Floating World Lost





Eugene N. Anderson


Published by University Press of the South, New Orleans, LA, 2007

Copyright held by University Press of the South;

any reproduction of this work, or part of it, requires their written permission.



For my children:

Laura, Alan and Tamar, who were there

Amanda and Robert, whom I wish had been.




Acknowledgements:  I remain always deeply grateful to the people of Castle Peak Bay, especially Choi Kwok-Tai, Cecilia Choi, and Chow Hung-fai, and to Marja L. Anderson and to Laura, Alan and Tamar.  A long list of friends and scholars who contributed to this effort includes Hugh Baker, Stanley Bedlington, Paul Buell, William Chan (special thanks for identifying the fish), Louis Cheung, Wolfram Eberhard, George Foster, Hill Gates, Greg Guldin, Ho Kaak-yan, Dell Hymes, William Jankowiak, Graham and Betsy Johnson, Hiroaki Kani, Arthur Kleinman, Rance Lee, John McCoy, James and Helen McGough, Jack Potter, Edward Schafer, G. William Skinner, Ambrose Tse, Barbara Ward, James and Rubie Watson, Arthur Wolf, Yan Yunxiang, and others.  David Akers-Jones, James Hayes, David Robertson, and Major and Mrs. A. M. McFarlane provided much help in the field.  Derek Davies and Nick Ludlow helped with photography.  Last but not least, research was supported by the U. S. National Institute of Mental Health, the World Health Organisation, and the University of California.  This book draws on an earlier work, The Floating World of Castle Peak Bay, about half of the present book being a reprint of that effort, and I am grateful to the American Anthropological Association (which published the earlier book) for granting permission.  That book was dedicated “to the boat people and those who have studied them,” and I wish here to make special note of the late Barbara Ward, superb ethnographer and wonderful colleague.









Castle Peak Bay, in the Western Territories of Hong Kong, was once home to a teeming population of fishermen and cargo carriers who lived on their boats and rarely came on shore.  They made up a separate community, calling themselves seui seung yan, “people on the water.”  They, and the shore-dwellers around them, considered themselves to belong to a separate ethnic group, defined by residence on the water more than by language or culture.  They dwelt in boats, or in pile houses built over the water.  They were part of a vast array of boat-dwelling people who wandered widely over the rivers and seas of China.

Most human cultures are divided into subcultures, which may be defined by class, caste, occupation, or anything else important.  The boat way of life was based on a subculture was defined by ecology.  This book explores the structure of this subculture, and the practice of life by the people who were the bearers of it.  The subtitle Variations on Chinese Themes speaks to this focus.  I describe the ways in which boat life varied from life on land, and then explain what this all means for the study of Chinese culture.

Today, the boat community is no more.  The boat families moved on shore and merged with the general Cantonese population of Hong Kong.  The bay itself no longer exists.  It was filled in, providing level space for high-rise apartments.  Descendents of sea captains now live twenty floors above the rocky fill, and commute to jobs assembling electronics or programming computers.  The bay, formerly isolated, has become part of the vast urban agglomeration that is modern Hong Kong.

It has been the lot of most anthropologists to see their communities change.  Some communities even disappear.  Few ethnographers, however, have seen the disappearance of even the physical setting of their field experiences.

In 1965-66 and again in 1974-75, my family and I lived at Castle Peak Bay, studying the fishing community.  I return to Hong Kong every few years—most recently in 1999—and revisit the fate of the world I once shared.   Every year, more bays are filled, fewer boats go out.  The same transformations have now taken place in nearby parts of Guangdong Province.  Even more remote areas are seeing the fishermen come to shore, often encouraged or forced by government programs.

In Hong Kong as a whole, virtually no one still lives on the water.  Even by 1974, most had already gone on shore to work in local factories and shops.  Since then, fishing has succumbed to pollution and overharvesting.  Lightering—carrying cargo from large ships to shore—has disappeared with the construction of deepwater container port facilities.  A tiny group of people is still employed in fishing, local cargo carrying, and ferry work.  The floating restaurants of old are now tourist venues, piered to the sea floor and staffed by people of land origin.

This book, then, records a vanished way of life.  Boat-dwelling fishermen still number in the millions in China, southeast Asia, and elsewhere, but the boat people of the Hong Kong area had their own special ways of being human.  I became deeply involved in these.

A political scientist reading one part of this book dismissed it: why bother with a vanished way of life?  A student recently commented on one of my classes that the class wasted his time with discussions of “extinct, out-of-date societies.”  Such comments (even when phrased in a tactlessly biased manner, as in the latter case) require an answer.

The first part of the answer is that the boat people have much to teach us.  Their knowledge of fish and fishing was unsurpassed.  Their society was well-run in spite of lack of government and police.  Their social values and ethics were striking and original, and suggest to me important possibilities for the human future.  I have drawn at length on their social theory.  Their songs were beautiful, their food was exquisite, and their boats were superbly engineered.  All this they managed in the face of incredible poverty and uncertainty.

Simply knowing that humans could create so much beauty and happiness under such conditions is perhaps the most valuable lesson the boat people have to teach.  Any creation of the human spirit deserves serious consideration, simply because we are all human, and we need each other.  We need to know and understand.  Quite apart from the scientific merits of cross-cultural comparison, the human value of appreciating human achievement needs to be served.

Above all, the boat people recall to us the ancient virtue of humility.  The boat-dwellers created lives of beauty, glory, and satisfaction in the face of overwhelming challenges.  They “raised a Heaven in Hell’s despite.”  They deserve to be remembered.

I shall generally describe the bay as it was in 1965-66.  This will be a freeze-frame—a single picture in a long record of change, flux, and uncertainty.  I shall use information from 1974-75 only to supplement the picture—to flesh out the story, and sometimes to record the beginnings of basic change.

In 1965, Castle Peak Bay spread a flat sheet of shallow water over the mudflats and shallows that separated Castle Peak itself from the mainland.  The peak, with the ridge north of it, had formerly been an island.  Tang Dynasty travelers sailed through the channel between the island and the mainland.  It had shrunk to Castle Peak Bay in the south and Deep Bay in the north.  The rest of the channel had filled with alluvium, becoming some of the most fertile farmland in China.

The bay was fringed with small, traditional-style houses and shops, wherein lived fish-dealers, farmers, craftsmen, and merchants.  The town of Tuen Mun San Hui (“fortified strait’s new market”) at the head of the bay was already invaded by a few multistorey buildings, and there were unostentatious—but horribly polluting—factories there.  However, much of the town preserved the appearance and economic structure of a Qing Dynasty village.  A Buddhist college just north of it, and a major Taoist temple complex halfway up Castle Peak, provided ties with elite religious traditions.  Far more important to the local people were the countless temples and shrines that enshrined folk deities and folk conceptions of divinity.

The bay was home to several thousand boat people.  In 1965-66, most of these lived by small-scale artisanal fishing.    The larger boats were trawlers; medium-sized boats lived by long-line fishing, purse-seining, or gill-netting; the smallest practised local hook-and-line fishing.  The largest boats might hire a few hands, but almost all boats were strictly family operations.  A big boat might hold an extended family, while a small one was normally a nuclear-family operation.  Brothers who owned separate boats typically cooperated, operating what were in effect family fleets of two, three or four boats.

Major fish and shellfish caught included groupers, soles, pomfrets, croakers, sharks, shrimp, squid, and crabs (see Anderson 1972 for the whole list).  The inordinate fondness of Cantonese for sea food guaranteed high prices, at least for good quality products, and made fishing a highly lucrative occupation for the fortunate.  However, it was difficult and dangerous.  Nets and lines were still drawn by hand.  Many boats were nonmotorized; handling the gear and rigging was a major task.  The boats were extremely well-built and seaworthy, but managing them in a winter storm on the open sea required both great strength and great skill.

Moreover, a long spell of poor luck or stormy weather could wipe out a family.  Most people were in debt to fish-dealers.  The government had introduced cooperatives that worked well and eliminated the worst of the economic danger.  However, all older fishers could remember times when debt burdens had been crushing.  Indeed, in 1965-66, many families had been ruined, and had moored their boats for good and sought shore jobs.  Elders remembered World War II as the most difficult time in their lives; many people, especially children, had starved to death.

Difficult times and demanding work bred independence and resourcefulness.  Also, fishing, unlike at least some kinds of farming, requires major decisions on a day-to-day basis.  A boat captain (invariably also the father of the family) had to decide whether to put out in rough weather; where to go; how many times to cast the gear; when to pull in; and much else.  He has to make life-and-death decisions on a regular basis.  A captain could not stay in port on every rough day, since that would court economic suicide; but going out in clearly worsening weather was even more dangerous.   He would therefore consult all adults in the family.  They would help with any decision.  This meant that women were sharers not only of all work activities but also of all major decision-making, and gave them a degree of equality and power very different from that typical of the local land-dwellers.  This, in turn, led to major differences in family and child-rearing behavior (Anderson 1970b, 1992, 1998; Ward 1965, 1966, 1969).

Inevitably, families operated as independent units.  Lineages did not exist; they could not be maintained.  Leadership in the boat communities was informal and weakly developed (Anderson 1970b).  Individuals emerged as local leaders in so far as they could resolve disputes, deal with shore authorities, and organize community activities; such people enjoyed informal power in proportion to their success in these activities.  Life on the water was thus notably libertarian–in some contrast to the more tightly organized life on shore, controlled as it was by lineage and government.

These environmental and social factors lie behind many of the cultural and religious practices of the boat people.  They faced an uncertain and risky world, a scattered population organized largely at the family level, and a terribly hard and demanding life.  Common sense would lead one to expect that religion would address those issues, and that they would find solace in active, devout religious practice.  Such was indeed the case.



Over the last several years, I have been reanalysing data from this field work.         My interaction with the boat people and their world is somewhat significant to this analysis.  Though I am rarely excited by “reflexive” ethnography, most of which strikes me as narcissism, I have to establish some bare facts at the beginning.

In 1965-66, I was a graduate student in anthropology at the University of California, Berkeley.  I spent 14 months in the field.  My interest in all aspects of Chinese life went back to my parents, who had talked to missionaries and other travelers.  I wanted to look especially at fisheries development and at knowledge of fish and marine life—what we now call “traditional ecological knowledge.”  In the field, I learned most about fish, fishing methods, fisheries development, and the fishing business, but religion came in a close second.  Boat religion was fascinating to me.  Moreover, more than any other aspect of boat life, it challenged my rather simple paradigms of social structure.  Medicine and medical problems, food, and nutrition came into the picture too, and dominated later field work.

I went into the field with my wife, Marja Anderson, who worked with me on nutrition and medicine (Anderson and Anderson 1973, 1975, 1977; Ho, Anderson and Anderson 1978).  Thanks to her tireless energy, we were able to find out a considerable amount about these matters, and also about women’s lives, so often unstudied in those days.  The first year, we had with us our baby daughter Laura.  We all returned in 1974 with two more children, Alan and Tamar.  Everyone was delighted that we had had a son.  The boat people were about equally delighted with the girls, but the land-dwellers sometimes esteemed sons well above daughters in the scheme of things.  Everyone, however, adored all three of our children.  The children themselves were happier and healthier in the field than at home, a tribute to this caring environment as well as to outdoor living and good food.

I was not a totally green hand at field work, having done short field research projects in Panama, Tahiti, and California.  However, I was inexperienced, and had not been in East Asia before.  I made the usual mistakes, and had some minor adventures.  One of the more printable mistakes I made in Cantonese was asking the “name” of a fish.  The people I questioned looked at me as if I were insane, and I quickly learned that the word for “name” (Cantonese meng) applied only to personal given names.  I had to ask:  “What fish is that?”  Also, I started off wrong by using the textbook greeting “Are you well?” (nei hou ma in Cantonese, ni hao ma in Putonghua.)  Older people were offended.  I learned the greeting was a translation from English, and was offensive to traditionalists, since the Chinese implication was “Are you all right!?”—implying that you look sick.  I learned that the proper Cantonese greeting was “Have you eaten?” around mealtimes, or “Where are you going?” to people moving around.  This last had its own pitfalls.  We had to buy water at a standpipe well.  So, when asked the question, I at first sometimes answered “I’m going to buy water.”  Again I got The Look.  It turned out that one uses the two words “buy water” (maai seui) only when getting water to wash the corpse of a deceased family member!  Just one of those idioms….

Our living conditions were not easy.  In 1965-66, after an initial pleasant but detached period in a house on land, we moved down into the midst of things.  We lived on a sampan, a small residential boat.  We never went far; the boat usually remained moored to a fish dealer’s floating business.  The fish dealer and his family became fast friends and great helpers.  In 1974-75 we lived on shore.

Sam paan means “three boards,” and a sampan is a boat built up three boards from the waterline—in other words, a very small boat indeed.  Ours was 15’ long and about 7 wide, with a cabin not quite 5’ high.  It was by no means the smallest inhabited boat—far from it.  We slept, dressed, and wrote in the cabin, cooked and laundered on the deck, but otherwise spent most of our time off the boat.  We bought food every day at the local markets, and cooked for ourselves on the boat’s built-up stern.

Both practicality and personal taste made us live strictly local style, learning from those around us how to cook, dress, wash, and care for the boat.  Marja became a superb Chinese cook.  We took baths in a basin in the cabin, heating the water over the little kerosene stove on which we cooked.  We had a tape recorder, useful for field work but also useful for playing music from home.  Marja soon adopted the loose black cotton clothes of the boat women, which proved the only garments comfortable and tough enough to be serviceable on the boat.  Little Laura put up with it all, and reveled in the constant contact with us and our neighbors.  By the time we left, her Cantonese was better than her English, but she forgot it all when we left the field.

In general, life on the boat was comfortable, and the bay was so beautiful and fascinating that we rarely found the time wearing except when we were sick.  However, we had our problems.  We were living on limited means.  We were not fluent in Cantonese.  During prolonged rains, the boat was cramped and uncomfortable, and staying clean became difficult.  June 1966 brought history-making rains and floods, and we looked and felt like drowned rats.  Typhoons passed through; one in fall of 1965 drove us to a typhoon harbor and kept us awake all night.  I had to hold down the heavy cabin hatch and other portables.  We left the boat in July 1966, just in time to avoid another storm.

Such minor problems aside, we had successful field sessions at Castle Peak Bay.  This was thanks largely to our wonderful field assistants, Choi Kwok-Tai (and in 1974-5 his daughter Cecilia), and Chau Hung-fai.  They were always ready for anything, incredibly tough under field conditions, unfailingly helpful, and full of knowledge of boat and shore ways.  The boat people themselves were hospitable and supportive, and the few who tried, rather naively and innocently, to cheat me provided more humor than trouble.  (Minor cheating was a game in Hong Kong, and we would all have a good laugh over it, whether I won or lost.)  Some land people were not so friendly, since I was identified with the despised water folk, but unfriendliness never went beyond an occasional dirty look or refusal of minor help.  In general, we were well accepted—a fact made clear by the rapid fall in prices charged to us in the markets; we were paying local-customer prices after the first few months.

The food at Castle Peak Bay was truly memorable.  In those days, the waters were clean, and rich in fish and shellfish.  Fish were kept alive in well-smacks—small old rafts with the bottoms knocked out and replaced with wire.  These floated along the shore of the bay.  Diners—most of them city folk who came out on weekends to enjoy the seafood—would shop around, buy the fish they wanted, and take it at a run to the nearest good restaurant.  That seafood was incredible.  I have never tasted anything so good since, except among the Native Americans on the Northwest Coast, who are equally blessed with a combination of superb fish and superb cooking skills.  I have since eaten in some fine restaurants in Paris, Rome, and Los Angeles, but they could not touch Muk Choi’s Garden of Delight or any of a number of other places at Castle Peak.

In 1974-75, we rented a house from an old friend, a boat person who had moved on shore some years earlier.  It was a humble structure of boards and stone, informally built.  It consisted of a downstairs room and a sleeping loft.  Almost all our living was done on the covered patio.  Here we set up our tiny kerosene stove, washed food and clothes, and did most work.  We were viewed as returning locals, and welcomed accordingly.  We were out for a shorter time, and I spent a fair amount of it doing library research or research in other settlements, so the field situation was not quite as of old; but we had close friendships all the same.

A veteran field worker, or a more industrious one, would have gotten vastly more data than I did.  Still, we learned much, and we had a wonderful time.



Then as now, I was primarily interested in human ecology:  how people relate to their physical and cultural environments.  I am primarily interested in how people understand these environments.  Thus, in Hong Kong, I studied everything from catching methods and names for fish to the relationship of boat-dwelling and boat size to kinship and family structure.  I had to know something about cognitive psychology, and a good deal about wider social relationships, including relationships of power.

I was becoming involved in what later came to be called political ecology (Robbins 2004).  Basically, political ecology is the study of how the above-mentioned relationships of people and environment are structured and influenced by relations of power—by politics.  My research was largely on knowledge—on cognition and its relationship to practice—but a good deal of political ecology emerges in the present book.  Andrew Vayda and Bradley Walters (1999) have criticized political ecology for tending to be too much politics and too little ecology; I take a holistic approach, trying to look at economics, biology, politics, psychology, religion, and other anthropological categories without prejudging which one is most important.  Perhaps, in the end, they are not separable in communities like Castle Peak Bay.  One must simply look at how people get along.

And one must look at culture, as adaptive mechanism and as expressive form.

Anthropologists agree that culture is that which is learned and shared in groups.  As such, it is shared behavior.  In so far as humans can share knowledge, it is shared knowledge; it is the set of facts, ideas, and rules that people have to have to operate in their society and to be accepted as functioning members of it.

Beyond that minimalist definition, anthropologists have seen culture as many things.  Culture has been portrayed as a ragbag of traits, “a thing of shreds and patches” (Lowie 1947:441, quoting Gilbert and Sullivan’s Mikado; he later described the remark as a sour comment on modern civilization, not a generality about culture—Lowie 1947:x).  Culture has been regarded as an ironclad set of unchanging rules (White 1949).  Others see it as consensus, universally shared knowledge.  Alternatively, ecological anthropologists often see culture as adaptation (McElroy and Townsend 1996).  Others describe culture as a set of rules or structures that allow people to order their lives and anticipate what others will do (Lévi-Strauss 1958; Frake 1980).  Many of us now see culture as general patterns or apparent regularities arising in and from practice (Bourdieu 1977, 1990).

Anthropologists have also differed as to what constitute the wellsprings of culture.  Marvin Harris held that culture was entirely determined by the exigencies of making a living (Harris 1968).  Boas saw culture as mainly about communication, thus focally a matter of language, art, literature, and communicative ritual.  (He never quite said this, but it emerges clearly from studying his works.)   Clifford Geertz (1973) sees culture as about constructing and sharing interpretive meanings.  Marxists, functionalists, and other theorists all have their views.

Commonly, today, culture is interpreted rather than simply recorded.  Geertz and his followers see culture as a text or series of texts whose meanings must be understood (Geertz 1973).  For them, one cannot merely list “culture traits” or cultural knowledge points.  It is the web of meanings that matters.

For the boat people, the first three theories in the list above—ragbag, ironclad, and consensus—will not stand.  Boat culture was no ragbag; it was well integrated around fishing and around Chinese traditions.  It was not ironclad; rather, it was highly variable, each family posessing its own ways and truths.  Finally, consensus would have been disastrous, since life depended on having available a cultural knowledge base that was far greater than that which one person could possibly learn, and also on every captain (i.e., normally, every head of family) being able to sail—independently—his or her ship. Some individuals were knowledgeable or expert in one thing, some in another, and the community drew on these different sorts of expertise.

The last four theories, however, all apply.  Culture was most definitely a web of shared meanings, and interpretation is necessary if one is to make sense of life on the waterfront.  Culture was adaptive, as proved by its constant reconfiguration to meet economic and environmental circumstances.  Culture included sets of rules and expectations that allowed one to behave appropriately and to anticipate what others would do.  And it most definitely guided and emerged from practice.  It could not be reduced to that set of rules; it was renegotiated daily.

Charles Frake (1980) called for an ethnography that would allow one to behave appropriately in the society in question—or at least to anticipate what others would do and would think appropriate.  His model was linguistic; he was thinking of a good language textbook, which should teach the student to use the language correctly and more or less fluently, in order to get food and shelter and friendship.

An ethnologist may need more than that from an ethnography, but certainly the ethnologist will need at least that much. If one cannot use the ethnography as a guide to how to get along in the society, or at least what to expect in the segment of society analyzed, what value can it have?  A specialized ethnography may be a guide to only one specific aspect of life (kinship, fishing, ritual…), but at least it should be a guide to something.  The result can be minimalist—a brief ethnography, an introductory language book—or detailed and advanced.

Most anthropologists hope that an ethnography will do more; ideally, it will provide insights into the human condition, enabling us to advance the general theory of society.  Cross-cultural comparison and resultant theory-builiding (Harris 1968, arguing against Frake) needs to be done, to allow us to understand the deeper roots of behavior and extracting the “story behind the story.”  It may also be desirable to critique neocolonialism or globalization or other wider matters.  Some ethnographers simply aspire to produce literary work (presumably for entertainment; Clifford and Marcus 1986).  Others, “reflexive” ethnographers, write out their own angst with the “ethnographic subjects” there only for backdrop.  These last two goals seem beyond the pale of descriptive ethnography.  The other goals are more relevant here.

Returning from the field, I wrote a simple Frakian ethnography, one that would give the rules for behavior. It would let the reader act like a boat person, or at least would let the reader know what was going on if she was watching the excitement, confusion, and activity of a Hong Kong waterfront.  No one was interested in publishing such a vast creation, or even the short and summary version of it that eventually resulted.  The American Anthropological Association did eventually make that summary available (Anderson 1970b).  The present book, however, does more and less; it adds wider and more theoretical considerations, but does not tell the reader all.  There is no sense learning the rules for a world that no longer exists.

There are, however, reasons to continue to work within a broadly Frakian mode.  Culture is practice, but not all practice is culture.  The same can be said for meaning.   Everything we notice is meaningful for us in some way, and thus saying that culture is “meaning” is not really saying anything.  Culture is adaptation, but not every adaptive behavior is cultural.  Moreover, not every culture trait is adaptive, except in the trivial sense of marking the bearer as “one of us.”  Millions of rules for ordinary behavior are purely arbitrary—even if they were adaptive once.  In the western world, sleeve buttons on coats were once for glove attachment, but this has been forgotten for so long that an urban legend has arisen to the effect that they are there to discourage the young from wiping their noses on their coat sleeves.  Chinese culture has similar “vestiges” (and similar legends to explain them!).

Frake’s approach seems the best at capturing what culture really is and really does.  Culture is the knowledge that allows us to do what we have to do, should do, or should want to do, within our reference groups.  It is the knowledge that allows us to understand what others are doing in those groups—and why we cannot always do the same things that they can do.  A good deal of cultural knowledge is not clearly or directly related to this goal, but it can still be regarded the background knowledge needed to allow us to perform.  The boat people’s fish taxonomy, for instance, not only allowed them to identify and talk about fish, it also allowed them to deal with markets and marketing.  Culture is not just knowledge or meaning; it is knowledge that is useful, or at least could be useful in the right context.

This approach allows us to speculate on how knowledge, rules, and ethics enter a culture’s repertoire, and get stored and encoded there.

In my early field work, in spite of considerable awareness of history and process, I followed the tradition of synchronic structural-functional analysis in which I had been trained.  Synchronic analysis focuses on one point in time, abstracting it from ongoing processes.  Structural-functional analysis seeks to discover social structures, and if possible cognitive structures too.  It then analyzes them in terms of the functions they perform within the society being studied.   This, of course, assumes that there are structures—social institutions that can be neatly described and mapped, like crystal patterns.  Individual dynamics, emotion, dialogue, argument, negotiation, conflict, and creativity—all the exciting actions that makes human affairs so absorbing—tend to be dismissed as “noise” in the system.  Indeed, it is “noise,” but it is what the Chinese call “heat-and-noise” (re nao)—the stimulating intensity of a healthy social scene.  It is not what electronic engineers call “white noise.”

Hewing to a structural-functional line led me to produce a rather thin, even skeletal, account of the social structure of the Castle Peak Bay community.  It was not that I was unaware of the other issues; I simply did not have a good theoretical language to talk about them.  I made use of history in my analyses, but not as much as I might have done.  I also remarked on the free-wheeling, individualist quality of boat life (as had all other writers on them), but I had no way to work it into an analysis.

Since then, like many of my generation, I have fallen under the influence of Pierre Bourdieu (1977, 1990), Barbara Rogoff and Jean Lave (Rogoff and Lave 1984) and other theorists of practice, dialogue, and negotiation.  For Bourdieu, in particular, structure emerges from interacting individuals’ practice. Kabyle Berber culture as described by Pierre Bourdieu (Bourdieu 1977) is very different from Castle Peak’s in many ways, but Bourdieu’s description of the ways that culture kept re-emerging from ongoing political and social game-playing is a perfect description of life on the Castle Peak waterfront.

Anthony Giddens’ concept of “structuration” (Giddens 1984) has also provided an extremely helpful guide to my thinking.  As all these scholars point out, human beings interact and engage in political (and other social) activity in pursuit of individual and family goals.  Overarching social institutions set boundaries but are themselves constantly being changed and reconstructed.  Instead of a smoothly running machine that magically keeps itself perfectly tuned (the structural-functional fallacy), society is a shifting, constantly varying “field” (Bourdieu’s term) in which individuals seek to maximize their welfare and success within the context they perceive.

Students of Chinese religion have been quick to apply practice models (often without explicitly saying so—e.g. Dean 1993).  Several recent studies of gifts and gift exchange (Yan 1996; Yang 1994), courtship (Jankowiak 1993), religion (Weller 1987), and other social institutions have incorporated practice theory, though not always making a point of it.

In particular, ethnicity has been analyzed by countless recent writers, almost all of whom contrast the shifting, flowing nature of ethnicity-in-practice with the rigid, essentialized categories beloved by bureaucrats, racists, and many earlier-day anthropologists.

I have thus re-examined boat-dwellers’ social patterns with this in mind (Anderson 1992, 1998; Anderson, Wong and Thomas 2000).

One problem with practice theory as presently found is that it is often thin on motivation.  Writers like Bourdieu and de Certeau were interested in establishing the approach.  Bourdieu’s motivational theory was simplistic; he seems to have felt that people act only to get wealth and power.  At least, he explained artistic taste, academic life, cultural reproduction, and similar phenomena from that point of view.  This gave his work a narrowly functionalist quality; institutions served the interests of power.  Art, for instance, was about snobbery, not about aesthetic experience (see the brilliant, highly critical review by Jon Elster, 1981).  Other practice theorists have simply not addressed the question; they seem generally to be interested in describing the practice, not in constructing broader theories of what lies behind it.

This does not give a satisfactory account of life on the water, where real-world concerns were immediate and often desperate.  One had to fish to survive, and one often faced life-or-death choices.  Motivation could not be ignored.  On a more broad theoretical level, practice is not done in a vacuum, and it is not always about power or wealth.  People want many things—social connectedness and security above all, and perhaps, now and then, a bit of fun as well.

One must thus draw on the literature describing culture as adaptation (see e.g. the review by Bates, 2004).  Humans need security first of all.  They must feel safe, and able to defend their group.  They must be able to make a living—to be reasonably sure they can get food without being killed in the process.  Once this is assured, they can go about the business of eating, sleeping, enjoying themselves, and dealing with everyday conflicts and problems.  But, also, to have an ongoing society, they must reproduce.  This involves far more than sex; children need care for many years.

Culture provides a storehouse of ways for doing all this.  It provides alternatives; humans are not locked into instinctive patterns.  Above and beyond that, culture typically integrates livelihood, group defense, emotional coping, and social rules into one system.  Usually, the system is loose and flexible, able to change fast when change is necessary.  This openness and flexibility is important.  The Chinese say that the flexible bamboo survives the storm that breaks the greatest pine.

Humans are creatures of emotion, and thus one cannot understand culture in purely rational terms (Milton 2002).  It is, among other things, a way of dealing with emotions.  Ideally, it can bring forth emotional harmony and satisfaction while providing ways of dealing with emotional clashes, tensions, and problems.  In the real world, where such tensions continually arise and shift, no culture can do more than provide some alternate possibilities for dealing with them.  No group reaches Émile Durkheim’s ideal of society as emotional union (Durkheim 1995).

Thus, any strong pressure from economic, military, emotional, or social factors can dynamically deform a cultural system and its social and cognitive structures.  This has happened to the boat people, as to others in this world.  Their history is a history of adapting to continual change.


Another thing one could learn on the Hong Kong waterfront is that culture is located in individuals.  Contra the classic Leslie White (1949) position, there was no ideal Culture floating in space (mental or otherwise).  As so often, Max Weber said it best:  “Interpretive sociology considers the individual and his action as the basic unit, as its ‘atom’…. [T]he individual is also the upper limit and the sole carrier of meaningful conduct….  [S]uch concepts as ‘state,’ ‘association,’ ‘feudalism,’ and the like, designate certain categories of human interaction.  Hence it is the task of sociology to reduce these concepts to ‘understandable’ action, that is, without exception, to the actions of participating individual men” (Gerth and Mills 1946:55; “men” is used, of course, in the old gender-neutral way, to mean “people”).

Each individual had his or her own take on Cantonese boat ways, was aware of it, and could get defensive about it when challenged.  Culture was shared understandings, shared meanings, shared rules—but the sharing was never 100%.  No one knew all the rules, and no one had quite the same knowledge base or set of opinions.  The result was something of a happy anarchy.  Sometimes the anarchy was not so happy, but always it defended as preferable to a tight rule-based social order of the sort attributed (often with exaggerated, negative comments!) to the land-dwelling Punti.  Some boat people were less anarchist, and wistfully envious of the good order on shore, but they were resigned to their society’s freewheeling code.  Above all, they were aware that it was different from other codes found nearby.  Those who genuinely wanted a more “grounded” order left the boats and moved on shore for good.

Yet the realities of boat-dwelling enforced some conformity.  A good simple example of rules and practice is provided by clothing.  The boat  people traditionally wore black cotton pants and shirts, and a basketry hat shaped like a mushroom (and often called a “mushroom hat”).  Marja Anderson quickly adopted this style, for the very good reason that no other available women’s clothing would allow easy movement and stand  the hard usage of life and laundry on the water.  I did not, because ordinary California male work clothes did the job well.  The vast majority of the boat people, including essentially all the women who actually lived on board, dressed in black cottons, but a large percentage of the men dressed in ordinary western-style work clothes, as I did.  (In those days, Hong Kong made a good deal of the world’s clothing, and cheap factory seconds in western styles were always available.)  Younger boat women who did not sail the open sea often dressed in western-style clothing.  The traditional outfit was adaptive, and also a culture marker; when it was not adaptive, it tended to get sidelined, and culture was left unmarked.  This nonmarking itself could be adaptive, when one wanted to avoid the stigma of being a “boat person.”

With all this play between adaptive conformity and preferred independence, the community was tight, most knowledge was shared, and people seemed strikingly uniform to the outsider.  The differences between people and families showed up in thousands of little ways, but in no one big way.  One could make generalizations about clothing, food, kinship, morals, fishing methods, and everything else, and these generalizations would be reasonably adequate as Frakian rules. They would miss something vitally important: the role of individualism and adaptability.  But they would let a raw outsider get along on the boats without too much trouble.  We know this is so, because we did it; participant observation works!  Many land people had moved to the boats in living memory, and did a better job than we did.

Culture is not about uniformity or about set rules.  It is about making individuals (within a reference group) act similarly enough that they can get along and get their jobs done.  It comes from the need of people to work and live together.

It works partly because of the broadly imitative tendency of the human animal.  People (and other primates) like to “ape” each other. In particular, humans like to imitate their immediate superiors—their parents, older siblings, and slightly older neighbors, to say nothing of movie stars and other high-status strangers.  No doubt this tendency evolved to make mutual aid more successful in primate groups, but today it keeps much of culture alive.

Yet one cannot make too much of this, and it certainly cannot explain cultural differences.  The specifics of a culture can be explained only through examining the history of the social group that bears it, especially their adaptive needs.  They have had to face threats and dangers, and to take advantage of economic and political opportunities.  How they deal with these has everything to do with the final outcome: the pragmatic rules for acting in the everyday real world (de Certeau 1984).  Anthropologists take a holistic view, and I certainly follow that practice, but “holism” does not mean that culture is seen as some sort of harmonious whole.  To look at the whole means to look at a shifting, diverse, protean pattern; one feels like William James’ newborn baby, confronted with a “blooming, buzzing confusion.”  Imposing a bland uniformity on it would be a hegemonic act (as Gramsci saw), and no boat person would do it.

However, culture is also about maintaining solidarity in the face of threat.  The boat people had to be uniform enough to present a fairly solid front to the hostile shore-dwellers.   In earlier years, they had had to resist much worse threats, climaxing in the Japanese occupation in World War II.  This defensive aspect forced on the boat world a degree of organization and uniformity that it would not otherwise have had.

Conversely, the needs of boat life kept culture flexible and individual.  Boats and fates were different.  Every head of household was a captain whose crew was his (or, occasionally, her) family.  Every day, in this region of storms, disasters, and human troubles, their lives depended on the captain’s decisions.  Every day, livelihood was at stake.  Life was chancy in a way it was not for even the most risk-taking of land people.  Even in the United States, sea fishing is a close second to mining, and far ahead of all other occupations, in physical danger.  The boat people had no radios, no radar, no phones on their fragile wooden boats.  They had to be able to make their own decisions, quickly, accurately, and without much chance for second thoughts.  They had to run their own ships.              For this, culture was the great guide—the repository of knowledge. But it had to be a storehouse of all sorts of knowledge, including all manner of alternative plans and planning methods. It could not be a bank of cut-and-dried rules.


Cognitive anthropologists today discuss such issues in terms of “cultural models.”  Throughout the history of anthropology, scholars have proposed knowledge structures of this kind.  The tradition goes back to the original “anthropology” book, Immanuel Kant’s Anthropology from a Pragmatic Point of View (1976/1796).  Kant saw anthropology in terms of what we now call social cognition.  (In fact, social cognitive psychology, like anthropology, stems from Kant’s work.)  He developed phenomenology, the study of the ways people perceive and experience the world and then organize it in their thinking.  Anthropologists and ethnologists, as well as psychologists, sociologists, and philosophers, subsequently developed more and more complex models of thought, based on Kantian ideas.

Much of the modern concern with such systems can be traced to the thoroughly Kantian writings of Claude Lévi-Strauss, especially the essays collected in his book Structural Anthropology (1958).  Lévi-Strauss differentiated the “conscious models” held by the people themselves from the models the ethnographer makes of their thought.  These were highly useful concepts, applied to the boat people by Barbara Ward (1965, 1966).

For Lévi-Strauss, as for most writers, kinship systems have been the archetypal models.  Kinship systems are widely shared and widely understood within a culture.  Every English speaker, even young children and recent immigrants, knows what “father,” “mother,” “daughter,” and the rest of the common terms mean.  They know that “mother” means “female parent”—the woman in one’s lineal family who is one generation above oneself.  They also know that “mother” is supposed to be loving, nurturant, caring, and so forth.  The conscious models tend to run heavily to these role expectations and emotional cues, while outsiders’ models (especially the statistical models that Lévi-Strauss described) tend to be more formal, cut-and-dried, and rigid.  They are often more formal and neat than any conscious model would be.  As Lévi-Strauss pointed out, this is not a weakness; outsiders’ models are not supposed to reproduce the whole messy reality, but to do a particular job.  Often, they are made to facilitate quick and easy comparison between languages or cultures.  They represent reality the way an airplane model represents a real airplane.

This means that outsiders’ models (and insiders’ too, for that matter) differ, according to what one wants from them.  A child’s toy airplane and a Boeing engineer’s test model can be equally good models—in that they serve their purposes equally well.

Barbara Ward pointed out that there are many “varieties of the conscious model”—each ethnic group had its own set (including unflattering stereotypes of the other ethnic groups!), and even each individual had his or hers.  From this diversity of models, the ethnographer must come up with an overall model that works for outsiders.

The outsider may wish to understand the other culture deeply, or may wish only to know how to say “hello” and ask for food.  There are, therefore, different outsiders’ models too.

Subsequent writers on cultural models have tried to define widely-held, widely-shared structures.  Kinship systems and other semantic systems remain the archetypes (see Kronenfeld 1996).  However, everything from “scripts” for entering an American middle-class restaurant (Schank and Abelson 1977) to love and marriage (Strauss and Quinn 1997) is fair game (David Kronenfeld 1996 and work in progress).  My own work has been largely on classification systems for plants and animals, and indeed the Ph.D. thesis resulting from my Hong Kong work was a study of naming systems for fish and other sea life (Anderson 1972).

Models may be highly structured and widely-shared domains like kinship systems.  They may be areas where the vast web of connections in the brain are tighter and closer—where one idea typically calls up a particular set of others (Naomi Quinn, personal communication; see also Strauss and Quinn 1997).  A model may be a set of rules for a particular bit of behavior; it may be a particular canonical story-line or program.

The farther such models get from kinship, the more heavy water they encounter—to use a marine metaphor.

Restaurant scripts vary much more.  Schank and Abelson’s pioneer study vastly underestimated the need for improvisation and local adaptation in restaurant scripts.  There are quite different ways of entering different kinds of American restaurants; fast-food-joint etiquette is quite different from that proper to a pizza parlor, a white-tablecloth gastronomy palace, or a buffet.  But at least we can still specify these.  One could write rules for all of them, such that a foreigner reading the rules could act appropriately.

Love and marriage are different matters.  There are, indeed, broad and vague expectations that most Americans share.  My student Kimberly Hedrick (unpublished data) has found that a wide variety of ordinary Americans were familiar with the standard stories—the age-old fables once known from folktales, now from Hollywood plots.  But they did not expect any real love affair to follow those scenarios.  They referred back to the stories for orientation and sometimes for understanding, but they knew that no one really lives that way, and that every love affair is a new, unique, surprising phenomenon.

Are there, then, cultural models for love and marriage?  Yes, but they are not at all the same kind of model as that we see in kinship systems.   Knowledge of English kinship allows one to label properly, every time, everybody to whom one is related.  Knowledge of love and romance clichés does not give that much advice.  They are mere skeletons of plots—sometimes suggestive, but no more than that.  Even a wider knowledge of love and romance—knowing, for instance, all the self-help literature on “communication,” “talking about feelings,” and so forth—does not help much with the wild, woolly, and thoroughly unique ride that is a real love affair.  One has to improvise.  There are cultural models telling musicians how to improvise—every jazz musician knows a whole set of them—but a lover is often on her own.  And, even in jazz, is a model-for-improvising really like a model for how to label kinfolk?

At this point, we have come full circle, and are back to practice.  Bourdieu, a sometime student of Lévi-Strauss, reacted strongly against the latter’s structuralism.  For the Kabyle Berbers, the house had to be structured in a quite specific and particular way, but no two houses looked alike.  Politics—including kinship politics—was even more improvisational.  People negotiated.  They developed Machiavellian tricks.  They created new and fluid ways of doing business.  Kabyle Berbers—and Hong Kong boat people—treated the informal political rules (yes, there were rules) as obstacles to be avoided or finessed.

If one wishes to use kinterms or fish names appropriately, one needs a simple, cut-and-dried model.  However, if one is trying to deal with an actual living kinsperson, manage a love affair, or play politics, one needs a model that specifies not only the rules (such as they are) but also the standard ways around them.  Not only that; one needs to know a great deal about the methods in which all the above are deployed, strategized, and adapted to particular situations.  To keep with the musical parallel, we have gone from “Twinkle, Twinkle, Little Star” to Dizzy Gillespie.

Cultural models thus blend insensibly into the whole web of thinking, planning, and acting.  No one can specify the point at which a bunch of ideas is organized enough to be a “model” rather than mere typical human woolgathering.  One can say only that some realms of human action are tightly organized around very well-structured sets of knowledge, while other realms must be the domain of improvisation, negotiation, and continual re-creation.  (The resemblance of “re-creation” to “recreation” is one that an English-speaking boat person would truly appreciate!)

Cultural models could include the various knowledge structures we all know:  classification systems, plans, strategies, goals, beliefs, vague ideas, tendencies, poems, songs, stories, routines, repertoires, dance patterns, canonical scripts, and so on.

Anthony Giddens (1984) coined the word structuration for the process by which negotiation of this sort leads to new structures.  Douglass North provided a broadly similar—but very differently phrased—account of how economic institutions come into being (North 1990).  Giddens sees social interaction as key.  North focuses on people’s efforts to reduce the costs (monetary or otherwise) of interaction by creating institutions that make it easier.  The boat people did not create any dramatically new social institutions, but they did create, and constantly rebuild, cultural institutions ranging from fishing methods to song styles.  Thus the cultural models were always changing, as the result of agency (Giddens 1984) taking shape in practice.

Obviously, I am creating observer’s models, though I try my best to stay close to the conscious models of my friends (see, again, Ward’s wonderful treatment of the relevant problems).  Unlike many interpretive anthropologists, I do not hope to get inside their heads.  However, unlike the more extreme postmodernists, I do not believe that all ethnography is fiction, or that cultures are hermetically sealed from each other.  I know I can still enter a boat community in southeast China and act appropriately; if I brush up my Cantonese, I can sit at a feast and talk about fish and fishing and the evils of government.

That may be a small accomplishment, but it is a real one.  It tells us that people can understand each other reasonably well, even across enormous cultural walls.  Perfect understanding is not a possibility; but, then, perfect understanding of anything—even of our own selves—is not for this world.


One example of cultural models in practice, and one very salient to the boat people, was response to oppression.  The boat people were near the bottom of Cantonese society.  This, in turn, was dominated in Hong Kong in the 1960s and 1970s by the British colonial system.  It is not too much to speak of the “British raj,” since many British colonial administrators were reposted to Hong Kong after India and other British colonies in Asia became independent, and many South Asians followed them as soldiers or tradespeople.

The Cantonese were not exactly under the yoke of colonial oppression—they had a great deal of freedom and opportunity—but they were not the rulers of their land, nor did they have much power to make policy.  The boat people, being the lowest-ranked Cantonese in the social hierarchy, had especially little opportunity in that regard.  On the other hand, the British system protected them from the worst of the oppression they had known under Chinese governments; at least that is what I heard from all parties.

In any case, they had need of the full panoply of “weapons of the weak” and “arts of resistance” (Scott 1985, 1998).

These did not exclude the use of force.  Two old cannons lay in tidewater by our boat.  When I asked about them, I was told:  “In the old days we used cannons to defend ourselves from pirates and the government.”  The equation of those two categories was noteworthy.  All older people had tales of defense through force of arms.  Pirates were the worst and most constant problem, but Nationalist government vessels in the old days acted like the pirates, as far as the boat people were concerned.  Then came the Japanese war, when the boat people had served heroically in resistance.  The Communists who then took over most of China were sometimes supported, sometimes opposed.  Life in Hong Kong had been peaceful since order was restored after World War II, but no one really expected this to last forever.  The violent and unsettled times of old had given the boat people a fatalistic expectation of change and upheaval.

The boat people resisted prejudice and oppression by retreating into their water world.  Barbara Ward has discussed at length their common self-description “dragons on water, worms on shore” (Ward 1965).  On land, the boat people were exposed to humiliation, abuse, and exploitation.  Not many years earlier, they risked physical attack, and even in 1966 most older women and younger children would not go more than a few feet from the water’s edge.  On the water, however, the boat people were the masters.  Their consummate knowledge of seafaring and the water world made them independent and powerful.  Any land-dweller who ventured onto the water was at the mercy of the boat people, on whom he or she had perforce to depend.

Also, the boat people were essential to the waterfront merchants, who lived either by buying fish or by selling goods to the boat-dwellers.  This did not stop abuse and insults, but it kept those unsavory features of life down to a minimum.  Also, the merchants had to help in defending the boat people against serious threats.  There was no love lost on either side between the boat people and the fish merchants, but both sides knew they depended on each other.

As James Scott would expect, resistance to prejudice was normally shown through a constant background noise of criticism and restatement of core beliefs.  Usually, this consisted of a stream of sour remarks about land people and self-praise of boat people.   Most commonly, land people were said to be mean, stingy, prejudiced, unfair, and rigid.  Boat people, in contrast, were honest, open, fair, reliable, and independent.  These particular stereotypes are interesting, both because they reveal boat values, and because they contrast with the boat people’s stereotypes of other ethnic groups (see Anderson 1967 for wider discussion of ethnic stereotyping in Hong Kong in the 1960s).

