kirby; Striffler). In the 1970s, this model was developed in Asia through companies like Charoen Pokpand (CP), run by a Chinese businessman who started his activities in Thailand and Guangdong (Davis). In 2012, CP controlled one-third of China’s poultry production and produced 1 million chickens at the same time in Thailand (Silbergeld:77).

The livestock revolution was brought to Hong Kong as early as the 1950s by two Jewish bankers from Iraq, Lawrence Kadoorie and Horace Kadoorie, who taught Chinese immigrants in Hong Kong—with the motto “Helping people to help themselves”—how to transform chickens into commodities, build cages, select breeds, hatch eggs, ventilate (as backyard poultry was replaced by closed farms), and vaccinate (particularly against Newcastle disease, that killed chickens massively without being transmissible to humans). They particularly raised the yellow-feather chicken (huángmáo jī 黃毛雞), which requires more days to breed but is more appreciated by Chinese consumers than ordinary broilers. Because of the boycott of Chinese products by the United States, this breed was raised in Hong Kong and sold to the Chinese diaspora in North America. While in 1949 there were 145 farms breeding around 1,000 chickens each in Hong Kong (Yeung), this soon expanded to more than 1,000 farms raising around 100,000 chickens each.

The poultry production in Hong Kong relied on the importation of eggs from China, which led British inspectors to control that ducks were actually raised in Hong Kong and not in the communist mainland.

[Many of the] ducks processed in Hong Kong and then exported to America came from eggs laid in China and brought to Hong Kong to hatch. Were the ducks from these eggs communist ducks or true-blue British ducks? The correspondence on the subject was voluminous before a solution was finally reached. Provided that an inspector was present when the duck was hatched, that he forthwith rubberstamped the duckling’s foot, and that on reaching maturity a further marking was put on the duck, then the ducks might be slaughtered, dried and admitted into the United States. (Grantham:166)

After 1997 and the emergence of bird flu, the threat from the border was formulated in terms of mutating viruses and not communist ideas. The new Hong Kong government under Chinese sovereignty decided to rely on massive Chinese imports and strong biosecurity measures to get rid of the threat of bird flu. It made generous offers to poultry farmers who abandoned their activities through a Voluntary Surrender Scheme.

However, consumers were attached to buying local breeds in live poultry markets and were ready to pay twice the price for a Hong Kong chicken. They invoked symbolic reasons (offering a live chicken to a guest is considered a sign of honor, based on its ritual consumption in soups during ceremonial meals) but also the fact that Hong Kong live chickens were more healthy and tasteful by comparison with chilled imported poultry that “tastes like wood” (Liao et al). The import of live poultry from China declined from 30 million in 2002 to 10 million in 2007, while the import of chilled poultry increased in reverse proportions, but the production of live poultry in Hong Kong was not entirely stopped, and there were 30 farms raising around 1 million chickens when I started my fieldwork in 2007. All these farms, located in the New Territories (the part of the continent between Hong Kong island and the border with China), kept their chickens confined, as it was forbidden to raise free-grazing or backyard poultry or to buy live poultry directly at the farm. All live poultry raised in Hong Kong or imported from China had to be transferred and controlled at the Central Market of Cheung Sha Wan before being sent to retail markets where consumers could buy them. Live poultry markets had to clean their stalls every night and stop their activity for one day per month so that viruses would not accumulate there. They were separated from the rest of the market, and customers were informed of the risks of consuming live poultry. The possibility to buy live chicken locally had been an integral part of the value of poultry consumption in South China, and it persisted as a form of resistance to the livestock revolution.

For comparison, Poyang Lake, in southern Jiangxi Province, is also monitored as an “influenza epicenter” because over 14 million ducks are raised there (Fearnley), almost one-fifth of China’s total duck production—it is forbidden to raise ducks in Hong Kong. For farmers in this area, the pertinent distinction is not between wild and domestic ducks, as in the discourse of flu experts, but between sideline ducks, which are raised free-range for personal consumption and which display a wide variety of species, and husbandry ducks, which are raised indoors with standardized species and sold to the market (Fearnley). The threat of avian flu (qinliugan 禽流感, in the words of Chinese farmers), exposes this local production to a global environment, transforming poultry from sources of food and income into signs of potential loss for the household.