Rarely, however, were these biased statements worked up into actual “arts of resistance.”  The boat people’s rich folk song tradition included only few lines on the issue.  Proverbs and verbal art also stayed close to broader ideals.  In fact, the boat people were rather idealistic and general in their verbal art; bitter commentary was reserved for ordinary conversation.

This was, at least in part, because of fear.  Until recently, they could have been in serious trouble if land people had overheard negative songs or proverbs.  Like the blues of African-Americans in the early 20th century, boat folk songs sometimes had a covert critical agenda that took some digging to ferret out.

More serious was the need to make a living.  Fishing is, at best, a chancy occupation.  Available alternatives were few, and usually not much of an improvement; cargo-carrying, passenger ferrying, boat vending, and smuggling all had their own insecurities.  The land people had the boat people in economic thrall, as will be seen.  The boat people had to protect themselves economically as best they could.  Whatever subsistence they got had to be gained from, and on, the water.

The theme of resistance occurs throughout this book, though not always stressed.  The boat people’s senses of self, their psychological integrity, depended on maintaining their sense of independence on the water.  This led to serious problems for those who had to move on shore.  They coped by swiftly assimilating to Punti society, but during the transition they often had troubled lives.

Thus the real defense of the boat people—not only physical self-protection, but defense of self-image and self-esteem—was retreat to the boats (or the pile-house communities built over tidewater).  Here, the boat people were in their own world, a world they understood and dominated.  It was closed to outsiders, and had its own ways and celebrations.  In it, they could ignore or cope with the outside.


Working with these perspectives allows me to do what I have always dreamed of doing:  Actually using the boat people’s own ideas and theories in my analysis.

However much anthropologists have been willing to learn from their subjects, they have rarely taken seriously their subjects’ views on society and culture.  There is a long roll of honorable exceptions, going back to early days.  A notable landmark was Paul Radin’s book Primitive Man as Philosopher (1957).  Radin not only documented serious philosophy among the traditional small-scale societies of the world, but he even suggested (timidly—one usually has to read between the lines) that modern philosophers had something to learn from the “primitives.”

Thus it came about that, in previous reanalyses of my boat-borne data, I have made deferential suggestions that we of the modern world could learn much from boat-dwellers’ family behavior (Anderson 1992), child-rearing (Anderson 1998), and ethics (Anderson, Wong and Thomas 2000).  I now wish to add politics and social theory to the list.  Anthropologists have been quick to appreciate the contributions of their subjects to the world’s store of art, music, economic plants, and medicinal lore; they have usually ignored the value of local social theory.  Yet, every group on earth has its theories of human interaction and group functioning.

Unlike Bourdieu’s Kabyle Berbers, the Cantonese talked about their society with self-consciousness and with an ability to pick out basic principles.  Bourdieu portrays the Berbers as uncommunicative about their social order.  They lived it, but they did not talk about it.  They did not have explicit folk theories of society.  By contrast, the Cantonese—not the least verbal people on the planet—talk much about society, both its realities and its ideals.  I suspect that the Berber are silent because of a kind of linguistic beheading; the language of learned, abstract discussion in Berberland was Latin in ancient times, Arabic from the 8th century onward, and French (in addition to Arabic) for the last century or so.  The Cantonese have always used their own language, or a closely related Chinese language, in learned discussion.

China possesses a huge and venerable literature on social questions.  Much of it was accessible even to the illiterate, because it was constantly being recycled through proverbs, plays, storytellers’ tales, and ordinary speech.  A reader of a manuscript of mine once commented on how preposterous was my claim that the boat people were aware of Confucian thought—because the boat people were “illiterate.”  Leaving aside the fact that many boat people were perfectly literate, this remark showed serious ignorance of Chinese folk culture.  The words of Confucius and Mencius were at least as familiar in old China as the words of the Four Gospels were in Europe and America a century ago. Every child who got through school memorized the Analects and the book of Mencius, and every community had enough educated people to guarantee that everyone, literate and illiterate alike, knew the more famous quotes from these sages.  Ideas of Confucius and Mencius, as well as of their most famous commentator, Zhu Xi (13th century), and their legendary rival Laozi, were common knowledge.

Confucian thought is social.  Unlike religious discourse, it cannot be dismissed as referring to another world, or to ideals impossibly lofty for this world.  The social theory of the Four Gospels is real enough, and sophisticated.  However, it is generally ignored, even by the most learned Christian writers, because Jesus is supposed to be talking about God and Heaven, not about this world.  Christian discourse has generally been about either individual morality or the abstract nature of Jesus and the Three-in-One.  No one could dismiss Confucian social thought this way.

Also well known at Castle Peak Bay were the basic tenets of Buddhism and Daoism, at least in so far as those influenced ordinary practice; the many Buddhist, Daoist, and folk-syncretist priests and temples in the vicinity made sure of that.  Much of this lore concerned the supernatural realm, not the human, but—as we shall see—the difference between these realms was not always seen as great, and discourse on the one often had much to do with the realities of the other.  Much of Buddhist and Daoist thought, like Confucianism, speaks quite explicitly to social questions.

Thus, the boat people had at their fingertips a quite impressive body of social philosophy.  They were not learned inquirers into the details of Mencian or Daoist thought, but they knew enough of the basic principles to argue them in everyday, pragmatic affairs.   I found their social thought to be liberating, exciting, and important.  It and its Mencian, Buddhist and Daoist roots have continued to influence me since.

Inevitably, the boat people had their own individualistic take on the Confucian-Mencian canon.  Land-dwellers tended to see Confucianism—via the thought of Zhu Xi and the Neo-Confucian tradition—as imposing rigid rules on society, especially the family.  The boat people did not see things that way.  They saw the Confucian tradition as imposing a set of general guidelines for social interaction:  be fair, be honest, defer to elders, be faithful to friends, and above all be responsibly independent.  They were aware of the many statements in the Analects and especially in Mencius that question authority or limit its scope.  They were quite aware that Mencius had not only advocated overthrowing bad rulers, but had defended this idea in public discussion with dukes and kings.  They saw the Confucian code as creating a social world that could dispense with government—a world in which internalized morality, enforced by informal means, could take the place of formal authority.  They were quite prepared to live with formal authority if it brought them benefits and was not too corrupt, but they reserved the right to rebel—or at least to pull up anchor and sail for better shores—if formal authority did not deliver.

Buddhist and Daoist ideas were interpreted similarly.  The relevant teachings counseled good behavior: probity, integrity, fidelity, honoring commitments, helping, and not imposing one’s will on others.  They held much wisdom anent getting along with the supernaturals.  They also taught independence and self-reliance, and strengthened the concept of a moral order as opposed to a forcibly imposed one.

Following Mencius, the boat people said that humans are basically “good,” in the sense of “eusocial.”  A person who did well by his family and neighbors was only doing the natural thing.  However, again following Mencius, they saw good upbringing as necessary too; like modern psychologists, they recognized that both genetics and environment are necessary to create the adult.  A bad person was clearly a badly-raised person.

This is, of course, not to say that the boat people were angels.  Like most  people and like all cultures, they had ideals and standards that were higher than anything they could practice.  Ideals were for emulation; one tried to be good, but recognized that the flesh was weak and that everyday needs meant frequent compromises.

In fact, a combination of high ideals with relaxed realism about practice was itself a part of the moral code.  Some ideals (such as utterly peaceful behavior) were only for the most self-sacrificing.  Other morals, such as sexual fidelity, were for ordinary good people, but the lack of them was not grounds for absolute ostracism.  Yet other moral stands, including honesty in financial dealings and generally peaceable behavior, were required of everyone, and failure to conform means expulsion from the community (or worse).

Similarly, the extension of a particular moral varied; treating everyone with equal warmth and friendliness was a high ideal; treating everyone so unless they were downright offensive was an expectation, but one very often honored in the breach; treating one’s own nearest and dearest with warm friendliness unless they very seriously offended one was an absolute standard.  Respect for this threefold division was one of the lessons I learned on the waterfront.

In short, I learned to see culture as a matter of rules.  These included explicit ideals, explicit rules for actual practice, vague rules of thumb, and areas where limits were set on possibilities but free play was expected within those limits.  I learned to see cultural knowledge as something that was learned and deployed for pragmatic reasons—not because it was “the culture.”  I learned to expect improvisation.

From this, I can here go on to speculate, more generally, about cultural systems.



Practice is what an ethnographer sees and records.  Structure, however, is real too.  Practice builds on it, and, over time, builds it (Bourdieu 1977; see also Bourdieu 1990; Piaget 1963; Rogoff and Lave 1984).  Cultural structures, systems, models, are dynamic things.  They never stand still.  Interaction is the direct means by which they form and re-form—a perception first (and best) enunciated by Wilhelm Dilthey (1985), but followed up by his student George Herbert Mead (1962), and subsequently by many others.  Jurgen Habermas’ work on discourse, interaction, and negotiation in politics (Habermas 1984) and especially by Emmanuel Levinas’ interactive morality (e.g. Levinas 1969) have influenced this book, as well as discussions with David Kronenfeld over the years (see Kronenfeld 1996).

Interaction between adults changes cultural institutions as people feel needs so to do.  Interaction also changes culture when the rising generation learns from their elders.  The young never quite reproduce what they hear; they must adapt, meeting new challenges and creating new ways of doing business.  Even if the objective situation does not force such a change, the need of young humans to rebel is certain to make itself felt.

Some cultural structures are hard, solid, firmly fixed, and almost changeless over time.  Basic kinship systems, basic grammar rules (the ones learned by age 10), basic terms for everyday objects such as body parts and rocks, and basic structures of food are particularly inflexible.

On the other hand, some things change with revolutionary speed.  Women’s fashions are particularly notorious for this, though they have certain underlying continuities and follow at least some rules (Richardson 1940).  Sexual behavior, gender roles, and related matters have changed profoundly in the United States since I first went to Hong Kong.  Political and economic life has altered, being progressively distorted by the growth of giant corporations, the progressive squeeze on natural resources, and the rise of state power, not only in the United States but throughout the world.

One can study structures of thought and action in such arenas only by keeping a very dynamic view of structure.  Kinship systems may seem frozen in beautiful crystalline arrays, if one looks at the short or even medium term; political systems are more like whirlwinds, created and maintained by dynamic forces and subject to sudden and dramatic shifts.  The wind-shear that creates whirlwinds has its metaphoric analogy in the political strife that creates political action .  Wind-shear, and human competition over resources and power, are always with us, providing underlying structure; the ways they play out in whirlwinds and politics are impossible to describe or predict accurately.

Language is the classic place to find fairly permanent structures.  Grammatical rules (the basic ones being firmly set) and vocabulary (all found in the dictionary) are combined to produce an unbounded set of new sentences.  These sentences, in turn, range from dully predictable (“How are you?” “I’m fine”) through moderately new but still conforming to a model (“I’m basically OK, but I have a headache that’s killing me, and I’m taking aspirin”) to completely new and weird (the famous example is Chomsky’s “colorless green ideas sleep furiously” [Chomsky 1957]).  Grammatical structure is preserved as sentences become more and more original, but meaning becomes more and more a factor of individual agency until eventually it is so idiosyncratic that communication fails.

Even kinship systems are subject to continual negotiation, but in regard to behavior and usage more than in regard to the basic structure of the system.  The boat people did not know all the words for different classes of kin, and did not usually behave according to “proper” Neo-Confucian principles, but they did all agree on and use the basic kinterms, and they did all agree on the basic expectations of support and deference that went with these.  Similarly, English has added “sibling” as a routine, universally used kinterm within my lifetime, and “parent” became standard usage only a generation earlier.  The people I called “parents,” and others of their generation, still spoke of their “fathers and mothers” or their “people” or “folks.”

We can establish a continuum, from kin terminology to local politics.  Lévi-Strauss concerned himself mainly with kinship terminologies and secondarily with other long-lasting systems, while his rebellious student Bourdieu concerned himself largely with local (and, later, national) politics.  Our theories first shape our field work and then are shaped by it—yet another example of interaction affecting systems of thought.

In realms where stability lasts over generations, we can speak of “tradition” and of “cultural models.”  In realms where change is faster than generational succession, we are better advised to speak of “rules of thumb.”  People learn from each other how to act and what to expect, but they cannot transmit much of this to the rising generation.

Music provides a good example of structuration.  Psychologists have written:  “The emergence of the human mind can be thought of as the magic of the symphony, with the biological substrates analogous to musical instruments, the information-processing operations analogous to the musical score, and the orchestration of these production components and the modulation by the social context as analogous to the conductor and musicians and the influences of the symphony hall itself” (Cacioppo et al 2004).  The anthropologist may add that culture determines what kind of score is found (concerto or Daoist ritual), how much the rules can vary with practice (singing-school rigidity or jazz improvisation), and how expert the musicians are (folk amateurs or New York Symphony).

For the last century or two, in Hong Kong as in the west, every generation has insisted on having its own music, different enough from the last generation’s to drive Mom and Dad into spluttering fury.  Yet the standard scales, and the basic structure of a tune, have persisted, in spite of several desperate attempts to displace them.  Chinese music maintains its systems; the west maintains its major and minor modes.  Twelve-tone music, Indian ragas, and the more African elaborations of the blues each had a season in the sun, but it was not a long or hot season.

Consider the evolution of one type of western music.  Scottish music clings to the pentatonic scale, the “Scottish snap” (syncopation on the lead notes of certain subphrases), and a whole set of grace notes, but otherwise has evolved from song to fiddling to rock band to concert music.  In all these cases, we see a basic structure that goes on and on and serves to define a tradition, and, building on it, an infinite and kaleidoscopic variety of practice and innovation.

Jazz melodies, chords, and note and chord progressions are fairly set.  On this minimal framework, jazz musicians built infinite variation.  The variation, in turn, is usually created by applying a few relatively simple rules and devices.   However, a very great or very insane jazz musician will break out of the mold, and do completely “impossible” and “illegal” things, to the frequent delight of audiences.


The questions to ask of any cultural system are, then:

What is the consistent, slow-changing, structured base?

How important is it relative to the rest?

What is built on top of it—a small, compact improvised system, or a huge loose one?

In this topwork, how does trivial, relatively rule-bound variation blend off into the wild and new?

At a deeper level, cultural structures have to build on biological capabilities.  Relatively specific neurobiological programs can usefully be called biological primes.  Humans naturally recognize close kin (as do all mammals), and tend to give them special care and attention but avoid having sex with them.  Humans clearly have some innate programming for language, though whether this involves a Chomskian “language organ” or a more general capacity for higher-order planning of symbol strings (coupled with throat and tongue adaptations) is a moot question.  Humans have an innate ability to classify things (Atran 1990; Berlin 1992), though, again, no one knows whether this is a true “classifying organ” in the brain or just a highly adapted ability to see detailed similarities and differences in things.  (Kant’s arguments about the latter were central to his invention of anthropology; see Kant 1978.)   Basic emotions and motivations are biological primes.

Individuals must deal with the world by using such biological primes to start off their processes of thought and action.  However, no one acts because of genetic orders, except in simple matters like breathing and eye-blinking.  Action requires building structures of thought on the biological bases.  Pace the sociobiologists, humans do not have instincts, above a very lowly level.  Human genes specify abilities to learn and adapt—not locked-in plans for behavior.  The genes guide us, vaguely, toward certain broad enterprises or courses of action, but mainly they guide us in learning how to deal with all manner of different, unexpected challenges.

Once individuals have developed thoughts, they interact.  This can lead to cultural construction; culture then systematizes widely-shared thought-patterns into cultural models.  Cultural models are most likely to occur where biological primes provide a reasonably specific ground, as in kinship, language, and probably music.  However, cultural models are also universal and intricate in the realm of religion, whose biological base is at best dubious and murky.

Thus, we can add yet another question to those above:             Do biological primes underlie a cultural institution, and, if so, how did people get from genetic plan to cultural rule?


Predicting practice is difficult; predicting it from structure is particularly so.  On the one hand, much of the practice that practice theorists have identified is merely a set of short-cuts and approximations (and even sloppy attempts at) more structured things.  On the other hand, practice, in a given situation, may simply take existing structures of the relevant system as one input—with the other inputs being anything and everything else, such as hunger or the weather.  Moreover, the boat people not only varied such things as religious ritual and kinship behavior according to immediate situational factors (rain, poverty, political worries, anything), but they would very often vary them just to be different!  Being known as an Individual was a point of pride on the waterfront.

To do this, however, one had to know something about the underlying structure, and one had to know when to stop.  Getting too far from the canonical ritual made it ineffective, and one had to know what could be varied and what could not.  Burning at least three sticks of incense, for instance, was mandatory in essentially every ritual.  The gods would not even recognize it as a ritual without that.  In kinship practice, one could get away without knowing the term for mother’s father’s elder brother, but everyone had to know the term for father’s elder brother, and had to know, also, that it was the proper polite address for any unrelated male friend of one’s father’s generation.

Other matters, however, were loosely structured simply because there was no close structure known, or because the cost of getting the requisite information was much greater than the benefit from having it.  Much interpersonal behavior with strangers, much casting of nets in new places, was done quite improvisationally, because one had to try in the absence of good information.



The problem of cultural models is made more interesting and complex in the present case by the reality suggested in the subtitle.  The boat people of Hong Kong were quintessentially Chinese.  They spoke a Chinese language, ate Chinese food, dressed in Chinese style, and certainly thought of themselves as Chinese.   In many ways, they fit the classic stereotypes that Chinese have of themselves and that Westerners have of Chinese:  family was extremely important, tradition was strong and lively, Confucian and Mencian taglines graced numerous occasions, group welfare was explicitly held more important than the welfare of individuals.

Yet, their culture was different—to many, astonishingly different—from other stereotypes of China.  They lacked lineages or large-scale family organizations of any kind.  They had relatively free, egalitarian marriages, usually contracted by—or at least with the knowledge and consent of—the couples getting married.  Their women were strong and independent.  Their men did not till the soil.  They had no fixed address.  The stereotype of Chineseness includes obedience to law and order, but the boat people were a rough lot, in both speech and action.  Some even slid over into piracy (Murray 1987, Antony 2003; the latter book is a superb source for negative stereotypes held by land people of the boat-dwellers).

Above all, they were strongly individualistic.  If there is one thing agreed on by all stereotypists, it is that the Chinese are communitarian, Americans individualistic (see Bond 1986; see also Turiel 2004 on the problems of such stereotyping).  Some celebrate the Chinese for this; Tu Wei-Ming, in particular, has devoted his life to a brilliant and insightful exposition of a Neo-Confucian (or Neo-Neo-Confucian) position (see e.g. Tu 1985), while many Americans yearn nostalgically for community (see Turiel 2004 for review), sometimes citing the Chinese as exemplary.  Others find it problematic; some East Asians are critical of the patriarchal family, and Chinese-American novelists such as Maxine Hong Kingston and Amy Tan have provided nuanced but essentially negative views of the the family-and-community world.  Asian-American family therapists Stanley Shon and Davis Ja were worried enough to write at length on how to mollify patriarchy and Westernize the family in immigrant Asian communities in America, to provide what they believe will be better family environments (Shon and Ja 1982).

Yet, at Castle Peak Bay, we lived with thousands of Chinese who were actually more individualistic and independent than Americans we knew.  This was not a difficult call or hard to assess; the gap was impossible to miss.  Most of the older boat people resisted the very idea of government, let alone communitarianism and conformity.  They had lived under Nationalist, Japanese, Communist, and British regimes, and they knew the workings of different forms of government very well indeed.  I once asked a group of men:  “Which of those governments did you find best?”  An elder in the group said:  “The Nationalists—because they couldn’t control us.”  The rest agreed.  This was only one of hundreds of remarks, by dozens of people, putting opposition to authority in the most thoroughgoing terms.

They set a high value on independence, ranging from the freedom to sail and fish wherever they pleased to the ideal of being (and being known as) a Real Character, as a unique and distinctive individual.  Unlike Americans, they did not set a high value on cultural conformity (cf. Turiel 2004, only the latest to note the contradiction of Anglo-American “individualism” with slavish conformity in matters of popular culture).  In fact, they set a low value on it, except for the basic morality of mutual aid and mutual support that served as the foundation of their social philosophy.  It was no accident that most of their songs were improvisational rather than learned by heart.

Families were a partial exception; the patriarchal family was alive and well.  But even there, patriarchy was observed to varying degrees.  One could observe a range from families almost as patriarchal as the stereotype (a stereotype that many land families actually approximated) to families where everyone had a voice, where punishment of children was unknown, and where everyone seemed to cooperate happily in spite minimal order-giving.  Children had to defer to adults, and adults barked commands in proper sailor fashion, but in most families no one ever seemed to lean on anyone else (Anderson 1992, 1998).

All observers not fooled by the “tribal” myth seem to agree that the differences between Cantonese boat people and Cantonese land people were due strictly and solely to ecological adaptation, and could be predicted therefrom (Anderson 1970b, 1972; Antony 2004; Ward 1965, 1966; and many references in all these sources).  Independence came from boat-management, just as food—although typical Chinese food—ran heavily to fish and shellfish.

However, the differences, ecological in origin though they might be, had gone on to become markers of a true subculture.  Thus, they were more standardized and generally shared than individual environmental adaptations would have been.  On the other hand, many of the adaptations disappeared rapidly when the boat people moved on shore.  I would guess that their descendents still lack lineages, but have otherwise been absorbed into the Cantonese mainstream.

Sailors who manage their own boats and have to be highly mobile simply cannot be as communitarian as rice cultivators.  Even on land, lineages routinely disappear in Chinese communities that lack large-scale landholding with intensive agriculture.  The patriarchal family, in so far as it produced the stresses and strains immortalized by novelists and sociologists (see Chapter 5), could not possibly have existed on the boats; people had to be mutually supportive, and had to be free to make split-second decisions.  When a father washed overboard, his wife or sons could hardly stand around waiting for a command decision on what to do.  Wives who were not free and equal managers of the boats were not fully effective in the fishery.  Defiance of government was natural for people whose livelihood depended on free travel, and often on “free trading” (what some would call smuggling).

Thus, Chinese culture proves highly changeable, according to immediate ecological needs.

What, then, are we to make of cultural generalizations and stereotypes?  Can an ethnographer state rules in such a universe?  The short answer is yes.  However, the rules have to be properly qualified, and grounded in more basic realities.

Lineages provide the clearest example.  All the local land-dwelling Cantonese had lineages.  None of the boat people did.  In the New Territories, this was a perfect open-and-shut matter.  (It was not in some other areas; at least some boat people nearer Guangzhou had lineages [Wu 1937], and many land-dwelling Chinese in other parts of China lack lineages.)  The explanation is simple and straightforward:  lacking land to administer, resources to manage, and even the possibility of having lineage halls or keeping records, the boat people neither needed lineages nor could maintain them.

Matters of individuality and family relationships are far more difficult to define.  The difference between land people and boat people was not open-and-shut.  There were gradations; in fact, the reality was a continuum.  Richer and more settled boat people often emulated the land-dwellers—arranging marriages, developing patriarchal relations with children, and working with government.  However, just as often, successful individuals simply used their money to maintain themselves in an even more flamboyantly individual and free-spirited lifestyle than they had had before!  Many of our friends and acquaintances boasted to us of doing exactly this.  Conversely, many far-from-affluent people had thoroughly conservative personal lives.

Particularly interesting was the ways that Neo-Confucian philosophy was used to justify individualism.  Tu Wei-Ming would have been fascinated to hear how the boat people could quote Confucius and, especially, Mencius to justify their lifeway.  Mencius’ writings include many passages about individual conscience, defiance of tyranny, and the individual and internal locus of morality.  Older and more experienced boat people knew these by heart.  Teik-Aun Wong, Lynn Thomas, and I (Anderson et al. 2000) have discussed at some length the ways in which philosophy can be bent to differing interpretations, according to the bias of the interpreter.

What are we to make of such matters?  How can anthropologists interpret and understand such subcultural and individual differences?

Many Chinese of the old school saw these boat people lifeways as deviant in the bad sense.  The stereotypes of boat people heard by Ward and myself, and found by Antony (2004), Kani (1967), and others, in traditional Chinese literature, speak to that.  For many land people, the boat people were just plain bad Chinese.

Many American anthropologists today might see the boat people as proving that culture is a mere invention of ethnographers, and that individual agency is all there is.

As the subtitle of this book suggests, I prefer to see the boat ways of life as variations on Chinese themes.  Culture is certainly real.  The boat people did conform when it was ecologically rational.  They had long ago discovered the most efficient clothing style, and they had all kept to it within recent memory.  They all ate Chinese food, drank Chinese alcohol, and spoke good if distinctive Cantonese.

Their ways of varying from more typical Cantonese patterns were all not only ecologically driven, but were also found, in a kind of latent or seed form, in mainstream (hegemonic?) Cantonese culture.  Other Chinese see the Cantonese as rough, independent, spontaneous, and prone to idealize an “original” character; the boat people merely took this a few steps farther.  And the boat people were certainly not the only Chinese in history to find in Confucian (especially Mencian) and Daoist thought a charter for independence and even defiance of government.  China’s literary history is studded with thousands of highly original individuals, many of whom justified their stubbornness in the same terms and from the same passages.

We are dealing with variation at several levels.  There were some open-and-shut differences between land and boat society.  Other differences are only differences of degree, or present a continuum from a land extreme to an occasional boat extreme.  Then, within the boat world, there were differences from community to community.  Finally, families and individuals all had their own interpretations and their own practice.

Of course, all this is deadly to the classic stereotypes, whether positive (as in Tu Wei-Ming’s writings) or negative.  Most Chinese do indeed act and think in a relatively communitarian manner, privileging society over the individual, as literally thousands of sober studies and analyses have demonstrated (Bond 1986 provides good, balanced reviews of some of them).  However, this communitarianism does not even begin to exhaust the cultural repertoire.  It is merely an aspect of the culture that happens to be preferred at one time and place.  Changing circumstances can cause a totally different interpretation of the culture, especially if already present in a latent or local form, to take over rapidly and dramatically.  (The “sexual revolution” in the United States in the 1960s was not a sudden new thing appearing; it was merely the elevation of a previously minority lifestyle to cultural dominance and hegemony.)

If one wishes to stereotype, I would suggest that Chinese culture really is defined and described by an individual-in-society perspective; sometimes the individual side comes out more strongly, sometimes the social side.  But this makes the cultural model so general and broad that it is no use as a stereotype.  One can generalize about subcultures and situations, but, even then, one must properly qualify the generalization.  Practice need not follow belief and ideology, and no two people think or act quite alike.

I have elsewhere pointed out practical applications of this (Anderson 1992, 1998).  Family counselors need not go outside Chinese culture to find egalitarian family models.  Educators need not go outside Chinese to find individual-tailored, nonpunitive, empowering education strategies.  Tu Wei-Ming need not worry about a tradeoff between Confucianism and individual rights; he can draw on an authentic and ancient Chinese tradition that combines them harmoniously.


A culture, then, is a set of plans for acting, not a rigid crust.  It is a set of useful ideas to draw on at will.  It is a tool kit, not a tool.  It has a structure, but the structure is that of a rather messily arranged toolbox.  Everything is in there, somewhere, but perhaps not in perfect shape and not in the expected spot.  Claude Lévi-Strauss’ image of the bricoleur—the shade-tree mechanic—is deservedly famous.  Not only the traditional consultant, but everyone, has to be a shade-tree mechanic when using his or her cultural tool kit.  The boat people found it necessary to draw on the shipboard tools—the fishhooks, spars, rigging-mending equipment—rather than the plough or the bricklayer’s trowel.  (And France’s counterpart to Home Depot calls itself the place for bricoleurs.)

One must, also, imagine a tool kit that blends at the edges into other, slightly different kits (Vietnamese, Thai…).  Moreover, the tools have the useful property of expanding or shrinking at will.  As every AI expert would agree, cultural tools are more like computer software than like hardware.  They can be stored in zipped files when not needed, or expanded by all sorts of patches and supplements and individual tinkering.

Or, to return to another metaphor, culture is indeed a “thing of shreds and patches.”  But the shreds and patches are woven into a wonderful patchwork quilt—one with the miraculous quality of changing to fit the needs and tastes of anyone who lies under it.  Somehow, it keeps us all warm.



A final note on cultural modeling, to remind us to be sober and modest, is the question of ethnographer authority.  Throughout the history of anthropology, and especially in the last couple of decades, scholars have questioned the right of ethnographers to produce “the” account of “their” people.  Who am I to talk about the boat people?  I spent something over two years in Hong Kong, including nine months living on a sampan, but I am, obviously, not a seui seung yan by birth.  I have no right to speak in an authoritative or authorial voice.

This book is not in any way “the ethnography” of the Hong Kong boat people.  It is a personal text, an account of one man’s experience in a few boat communities.  It is not intended to be a definitive or canonical text.  On the other hand, I know a great deal about the boat world, and I was there during a critical time: the period of its decline and disappearance.  This gives me not only the right, but (I think) the obligation, to record what I know of the brilliance and troubles of the bay’s last years.

I offer, then, what is, and can only be, a personal view, the view of a sympathetic and experienced outsider.  I was not there long enough to become an insider, but I was there too long to be a coolly “objective” outsider.  Thus, I cannot pretend to speak ex cathedra on boat society.

I tried my best to remedy the lack of boat people’s own voices by recording the boat people’s songs in some detail (Anderson 1975), and by recording life accounts (Anderson 1970b:226-248).  Unfortunately, in those days, publishers were not interested in such materials, and I did not follow up as I should have done.

I have also tried to stick as closely as I can to factual matters that could be objectively verified:  fish species, dollar amounts, number of people on a boat, actual behavior at religious ceremonies.  In this book and in my previous publications, I have used the boat people’s own concepts, ideas, and often exact words.  In particular, I have tried to keep interpretation as close to my boat friends’ own words as I could; I am too uncomfortably aware of my own limitations to indulge in Geertzian flights.  This has gotten me criticized in the past for not imposing enough “theory” (whatever that might mean) on the “raw data.”  Today, I hope that the many critiques of the “authoritative voice” have made us more sensitive to the aridity (if not immorality) of presenting ex cathedra theory and interpretation in one’s own authoritative voice, while silencing one’s subjects and denying them even the exiguous existence of “aggregated statistical data.”

I wish I could present my friends as living, breathing human beings.  The generation of boat men and women who remember the old ways is now as old as I am, and—to my knowledge—none has stepped forward to write the great novel or great memoir we need.  My own few words can be no more than a stopgap, while we await that hoped consummation.







The Cantonese fishing people of coastal Guangdong Province maintained until recently a highly distinctive and highly traditional form of life.  For centuries, they have been regarded as a distinct minority within Cantonese culture.  Until recently, most of them lived entirely on their boats; virtually all the rest lived on huts built over tidewater.  They lived by fishing, shellfish and seaweed collecting, cargo-carrying, and small-scale trade.  Their speech often differed slightly from that of land neighbors (McCoy 1965).  Their clothing and foodways also varied slightly from land norms.  Last, but not least, their religious practices differed substantially from those of the land people.

The lifeway of the boat people is dramatically different way of being Chinese.  In current usage, they are typical “Han” Chinese, though the Cantonese traditionally called themselves “people of Tang” rather than “people of Han.”  The terms refer to the Han and Tang Dynasties; the Chinese Empire took its mature form in Han (206 BC-220 AD), but the Cantonese feel they were not really Sinicized until Tang (607-960), although Guangdong was actually conquered and occupied under Han.  (And Han tomb models of boats show the boat people were living then pretty much as they are now.)  The “Tang” label is so persistent that Grant Street, the main thoroughfare of San Francisco’s Chinatown, is known throughout the Chinese world as “Tang people’s street.”  (This is T’ong Yan Kai in Cantonese.  Usage has changed, and now Cantonese, along with others who traditionally speak languages of the Chinese family, tend to call themselves “Han” Chinese.  On the ethnically diverse history of south China, see Eberhard 1942.)

The boat people fit squarely into that universe.   They merely presented a set of variants, bearing a Wittgensteinian family resemblance (rather than a simple set of changes in one pattern).  Each family had its own way of being Chinese and of being seui seung yan.  Also, each family changed and adapted over time, so today’s range of knowledge and practice might be varied tomorrow as circumstances altered.


The land people called themselves Punti (pun tei), “locals,” meaning Cantonese-speakers in contrast to the Hakka, Hokkien, and other non-Cantonese groups .  The boat people were generally not considered  Punti, because of their lifestyle.  However, the boat people were Cantonese-speaking locals, and were thus Punti on the dotted line. I avoid the word because of these confusing webs of meaning.

In imperial times, boat people were legally constituted as a separate ethnic group.  They were subject to discrimination; until late Qing, they were not allowed to sit for the Imperial examinations.  They were not allowed to move on shore, though they could, in many locations, build houses on piles over tidewater.  This had been done for centuries at Tai O, on Lantao Island, across a narrow strait from Castle Peak.  Castle Peak was following suit in the 1960s, with a rapid rise of handmade shacks built over the mudflats.

Boat people probably provided many of the sailors, and much of the expertise, that made China a great sea power in dynastic times.  Contrary to a myth of China’s maritime insignificance, Gang Deng (1999) has shown that China was a leader in seafaring, marine technology, and exploration, especially in early Ming but also at other times in history.

In spite of this, land-dwellers have generally regarded boat people with prejudice and contempt.  Boat people suffered discrimination in Imperial times and subsequently.  Antony (2003:12) summarizes the traditional stereotype of boat culture:  “It did not share in the traditional Confucian values of honesty, frugality, self-restraint, and hard work; rather, it espoused deception, ambition, hedonism, recklessness, and getting ahead by any means.”  Of course, the boat people reversed this; they had the exactly same stereotype, but with themselves as the honest workers, the land-goers as the devious amoralists.  Boat people frequently expressed this to us at very considerable length.  The one place they agreed with the old stereotype was on “recklessness.”  The boat people admitted they were more risk-taking, less cautiously routine-bound—but, instead of “reckless,” the words they used were “courageous,” “free,” and “on our own.”

For decades or centuries, the boat people have been claimed as more “aboriginal” or “tribal” (in some sense) than the land people.  This belief was shown to be false decades ago (Anderson 1970a, 1970b, 1972; Ward 1965, 1966; both these authors review a great deal of earlier work), but it still surfaces on occasion.  In fact, the boat people clearly have some “aboriginal” brackground, but non-Han ancestry is general in China’s far south (for a qualified view of the boat people from this perspective, see Ho 1965).  Linguistic, historic, and cultural evidence suggests that southeast China was largely Thai—with many other linguistic groups, such as Yao and Miao—before, and for a long time after, the Chinese conquered it in the Han Dynasty (206 BC-221 AD).  For instance, the delayed transfer marriage still practised in the Guangzhou area (Stockard 1989) is a Zhuang custom (Jeffrey Barlow, unpublished ms.; the Zhuang are a Thai-speaking group now dominant in much of the Guangsi Zhuang Autonomous Region just west of Guangdong).  It has probably been retained by former Zhuang who have become Cantonese, rather than having been borrowed by Chinese migrants (see Stockard 1989).  In Cavalli-Sforza’s studies of human genetics, southeast Chinese were found to sort genetically with southeast Asians, not with northern Chinese (Cavalli-Sforza 2001).  This being the case, contrasting “aboriginal” boat people with “Han” land people is completely wrong.

Often, unfortunately, “tribal” claims for the boat people are coupled with the use of the highly offensive ethnic slur “Dan” (or Danjia; Cantonese Tanka; see e.g. Siu and Faure 1995; Ye 1995).  This is supposedly a “tribal” or “ethnic” name of the boat people.  It goes back to imperial times, but it seems purely an outsiders’ term with an insulting connotation.  It is not always insulting, and lately it has been rehabilitated by boat people from Tai O and elsewhere, following the rehabilitation of “Black” and “Chicano” in the United States.  This fascinating bit of sociolinguistics was conveyed to me (in an email of May 15, 2002) by a young Hong Kong sociologist, Ho-fung Hung, who studied Tai O and its recent conflicts and activist politics (Hung 1998, 2001a, 2001b).  There are several folk etymologies purporting to explain why the word tan, “egg,” came to be an ethnic slur word for Cantonese boat people:  Their boats look like eggshells, “Tan” was a “tribal” name, “tan” is their pronunciation of some other word (various words are suggested), and so on.  These speculations are no doubt as valid as most folk etymologies.  The boat people referred to themselves as seui seung yan (Putonghua shui shang ren “people on the water”; seui seung yan was pronounced soi song yan in Hong Kong boat dialect) or by similar descriptive phrases.  In English, the people in question have always been called “boat people.”  In the 1970s, this term came to be used for the Vietnamese refugees who fled by boat from Vietnam.  Many of these came to Hong Kong, and obvious confusion followed.  In this book, of course, the term “boat people” applies to the boat-dwellers native to the Hong Kong area, not to the Vietnamese.

In fact, there was continuous movement from land to water and from water to land.  The water people were never a separate group.  My interviews and life histories revealed that many people traced their descent back to shore-dwellers who had taken to the boats during periods of unrest over the last couple of centuries—often within the last two to three decades.  World War II, in particular, drove many land people to the boats, where life was safer than on shore.  Conversely, boat people have been moving onto shore for millennia, often in substantial waves, when opportunity or local decree encouraged them (see Antony 2003 for the most recent historical treatment of these millenia-old patterns).

They did, however, maintain a separate lifestyle with many cultural features with linked them with other boat people in China, and differentiated them from the land-dwellers.  Most of these features can be understood as adaptations to a seafaring, boat-dwelling life (Antony 2003; Ward 1965, 1966).

However, successive governments, both local and national, constructed them as a separate ethnic group—greatly oversharpening and exaggerating a real subcultural difference.  The boat people became a “separate tribe,” as if they were as culturally distinct as the Tibetans or Mongols.  The only real difference, howeever, was that the boat people lived on boats (or pile houses over water) and worked at water occupations.  Even such clear cultural markers as lack of lineages, special accents, and independent ideology were not hard-and-fast distinctions from the land world; some land groups lacked lineages and had similar accents and attitudes.  Costume, folk songs, and a few other such usages did set the boat people apart, but not very dramatically.

Yet the boat people received prejudice; often they had been forbidden to live on land or—locally at least—to sit for imperial examinations.

They came to see themselves as “dragons on water, worms on shore” (Ward 1965, 1966).  The double terminology is critical here:  they were held down, victimized, and unable to control their destinies when on land, but on the water they were unquestioned rulers.  Land people could travel on water or ship cargoes only through boat-people agency.  On the water the boat people were fully in control of the human or social side of their lives, though the weather and seas were still out of their control.

Thus, much of boat culture and discourse was about control and its negotiation.  A sense of control over one’s life is a powerful human need (Anderson 1996), and the boat people had it on water, lost it on land.  Their lives had to deal with that constant negotiation of a major, basic human psychological mainstay.

The boat people of the central Guangdong coast formed a large social network, united by constant shifting back and forth.  Until the rise of the Communist government, local and national borders meant little to them, and even after 1949 it was always fairly easy to weigh anchor and move.  The Hong Kong border was permeable.  Tens of thousands of boat dwellers had fled at one time or another from the unrest and poverty that were endemic in China during the 20th century.  Particularly large waves came to Hong Kong during the famine years of 1958-61, when millions of Chinese starved to death.  A steady trickle came through after that, until prosperity and the call of shore life made China more attractive during the 1970s.

The main source area of Hong Kong boat people, both long-established and newcoming, was Hong Kong’s immediate neighbor region:  the two-county area of Baoan (Po On in Cantonese, and so called henceforth in this book) and Dongguan (Tung Kun).

The distinctive boat-dweller accent of Hong Kong seems to be, basically, the accent of that area.  It is very close to standard Cantonese (the urban accent of Guangzhou and Hong Kong).  There are some patterned differences.  Most obvious is the substition of  –o– for –eu– (as in soi song yan, above).  “Hong Kong” probably got its English name via boat people, because it is Heung Kong in standard Cantonese.  (Poetically “Fragrant Harbor,” literally “incense port” because sandalwood was once shipped from it, Hong Kong is Xiang Gang in Putonghua.)  The boat people changed initial ng- to something like a w- (ngo “I” becomes wo).  Some final consonants were dropped or assimilated (all final nasals became –n).  As in colloquial Hong Kong Cantonese generally, unaspirated ts before vowels is sounded like English j (tsai “little one” is pronounced jai).  There were also some tonal quirks.  I have not indicated boat pronunciations in this book; I simply use standard Cantonese forms (see Anderson 1970b and McCoy 1965 for full technical descriptions).

Many boat people came from farther upstream, around Guangzhou city.  A few came from still more distant counties.  In Tai Po Kau, on the east side of Hong Kong’s New Territories, there was a large community of boat-dwellers speaking a dialect of Southern Min (Hokkien), a completely different language from Cantonese.  Their dialect was spoken from northwest Guangdong Province into neighboring Fujian.  These people kept to themselves, and had their own style of boats, their own foods and customs, and their own fishing techniques.  They disappeared as a community in the 1970s and 1980s, with their dialect and culture still unrecorded (see Anderson 1970b:257-259 for what little is known).  They constituted no more than 3% of the boat people of Hong Kong; the rest, including all those I knew at Castle Peak, were Cantonese.

In Hong Kong, the boat people were concentrated in a number of good harbors.  The biggest and most famous concentration was at Aberdeen, on the outer side of Hong Kong Island.  (The main city is on the inner side, facing the mainland.)  Others were at Kowloon Harbor, Deep Bay, So Kwun Wat in Hong Kong, and elsewhere.  Barbara Ward, the only other person to do prolonged participant-observation research on the Hong Kong boat people, worked at Kau Sai, a small settlement on the east side of the New Territories; Castle Peak Bay, a larger settlement on the west side, was a quite different place.  Outside of the British colony, there were huge concentrations of boat people all around the Pearl River estuary.  In 1966 we visited Macau, then a Portuguese colony, where we observed huge concentrations of boat-dwellers.  In 1978 I visited Guangzhou, and saw what little was left of the boat culture there; by then, most of the boat-dwellers had moved on shore there (as in Hong Kong).  Smaller concentrations existed all along the coasts and major rivers of Guangzhou and Guangxi.  Elsewhere in China were perhaps three million boat people, speaking the languages of their own areas (see e.g. Worcester 1959).  Almost everywhere, the boat-dwellers spoke the same dialects and languages as the shore people nearby.   Nowhere did they have their own language.

This is true throughout Asia; almost every country with rivers and coasts has its own boat people, all local, all speaking local languages or at least languages derived from home ports not too far away (Sopher 1965—eked out by my own observations and conversations all round southeast Asia).  The boat-dwelling adaptation is clearly something that appeals to all sorts of people; representatives of many of the ethnic groups of Asia have taken to the sea.  Some, such as the Sama (Randall 1977) and certain Bugis trader groups, have become long-distance voyagers and merchants.  Humans are usually creatures of the land, but here they have become as bound to the water as are the fish that supply much of their livelihood.

Speculations that the boat-dwellers are some sort of distinct “tribe” or “tribal group,” or otherwise have separate origins from the land people, have often been entertained (see Sopher 1965).  All the evidence suggests otherwise:  that they are simply representatives of the shoreline groups whose languages they speak.  (The Moken “sea gypsies” of the Malayan peninsula, who speak a distinctive dialect, are a partial exception; see, yet again, Sopher 1965).  The Sama have dispersed from the south Philippines to Sulawesi and elsewhere, but their origins are clear and they know their homeland.  Vietnamese boat people speak Vietnamese.  Japanese ones spoke Japanese when they existed as an identifiable group.

In Hong Kong, the boat people are clearly Cantonese in language, culture, and—usually—origin.  We talked to many who knew from family tradition that they were recently land-dwellers, only a couple of generations away from rice farming.  However, the boat lifestyle predates the Sinicization of the south.  Archaeology reveals that there were boat-dwellers as early as 1500 BC in Hong Kong (as I learned from archaeologist William Meacham, while observing with him the ongoing excavation at an obvious boat-dweller site on an isolated island).  These would probably have been Thai-speaking.  On the other hand, Austronesian voyagers from south China apparently settled Taiwan and later the South Seas (Bellwood 1997), and it is tempting to see them as related somehow to early boat people.  Doubtless these early boat-dwellers represented many linguistic communities.

Boat people much like modern ones were certainly present by Han, as we know from tomb pottery.  People throughout the Han empire had the custom—incomparably useful to archaeologists—of making small pottery replicas of the things of this world, to accompany the dead into the other one.  From this we know that the Han world had woks, pagodas, “sharpei” (sha pi) dogs identical to modern ones, and sampans almost exactly like the one on which we lived.  Even minor details of style are the same.  These models show the oldest fore-and-aft rigging and watertight compartments in the world; evidently the South China Sea was an area of maritime innovation, and the lack of subsequent change indicates early perfection rather than late conservatism.

After Han, we get occasional flashes of light in what is usually a profound darkness.  Boat people are known to have existed all over the major rivers and coasts of China from a very early time.  Their folk songs, often clearly related to modern songs, were recorded by poets.  In the 16th century, Feng Menglong recorded some boat folk songs similar to ones I recorded in 1965-66 (Anderson 1975).   Even earlier, the boat folk songs of the lower Yangzi inspired Tang Dynasty poets to imitate them in elite style.  Cui Hao (704-754) gives us an exchange between a girl and a young man, of a type that could still be heard on Castle Peak Bay in 1965.  The girl sings the first verse, the boy the second.  Hengtang and Longpole (Janggan) are near Nanjing.


“Where are you from good sir

this maid is from Hengtang

I ask while our boats are moored

perhaps we’re from the same place


My home overlooks the Great River

I travel along the Great River’s banks

I’m from Longpole too

but I’ve never seen you before.”  (Red Pine 2003:36-37).


In the 18th Century, the poet and writer Shen Fu lived in Guangzhou for a while, and had a romantic affair with a boat girl.  He describes it in one of the most exquisite and heartbreaking short accounts, in Chinese or any other language, of a lost love.  This account was published in a small book titled Six Records of a Floating Life (Shen 1983); only four of the chapters ever saw the light.  He includes some general ethnography, including the significant note that the word for “what?” was mie (Putonghua for “rice grain”).  Mi ye is still the boat pronunciation of the Cantonese mat ye? (“what?”).  The title of this book and its predecessor The Floating World of Castle Peak Bay (Anderson 1970b) are both based on Shen’s, whose title, in turn, refers both to literal floating and to the Buddhist image of this transient life of ours as a “floating world.”

After Shen, accounts get thicker, and have been synthesized in historical studies by Ho (1944), Kani (1967), Ye (1995), and others.  These are generally marred by a too-literal reading of local gazetteers and similar sources that allege “tribes” of boat people.  The boat people do not have tribes, or anything similar.  The “tribes” were probably followers of a particular local leader, or people who came from a certain area and named themselves from it.

In the early 20th century, sociologists discovered the boat people.  Wu Yuey-len (Wu Yuelin in modern transcription) and Chicago-trained Chen Suching (Chen Xujing), and later their student Ho Kaak-yan (Ho Ge’en), were associated at Lingnan University in Canton.  They carried out extensive descriptive sociology of the sort done at the University of Chicago in those days; in fact, it was excellent ethnography.  They described everything from economics to folksongs (Chen 1933, 1935, 1946; Ho 1944, 1965, 1966; Wu 1936a, 1936b, 1937).  Ho later moved to Hong Kong and taught at the University of Hong Kong, where I interviewed him; Chen and Wu stayed at Lingnan; Chen survived the Maoist era and lived to a venerable old age.  Barbara Ward carried out major research on the boat people of Kau Sai, Hong Kong, in the 1950s and 1960s (Ward 1955, 1958, 1966, 1965, 1966, 1967, 1969).  The economist Frank H. H. King (1955) and the linguist John McCoy (1965) worked with her there for short periods.  Ward’s pathbreaking, brilliant, and richly theorized ethnography was tragically cut short by her untimely death, before she could publish more than a few articles.  Her interests ranged from child-rearing to livelihood to history.  She was instrumental in showing that the boat people were not a separate group but were simply Canontese on boats.  More challenging to theory was her work on the boat people’s own ideas about their social system (Ward 1966), of which more below.

Many popular writings and photographs were devoted to the boat people.  Col. V. R. Burkhardt published a long series of columns in the South China Morning Post, later compiled into three volumes (Burkhardt 1953-58), that contains much excellent material.  Less excellent were the vast numbers of photographs, film, paintings, transient writings, and other materials that exploited the “picturesque, quaint” boat people.  The old sailing junk was a photogenic craft, and became a symbol of Hong Kong.

After that, the research reported in this book is the major published material.  (See also Hung 1998, 2001a, 2001b, and references therein.  Anderson 1970a, 1972 provide full bibliography of boat people studies through 1970.)  The boat people no longer exist as a major community in Hong Kong, but many remain along the rivers and coasts in more remote parts of Guangdong.  The opportunity for studying the traditional culture is now gone forever, but opportunities now exist for studying the modern vicissitudes of those who make their living on great waters.



Like the boat people, Castle Peak Bay has a long history and a controversial past (Balfour 1941; Lo 1963; Ng 1961).  When Castle Peak and its environs were first noted, in the Tang Dynasty (618-907 AD), Castle Peak was separated from the mainland by a strait that connected the present Castle Peak Bay with Deep Bay.  Later, the strait filled, and Deep Bay became shallow.  Finally, from the 1960s onwards, Castle Peak Bay progressively shrank, as more and more fill encroached on it.  By 1990 it had disappeared.

The strait was known in ancient days as Tun Mun (or Tuen Mun, “Garrison Strait”; mun, literally “gate,” means “strait” or “passage” in Cantonese nautical usage).  This name may have been used later for the strait on the other side of Castle Peak, a still-open channel separating it from Ling Ting Island.  The area was once of the utmost strategic importance; it guarded the north corner of the Pearl River estuary, while Macau guarded the south.  Thus, in the Tang Dynasty (620-907), a garrison was established in the area. The great Tang writer Han Yu supposedly passed through the area, though details are shadowy.

According to local legend, a mysterious monk, Pei Tu (“Cup Measure”), who sailed in a winecup and could disappear at will, founded Castle Peak Monastery in the fifth century.  In fact, the monastery was founded in the early 20th century.

During the Nan Han Dynasty in the 10th century, records speak of “Tan” people in the area, using the same character as that in “Tanka” (dan jia; Lo 1963:23).  In the Song Dynasty (960-1279), the area remained a port and checkpoint.  Local legends once again take us farther:  the last boy emperor of Song passed through and found brief refuge on Lantao Island, where his guardian and uncle is still revered as divine patron of the community of Tai O.  More securely attested is the passage of Jorge Alvares in 1513; this Portuguese navigator touched at Ling Ting Island and was the first European to see castle Peak.  The Portuguese fortified Lantao, but the Chinese drove them out, and they consolidated their position in Macau instead (Braga 1956).  The local gazetteer recorded that the Portuguese raided the surrounding countryside, capturing and eating children (Ng 1961).  Such claims of cannibalism occur again and again in Chinese history, with no factual evidence presented; the view of “barbarians” as cannibals merely a standard trope, as it is in European and Euro-American travel lore (see Chong 1990).

Piracy increased, notably at the fall of Ming (1644), when Ming loyalists took to the sea and harassed the coasts for years (Anthony 2003; Murray 1987).  Some were theoretically under the direction of the Ming loyalist leader Coxinga.  (The name is a European rendering of kok sing a, Southern Min for “Imperial Surname Kid,” a slangy reference to his being granted the Ming imperial surname for his services in trying to save the dynasty).  Japanese, local boat people, and shore people turned pirate were also involved.  From 1662 to 1669, the Qing government depopulated the whole coast, the people being forcibly resettled inland.  The local economy never recovered until modern times.

The British leased the New Territories in 1898.  The area remained rural and agricultural, raising high-quality rice and fruit, until after the Japanese devastated it in the Second World War.  From 1950 onward, urbanization spread explosively from Hong Kong and Kowloon.  By 1997, when the New Territories reverted to China, the southern New Territories, including Castle Peak Bay, were a vast sprawling city.  The northern New Territories were still rural, but no longer agricultural; the labor force had gone to the city or to England and Europe to work.

Much of the urbanization of Castle Peak was driven by refugees from China.  These came largely in three waves:  1949 (during the civil war that brought Mao Zidong to power), 1957 (initial communization of rural lands), and 1960-61 (famine following the failure of the Great Leap Forward).  The last of these was by far the largest as far as Castle Peak Bay was concerned, for it brought floods of desperate people, most of them boat people, to the bay.  Most of these were from nearby, especially from Po On (Bao An), the Chinese county (xian) just across the border; most of its boat people had fled, according to informants.  Many more came from Tung Kun (Dong Guan), the next one north, and from other counties around the Pearl River estuary.  Many brought functional boats and kept fishing, but most had fled in aging and decrepit craft, and had no capital.  They could only squat in houses built of scrap lumber on the waterfront, often on pilings in the water.  They were destitute, and many children died of malnutrition.  Finally, from 1965 and especially from 1970, expansion of factories and shops in the area provided employment.

The town of Tun Mun Kau Hui (“Garrison Strait Old Market”) was a gray-brick town on the old creek entering the bay.  An eastward extension, Tun Mun San Hui (“Garrison Strait New Market”), began life about 1945 on a mudflat at the head of the bay.  It grew with spectacular speed.  By 1965 it had a government clinic, a market, a school, and other amenities.  By 1974 it had expanded onto bay fill and acquired many factories and high-rise apartment blocks; by 1997 it was a huge city.

In the 1960s and 1970s, there were several thousand boat people in the bay area.  No exact figures can be given, for people were always moving in and out, shifting to shore life, shifting back to boat life, or otherwise acting as a truly “floating” population.


The boat people were dependent on the natural world for life and livelihood.  Earning their income from the sea and much of their food from local land, a the mercy of weather, and required to navigate dangerous passages, they had to know a great deal about their surroundings.  This section provides a brief introduction to those surroundings.

The New Territories and Kowloon comprised a peninsula.  Most of it is low but rugged granitic mountain land.  Castle Peak is just under 2000’ high.  Overall, the coast of the region had been sinking, producing many bays and islands.  The islands are drowned mountains.  Largest was Lantao, across a strait from Castle Peak.  Most flat land is recent alluvium, deposited by runoff from slopes deforested by human activity over the last two thousand years.  In the torrential rains that follow typhoons and summer storms, topsoil washes in floods from cleared land, leaving gullies and bare half-decomposed granite.  The granitic soil, under tropical conditions, rapidly loses fertility, and becomes laterite or simply bare rock.  Attempts to reforest—usually with pine (Pinus massoniana)—were rarely successful, since even this almost-indestructible tree could barely hold on in such conditions.  Moreover, farmers regularly burned the mountains, to clear brush and eliminate snakes and other wild animals.  Where this did not go on—as in fengshui woods and forest reserves—the natural semideciduous tropical forest eventually reasserted itself.  A magnificent wood at Tai Po Kau Forest Reserve now reminds Hong Kong of what it has lost.

The peninsula is tilting slowly from northwest to southeast, so new lands are emerging from Deep Bay.  Because of this and siltation, the bay is no longer deep.

Hydrology is dominated by the distant but powerful Pearl River, whose waters pour in a vast muddy stream through the estuary to the west.  The complex pattern of islands and rivers causes strong currents.  Narrow straits constrict the flow, forcing tidewater to race through fast and powerfully.  Turbulent currents boil up where they collide, as in the “Meeting-Waters Strait” east of Castle Peak.

Hong Kong has a classic monsoon climate.  Rainfall is concentrated in summer.  Annual rainfall is 70”-80” per year, but variations are substantial.  Hong Kong (average 86”) had 35” in 1963, about 160” in 1966.  About 25 typhoons and tropical storms affect the area per year, with one or two hitting dead on, producing violent winds and rains that may deposit 25”-30” of precipitation in two or three days.  Temperatures range from freezing (Hong Kong is the only place in the world’s tropics to have freezing temperatures on the coast) to nearly 100<o> F.  Summers are unbearably hot and humid.  Storms keep fishermen in harbor; it is a time of boredom, worry, and misery from soggy clothes that never get quite clean or quite dry.  Late falls and winters are dry and clear, usually delightful, though cold drizzle may go on for a day or two.  Fish run, and there are few storms to keep fishers in harbor.  This is the time for festivals, weddings, and general enjoyment.  Spring is a time of fogs and of sudden temperature shifts as the dry north wind and wet southeast wind battle for control.  Sickness is rampant at this season, and the boat people blame this wild fluctuations in weather, as well as the relative lack of green vegetables at this time.  Early fall brings typhoons.

Farm villages, built of hard-fired gray brick with gray tiled roofs, had changed relatively little since the Qing Dynasty.

The rural lands around the bay produced, in the 1960s and 1970s, rice, Chinese cabbages, ducks, pigs, chickens, and fruit.  More specialized crops included sugar cane, sweet potatoes, eggplants, peppers, papayas, lotus, water spinach, and much more.  Geese and water buffaloes provided a distinctive local touch.  Cantonese cultivators grew the high-quality rice and many of the vegetables, but Hakka farmers in the higher and hillier areas were also important, especially as vegetable and fruit gardeners.  The more level areas were incredibly productive.  Chinese cabbage yielded eight or more crops a year; rice two or three.

Rice yields were up to 2500 pounds or more per acre per crop, a substantial yield in those days.  Good yields in the Pearl River area today are closer to 10,000 pounds per acre per crop.  But rice is no longer grown in the Castle Peak area.  Saddest of all, when I was last in the area, I was told that the distinctive, flavorful, salt-tolerant rice of the northwest New Territories had become extinct.

Thanks to all this farming, food was abundant, cheap, and superb.  Farmers took enormous pride in the quality of their meats, grains, and vegetables.  Good quality brought high price; a pound of tender, succulent mustard greens brought about four times the price of a pound of old and tough ones.  The pork and rice of the area were deservedly famous.  Eggs were fresh and came from chickens and ducks fed on natural, varied diets—often largely weeds and bugs, since these birds were the main pest control agents on the farms.



The cultural ecology of south coastal China is a complex, elaborate system, brilliantly effective at producing human food, miserably unsuccessful at preserving forests and wildlife (Anderson 1988; Anderson and Anderson 1973; Elvin 2004; Marks 1998; Wen and Pimentel 1986a, 1986b; we, with Wen and Pimentel, emphasized the benefits to humanity, but Marks and Elvin emphasize the damage to wildlife).

Farmers directed rainwater and stream water into irrigation channels—vegetables higher up, then rice lower down.  The rice requires exquisite control of water levels, so it can be grown only where water is always available but can always be turned off or drained off.  Burning the hills and fields released nitrogen fixed by legumes and other plants.  More nitrogen was fixed by blue-green algae growing in the paddies.  This nitrogen was constantly recycled through animals and concentrated in their manure, used for fertilizer.

Still lower, where water was more difficult to drain, were buffalo and duck farms.  Below them, where water never drains away, were lotus and fish farms.  Then came brackish shallows, with shrimp and oyster farms.

At every stage, composting and recycling of all wastes—from human and animal dung to old ropes and sandals—continued to enrich the system.  Over the millennia, this process has massively transferred soil and nutrients—especially nitrogen—from the uplands to the deltas.  The mountains are eroded, often to bare rock, while the delta lands build rapidly outward to sea, and become ever richer with accumulated silts.  Much of the Pearl River Estuary area is sa tin (sha tian), literally “sand fields,” but, in local usage, “polder land.”  These fields have been reclaimed from the sea, as shallows fill up with silt.  This process continues actively today.   My friend Hugh Baker, with his inimitable turns of phrase, summarized my findings by saying that “it must make nitrogen atoms unfortunate enough to be in other parts of the world feel unloved” (Baker 1989:661-662).

Inevitably, some nutrients escape this almost perfectly closed system.  These do not escape for long.  They go to feed the fish that, until a generation ago, produced one of the greatest and most diverse fisheries in the world.  The boat people were those who caught these and brought them to land—thus closing the final loop in this endless chain.

The South Chinese rice agricultural system is certainly one of the wonders of the world.  Pace Mark Elvin and Robert Marks, south China was better than other areas even in its dealings with elephants and tigers; Native Americans seem to have exterminated the American equivalents without maintaining much population density or supporting a civilization.  There are still tigers and elephants in southern China, in spite of the hundreds of millions of people, but the last American elephant died (probably by spear-point) thousands of years ago.

The South Chinese system does not, however, reflect a society of rich farmers.  It was the product of a dictatorial imperial system that encouraged farming and farm technology but discouraged radical change (whether ecological, economic, or social).  Given the impossibility of getting ahead any other way, the farmers had to work harder and harder to wring a livelihood out of tiny parcels of land—parcels that got steadily smaller as population grew.  (The system survived by the growth of land seaward, but population growth usually ran faster than land area growth.)

As pointed out by Philip Huang (1990), this resulted in something rather similar to the “agricultural involution” that afflicted colonial Java (Geertz 1963).  Lack of capital and lack of legal opportunity led to “biological” innovation rather than mechanical or technical (Hayami and Ruttan 1985).  Farmers had to develop what they could develop without much cash and without any political freedom to experiment with new political-economic institutions.  Moreover, farmers had to have many children, to provide a labor force for the farms.  The typical completed family was around seven children (our data, but they are quite representative of old China; see e.g. Eastman 1988; Huang 1990).  With so many children, a family could not save or get ahead without some special circumstance.

Moreover, the ecological damage highlighted by Elvin and Marks was real enough.  Constant burning afflicted the high uplands.  This was terribly destructive to forests and terribly causative of flooding and erosion, but it did release nutrients in the form of ash and mineral salts.  Rains washed these nutrients off the high hills.

This is not to say the system stagnated; it did not.  It steadily grew, changed, and improved.  But the changes and improvements were small, and were of a “biological” and involutional nature, so they did not greatly help the burdened peasantry.  They enabled more and more people to live, but not many lived better.  Any catastrophe (such as the Taiping Rebellion and the wars of the 20th century) pushed millions over the brink into starvation.  Millions more died of endemic diseases such as malaria and bubonic plague (Carol Benedict’s superb work on plague, 1996, now a classic of medical history).

The system was at once a masterpiece and a deathtrap.

Within it, the boat people played a key role.  They caught the fish that provided 90% of the animal protein along the coast (Anderson 1988; Buck 1937).  They provided the water transport, essential to the system at all levels.  They were, thus, one group within a tightly-locked ecological and economic system, not a separate and independent people.  Yet, they had the independence of the free drifting life.  They were fully aware of the contradiction between these two aspects of their reality, and they felt the resulting tension; this fueled their constant talk of “independence.”  Cognitive dissonance was audible in many a waterfront discourse on such matters.

As the world runs out of resources, we all will have to return to something like the south Chinese system—intensive, sustainable, self-renewing.  One hopes we can do it without the ecological devastation of the wild, the totalitarian central government, and the local problems of violence and conflict.   Realistically, we probably cannot.  If we can, it will be because we learn from the south Chinese experience.






Castle Peak Bay, as of 1965, was a triangular bay, the head being at the north end.  This end was largely mudflats.  Wrecked boats settled on it, and a typhoon shelter (a mudflat except when typhoon waves filled it, and adequate only for small craft) took up part of its space.  By 1974, all this area was filled, but the rest of the bay—from the center south—was still a fairly deepwater harbor, with boats ranked in neat rows after the day’s fishing.  A small island, Mouse Island on maps, lay in the east central part of the bay.  Much beyond it, the roadstead was too open for anchorage.  The boat people called this island “Little Ghost Island,” because they abandoned dead children on it—not from poverty, but from adherence to an ancient tradition that no one seemed to want to talk about, partly because there was a rumor that in the past the children had not always been completely dead.  By 1965, most children were buried on shore.

The waterfront was lined with settlements.  On the remote and roadless south part of the west shore, and on around to the west side of Castle Peak, only a few houses and tiny communities existed.  They were as much as five miles from a road, and were isolated and highly traditional, often without electricity or other modernizations.  By contrast, the east side of the bay, on the main road from Kowloon through the western New Territories, was densely settled.  Shops, restaurants, and temples clustered around the wharf, which extended from the middle of the east shore out a few yards into the bay.  It provided mooring only for sampans.  To reach regular boats, one hired a sampan, rowed by a boat woman (very rarely a man).  These water taxis were a highly distinctive feature of life in all the boat settlements of those days.  The sampan women were in a position to know everything about everyone, and were prime sources of gossip.

Around the head of the pier, a cluster of shops, restaurants, and daily market stalls catered to the ordinary needs of the boat people.  Above the head of the pier, on a rocky hill, stood the Three Sages Temple, the main temple for land-dwellers.  The main one for boat people was the Tin Hau temple across the bay; another Tin Hau temple, near the Three Sages, was built in the late 1960s.

Along the bay southward were the fish dealers’ huge barnlike buildings.  More will be said of them.



The standard greetings in Cantonese are heui pin tou a? “Where are you going?” and sik jo faan mei? “Have you eaten yet?”

Where did people go, and what did they do and eat there?

Well before dawn the big boats slid out of harbor.  One by one the diesel engines roared into action, or, on sail-driven boats, creaking and heavy thuds showed that men and women were hoisting the rigging and the huge five-sided sails.  The cries of sailors giving orders waked sleepers.  If the weather was fair, hardly a boat is left in harbor by 6:00 a.m.

On the boats, each person began his or her appointed tasks.  The head of the family—normally the oldest man who is still actively working—directed the whole operation.  He guided the boat around other boats, rocks, and dangerous eddies.  If the boat was big and modern enough to have a regular captain’s wheel, he was at the wheel.  The second-in-command was usually the head’s eldest son or younger brother, but often his wife or even his mother.  He or she operated the rudder.  The other adults worked with the engine or raised the sail, and put gear in order.  A child was entrusted with the necessary task of placing three burning sticks of incense in the bow for the Bow Spirit, at the side of the boat for the Sea Surface Spirit, and at the mast for the Sky Spirit.

By dawn the boat was well out of harbor.  The women prepared breakfast, the senior woman directing the others.  People drifted to the cooking area; a bowl was sent aft to the tillerman at the rudder.  Breakfast was boiled rice with soy sauce, and, unless times are hard, with steamed or leftover vegetables and some sea food from yesterday’s catch.

Then the long day began.  Fishing went on without intermission till late afternoon.  Some time between 4:00 and 7:00 p.m. (depending on season and weather), the boat was back in harbor.  The catch of the day was loaded in a rowboat and brought to the laan.  The fisherman would be paid off—usually getting very little, for the laan kept most of the money to pay off the debts that inevitably accrued, at very high interest rates, from the fishermen’s needs for loans for boats and repairs.  The fisherman would go shopping at the pier head.  The man would buy small necessities for dinner and the household, return to his rowboat, and go back to the main boat.

The women would cook dinner—rice, vegetables, sea food.  Always, some of the catch was too small and misshapen to bring a good price, and this became the featured dinner dish.  Both men and women were expert at picking out fish and shellfish that were small and unlovely but exquisite-tasting.  They were also expert at cooking them by whatever method brought out the very best flavor.  I have never had anything like the fish dishes they cooked; no four-star restaurant can touch them.  (I have given extensive accounts of local food in Anderson 1970b:33-36; Anderson 1988; Anderson 2003.)

Everyone would then relax.  They might make nets, buy snacks from rowboat-plying hawkers, or tell stories.  They would certainly talk over the day’s affairs and waterfront news.  By 9:30 all were asleep.


This was the normative day.  By 1965, it was already only a memory for many:  the impoverished refugees who had to work at odd jobs, beg, or simply sit and pray and starve.  By 1974, the vast majority of the boat people were working shore jobs, mostly in the factories that sprang up after 1960 in San Hui.  By 1999, when I last visited the area, the boat way of life was a fading, nostalgic memory of people my age.

Even in its time of dominance, it was not the universal order.  Storms kept the boats in port all too often, especially in late summer and early fall.  Emergencies—especially engine repair—did the same.  Major festivals gave people an excuse not to fish.  At New Year the boats were idled for days.  Even the industrious knocked off three days, and many of the less industrious took off two weeks.

There were also those boat people who were not fishermen.  There were cargo handlers and lightermen.  There were boat hawkers and sampan girls.  There were shoreline workers:  longshoremen, restaurant workers, entertainers, small-scale fish retailers, odd jobbers—people who lived on boats still, but did not fish.  There were even some teachers and skilled repair workers who lived on boats.

This was no new thing.  History records not only that there has been constant movement back and forth from shore to boat, but that boat people have worked longshore jobs since time immemorial.  The boat people of nearby Shanam, in the 1920s, worked largely in salt-making (Wu 1936a, 1936b, 1937) and related shore jobs.

Boat people in harbor or on shore could indulge in the greatest of Cantonese pleasures, yam cha (“drinking tea”).  This ritual, now well known in California, involves going to a restaurant to drink tea and eat snacks.  In urban restaurants, it involves whole families sitting for hours around a table, gorging on wonderful tim sam (“dim sum,” small snacks).  “Tim sam” literally means “dot the heart”; compare the English “hits the spot.”  Drinking tea on the Castle Peak Bay waterfront was a different world.  The restaurants were small, rough, and built out on piles over the water.  (Castle Peak had a “floating restaurant,” actually moored on piles, but—like other, larger floating restaurants—it catered more to tourists than to local people.)  Yam cha was largely a man’s world, and largely about drinking tea—no one could afford snacks beyond minimal ones.

The senior man of a family would have his regular table.  A senior man might have his grown sons or other family members, including his wife or other senior women, but young women and girls were not allowed.  For a community leader, his regular table was his office.  A leader would be at his table at a well-known time, and anyone wishing to speak to him would know when and where to find him.

Land and water people came together over tea, signed contracts, made deals, discussed economics and weather and pleasures, planned coming festivals.  The fishermen talked endlessly of how the fish were running, and who was catching what.  Drinking tea was thus part of work, not just a pleasure.  The Cantonese rule that deals had to be closed over drinks and food was one of those cultural rules that brooked no exceptions, and so a negotiation too small to be worth a dinner had to go on and to close over tea.

While the man drank tea, his wife went to market.  Crowds of women rode the local bus (never the fancier long-range bus, though it stopped at all the right places).  Within recent memory, shopping had been a senior man’s job, for women and young people were afraid of the prejudice and harsh treatment they got on shore.  Some still left it to the men.  Most, however, were willing to go as far as the public market in San Hui; it was safely near the bay.  However, boat women who were loud and volatile on the water became quiet and reserved.  Children clung close and tried to be invisible.  This was land people’s territory.

At San Hui, the boat people bought vegetables, hardware, clothing, and, very rarely, meat.  (More routine needs like rice, eggs, oil, and medicinal herbs could be bought on the pier.)  Such local items as papayas, roasted chestnuts, and other fruits and nuts were rare treats.

Everyone then returned to the boats, to spend the day washing, cleaning, and mending or making nets.  Children played, often swimming in the harbor.



Food was important to the boat people, not only as sustenance but as marker of social occasions and as one of life’s greatest pleasures.  The boat people ate well, and spent a great deal of time cooking, eating, and talking about food.  Like other Cantonese, they spent a good deal of their discretionary income on dining out.  This was not, however, merely for pleasure; dining had major social functions.

The ease of getting fish in those good years meant that few active fisher families lacked for basic sustenance and for high-quality protein.  Valuable fish were sold for many times their weight in rice; less valuable catches were the family dinner.

This was fortunate, for the caloric needs of the boat people were enormous.  Anyone fishing on the winter sea faced horrific conditions.  Fishers had to draw nets by hand, handle awkward rigging and rowboats, and move heavy equipment and heavy fish-baskets around the deck—all on a pitching, rolling boat in freezing wind and soaking cold spray.  Summer, with its enervating heat, frequent diseases, and constant thunderstorms, was not much better.  Under such circumstances, an adult man needed at least 5,000 calories for a long day’s work, and we saw people eat more.  I was once challenged on this by a critic who had looked up “fishing” in a table of nutritional needs and found it required only 3,000 calories.  Of course, the table-maker was thinking of recreational fishing.  I had to point out, patiently, that dangling a hook and line in a lake was not quite the same thing as facing a winter storm on the South China Sea.

One typical breakfast for an adult man facing a winter day was fat pork ribs, heavily sugared, and washed down with raw alcohol.  This tasted to the men as awful as it sounds, but it was often the only way they could choke down enough calories in the few minutes available.  Work without breakfast could easily mean death from hypothermia and exhaustion.  Lunch and dinner involved several bowls of rice, with sea food, soy sauce, ginger, and perhaps a bit of something else.

Vegetables were in short supply; they were expensive (especially in late winter) and hard to store.  The boat people knew all too well the result of going for weeks with almost no vegetables or fruits:  sores that did not heal, sore mouths, infections, and all the other symptoms of scurvy—the sailors’ eternal scourge.  Knowing nothing of vitamin C, they believed it was caused by yit hei (“heating qi”), but they knew the cure well enough; experience showed which vegetables and fruits were most effective.  They saw this as proof that the foods had cooling qi.  We noted that the foods were those highest in vitamin C content.

Serious food was divided into faan (cooked rice) and sung (prepared items to put on it).  In addition to these came ko or kuo (fruits and nuts), tim sam, noodles (min), and other snacks; children away from home were the main snackers.

Social occasions were marked by special foods, usually high in animal protein.  First in importance were the choice sea foods of the area.  These were readily available.  Castle Peak Bay had the reputation of having some of the best—the water was clean and the fish abundant.   The standard shore dinner included shrimp, crabs, and fish; the favorites among the latter were the groupers (sekpaan, Epinephalus spp.), with the pomfret second.  Other fish, as well as fish roes, cuttlefish, squid, and other sea creatures, followed in order, with the lowly sharks and rays being among the least choice.  Good seafood was kept alive in clean water.  A dead fish or shrimp was a second-rate commodity.

These foods were usually used in feasts that were primarily to entertain and interact.  Feasts for more structured occasions, especially religious ones, demanded certain specific land foods.  Chicken was the most general of these, being virtually a marker of ritual or ceremonial occasions, from weddings to ancestral sacrifices and from store openings to community gatherings.

Pork was virtually universal at all land people’s meals above the plainest level.  The boat people got far less of it (and far more sea food), but still had it on the table at every special occasion of any great importance.  However, unlike the land people, they did not so often have it at dinners that were merely fun and good fellowship.  Moreover, the land people often had large hunks of roast pork, by itself; the boat people usually stretched it by cutting it into small pieces and cooking it with less expensive vegetables.  Duck, rare and exceptional for everyone, was also rarer on boat decks than on land tables.  Finally, expensive, choice vegetables were definitely a land luxury; the boat people had few enough of even the cheap kinds.  Only at large weddings and other truly exceptional feasts did mushrooms or blanched chives or small cabbages enter the boat world.  Certain foods, such as shark fin, duck with cashew nuts, and steamed freshwater mullet, were especially expensive, and thus marked truly exceptional occasions.

A boat feast, then, was usually of sea food with chicken dishes and a dish or two of cut-up pork with vegetables.  A land feast was largely chicken and pork dishes, with some expensive sea food and some expensive, luxurious vegetables.  The actual code was far more complex than this, and one could write a book on the specifics of social-occasion marking through food in Cantonese culture (see Anderson 1988, Watson and Watson 2004).

For the boat people, fish had to be cooked a certain way.  Groupers were steamed and eaten with ginger and other flavorings.  Pomfret was steamed and usually covered with a white sauce.  Sea bass was chunked and stir-fried.  Yellow croakers went into sweet-and-sour sauce (the only time anyone made sweet-sour sauce on the waterfront, sweet-sour pork being considered uncouth).  Crabs required black bean sauce or red vinegar.   Fresh shrimp were generally simply boiled, or stir-fried with various flavorings.  People were surprisingly conservative about these methods.  A diner choosing to cook a fish in a non-canonical way had to explain why he had “wasted a perfectly good fish.”


Patterns of clothing in rural south China distinguished different ethnic groups.  The boat people wore loose black cotton pants and shirts, and mushroom-shaped hats made of hard bamboo slips or palm leaves.  The black cotton cloth was light but incredibly tough, withstanding salt water and heavy wear.  In winter, people adopted the layered look, with thin undergarments and wadded jackets.  For formal occasions, dark blue often substituted for black.

In the mid-1960s, most people still wore traditional clothing.  By the 1970s, cheap western-style clothing had become rather common.  It had to be very cheap to compete, because the old-fashioned outfits outlasted western-style garments by months or years.

Children wore brighter colors, and young ones added many protective charms, including tiny models of floats.  Most women and many men wore jade bracelets, because in an accident these would shatter instead of one’s arm bone (a religious belief, not a scientific fact; most people could recite counter-cases).


Recreation usually involved socializing, eating, drinking, and relaxing.  The young engaged in horseplay—chasing, swimming, pickup ball, and miscellaneous energetic activity.  Land people included in their stereotype of boat people an odd belief that the boat people could not swim.  Anyone who visited the bay on a warm sunny day could see the error of that.

More serious time-killing usually took the form of gambling; older women in particular played mah-jongg and related games by the hour.  A standard way of saying someone was living the easy life was to say “he/she can play mah-jongg all day.”  Mah-jongg was a noisy game, usually played on a table of thin, resonant wood.  It was an older adults’ game; playing it was a privilege of seniority.  Younger adults did not play.  No one won or lost much money; now and then someone would win HK $100, but this would be lost again in a few days.

On the other hand, the boat people were largely a sober lot.  Robert Antony in Froth Floating on the Sea (2003:140-150) provides lurid accounts of the activities of South Chinese pirates—boozing, whoring, gambling, and otherwise acting like pirates.  (Those of the Caribbean, however idealized by Disney, were, in reality, the same or worse.)  The boat people of Castle Peak Bay told the same stories of such rough elements, but they were quick to distance themselves from such lifestyles.  Their difficult life, demanding not only hard and dangerous work but also extreme thrift, honesty, and mutual aid, allowed very little recreation.  Almost all profit went to pay off loans.  True, what little was left very often went to gambling and alcohol, but that was more about maintaining social ties than about debauchery.  Tiny sums of money, a few days a year, hardly permitted debauchery.  Even the relatively well-off could not indulge often.  The boat people carefully calculated feasting and drinking, and gambling to a lesser extent, to maximize social ties.  These activities were not mere amusement.

Music delighted everyone.  Radios, festival music, instrument solos, and folksinging all made the waterfront a vibrant, wonderful cacophony that could be heard a mile away.  During the day, one was usually able to listen to several radios, each playing at top volume, each tuned to a different station—and often someone was singing over it all.

Well into the 1960s, the boat people were known for their “salt water songs” (haam seui ko; Anderson 1975; Wu 1937). These songs had been famous throughout south China for centuries.  These included some traditional folksongs and poems (with set texts), but most were freely improvised, blues-like sequences.  Singers, usually young, displayed great musical and poetic ability.  Singing was a typical feature of courtship (usually of the “light love” variety), but salt water songs also included funeral laments, marriage songs, complaints about hard life, and simple time-killing songs for dull hours on the water.

In 1965, when I began to record salt water songs, many people knew a rich selection, but radio music had already made great inroads.  By the 1970s, salt water songs were dead, and only a few people remembered them.  One of the greatest folk song forms in the world, a tradition fully comparable to American blues or Scottish ballads, had disappeared without trace.  To my knowledge, my recordings, done on a wretched old tape recorder that barely played, are all that survives (the tapes are deposited in the Hong Kong Museum).

The salt water songs were a true expression of boat culture: beautiful, intense, personal, and usually highly improvisational.  They combined classic themes and images in ever-new ways, the better to express deep truths in a language that was at once highly metaphoric and strikingly direct.  They ranged from gloriously ribald to delicately sensitive.  They treated the full emotions of the singers, revealing much that could never be spoken.  They focused boat life and experience into a single rich form.  They were one of the great achievements of the human spirit, and they should never have died.  Nothing like them will ever appear again.



Predominantly fishermen, the boat people had an amazing knowledge of marine life.  Dr. William Cheung, the Stanford-trained fisheries expert who generously provided most of the identifications for me, had kept a running record of the knowledge, and we estimated that the boat people of Hong Kong knew over 1,400 named taxa of marine life.  I recorded about 400 names at Castle Peak Bay.  These covered everything from seabirds, sea snakes, and sea turtles to shrimps, sea cucumbers and even humbler bottom fauna.  (All the names can be found in Anderson 1972; I will stick largely to English here.)

Like most people, they recognized life-form categories:  birds (jeuk), snakes (se), turtles (kuai) fish (yu), decapod crustacea (ha, including shrimps, prawns and lobsters), crabs (hai), and snails (lo).  These were basic large divisions of the animate world.

They also had a range of smaller but comparably basic divisions—small life-form categories, or merely the residual taxa that Berlin called “unique beginners.”  (No one has ever really solved the question of how to regard these terms.)  These included Cantonese categories exactly equivalent to English labels: clams, squids, cuttlefish, octopi, sea cucumbers, oysters, roaches, horseshoe crabs, starfish, jellyfish, and sea urchins.  Except for squids with a very few categories, these taxonomic units were not further subdivided.  There was only one marine type of each.  (The roach had a land exemplar, however; water roaches—seui jaat— were crustacea, Lygyda spp., but seen as marine forms of the cockroach, kak jaat.)  Porpoises and dolphins were often called “sea pigs,” which would make them another sea exemplar of a land category.

Cantonese had no categories equivalent to “shellfish,” “crustaceans,” “mollusks,” or “cephalopods.”  The boat people simply saw no reason to lump highly disparate creatures under broad headings of that sort.  English keeps “squids” and “cuttlefish” separate, as Cantonese does, though the learned scientific term “cephalopod” has gained some currency.

One category somewhat problematic to me at first was lung, “dragons.”  The boat people knew, and claimed to see on occasion, three or four types of these.  Most of these remained invisible to me, but one, the cham lung (“sinking dragon” or “deep-diving dragon”), was quite visible; it was the sturgeon.  The tentacles and distinctive scales of this strange, primitive fish made it seem more dragon than fish, though it was uncomfortably intermediate between these two categories (see Chapter 7).

In general, Chinese classification is strikingly similar to English, and thus very different indeed from many classification systems around the world.  Not only does it have more or less the same structure and categories; it is also rather complex, with many terms at each level.  It has many folk species, for instance.  This contrasts sharply with North American Native systems, including the Haida, Nuu-chah-nulth, and Yucatec Maya ones I have studied.  To my knowledge, all Native American systems are very heavy on folk generics, very light on life-form categories and folk specifics.  They are “shallow” taxonomies.  Apparently it is easier to learn a simple list of names until the list becomes overwhelmingly large (Hunn 1990).

I suspect the deeper and more complex Chinese system results from China’s long history as a mercantile civilization.  I think that this history has produced particular pressures to name certain beings in certain ways—as a similar history certainly did in Europe.  Plant names can be more easily traced than fish names in Chinese history (Anderson 1991), and thus provide a valuable comparison.  Sources from the earliest period of Chinese history show a Native American type of plant taxonomy, with many one-word terms comprising folk generics.  Folk specifics were added steadily over time, as the Chinese discovered more and more plants that were obviously related to familiar plants.  For instance, the familiar willow lent its name to the Near Eastern pomegranate, which arrived in the early medieval period and was labeled “stone willow.”  Then the Portuguese brought the guava from South America in the 16th century; the guava was called the “barbarians’ stone willow” because of its similarity to a pomegranate.  The Portuguese were “barbarians” not just because of Chinese prejudice, but also because of less-than-lovable behavior; more than a few were smugglers and swashbucklers.

Brent Berlin points out that people classify under one terminological head those organisms that appear to make up a “natural” group to members of the culture (Berlin 1992).  Berlin, like Scott Atran (1990), tends to think such groups are natural—they are typically biological groupings, given by shared evolution and genetics, and people see them as such.  Where the folk are clearly not doing this, as in classing whales with fish, it is because of understandable mistakes; numerous similarities link whales and fish.  The boat people realized that porpoises and whales are “like pigs” internally.  They classed them with fish on the basis of surface similarities, but also called them “sea pigs.”  As in the case of the sturgeon, this appearance of belonging to two opposed basic categories made these mammals anomalous, and thus uncanny (see Chapter 7).

Other anthropologists, especially those who have studied systems radically different from English or Chinese, see taxonomy as much more arbitrary, much less grounded in “real” biology (e.g. Ellen 1993; Forth 2004).  Obviously, the anthropologists’ experience matters here[1].

Looking over world systems, one cannot miss a universal tendency to “carve nature at the joints” (as taxonomist folklore has it)—to see real biological divisions.  Such divisions tend to make sense; they capture useful truths.

However, when other, even more useful truths are captured by a quite different classification system, people may ignore whatever joints nature may have.  For instance, some systems noted by Ellen and Forth class bats with birds, but class flightless birds as something else entirely.  This makes perfect sense if flight is a key characteristic but other obvious biological markers (feathers, beaks, etc.) are not.

More interesting are basic differences that these and many other systems make between spiritually powerful creatures and other, lesser ones.  We shall see below that the boat people have such a division, but they do not make it part of their basic taxonomic system.  Many other peoples do.  However, Native North American groups such as the Haida made a far more serious issue of spiritual-power classification.  One cannot understand their system without knowing both local biology and Haida religion.  This did not keep them from having an enormous and biologically very sophisticated taxonomy and knowledge base for marine life.

Even scientific taxonomists provided us with some groupings that seemed reasonable, but turned out to be genetically far off base.  Fungi were included with plants (they are genetically closer to animals).  American vultures were lumped with African ones (American vultures are storks; African vultures are close to eagles).  So, when a society finds it expedient to recognize ceremonial value, fightlessness (of birds), or stem shape (of plants) as basic, we should not be surprised, even if it means “carving nature at joints” totally different from the joints recognized by scientists or folk English speakers.

The boat people found it to their advantage to develop a classification system close to the scientific one.  This interested me, and I asked them to explain.  They said that not only do all sharks, all groupers, all squids look alike, they also act alike and often taste alike.  If you want to catch new species, you can make certain predictions based on the behavior of the ones you know.  If you want to sell them, you can expect to get about the same price, or at least be in the same pricing universe; sharks are very cheap except for the valuable fins, groupers are all valuable, squid are in between.

Broad, basic terms of the sort described above were broken down into what my teacher Brent Berlin calls “folk generics.”  These are the basic terms of reference to animals and plants:  one-word or two-word, easily-learned names that can be applied to a group of very similar organisms.  These are sometimes broken down into “folk species.”  In American English, “trout” and “shark” are folk generics.  Folk species within these categories include brook trout, brown trout, rainbow trout, white shark, sand shark, and so forth.  Normally,  a folk species name is made by adding an adjective to a folk generic.

Such categorization is dynamic.  Most Americans have “salmon” as a folk generic, with chum salmon, pink salmon, and so on as folk species.  These Americans see salmon as very similar, and probably very few can distinguish the species.  By contrast, the Northwest Coast Native peoples never recognized such a generic.  They have always seen the many species of salmon in their waters as totally different fish.  They find it preposterous that anyone should lump them.  This pattern has spread to Anglo settlers, who now use “chum,” “pink,” “sockeye,” and so on as folk generics, and use the word “salmon” only when referring to aggregate economic data and the like.  They laugh at tourists and other ignorant folk who cannot instantly see that a sockeye is totally different from a pink.  (This paragraph is based on my field work in British Columbia, 1984-1987.)

The boat people’s fish terminology turned out to be as neat, straightforward, and well-organized as kinship, but there were huge differences in people’s knowledge of it, leading to smaller but real differences in people’s classification systems.  Some were experts in one thing, some in another.  Trawlermen knew all sorts of small, scruffy marine items that hook-and-line fishermen ignored.

Hook-and-liners knew the fine distinctions between the many grouper species (or rockfish; sekpaan in Cantonese; to scientists, Epinephalus spp. and close relatives).  Trawlermen just called them all “red” or “green” or “spotted groupers.”  This was one case where one could see why nature was carved at the joints.  The hook-and-liners explained to me that each sekpaan has its own behavior:  its own favored habitat, depth, wariness of the hook, favorite bait, and so on.  They could give excellent species-specific accounts of these matters.  The trawlermen caught groupers in huge net hauls, and did not have any occasion or reason to know the differences between different species.

In all cases, folk generics corresponded to scientific taxonomic units, but the units varied in size.  Sekpaan corresponded almost perfectly to the scientific genus Epinephalus, but sa yu, “sharks,” corresponded to several families—all the fish we call sharks, hammerheads, and guitarfish in English.  The boat people were aware of real biological relationships, but they were not concerned with making folk genera comparable in extension.

They were equally well aware of real biological differences.  For instance, they recognized that the almost exactly similar mudskippers and walking gobies were quite different biological entities.  The hook-and-liners recognized different species of sekpaan that biologists could barely distinguish; the hook-and-liners knew, moreover, that one species occurred on very rocky and rugged bottoms, while another virtually indistinguishable one preferred pebbles and sand.  Marine biologists, if they knew this, knew it only because they had learned it from the fishermen.