In August 2009, I worked in a farm that had been infected by H5N1 6 months before, situated in the village of Yuen Long, in the north of the Hong Kong New Territories. The owner of the farm, Wang Yichuan, was the head of the Hong Kong Poultry Farmers Association. He communicated regularly with the media, who described his enterprise as a “model farm” (mofan nongchang 模范农场). Accepting an anthropologist to work in his farm was probably part of his media communication. A former truck driver, he had bought this farm in 1994 after he read in the newspaper about a Chinese man from Singapore who became rich by raising chickens. His wife, originating from a poultry-breeding family, advised him not to engage in this business, which she found tiring and risky. Wang’s first difficulties came from the treatment of the waste, which triggered complaints from the neighborhood. Then, after 1997, they came from bird flu. It was a common occurrence to find 10 dead chickens per day, he said, but when he found 200 on December 6, 2009, he realized that there was something wrong. Even more alarming, half of these dead poultry were “sentinel chickens” (shaobingji 哨兵鸡): unvaccinated chickens placed at the ends of the rows of cages, with a rate of 60 sentinels for a flock of 3,500 broilers or 500 breeders. Their death signaled that the H5N1 virus was in the farm.

Consequently, he was obliged by the government to kill and bury the 70,000 chickens living on the farm as well as 25,000 fertilized eggs, clean all the equipment, and change the nets. “There is not one feather left,” he proudly told me. He was promised compensations but never received them. When I worked at his farm in August 2009, he raised 30,000 chickens (his license allowed him up to 100,000) and only employed four workers; but he had bought another farm in mainland China, where he employed four other workers, and he was regularly crossing the border for his business. “Bird flu is a risk I am ready to take,” he told me, “because I like the poultry business. You can lose a lot, but you can also earn a lot!”

Working on Wang Yichuan’s farm made me realize that biosecurity is not only a global technique of preparedness for future epidemics but also “a more banal present-tense, enacted regime of corporate governance, alongside a subtly inculcated ethic for living amid industrial animals” (Blanchette:651). Most of the measures recommended by the government after the outbreak had not been implemented. There was a pond for truck wheels to avoid viral import from other farms, but ponds for boots between the buildings of the farm were bypassed by workers. Wire nets to avoid contact with wild birds had been bought, but they were still lying on the floor, and sparrows were entering the buildings easily. Experts had also recommended mixing ordinary chickens and sentinel chickens so as to avoid the concentration of viral load at the end of the rows, which led workers to be more attentive to their distinction in the absence of separation by space or species. Sentinel chickens were sold to the central markets at the end of the week just as ordinary chickens if they were not sick. They had become part of the daily life of the farm and could even be eaten at the end of the day by poultry workers.

When I followed the truck driver from the farm to the central market, we passed by a series of containers prepared to leave on boats and inscribed with the motto “We carry, we care.” Poultry farms have traditionally been places of care in an intermediary space between hatcheries and markets. But because of biosecurity measures, they have been conceived as confined spaces where commodities are carried under a growing state of vigilance. The “carrying capacity” (Swanson) of the Pearl River Delta had been put under stress by the threat of avian influenza: once a coexistence between humans, poultry, and wild birds (Liu), it is now a competitive economy in which the Hong Kong brand maintains its value.

If Wang Yichuan’s farm, at the edge of the delta, was still running activity between hatcheries and markets—and despite its opening to wild birds—the Kadoorie Farm, situated in the middle of the New Territories, kept its chicken population entirely closed. In 1997, with the emergence of H5N1 and the threat on poultry farming activities, the Kadoorie Farm could not sell its poultry products anymore, and it became a center for the conservation of local breeds. Its leaders were proud to say that during the Cultural Revolution, Chinese pure breeds had disappeared from mainland China, thus giving their farm the role of a repository for local breeds that could be sent to China to be raised anew.