Clearly, Berlin is right about at least some folk biologists.  They do know “nature” very well indeed, and know how to carve it at joints that are biologically real .

Folk specifics usually corresponded closely to biologists’ species, but sometimes a folk specific lumped a whole set of biological species, especially in the case of small organisms whose biological species look much alike.  Shrimps and small, worthless fish were the main examples.

Unlike the Maya (whom I have worked with since), the boat people had many folk generics divided into folk species.  In addition to groupers and sharks, there were species of rays, croakers, soles, seabass, and many other common types. They coined many names that were, or functioned as, folk specifics by adding adjectives to life-form categories:  “spotted shrimp,” “dragon shrimp” (lobster), and so on.

Turning a term like sa yu into a folk species term involved dropping the life-form category yu and adding an adjective, as in pak sa, “white shark,” apparently identical to the western white shark (Charcharodon carcharias).


This biologically-grounded taxonomic system was the basic system—what Berlin calls a basic or general-purpose taxonomy.  It was the one always used when what one wanted to do was simply give a name to something.  It was the system always used to supply a name when I asked “what’s that”?  (Not, you remember, “what’s its name?…”)

However, the boat people used other ways of classifying fish.  These were all ad hoc, functional systems.  They cross-cut the biological taxonomy.  I recorded the following:

–Price.  The price mattered a great deal.  So everyone knew which were the high-priced fish, and even the exact prices that the better ones brought from the dealers and in the markets.  Many were worthless, or good only as “duck feed fish.”   Some fish were good only for salting as haam yu “salt fish.”

–Catching method:  trawler fish, hook-and-line fish, gill-netter fish, and so on.

–Life cycle stages.  Amazingly common and uniform around the world is a tendency for fishermen to divide common fish into life cycle stages.  In English, for instance, salmon are fry, then parr, then smolt, then finally mature salmon.  Eels begin as elvers.  Similarly, for the boat people, tiny fry were “flowers” (fa), and there were size terms for sole and a few other species.

–Life type.  Oysters were differentiated by where they grew:  rock oysters, tile oysters, and so on.

–Metaphor.  Very rarely, a human was called a shark or a dragon.  One might have thought the boat people would more often use marine metaphors, but they did not.

–Religious importance.  Several species were san yu, “sacred fish.”  This very important category will be discussed in Chapter   .

These ad hoc functional schemes were never used to produce ordinary conversational or formal names.  They arose only in context.  On the other hand, they were important, formalized, and frequently discussed—especially the price system.

The boat people certainly recognized that the basic taxonomy was grounded in nature—in the reality of the fish and their obvious relationships or similarities.  They recognized that this was a totally different classification method from that used in the cross-cutting systems, which were based entirely on human concerns.  The boat people thus paralleled the thinking of Claude Lévi-Strauss (1962) in his differentiation of “scientific” thought and “wild” thought.  Scientific thought tries to understand the nonhuman world in its own terms.  Wild thought understands the nonhuman world in strictly human terms, projecting human categories onto it, seeing it only as a site for human use-values and metaphors.  The boat people’s sharp distinction between the basic taxonomy and the ad hoc cross-cutting systems closely parallels this division.

The Yucatec Maya, whom I have studied since, have a similar distinction, but it seems less sharp.  In fact, recording the related Itzaj Maya language, C. A. Hofling and the Itzaj Maya scholar Felix Tesucún (1997) do not distinguish a basic from ad hoc, special-purpose systems.  They provide only one taxonomy, incorporating functional, human-centered categories as part of the whole system.  The human-centered terms are subcategories within a basically nature-centered taxonomy.  Since one of the two is a fluent native speaker, this must be psychologically real to the Itzaj.  It would certainly not be psychologically real for the Yucatec, who do distinguish basic from special-purpose taxonomies.  However, even among the Yucatec, the division is not as clear as among the Cantonese.

The whole nomenclatural system provides a wonderful simple case of a cultural classification system.  The terms hang together neatly, are organized in sharply defined cross-cutting categorization modes, and apply quite precisely to the natural world.  People use the terms in a flexible, adaptable, but straightforward fashion, to talk about fish.  Practice enters, especially in the cross-cutting ad hoc systems.  These change rapidly according to felt needs in interaction and discourse.  On the other hand, the basic structure of Chinese fish taxonomy has remained the same for thousands of years, as we know from classical texts.

This makes a very neat model of culture:  long-lasting structural framework, shorter-term changes and adaptations, and instant accommodation to the much more fluid, rich, complicated world of actual discursive practice.  (On the general theory all these matters see Hanks 1990.)  This may be one of the most important theoretical conclusions I learned from my boat friends.

It also makes clear the degree to which traditional people know their environment.  The boat people’s system was biologically sophisticated and knowledge-rich.  I could give only a brief summary of their incredible working knowledge even in my more detailed books, and will not even attempt it here.  Suffice it to say that one could easily fill an encyclopedia with Cantonese fishermen’s knowledge of sea life, to say nothing of ships, winds, waters, and the rest of their environment.

Comparing the boat people’s system with modern biological knowledge helps appreciate how much they knew, and also what we can learn from it, but the comparison is also interesting for what it tells us about human thought.  This being the case, we need comparisons of the boat system not only with the biological, but also with the Maya, the Haida, the various systems studied by Ellen and Forth, and, last but far from least, with classical Chinese fish lore.  Such comparisons will have to await further research.

The currently popular myth that “indigenous knowledge” is “culturally constructed,” and thus not similar to or close to modern biological knowledge, does not stand up under investigation.  Most fishermen’s knowledge of marine environment and biology was scientifically impeccable, and modern marine biologists learned a great deal from the boat people.  The boat people sometimes phrased it differently, but it could be translated into bioscientific terms with ease.  Cultural classification systems are, of course, culturally constructed (by definition), but that does not mean they are out of touch with reality.  Many anthropologists seem to think that there is an opposition between cultural construction and “reality” (whatever that is).  Of course not.  Culture is adaptive; it is the knowledge we need to live our lives.  It has to be realistic.  It is never a perfect reflection of reality; that would require perfect information, and humans never have that.  Culture provides useful knowledge, not perfect knowledge.

Yet, equally interesting, and common in China as elsewhere, is the combination of highly perceptive and expert biology with a number of beliefs that seem flagrantly and obviously counter-evidential.  Dragons and sacred fish were important to the boat people.  The boat people sometimes saw dragons writhing in the clouds of a typhoon or violent storm; fear made the vapor seem malignantly alive. Classical Chinese biology—like folkbiologies around the world—includes many other strange creatures, ranging from phoenixes to crane-riding Immotals.   The sacred fish were real enough, but were fish that seemed anomalous in some way, and thus uncanny; they were thus out of the normal realm.

Very often, such “mistakes” arise by logical extension of reasonable beliefs that approximate reality.  I have described elsewhere (Anderson 1996) how Chinese ideas of nutrition and of landscape evolved naturally from correct observations.  Such beliefs as are incompatible with modern bioscience were logically deduced from correct observations; the logic was imperfect only because the observations were limited by the investigative and theorizing capacities of a premodern society.

We need more and better analyses of how such knowledge systems arise, and how they serve to mediate between people and environment.



To live, the boat people had to make money.  Chinese society has been market-oriented for millennia, and much ink has been spilled over the question of why China did not develop capitalism before the west did.

The boat people had an ambiguous view of money.  This showed clearly in a proverb they very often cited:  “Money is like dung; honesty and goodness are worth a thousand gold.”  Money is dung—but is the measure of value and goodness, too!  The ambiguity is, however, less real than appearance suggests.  In fact, money per se is like dung; misers and penny-pinchers got no respect.  Money used for sociability, community, and mutual aid was a different thing.  Using money thus was the very soul of honesty and goodness.

The proper boat person, especially the male, thus immediately converted any spare wealth into social capital (a phrase I used in my original Castle Peak monograph, before Bourdieu made it famous [Anderson 1970b:83]; maybe I should claim credit).  He throws feasts, supports festivals, cultivates friendships, enjoys himself, and helps others out.  Such a man is gaining min, “face,” of which more anon.  Actual behavior ranged from almost complete indifference to money to grasping skinflint attitudes, but most boat people, especially leaders, stood close to the ideal.

Poor people had mixed standing.  Unfortunate or hard-working ones were pitied.  Lazy, dishonest, or incompetent ones were despised—not so much because they were immoral as because, in the words of one land-dweller, “the poor can’t help you.”

Rich people did not get automatic respect.   They would have to deal with gossip to the effect they got their wealth dishonestly, especially if they did not spread it around.

The ideal, then, was to get money industriously; grasping or too-clever tactics were acceptable.  But, however the money came, one should share the good fortune with as many people as possible, and in as helpful a way as possible.  Helping with religious festivals was perhaps the most prestigious.  The believers saw these festivals as the best hope for guaranteeing good weather, good health, and good luck to the whole community.  The skeptics saw that at least the festivals brought the community together as nothing else could.

Traditionally, the boat people of the western New Territories made their living on the water and from the water.  Most of them were fishers (for full detail on fishing, too specialized to cover here, see Anderson 1970b, 1972).  The rest made their living as cargo-carriers, longshoremen, fish market workers, and other shoreline occupations.  A large number, often women or older people with limited options, made their living carrying passengers in sampans.  Many others lived as boat-based vendors.  They rowed among the moored fishing boats during the evening, selling everything from vegetables and oil to firecrackers and incense.  Like vendors in Elizabethan London, they had beautiful, musical calls to advertise their wares; these calls were among the most characteristic and nostalgic sounds of quiet evenings on the water.  Cries were more numerous and varied in the past, said the fishermen, who added that they deeply missed those old sounds.

At Tai O, but not at Castle Peak, some boat girls made their living by singing songs and otherwise entertaining young men.  Sex was not necessarily involved, nor was it excluded.  No one considered these girls to be prostitutes, and no stigma attached to their occupation.  Quite the reverse; by singing folk songs and sometimes classical Chinese poems, they brought highly valued cultural delights to the waterfront.

Within recent memory, smuggling and piracy had also been important occupations (Antony 2003; Murray 1987).  Piracy was long extinct in 1965, but smuggling opium and running refugees out of China were still side occupations for a few fishermen.  I had no opportunity to investigate this, for obvious reasons, but I was told by reliable friends that it was rare at Castle Peak.  Most such activity was concentrated in the bigger, more urban ports.

At Castle Peak, all these nonfishing boat people, whatever their occupation, lived on sampans or on stationary craft.  At Tai O, most boat people lived on houses—built and maintained like boat cabins—on pilings over tidewater.  The tidal “streets” left between rows of houses became mud at low tide, neither navigable or walkable.  These streets were known as chung (no character known), a word that sounds to me suggestively like the Thai equivalent klong.  (There is a theory that klong is cognate with jiang, the traditional Chinese name of the Yangzi River.  Central China still has some Thai speakers, and, as noted above, the Hong Kong area was probably Thai-speaking—in the broad sense of the word “Thai”—within early historic times.)

Life on the old boats was wonderful, vibrant, and fascinating.  It was even beautiful, for the boat people kept their boats clean, varnished, and shipshape.  The old craft were striking enough to show up in every tourist photograph of Hong Kong—even long after they had disappeared except as photographers’ props!  Many of the furnishings—jars, tackle, pans, altar sculptures—were true works of art.

By 1965, unsuccessful fishermen and poorer refugees from China were moving out of the fishery to work in the factories, shops, and cafes that were opening in San Hui.  No one we talked to did this voluntarily; all were forced into it by being unable to raise capital to maintain a boat.  After that time, decline set in, until the fishery ended.

In 1965 and 1966, the fishery still employed the largest number of people of any activity around the bay.  Some 400 species of marine life were more or less commonly found, and another several hundred potentially occurred somewhere in the area.  I recorded about 100 species commonly taken and sold at Castle Peak Bay.  Most important commercially were groupers, soles, croakers, goldenthread, and shrimp.  Pomfret, crab, and flathead made up a more specialized group of species.

Huge old hand trawlers—magnificent sailing-junks from another era—dominated the skyline.  The newer, sleeker, diesel-powered trawlers, net-hauling booms lowered for fishing or raised out of the way, moved around them like wolves among bears.  Trawlers drag huge nets over the bottom, or close to it, and catch mostly trash fish and larger but less valuable species; the slow hand-trawlers specialized in catching vast masses of tiny sea life used for duck feed.

Below and around and among these were the small gill-netters that stretched long curtains of nets from the water surface, blocking the paths of migrating schools of fish.  They made most of their money in spring, when the yellow croakers (Pseudosciaenia crocea) ran north to breed in the Yangtze area and the China Sea (on this fishery, see Chu 1960).  By 1975, the yellow croakers were declining, victims of pollution and overfishing.  Gill-netting was a dying art.  Sweet-sour yellow croaker, one of the superb signature dishes of Castle Peak Bay restaurants, was almost a thing of the past.  A few purse-seiners, similar to gill-netters but using smaller nets that surrounded fish, worked the shallows; they caught small amounts of small fish.  This was a chronically low-paying fishery; purse-seining paid much better in the eastern New Territories.  (Barbara Ward’s community, Kau Sai, in the east, was largely a purse-seiner port).

Smaller still were the hook-and-line boats that specialized in catching high-quality larger fish for the fresh-fish market.  Some of these boats were no more than sampans, but they had cabins and housed families.  Many of these were not mechanized.  The ancient five-sided mat sail, amazingly adaptable for sailing close to the wind, held its own among the aging bus engines and other retooled motors that drove most of the fleet.  One reason was that the engines constantly broke down, and the trouble and expense of repairing them ruined many a family.  Therefore, less affluent or more traditional families avoided them.

In the fresh fish market, hook-and-line fish took pride of place, since net-caught fish were dead on arrival.  This allowed the tiny hook-and-lining sampans to make a living.  What they lacked in catching power and range they made up in value of product.   Some species fetched many times as much as others, and fashions tended to change; the fishers had to keep up on fickle tastes.


The number of people on a boat was a function of its size.  The hand-trawlers had, by my census count, up to 47 people living on them.  I heard of junks with 60 to 80 residents.  Often there were passengers as well.

One other type of boat was the houseboat.  In addition to old wrecked trawlers and other craft still used for housing, there were many houseboats, some large enough to serve as schools and meeting halls.  They had large square cabins, and could not travel; they were anchored in place, except when ponderously towed to shelter during typhoons.

One important type of houseboat was the live fish boat.  Several of these lined the southeast shore of the bay, where they sold live seafood to gourmets.  Tied to the boats were small, low barges with the bottoms replaced by chicken wire; sea water circulated freely for the fish, shrimp, crabs, and lobsters.  (Such boats are known as “well-smacks” in England.)  The live fish boat owners lived in their small houseboats, surrounded by these well-smacks, and bought the fish from specialized small fishermen (mostly hook-and-liners) for resale to the connoisseurs.

Life on the boats was not necessarily poor or uncomfortable.  Medium-sized ones had chickens and a dog or cat.  Big ones might add ducks, geese, and even a pig for New Year.  Flowering plants and medicinal herbs in pots graced the stern, and many sterncastles were magnificent floating gardens.


The fishermen sold most of their catch to the fish buyers.  Here and elsewhere, when talking about sales and economic interaction, I use the term “fishermen,” because dealings with the buyers and similar powerful shore folk were invariably carried out by the senior men.  The women had a full share in decision-making and fund allocation on the boats, and did their share of the fishing; indeed, the boat world was the least patriarchal in Hong Kong.  But land people, especially big businessmen, were dangerous; women avoided them.

Castle Peak Bay was especially interesting as a last holdout of the Imperial Chinese merchandising institution of fully private fish-buying.  Elsewhere in Hong Kong, and of course in China, government had stepped in to market the fish.  Only at Castle Peak did the fish buyers still control the industry.

The fish buyers’ dark, barnlike structures lined the southeast part of the shoreline.  The sheds, and by extension their owners, were known as laan (“pen,” and by extension “dealing place, auction shed”).  The backs of these structures projected over the water, raised on high pilings.  Boats came in and offloaded fish in baskets, which were carried up flights of steps (underwater at high tide) to the main shed floor.  The other end of the laan opened on the main road. These dealers—there were 18 of them in 1965, fewer in 1974—bought fish directly from the fishermen, packed and graded them, and resold them to middlemen from the cities, who came out at evening to buy lots of fish.  These they would put in baskets and leave on the curb, there to be picked up in the wee hours by the dealers’ trucks.  These city buyers, in turn, sold the fish to retailers in the urban markets.  Caught between the powerful laans and the sharp urban retail dealers, and forced to drive trucks in the wee hours, these city buyers did not have an easy life.  The laans took 6% of the city buyers’ price for themselves, remitting the rest to the fishermen—but only after debt repayments had also been extracted.  Debt repayments were calculated such that the fishermen’s usual margins were barely enough to buy fuel and food.  Anything beyond that required borrowing still more from the laans.

Laans preferred loyal and honest customers, and gave much better terms to such, to hold them.  On the other hand, they quickly stopped all credit to deadbeats, driving them out of the fishery.  The laan owners knew the credit-worthiness of every established fisher family on the waterfront.  Newcomers had a hard time of it unless they could establish a reputation.  Conversely, a fisherman with a reputation as both a good fish-catcher and a thoroughly honest man could get special terms from the laans.

The laans were cordially hated by the fishermen, who uniformly accused them of colluding to fix prices.  They also supposedly gave short weights; a proverb said that “a laan’s catty is eighteen ounces.”  (The normal catty has sixteen ounces.  This proverb was not restricted to the waterfront—it was a “frame text,” and producers all over Hong Kong used it to describe purchasers of every stripe.  But it was not a hollow figure of speech.  I personally saw some rather shady weighing carried out; see King 1955 for a study of pricing in a more free-market boat community).  Laans often sold fuel and other needs, and, like many other merchants on the bay, they charged top prices[2].   Also, the laans were not friendly to boat people.  In 1965, one lost his temper and said “I can whip you dirty Tanka like dogs.”  The fisherman sued him for abuse, and won, but that was something that could not have happened even three or four years earlier.

Moreover, the laans were the fishermen’s source of capital, and charged high interest rates—theoretically 20% per year, actually higher, though terms were flexible (they did not want to put fishermen out of business permanently).  In practice, fishermen in debt to laans almost never escaped.  The interest rates were too high, the capital needs too great and too oft-recurrent.  Engines, in particular, constantly needed fixing, and the boat people—unlike some groups I have worked with—were not “natural-born” mechanics.  They tried to repair their engines, but rather than working with the cool efficiency of the people I know in south Mexico, the typical fisherman sweated, cursed, cut corners, became angry and threw bolts, tinkered without much enthusiasm, and eventually had to buy a new engine.  Even without this, expenses constantly added up, and prices stayed low.

A fisherman could not normally change laans.  The laans had a strict rule that they would not pick up another laan’s fisherman, unless said fisherman was thrown out by his usual laan.  Since no laan would do this unless the fisherman was a deadbeat, fishermen did not switch laans.  Once in a great while an extremely lucky and honest fisherman could get away with a switch, but normally this was impossible.

The fishermen therefore pressured the government to put in a free auction market, which was done around 1970.  By 1974 it was in full operation, but was not as revolutionary as hoped.  The laans gave slightly lower prices than the market; however, they made up for it by giving easier credit to well-known clients (known to be good risks) than the government loan programs did.  Moreover, the laans were only too happy to loan money for weddings and other festivals, which the government found too frivolous.  Therefore, at first, many fishermen stayed with the laans.  This slowly changed, but before the laans could die out, the fishery ended.

In the 1940s and 1950s, the British colonial government became convinced that producers’ cooperatives were the best way to develop small-scale rural production in the colonial world.  Producers would join coops, pay a small fee for handling fish, and also pay 6% of their income into the general fund.  They could then borrow such money as was available, and pay it back at low interest rates.  Gradually, coops could get larger and better off, if managed well.  Throughout the British Empire, they became something of an article of faith, but, alas, the sun set on them when it did on the Empire itself.

In Hong Kong, the movement to cooperativize farmers and fishermen was most identified with G. A. C. Herklots, a biologist and agronomist.  Herklots had written a useful book on fish, which we used in the field both for reference and to study boat people’s ability to recognize the fish from the black-and-white drawings (Herklots and Lin 1961).  Among other things, Herklots saved many lives by managing to grow vegetables while interned in a Japanese death camp in Hong Kong in World War II (Herklots 1972; Herklots and Lin 1961).  By 1965, many farmers and fishermen were in savings-and-loan coops for developing production, and in “better living societies” (coops for developing housing and other amenities).

Such activity was only beginning at Castle Peak Bay.  The government had been slow to move in, and the laans were implacably hostile.  Fishermen who went with the coops found the laans stonewalled them when they requested credit.  This kept the coops poor (very few fishermen could join), and thus kept the threat credible.  By good fortune, we were in Castle Peak Bay when things turned around.  Thanks to a great extent to the tireless and self-sacrificing work of Choi Kwok-Tai, at first clerk and later director of the cooperative office at the bay, the coops finally turned a corner, and acquired enough successful members to have significant money to loan.  Fishermen—again, only men engaged in coop business—rapidly turned away from the laans.  Between coops and the new market, the laans found themselves badly squeezed, and they were a shadow of their former selves in 1974-75.  Even the advantage of providing loans for festivals was not so significant, for secularization and declining fish catches were combining to reduce festival activity[3].

Fishing was a living; the Hong Kong market was almost infinitely large, exports were developing by 1974, and Cantonese love good fish better than all other foods.  However, fishing did not pay well.  In general, a fishermen either caught great amounts of poor-quality, low-priced fish (hand-trawlers, purse-seiners), or small amounts of high-priced ones (gill-netters, hook-and-liners).  A top-value grouper weighing a kilogram brought as much as a ton of fish sold for duck feed.  Only the power trawlers had the chance to make a real killing, as when one man got impossibly lucky and caught a whole shoal of shrimp at one shot—“netting” $30,000 Hong Kong dollars, almost all of which went to pay off his debts.  But even the power trawlers barely survived, because of the high costs of fuel and engine repairs.  Also, China charged taxes and license fees for fishing on their waters (or fines for poachers—poaching was common), and these could be substantial.  Above all, repaying the laans for loans took most of the profits, unless one was in the coop.

Worse was the steady decline of the fish.  Pollution and overfishing eventually destroyed the Hong Kong fishery.  From 1965 to 1975, the loss was spectacular; the bottom had dropped out of the fishery.  Local seas were virtually sterilized by chemical wastes from industries.  Migrant schools—even the vast runs of yellow croakers, streaming north in millions to breed near the Yangzi mouth—had been fished to extinction by boats all along the China coast.  The high-seas fishery had been exhausted everywhere within reach of Castle Peak boats; fish for the markets of Hong Kong now came from farther away, from the waters reached by the huge, powerful boats out of Aberdeen and out of Guangzhou.  By 1990, there was virtually nothing left in the waters once plied by the Castle Peak fleet; a few boats caught minnows in the polluted waters.  The few surviving Hong Kong fishermen ran out to deep water, in the South China Sea and beyond, there to catch fish that sat on ice (or, at best, stayed alive in dirty tanks on board) for hours or days before reaching the restaurants.  This, of course, caused sorrow among old-time gourmets.  There is certainly no comparison between the fish of Castle Peak Bay in the old days and the so-called “fresh” fish of modern Chinese restaurants.

The lack of fish conservation stood out as an anomaly in rural Hong Kong.  Management of the rice agricultural system was incredibly careful, precise, and fine-tuned (Anderson 1988; Anderson and Anderson 1973; Marks 1998).  The system was sustainable over the long term (though probably not indefinitely), and involved conservation and recycling of a highly sophisticated order.  It was very hard on the forest environment and the large animals therein, as Robert Marks (1998) and Mark Elvin (2004) have pointed out, but otherwise it was a model of sustainable agriculture, and as such has much to teach the world (Anderson 1988).  A fully integrated part of it was the exceedingly sophisticated management of pond aquaculture and of oyster-farming in shallow bays.  So, fisheries management and conservation was no new or radical idea.

In Malaysia in 1970-71, we studied an area where Chinese fishermen had parceled out the shore and shallow waters, and enforced their community fishery space with savage thoroughness, burning boats and killing fishermen who poached (Anderson and Anderson 1978).  The Malaysian government ended this system—it was too violent, as well as being too much an independent and uncontrolled regime in a centralizing nation—but did not substitute anything effective in its place, so the fishery collapsed.  However, the system worked very well indeed for a long time, showing that Chinese fishermen are no less able than others around the world to establish sea tenure.

Why, then, did the south Chinese boat people not do this?  The answer is simple:  their drifting and politically powerless life.  Under modern governments as under the Qing Dynasty, they had had no political power, no way of establishing ownership, and no way of enforcing communal tenure.  Moreover, they had had to move at a moment’s notice from port to port, depending on the political situation.  Subjected to erratic clearances from the shoreline in Qing (Antony 2004; Murray 1987), and to chronic violence and governmental change in the 20th century, they had rarely had opportunities to develop stable communities, let alone conservation.  Even the stable communities, such as Tai O, had limited power over their own resources.

As I slowly accumulate more ethnographic experience, I wonder more and more at the ability of some writers to generalize about the environmental wisdom—or any other attribute, for that matter—of “the Chinese,” “indigenous people,” or even “all humanity.”  Examples like the present one show that the same people, under different conditions, can be superb (though never perfect) managers, indifferent managers, or not managers at all.  I have found this range everywhere that I have worked.  The huge literature on fisheries management, ranging from ethnographies (such as Maria Luz Cruz Torres’ superb Lives of Dust and Water, 2004) to conservation and comanagement work, documents similar things all over the world.  Moreover, many boat communities—Tai O included—depended on fish runs.  Vast fish schools migrated from the South China Sea along the coasts and up the rivers.  No one community could exert any meaningful control over these.

Formal national and colonial governments failed because they did not really care deeply about the boat people; because they were ideologically committed to other styles of governing; and because they too faced instability and violence.  Still, one must fault these governments for their indifference to conservation.  China’s Communist government has drawn down the conserved and accumulated wealth of thousands of years, wasting it in a few decades.  China faces a bleak future (Brown 1996; Chetham 2002; Edmonds 1998; Smil 1984).  Capitalism in Hong Kong proved no better; the market never internalized the costs of pollution, resource depletion, and destruction of producer livelihoods.  A truly free market should do these things, but somehow no market is perfect enough.

At any time, the barest minimum of investment in fishing regulation and pollution control could have saved a fishery and a way of life.  Hong Kong’s relentless commitment a chimerical “free market,” and China’s Communist commitment to maximizing immediate production whatever the future cost, were equally effective at wasting a multi-million-dollar resource that could have been saved for a trivially tiny sum.  Thus does political correctness, capitalist or communist, destroy everyone’s benefit, sooner rather than later.






The household of the boat people was, of course, the boat.  When the boat people used the term ka (jia, “household”) it was the boat and its occupants that they meant.  The live-fish boats and the huge old trawlers often had an employee or a few employees, on board, but in other cases the occupants of a boat were close familial kin.

Boats of friends and relatives anchored or moored together.  As Ward (1955) found in Kau Sai, the question of which boats moored together was an important consideration.  The boats did not anchor randomly in the harbor.  They formed neat rows, neater in fact than village streets, and the patterns of anchoring were significant.

Most boats were isolated by water, but they moored near friends and relatives.  Some boats tied up to each other, so that groups became, in effect, one enlarged household.  Brothers, in particular, not only moored together but warped in so that the decks became one single living space.  When five or six trawlers were thus joined, the space was as large as a suburban house lot.   Father-son groups also fused this way, and so, less often, did uncle-nephew groups, brothers-in-law, and other close kin by birth or marriage.   At a festival, boats from outports gathered at Castle Peak, and long-lost kin came together; the united deck space might include 20 boats—true family fleets.  After their father’s death, brothers usually “drifted” apart eventually, but they reuinted at such occasions.  However, no fleet was more than two generations deep.

In the everyday unions of two to six boats, women cooked together; men ate together and discussed fishing and fishing grounds and financial details far into the night.  Children formed one playpack, swarming over all the craft.  The boat people, agile sailors, climbed with ease over and among the gunwales, rigging and tackle piles, barrels, and other impedimenta, or leaped small gaps between craft.  Important business transactions took place over dinner on the boat of the eldest man—the father or oldest brother.  The boats went out to fish together, and often fished close to one another, ready to help in an accident.

As a man’s sons grew up and married, he was expected to help them obtain boats.  This involved considering not only finances, but labor power.  A hook-and-liner could break off on his own as soon as he married.  (Single-person management of a boat was essentially impossible.)  A trawler needed a larger crew; a son had to have enough sons of his own to run a big boat, or an elder brother might split off with his younger brothers for crew.  Labor power, as well as space, determined the fact that small boats held nuclear families but big ones held extended ones (statistics are found in Anderson 1970b:88).  Only the smallest trawlers held nuclear families, and then only as a temporary expedient—at least, so the families hoped.

Families on the largest boats consisted of parents, children, sons and their wives, and grandchildren.  After a father’s death, brothers stayed together on the same boat only long enough to marry and branch out on their own.  Polygamy was theoretically possible, but found on only about 6% of boats, and never involving more than two wives.

Boat families were large in the old days.  Crew members were precious, and, as one person put it bluntly, “it costs less to raise crew members than to hire them.”  In 1965-66, the average family had five children, and the average completed family had about seven.  By 1975, this was falling fast; many young families were stopping at two, and few indeed were having more than four.  I was watching a demographic transition happen before my eyes (as I have since done, again, in southern Mexico).


The wider circle of kin remained important to boat people, though apparently less so than to the land people in the huge lineage villages nearby.  Boat kin beyond the immediate family tended to wind up in different ports, far away and out of touch.

However, kin ties were as strong as the exigencies of a “floating life” would allow.  Kin in general were hingtaai (xing di, “older and younger brothers”).  This phrase meant, metonymously, all kin.  No one used the formal Chinese words for extended families or kindreds.  Relations within the kindred were, indeed, like those between siblings—close, warm, and supportive.

Unlike the land people, the boat people had no lineages (tsou, Putonghua tsu; see Ward 1965, 1966; in some other parts of south China, however, boat people did have lineages; Wu 1937).  Within sight of some boat settlements were the huge lineage villages of the New Territories, famous throughout the anthropological world as perhaps the largest, deepest, and most formalized kinship groups found at ordinary-folk level (see Baker 1968, 1979; Freedman,1965, 1966; Potter 1968; Watson 1975; Watson and Watson 2004).

The boat people gave as reasons the facts that they had no land or equivalent estate to administer; no place to put lineage halls; no fixed address to call home; and neither a reason to group hundreds of kin together nor a way to do it.  Some mentioned also that they were generally illiterate, and so could not keep records.  (Being illiterate, they had not read the British Africanists, and so did not know that many lineage-based societies are nonliterate—the elders keeping track, in their heads, of thousands of kin.  Ancient China, before writing became common, was such a society—though its lineages were simpler than, and different in structure from, the huge lineages that grew up with widespread writing, printing, and middle-class society in and after the Song Dynasty; see Chang 1986.)

The basis of the great lineages is communal estate-holding—especially land, but also other wealth (Baker 1979; Cohen 1976; Pasternak 1972).  The foci of such great lineages are their ancestral starting-points and ancestral halls.  Lacking all the above, and lacking any need for lineages, the boat people did not bother.  Other Chinese fishermen often lack lineages (e.g. Diamond 1969), and, as all social historians of China know, north China lacks the huge, deep lineages of rural Hong Kong.  Indeed, even the other minority ethnic groups of Hong Kong tend to have small, shallow lineages.  Pasternak (1972) provides a land-based tale of lineage loss due to land ownership changes.  He called it “agnatic atrophy” (Pasternak 1968, 1972).  It may have been the original state of his sample village, rather than a derived condition, but, in any case, it was clearly more functional than historic—more a matter of present needs than of historical contingency.  As noted in Chapter 1, many boat people trace their descent to recent land-dwelling families that gave up their lineage identification to go on the water.

The Chinese fishermen we studied in Penang, Malaysia, maintained their ancestral lineage connections only if they were politically and socially active, and then they would pay a fee to join a “lineage association”; ordinary fishermen did not do this, and their lineage connections gradually declined into forgetfulness.  We were watching agnatic atrophy in process.

The limits of the family—the hingtaai unit—are therefore impossible to define.  A family blends off into a realm of forgotten distant cousins.  The patrilineal ancestry will be remembered as far as living memory goes.  An elder will remember his grandfather, perhaps his great-grandfather.  Beyond that, there is only haze.  Only one man in my sample knew much about his patrilineal family six generations before, and that only because his ancestors had left the land for the water in that generation.  Matrilineal kin are even more quickly forgotten.

Therefore, hingtaai rarely included anyone beyond one’s grandfather’s brothers and their children, and even they were usually a rather shadowy lot.  Father’s brothers, by contrast, were intensely involved in one’s life, and mother’s brothers often were.  The boat people were strongly patrilateral, but lack of lineages and closeness with in-laws made matrilateral kin more important than they were among many land-dweller families.

Ancestor images (or tablets, for a few highly literate families) were thus not stored in a temple indefinitely, let alone in a lineage hall (as among the land people).  The ancestors were commemorated on the boat shrine until no one was left alive who remembered them.  Then they were “sent to the sky”:  the images or tablets were ritually burned.  Memory of them faded out in another generation or two.  We later found the same thing in Malaysia and elsewhere; it is, in fact, a general rule for Chinese families that are not closely enmeshed in lineage affairs.

A consequence of this was that boat knowledge of Chinese kinship terminology was limited to these close relatives.  More distant relatives have their own names in Chinese, but the boat people simply extended the terms they knew.  If they needed to be more accurate, for ritual purposes (usually at weddings or funerals), they consulted an expert; a few elders knew the entire system, and shared their knowledge at need.

The Chinese kinship system is the limiting case of a descriptive system.  There are separate terms for father’s elder brother, father’s younger brother, mother’s elder brother, mother’s younger brother, father’s father’s elder brother, father’s father’s younger brother, and so on, for a total of well over 100 terms (see e.g. Feng 1948).  The boat people knew terms for lineal kin, brothers and sisters of direct patrilineal kin, and usually brothers and sisters of matrilineal direct kin too.  They knew terms for in-laws only for the spouse’s immediate nuclear family.  Beyond this limited circle, they knew few or no terms.  Two factors made this simplification desirable.  First and most obvious, the boat people lived in a world of small linear families, not the world of huge lineages where one dealt routinely with distant kin.  Second, the boat people tended to assimilate any distant relative to close-relative categories, as a conscious or “preconscious” way of building solidarity.  The Chinese saying “all men are brothers” was  well known, and considered highly relevant, on the waterfront.

Actual terms of address and reference were not the formal kinterms.  Hingtaai, the formal term for one’s older and younger brothers, was almost always used to mean “relatives,” not brothers.  In ordinary address and reference, brothers were taai lou “big fellow” and sai lou “little fellow.”  One’s son was sai lou tsai “little fellow-let”.  One’s wife was addressed by her familiar name and referred to as lou p’o “old woman”; husband was “old man.”  And so on.  Meanwhile, as in almost all societies around the world, respectful kinterms were liberated for use to any respected adult.  A man of one’s father’s generation was politely called “father’s elder brother”; of grandparental generation, “father’s father’s elder brother”; and so on.  See Anderson 1970b:94-96 for fuller account.

More equivocal was mui tsai “kid sister” (mui “younger sister,” tsai affectionate diminutive particle).  This could be an affectionate term for an unrelated girl or a very unrespectful term for a female servant or slave.  In Hong Kong, as in Imperial China, girls, often boat girls, were still (illegally) sold to wealthy people as “maids”—de facto slaves.  On the waterfront, however, mui tsai was affectionate enough that in the 1960s our little daughter was kuai mui tsai “kid sister foreign devil” (lit. “ghost kid sister”) to everyone.  By the 1970s language was more polite but less colorful.


A proof of this desire to be solidary with one’s immediate world was the network of fictive kinship.  Sworn blood brotherhood (of the sort widely found in China) was one form of this, though it was fairly rare.  More often, was contract (k’ai) kinship, an equivalent of godparenthood in the United States or compadrasco (coparenthood) in Mexico (Foster 1961).  Among the 47 families of my 1965-66 sample, 12 individuals had k’ai relatives, and three of these were in one family.  Most were boys.

Usually, a child would get a k’ai ye (godfather) or k’ai maa (godmother) if the child’s health or other fortune was poor.  The child would then inherit his luck or good fortune from the k’ai parent instead of his or her own parent, whose fortune was presumed to be incompatible with the child’s.  This did not mean the parent’s or child’s fortunes were necessarily bad.  Fortunes could be both perfectly good, but mutually incompatible—as blood types can be, in biomedicine.  The reasons were usually, but not always, astrological.

The k’ai parent would, therefore, be someone with notably good health, or luck, or astrological compatibility.  Such matters were beyond the knowledge of most fishermen; spirit-mediums diagnosed the need, and selected possible k’ai parents.

However, normally, friends and close associates of the child’s parents received the call.  The institution thus formalized and secured links of mutual aid.  It was not so explicitly an institutionalization of this as compadrasco is in Latin America (Foster 1961 and my own experience), but it certainly had the same effect.  For the fishermen, contract parenthood was symmetrical in Foster’s terms; it was always contracted with social equals.  Among the land people, it was usually asymmetrical; more powerful people, and even gods, were called.  In the Chinese Malaysian community where we did research in 1970-71, though the people were fishermen, contract parenthood was not only usually asymmetrical, it was usually religious; gods were the main k’ai parents, with Kuan Yin being by far the most popular.

Once a spirit medium or similar authority suggested contracting fictive kinship, the family selected a likely candidate.  Then the child’s mother went to the woman of the selected household to negotiate.  If all went well, the families worshiped ancestors together and drew up the contract.  For the boat people, a verbal contract sufficed.

Unlike a Latin American compadre, a k’ai parent had little formally to do.  He or she had to give the child lucky money at festivals, and offer presents and receive respect at the child’s marriage.  The child had to pay respects, and mourn as regular kin did at the k’ai parent’s funeral, then continue to worship the parent’s soul (usually by leaving a tablet in a paper-goods shop to receive prayers).

The contract relationship was a mixed bag.  It was not quite real kin, and seemed somehow to compromise family.  A symptom of tension was the use of k’ai taai (fictive younger brother) as a term of abuse, an obscene term for a passive homosexual.  This was a deadly insult.  In actual reference to fictive kinship, one spoke only of father, mother, son, and daughter, not of siblings, so this particular kinterm had no legitimate use.


Since the lack of lineages and need for support heightened the importance of affinal kin, it follows that marriage was the consummately important social act (in more ways than one).  Typically, a family would seek a spouse for their child from the circle of neighbors, good friends, and coworkers.  More than in fictive kinship, parents explicitly wanted to shore up social linkages by thus formalizing them.  There was rarely a shortage of candidates.  Exogamy did not extend beyond the family line, many different families lived close together, and the boat people moved around enough to have friends in many ports.

Marriage among the boat people was so far from the traditional Chinese model that half the marriages in our immediate experience were contracted “for love” (though parental approval and help was always sought eventually).  Moreover, many of the marriages “arranged” by parents, in the classic Chinese form, turned out to have been suggested by the young people in question!  The parents duly formalized the negotiations over dowry and other payments.  Even genuinely arranged marriages—arranged, that is, without the couple’s suggesting them first—rarely involved strangers.

Still less were valued sons and daughters married off to children or aged rich men.  Such things had, indeed, happened in the past, when families had been reduced to starvation and had had to marry off a daughter or send her off as concubine or as a servant to be married off by her masters.  But by 1965 everyone was well enough off to avoid such dreaded circumstances.  The endlessly repeated literary and movie cliché of the bride who meets her husband at the wedding, only to find that he is a child, an ancient, or a madman, was far from boat experience.  Still farther was the concept of “minor marriage”—marrying young children to each other and raising them together (A. Wolf 1995; M. Wolf 1968, 1972).

Given the free and easy lifestyle of the boat people, and the completely accepted institution of “singing girls” at Tai O and elsewhere, virginity was by no means required or universal.  It was theoretically preferred.  I have no reliable data on how common virgin brides actually were.  My data on the sex lives of the boat people are extremely extensive and extremely unreliable!  Among sailors, whether in southern California or in southern China, jokes, insults, teasing remarks, stories, gossip, wild generalizations, and rank speculation about sex, and about individuals’ sex lives, are the classic theme of conversation over a drink or two (or more).  Castle Peak Bay certainly fit (and overfulfilled) that stereotype.  Unfortunately, the resulting excruciatingly detailed and extensive picture was only accidentally in contact with reality.  Not only did people exaggerate, dissemble, and invent; they had an elaborate tradition of erotic folklore to draw on.  Many stories told as fact could be traced to Chinese novels and classic short-fiction collections.  Even without running down specific stories, one could easily find in such literature the rules for insults, boasts, and other relevant folk forms.

We did get enough reliable data to know that sex was fairly free, uninhibited, and joyous among most couples, including at least some unmarried ones.

A double standard existed; men often visited the Tai O girls after marriage, but wives were expected to remain faithful, and usually did so in our experience.  One couple close to us produced a son when the man was 58; his wife was substantially younger, but still well over 40.  We were aware this man was not averse to a visit to Tai O on occasion, also.

Once a young person finally decided to settle down and commit to marriage, the search for a suitable partner was usually not long or difficult.

Some people said that arranged marriages were the rule until recently.  Rich boat people, in particularly, supposedly contracted economic relationships such that the bride and groom were strangers.  We knew of cases of runaways (see above), including some in which brides fled from arranged marriages that did not work.  However, boat society has always been free and open, with love liaisons not involving marriage being well known (recall Shen Fu’s Six Records of a Floating Life; Shen 1983).

Mothers usually negotiated arranged marriages.  Men would lose too much face if the negotiations fell through, so they did not deal with such matters.  Here, as in many other cases, women were by no means unaware of the power this gave them, and sometimes gloated openly about it.  We recalled the equally neat trick by which women convinced men that higher-status people ate last—thus insuring that the “low status” women and children got the food when it was hot and freshly cooked.  The men were rather ruefully appreciative of the dodge.  They worked similar tricks on the land people, using “low status” to excuse a number of irregular behaviors.  Truly, low status has its rewards, especially if the low-status individuals can outsmart the higher ones, as James Scott reminds us (Scott 1985, 1998).

The boat people did not usually use go-betweens, since there was not enough money to reward them.  Sometimes, a relative served as go-between, or at least alerted a family with a marriageable youth of the existence of an available and desirable potential mate at such-and-such a port.

Chinese traditional families were supposed to take horoscopes into account in deciding whether a match was suitable.  Nobody at Castle Peak seemed to take horoscopes very seriously.  Wolfram Eberhard (1963) found that in Taiwan people tended to use horoscopes as excuses to set up, or break off, engagements, when the real reasons were personal.  I suspect the boat people did the same, but people we knew simply did not take the horoscopes seriously enough for this to become an issue.

Within the boat community, almost all marriages linked families in the western New Territories ports.  People married fairly often within the Castle Peak community, but were just as apt to find spouses in Tai O or other nearby ports.  Much more rare was a marriage to Aberdeen or any other more distant bay.

The boat community was almost strictly endogamous.  People tended to marry individuals from the same general category of the boat world.  Trawler boys usually married trawler girls, and gill-netters married gill-netters.  However, this was very far from a strong rule—merely a result of differential knowledge and networks.  More serious was a tendency for people of similar economic status to marry.  A marriage of rich and poor did not seem a good bet for lasting success, according to waterfront opinion, borne out by alleged cases.

A few marriages of boat girls to land-dweller’s sons had recently taken place.   Only one land woman had married a boat man, and she promptly left him—the life was too hard.

Once married, a girl moved to her husband’s boat or house.  She then became part of the family.  She was still something of an outsider, however, till she bore a son, and thus contributed to the patriline.  This was less an issue than among most Chinese (including the fishermen we studied in Malaysia; see also M. Wolf 1968, 1972), but was still a very serious matter.  However, a wife who bore only girls was thoroughly accepted after a few years, though the family would always be somewhat regretful of lacking a son.

After marriage, relations with in-laws were close, often coming to approximate parental relations even for men—let alone for women, who, in China’s patrilineal and virilocal world, almost always came to live with their husbands.  Exceptions occurred when women had large, independent money pools—something that normally could happen only if the woman in question had been a successful singing-girl.  The close, supportive ties of in-laws were a point of difference from the land people, who found in-laws a frequent cause for tension and distancing (see e.g. Baker 1968, 1979; Freedman 1966).  In fact, we knew several cases in which a man’s closest friends, coworkers, and allies were his wife’s immediate kin.