The visitor who enters the Kadoorie Farm today sees botanical gardens, wild birds, crocodiles, monkeys, and, preserved from the gaze of visitors, 2,000 chickens with a warning: “The Chicken Display House will be closed until further notice to ensure the chickens at the Kadoorie Farm and Botanical Gardens are protected from any possible outside contamination while bird flu concerns still exist in Hong Kong.” The Kadoorie Farm has its own system of alert, more severe than that imposed by the government to other poultry farms, with three levels (vigilant, serious, urgent). Indeed, in case of an outbreak of bird flu in the surroundings of the farm, the cost of culling would measure not the value of the meat but the genetic knowledge preserved by decades of selection.

I met Shing Tam-Yip when he was the head of the breeding team, taking care of the 2,000 chickens and nine pigs. A passionate bird-watcher and plant scientist trained at Hong Kong University, he had wished to build his own farm, but the environmental impact assessments were too stringent, and he accepted the job offer from Kadoorie Farm. He told me that before 1997, the selection of the purest breed was a public ceremony but that it became hidden after 1997 for safety reasons. Selection consisted in sexing the males from the females, ringing the females, preserving the males who had the highest value, and destroying the rest of the males. Shing contrasted the killing of day-old chicks for selection to the massive killing of poultry as adds preventive measure against bird flu: “We use CO2. This is not torture. For ten seconds they shake a lot, but after twenty seconds it is silent. When they killed poultry at the central market of Cheung Sha Wan, the quantity of gas was not enough. Poultry died after a very long time. It was really torture. People watching on television felt distress.”

It is interesting to contrast the views of chickens by Wang Yichuan and Shing Tam-Yip. For Wang, chickens are commodities whose circulation must be secured, and the loss of the flock is a risk that must be anticipated. For Shing, they are tokens of a species whose loss would be a disaster and whose preservation justified the killing of young chicks through sexing. Shing told me: “In mainland China, people don’t have the concept of species; for them, it is just meat.” While for Wang wild birds are pests that can introduce viruses in the farm and must be repelled by nets, Shing compared chickens to wild birds as different species to protect. He showed me the refuge where wild birds were healed after being caught in smuggling activities before being released. The Kadoorie Farm raised domesticated chickens as well as wild birds, but only wild birds were released, while chickens remained on the farm.

Thinking about bird release then led me to a different view of birds: not as sentinel devices signaling potentially pandemic viruses or as commodities in a process of carrying and care, but as potential ghosts in an unstable interaction. While I was working at the poultry farm, I constantly thought about the possibility of liberating chickens from their reduction to commodities, which led me to follow their transportation to the market, thinking about the tension between carrying and care. But when I heard about the release of wild birds by bird-watchers, I came to think about what it really means to liberate a bird from the livestock revolution. How can the gesture of releasing birds open new possibilities of interacting with birds, and what does it reveal of the logic of sentinels?

Bird-Watching in Sites of Bird Spirits

Taking the perspective of bird-watchers, I came to think differently about the work of virologists and poultry farmers when they monitor Hong Kong as a “reservoir” for avian influenza. An “avian reservoir” is not only a site of viral mutations under surveillance but also a set of diverse living beings who must be monitored and inventoried. Therefore, sentinels are not only technical devices sending warning signals but also collectives of humans and more-than-humans caring for their environment (Keck). Because they are sensitive to signals, they can detect invisible beings such as viruses or ghosts.

The Hong Kong Bird-Watching Society (HKBWS) gathers more than 1,000 members all over the territory, where more than 500 species can be observed in different environments such as the sea, the forest, or the wetlands. The major site for observation is the Mai Po Nature Reserve run by the World Wildlife Fund, where migratory birds roost in the wetland of the Pearl River Delta. Fishermen used to grow shrimps in the ponds using the sea tide as a regulator (a technique imported by immigrants from South China in the 1950s called gei wai 基圍), but now they are employed by the World Wildlife Fund to preserve these habitats for wild birds. With 10,000 migratory birds feeding in the marshes every year, it became a Wetland of International Importance under the Ramsar Convention in 1995. It shares the properties of the rice paddies—a mix of wild birds, water pests, and interested humans—but for the benefit of birds, not for human consumption.

Mai Po has thus become a model for the simulation and surveillance of avian influenza, within a colonial and postcolonial genealogy of natural reserves. It was a military territory at the time when the HKBWS was founded by British officers in 1953 on shared affinities between “nature lovers” and military soldiers (Tsing:133). While under the colonial rule, British officers looked at birds while preparing for the invasion of the Chinese army into Hong Kong, and postcolonial experts watched over dead birds as signs of a potential pandemic coming from China. The tipping point of the discordant event had changed (Khan), but the function of the delta as a sentinel post had been maintained.

In March 2004, the government closed the Mai Po reserve because a wild bird infected with H5N1 had been found within a 3-km radius of the premises. This decision was highly criticized as an excess of precaution, but it was repeated almost every year, with a 21-day ban imposed on the reserve when infected birds were found in its vicinity. Hong Kong bird-watchers argued that since wild birds found with H5N1 on the territory were resident species and not migratory, it was irrational to close Mai Po and not bird parks or bird markets in Kowloon, where urban visitors might be in contact with bird feathers or feces. Bird-watchers used ornithology and epidemiology to criticize the blaming of wild birds out of fear of avian influenza. They wanted to show that they managed and monitored the reserve in a more rational way than how the government handled avian influenza on the whole territory, thus becoming sentinels of environmental biodiversity rather than sanitary biosecurity.

Chiu Ying Lam was the first Chinese member of the HKBWS in 1976 and was elected chairman in 1997. Trained in astronomy, he was also the head of the Hong Kong Observatory. He made bird-watching a popular practice not just for the colonial military elite but for the new middle class discovering leisure and nature, following the model of Taiwanese bird-watchers (Weller). Chiu Ying Lam told me that he started bird-watching when he was young in the cemeteries of Happy Valley, where tombs are divided into Catholic, Protestant, Muslim, and Chinese areas in a dense urban neighborhood. “Cemeteries are good places for birds and ghosts. It was my first bird-watching, I was 27. Suddenly it was opening a door: I was seeing the living objects. It was opening my heart. Before I was studying physics and mathematics—dead things. I was blind. Our organs are opened but the signals are filtered out. I became an addict of bird-watching, I wanted other people to see, it was like preaching” (interview, Hong Kong, December 2008).

This narrative of eye-opening is interesting to connect to the perception of ghosts. In the Chinese tradition, ghosts are the spirits of the dead who cannot find a place in ancestral lines and receive sacrifice (Sterckx). The perception of birds on the borders of Hong Kong is associated with the death of refugees who tried to enter at the cost of their lives. Hong Kong can thus be compared to another site of bird-watching on the coast of Fujian, the island of Kinmen, used by Tchang Kai-Shek as a garrison to prevent the invasion of Taiwan by the People’s Liberation Army. Because the army has occupied this territory until the uplifting of the martial law, the landscape has been preserved from development, and bird-watchers come there from Taiwan and China to see birds that have disappeared elsewhere. There are many narratives about ghosts linked to the perception of birds in Kinmen (Szonyi). Watching birds on the seashore reminds local people of the soldiers who died while defending the territory. They are military sentinels in the sense that, while looking at birds, they perceive the ghosts of those who died in a conflict.

Chiu Ying Lam articulated this connection between birds and ghosts in a specific way. He said the Chinese were interested in birds as meat and the British in birds as species, but he was interested in what happens when he looks at a bird and a bird looks at him. Of course it is rare to be looked upon by a bird, but the gaze of the bird was often mentioned by Hong Kong bird-watchers when they showed me their photographs. While domestic poultry can be killed without looking back, wild birds send back the gaze as a sign of potential revenge. While poultry farmers exchange meat for care, bird-watchers exchange gaze for protection, but both kinds of interaction can turn to violence if the environment is not secured. Geographer Jared Diamond talks about the “lethal gift of livestock” to describe how animals bring pathogens to humans when they are domesticated too intensely; but this view remains too global and technological as it depends on virologists’ perspectives. If we take the perspectives of farmers and bird-watchers, we can rather say that the ghostly apparitions of wild birds in markets and farms remind humans of the uneven exchange of domestication.