The stated ideal of boat families—described to us repeatedly, in rather set terms, by dozens of people—was for everyone to get along.  In-laws in particular had to be supportive and warmly helpful to each other.  Clear lines of authority should be maintained, but an elder should listen calmly, respectfully, and attentively to everyone’s opinion—expressed in due order of seniority, and with due consideration and temperateness.  Everyone was supposed to be even-tempered, calm, and thoughtful.  Aggressive and impatient behavior endangered everyone on a boat in rough waters, or even in the harbor.

Some families really did live up to this irenic ideal.  We had tea every morning (in 1965-66, and frequently in 1974-75) at a tiny café-cum-residence near our boat.  It was the home of a widow—of land origin—and her boat-born daughter-in-law.  (The latter’s husband usually worked outside the home).  The two women were both gentle, affectionate, retiring souls who adored each other and adored the daughter-in-law’s children (who were, indeed, model children by any standards).  In the hundreds of hours we spent absorbing tea, we never saw anything but affectionate and gentle behavior in that household.

However, families with such close relationships represented one end of a continuum, the other end being represented by dysfunctional families with considerable fighting, abuse, and even illegal violence.  In between were the normal families, usually harmonious but often argumentative or burdened with one or two distant and suspicious members.

Some families even exemplified the tense, distrustful relationships between in-laws described by Hugh Baker.  The lack of lineages and lineage villages, however, meant that in-laws were never real enemies.  They often were real enemies in the old walled lineage villages that Baker studied, because lineage exogamy forced men to find wives in other villages, and neighboring villages were almost always competing over land.  This often led to violence.  Wives were suspected of, and sometimes clearly guilty of, helping their natal families in enemy towns.

The dynamic was exactly the opposite among the boat people.  One needed all the helpers and allies one could get.  One never knew when one would need to be rescued in a storm, helped with a large catch of fish, or helped with a quick small loan to fix an engine.  One did know that all these needs would arise, sooner rather than later.  Therefore, the pressure was all on the side of being as solidary as possible with in-laws—and with anyone and everyone else nearby.

At first, though, the plight of a new daughter-in-law was hard, often very hard.  As junior woman, she had to do the heaviest and dirtiest work, and patiently accept criticism from her husband’s mother.  As noted above, in-laws were much closer on the boats than among many land people, but this was still no enviable situation.  It was the subject of several of the boat folk songs (Anderson 1975; Wu 1937).

However, the tendency to marry among close friends and neighbors usually reduced the tensions innate in the Chinese family between an in-married girl and her parents-in-law (cf. Baker 1968; M. Wolf 1968, 1972).  Therefore, the classic Chinese drama of the wronged and abused daughter-in-law played itself out only occasionally among the boat people.

Still, it did happen.  Brides fled back to parents to escape impoverished, wastrel, or abusive husbands.  The boat people of Castle Peak Bay being a notably non-abusive lot, all the runaway cases known to us involved economic problems.

We heard of at least one case of attempted suicide—a more socially tense case involving a girl from a poor family married into a rich one and thereafter insulted and hounded.  Suicide of a daughter-in-law is probably the commonest type of suicide in China, and occurs as a trope throughout the history of Chinese literature.  Such a suicide is not just a tragic act; it directly damages the abusive offender—usually the husband’s mother.  Not only do abusive individuals lose face; the unfortunate girl becomes a “wronged ghost,” haunting those that have driven her to death, driving them to madness and misfortune.

A woman’s lot was hard work, but fairly equal partnership.  On the fishing boats, she was high in the command chain, and she had to manage boats and haul nets or lines with the men.  If not in a fishing family, she would still be actively engaged into economic activity; no one could afford a nonworking wife.  This gave her considerable power and independence.   She could not be relegated to the shaky status of fishing wives in some parts of the world (Cruz Torres 2004).  Even among boat families that had forsaken the water to live and work on land, women had to work—not only around the house, but usually for cash—to make ends meet.  They had to have decision-making power, and they had it.  The Confucian ideal of women obeying “fathers when young, husbands when married, and sons when old” had less to do with reality than in some, probably most, Chinese spheres.  In particular, obeying sons when old was conspicuously absent; sons deferred to mothers throughout life.


Each boat family was theoretically an independent and sovereign unit, regulating its affairs with little interference from anyone.  Even brothers could only suggest, advise, and agree.  Only a father with a fleet of sons could genuinely command other boats, and even he had that right only as long as his sons allowed it; very often, they did not.  All they had to do was sail to another port, or even another side of Castle Peak Bay.

On the boat, there were strict rules of authority, which made life very much easier.

First, each boat had one senior man who served as captain.  This was the eldest man on board, unless the eldest was too old and frail, in which case he would give over command of his boat to a son, or live on a son’s boat with the son as commander.  Since the eldest son would normally have long ago gone his own way, the caretaking son was often a younger one.  The transition from captain to elder was a gradual one, managed with a minimum of friction, but—be it noted—not totally without friction.  A man who was losing control and wanted to maintain it would try hard, sometimes tyrannizing his sons, who would in turn try to maneuver him—usually with great tact and gentleness, but sometimes with blunt frustration—into giving over command.

Occasionally, the strict seniority rule was broken by an older brother deferring to a younger and more assertive or competent one.

Second, other adult men and women advised, in order of seniority and gender.  A son listened to his mother until fully adult, after which the mother was supposed to defer.  As long as the father was alive, father and mother stood together in authority, but once a father died, his widow was supposed to defer to her son, unless he was still quite young.  This was, of course, the classic Confucian pattern, and it was universally followed.  However, it provided plenty of room for play; how old is “still quite young”?  And a mother with strong character made her will known quite emphatically, no matter how old her son was.

A wife, especially a captain’s wife, had considerable real authority, and many of them “twisted their husband’s ear” thoroughly.  (The idiom, of obvious derivation, is the standard Cantonese phrase for the condition.)  A man with an ear-twisting wife got compassion and some sympathetic joking from his peers; most of them knew how it felt.

Third, a captain was expected to listen to advice and counsel from other adults, and take it when it was good.  They, in turn, had to respect his final decisions, as well as commands that he had to make without consultation when faced with the need for immedate action—a daily event on the open sea.

Fourth, women controlled cooking, child care, and related aspects of home life.  They did not take the lead in off-boat matters (politics, economic transactions, and so on), but they provided advice and back-up assistance.

The women’s status hierarchy, like the men’s, was a matter of seniority, but here confusion reigned.  Theoretically, an older son’s wife ranked above a younger son’s wife, but if the older son’s wife were (herself) younger than the younger son’s wife, or (much more serious) if the older son’s wife had not born a son and the younger son’s wife had, a real problem in rank presented itself!  Such contingencies virtually guaranteed that the couple in the anomalous position would get their own boat.

On a large fishing boat, the senior man would be stationed at the wheel or on lookout on the cabin roof.  The next senior man (the oldest son, in the normal situation where the captain was the father of the family) would take the wheel when the senior man was otherwise occupied.  Otherwise this second-senior man would be managing the rudder, rigging, or fishing tackle.  The third man would be in charge of the engine.  The fourth would be the usual manager of fishing tackle.

The senior man’s wife would direct cooking and mending; the next woman in line would do the least unpleasant cooking and cleaning operations; and so on down to the youngest adult woman, who would swab decks, do heavy laundry, haul water, and help fish.  Children and oldsters would be mending nets, washing vegetables, and looking out for younger children.  Everyone knew his or her exact place and task responsibiilities, and if there were any question, a council of seniors would resolve it.  Sometimes this council was convened when a family fleet moored together in the evening, and brother captains discussed the issue.

Boats, whether in Newport Harbor or Castle Peak Bay, have to be managed with exquisite and thorough order and care.  Loose lines of authority, poor task management, and shirking were immediately disastrous.

On the other hand, inflexibility was even more rapidly fatal.  Discussions over dinner were an institution never ignored.  Consultation, along proper lines of authority, was not only real but universal.  People followed the Confucian virtues of harmony, cooperation, and respect, not only because they came from a Confucian tradition, but—much more importanly—because they lived and moved in the world that produced Confucius and his ideas.  China had changed in the 2500 years since the sage taught, but some aspects of it had changed very little.  The need for a household to regulate its affairs stood high on this list.

Naturally, this allowed much variation.  Some boats were run in a quite patriarchal and authoritarian style; others were so democratic that we wondered how anyone ever decided anything.  The vast majority were somewhere between these extremes, but the spread was great, and no obvious central peak presented itself.

As people became too old to work, they slowly faded from the world of the living.  Death, like maturation, was a gradual process.  Command devolved increasingly on the next generation.  The elder slowly sank into quiet.  He or she did the shopping on shore, takes care of young children, tells stories, dreams, and waits patiently for the afterlife.





Child-rearing on the boats turned out to be one of the most interesting and thought-provoking things we studied.  Having children of our own, and spending two full years in local households with local families, we had a unique opportunity to observe.  Our two older children were old enough in 1974-75 to be real field workers, and both were (and still are) first-rate observers and judges of humanity.  They were well accepted into local playpacks, and became treated more or less like any other neighborhood kids.

This was one of the areas where one could learn important data from a society now gone.  The distinctive family behavior on the waterfront challenges stereotypes—both Chinese and western—of “the Chinese family.”  In household behavior as in much else, the boat people had a quite distinctive way of being Chinese.  I have told the full story elsewhere (Anderson 1992, 1998), but some repetition will be useful here.

Much of the literature on Chinese families presents a picture characterized by authoritarian patriarchy, distant and stern communication, and silent obedience alternating with wild break-out.  Chinese novels often present this picture, from Cao Xueqin’s great 18th-century novel Hong Lou Meng (translated as The Story of the Stone; Cao 1973-86) to Ba Jin’s 20th-century classic Jia (“The Household,” translated as The Family, Pa Chin 1964).  Western accounts are usually conformant to this image.  Some even stress violent wife and child abuse and other pathologies, maintaining they are typical and inevitable consequences of the patriarchal family (Eastman 1988 is particularly dramatic, but there are many other examples, both Chinese and western).

As one might expect from all that has gone before, the boat people were as far from this stereotype as one could be and still remain identifiably Chinese.  The patriarchal Confucian family was alive and well, but in a very open and even nurturing form (Anderson 1992, 1998).

Picking up from the last chapter, I can begin by continuing the account of child-rearing on the water.  Children slowly and gradually got more rights and more responsibilities.  Every new task responsibility meant more freedom, authority, and ability to act independently.  This process started when toddlers were old enough to help mend nets.   It ended only when a man finally took over from his aging father; the two could be 60 and 80 years old by this time.  The man would still be learning from his father.  Thus, there were no sharp breaks: no initiation rites at young adulthood, no sudden changes at marriage.

An eight-year-old could mend nets or bring water and bait.  A twelve-year-old did simple fishing.  A sixteen-year-old boy would get a chance to steer the boat, while a girl would be cooking.  Usually, children were ordered to do something; parents expected them to know, from helping and from observation, how to do it.

The plan produced the most problem-free adolescence I have ever seen.  Teenagers had a physically difficult and demanding life, but at least did not have to face the trauma of being child and adult at the same time.  They moved rather tranquilly up the ladder of command, pushing to have more authority, but aware that they would get it only when they could do more responsible tasks.

Boys got more freedom, less discipline, and better care than girls.  They also got less work; girls had to help fish and manage the boat and had to help with women’s tasks.  The disparities were not as great as in many or most Chinese communities (see e.g. Croll 2000), but they were certainly real and obvious.  Some parents dressed boys in girls’ clothes to make them less attractive to deities that might want to steal them.  Boys got more good luck charms to wear:  tiny models of life-saving floats; sea horse eyes; jade bracelets; tiny strings of old coins.  Boys were more outgoing; girls turned inward and expressed less emotion; this was the expectation, generally fulfilled, though to varying degrees.

Education was intensely moral.  Actual verbal instruction was heavily focused on moral training.  Learning about fish took place within this moral discourse.  From birth, a child learned that he was part of a wide family group, and that his or her will was more or less subservient to its collective interests.  Most of the people in the child’s world taught this more or less consciously.  Mothers passed babies around to be held by everyone.  All the women and girls on the boat or in the household “mothered” all the children.  From toddler age, children are involved in playpacks that included everyone in the extended family, and, on shore, in the whole neighborhood.  Except on the smallest and most socially isolated boats, the child has close relationships with many persons, but no single outstandingly close one-to-one relationship.

Life on the boats, or in any traditional Chinese household, cannot center on the will of the baby.  A mother nurses when she has time, not when the baby is hungry.  She, or any other caregiver, dresses, washes, puts down for sleep, gets up, plays with or quiets the baby when time is convenient—not when the baby wants it.  The baby soon learns to be philosophical about all this; crying does no good, and smiling or cooing does rather little.  This does not mean the baby gets little care; quite the opposite.  Usually, several female caregivers are around, and the father and his close male relatives often do their share.  The baby most often cries in a desperate and totally futile attempt to be left alone.  Wu (1937:812), describing the same syndrome, reported that boat children were “spoilt.”  Wu’s value-judgement shows that a traditional Chinese male was as surprised by the high level of caregiving and low level of punishment as we were.  The actual effect of this system is not to “spoil” the child but to teach him or her that his or her will is subservient to that of elders.

The pattern continues with age.  Boat children could not develop exploratory and independent abilities early; they would fall overboard and drown if they did.  They could also get caught on fishhooks or other deck gear.  They were thus tied to the mast and kept as quiet as possible.  Caregivers discouraged early walking.  Someone was always around to pick up the child.  When a child fell overboard—which happened often, in port; people were more careful on the open sea—everyone within 50 yards, whether related or not, instantly went to help.  (A strange story among land people held that boat people would not save a drowning person.  This was no more true than the other common tales told about the boat-dwellers.)

Mothers nursed babies for about two years, never less than eighteen months and sometimes up to four years.  Caregivers fed children solid foods from eight months on.  Children did not feed themselves until three or four.

A distinctive and vitally important feature of child-rearing among south Chinese was the meitaai or “carrying belt.”  This was a square of cloth that held the baby to the mother’s back.  It was tied on by ribbons, starting from the four corners, and tied together in front.  This left the mother (or other caregiver) free to do anything and everything, while producing no special strain on her.  Babies were carried in the meitaai until three or four.  All south Chinese used this device, but the boat people used it much more than the land people, since they needed to keep the child immobilized as much as possible.

The meitaai gave the child a tremendous sense of security and reassurance.  A baby or toddler crying inconsolably would always quiet down immediately if backpacked.  This was true even when the crying was due to real hurt, physical or psychological  Weaning from the meitaai was, strikingly and very obviously, more traumatic for the child than weaning from the breast[4].  Thus, severely troubled children were sometimes backpacked when they were six or seven years old.

Girls recently weaned from the meitaai almost always got toy meitaais, which they used to carry dolls and stuffed toys.  By the time they were five or six, they often had a real baby—a new younger sibling—to carry.

The age from three to six was traumatic and difficult; the child was weaned from breast and meitaai, but had no freedom yet.  Crying and aggressive behavior increased.  By six, however, the child was undertaking the simpler adult tasks.  Boys began to fish; girls began to care for siblings.  With this came a vast increase in freedom and opportunity to learn, explore, and interact with the wider world.  Crying and aggression dropped off, in proportion to the steady increase in rights and duties and in playpack involvement.  With age, playpacks got larger, more diverse, and more cheerful.  The playpack, inevitably mixed-age because older siblings are taking care of younger ones, became a major focus of socializing and learning.

Playpacks were not very gender-segregated.  Boys and girls played with dolls, girls and boys explored and scuffled.  Both improvised elaborate games with bits of string and wood and paper.  Somehow almost everyone managed to find a kite to fly in spring.  Teasing and taunting were common, and, in all innocence, boys and girls alike used the rough waterfront language in perfectly ordinary settings.  Children knew an amazing range of songs, rhymes, chants, and other folklore.  All this is lost without trace; we could not record it, partly because the subjects would never hold still.

Many boat children were notably less aggressive, less noisy, and generally better behaved than most land-dwelling children of the area.  This was especially true of the recent refugees from mainland China, a very traditional population.  However, a substantial minority of the boat children were toward the other extreme; they resorted to begging, yelling insults, and otherwise acting bad, as long as they were out of parental control.  This earned them the sobriquet of “water rats” from the better behaved.  Overall, one can see, in this range, the independence and individual difference so visible also in adults.

The boat people reacted with horror and with immediate community intervention to threats of genuine abuse.  Senior family members stopped it in no uncertain terms.  Failing the availability of senior family members, bystanders would quietly, tactfully distract and calm a parent who appeared to have lost his or her temper to the point of physically harming a child.

By the age of eight or ten, a child rarely needed punishment.  The message of overwhelming family dominance had penetrated.  A simple verbal command almost invariably received rapid compliance.  If it did not, a most impressive verbal fireworks display could result, and no one ignored that.  Physical punishment, reserved for the most serious cases, consisted of a light slap, or in extreme situations a light but painful spanking across the backs of the calves with a flat board or thin stick.  Parents inflicted this for only three situations:  fighting, extreme insubordination (very rare), or dangerous neglect of a younger sibling.  Girls got more punishment than boys, partly because of patriarchal bias, but partly because girls usually did most (though far from all) of the sibling care, and thus were at risk of committing neglect.  If a girl neglected a baby brother, she was in real trouble, worse than anything that happened when she neglected a baby sister.

We saw such punishment about once a week, and in an average day our research and marketing rounds took us past hundreds of households, where life was mostly lived in the open.  The lack of conflict was striking.  Moreover, such conflict and parental problems as occurred were concentrated among certain families, mostly boat people who had failed at fishing but had not fully moved onto shore, and who lived in wrecked boats in the typhoon shelter.  One group in particular made up something of an outcast colony, and had continual (though usually minor) problems with children.

The land people, though somewhat stricter than the boat people, followed the same principles of childrearing.  Far from the beatings alleged in some Western literature on China, corporal punishment was almost unknown at Castle Peak, beyond the spanking on the backs of the legs previously noted.  British colonials frequently commented to us on how amazing it was that “Chinese children are so well-behaved” when they were so “spoiled” (and sometimes added it was proof that the Chinese were an inferior race, naturally docile).

In fact the secret was the deliberate ignoring of the infant’s will, discussed above.  The land people saw this as simply part of socialization for civic life.  The boat people, with their value on independent agency, saw a contradiction, but dealt with it by saying that young children had to be restrained or they would have accidents, such as falling overboard.  Thoughtful parents went on to say young children did not have and should not have much agency, and, in any case, they were part of the family, a structured world where order had to reign.  The famed independence of the boat-dweller had to be earned, by growing in responsibility over the years.  With independence as a promised reward for clear evidence of responsible behavior, children were motivated to take on more duties.

It is not hard to see how such a child-rearing method produced well-behaved young without harsh discipline.  This was especially true on the boats, where a child had no escape.  Land people had to be more disciplinarian because their children roamed the whole neighborhood in playpacks.  But even the land people of Castle Peak Bay were, by Chinese or world standards, gentle parents.  They made up in constant supervision what they lacked in harshness.  Adults or older children were always around, and would discipline any local child that was out of line, sometimes even perfect strangers.  This certainly had much to do with maintaining tranquility and keeping children socialized.

Some particularly articulate parents put the matter in philosophical context.  The independence of an adult had to be won.  It was not something that came with the turf (as it does for many American teenagers).  It came only with demonstrated competence and responsibility.   The road from restrained infancy to free senior adulthood was a gradual progression.  Each stage had to be won by demonstrating the ability to handle more duties with responsibility.

Crime and trouble were not teenage specialties.  Crime was very rare at Castle Peak in any case, but when it occurred, adults were the ones involved; some families were pirates by trade and tradition.  Teenage gangsterism was, however, not unknown (and not even rare) at some more urban ports.

Few children got to school in 1965, but the number was rising, and by 1974-75 all children got primary education.  Some were beginning to go on, and boat people in more urban ports were beginning to go to college.  A government survey of 187 Castle Peak children in 1964 showed most had no schooling, and very few girls got to school at all.  By 1965, people recognized the need for education.  A rapid increase in the numbers took place.  The major reason to move on shore was to get access to schools, and those who moved on shore took in countless related children whose parents were still on the boats.  Bright girls sometimes got to school instead of duller brothers—a really dramatic new turn.  At this time, school was cheap or free, but parents had to buy uniforms, supplies, and books.  We knew families in which everyone went essentially without food for days in order to buy these for the brightest children.  Those bright ones had hope; their siblings had nothing but hunger.

Cultural reproduction thus took place through a variety of informal channels, and occasionally the formal one.  Most education was through practice (cf. Bourdieu and Passeron 1990; also Chaiklin and Lave 1993, Lave 1988; Lave and Wenger 1991; Rogoff and Lave 1984).  Children learned not through formal instruction but through watching and involvement—what Barbara Rogoff now calls “intent participation,” a process which “involves observing keenly and pitching in to shared endeavors” (Rogoff 2004).  She defines this as “horizontal and collaborative….  Experienced people guide while participating, and learners take initiative….  Motivation revolves around the chance to contribute to important activities….  Assessment occurs during shared endeavors, to aid learning.”  This description of learning among the Maya of Mexico is perfect for the boat people of Hong Kong—and, for that matter, for the Maya that I too have lived with in Mexico.

Going beyond that literature, I speculate that the boat people are typical humans in their embedding of instruction in a moral framework that stresses the virtue of the learning.  People want to learn only when they think the data are important socially.  They must be emotionally involved in the process, and they must think it is in some sense “right” or valuable.  So, when the boat people went on shore, they immediately dropped the old knowledge, forgetting even what they knew; it had ceased to have meaning for them.


For a few decades in the late 20th century, a debate arose between those who felt emotions were innate and shared by all higher animals and those who saw human emotions as complex, basically learned, and subject to heavy cultural conditioning.  Most earlier psychologists fall into the former category.  Many anthropologists fell into the latter.

Freud and other early psychologists introduced many new ideas that made emotions look less than simple and straightforward.  Freudian theory forced social scientists to look at even the simplest emotions as highly conditioned by early experience.  By the 1930s, anthropologists were studying “culture and personality” (Wallace 1970), basically the interactions of innate feelings with cultural norms and structures.

A later generation of anthropologists, often those trained by such culture-and-personality theorists as Beatrice and John Whiting, concluded that the feelings might not be innate.  They held that some, or even all, emotions were culturally constructed.  Richard Shweder (1991), Catherine Lutz (1988), and their associates were leaders in this field.  Lutz argued for the utter strangeness to westerners of the Micronesian emotion of fango.  However, she weakened, if not destroyed, her case, by doing a perfectly good and clear job of explaining it.  It turned out to be much like ordinary familial love as practiced in the world at large.  (It seems to me to be a social construction of parent-child love.  Having done field research in Tahiti, I have some experience with the closely related Polynesian concept of arofa.)

Others drew from other theoretical perspectives.  The Polish linguistic anthropologist Anna Wierzbicka argued for the uniqueness of such cultural states as Polish feelings of nostalgia (Wierzbicka 1999), but her description makes it sound like other cultures’ ideas of nostalgia.  Sociologist Rom Harré insisted on the fundamental difference of French, Spanish, and English emotions (Harré 1986), but his evidence is not convincing.

Time seems to have softened radical claims.  Few if any cultural anthropologists believe now in purely innate feelings unmodified by culture.  A few biological anthropologists probably still do, but certainly the evolutionary psychologists and those influenced by them have moved beyond such a position (Barkow, Cosmides and Tooby 1992).  On the other side, few have followed the extreme social constructionists.  Most scholars now allow innate feelings to exist.

To make a long story short, the intermediate position turns out to be correct.  Clearly, people do have deep innate feelings.  Rage, affection, loyalty, fear, interest, and parental affection are shared with all mammals and are clearly genetically specified in a reasonably tight way.  Feelings like guilt and disgust may require human intellect, but are arguably panhuman.  However, humans are thoughtful animals, and highly social ones too.  So one expects to find differences in the subtleties and in the cultural interpretation of emotions.  Shorn of their extreme rhetoric, Shweder, Lutz, Harré, and others certainly show that people in different cultures, and for that matter different individuals within the same cultures, have different phenomenological worlds.

Psychologists have long studied the interaction of emotion and reason.     From the 1950s onward, Albert Ellis had built up increasing mountains of evidence for the success of rational-emotional therapy (RET; Ellis and Blau 1998).  This method depends on the recognition that people are creatures of both reason and emotion, and that human action involves integrating these.  Many life problems come from mistaken cognition leading to misdirected emotion.  Aaron Beck’s cognitive therapy (Beck 1999 is perhaps the most anthropologically interesting, if not the most focal, of his many books) and several variants of it essentially follow Ellis’ perception.

Experimental psychology followed in due course, teasing out the neurological pathways that link emotion and reason.  It turns out that these are closely integrated in the brain (Damasio 1994).  Social perceptions, and all normal social functioning, depend utterly on successful performance of this integration task.  It also appears that some degree of feeling—“emotions” in a very basic sense, or at least moods and animal reactions—influence all cognition, even the most apparently neutral (Zajonc 1980).  A constant low-level play of interest and disinterest, evaluation for good or ill, and responses of affection or fear color all thought.  Reactions thus involve cognitive evaluation of situations, but also emotional reactions to them, combined in often predictable ways; fight-flight response, for instance, turns on evaluation of a threat and of how best to cope with it (Bandura 1982, 1986).  The situation at hand is evaluated in the light of prior experience; one flees, cowers down, or fights back.

Culture and individual cognition can thus influence emotional states, and the experience of emotion.  Mental action is a constant, shifting, rich fabric in which knowledge, feeling, and will are interwoven and color-mixed as in a Chinese textile.  However, predictable basic human reactions provide the warp, supporting and defining the whole.  Culture weaves patterns over the warp strings, and holds the fabric together.


All this is preface to a consideration of some claims made by western observers about Chinese emotional states.

First, westerners, and some Chinese as well, often claim that Chinese are among the peoples who are so intensely social that they do not experience the world the way “westerners” do.  According to a stereotype that goes back to the ancient Greeks, “Asians” are group-oriented, in contrast to the individualistic, freedom-loving peoples to the west of them.  These western individualists were first the Greeks, then “Europeans,” finally the ambiguous and unbounded “westerners.”  In the modern version, as espoused by scholars such as Richard Shweder, “Asians” are group-oriented, think mostly of the welfare of the collectivity, and do not stop to analyze and think independently about their cultural rules.  “Westerners” are “individualists,” verbal, self-conscious, self-analytic, and well-contained.  (See Shweder 1991 and references therein; see Anderson 1998 and Anderson et al. 2000 for China citations. For the general case of Western and Chinese stereotypes of their culture and of Chineseness, see the very important historical study by Lung-Kee Sun, 2002.)

My professor in these matters, John Pelzel, a superb teacher, used to say (in lectures, 1961) that, according to this view, westerners are like steel-plated balls, while Chinese were like amoebas, extending pseudopods till “there is more of them in the pseudopods” than in their own selves.

An interesting question is just who these “westerners” are.  The term is very general, but in most cases the scholars obviously mean urban, university-educated Americans and West Europeans.  Small-town Midwestern Americans, or East European villagers, might make a very interesting comparison set.

Be that as it may, there is no doubt that Chinese, in general, are indeed more strongly social in their values and behavior than such urban educated western folk, and are less self-consciously individualistic.  Michael Bond (1986), for example, assembled many data-rich papers that prove this beyond reasonable doubt.

However, the contrast between “westerners” and “Chinese” is clearly overdrawn in much of the literature, as Sun (2002) shows in extenso.  The Hong Kong boat people prove that Chinese can be reasonably individualistic.  They do not, to be sure, inhabit steel-plated balls, but neither do westerners outside of cultural psychology books.

Scholars of such matters would do well to consider Chinese literature.  On the one hand, it almost uniformly describes highly prosocial values, and reflects a patriarchal order.  This order seems alien and oppressive to most westerners.  However, it also seems oppressive, though not alien, to the Chinese writers themselves.  Modern novels, in particular, usually criticize it.  Their criticism proves they take it seriously.  (See, for instance, the writings of Lu Xun, Ba Jin [Pa Chin 1964], or Maxine Hong Kingston.)

Yet, on the other hand, Chinese literature reflects an emotional world strikingly familiar to the western reader, once he or she adjusts to the unfamiliar order imposed by kinship and family structure.  The most revealing single example is Cao Xueqin’s great 18th-century novel The Story of the Stone (Cao 1973-1986).  Cao describes in exquisite detail the lives of teenagers growing up in a particularly strict Neo-Confucian family.  They have to conform to norms particularly rigid even by traditional Chinese social standards.  Yet, they are all individuals, and they are all highly self-analytic and self-conscious.  They experience a whole range of emotions familiar to American readers, including the desperate urge to rebel against family strictures.  They resemble American teenagers closely enough that anyone spending time with a group of bright American young can soon find close matches for Cao’s characters.

One might also consider the relation between rhetoric and reality.  Certainly, Chinese rhetoric, from classical literature down to village discourse and beginning school classes, favors the Confucian position.  Obedience, prosocial behavior, conformity, and hierarchic relationships of deference receive constant attention.  But historians know that when a government continually passes laws against some particular behavior, that means the behavior is commonly problematic, not that it is eliminated.  Many Chinese literary works, notably Wu Chingzi’s novel The Scholars (Wu 1957) and Pu Songling’s short stories, dissect the enormous disparity between rhetoric and behavior.

A reader of these or any of thousands of other Chinese novels, stories, plays, operas, and poems cannot escape feelings of great familiarity.  China certainly produced its share of dour advocates of conformity.  However, it also produced its share of selfish and amoral people.  Perhaps most interesting, it also produced prosocial but highly independent thinkers—classic “individuals” as idiosyncratic as anyone westward.  One recalls literary figures such as Mencius, Tao Yuanming, Li Bai, Zheng Xie, Yuan Mei, and countless more.

Even the hypersocial preachers of conformity and dogma often turn out to be rather misconstrued.  Zhu Xi in the 1100s taught, and to a great extent originated, the strict Neo-Confucian rules that often seemed (in future centuries) to be an enormous leaden weight on Chinese civilization.  Yet Zhu himself was a loner and outsider, rulemaking from a social-critical position based on Mencius’ rebelliousness.  He spent a good deal of his life in a hut perched partway up a thousand-foot sheer cliff in the remote mountains of Fujian.  One wonders what he would have thought of the rules’ latter-day petrification.

Ethnographies of China that focus sensitively on feelings (e.g. M. Wolf 1971; Gates 1987) seem, in general, to conclude that the human feelings of Chinese in Chinese families seem to be what one would expect of ordinary people subjected to such realities.  The differences between Chinese and American life are clear and sometimes deep, and usually in the postulated directions.  Chinese family structures have their effect on human relations.  On the other hand, humans are indeed siblings beneath the epidermis.

Rhetorics of conformity and individualism may also outrun people’s realities.

Chinese of the old school—including literally everyone in Castle Peak—loved to justify anything they did by recourse to the words of Confucius, Mencius, Zhu Xi, or any other ancient authority who might be relevant.  Doctors quoted ancient medical works, educators quoted old education manuals.  However, we noted that people justified quite different, and even opposite, courses of action by recourse to the same sages, and even the same passages.  Clearly, more directly relevant reasons had some bearing here.

Conversely, Americans love to justify anything they do by some directly functional reason, especially an economic one.  However, much of what is so justified is so far from economically sensible that one must look for other reasons.  One often finds that the reason is sheer cultural tradition.  For instance, Americans continue to wear Northern European clothing styles in southern California; the Native Americans sensibly went almost naked in that hot land.  Frequently even the American pretense of “rational self-interest” breaks down, and Americans admit that they are seeing an admittedly awful movie, or following an admittedly ugly and uncomfortable fashion, solely because “everybody’s doing it.”

Chinese concepts of intelligence foreground not only intellectual competence, but also “interpersonal intelligence” and rote memory, while American folk concepts of intelligence feature verbal skills more than Chinese concepts do (Sternberg 2004:335).

One should not confuse the rhetoric of folk explanation with the reality of emotional and intellectual experience.


Another problem concerns privacy and space.

A small fishing boat provides a minimum of space and privacy.  Even on shore, a fisher family’s simple home is little better.  A one-room hut, designed like a boat cabin or built as a square block, jams up against many others on a crowded shoreline.

Humans are amazingly social creatures, but such conditions tax their abilities to get along.  Formerly, some authors suggested that humans are territorial, or cannot tolerate crowding, and I found in Hong Kong an ideal test of that theory, since obviouisly people were forced into a minimum of space (Anderson and Anderson 1973).  The boat people had it easy.   In 1965, when recent refugees from China jammed the city, I visited a room in the oldest part of Kowloon that had 57 people living in a space of approximately 750 square feet.  They had decked the room into two storeys (each about 4 feet high) and they slept in shifts.  Yet they got along.  In Malaysia, fisherman friends of mine who had gotten relatively affluent built large houses—and used the opportunity to bring all their families in.  The result was as much crowding as in any hut.  They were proud of it.

Yet, inevitable stresses occurred.  Children played noisily.  Competition for scarce deck space was inevitable, as people worked on critically important projects—net-mending, trap preparation, bait-cutting.  Cooking filled the air with smoke, which dispersed on an open boat but was difficult to bear in a small house.  The bathroom of a boat was the starboard side of the sterncastle—a hole allowed wastes to escape—and small boats had no way for bathing or defecating individuals to hide completely.  No one looked; people condemned any staring.   Sex could hardly be private on a rocking boat.  Couples were very quiet, often sitting up (in the position of Tibetan yab-yum figures) to avoid noise.  But, inevitably, a small boat would sometimes shake and creak, conspicuous in the still night.  People studiously ignored this, or sometimes teased a particularly amorous couple.  Privacy was respected as much as possible, but no one was under the illusion that total privacy was possible.

Beyond these, rules for respecting space and privacy were clear and sharp (Anderson 1973—I have benefited since that date from discussions with Rance Lee, Robert Moore, and Ambrose Tse):

First, functional subdivision of space.  Sleeping space was inviolate.  Other activities have their defined areas.  Socializing was in neutral space:  front room of a house, center or aft deck of a boat.

Second, time was loosely structured; people worked around each other, negotiating.

Third, people tolerated almost any amount of noise.  The bay was quiet, but life on shore, especially in the towns, was very noisy.  People adjusted, learning to sleep through everything from crying children next door to firecrackers at New Year.  Thus, the constant pressure to “quiet down,” typical of American towns, did not occur.

Fourth, children could be disciplined by anyone, as noted above; this was necessary for close-quarters living.

Fifth, emotion was warm and passionate, but kept quiet except in its proper—usually private—place.  People did not tolerate either coldness nor promiscuously open emotionality.

Sixth, family and public rules of deference, respect, and leadership (discussed in the preceding chapter) clearly had much to do with living in close quarters.

These and other mores allowed people to live together in harmony and cooperation, and to enjoy social life without undue stress.  The rest of the world could well learn from these simple rules.


Another question currently debated is that of maternal sentiments.  Is mother love a human universal?  Do the Chinese feel it?

Nancy Scheper-Hughes (1992) denied the existence of innate feelings of mother love or motherly attachment, seeing these as pure cultural construction.  (No one seems to have even considered fathers.)  Scheper-Hughes saw mother love as a western bourgeois idea.

Darwinian theory requires us to think seriously about this.  Even before Darwin,  scholars observed that raising children is so difficult and demanding that essentially no one would raise a child without a powerful innate drive to do it.  In view of the fact that the human population is six and a half billion and increasing at an explosive rate, it seems rather strange to conclude that parental love is a fiction.  Moreover, the western bourgoisie are almost the only group in the world who are at “zero population growth.”  One would expect that if mother love were a western bourgeois notion, only the western bourgeois would go to the trouble of raising children; surely no one else would.  As Nasir ad-Din Tusi put it in (or about) 1235 C.E., long before Darwin:  “An example of natural love is that of the mother for the child:  if this class of love were not innate in the mother’s nature, she would not give nurture to the child, and the survival of the species could not conceivably be effected” (al-Tusi, tr. Wickens, 1964:197).

Feelings, however, would naturally be expected to vary in strength between individuals.  On the genetic level, Darwinian theory leads us to expect variation in all genetically guided traits.  At the phenotypic level, Scheper-Hughes is surely right (and has a long line of forebears) in maintaining that poverty, alcohol and drug abuse, chronic sickness, and general desperation often play against any maternal instincts, hardening many women toward their young.  Bowlby’s attachment theory, and other attachment-through-experience theories, add that mothers who have little contact with their children, such as the elite mothers of Europe in ancien regime days (Hrdy 1999), have little chance to bond and thus to develop full maternal sentiments.

Arthur Wolf, who has written some of the best, most thorough, most insightful, and most psychologically informed ethnography of China, has argued that “maternal sentiments” are not strong in at least some parts of China, and may be nonexistent for many women (A. Wolf 2003).  This conclusion is based on his research on minor marriage:  giving of child-brides or even infant-brides, to be raised by the child-husband’s family.  Working in a part of Taiwan where minor marriage was common, he found that mothers usually gave up their daughters without undue stress or apparent emotion.  From this he concluded that mother love in humans is at best fickle, if it exists at all:  “…bearing a child, nursing it, and caring for it are not alone sufficient to arouse maternal sentiments….  [other authors] may be wrong in arguing that maternal sentiments are ‘a major component of human nature’” (A. Wolf 2003:S40; the scare quote is from John Bowlby, who argued for strong attachment between mother and infant; Bowlby 1988:165.  It is well to remember Bowlby’s point that bonding depends on good circumstances around and after birth.)  Wolf argues that even well-to-do women gave up babies, apparently only for calculated financial advantage, and appeared to do it without concern:  “To be confident that Taiwanese women did not find giving a child away particularly distressful one would need to have observed their concurrent behavior.  I never had an opportunity to do so because by the time I arrived in Taiwan in 1957 adoptions were not common and no longer handled in the traditional manner.  But I have talked to many women who gave their daughters away, and my impression is that this was not a wrenching experience” (A.Wolf 2003:S39).  Wolf was dealing with memories of a long-distant past, and apparently did not make detailed psychological enquiries.

Wolf takes a consciously strong position, stronger even than such questioners of mother love as Nancy Scheper-Hughes (1992) and Sarah Hrdy (1999).  Unlike Scheper-Hughes, who documents the horrific lives of her apparently uncaring mothers, Wolf (like Hrdy) dismisses poverty and harsh life as an explanation.  Wolf and Hrdy maintain well-to-do women were just as guilty—a claim for which one might wish for some direct ethnographic evidence (Hrdy cites only historical records, many of them polemics of exceedingly dubious veracity).

To be sure, many well-to-do women gave up girls in old China and Europe, and Croll shows that many well-to-do women abort girls today.  However, the vast majority of abandoners were either desperately poor or in danger of becoming so.  Wolf (and Croll) somewhat downplay the uncertainty of life in the old days.  Unlike East Asia today, East Asia two or three generations ago was a land of falling economies and falling expectations.  The poor could not afford to give dowries for their daughters, and might let girls die if they could not—as the boat people did—reform the dowry rules.  Even the rich feared the effects of multiple dowries—on top of revolutions, market crashes, wars, famines, and epidemics.

A classic European stereotype, common in the 19th and early 20th centuries, alleged that the Chinese, like other “inferior races” suitable for colonial takeover, were deficient in human feeling, notably mother love.  Writers such as Arthur Smith (1894, 1899) held forth on the callousness and cruelty of the Chinese.  The 19th century authors (Smith is typical) spoke more of infanticide, extremely common in those days (and far from unknown today; Croll 2000) and almost always directed against girls.

Infanticide (usually or always female) and selling off of excess girls had been done in the past, to the enormous heartbreak of the parents, who did it only under family pressure (in cases we knew; see below).  Selective neglect of girls existed in the 1960s.  Girls got less food and more trouble.  However, in many families, a promising girl got more care than a not-so-promising boy.  By the 1970s, girls got full care.  The economic situation had improved; easy access to education had guaranteed that a girl could potentially be as productive as a boy (or more productive if she were brighter).  Gender differences had by no means disappeared, but they had become less obtrusive; they were no longer dangerous to health.

The horror stories of “typical Chinese” callous infanticide and neglect of girls (Croll 2000; Smith 1894; Wolf 2004) certainly did not apply at Castle Peak.  Children were universally loved and valued.  Some families were certainly better and more caring parents than others, and indeed a few parents (there as elsewhere) beat or starved or abused their children, but literally no one among the thousands of people we knew exhibited the callous attitudes that westerners so love to attribute to all Chinese (see e.g. Eastman 1988; Smith 1894).  This is not to say that the sober data recorded by Croll are incorrect—they are, alas, all too true.  Patriarchy was alive and well at Castle Peak.  However, the flamboyant stereotyping by writers like Arthur Smith take the worst of Chinese behavior as the rule.

The Cantonese boat people (and even the Castle Peak Bay land people) may, however, have been unusually far in the other direction (see further discussion in Anderson 1992, 1998).  Certainly, intensive ethnography of some other Chinese communities shows far more severe attitudes toward women—especially in-laws—and children (see e.g. Gates 1987; A. Wolf 2004 and references therein; M. Wolf 1968, 1972), and ethnographic novels like Ba Jin’s, provide an insider’s view that is no less revealing.  Marja Anderson and I ourselves studied a Chinese community in Malaysia that was much more like those harsh communities than like Castle Peak Bay (Anderson and Anderson 1978).  Perhaps significantly, it was largely Hokkien, thus linguistically and culturally closer to the communities known to Gates, Smith, A. Wolf, M. Wolf, and others.

Like minor marriage, infanticide was essentially a way of dealing with the fact that girls marry out, and require dowries.  The parents must not only pay to raise a girl who will never do much for them, but must then pay to get her married off.  Many families cannot afford this, and many more could do it only at the price of substantial economic stress.  They thus kill the girls if they are desperate, or, if less desperate, sell them or adopt them out.  This gave rise to the muitsai (“little sister”) system, in which the girl goes out as a servant or concubine.  More charitable (at least in theory) was minor marriage.

Wolf is, of course, aware of the social-structural and economic reasons for these practices, and his nuanced and comprehensive ethnography is the antithesis of Smith’s crude racism.  He is careful to add European data to show that Europeans too (at least in the past) got rid of children at a high rate[5].

Scheper-Hughes allows poverty and hardship as causative of lack of maternal feeling, but still insists on the reality of the lack.  She appears not to have interviewed her cases according to standard psychological or psychotherapeutic interview methods.  Comparable research by trained psychotherapists known to me (who would wish anonymity here), as well as my own research in clinical settings, in the United States and Canada, indicates that many women do indeed so respond to poverty and substance abuse.

However, concealment or even suppression of real and intense feeling is also common, apparently much commoner than actual callousness.  Often, such hiding of emotion is related to extreme depression.  Therapy can release the hidden emotion in a dramatic, explosive way, as I have repeatedly seen in both clinical and nonclinical contexts.

The sort of brutalization described by Scheper-Hughes makes people withdraw more and more from relationships and emotional engagement, because of fear and stress (Bandura 1982, 1986 explains the general case).  Maternal attachment is probably the last to go, but it too must go eventually, in a few cases.  It is possible that the Taiwanese mothers of whom Wolf speaks were brutalized much more than they cared to admit.  It is also possible that hiding emotions had simply become second nature to them.

Barbara Tuchman (1978:51-52) has commented on the apparent lack of deep attachment of elite parents to their young in medieval Europe—though her claims are somewhat belied by the assumptions of maternal love in medieval literature such as the Lays of Marie de France.  Medieval fathers, though not mothers, tended to be absent much of the time, and often had little chance to bond with children.  So they may indeed have had little parental concern—and of course they were the ones writing the books (with rare exceptions, such as Marie).  On the other hand, Tuchman’s book and other sources document great kin solidarity in those times.  It extended even to adopted, illegitimate, and distant relatives.

In short, the picture of attachment is not simple; attachment varies with personal and cultural conditions, as Bowlby discusses in detail. Bowlby does not have the last word on attachment, but his findings from thousands of data points gathered by meticulous and psychologically informed research can be qualified only by superior data.  On the one hand, it now seems that bonding is a slower and chancier process than Bowlby thought (see esp. Hrdy 1999).  On the other hand, an absolute lack of maternal feeling in a large percentage of otherwise normal women seems beyond Darwinian possibility, and is certainly beyond ethnographic demonstration.