Chiu Ying Lam came to this conclusion when he talked about bird release. Buddhist practitioners used to buy wild birds in the bird market and release them in adjacent parks to produce merits—a traditional practice in China coming from the release of birds by aristocrats to thank them for the quality of their songs (Handlin Smith). Bird-watchers showed that these birds often died from the release because they were kept in cages under stressful conditions and then released in an improper environment. Buddhist authorities launched a public campaign with posters showing a bird flying and gradually becoming a cadaver, thus transforming release life (fangsheng 放生) into release death (fangsi 放死). Bird-watchers working at the Kadoorie Farm suggested organizing ceremonies of bird release with wild birds captured on the border for smuggling, sheltered at the farm, and then released with GPS to follow their development. But Chiu Ying Lam noticed that these “scientific releases” were organized on Sunday and that Buddhist practitioners who were invited did not show up. He suggested organizing such releases in Buddhist or Taoist temples to meet the spiritual demands of ordinary people.

Bird release has become a highly debated topic in Hong Kong and Taiwan because it reveals the contradictions of the livestock revolution. If poultry becomes sick from industrial conditions of breeding and threatens to transmit deadly pathogens to humans, why not release birds from the market to diminish the suffering produced by the industry? Such was the reasoning of the ordinary Buddhist practitioners I met. But the Buddhist authorities, advised by bird-watchers, showed that releasing birds rather amplified the threatening invisible entities. Bird release produced viruses or merits depending on the carrying capacity of the lands where they were accumulated. When releasing a bird, one experiences the contradictions produced by the livestock revolution between birds as living beings and birds as commodities, since one does not know if the gesture of release will produce merits or viruses.

After the massive killings of chickens to eradicate avian influenza, Buddhist authorities used to pray for the souls of dead birds at the borders of the territory, because the territory needed to be cleaned of its bad energies. While biosecurity interventions appeared as a form of sacrifice that fixed living entities in the central market to clean the poultry industry from its impurities, watching wild birds on the borders of the territory was a way to interact with the ghosts of those who had been killed to preserve the territory. Birds observed on the borders of the Hong Kong territory reminded bird-watchers of the poultry killed to clean the territory from avian influenza and of the refugees killed in trying to join the British colony. While virologists tried to fix migrating birds through lines of viral descent, and while farmers tried to fix poultry through biosecurity measures, bird-watchers were attentive to the uncertain interaction with birds as potential ghosts escaping lines of descent and control.


This paper has multiplied the views of South China as a sentinel for avian influenza pandemics by taking different perspectives on the interactions between humans, birds and viruses and sentinel post, sentinel device, and sentinel territory. South China thus appears as a potential origin of the livestock revolution, which is one of the global indicators of the Anthropocene, and as a site of ghostly apparitions, which is the local perception of the threat of poultry farming. If the accumulation of livestock has produced an amplification of viruses but also a proliferation of ghosts, the gesture of bird release is an attempt to solve this contradiction by transforming livestock accumulation into lines of viral mutations or into personal intentions of ghosts. I have tried not to choose in this alternative between a secular and a nonsecular, or between a global and a local view of the livestock revolution, because I wanted to remain as close as possible to the uncertainties of life under the livestock revolution. Rather than ask how viruses mutate in an animal reservoir, I have asked how invisible entities appear in more-than-human interactions as signs of future threats.

Parallel to the logic of indicators, which follows the risks of the livestock revolution in a quantitative way, I have suggested a logic of sentinels to follow what happens on the thresholds of domestication. The Anthropocene has produced a logic of “too-much-ness” visible in the numbers of poultry raised in Hong Kong farms (Viveiros de Castro). It is a logic of confinement, concentration, and integration in which the increase in the quantity of poultry increases the risk of transmission of diseases. The logic of sentinels, by contrast, is attentive to the possibility that some living beings do not follow the general rule, which can be characterized as a logic of “some-ness.” A saying in the world of influenza experts is “Sick birds don’t fly,” arguing that the transmission of H5N1 from Asia to Europe goes along roads of poultry smuggling rather than migratory routes. This sentence is catching the imagination, and so is the Buddhist poster of released birds turning into cadavers. If some birds released for spiritual purposes are infected with flu viruses, this connection between spirits, viruses, and industrial chickens can be captured by sentinels because they are attentive to the possibilities of bird movements. By releasing birds into the wild, Buddhists and bird-watchers reveal the feral aspects of the livestock revolution in South China, thus exploring the threshold of domestication as a space of apparitions, frictions, and negotiations that can potentially be inhabited as a sentinel territory.