There certainly was a horrific prevalence of female abandonment, infanticide, and neglect in eastern Asia.  Elisabeth Croll, one of the best and most seasoned ethnographers of that tragic region, has summed up the facts (Croll 2000).  Croll establishes that many families, even in the affluent and secure present, still allow girls to die.

However, the questions of who makes the decisions and how the mothers feel remain to be answered.  Marja Anderson and I had many conversations with women who had given up young daughters, and a few with women who had seen their daughters summarily eliminated.

Usually, the boat people strongly oppose the open expression of negative emotions, particularly grief, and most particularly grief over a hopeless situation.  They have an extreme aversion, backed up by sanctions, against expressing such emotions openly.  However, boat people, in many situations, do feel it appropriate to open up.

Our interviewing was informal, private, detailed, sustained, and gentle.  We provided a calm, quiet situation and got the women to talk—minimizing actual questions.  This eliminates the possibility that we led the women into emotionality by our questioning pattern.  Actually, we tried with considerable success to get women to talk on their own, without questions intruding.  We also were privy to several interactions in which we did no questioning at all.

In such situations, after several minutes to hour, all women without exception would become highly emotional, and these interviews almost invariably ended in sobs.  These women, especially those involved in infanticide, had clearly never even begun to resolve the grief and pain.  They kept silent, dealing with it as best they could.  One saw why silence about grief was so necessary in old China.  No one could have borne the suffering otherwise.  Explicit statement of this truth is, in fact, a standard trope in classical Chinese literature.

We observed one dramatic, and thoroughly public, confrontation between a mother and her adult daughter whom she had given up shortly after birth.  Being public, this had to be managed through improvised folksong, the appropriate vehicle for expressing grief (e.g. at funerals and wakes, as well as situations like this one).  It went on for about two hours.  The singing broke into rough chanting, alternating with weeping and sobs.  Quite apart from the emotion of the two participants, the watching crowd all clearly recognized the emotions as normal, and empathized thoroughly; many of the women joined in the tears[6].

In these cases, the decisions to give up or kill girls had been made by the family as a whole—which usually meant, in reality, the husband’s parents.  They would persuade or force the husband and wife.  Some wives and mothers stood steadfast, but more often economics intervened; one had to face the fact that keeping one more child could mean that all went without, and all were in danger of dying.  We knew families where this was a genuine and serious concern.  They usually kept all the children anyway, and took their chances—but that was in the expanding Hong Kong economy of the 1960s and 1970s.  Even so, we saw children die of malnutrition.

One can only conclude, from this and other evidence (including the comments by other scholars appended to Wolf’s article—those by Chinese scholars in particular), that maternal sentiments are usually real and powerful, in China and elsewhere.  People bear what they can, and often bear in silence.

George Bonanno (2004) has recently reviewed a vast amount of evidence and concluded that people may feel grief very deeply and yet survive through sheer grit:  “Resilience to the unsettling effects of interpersonal loss is not rare but relatively common, does not appear to indicate pathology but rather healthy adjustment, and does not lead to delayed grief reactions” (Bonanno 2004:23).  Even post-traumatic stress disorder is much less common than the incidence of traumatic stress would lead us to expect.  Yet, Bonanno makes it clear that for these resilient copers the grief and suffering are real.

One might expect that parents would engage in psychological distancing if they expected the child to die.  Scheper-Hughes found this in Brazil, and I have observed at least the attempt to do it not only in Hong Kong, but also in Mexico and elsewhere.  However, parents’ success at this is generally (though far from always) poor.  Folk wisdom in Mexico holds that women have to tough it out (aguantar) as best they can, and for their own sakes and others they need to hide emotion.  The fact that hiding is deliberate, forced, and difficult is universally recognized in folksongs and other popular culture (Anderson 2005).

Certainly, culture can fine-tune or even create emotions.  Human feelings are subject to all manner of modifcations, cultural and otherwise.  The point here is not that culture is irrelevant or that emotions are some essential quality in the human animal; extreme innatism simply does not stand up under research.  However, it seems doubtful, if only on Darwinian grounds, that any large segment of the human race is without parental sentiments.  Certainly, we never encountered Chinese mothers lacking in maternal concern.  No doubt some Taiwanese women stifled any feelings thay may have had; perhaps some were indeed without feelings.  Taiwanese society is different enough from Cantonese that it may be less emotional, more disengaged—but the many ethnographies of Taiwan, including those by Wolf and his group, do not seem to show this.

People are tough.  They survive.  Hurts can be stuffed down, even more effectively than Freudians and Bowlbians knew.  Chinese mothers, and many more in this imperfect world, must often make horrible choices deliberately and with clear eyes, and keep their mouths shut about the consequences.  So does an animal that coolly gnaws off its leg to escape a trap.

I wish all social scientists knew Bertolt Brecht’s long and wrenching “Ballad of the Infanticide Marie Farrer.”  It ends:

“…her sin was heavy, but her pain was great.

Therefore I beg you not to fall in scorn;

We all need help from every creature born”  (Brecht 1947:26, my translation more or less following H. R. Hays’ on p. 27).















The boat people moved in a wider world, in which political identifications of various kinds cut across the boat-land distinction.

Whether maritime or land-based, a person from the Castle Peak area identified himself or herself as a Ching Saan yan, a person of Green Mountain (Castle Peak).  The area so labeled was defined by marketing structure.  As William Skinner pointed out (Skinner 1964-1965; see also the papers in Skinner 1977), China’s market geography was and is a complex set of polygons, each polygon fitting with its neighbors in theoretically neat tiling-patterns.  These polygons are centered on marketing areas.  They form a nested hierarchy, from local markets to national capitals.  Hugh Baker (1979) pointed out that, in reality, the polygons were often more like circles, separated by splinters of unused and uninhabited land; these slivers served, quite literally, as grounds for competition.  Villages tried to expand cultivation into these debatable zones.  Before the 20th century, neighboring villages often fought over the right to expand into such borderlands.

People marketed at a certain town, and within its sphere they contract marriages, transact business, find cooperation and sociability, and organize festivals.  Skinner’s use of “marketing” as the defining term betrays a certain economic determinism; a sociologist might speak of “interaction.”  But “marketing” was the Chinese term, too, as the name of San Hui (“New Market”) makes clear.  San Hui was a typical “standard market town,” center of such an area.  As such, it could supply everything that anyone needed daily or weekly, and almost everything that ordinary people ever needed.

Above it in a hierarchy of markets was Yuen Long, the central town for the western New Territories.  (Skinner’s student John Young, 1974, provided a superb study of its marketing functions.  This book must be read by anyone interested in Chinese marketing in earlier times.)  Yuen Long supplied the few things that ordinary people occasionally needed but that San Hui was not big enough to provide.  These included special government services, full emergency medical care, major manufactured and crafts goods, and such specialized services as poultry-castration (only on market days!).  As a large town, Yuen Long was part of a rotating series of major markets.  It was, if I remember aright, a “1-4-7” town:  major markets were held, in a huge open space at the edge of town, on the 1st, 4th, 7th, 11th, 14th, 17th, 21st, 24th, and 27th days of the month.  The nearest towns would then be a 2-5-8 and a 3-6-9 town.  Thus, a triangle of coordinated markets was established.  The poultry castrator, and other specialized vendors of goods and services, traveled around the circuit, resting or going to ordinary everyday markets on the zero days.  Such towns supplied things that ordinary people needed only once or twice a year (as well as all the ordinary needs).

Still higher on the scale of markets was the great Kowloon-Hong Kong metropolis, valuable to Castle Peak residents as a place for rare, exotic recreation and as a place to sell fish.  In the 1960s, many people had never been there, or had been only once.  Travel was much more common by 1974.  Such great centers had not only daily and yearly needs, but everything anyone might ever need.  To the goods and services available at lesser market towns, they added the goods and services that a normal person needed only once in a lifetime, and goods and services that special people (big businessmen, political leaders, and the like) needed more regularly.

Below San Hui in the hierarchy came local markets like the bay pier.  Here, one found items like rice, oil, salt, and tea, as well as incense and religious services, restaurants and cafes, and fish dealers—in short, the things that everybody needed every day, but nothing much beyond that.

Smallest in the marketing hierarchy were the informal daily gatherings of people at crossroads, to trade or haggle over a few vegetables, fish, or herbs that their gardens had provided, and to sell home-made snacks.  Pedlars brought folk medicines and cheap cloth.

The local markets defined three sharply separate subsections of the Castle Peak residents:  the east bay, the west bay, and the San Hui area itself.  The fishermen almost all lived in the east bay, focused on the pier.  Boat people in the other two areas had given up fishing.  They found work in the factories and service jobs around San Hui.  By 1974, the factories were booming; the bay had already been half filled.  Public apartment housing had been constructed on the former tideflats.  Private houses decorated the higher ground.  All the unfortunates living in wrecked boats or shacks in the 1960s were now well housed, usually in the new public housing, and well employed.  One exceptionally bright girl we befriended when she was a starving child in 1965 was a waitress in a restaurant in 1975, and moving up toward better opportunities.

The outcaste community described in the previous chapter had also disappeared by 1974.  Its people found homes on land, and joined the factory and service workforce.

Within the fishing community, boat type influenced sociability.  Trawlers socialized with trawlers, hand-liners with hand-liners.  Religious associations, coops, and informal leadership groups cut across these lines, and everyone met in the shops and restaurants of the pier area.

The boat people were highly mobile, and had constant interaction with people of Tai O, Ma Wan, and a couple of dozen other small or large ports of the western New Territories.  If the fishing were good or if relatives were there, a family might temporarily move to one of those ports.  Land people were more apt to have ties with other land communities, and with the larger centers, Yuen Long and Kowloon.

A community was thus less a matter of physical residence than of social interaction, and social interaction was defined, in large measure, by marketing and commercial links.

However, this did not quite exhaust the structuring of community.  Skinner tended to see religious structures as simply mirroring the marketing ones.  Some did, notably the cult of the “Earth Gods” (spirits of the locality) and, of course, the gods of the towns themselves.  However, other cults cross-cut the marketing grain, and provided another way of linking people.  At Castle Peak Bay, not only the local boat people but all the boat people from Kowloon westward came to the Tin Hau temples.  Land-dwellers from wide-flung areas came to the Three Sages Temple near the pier, and land-dwellers often worshiped Tin Hau, also.  Devout or desperate worshipers traveled farther afield, to famous temples at Tai O, Kowloon, and elsewhere.  Conversely, urbanites came to the great Taoist and Buddhist establishments in the area—but locals rarely did, for these were more sophisticated temples, for the well-educated.  Religion also integrated boat and land people, and was the only institution that did so (see following chapter).  Thus, religion served to weave the social strands into a fabric, by integrating people otherwise separated by ethnicity and market boundaries.  Judging from my experience and reading, this finding holds true throughout China.

In addition to the impersonal ties of marketing and temple membership, the world outside the family was bound by a spiderweb of friendship links.  The Cantonese word p’ang yau, generally translated “friend,” actually means “acquaintance.”  At Castle Peak, at least, it meant anyone that one knew well enough to engage in conversation without some special occasion to serve as an excuse.  By contrast, “friend” is hou p’ang yau (“good acquaintance”).  These are people with whom one trades, has tea, celebrates birthdays, talks endlessly, plays mahjongg, and generally enjoys good fellowship.  Friends were expected to look after each other; mutual aid and care were universal.  Borrowing was usually from kin, because neighbors might not be neighbors long in this mobile world, but still a great deal of borrowing and mutual favor-exchanging took place between unrelated friends.

Acquaintances and friends could be quite shameless about asking for favors, but this was reasonable in the “floating world.”  The asker knew that the chance might never come again.  The person asked knew that he or she might desperately need the other’s aid in a day, or a month, or a year.  One needed all the contacts one could get.  One never knew whose boat would be the nearest when one’s own boat wrecked.



Leadership in the boat community was a matter of some ideological ambiguity.  In nothing did the boat people agree more totally and thoroughly than in their rejection of all top-down control.  They resented government, police, and domination by local powers, whether legal (town officials) or informal (fish-dealers, local rich folk).

Of course, necessarily, they did not live as anarchistically as they talked.  They submitted to the British government, because it generally facilitated their life and fishing, rather than interfering with those.  They recognized that some sort of government was necessary to maintain markets, mark harbors, and so forth.  But they did not submit quietly or eagerly.  They talked nostalgically of unsettled times in the 1930s and late 1940s when they were outside effective control by any government.  They justified such opposition by quoting the more libertarian passages of Mencius.  That philosopher justified rebelling against evil rulers, and argued for liberty of conscience to a degree very rare in world literature.

Often, the boat people sounded strikingly similar to China’s home-grown anarchist philosophers of the early 20th century (Scalapino and Yu 1961).  The boat people had never heard of Kropotkin or Emma Goldman, but Mencius was quite enough.  Surely, few people in human history were so uncompromisingly against all hierarchy, all top-down control, and all interference with personal liberty.  This was not the American “libertarian” philosophy, which opposes big government but idealizes big business, including government support thereof.  The boat people opposed big business, big government, and anything else above the grassroots level, with equal fierceness.

A long history of free-trading, smuggling, and outright piracy lay behind all this, but it was primarily the result of the needs of the fishery.  A boat had to be a little republic.  People had to be self-sufficient, self-reliant, and self-determining.  They had to be able to fish where they wanted, and to control their own fleets.  They had to be able to defend themselves against pirates and rivals, without waiting for distant and sluggish government agents to do it.

Almost equally important, in the minds of the boat people, was the eternal and inevitable dominance of government by land-dwellers.  Even those boat people who did not hate and fear the land people were most unenthusiastic about this dominance.  The land people had always used their position of authority to impose unfair terms, and the boat people thus wished to escape land-based control as much as possible.

However, no human community can really function or survive without a leadership structure.  Consistency thus could go only so far, and stated principles that sounded anarchic had to be qualified strongly in the real world.  The boat people made use of government when it suited them, as in the cases of the cooperatives and the government fish market.  Increasingly, they came to rely on government schools, government harbor and channel controls, and other infrastructure services.

Most of the fishermen at Castle Peak had come as refugees from Communist China.  Almost all came for economic, not political, reasons.  They described rations of two pieces of cloth a year and a catty of rice a day; neither were remotely close to adequate for a fisher’s life.  They continued to fish in Chinese waters, under license, paying a small tax.  However, in late 1965, just as we moved onto our boat, the Communist government closed all waters.  The fishermen thought is was largely an attempt to make them come home.  They did not; they simply poached.  Rolex and other imported watches began to appear, and the following dialogue took place between myself and one fisherman:  “Oh, nice new watch.”  “Yes, it’s for the cadre.”  “But I thought there was no bribery in Communist China…?”  “Oh, yes, but this is only a little gift.”  So fishing went on.  “But doesn’t this mean money is more important than government policy?”  “Of course, what else?”  The cadres did insist on the boats having Chinese licenses.

The boat people were neutral about the political philosophies of Communists, British imperialists, Americans, and others; they disliked all governments and liked all people.  In spite of the years of anti-American propaganda, we had little difficulty with the refugee community.  Only those who maintained active Chinese Communist party membership and organizing enterprise saw us as dubiously desirable.

The boat people viewed Hong Kong’s marine police with fear and loathing.  These police interfered with smuggling, collected the high license and harbor fees required by Hong Kong, and had to enforce the countless minor maritime laws adapted for bigger and more demanding vessels:  laws requiring formally trained personnel, special equipment, etc., to be on board.  The boat people accepted gladly an increasing range of government services, however:  typhoon signals, typhoon shelters, land-based police who maintained order and stopped gangsters from abusing boat women on shore, public health measures, public education, and other trappings of the modern state.              More important than these, at least in 1965-66, was the actual leadership within the boat community.  (By 1974-75, boat life was in decline and government power on the rise.)  Needless to say, boat community leadership could not be formalized, or even overtly recognized.  It was none the less real.  People recognized certain fishermen as leaders, using the terms a yat, “number one,” or taai yan, “big man.”  Marshall Sahlins, theorist of the “Big Man” (Sahlins 1972), would have been pleased.  Sahlins used the term exactly the way the boat people did:  for a man who waxed large in the social picture through his own efforts.  (Sahlins further noted that such a man was often literally big, thanks to many a feast.  The boat people, hard workers that they were, almost never got fat, even when socially considerable.)  Such terms were more often used for powerful land people, and with no friendly implication, but the boat people admitted that they too had their big men.  I found it sometimes difficult to get them to admit this, even when no one had the slightest question that the phenomenon was real.

Many women had real political and social power, but all recognized leaders were men, and the public sphere was a male realm.  On the other hand, everyone knew which women had power, and everyone knew that the wives of male leaders were often full or nearly-full partners in much of the leaders’ decision-making.

The path of a boat leader was of a strictly grassroots order.  A successful boat captain—that is, male head of family—would find himself sought out by relatives and neighbors for help.  If he succeeded in solving their problems, he gained a reputation, which could spread.  Most such grassroots leaders remained known only to close friends and immediate kin, but some rose higher.

The boat community of Ma Wan Island, a small, isolated island a few miles from Castle Peak, recognized as leader and spokesman the dynamic young Yip Yautaai.  He had begun as a particularly caring and attentive solver of neighbors’ problems.  He soon found himself spokesman for the community—a very poor and remote one, poorly served in all respects—in its attempts to improve.  He organized a coop and a Better Living Society that involved all the fishermen.  He got government services and assistance.  He then succeeded in finding nongovernmental charitable organizations that would help with housing and livelihood.  Soon the fishermen moved from their boats into well-built apartments on shore, built with the help of Canadian and other charitable associations.  Throughout all this, Yip remained modest, friendly, and unassuming. Yip’s leadership and ability in all this made him known and respected throughout the boat communities of the western New Territories.  He was widely held up as a living model of a boat leader.

Yip’s achievements were unusual, but effective leaders could be found in many ports.  Castle Peak remained somewhat unorganized; there were many small leaders but no powerful ones.  This had much to do with the divisions of the fleet and the lack of major community-wide problems to solve.  There were many boat types and many sub-communities.  Leaders rose to the level of representing 20 or 30 boats, but rarely could find much scope to move higher, or much competitive advantage over other leaders.  Of course, it also had much to do with the fact that a leader of 20 or more boats attracted many attempts to cut him down to size!

A leader had to deal with issues of face (min, Putonghua mian).  This Chinese term has developed a quite wrong connotation in English.  English-speakers have adopted the word to refer to purely superficial, often hypocritical, social slickness and prestige.  This is not how the term is used in Chinese, or at least in Cantonese (see Moore 1981).  For Cantonese, the term refers to social reputation based on actual skill at negotiating with fairness, honesty, and competence.  To be sure, a well-to-do individual with competence, honor, and connections can have face and still be a shady operator.

However, a boat person—lacking wealth, important connections, and chances to develop suaveness—could only rely on native honesty, reliability, and social skills.  He also had to have considerable courage.  He had to be level-headed, thoughtful, and judicious.  The risks—social and even physical—were not inconsiderable.  A man who blew his cool would never be a leader; he would be more apt to lose his boat.  Also, a leader had to be a good listener, since he had to deal with all sorts of people from all social spheres.  The boat world had its shady dealers and people who cut moral corners, but they had little face.  Many feared them, but few respected them.

In so far as “face” implied a slick surface as well as an integrity beneath, the term lin (lian) was pressed into service, to refer more specifically to the latter (see Moore 1981, 1988).  Face was a social quality, lin a more personal one.

Face comes from successfully navigating the difficult and slippery world of social interactions and transactions.  It is a quality that emerges from those.

To gain face, a leader had to resolve social dealings, conflicts, and disputes in such manner that others also gained face.  The typical case was a conflict among the leader’s followers, or between a follower and an outsider to the circle.  If the leader could resolve such a conflict to everyone’s satisfaction, leaving no one feeling wronged and embittered, all parties—especially the leader himself—gained face.  If the leader left the parties unsatisfied and angry, he lost face.  Negotiating non-conflict situations, such as marriage negotiations, government dealings, and organizing for festivals, also involved the opportunity to gain or lose face.

Losing face also came from any public shame, especially being discovered in a dishonorable or compromising action.  Deadly, also, was being played for a fool—wan pan chat, “played for a stupid prick,” in the blunt language of the waterfront.  A pan chat was one who had tried to pull off a social deal, but had been outsmarted because of his own stupidity.  This was even worse if he had tried to cut moral corners in the process.  Even trivial matters like not being able to respond cleverly to an insult, or being cheated in a minor transaction that anyone should have handled, made one a bit of a pan chat.

Being a running dog (informer), cheat, or deadbeat were also sure ways to lose all face in short order.  Being chou, “coarse,” here meaning “socially insensitive,” was costly to a leader, though not necessarily face-losing to an ordinary person.  A leader had to have sensitivity and polish, and had to keep calm and balanced.  General bias and uncontrolled anger were disliked in everyone and fatal to a leader’s position.  By contrast, bias in favor of one’s clients and carefully stage-managed anger in defense of them were desirable.  (It gave one face to display obviously stage-managed anger, but genuinely losing it and getting seriously angry was a face-losing proposition.  Few facts give more insight into waterfront sociability than this one.)

Crumbling under pressure was always fatal to leadership.

Honesty and fidelity to one’s followers was obviously desirable, but honesty in land-based terms was not necessary.  Yip Yautaai and many other leaders were model citizens by any standards, but many leaders were, or had once been, smugglers, pirates, and the like.  Such men were feared and not always trusted, but they were often leaders all the same; land-people’s standards of good behavior did not necessary mean anything one way or another.

Finally, a boat leader had to have something not at all related to face:  he had to be a distinctive “character.”  An ordinary, conformist, pleasant sort would never be a leader.  A leader had to have a reputation for something distinctive:  pungent wit, great knowledge of and insight into people, great knowledge of folk songs and stories, hard drinking, bonhomous feastmanship, or the like.  He had to stand out as an exciting, interesting person, revelling in life and in originality.

In short, a leader was subject to countless democratic pressures.  He had to be successful and yet not proud or haughty.  The boat people took delight in cutting down to size anyone who overreached himself.

Someone who could successfully deal with such issues soon became the spokesman for his group in dealings with outsiders, be they other boat people, government workers, businessmen, or wandering anthropologists.

A small-time leader tended to be trapped by the size of his operation; he could do only so much with his resources.  Yip, a less than affluent member of a modest community, was exceptional in being able to parlay sheer personal skill into a great deal of improvement for his people.

Leaders developed regular patterns of morning tea-drinking, so that they could be found by anyone needing their help.  A leader would have regular hours at a particular table in a particular restaurant.  Morning was the time for this, because the leader would usually have to go out to fish, or else would be taken up with social and political dealings later in the afternoon.  Part of the responsibility of a leader was attending weddings, religious festivals, and other rituals involving group members, so he had a busy life after morning teatime.  Certain tea shops and restaurants attracted leaders, others did not.  The attractive ones were those that were large, open almost all the time, and welcoming to boat people.  Boat leaders knew the restaurants to visit in Yuen Long and Kowloon as well as in all the western New Territories ports.

The exigencies of a world of highly independent people, who valued their agency and took no orders from anyone, inevitably and necessarily elevated such negotiation-experts to leadership positions.  The situation created a specific morality:  one valuing conflict resolution above almost all other things.  Independence and autarky guaranteed that conflicts would be frequent.  The needs of mutual aid and mutual dependence, and of harmony, guaranteed that solutions to those conflicts would have to be found.  This in turn guaranteed that good problem-solvers would be the pillars of the community, and their social skills would be highly valued.  The boat people were quite aware that an independent community of independent people needs more, not less, self-conscious harmony-maintenance than a community of policed conformists.   The dangers of a Hobbesian meltdown were obvious.

Thus, the boat communities of Hong Kong were amazingly orderly; of course, the efficient colonial government was also responsible, but the boat communities were largely self-regulating.  However, without the police and other trappings of modernity, leadership might not have had such beneficent consequences.  Antony (2003) describes the rise of several pirate leaders in the South China Sea who seem to have risen by similar means—in addition to ruthlessness and predatory savagery.

Other ethnographies of China have described somewhat similar processes.  In particular, the excellent, sensitive accounts in Kipnis (1997) and Yan (1996) relieve me of the necessity of giving a complete account of networking in Chinese communities.  However, in more ordered, hierarchic communities the focus shifts to drawing on “connections” (see below) and manipulating hierarchical relationships and formal political rules.  This was true of land-dwelling communities in Hong Kong as well as in Kipnis’ and other communities.  This gives a very different flavor to the process.  At best, it is more tightly bound to formal political and social rules (see Kipnis’ account).  At worst, it becomes corruption and manipulation (Yang 1994).  As in other cases, one can see the boat people of Castle Peak Bay as set within a continuum:  from accommodation to formal political rule, with corruption and favoritism as the pathological extreme, to complete grassroots autonomy, with piracy as the pathological extreme.

Leaders, of whatever sort, had to build up networks of friendship and acquaintanceship (pang yau), obligation (Putonghua guanxi, a word rare in its Cantonese form), and mutual good feeling (kamching, Putonghua ganqing).  These were the common coin of interaction.  A leader would throw feasts, invite friends and other leaders to dinner, help with religious rituals, provide charity, help in construction or engine repair, and otherwise share.  Most critical were feasts—celebratory dinners, often in restaurants.  (See Sahlins 1972 on the worldwide importance of special food in such matters.  Ordinary dinners would not count.)  More directly useful to clients were such matters as bringing a problem to the attention of the proper powers, and using one’s authority to make sure they attended to it.  An ordinary person could not get a permit to build a shack on shore, for instance, but a leader might.  A leader had to be able to travel to Yuen Long and elsewhere to represent clients too poor or occupied to travel far.  Dealing with the authorities in Yuen Long was frightening to the boat people.  I have seen veteran leaders shake and turn pale in a Yuen Long office.

A boat leader thus had to have some money, but surprisingly poor people got very far.  Yip Yautaai started with little more than a tiny boat and a change of rather worn clothes.  Our sample of leaders disclosed no correlation with wealth.  Better-off boat people had usually accumulated enough enemies to offset the advantages of money.

Perhaps commonest of all, though most unobtrusive, was providing information.  A leader was a node in the vast communication network that linked the whole boat world, from Guangzhou to Aberdeen and Macau, into one great linkage.  Clients expected their leader to be the main source of valuable information about the wider world.  Other leaders expected him to be the source of new and important items that his client network had discovered.  This could be anything from new fishing grounds to an impending marriage or a threatened government crackdown.

Sometimes leadership was courted, sometimes shunned.  In neither case was success automatic.  Leadership was negotiated with the potential clients.  They would not put up with an unskilled social dealer.  Conversely, they would force a consummately skilled man into the role, whether or not he wanted the rather onerous responsibilities.

Leaders we knew varied from people who were simply outgoing, friendly, and cheerful to people who were quietly competent—people who could talk a suspicious official into giving an honest, sympathetic hearing to problems that would have normally been beneath notice.

A trivial and thoroughly typical case exemplifies leadership at work.  A trawler from Castle Peak accidentally ran over the nets of a gill-netter from Ma Wan, ruining them.  They argued over payment; it was a clear accident, but the nets were a major loss.  The trawler was a friend of a minor Castle Peak leader, Leung Taai; the gill-netter was part of Yip Yautaai’s community.  Thus, the conflict went right to those two men.  For days they became more and more edgy, seeming to look for trouble with each other while dreading the possibility of finding it.  Both these leaders were good friends of my field assistant and myself, so we arranged a feast and invited them.  Over drinks (and they could drink as only sailors can), they worked out a compromise deal, “to give you [my field assistant and I] face.”  They were helping us become leaders too, by letting us provide the excuse for them to settle their grievance.

Grassroots leadership of this sort was successful at reducing trouble, and the local police (land-dwellers) told us that the boat people gave them less trouble than the land people did.  But grassroots leadership proved totally helpless in the face of overfishing.  The fishermen told us explicitly that they knew the fish were disappearing, but simply could not stand to put themselves under the control necessary to stop overfishing.  Unlike so many grassroots communities round the world (Anderson 1996; Berkes 1999), the fishermen could not agree on limiting their fishing.  It was a suicidal liberty.  Within a generation the fishery was destroyed, and the boat world was gone forever.



All this brings up the question of moral behavior.  Since that has been treated in great detail elsewhere (Anderson et al. 2000), I will be brief here.

A good person (hou yan) had the characteristics given above for a leader, but did not need to have—and did not normally have—so much social skill or personal courage.  He or she was responsible, first of all:  a good fisher, good family man or woman, good participant in the endless small exchanges of mutual aid that held the waterfront together.  A good person incurred obligations and discharged them faithfully.  Intelligence got praise, but more praise went to the person who thought carefully and judiciously through problems, listened to others’ opinions, and came to a sensible conclusion.  Respect of others, especially one’s elders, was absolutely essential.  On the other hand, the boat people only modestly favored the Confucian virtue of filial piety.  Deferring to elders in the family was assumed; blind obedience or excessive deference to them brought no praise (see previous chapter).  A good person was not somebody who did anything blindly.  Goodness lay in actively, thoughtfully helping others.

Goodness also involved tolerance.  People lived at close quarters.  They had to tolerate a sometimes overwhelming amount of noise, smoke, and other human annoyances.  They had to have firm rules for dealing with behavior that was completely out of line; everyone would intervene in cases of real danger, but everyone would leave everyone else alone otherwise.  This meant not noticing a most impressive range of human behaviors.  Not only defecation, but also sex, frequently could not be hidden on rocking boats or in thin-walled, close-packed houses.  The hou yan was imperturbable and tactful.

A bad person was an m hou yan, “not good person.”  People carefully avoided using words that literally meant “bad” and “evil.”  Such words were too insulting, but, also, using them could call up dangerous supernatural evil forces.  If the politely circumspect “not good” was inadequate—i.e., if the speaker were truly angry—the villain in question became a “rat,” “dog,” “devil” (kuai), or other nonhuman.  Only truly evil deeds (not persons or things) were “evil.”

“Not good” people were aggressive and tactless, but above all irresponsible, untrustworthy, and unhelpful.  They did not play their part in the endless round of incurring and discharging obligations.  Perhaps the ultimate badness was to fail of an obligation.  Any dishonesty, faithlessness, laziness, and violently aggressive behavior attracted extreme negative judgement.  Acts of this sort became known to the entire waterfront within hours of taking place, a fact which certainly had a major effect on social order.

A different kind of badness was “ignorance” (mou man “lacking culture,” or equivalent words).  An ignorant person was rude, disrespectful, and mean.  He or she was openly intolerant, insulting, or shameful.  He or she carried gossip, said stupidly tactless things, and hurt others’ feelings for no reason.  This, of course, loses face.  At first I wondered at the term “ignorant”—what did education have to do with it?  I learned that mou man refers not to schoolbook education, but to learning proper social behavior.

Illegal behavior did not necessarily make one bad.  Smoking opium and other criminal behavior led eventually to ostracism, though a bit of such was tolerated in some quarters.  Running snakes (smuggling refugees out of China), however, could make one a good person, not a bad one, though it was a crime; everyone believed boat people in China were suffering and needed any help they could get.

Heavy drinking received no censure.  It went with the sailor’s life, and people put away incredible quantities of alcohol at feasts.  (At other times, alcohol on the boats was  rare.)  An alcoholic would have been censured, but we literally never heard of or saw an alcoholic among the boat people, and we knew over a thousand people.  We knew one alcoholic shore-dweller very well; in 1965-66 he was seriously alcoholic, but just before we returned in 1974, his doctor told him to stop, and he stopped—just like that.  We knew, barely, a few other alcoholics, but the number was incredibly small for such a large community.

The secret of this low incidence of alcoholism in a hard-drinking town was simple; it lay in a set of rules that were enforced by unanimous and condignly enforced social opinion.  Drinking was absolutely confined to adults.  Heavy drinking was confined to senior adults.  Drinking, beyond an occasional beer with dinner or peg of raw alcohol on a cold morning, was confined to feasts, and these had to be for a serious occasion (religious or personal).  Drinking had to be with a large amount of fairly oily food, and both the drinking and the eating had to be spaced out over several hours.  These rules sufficed.

Far more serious was sorcery.  Everyone feared it above almost all things.  It was correspondingly rare. We knew two people seriously accused of it.  Both of them gladly admitted to the charge (see following chapter).   They were clearly mentally deranged, but people called them m hou yan rather than song (“crazy”), because their sorcery defined them.

Cynical proverbs, recited when someone let someone down, may be a good way to show values in action.  “Feelings are thinner than paper,” said one.  “Cantonese have no pity” occurs in a song.  Conversely, “even thieves have morals” reminds us that even bad people can be good; the Mencian idea that everyone was good deep inside, however much education they needed to bring it out, was the most widely known, widely cited, and widely held of Mencius’ teachings.  Even those who were unclear about his name knew that the sages held that “people are good.”

All this is very Confucian, but it is not the Confucianism of modern elite discourse.  It is, in fact, a quite different construction of the morality discussed in Confucius, Mencius, and other classic Confucian writers.  Since the days of Zhu Xi (1130-1200), the great Neo-Confucian philosopher who did much to define Chinese moral culture, elites have often interpreted Confucianism to mean obedience of inferiors to superiors.  Yet, the classics spent at least as much time discussing fairness, justice, mutual aid, community spirit, civic culture, and integrity.  They also couch deference in terms of mutual support and good (virtuous) order, not in terms of blind obedience to any and every order.  As in matters of family, the boat people had a perfectly valid, thoroughly Confucian, and thoroughly Chinese way of being good—but it is, once again, very far from the stereotypes promulgated in certain books, both Chinese and western, in recent decades.



In recent decades, ethnicity has been as overwhelmingly central a topic for ethnography in China as kinship was in the 1960s and 1970s.  Especially important has been study of the differences between the highly formal and often arbitrary “minority” codings and discourse of China’s Communist government and the far more fluid, diverse, complex relationships that the ethnographers find on the ground.  This issue has led to the production of an enormous literature (e.g.—and this is only scratching the surface—Blum 2001; Brown 1996; Constable 1996; Gladney 1991, 1998; Harrell 1995, 2001; Litzinger 2000; Mueggler 2001; Schein 2000).  These studies have been important far beyond the boundaries of China, because they provide, collectively, such a superb picture of the ways an arbitrary top-down bureaucracy makes a simplified, authoritarian grid from a rich and intricate net of actual relationships.

Ethnic relations in waterfront communities in old Hong Kong were characterized by great fluidity.  It is no accident that the modern performative, anti-essentialist concept of ethnicity was pioneered by Fred Blake in his brilliant and pathbreaking studies of ethnic relations in Hong Kong (Blake 1981).  Blake’s pioneer work came well before the “postmodernists” and their ilk “discovered” this notion, and deserves more recognition than it has received.  Blake, in his studies of a Hakka town and its relations with Cantonese and other Hong Kong groups, found that individuals could and did “pass” as members of various different ethnic groups by manipulating symbols:  clothing, food, accent, association, and so on.  Often, ethnicity was a role like a role in a play, to be assumed when strategically useful.

In general, the findings of all these scholars fit with the views of James Scott (1998) and George Orwell: the power state simplifies not only for ease and convenience, but also to maintain control through a type of mental brutality.  All the authors see ethnicity as emerging from interactive practice, not as some sort of essentialist bred-in-the-bone reality.  Most see the Communists as using ethnicity for control, and being highly arbitrary in their application of ethnic labels.  For instance, the same people may be classified in two totally different ethnic groups when divided by a state line; it may literally happen that brother is set against brother, officially.  Native Americans may recognize parallels along the US-Canada border.

In China, as elsewhere, “ethnicity” is a political thing, not a cultural reality.  Ethnic groups are defined by politics.  Usually there is some cultural basis.  In China, most “minorities” and Han Chinese groups are linguistically defined, or are supposedly so.  Obviously, speakers of the same language who are classed separately on different sides of a province boundary may have something to say about that.  At least China seems free of such monstrosities as the American ethnic group “Asian-Pacific Islander,” which lacks any cultural, linguistic, political, or genetic unity, and was established purely for convenience by the U. S. Census Bureau.

The boat people are, of course, an “ethnic group”—not recognized as a minority—defined by residence and subculture.  One is reminded of the “gypsies” or “traveling folk” of the British Isles; the ethnic labels include not only Roma but other, quite unrelated, traveling groups.  Only the mobile lifestyle provides a unity.

Within the recent literature on Chinese minorities and ethnicity, there is a range of opinions.  Some scholars, such as Harrell, have a general sense that ethnicity is real, but much more complex and linked to real-world practice than the Communists allow.   Other scholars appear to see ethnic labeling as inevitably arbitrary, a reductionist classification that may or may not have any basis at all.  Certainly, languages, costumes, and other cultural markers differ, but they may not covary, and they may not go with locally recognized ethnic labels.

Many, perhaps most, are critical, explicitly or implicitly, of the zoo-like exhibitions of “minority culture” that the Communists have long favored.  On the other hand, these shows may at least having some value in keeping cultural arts and expressive means alive, and even allowing survival of minorities that might otherwise be suppressed.

I have little to add to these excellent studies, partly because I have not done much research in Communist-controlled China.  I have already touched on the nature of boat people ethnicity, and on relations with other ethnic groups, and with governments (see above; also Anderson 1967, 1970b).  Suffice it to say that the boat people are among the least marked and most arbitrarily defined of ethnic groups, and that the classification of them as a “tribe,” the “Tan,” with a long “history,” is a particularly extreme case of ethnic essentialization and the invention of an “imagined community” (Benedict Anderson 1991).

Studies of such groups, defined by geography, ecology, village, or other identities, are much rarer than studies of ethnic groups marked by truly different culture.  An excellent comparative volume edited by Tao Tao Liu and David Faure (1996) does consider occupational and geographic subcultures along with “minorities.”   Also relevant are Dru Gladney’s studies of Chinese Muslims (Gladney 1991, 1998).  These have no linguistic, social, political, or cultural unity.  They are not even necessarily Muslims, since those who have fallen away from the faith may still count, for an uncertain number of generations.  Yet they are properly bureaucratized as the “Hui Minority,” with all the appropriate formalization.  Gladney supplies much detail on the inevitable problems that arise as the government tries to deal with this amorphous group, while the Hui try to revitalize their Islamic faith and society.

Chinese essentialization of ethnicity began long before the Communist government.  Labeling and arbitrary bureaucratic classification of ethnic groups has a tradition thousands of years long in China, going back to the Zhou and Han Dynasties’ attempts to label the “Rong barbarians,” “Hundred Yue,” and so forth.  It has gotten more rigid over time, the Qing Dynasty being a watershed period for ethnic sclerotization.  (A masterful, but alas unpublished, manuscript by Jeffrey Barlow on the Zhuang makes this point particularly well.)  The Communists merely continued an age-old tradition, adding Stalinist “nationality”-classification methods to the pool.

In general, it is probably safe to say that one labels people in order to deal with them in batches.  They become nameless, faceless exemplars of their ethnicities, and can thus be consigned to particular places, roles, castes, statuses, or other assigned seats.  Ethnic labeling in China always went with a certain degree of hierarchic thinking.  The old imperial system referred to minorities as “barbarians,” who could be “raw” (least like Chinese) or “cooked” (closer to, ideally assimilating to, Chinese).  The Communists substituted a Marxist-Leninist-Morganian rhetoric of stages in cultural evolution.  In practice, it seems to be little different from the old classification system.  The Han defined themselves as the highest on the Marxian evolutionary scale, so “raw” morphed into “backward.”

One therefore expects, and finds, that labeling sailors, fishermen, and longshore folk as “Tan” or “boat people” would have much to do with administrative and social convenience.  Indeed, it allowed land people, especially governments, to deal with the boat people as a separate, specific, unified population, one that could be dealt with as a unit.  It allowed the whole floating population to be held down, forced into one uniform low-status category.

This, of course, led to resistance, and thus to constant tension between land and sea.

Relations with the land-dwelling Cantonese—the Punti—continued hostile until the boat people disappeared as an identifiable group.  All boat people old enough to talk to us remembered a time when women and children were not safe on shore.  They would be beaten or even carried off and raped or sold.  (This was rare in Hong Kong, but said to be routine in pre-Communist China.)   In 1965, most women and children still refused, out of fear, to come on shore, except to walk in broad daylight along the most fisher-frequented parts of the immediate waterfront.  Almost none dared a trip into San Hui town.  In 1965-66, land children (of the sort described as mou man—this is, in fact, how I learned the phrase) still yelled taan ka lou (“Tanka guys”) after boat people, and sometimes threw pebbles or small sticks at them.  (The same bad children acted the same to foreigners, in which case the cry was faan kuai lou.  Literally “foreign ghost fellow,” this is the phrase translated as “foreign devil” in the older literature.  The phrase may refer more to ghost-like pallor and wide pale eyes than to demonic behavior.)

By all accounts, the situation by 1965 was already better than it had been a few years previously.  The change from 1965 to 1975 was dramatic.  However, even in 1975 boat women still stayed close to their menfolk or to a group of other women.  In spite of the waning of extreme tension, hostility between land and boat people—and other ethnic groups as well—was obtrusive, and warped behavior along the waterfront.  Otherwise gentle and kind people, of all ethnicities, became tense, irritable, and verbally aggressive when walking the narrow line of the beach.

The land people held the boat people to be coarse (chou), holding up their sailors’ language and hard drinking as proof, as well as the rather free and easy sexuality of the floating world.

A point of more historic significance was that the boat people lacked fixed addresses, lineages, or other marks of proper order.  Not only boat people throughout China, but also other Chinese groups, received prejudice and discrimination if they had no fixed address and no lineage base.  Strolling actors, muleteers, migrant laborers, and even wandering holy ascetics all came in for their share.   All such people had been subjected to discrimination in early Imperial Chinese times, being forbidden to take the imperial examinations, marry proper citizens with fixed addresses, go to government schools, or hold official posts; they were a true caste in those days.  Such discrimination was already waning long before the Qing Dynasty fell, but it survived locally into the mid-20th century.

Finally, the boat people suffered prejudice and discrimination because of poverty and lack of education, as well as a history of wandering and piracy.

The boat people’s corresponding negative stereotype of the land people was that they were mean, money-grubbing, snobbish, and out to take the boat people for anything they had.  Friendship between land and boat people was correspondingly rare, but it did happen, because there was clear and inescapable economic dependence; the land economy depended on the fishery as much as the boat people depended on the land for rice, oil, and incense.  The boat people had to sell their fish through land people, and the land people had to buy those fish and hire boat people to handle it.  The religious life of the community depended on harmony and cooperation between the groups.

Tai O had much better relations between boat and land people than Castle Peak; in fact, it was hard to see any tension at all, though some existed.  Conversely, relations were worse in some eastern New Territories ports we knew.  We visited ports in which we saw boat men (let alone women) attacked verbally and threatened physically.  This very rarely happened at Castle Peak, where mutual dependence, legal protection, and the presence of a very large number of very tough boat people inhibited such behavior.

Relations with Hakka and Hoklou were tense and distant, but those groups had few representatives around the bay.  A number of Teochiu lived in the area, but they spoke Cantonese as well as Teochiu, and the boat people did not see them as a separate group.  The Teochiu (called Chiuchau in Cantonese, Chaozhou in Putonghua) came from the hinterland behind Shantou in northeastern Guangdong province.  They had a reputation like that of the Scots in the British Empire: rough highlanders, fond of emigrating to make their fortune through hard work and extreme thrift.  Their music was high-pitched and played on small instruments, which caused the Cantonese to say that it sounded like tsi kei kau tsi kei—“each one for himself.”  Thus, asking for separate checks in a restaurant—a rare phenomenon in Cantonese society—was called “Teochiu music.”  Needless to say, this stereotype was no more true than others.


The land people differed among themselves more than the boat people did; land society had much greater extremes of wealth and poverty, conservatism and liberalism, education and illiteracy, openness and intolerance.  Many land people showed absolutely no prejudice against boat people or anyone else.  Certainly, few if any were as harsh and greedy as the boat people’s stereotype.  In many ways, the land people of Castle Peak Bay differed only in degree from the boat-dwellers.  Like the boat people, they were an independent, friendly, cheerful, sociable lot, with a society based on mutual aid, and with an amused cynicism about government.  They were, on average, closer to the Neo-Confucian ideal of order, law-abidingness, and patriarchal family relations than the boat people were, but even here one saw considerable overlap, and very little indeed that was close to the extreme conservative ideal.  Most of the land-dwelling Cantonese of the Castle Peak area lived far from lineage villages or other conservative centers.  Correspondingly, Castle Peak was a very open place, on land as on water.

I observed this continually as boat people and Teochiu “passed” with success and aplomb, thinking nothing of the matter.  The ones I knew in the 1960s were perfectly clear about their root ethnicity, but by the 1970s many were not so sure.  Young people growing up did not know whether they were more Teochiu or Cantonese, more boat or Punti—and many, probably most, of them did not care.  They were content to be members—not necessarily very focal or typical members—of two or more ethnic groups.  Younger boat people who had moved on shore in the 1960s, as children, often thought of themselves as ordinary Cantonese.  Older Teochiu who maintained their ethnicity with some militance found that their older children were saying “well, I’m really Teochiu, but I have become pretty much Cantonese,” while their younger children were not even that sure of their roots.  This was disconcerting to the old, but they adjusted.  Similar processes slowly dissolved the Shanghainese and other groups in Hong Kong, and have gone a long way toward dissolving the various mainlander groups in Taiwan[7].

As previously noted, Teochiu reaffirmed their identity by getting together to eat traditional Teochiu food.  The boat people had no such resort.  Their foods were either widely shared or impossible to get without catching them oneself.  In any case, they had no incentive to maintain an ethnic identity once they were on shore working land jobs, so they gradually merged into the Cantonese mainstream.










Adaptation to a floating life has profoundly shaped the maritime subcultures of East Asia.  None has become more distinctive or more marine-adapted than the south Chinese boat-dwellers.  Whether we examine food, clothing, kinship, or social relations, all the differences between boat and land people seem clearly due to immediate ecological factors.  One need not invoke mythical tribal origins.

One could, in fact, predict, or at least explain, all of boat life by considering how traditional Cantonese society—a family-oriented, rural, tough, self-reliant Chinese social order—had to adapt to accommodate long-range traveling and fishing in a highly mobile world where people had to shift, frequently and irregularly, from port to port.  As such, the floating world provides a sort of limiting case of Chinese society.  It reinterpreted Confucianism, and to some extent Taoism, Buddhism, and the broader Chinese folk religious ground from which those grow.  It managed without land or buildings.  It was self-governing, even to the neat rows in which the boats moored every night, without duly constituted authority of any kind.  It created its own songs and peddlers’ cries and children’s rhymes.

The boat subculture was basically defined by residence and occupation—by cultural ecology.  Power relationships—political ecology—took over in making the boat people into underdogs.  Chinese prejudice always cut against anyone of no fixed address:  anyone who drifted, who did not have a proper lineage hall or residence.  The boat people tended to be less educated, more rough, more vocal and violent than “proper” land people.  Thus, they became a despised minority, and suffered accordingly.

They resisted, in all the ways described by James Scott (1985, 1998).  They developed their own world and ways.  They developed an ideology of independence and self-reliance.  This, however, could not be maintained when they had to deal with the world of fish dealers, police, and other land powers.  Thus, their self-image was of “dragons on water, worms on shore.”

Thus, from them, we learn yet another sad lesson in the endless story of human intolerance and inequality.  Yet, we also learn a more hopeful—if far from totally hopeful—lesson of resistance to those most evil of all human evils.

The Cantonese boat subculture provided real ethnic identity.  As in other ethnic cases, no one could put a boundary around the ethnic group in question, or say for sure where it began and ended.  The subculture had its own structure and principles, but these were flexible and were constantly modified by practice.

As we know from the vast literature on ethnicity and identity in China, ethnic groups are politically defined, and are defined in and by opposition to other ethnic groups.  The boat people’s structural opposite was the land Cantonese (“puntei”) cultural group.  The boat people were also situated in a wider grid, where they contrasted not only with their immediate structural opposite, but also with Teochiu, Hoklo, Hakka, British, Portuguese, and other Hong Kong groups recognized in popular speech and popular cultural awareness.

On the other hand, the boat people were not merely the figment of some bureaucrat’s imagination.  Their culturally distinctive ways were real, important, and salient.  Most were, in fact, matters of life and death:  knowledge of fishing and boatcraft, knowledge of seafaring and swimming, knowledge of community dynamics in a community lacking formal leadership.  Other cultural ways were not so obviously vital, but certainly had a major adaptive function; the worship of Tin Hau, for instance, was essential to hold the community together.  To regard the boat people’s cultural distinctiveness as mere political discourse would be to deny the boat people their record of stunning success at adapting to a terribly difficult and demanding world.  Their modern definition and status may indeed have been due to bureaucrats asserting power, but their wider worldview and behavior were not.

The same may be said for all China’s ethnic groups; those who analyze their definition and existence solely in terms of modern politics deny them a history and a record of accomplishment.  Minority people are not merely names or bureaucratic categories; they are living people, and they do have distinctive ways of living.  These groups have a long, complex record of achievement, and it should not be trivialized—whether by governmental reduction to tourist amusement or by anthropological reduction to mere “discourse.”

The boat subculture thus provides a powerful argument for culture as adaptation, and against cultural irrationalism, especially the postmodern form that sees everything as arbitrary “cultural construction” formulated without feedback from a real world-out-there.  People make decisions, and the boat subculture is the result of thousands or millions of choices; but the choices are not made by airy intellectuals in a dream-world.  They are made by people trying to make a living in a highly demanding environment.

The nature of ethnographic generalizations then concerns us.  I attempted in my previous works on the boat people to give a set of rules that would allow an outsider to anticipate appropriately (Frake 1980) what he or she might see—to know the general rules everyone was more or less expected to follow, and to understand, in a minimal way, what the boat people were thinking when they invoked those rules—why they saw the rules as appropriate in general and in the specific situation.

In the present book, I have tried to summarize, from those earlier works, some conclusions of possible interest.  Such generalizations are, broadly, true—they describe the life of most boat people, and the life of the community.  They do not take into account such matters as variation by income, boat type, personality, and, above all, gender.  I have, at least, been able to say something about that last, and about male-female relations in a world very different from that of the stereotypes of  “China.”  Male-female relations were complex among the boat people, but, broadly, more egalitarian than in any other traditional Chinese community I know through experience or description (though the Hakka studied by Fred Blake come close).  No could mistake the boat people for feminists—the patriarchal family was alive and well—but women had a great deal of independent power and agency, and men had to take women’s ideas, views, and independent behavior into account.  Children, too, started with rights equivalent to their duties, and continued to acquire both together as they grew.

The boat people, and for that matter the Cantonese, put all generalizations about China to the test.  Social historians and cultural psychologists need to consider individual and subcultural variation in their own models.

I have thus made several kinds of generalizations.

Some are universally true for all boat people, and more or less distinguish them, such as the lack of lineages.  The boat people are only one of many Chinese groups that lack lineages.  The extensive literature on Chinese kinship has focused largely on lineage dynamics.  Historians have studied the rise of lineages in the medieval period, and the sudden and rather dramatic way that lineages took their final form in the Song Dynasty (see Mote 1999).  Ethnologists have written vast tomes on the lineage.  A brief flurry of excitement arose in the late 1960s and early 1970s, as Chinese subcultures without lineages led scholars to consider why lineages might arise and disappear, fairly rapidly, because of economic and ecological circumstances (Anderson 1972; Cohen 1976; Pasternak 1972).  This solved some major questions, but the wider issue awaits treatment by contemporary theorists, who can draw on advances in the study of cognition and practice.

Others are statements about typical or modal behavior, and in this case I have tried to indicate that many boat people did not conform; for instance, I can say that children were generally very obedient to parents, but there were families where this was most conspicuously not the case.  Their quarrels were impossible to miss on the crowded waterfront.

Other generalizations are statements of what the boat people themselves said—their discourse, in Foucauldian terms.  Such statements might or might not reflect objectively verifiable behavior.  A case in point is the personal independence that has been a theme of this book.  The boat people and land people all agreed that this truly distinguished the boat world from the land world.  The boat people stressed it in discourse.  Visible reality was a more qualified, nuanced situation—not an open-and-shut difference, but a continuum, with broad overlapping.  Certainly, many boat people not only were independent and nonconformist, but openly revelled in acting that way, sometimes with the explicit goal of shocking the land people!  These swallows may not have made a summer.  Many boat people were as sober and deferential as any land Cantonese, and were not particularly independent.   Psychological research, however, did confirm the general claim (Anderson 1992, 1998).

Barbara Ward saw the boat people in terms of cultural models (Ward 1965, 1966).  Again, one may question the generalization involved in defining a “cultural model.”  Each person has a loose network of rules, defaults, more or less organized facts, and, in general, guiding principles for action.  How much these are systematized into a model—one that could be formalized in mathematical statements—may be doubted, and, indeed, Barbara Ward saw models as rather rough-and-ready affairs.

The boat people certainly had their conscious models, as Ward reported.  Observers—ethnographer, land-dweller, government agents—had other models, often very different from each other.  And one could, with Lévi-Strauss, contrast statistical models with structural ones of all sorts.  All such models are generalizations, and, as we know, “all generalizations are false.” In short, cultural models do not exist in external reality.  They exist as conscious models in the heads of the culture-bearers or the ethnographer.

Each individual has a loose collection of more or less structured lore in his or her head, and this lore includes cultural knowledge, more or less systematized.  However, this “model” may be rather different from the model held by a next-door neighbor, and still more different from the model held by a spouse (of different gender) or teenage child.  The external observer’s model is yet more distinct.

One can make statements that are well supported and well documented, and that are beyond reasonable debate.  These never make up a cultural model, however; we have to resort to tentative formulations.  I have tried to keep these to a minimum in this book, with the result that I can stand confidently behind each statement.  It also means I cannot sell my work as an example of insightful, exciting, or brilliant interpretation or insight.  However, in documenting a way of life that has vanished, one is well advised to stick to clearly accurate data.  No one can check claims (except with the few other published sources that are actually based on ethnographic study).

Discourse—what people say to each other—may or may not reflect real basic values.  Ethnographic questioning and observation can go a long way to finding the differences, and understanding what values really animate behavior, but can never completely resolve the differences.  Conflicting values—Confucian morality, immediate wealth maximization, divine protection, community welfare, and many more concerns—are variably and contextually expressed in behavior.  In the end, the ethnographer can only record what people say and how well that predicts what they do.

Boat society fits perfectly into Anthony Giddens’ view of structuration (Giddens 1984).  Individuals make choices.  They make these with constant attention to their environment, both natural and human.  Class, caste, gender relations, historical contingency, and other social facts influence—and often determine—their choices.  Society changes according to how those choices are made.  Structure emerges, but it is a dynamic system.  It may stay very fluid. If it crystallizes into a frozen array, it may recrystallize in a very different pattern in a few years.

One can describe a social system, or a subculture, at one point in time:  Castle Peak Bay from 1965 to 1975, for example.  But this short passage is not the story.  The story is a longer one, a long tale that goes back at least 3500 years, to the sites on Cheung Jau and other islands.  One part of the story ended in the late 1980s and early 1990s, when Castle Peak Bay was filled, ending forever its floating world.  But boat people still endure, along more remote coasts and rivers of China.


The ideology expressed by almost all boat people at Castle Peak was one of independence, but also of mutual aid, mutual reliance, and mutual support, growing from a dense and continual webwork of interactions, but all done without compulsion.  Kinship and spatial relations structured most of these interactions, but one attempted to maintain the widest network possible, especially if one had aspirations to leadership.  One needed to interact with land people of all sorts as well as with other fishers.  However, everyone strongly opposed the fell hand of authority.  People should do all this on their own.

We watched this ideology change in the face of reality.  The cooperatives were beginning to achieve acceptable status in 1965.  By 1975, cooperatives, schools, the government fish market, public health clinics, and other facilities had transformed waterfront views of government.  No longer was government an unqualified evil.  In fact, people admitted that they needed it.  The ideology of independence and self-reliance persisted, but in qualified form.

We might have foreseen, in this, the end of the boat world.  Perhaps it could not survive once it admitted that land-based organizations could be powerful forces for good.  In any case, almost immediately after the boat people came to depend on government economic and social arrangements, they moved on shore, blended into the Cantonese mainstream, and disappeared as an identifiable group.  Fishermen persist as a separate group at Tai O and elsewhere, but may assimilate in the near future.

A theory of society lay behind the older view.  Obviously, to summarize it and generalize it is to do violence to it, because it was based so much on individualism and individual interpretation.  Yet, discourse on the waterfront had common themes and common agreements.  Everyone took as true a few well-known claims, and followed similar paths in deducing the rest of the theory from these.

First, humans were held to be preeminently social.  The extreme form of American individualism—the self-made man alone against the world—would not have been widely accepted; the boat people knew such individuals, but did not like or respect them greatly, seeing them as too antisocial.  People loved to be around each other.  Life’s greatest joy lay in relaxing, drinking, feasting, celebrating, or just talking over tea, with one’s friends and fellows.  The love of humans for ordinary casual socializing seemed self-evident.  No one would think of denying it.  The loneliness, isolation, and monotony of life on a small boat was enough to drive some people mad—literally—and few indeed were the people who enjoyed that.  When—as often happened—mutual aid began to shade off into exploitation and using, moral alarm bells rang, and waterfront gossip dealt with the issue as best it could.

Second, people depended on each other, quite apart from joy in company.  On the water, every man’s or woman’s life depended on instant, caring, thoughtful, responsible action by others.  It is really difficult, at this remove, to provide the reader a sense of just how fast a tranquil fishing day could turn into a desperately dangerous situation.  Small boats, rather round-bottomed to increase maneuverability, were at immediate and total risk if a storm came up suddenly.  Storms on the South China Sea were sudden, unpredictable, and violent.

Thus, reliability and responsibility were the keys to all virtues, the basis of ongoing human interaction.  One could, and did, celebrate at New Year with anyone and everyone, but one’s everyday life had to be spent with people one could implicitly and absolutely trust.  This fact alone made a command-based society impossible to envision. When a typhoon hit, or when pirates attacked, no one could wait around for government mandates.  No one could rely on people without a long-standing claim of friendship and relationship.  Even the slower emergency created by a run of bad luck—several days of bad weather, a broken engine, a mast swept away—required prompt, responsible, cooperative action.  One had to rely on family and friends for loans and support.

Third, given that, people had to have tremendous autonomy and agency.  They had to be free agents.  A leader, especially, had to achieve a reputation for being a bit unpedictable, even a bit crazy.  If he were too predictable, he would seem like a dull stick.  Worse, others would soon figure out how to take advantage of him.  If they never quite knew what he would do, they never quite dared to risk a bad encounter.

On the other hand, independence had its very strict limits.  For one thing, the boat world enforced a type of surface conformity by ecological necessity.  Everyone had to dress pretty much alike, because there was, in traditional times, only one best way to dress; loose black heavy-cotton clothes and straw eye-shading hat were so superior to anything else that no one really had much option.  Everyone had to eat about the same, run a boat in similar ways, and talk the same dialect.  Thus, to stand out, one had to be self-conscious about cultivating idiosyncracy.

For another, independence could not extend to life within the boat household.  There, the senior man had to be able to command, and everyone had to know his or her station.  On the other hand, the extreme value on individual agency made it unthinkable for a senior man to command in an arbitrary fashion, without consultation.  Still less would he be able to make people do what was not in their station.  Tradition governed the roles of each person.  Seniority and gender defined these, and led to a smooth order that worked well and could not reasonably (let alone arbitrarily) be challenged.

In short, independence and firm order both derived from a deeper principle:  Life depended on reliability, responsibility, and mutual respect.  It also depended on a certain amount of grassroots solidarity, since people had to unite to deal with emergencies and to oppose the land people.  (Tension with the land people was a fading memory by 1974, but not entirely gone.)

Thus, as we have seen, the boat people adopted a form of Chinese traditional culture that fit their needs.  In doing so, they created a thoughtful alternative to many Confucian and other ideas.  They followed Confucianism in seeing basic human “goodness”—eusociality—and in a political philosophy of personal integrity and of rebellion against tyranny.  They followed Daoism in their awareness of the inevitability of change, and their flexible, sometimes fatalistic, but always creative way of dealing with it.  They followed Buddhism in seeing a need to build merit with beings beyond human and visible realms.  They failed to control overfishing, defeat poverty, or to build stable communities; they succeeded in creating a flexible, consciously planned social order that enabled them to survive.

The extreme individualist morality that follows from rational choice theory and its relatives would thus not have been acceptable.  It would not have worked.  People sometimes asked us how American society could hold together with its constant shootings, divorces, and sexual peccadilloes.  The view of American society here implied was derived entirely from Hollywood movies, so it bore little resemblance to reality; the point is that genuinely anarchic personal relationships seemed completely beyond the pale.

And, in fact, I cannot see how theories of “rational self-interest” can stand up under the test of explaining life on the boats.  Individual competition of the sort postulated as panhuman by the more extreme sociobiologists (e.g. Ridley 1996; Dawkins 1976) would have made the boat world quite impossible.  Evolutionary ecologists who study fishing have learned to accommodate cooperation in their models; Michael Alvard and David Nolin, for instance, provided a superb account of just how far rational choice and evolutionary psychology can take one in explaining cooperation in a fishery (Alvard and Nolin 2002; Aswani 1999.)  Alvard’s observations provide what seems to be a perfect simple model of how a complex network of undirected cooperation could become established in a fishery.  Aswani shows that highly competitive people can still figure out how to control and regulate their fishery.

Ironically, the boat people of south China never had the opportunity to regulate their fishery; they had to make do with creating a complex, intricate adjustment to the reality of multi-stock fishing without hope of rational management.  Other Chinese fishermen (e.g. in Malaya; Anderson and Anderson 1978) were luckier, and could invoke sea tenure, privatization of beds or lagoons, and regulation of offtake.  Once again, adaptation to local political and ecological circumstances was the rule.  Instead of broad generalizations about behavior, we can generalize here only about the vectors and motives that lie behind, and modify, behavior.

Conversely, the boat people’s social theory, like their social reality, was far from the extreme communitarianism of Leslie White (1949), Talcott Parsons, or others who see Culture or Society as a glorious, perfect, beautifully structured machine that reduces individuals to mere moving parts.  The boat people rejected—vocally—the more extreme forms of Neo-Confucianism.  Many shore-dwellers strongly advocated the family in which a wife was utterly subservient to her husband, a son to his father, and a younger brother to an older.  As common in the area, and as strongly advocated, were the communitarian ideologies of Maoism, Chinese nationalism, and the like.  The vast majority of boat people had no use for any of these moralities.  They were willing to tolerate such ideas, so long as no one tried to impose them on boat households, but they were not converted.  Even the few Communist boat people were a notably independent group, and we were not the only ones who wondered how they would fare under an actual Communist government[8].  Suffice it to say that all of them had fled from China.  Perhaps those left behind were more communitarian.


Today, Chinese traditional society, even the relatively loose manifestation of it seen in the boat world, seems oppressively patriarchal to many people in the western world and in modern China itself.  Indeed, it often was oppressive even by the standards of its own practitioners.  People told stories of model households in which everyone worked together in harmony without friction or oppression.  The reality was almost never so smooth.  And things had been worse in the past, when female infanticide or sale occurred (only under desperate circumstances, however, according to all our sources).  Hong Kong today has rejected the more stern variants of the Confucian order; the boat people blend in partly because society has come to meet them, and even passed them, in its move to a more liberal order.

The boat world was also subject to evils that followed from autarky:  overfishing, piracy, lack of services, lack of any certainty to life.  The boat people realized that these were the costs of autonomy, but fatalistically bore them, regarding the benefits of autonomy as more than offsetting.  Yet, given the conditions of China in the early and mid-20th century, torn by famine, war, and government rampages such as the Great Cultural Revolution, the costs seemed high indeed.

In short, the boat life was not without enormous problems and costs.  Romanticising it is a temptation, at this late date, but one that must be resisted.

Yet, the possibility of learning from the floating world, and benefiting from its achievements in everything from maritime technology to song, is certainly clear.  The world has benefited greatly from fore-and-aft rigging, watertight compartments in the holds, and the compass—all inventions of Chinese sailors.  We can now learn from the boat view of society and from its useful practical corollaries in such areas as child-rearing and dealing with close crowding[9].

They followed their own way in creating a social order, a folk song tradition, an aesthetically exquisite world of polished wood and carefully arranged tackle, a cuisine that used common ingredients to produce the best food I have ever tasted.  It was a beautiful, valid, exciting, rich, full way of life.

Without that richness, they might not have survived.  The mysteries of why we need beauty and song are locked in unexplored regions of the brain.  Perhaps, when these are unlocked, we will learn that the salt water songs and the careful arrangement of dark wood and earthenware tones on the foredeck were as necessary to life as the mutual aid and respect that they facilitated.

Through it all, the boat people of south China remained their own persons—brave in the face of constant danger, tough in the face of incredible hardships, friendly and open in the face of endless war and oppression.  A bluff, cheerful man would tell me soberly of watching people starve and of fleeing under fire from the Japanese.  A poised, gentle, centered woman would break into hysterical sobs as she talked of watching her children starve to death one by one until she finally gave up and let someone abandon the latest baby on a deserted island.  Somehow, people survived, and created from centuries of terror and suffering a valid, beautiful, indomitable lifeway.

All that is gone now, and a new and less adaptable world must face the unimaginable horrors of the future.  Population growth coupled with resource waste has destroyed the fishery.  These same forces now doom the vast majority of humanity to the conditions the boat people faced in the mid-20th century.

Perhaps the example the boat people set will not be totally lost.  Perhaps their creativity and thoughtfulness in the face of adversity will give some meaning and some inspiration in the hard years.  In the end, the human spirit endures.











I recorded several traditional songs (Anderson 1972).  Especially important were the “salt water songs,” the traditional folksongs unique to the boat people.  These contrasted with several other song types:  classical poems, funeral laments, children’s songs, popular songs, and so on.  I published the Chinese texts and literal translations of my collection, with detailed commentaries, in Chinoperl News (5:8-114, 1975; see also Anderson 1972 for further detail, including some translations of songs recorded by Chinese investigators in the 1930s).  I have retranslated a few of the songs—less literally, more idiomatically than in 1975—for this book.  Many songs turn on tedious dissection of characters, or were badly sung, or were formless dialogues, and have been left out of this selection.

Most lines ended with a long-drawn-out aaaa, not indicated here.

The words give little indication of the beauty of these songs as sung on a quiet night on the still bay, with huge stars overhead and almost no other light.  Fung Sam-Jau in particular had a beautiful singing voice.  The experiences were unforgettable.

Most of the songs were recorded from Fung Sam-Tsau and Cheung Ngau-Tsai.  These were highly intelligent, paradoxical men.  Both were proud, independent, and difficult—standard boat characteristics, but ones they they took to extremes.  Both loved to display their knowledge of the old songs, but also demanded more and more money for less and less singing.  Both were unreliable about times and amounts of work.  Many boat people had a very flexible sense of obligation to set appointments, since they often found it necessary to race to another port because of a fish run or a family emergency; they might not return for days.  But Fung and Cheung were exceptionally hard to pin down.  In the end, they loved their songs, and we heard their repertoire.

We heard many women singing, and women were traditionally the main bearers of the tradition, but, tragically, women refused to record.

No one else seems ever to have recorded salt water songs, and my few tapes, miserably recorded on a cheap reel-to-reel tape recorder, are now all we have of a vanished tradition.  My original tapes are deposited in the Hong Kong Museum.  I understand that boat folk songs still exist in remote parts of south China, but, if other Chinese folk material gives any guide, the songs have probably been vastly changed by Communist policies, radio and TV singing, and other external influences.

One stray line, from an otherwise hopelessly untranslatable song, is too good to miss:  “A man of roads and villages can’t claim the sand and dust.”  A wanderer can’t be too proud; he owns not even the dust he treads.  This beautiful saying, which harks back to classical poetry, may stand as a final word on the boat people’s world and its poetry.


The Vegetable Song

Sung by Fung Sam-Tsau and Ng Kam


The hollow-hearted water spinach grows on the water.

The falling eggplant splits like a mouth breaking into a smile.

The water caltrop is solid as a fool’s block head.

Watercress and spinach cool the heart.

An onion has a head without a tail;

With a head but no tail, how can a man be called a hero?

Mustard and collards are both greens;

Avoid wrong, don’t gamble, don’t fight.

Celery and celery shoots:  Work hard and win,

Work diligently and then rely on your sons.

Ivy-like bitter melon doesn’t climb beanpoles.

After it’s picked, the vine remains.

Get pork when fresh; get a girl when she’s young.

If you want her, take her this year.

Like onions chopped on a block, you won’t last long.

Water to voyage on, none to drink;

But in the cabin, the lovely girl is a barrel of fresh water.

The foreign month is four weeks;

People without money are never mentioned.

Take a water chestnut, skin it, leave the stem,

Cantonese people have no human feelings.


The hollow stems of the water spinach symbolize an open mind here, in contrast to the stuffed-up mind of the blockhead.  Watercress and spinach are cooling in humoral medicine, and remind us to keep cool in trouble.  “A dragon’s head, a snake’s tail” is another variant of the next couplet; it means that one starts bravely on an enterprise, but peters out.  The next couplets turn on puns:  kaai “greens” and kaai “avoid,” kan “celery” and kan “hard.”  (The tones do not match, but tones are not sung in salt water songs.)  The bitter melon is self-reliant.  The foreign month and water chestnut seem to be used only to make up a rhyming line.  “Cantonese” here means “people in general,” though it is also a bit of a slap at the land people.



Fragment of a Courting Song

Sung by Cheung Fuk-Luk


Because I have no money I’ll float to another place;

The poor always have to drift and wander.

Because I have no money, I have to struggle all night for a living,

Yes, through the night, because I am poor.

Corn has layers of silk, you can’t see its face.

A Buddha-image has a mouth but speaks no words.

It’s an old saying:  Without money, you suffer wind and sun.

I am too poor for us to unite.

My hat is full of holes; I put it on.

My clothes are dirty and ragged, but I make do.

Take a cross, add a hook, join a left stroke,

If I had many helpers, I’d get wealth and property.

Yesterday I made a friend and we talked together;

Even a thousand gold won’t buy a real friend on the boat.


The cross, hook and left stroke write the character for “wealth.”  This song would be sung by a man courting a woman, wishing he could have her as a “real friend” on his boat.  She would sing an answer, either rejecting him or saying money is mere dung and she loves him too much to give him up.



Salt Water Song

Sung by Fung Sam-Tsau


Once more I ask the girl to go with me;

When she’s ready, she rows the boat here.

A new abacus and you pinch the wood;

You’re a sweet girl but you’re counting wrong.

“Up the bay to buy a carp, down the bay to steam it,

Steam it with bean paste and bean sauce, give me money too, I still won’t love you.”

The dried oyster opens as wide as Tai O channel.

If you want to “fire the gun,” row your boat here.

Step up and cut flowers, the boy over here,

But if he’s broke, the girl stays on her own side.

I start out to fish, see great waves,

The waves strike and sink the boat.

People have to grab the boat rail.

I buy a bowl of chicken soup for the girl,

As long as I live I will not forget.

The north star faces south.

The moon darkens, dark clouds cover silence.

The Sky God says “Open”!  Through the dark clouds we can see the sky.

Row with the sweet, pull the hand lines!  No good working too hard;

The dragon boats are tied, but we will soon know who loses and who wins.


An “abacus-pincher” is a penny-pincher; the girl—a professional entertainer in this song, not a prospective wife—cares only for his money.  The quoted lines are supposedly sung by the girl; she would reject even his best cooking attempts.  The dried oyster is her vagina, not at all flatteringly described; “firing the gun” and “cutting the flowers” are phrases for sexual intercourse, the violence of which makes the great waves.  The opening sky refers to hope that the courtship will lead to something.  Dragon boats race on the fifth day of the fifth lunar month; till then they remain tied up.  The implication here is that the girl has many competing suitors, and we will soon know who wins her.



Salt Water Song

Sung by Tam A-Kau and Fung Sam-Tsau


The girl has sea snails in her boat.

Her lover goes there, attracted.

He likes the sound of her voice;

All through the night he doesn’t interrupt her.

Buy salt-hay to tie big crabs;

If they escape, the hay’s wasted.

Little sister, go to the temple to cast lots.

The lot tells you to go with me.

It says you are still strong.

A new bamboo pole, seventeen or eighteen feet long:

First test the water, then see the girl’s heart.

I buy good pork but can’t afford two pieces,

And my little Castle Peak boat has no engine.

Miss Kam Tai goes out to fish, gets seasick and throws up—

Drop her off at Castle Peak with me to look after the house.


The lot in question is a stick that stands up, phallically, when several sticks are shaken together; the symbolism is obvious.  The pole for pushing the boat along is also a phallic symbol, so obvious that boat songs have used it for centuries.  Classical Chinese poets immortalized several examples.



Fish Name Song

Sung by Fung Sam-Tsau and Cheung Fuk-Luk


The flounder has eyes on only one side.

The mud flathead’s body wriggles.

The goldenthread has a red body and white belly.

The long-tube fish swims backwards or upside down.

The hairtail comes up all rubbed with white powder.

The yellow croaker’s head wears silver and gold.

The anchovy becomes extremely large.

The sawfish becomes king.

The white-rice fish and silver fish, cut open, have no blood.

The stonefish, salted, gets its scales and head cut off.

The barracuda leaps out like an arrow.

The flying fish is a flying boat on the sea.

The purse-seiner catches minnows near shore.

Shad and ginkgo-fish swim in open sea.

Sea mullets, gray mullets, stay near shore,

Little croakers come up from stony duck-mud.


Just fun—few double meanings here.  The fish are well described.  The sawfish, a sacred fish, is the most powerful of all sea creatures.  The huge anchovy is probably a joke—an ironic “fish story.”


Ten Women Floating on the River

Sung by Fung Sam-Tsau


The first old girl floats up to Hang Mei,

The water dries up slowly, and she asks a young man to push.

The second good woman floats to the fast current,

The fast current breaks her water jar.

On the water she uses her hand to hold it,

On the bank a cook uses his bamboo pole to carry it.

The third woman floats to San Kei,

The water dries out where the bottom is stony.

The fourth floats to San Sa,

Where some go to sea and others sstay home.

The sixth old girl floats to the First Channel,

Where some houses are on stone piles and some on wood stakes.

Of course behind the shops is a stream.

The seventh floats to Tai O pier,

Where once Miss Twenty would really give you something!

She’d ferry you at any hour, no matter how late.

The eighth floats to Tai O mooring,

Where a ferry poles to and fro.


The places are in or near Tai O, famous for its singing girls.  The obscene double-entendres should be obvious enough.  The drying water is the lubrication on the woman’s vagina, and so the song begins.  The “fast current” that “breaks the water jar” means just what Freud would have said it meant.  And so on.  Unfortunately, Fung could not remember all the song.

The word used for the women literally means “elder brother’s wife,” a somewhat respectful but very colloquial term for older women.



The Seasons

Sung by Cheung Ngau-Tsai


In spring, soon, people get married.

Spring color, flower scent, spread far off.

Spring wandering on fragrant grass is delightful.

“Spring Girl” teaches her son clearly and rightly.

The spring swallow in the roofbeams builds its nest.

In a spring hut, beadwork grows like jade fruit.

In the Spring and Autumn Period, countries fought with spears.

Spring love achieves happiness, but you pay for it later.

In the spring curtains and decorations the bridegroom meets the lovely lady.


Summer potted-plant leaves droop and doze,

On the summer mountain hay dries in heaps.

The Xia Dynasty’s Emperor Yu made channels and stopped floods.

Summer is very hot, people take their clothes off.

Sume days are long.  In the house

Summer makes one go as slowly as possible to the kitchen!

Summer lotuses grow, all flowers open.


Fall tangerines fill their branches.

Fall night rain is just the right time.

Fall colors come once a year.

The fall wind blows, stirring up clothing.

Fall exams come; people tell stories

Of “Fall Fragrance” and other famous things.

Fall moon rays shine, rain can still fall.

Fall sorrow: a stranger on the road—but he might still meet a friend.

In fall, girls smile as they walk, thinking secret thoughts.    The fall chrysanthemum grows with no thorns.


In winter, potted chrysanthemums look to the clear moon.

Winter makes all fresh and bright.

The East Mountain is red, its top purple and firm.

Winter cold freezes hardest in Beijing.

The eastern household lays out its lineage graves, small tower, and wall.

From winter solstice it’s 106 days to the spring festival.

Winter fruit, at the New Year, stands piled high.

In winter sun the apricots flower proud in snow and ice.

Winter-mushrooms are soft and smooth, everyone likes them;

Here’s joyful wine hidden till winter—Show what you can do with it!
Every line here starts with the season’s name, or a homonym (Xia for summer, east for winter).  The plants and animals are auspicious and thoroughly conventional.  This song is notably more refined than the preceding ones; Cheung Ngau-Tsai knew some classical poetry and preferred fairly literate diction.  But sexual references are far from absent.  Cantonese maintain that fall rain is an ideal time for lovemaking.  The colors (“color” can mean “sex”) and wind-lifted clothing drive the image.



Teasing the Bride

Sung by Fung Sam-Tsau


He new bride, veiled, makes bridal tea very hot,

Her left hand holds the cup, her right hand pours.

The first cup for her father-in-law to drink,

The second for her mother-in-law.

The third is for her husband’s elder brother,

The fourth for his elder brother’s wife.

The fifth and sixth are for the husband’s younger brother and his wife.

The seventh for friends and relatives.

The eighth is for her husband—

On the bed very good, mandarin ducks rose one.

The new bride makes tea with Tai O river water

and offers cakes for her father-in-law.

When mother-in-law instructs you, don’t answer back.

As for your husband, if you don’t care about him, whom will you care about?


The italicized words were sung in English.  This song is a straight instruction to the girl on how to serve—until the last lines.  Mandarin ducks are a symbol of wedded bliss.



The Bridal Feast

Sung by Fung Sam-Tsau

A pair of chopsticks iin the middle;

Next year when you have a son, invite me to eat ginger.

Pure sweet crisp meat on the bones of a fat chicken.

If you know what I mean we’ll all eat together.

The girl is pretty and good, a worthy match for the gentleman,

A properly raised girl with all the virtues.

Four sweet oranges grow on one branch,

The wife picks them and the husband peels them.

He gives them to her,

They eat and drink without separation.

If they mention separation, their parents are sad at heart.

The kite has sons, but they desert him;

When full-grown they fly off to the sky.


A feast with ginger celebrates a son’s birth; ginger is stimulating and healthy for a new mother.  The kite (the predatory bird is meant here, not the paper imitation thereof) is, of course, a contrast to the faithful pair.  Kites were an abundant pest in the harbors of Hong Kong.



Wine and Vegetable Song (another wedding feast song)

Sung by Fung Sam-Tsau


The mushrooms are delicious, sweet in the bowl.

The good-hearted Jen Gui went and conquered the east.

Di Qing had only three arrows left;

With them he shot three heroes. Fried dry shrimp look like gold tidbits.

Take up the bowls and stir all together.

This truly is the work of a master chef;

When people eat this, they cannot worry.

The whole duck is delicious in the bowl.

The wholly pure phoenix marries the gold dragon.

They will have sons, educate them, make them valuable men.

The whole household eats; all will be heroes.

In fall the tangerines are green and tender,

In fall comes the night rain; sons are made easily.

In fall fine scenery appears:  range after range of mountains.

In fall come the tangerines, and more babies will be born.

In summer the little tangerine grows buds;

In summer comes the dragon boat festival, and flowers open.

In summer days the wind comes, pearls fall in showers.

Next year there will be tangerines, and also clear tea.


A description of the feast segues into a short version of the following song.  The tangerines and tea in the last line are thank offerings for a son’s birth.  The sexist preference for “sons” is all too clear in this song, and all too typical.  Other wedding songs, not translated here, express the Cantonese ideal of “five sons, two daughters.”  (At least, a few daughters are desirable.)  Indeed, completed families in our sample did have seven children.



The Tangerine Song

A praise poem for a girl.  Sung by Cheung Ngau-Tsai.  (Mistranslated “orange” in Anderson 1967 and 1975, foolishly following Hawkes’ 1959 mistranslation; at least Anderson 1975:47 supplied a corrective note, thanks to editorial help.)


In spring the little tangerine tree’s roots grow below ground.

Spring comes with moderate rain.

In spring warmth, last fall’s branches bloom like brocade.

The spring tangerine flowers out like a phoenix lute.


In summer the little tangerine grows shoots. On Dragon Boat Day it puts out flowers.

The summer moon shines, the wind comes, branches and leaves sway—

Next year the little tangerine will need supporting stakes.


In fall the little tangerine has green fruit.

In fall in the night rain the seed should put forth shoots.

In fall after wind, the moon shines lovely on the hills.

When fall comes again, the tangerine will again grow larger.


In winter the little tangerine’s fruit is perfectly red.

At winter’s end, rain comes with festivals.

Winter months are cold; the flowering-apricot needs special care;

In winter the tangerines fill the whole orchard with red.


The tangerine is so auspicious that its colloquial name is “lucky fruit” in Cantonese.  Its red color is the perfect color of life, love, good fortune, and all wonderful things.  The tree is considered so exquisitely beautiful that people try to grow them in pots in the house.  This is why “orange” is a sorry mistake; the orange does not have these strong associations, though oranges too are auspicious.  Here, the tree is a metaphor for a young girl, marrying, making passionate love, and getting pregnant; see the “fall night rain” above.  The tree needs supporting stakes because of its heavy fruit—a lovely image for a pregnant woman.  “Flowering apricot” is a homonym for “little sister,” a pet or colloquial term for any young woman.  The flowering apricot flowers in midwinter, and needs protection from extreme cold.  Here, the special care implies care for a woman giving birth.

A famous poem in the Songs of the South, from around the time of the Han Dynasty, praises a young man using the tangerine tree as an image (Hawkes 1959:76-77).  The present poem is strikingly close to that one in imagery and rhetoric.  Since the boat people frequently learned classical poetry, via educated patrons of boat singing-girls, this is not as surprising as it might seem.  Cheung Ngau-Tsai, in fact, knew and could chant poems by Li Bai and other classic poets.







These life histories (except for #3, which I collected, in English) were collected by Choi Kwok-tai, with or without my assistance.  Mr. Choi worked with the fishermen and was a trusted friend to most of them.  Mr. Choi, Chow Hung-fai and I translated them.  Mr. Choi put them in third person, because he shortened them somewhat in recording, but they were almost word-for-word.  We all edited them slightly for the initial, almost verbatim translations in Anderson 1970b.  I have now re-edited considerably, to shorten them and disguise identities, but I have tried to keep the narrators’ turns of phrase, such as “fac[ing] the challenge of sexual desire.”  All names are pseudonyms.  All the participants must now be long deceased, or, if not, impossible to trace.


  1. The Nouveau Riche


Yat was a native of Po On [a county just north of Hong Kong].  His family has been fishing at least four generations.  His father ran a gill netter, and he lived happily under his parents’ care.  He gradually came to work more and more with his father.  His father employed a private tutor, teaching the four children on the boat, when Yat was 14 years old.  He had only a few months of study, because his father decided to retire early, being sick of fishing.  Yat took over the boat when he was 15.   His parents arranged his marriage the following year.  He then built a shrimp trawler, with some money borrowed from a fish dealer.  The family fished at Po On, Tung Kun, Macau, the Pearl River, and elsewhere, and especially they came to fish the yellow croaker runs at Tai O.

At that time Tai O had a great yellow croaker fishery.   When the fishermen came ashore, they could enjoy themselves.  The restuarants were crowded, and plenty of singing girls were there.  Yat was strong and well off, so he faced the challenge of sexual desire.  It was difficult for him to resist.  Fishermen in port would form groups, wandering among singing establishments, enjoying whatever happened.  His wife became worried, and got a concubine for him, but he still visits his old friends in Tai O when there on business.  Fortunately, he could keep his business up and work hard.

Hong Kong was occupied by the Japanese in 1941.  Seeing that war had broken out, thieves and robbers appeared like forests [as thick as trees in a forest].  It was hard to survive.  Yat, with his family, went to Tung Kun and hid in a small channel.  In two months he returned, and fished successfully in spite of pirates, which he discouraged by means of guns and cannons.

After the Japanese surrendered, the British government organized the Fish Marketing Organization.  Yat supported this and helped develop a unit at Tai O, but the unit went bankrupt.  In 1958, he mechanized his boat.  He then handed over his fishing business to his second brother and came to Castle Peak to buy a wooden hut and set up business on shore.  He sold rice, groceries, and salt fish, but went bankrupt soon because of lack of capital and experience.  But he had made contacts, so switched to selling diesel oil.  He started selling only a few drums of oil; lacking capital, he did badly, but he built up trust.  He could borrow some money.  More and more boats were mechanizing.  In the 1960s and 1970s, he and his family sold diesel, water, kerosene, and related products.  He became rather well off, and eventually moved to more urban parts of Hong Kong to retire in modest comfort.


  1. The Child Buyer


Yi was born in Tai O to a Po On family.  Of his three sons and one daughter, two sons were bought.  In traditional times, if a couple had no son after three to five years, people would ridicule them.  There was a saying:  Thirty no wife, forty no son—alone half your life!  So when Yi’s wife had no son after three years of marriage, he bought two sons.  The price of a baby boy at that time [mid-1940s] was very high.  Sons came from poor families who had more children than they could support, or from gamblers or opium addicts.  Intermediaries obtained a commission for selling these sons.  Another source was women whose husbands were away for long periods, leaving their wives, who were sometimes tempted.  They would escape to distant areas, and sell their illegitimate children.

Yi’s father ran a large sailing trawler.  The family fished around the Pearl River, catching croakers, shrimp, and the like.  Living was not too good, but he enjoyed his childhood.  He went to school for five years, from 11 to 16.

One cold night when he was a young man, the Japanese attacked.  Bombs and shots were everywhere, making the whole fishing port a battle scene.  The fishermen were living peacefully and did not know what was going on, since they did not read newspapers from one year to the next and knew nothing of the outside world.  They thought the war was a punishment from the Sky.  They had to flee for their lives, many being separated from their families.  Luckily, Yi was not asleep. He raised anchor and fled under the shots.  He passed a night on a deserted island, then fled to Tai O.  Japan occupied south China; the ports and markets were closed to him.  Killing, shooting, and robbing went on and on.  The fishermen could survive in Hong Kong Colony, fishing south of the main islands.  But soon the Japanese took Hong Kong.  The family moved to Castle Peak.  His father got him to marry, in spite of the war, because he was of age.  The boat was too large to opearte safely, so they bought a gill-netter and fished pomfret, goldenthread, and croaker.  Fish were common, prices low.  Once the Japanese police caught him, assuming he was a secret agent.  Only the lucky appearance of an interpreter allowed him to explain himself.

After the Japanese surrender, Yi’s mother got cholera.  At that time the fishermen did not understand this disease.  Cholera was thought to be a stomach ailment, and was treated by Chinese doctors.  With the help of the Sky, his mother survived.  She was so weak that other sicknesses came, and she lay in bed for ten months.  He had to stop fishing and pay medical fees.  His living was very bad. He went into debt and could not mechanize.  Eventually he got a mechanized boat on a loan, but repairs ruined him.  He stopped fishing and worked for a live fish dealer.  His wife and daughters-in-law made nets.  Eventually, with hard work, savings, and a network of loans from extended fammily members, they saved enough to get a bouat and get back into fishing.


  1. The Educated Man


Saam was from Cheung Chau; the family believes it came from Toisan five or six generations ago.  His mother had some education and insisted on living on land to send Sam to school.  His mother died in 1942.  The family did not do well, and turned to the land about 1950.  His father stayed at Aberdeen, while the children were raised at Cheung Chau.  From 1952 to 1955, when his grandfather died, Saam studied in Hong Kong.  Scholarship, encouragement, and a lack of opportunities for the uneducated kept him in.  He began to read enormously, for “escapism, amusement, and necessity.”  In 1955 he became the support of his family, thanks to a small scholarship.  He got tuberculosis but recovered, went through the university, and earned a BA and diploma in Education in 1961.  He then went to teach in Yun Long.  In 1964 he was able to go to England and study for a year at the University of London.  Returning, he got a post training teachers in Kowloon, which he was still doing in 1966 [I did not find him in 1974].

Saam thought that he and a lawyer of whom he knew were the only boat people to graduate from university.  He thought Chinese “overprotect” children and do not expose them to emotional conflicts, but that the carrying belt gives the children emotional satisfaction and thus helps prevent juvenile delinquency.  England convinced him that shouting at and frightening children can harm them permanently.


  1. The Fish Seller


Sei was local born.  He was the eldest of eight.  His grandfather fished and also carried fish for Macau and Hong Kong dealers.  He took to running a small ferry.  When Sei’s father was grown, small motor-powered boats took over the industry.  His father could not carry on the ferry business, and returned to fishing.  At that time [1910s-1920s], boats were few, so fishing was not bad.  At Castle Peak Pier, fish dealers and shops were scattered around the area.  Sei’s father was cheated and oppressed by his dealer, so switched to another, and eventually wnet into dealing for himself.  He and his brothers and cousins knew only the first ten numbers, so they had to draw pictures of the fish when they wanted to list fish.  Eventually, at 12, he got to school.  Schools taught by rote and were oppressive, so he tried to escape, but his father needed business help and made him stay.  Eventually the fish dealership began to pay, and the family moved on shore.

The Japanese then occupied the colony.  Pirates and rascals took the opportunity to rob and raid.  Lives and properties were in danger.  Their dealership failed.  The Japanese destroyed the pirates, but “spies” for the Japanese could raid and murder.  The family went to isolated islands to fish for a living from  small hand-liners.  They barely survived.  They lived on rice gruel and sweet-potato meal [the most despised of foods].

Finally, the Japanese surrendered; everyone rejoiced.  Sei’s father became a fish dealer on the far side of the bay.  Sei’s younger brother—the second son in the family—became a seaman; his third worked in a shop in Yun Long.  Sei and his unmarried elder sister carried on the father’s old business.  The seaman came home and joined in.  They succeeded, and threw a celebration that “set a record” for local display.

In 1947, the smuggling business was very active.  It increased steadily.  Sei started a live fish business for the smugglers.  He got capital, and began to loan it out.  He used crooked scales.  His income was “like a waterfall.”  He became creditor on more than a hundred boats.  They celebrated many weddings and feasts.

[He took to gambling and other shady pursuits, got a disease—tuberculosis or sexually transmitted disease—and came to a sad end.]


  1. Misfortune


Ng was a native of Po On.  He had a younger brother and five sisters.  Around 1960, the brother was arrested in China and jailed for interfering with economic policy.  He died in prison two years ago.

Ng had eight sons and two daughters, and fished with a motorized trawler.

Ng’s father’s business had not been good.  Ng went to school for a short time, but the boat was short of hands.  His lack of schooling made him feel the pain of ignorance, so he was sending his sons to school in the 1960s and 1970s.  The family worked hard, but their boat was wrecked in bad weather.  They joined an informal savings-and-loan group and got their boat back to functional state.

They were out fishing one day when they saw an ordinary-looking hand trawler.  They came near it.  Suddenly gangsters from it leaped onto their boat and stole all their money and equipment.  Once more they were penniless.  Relatives loaned them money and they set up again.

Before they could get ahead of debt, the boat wore out.  At this time Japan conquered China.  Ng borrowed money from arms-runners and ran arms from Hong Kong to loyalists fighting the Japanese in Guangdong.  As soon as he got enough money, he stopped this dangerous business, married, and went back to fishing.

Then the Japanese took Hong Kong.  The Japanese controlled the land and knew that spies and arms had to come by sea, so they terrorized the boat people, stopping boats to search, rob, and in suspect cases kill families and burn the boats.  Sometimes they would stop a boat, rape the young women, tie everyone with barbed wire, and set the boat afire.  The fishermen were “as frightened as birds scared by arrows, and could only sit and wait for death.”  When the fishermen saw Japanese boats, they would cut their nets and flee.

Once Ng was at a small island.  The Japanese came up suddenly.  The fishermen had no time to escape.  The Japanese charged with rifle and bayonet.  Ng had been asleep.  He was wakened by the noise.  He forgot his own life and fired a pistol.  The Japense—two soldiers—jumped into a sampan and fled for shore.  The fishermen could not follow them into the shallows.  The Japanese fired a machine gun. Ng and his family swam for shore and hid in a paddy field.

The family mended the boat once again—it was full of bullet holes—and went to Macau.  They did not know the waters.  They lived on bare rations.  Fortunately, the Japanese surrendered before Ng and his family starved to death.  They had no equipment left for fishing; they lived by carrying rice.

China closed its waters to unlicensed boats in 1957.  Ng mechanized, but his engine was constantly going out of order, so poverty came again.  A fish dealer loaned him money.  “The bad fate was the will of the Sky, but at least the spirits stretched out a hand to me”:  In the third lunar month [just before I met him] he caught a school of whte croakers worth $30,000 Hong Kong dollars.

But misfortune was not finished.  In 1965 he was secretly reported to the Chinese government as having poached on their waters.  He was caught in China, taken before a people’s court, and told to confess the amount of fish he had “stolen.”  He refused, and was jailed for a month and fined ten barrels of diesel fuel.  But he went back to fishing, did well, paid off his debts, and married off a son.  Finally, after being despised for poverty, he was respected.  [Mr. Choi here quoted one of the Chinese proverbs he loved so much: “The success is called king, the loser is called thief.”]


  1. The Orphan


Luk came from Pun Yu near Guangzhou.  His father was a very poor fisherman who could never succeed, so shipped out as a crewman.  Luk was one of three sons; the other two died of hunger and disease.  His mother died of starvation when Luk was ten.  His father could no longer fish, so Luk was taken as a boat hand by an uncle.  He was thus completely alone; he got no care or attention, and was laughed at by the other boys.  He had no one.  He saw his father only at a few festivals; when they met they would embrace and cry loudly, expressing their emotions.  His uncle then came into difficulties.  Sick and unable to support his own family, he sent Luk on to another uncle.  At 13, Luk began to serve this man; his wages were a silver dollar a month.  He sorted fish, swabbed decks, and mended equipment, working up to 18 hours a day.  After two years his wages went up to two silver dollars, then three.  Finally, another relative hired him at a good wage. He married at 23 and kept working.  Eventually, he saved enough to buy a boat, fish, carry rice, and otherwise work his way up.  His father grew sick and died; Luk spent all his savings for treatment and then for the funeral.  After that, he came down to Caastle Peak and mechanized his boat.


  1. The Shady Dealer


Chat came from Chungshan.  His father was a poor farmer and foodseller. When he was eight months old, Chat was adopted out to a childless boat family.  At 17 he married and took over the business.

The Japanese came, and he shifted to Castle Peak.  Illegal gangs terrorized fishermen.  But Chat associated with them and became friendly so was not bothered.  In the war, he set himself up running arms.  The Japanese cracked down, but Chat’s underworld contacts informed him and he escaped to Tung Kun.  He anchored at a different place every day, returning to Castle Peak after the war.  He went into dynamite fishing, another shady occupation.  When the Communists took over China, once again arms-running was profitable, and Chat smuggled arms from Hong Kong to rich merchants and mobsters.  He moved on land, built a wood hut, and sold firewood, rice, and groceries, but went bankrupt.  He fell back on carpentery, building illegal squatter structures.  Eventually he got his old shop back, and turned it into a café.  He then started to make dubiously valid identity cards, and otherwise deal less than legally.  He was made a representative for the local fishermen and for the Nationalists (on Taiwan).

His wife worked the café while he worked construction. He drank heavily and took a concubine.  His wife objected, and just at this point he lost his money in a bank failure.  [Chat was one of the least stable and most “hard” boat person I knew—a strange, troubled, dark personality.  When we interviewed him, he was living in poverty in a small room, spending much of his time dreaming of the past.]



  1. The Leader


Pat was a Po On man who lived most of his life at Tai O and Castle Peak.  When Pat was eight, his father died.  At 16, Pat took a job as boat hand.  He took part in the great seamen’s strike of 1925.  After that, he borrowed money and fished the Pearl estuary.  At this time he married.

By then he was already respected among fishermen, and served as local representative.  When the Japanese came, his fishing was ended.  Loyal to China, he was hounded by the Japanese and reduced to starvation.  For six years he and his family lived on sweet potato stems, rice husks, and rarely some sweet potatoes or thin rice.  His wife and two of his three daughters starved to death.

After the war, he remarried, and carried grass for mats from Chinese areas to Hong Kong, returning with salt or sugar, not bothering with minor formalities such as customs duties.  The Communist government pursued him for this and for his role in Nationalist local government, so he came to Hong Kong with a new boat, and mechanized there.

He worked his way up as representative as a group of relatives and anchorage neighbors.  He got into a conflict with land people when an outhouse was built in front of his mother’s grave, ruining its fungseui and endangering the living.  Eventually, after sharp conflict, he found a place to move her body.

[As of 1966, Pat, one of my most constant companions and superb consultants, was a bluff, powerfully built, extraverted man of 59.  He was a good talker,and had a Rabelaisian capacity for food, sex, alcohol, jokes, folksongs, and living in general.  Unfortunately, his fortunes declined, and by 1975 he was frail and living in poverty again.  He died soon after.]



  1. The Beaten Man


Kau was from Pun Yu.  His family ran an old sailing trawler.  Living got worse as China restricted its waters after 1957.  Men of his surname and home area built up a network of large trawlers.  (Another, intermarried surname group lost some membership to opium smuggling and smoking.)  In the Qing Dynasty the “Kau” trawlermen were rich and powerful, and some highly educated.  Their large, capacious boats proved good for smuggling.  More honest ones turned to mullet, caught when these fish run down from the rivers at the first cool weather.  The trawlermen could throw huge festivals, educate their children (“even the girls”), and marry well.  Many men went on shore and invested in land businesses, while the more beautiful girls could marry well-off land-dwellers.

The Japanese conquest drove Kau’s family to Macau.  Search for food forced them to sail back to Japanese-controlled areas, where they were once accused of smuggling arms and tied up.  Only the fortunate appearance of another Japanese ship with an interpreter on board saved them [the second case in our narratives of this common situation].  Sailing trawlers were suspect.  For one thing, they carried dynamite for fishing, and could sometimes use it on pirates or smuggle it to loyalist soldiers.  The experience with the Japanese was thus frightening; “when you talk to the tiger, you are frightened right then.”

Times were bad.  They sold their gold ornaments for food.  Kau’s elder brother got into fish dealing, was caught in China, and sent to a distant camp for reeducating such capitalists.  Others also went into fish dealing.  Kau and one brother carried on the fishing.  A typhoon destroyed the huge old boat, and Kau survived on tiny Mouse Island.  His wife prayed all night to the Sky.  They went on shore, but could never get money together to get a new boat or start a business.  They survived from relatives’ incomes, and Kau retired.  [Still fairly young and in good shape, Kau had given up; he spent his time brooding.]



  1. The Restauranteur


Sap was a young man, born locally to a Po On family.  In the late Qing Dynasty, unrest was severe in Po On.  Violence, piracy, feuding, and crime were common. Deserted islands became criminal hangouts.  So Sap’s grandfather came to Castle Peak, where he fished with a casting net [a small net thrown from a sampan—an exceedingly humble way to make a very slender living].  At that time, Castle Peak Bay was almost deserted.  The family caught good fish and sold them at the old market at the foot of the Peak.  Their living was stable, but the boat was overcrowded.  The grandfather erected a small wooden hut on shore.  The youths went out to sea often, but were better off than living on the boat.  Some of the family took up shore occupations.

The Japanese occupation of Hong Kong came when Sap was very young.  His father could not get hemp to make nets.  He sailed to occupied territory to find some, was caught by pirates, and was killed.  Three rival pirate gangs feuded over the extortion racket in the area; Sap’s father paid off one, and was killed by another.  Sap’s mother worried at her husband’s failure to return, and eventually got the story from an eyewitness, a fisherman.  The body was found and buried.

She continued fishing with the help of the children.  They were reduced to sweet potato meal, and even that did not always come.  Family solidarity saved them; relatives supported each other.

After the war, tourists began to come to Castle Peak.  There was no road around the bay, and visitors to Castle Peak Monastery had to be rowed across from the pier.  Sap’s mother took to rowing them.  By this time the sisters were old enough to attract many young men to the boat, and the mother, too, met many gentlemen from the land.  [In short, they became singing-girls, not necessarily a low-class occupation and not necessarily involving sex.]  Under the influence of educated tourists, the mother sent Sap to school for a full seven years.  He then worked as an employee on a live fish boat.

At this time, “high-class” urban men liked to hire boats to go out fishing.  The better boats prepared food and had pretty girls to serve it.  Sap’s two sisters—who by this time had friends in high places—got enough in this way to build two small houses on shore.  Sap sold firewood.  When local wood became scarce, he thought of converting his woodyard to a restaurant.  Business was not bad, but some of the customers were rough.  He had lived on the waterfront all his life, and [modestly put, with no boasting] “I can handle such people.”  [Sap was small and wiry, but quick and strong, not a man to fight against.]

A year later he formed a partnership with a cook who had run a movable food cart for some years.  The two opened a fish shack.  Slowly, it grew in reputation, and became a highly successful restaurant, the “Garden of Delight.”  Eventually it moved to much more capacious and showy quarters, up the bay.

[By 1965, it was already famous as far as the urban areas of Hong Kong for the incredible quality of its sea food.  By 1974, the quality of the “Garden of Delight” was showing the effects of volume feeding and the decline in local sea food quality, but the restaurant was still far above the sea food palaces of Hong Kong today.]



  1. The Convicted Man


Yau came from a hand-lining family.  He had no education, and his family had always been poor.  He worked on other people’s boats and on ferries.

In 1959, the Chinese government across the border called on fishermen to celebrate the Midautumn Festival and see the wonders of the new regime.  Yau went and enjoyed himself thoroughly the whole day.  In the late afternoon he was seized without warning, driven to the local town, and kept overnight.  He was brought to court the next morning and charged with murder.  This was the first time he heard anything to indicate why he was held.  The fishermen of the area had charged him, as he understood it.  He admitted that he had killed pirates during the unsettled times of World War II, but said he had never hurt anyone else.  He was questioned off and on, day and night, by different courts and officers, for nine days.  Sometimes he was interrogated seven times a day.  No on ecould establish his guilt.  He was kept in the charge room of the local police station for two months.  He was told that this was to allow fishermen to bring charges or petition for his release on grounds of innocence.  No one appeared with either of these, so he was  held till declared guilty.  He was locked in an old hut north of Guangzhou, and made to work, managing a fish-rearing pond.  Twice a day there were “re-education” meetings.  He was moved frequently to different places.  He became sick, with vomiting and pallor, and was treated in hospital.

In 1962, he was released, and returned to Castle Peak.  He held his wife tightly and kissed her.  He felt he was in another world.  He was so miserable and weak that it took him five months to recover, and the medical treatment almost ruined the family.

[This story is strange; murder would have drawn a much harsher punishment.  The tale remains inexplicable, and how much is true must be left to speculation.  He was certainly imprisoned, but why and under what circumstances remains unclear.]












Al-Tusi, Nasir ad-Din.  1964.  The Nasirean Ethics.  Tr. G. M. Wickens.  London:  George Allen and Unwin.


Alvard, Michael S., and David A. Nolin.  2002.  “Rousseau’s Whale Hunt?”  Current Anthropology 43:533-559.


Anderson, Benedict.  1991.  Imagined Communities.  Rev. ed.  London:  Verso.

Anderson, E. N.  1967.  “Prejudice and Ethnic Stereotypes in Rural Hong Kong,” Kroeber Anthropological Society Papers 37:90-107.


—  1970a.  “The Boat People of South China.”  Anthropos 65:248-256.


—  1970b.  The Floating World of Castle Peak Bay.  Washington, DC:  American Anthropological Association.


—  1972.  Essays on South China’s Boat People.  Taipei:  Orient Cultural Service.


—  1975.  “Songs of the Hong Kong Boat People.”  Chinoperl News 5:8-ll4.


—  1987.  “Why Is Humoral Medicine So Popular?”  Social Science and Medicine 25:331-337.


—  1988.  The Food of China.  New Haven:  Yale University Press.


—  1991  “Chinese Folk Classification of Food Plants.”  Crossroads 1:2:51-67.


—  1992.  “Chinese Fisher Families:  Variations on Chinese Themes.”  Journal of Comparative Family Studies 23:231-247.


—  1996.  Ecologies of the Heart.  New York:  Oxford University Press.


—  1998.  “Child-raising among Hong Kong Fisherfok:  Variations on Chinese Themes.”  Bulletin of the Institute of Ethnology, Academia Sinica 86:121-155.


—  2005.  The Lords of the Forest.  Tucson:  University of Arizona Press.


Anderson, E. N., and Marja L. Anderson.  1973.  Mountains and Water:  The Cultural Ecology of South Coastal China.  Taipei:  Orient Cultural Service.


—   1975.  “Folk Dietetics in Two Chinese Communities, and Its Implications for the Study of Chinese Medicine.”   In  Medicine in Chinese Cultures, ed. by Arthur

Kleinman; Peter Kunstadter; E. Russell Alexander; James L. Gale.  Washington:  U. S. Dept. of Health, Education, and Welfare.  Pp. 143-176.


—   1978.  Fishing in Troubled Waters.  Taipei:  Orient Cultural Service.


Anderson, E. N.; Teik Aun Wong; Lynn Thomas.  2000.  “Good and Bad Persons:  The Construction of thical Discourse in a Chinese Fishing Community.”  Bulletin of the Institute of Ethnology, Academia Sinica 87:129-167.


Antony, Robert J.  2003.  Froth Floating on the Sea:  The World of Pirates and Seafarers in Late Imperial China.  Berkeley:  University of California Press.


Aswani, Shankar.  1999.  “Common Property Models of Sea Tenure:  A Case Study from Roviana and Vonavona Lagoons, New Georgia, Solomon Islands.”  Human Ecology 27:417-453.


Atran, Scott.  1990.  Cognitive Foundations of Natural History.  New York:  Cambridge University Press.


— 2002.  In Gods We Trust.  New York:  Oxford University Press.


Baker, Hugh.  1968.  Sheung Shui, a Chinese Lineage Village.  Palo Alto:  Stanford University Press.


—  1979.  Chinese Family and Kinship. New York:  Columbia University Press.


—  1989.  Review of The Food of China by E. N. Anderson.  China Quarterly 119:661-662.


Balfour, C. S.  1941.  “Hong Kong Before the British.”  T’ien Hsia Monthly 11:330-352, 440-464.


Bandura, Albert.  1982.  Self-Efficacy Mechanism in Human Agency.  American Psychologist 37:122-147.


—  1986. Social Foundations of Thought and Action. Englewood Cliffs, NJ: Prentice-Hall.


Barkow, J.; Linda Cosmides; John Tooby.  1992.  The Adapted Mind.  New York:  Oxford University Press.


Bates, Daniel.  2004.  Human Adaptive Strategies:  Ecology, Culture and Politics.  3rd edn.  Boston:  Pearson, Allyn and Bacon.


Beck, Aaron.  1999.  Prisoners of Hate:  The Cognitive Basis of Anger, Hostility, and Violence.  New York:  HarperCollins.


Bellwood, Peter.  1997.  Prehistory of the Indo-Malaysian Archipelago.  2nd edn.  Honolulu: University of Hawaii Press.


Benedict, Carol.  1996.  Bubonic Plague in Nineteethn-Century China.  Stanford:  Stanford University Press.


Berlin, Brent.  1992.  Ethnobiological Classification.  Princeton:  Princeton University Press.


Blake, C. Fred.  1978.  “Death and Abuse in Marriage Laments:  The Curse of Chinese Brides.”  Asian Folklore Studies 37:13-33.


—  1981.  Ethnic Groups and Social Changes in a Chinese Market Town.  Honolulu:  University of Hawaii Press.


Blum, Susan.  2001.  Portraits of “Primitives”:  Ordering Human Kinds in the Chinese Nation.  Lanham, MD:  Rowman and Littlefield Publishers.


Bokenkamp, Stephen R.  1997. Early Daoist Scriptures.  Berkeley:  University of California Press.


Bond, Michael (ed.).  1986.  The Psychology of the Chinese People.  Hong Kong:  Oxford University Press.


Bonnano, George A.  2004.  “Loss, Trauma, and Human Resilience:  Have We Underestimated the Human Capacity to Thrive after Extremely Aversive Events?”  American Psychologist 59:1:20-28.


Bourdieu, Pierre.  1977.  Outline of a Theory of Practice.  Stanford:  Stanford University Press.


— 1990.  The Logic of Practice.  Stanford:  Stanford University Press.


Bourdieu, Pierre, and J. Passeron.  1990.  Reproduction in Education, Society, and Culture.  2nd edn.  Tr. Richard Nice.  London:  Sage Publications.


Bowen, William.  1992.  The Americanization of Chinese Medicine.  Ph.D. thesis, University of California, Riverside, Department of Anthropology.


Bowlby, John.  1988.  A Secure Base:  Parent-child Attachment and Healthy Human Development.  New York:  Basic Books.


Braga, J. M.  1956.  China Landfall, 1513.  Hong Kong:  Karel Weiss.


Brecht, Bertolt.  1947.  Selected Poems.  (Bilingual edition with translations by H. R. Hays.)  New York:  Grove Press.


Brown, Lester.  1996.  Who Will Feed China?  Wake-up Call for a Small Planet.  New York:  W. W. Norton for Worldwatch Institute.


Brown, Melissa (ed.).  Negotiatiing Ethnicities in China and Taiwan.  Berkeley:  University of California, Institute of East Asian Studies.


Buck, John Lossing.  1937.  Land Utilization in China.  Chicago:  University of Chicago Press.


Burckhardt, V. R.  1953-1958.   Chinese Creeds and Customs.  3 v.   Hong Kong:  South Chinese Morning Post.


Cacioppo, John T.; Gün R. Semin; Gary G. Berntson.  2004.  “Realism, Instrumentalism, and Scientific Symbiosis: Psychological Theory as a Search for Truth and the Discovery of Solutions.”  American Psychologist 59:214-223.


Campany, Robert.  2002.  To Live as Long as Heaven and Earth.  Berkeley:  University of California Press.


Cao Xueqin.  1973-1986.  The Story of the Stone.  5 vols. (last largely by Gao E).  Tr. David Hawkes and John Minford.  Harmondsworth, Sussex:  Penguin.


Cavalli-Sforza, L. L.  2001.  Genes, Peoples and Languages.  New York:  Farrar, Straus and Giroux.


Chaiklin, Seth, and Jean Lave (eds.).  1993. Understanding Practice:  Perspectives on Activity and Context.  Cambridge:  Cambridge University Press.


Chang, Eileen.  1955.  The Rice-Sprout Song.  Hong Kong:  Dragonfly Books.


Chang Kwang-chih.  1986.  The Archaeology of Ancient China.  4th edn.  New Haven:  Yale University Press.


Chen Suching (Chen Xujing).  1933.  “Account of the Shanam Tanka.”  Lingnam University Journal 3:1-151.


—  “The Origin of the Tanka.”  Nankai Social and Economic Quarterly 8:250-273.


—  Danjia de yenjiu (Tanka Researches).  Canton:  Lingnam University Press.


Chetham, Deirdre.  2002.  Before the Deluge:  The Vanishing World of the Yangtze’s Three Gorges.  New York:  Palgrave MacMillan.


Chomsky, Noam.  1957.  Syntactic Structures.  Hague:  Mouton.


Chong, Key Ray.  1990.  Cannibalism in China.  New York:  Longwood Academic.


Chu, C. Y.  1960.  “The Yellow Croaker Fishery of Hong Kong and Preliminary Notes on the Biology of Pseudosciaenia crocea (Richardson).”  Hong Kong University Fisheries Journal 3:111-164.


Clifford, James, and George Marcus.  1986.  Writing Culture:  The Poetics and Politics of Ethnography.  Berkeley:  University of California Press.


Cohen, Myron.  1976.  House United, House Divided:  The Chinese Family in Taiwan.  New York:  Columbia University Press.


Constable, Nicole (ed.).  1996.  Guest People:  Hakka Identity in China and Abroad.  Seattle:  University of Washington Press.


Croll, Elisabeth.  2000.  Endangered Daughters.  London:  Routledge.


Cruz Torres, Maria Luz.  2004.  Lives of Dust and Water.  Tucson:  University of Arizona Press.


Csordas, Thomas.  2004.  “Asymptote of the Ineffable:  Embodiment, Alterity, and the Theory of Religion.”  Current Anthropology 45:163-185.


Damasio, Antonio.  1994.  Descartes’ Error.  New York:  Putnam.


Dean, Kenneth.  1993.  Taoist Ritual and Popular Cults of Southeast China.  Princeton:  Princeton University Press.


DeBernardi, Jean.  2004.  Rites of Belonging:  Memory, Modernity and Identity in a Malaysian Chinese Community.  Stanford:  Stanford University Press.


De Certeau, Michel.  1984.  The Practice of Everyday Life.  Berkeley:  University of California Press.


De Groot, J. J. M.  1892-1910.  The Religious System of the Chinese.  Leiden:  Brill.


Deng, Gang.  1999.  Maritime Sector, Institutions, and Swea Power of Premodern China.  Westport, CT:  Greenwood.


Diamond, Norma.  1969.  K’un Shen, a Taiwan Village.  New York:  Holt, Rinehart, Winston.


Dilthey, Wilhelm.  1985.  Selected Works. Vol. 1, Introduction to the Human Sciences.  (Tr. R. Makkreel and F. Rodi.)  Princeton:  Princeton University Press.


Douglas, Mary.  1970.  Natural Symbols.  London:  Barrie and Rockliff.


Durkheim, Emile.  1995.  Elementary Forms of the Religious Life.  (Tr. Karen Fields.)  New York:  Free Press.


Eastman, Lloyd.  1988.  Families, Fields and Ancestors.  New York:  Oxford University Press.


Eberhard, Wolfram.  1942.  Lokalkulturen im alten China.  Leyden:  Brill.


—  1958.  Chinese Festivals.  London and New York:  Abelar-Schuman.


—  1963.  “Auspicious Marriages:  A Statistical Study of a Chinese Custom.”  Sociologus 13:49-55.


Edmonds, Richard Louis (ed.).  1998.  Manging the Chinese Environment.  Oxford:  Oxford University Press.


Ellen, Roy.  1993.  The Cultural Relations of Classification.  Cambridge:  Cambridge University Press.


Elliott, Alan.  1955.  Chinese Spirit-Medium Cults in Singapore.  London:  Athlone Press.


Ellis, Albert, and Shawn Blau (eds.).  1998.  The Albert Ellis Reader.  New York:  Citadel.


Elster, Jon.  1981.  “Snobs.”  London Review of Books, 5-18 Nov., pp. 10-12.


Elvin, Mark.  2004.  The Retreat of the Elephants:  An Environmental History of China.  New Haven:  Yale University Press.


Farquhar, Judith.  1994.  Knowing Practice:  The Clinical Encounter of Chinese Medicine.  Boulder:  Westview.


Feng Han-yi.  1948.  The Chinese Kinship System.  Cambridge, MA:  Harvard-Yenching Institute.


Feuchtwang, Stephan.  1991.  “A Chinese Religion Exists.”  In An Old State in New Settings:  Studies in the Social Anthropology of China in Memory of Maurice Freedman, ed. by Hugh D. R. Baker and Stephan Feuchtwang.  Oxford:  JASO.  Pp. 139-161.


Forth, Gregory.  2004.  Nage Birds:  Classification and Symbolism among an Eastern Indonesian People.  London:  Routledge.


Foster, George.  1961.  “The Dyadic Contract:  A Model for the Social Structure of a Mexican Peasant Village.  American Anthropologist 63:1173-1192.


—  1994.  Hippocrates’ Latin American Legacy:  Humoral Medicine in the New World.  Langhorne, PA:  Gordon and Breach.


—  1998.  “Disease Etiologies in Non-Western Medical Systems.”   In Understanding and Applying Medical Anthropology, ed. by Peter Brown.  Mountain View, CA: Mayfield Publishing Co.  Pp. 110-117.


Frake, Charles.  1980.  Language and Cultural Description.  (Ed. Anwar S. Dil.)  Stanford:  Stanford University Press.


Freedman, Maurice.  1958.  Lineage Organization in Southeastern China.  London:  Athlone Press.


—  1966.  Chinese Lineage and Society:  Fukien and Kwangtung.  London:  Athlone Press.


Furth, Charlotte.  1999.  A Flourishing Yin:  Gender in China’s Medical History, 960-1665.  Berkeley:  University of California Press.


Gates, Hill.  1987.  Chinese Working-Class Lives:  Getting By in Taiwan.  Ithaca, NY:  Cornell University Press.


Geertz, Clifford.  1963.  Agricultural Involution.  Berkeley:  University of California Press.


—  1973.  The Interpretation of Cultures:  Selected Essays.  New York:  Basic Books.


Gerth, Hans, and Mills, C. Wright.  1946.  From Max Weber.  New York:  Oxford University Press.


Giddens, Anthony.  1984.  The Constitution of Society.  Berkeley:  University of California Press.


Gladney, Dru.  1991.  Muslim Chinese:  Ethnic Nationalism in the People’s Republic.  Cambridge, MA:  Harvard University Press.


—  1998.  Ethnic Identity in China:  The Making of a Muslim Minority Nationality.  Fort Worth, TX:  Harcourt Brace College Publishers.


Habermas, Jurgen.  1984.  The Theory of Communicative Action.  (Tr. Thomas McCarthy.)  Boston:  Beacon Press.


Harper, Donald.  1998.  Early Chinese Medical Literature:  The Mawangdui Manuscripts.  London:  Kegan Paul International.


Harré, Rom (ed.).  1986.  The Social Construction of Emotions.  Oxford:  Blackwell.


Harrell, Stevan.  2001.  Ways of Being Ethnic in Southwest China.  Seattle:  University of Washington Press.


— (ed.).  1995.  Cultural Encounters on China’s Ethnic Frontiers.  Seattle:  University of Washington Press.


Harris, Marvin.  1968.  The Rise of Anthropological Theory.  New York:  Thomas Y. Crowell.


Hawkes, David.  1959.  Ch’u Tz’u, Songs of the South.  Oxford:  Oxford University Press.


Hayami, Yujiro, and Vernon Ruttan.  1985.  Agricultural Development.  2nd edn.  Baltimore:  Johns Hopkins University Press.


Herklots, G. A. C.  1972.  Vegetables in South-East Asia.  London:  George Allen and Unwin.


Herklots, G. A. C., and  S. Y. Lin.  1961.  Common Marine Food-Fishes of Hong Kong.  Hong Kong:  South China Morning Post.


Hinrichs, T J.  1998.  “New Geographies of Chinese Medicine.”  Osiris 13:287-325.


Ho Kaak-yan (Ho Ge’en).  1944  Account of the Tanka.  Canton:  East Asia Publishing Co.  (In Chinese.)


—  1965.  “A Study of the Boat People.”  Journal of Oriental Studies 5:1-40.


—  1966.  “The Tanka or Boat People of South China.”  In Symposium on Historical, Archaeological, and Linguistic Studies on Southern China, Southeast Asia, and the Hong Kong Region, ed. by F. S. Drake.  Hong Kong:  Hong Kong University Press.  Pp. 120-123.


Hrdy, Sarah Blaffer.  1999.  Mother Nature:  A History of Mothers, Infants, and Natural Selection.  New York:  Pantheon Books.


Hsu, Elisabeth.  1999.  The Transmission of ChineseMedicine.  Cambridge:  Cambridge University Press.


— 2001a.  “Introduction.”  In  Innovation in Chinese Medicine, ed. by Elisabeth Hsu.  Cambridge:  Cambridge University Press.  Pp. 1-11.


— 2001b.  “Pulse Diagnostics in the Western Han: How Mai and Qi Determine Bing.”

In Innovation in Chinese Medicine, ed. by Elisabeth Hsu.  Cambridge:  Cambridge University Press.  Pp. 51-91.


Hsu, Elisabeth (ed.).  2001.   Innovation in Chinese Medicine.  Cambridge:  Cambridge University Press.


Huang, Philip.  1990.  The Peasant Family and Rural Development in the Yangzi Delta, 1350-1988.  Stanford:  Stanford University Press.


Humphrey, Caroline.  1996.  Shamans and Elders.  Oxford:  Oxford University Press.


Hung, Ho-fung.  1998.  “Thousand-year Oppression and Thousand-year Resistance.”  Chinese Sociology and anthropology 30:75-99.


Hung, Ho-fung.  2001a. “Identity Contested:  Rural Ethnicities in the Making of Urban Hong Kong.”  In Hong Kong Reintegrating with China:  Political, Cultural and Social Dimensions, ed. by Lee Pui-tak.  Hong Kong: Hong Kong University Press.  Pp. 183-201.


Hung, Ho-fung.  2001b.  “Imperial China and Capitalist Europe in the Eighteenth-Century Global Economy.” Fernand Braudel Center Review 24:473-513.


Hunn, Eugene.  1990.  Nch’i-Wana, The Big River.  Seattle:  University of Washington Press.


James, William.  1903.  The Varieties of Religious Experience:  A Study in Human Nature.  New York:  Longmans, Green and Co.


Jankowiak, William R.  1993.  Sex, Death, and Hierarchy in a Chinese City.  New York:  Columbia University Press.


Kani, Hiroaki.  1967.  A General Survey of the Boat People in Hong Kong.  Hong Kong:  Southeast Asia Studies Section, New Asia Research Institute, Chinese University of Hong Kong.


Kant, Immanuel.  1978.  Anthropology from a Pragmatic Point of View.  (Tr. by Victor Lyle Dowdell.  German original 1798.)  Carbondale:  Southern Illinois University Press.


King, Frank H. H.  1955.  “Pricing policy in a Chinese Fishing Village.”  Journal of Oriental Studies 1:215-226.


Kipnis, Andrew.  1997.  Producing Guanxi:  Sentiment, Self, and Subculture in a North China Village..  Durham, NC:  Duke University Press.


Kleinman, Arthur.  1980.  Patients and Healers in the Context of Culture.  Berkeley:   University of California Press.


— 1986.  Social Origins of Distress and Disease:  Depression, Neurasthenia, and Pain in Modern China.  New Haven:  Yale University Press.


Kleinman, Arthur; Peter Kunstadter; E. Russell Alexander; James L. Gale (eds.).   1975.  Medicine in Chinese Cultures.   Washington:  U. S. Dept. of Health, Education, and Welfare.


Kong, C. Y.  1996.  Huihui Yaofang.  Hong Kong:  C. Y. Kong.


Kronenfeld, David.  1996.  Plastic Glasses and Church Fathers.  New York:  Oxford University Press.


Langer, Ellen.  1983.  The Psychology of Control.  Beverly Hills:  Sage.


Lave, Jean.  1988.  Cognition in Practice.  Cambridge:  Cambridge University Press.


Lave, Jean, and Etienne Wenger.  1991.  Situated Learning:  Legitimate Peripheral Participation.  Cambridge:  Cambridge University Press.


Levinas, Emmanuel.  1969.  Totality and Infinity.  (Tr. Alphonso Lingis.)  Pittsburgh:  Duquesne University Press.


Lévi-Strauss, Claude.  1958.  Anthropologie structurale.  Paris:  Plon.


—  1962.  La pensée sauvage.  Paris:  Plon.


Litzinger, Ralph.  2000.  Other Chinas:  The Yao and the Politics of National Belonging.  Durham, NC:  Duke University Press.


Liu, Tao Tao, and David Faure (eds.).  1996.  Unity and Diversity:  Local Cultures and Identities in China.  Hong Kong:  Hong Kong University Press.


Lo Hsiang-lin.  1963.  Hong Kong and Its External Communications before 1842.  Hong Kong:  Institute of Chinese Culture.


Lowie, Robert.  1947.  Primitive Society.  New York:  Liveright.


Lutz, Catherine.  1988.  Unnatural Emotions.  Chicago:  University of Chicago Press.


MacElroy, Ann, and Patricia Townsend.  1996.  Medical Anthropology in Ecological Perspective.  3rd edn.  Boulder, CO:  Westview.


Malinowski, Bronislaw.  1948.  Magic, Science and Religion.  Glencoe, IL:  Free Press.


Marks, Robert.  1998.  Tigers, Rice, Silk and Silt.  Cambridge:  Cambridge University Press.


Martin, Katherine Gould.  1978.  “Hot Cold Clean Poison and Dirt:  Chinese Folk Medical Categories.”  Social Science and Medicine 12:39-46.


Mauss, Marcel.  1990.  The Gift.  (Tr. W. D. Halls.  French original 1925.)  London:  Routledge.


McCoy, John.  1965.  “The Dialects of Hong Kong Boat People:  Kau Sai.”  Journal of the Hong Kong Branch of the Royal Asiatic Society 5:1-19.


McKnight, Brian.  1981.  The Washing Away of Wrongs.  Ann Arbor:  University of Michigan Press.


Mead, George Herbert.  1962.  The Social Psychology of George Herbert Mead.  (Ed. Anselm Strauss.)  Chicago:  University of Chicago Press.


Moore, Robert L.  1981.  Modernization and Westernization in Hong Kong:  Patterns of Culture Change in an Urban Setting.  Ph.D. dissertation, Dept. of Anthropology, University of California, Riverside.


—  1988.  “Face and Networks in Urban Hong Kong.  City and Society 2:50-59.


Mote, Frederick.  1999.  Imperial China 960-1800.  Princeton:  Princeton University Press.


Mueggler, Erik.  2001.  The Age of Wild Ghosts:  Memory, Violence, and Place in Southwest China.  Berkeley:  University of California Press.


Murray, Dian.  1987.  Pirates of the South China Coast, 1790-1810.  Stanford:  Stanford University Press.


Needham, Joseph, and Lu Gwei-Djen.  2000.  Science and Civilisation in China.  Vol. 6: Biology and Biological Technology.  Part VI:  Medicine.  Ed. by Nathan Sivin.  Cambridge:  Cambridge University Press.


Ng, Peter Y. L.  1961.  The 1819 edition of the Hsin-an Hsien-chih:  A Critical Examination with Translation and Notes.  M. A. thesis, History, Hong Kong University.


North, Douglass.  1990.  Institutions, Institutional Change, and Economic Performance.  Cambridge:  Cambridge University Press.


Nowak, Margaret, and Stephen Durrant.  1977. The Tale of the Nishan Shamaness.  Seattle:  University of Washington Press.


Ogden,  and I. A. Richards.  1925.  The Meaning of Meaning.


Pa Chin.  1964.  The Family.  Tr. Sidney Shapiro.  Peking (Beijing):  Foreign Languages Press.


Pasternak, Burton.  1968.  “Agnatic Atrophy in a Formosan Village.”  American Anthropologist 70:93-96.


—  1972.  Kinship and Community in Two Chinese Villages.  Stanford:  Stanford University Press.


Piaget, Jean.  1963.  The Origin of Intelligence in Children.  (Tr. Margaret Cook.)  New York:  W. W. Norton.


Pillsbury, Barbara.  1976.  “Doing the Month”:  Confinement and Convalescence of Chinese Women After Childbirth.  Paper, American Anthropological Association, annual meeting, Washington, D C.


Potter, Jack.  1968.  Capitalism and the Chinese Peasant.  Berkeley:  University of California Press.


—  1978.  “Cantonese Shamanism.”  In Studies in Chinese Society, ed. by Arthur Wolf.  Stanford:  Stanford University Press.  Pp. 321-345.


Qin Cicong.  1995.  “Jurchen Literature under the Chin.”  In China under Jurchen Rule, ed. by Hoyt Tillman and Stephen West.  Albany:  SUNY Press.  Pp. 216-237.


Radin, Paul.  1957.  Primitive Man as Philosopher.  New York:  Dover.


Randall, Robert.  1977.  Change and Variation in Samal Culture:  Making Plans to “Make a Living” in the Southern Philippines.  Ph.D. thesis, Dept. of Anthropology, University of California, Berkeley.


Red Pine.  2003.  Poems of the Masters:  China’s Classic Anthology of T’ang and Sung Dynasty Verse.  Port Townsend, WA:  Copper Canyon Press.


Richardson, Jane.  1940.  Three Centuries of Women’s Dress Fashions:  A Quantitative Analysis.  Berkeley:  University of California Press.


Robbins, Paul.  2004.  Political Ecology.  Oxford:  Blackwell.


Rogoff, Barbara.  2004.  Learning through Intent Participation in Cultural Activities.  Paper abstract, American Anthropological Association, Annual Meeting, Atlanta.


Rogoff, Barbara, and Jean Lave (eds.).  1984.  Everyday Cognition.  Cambridge, MA:  Harvard University Press.


Sahlins, Marshall.  1972.  Stone Age Economics.  Chicago:  Aldine.


Sangren, Steven.  1987.  History and Magical Power in a Chinese Community.  Stanford:  Stanford University Press.


Schank, Roger, and Robert Abelson.  1977.  Scripts, Plans, Goals, and Understanding: An Inquiry into Human Knowledge Structures.  Hillsdale, NJ:  Lawrence Erlbaum.


Schein, Louisa.  2000.  Minority Rules:  The Miao and the Feminine in China’s Cultural Politics.  Durham, NC:  Duke University Press.


Scheper-Hughes, Nancy.  1992.  Death without Weeping:  The Violence of Everyday Life in Brazil.  Berkeley: University of California Press.


Scott, James.  1985.  Weapons of the Weak.  New Haven:  Yale University Press.


—  1998.  Seeing Like a State.  New Haven:  Yale University Press.


Sharp, Henry.  2001.  Loon.  Lincoln:  University of Nebraska Press.


Shen Fu.  1983.  Six Records of a Floating Life.  Tr. Leonard Pratt and Chiang Su-hui.  London and New York:  Penguin.


Shon, Steven, and Davis Ja.  1982.  “Asian Families.”  In Monica McGoldrick, John Pearce, and Joseph Giordano (eds.), Ethnicity and Family Therapy.  New York:  Guilford.  Pp. 208-228.


Shweder, Richard.  1991.  Thinking Through Cultures.  Cambridge, MA:  Harvard University Press.


Siu, Helen, and David Faure.  1995.  “Introduction.”  In David Faure and Helen Siu (eds.):  Down to Earth:  The Territorial Bond in South China.  Stanford:  Stanford University Press.  Pp. 1-20.


Sivin, Nathan.  2000.  “Editor’s Introduction.”  In Science and Civilisation in China.  Vol. 6: Biology and Biological Technology.  Part VI:  Medicine.  By Joseph Needham and Lu Gwei-Djen.  Cambridge:  Cambridge University Press.  Pp. 1-37.


Skinner, G. William.  1964-1965.  “Marketing and Social Structure in Rural China.”  Journal of Asian Studies, 24:3-43, 195-228, 363-400.


Skinner, G. William (ed.).  1977.  The City in Late Imperial China.  Stanford:  Stanford University Press.


Smil, Vaclav.  1984.  The Bad Earth:  Environmental Degradation in China.  Armonk, NY:  M. E. Sharpe.


Smith, Arthur.  1894.  Chinese Characteristics.  New York:  Fleming Revell.


—  1899.  Village Life in China.  New York:  Fleming Revell.


Sternberg, Robert J.  2004.  “Culture and Intelligence.”  American Psychologist 59:325-338.


Stockard, Janice.  1989.  Daughters of the Canton Delta.  Stanford:  Stanford University Press, 1989.


Strassberg, Richard E.  2002.  A Chinese Bestiary.  Berkeley:  University of California Press.


Strauss, Claudia, and Naomi Quinn.  1997.  A Cognitive Theory of Cultural Meaning.  Cambridge:  Cambridge University Press.


Sun, Lung-Kee.  2002.  The Chinese National Character:  From Nationhood to Individuality.  Armonk, NY:  M. E. Sharpe.


Sutton, Donald S.  2004.  “Shamanism in the Eyes of Ming and Qing Elites.”  In Heterodoxy in Late Imperial China, ed. by Kwang-chih Liu and Richard Shek.  Honolulu:  University of Hawaii Press.  Pp. 209-237.


Tao Jing-shen.  1976.  The Jurchen in Twelfth-Century China:  A Study in Sinicization.  Seattle:  University of Washington Press.


Topley, Marjorie.   1970.  “Chinese Traditional Ideas and the Treatment of Disease:  Two Examples from Hong Kong.”  Man 5:421-437.


—  1974.  “Cosmic Antagonisms:  A Mother-Child Syndrome.  In Religion and Ritual in Chinese Society, ed. by Arthur Wolf. Stanford University Press.  Pp. 233-249.


—  1975.  “Chinese and Wetern Medicine in Hong Kong:  Some Social and Cultural Determinants of Variation, Interaction and Change.”   In Medicine in Chinese Cultures, ed. by Arthur Kleinman; Peter Kunstadter; E. Russell Alexander; James L. Gale.   Washington:  U. S. Dept. of Health, Education, and Welfare.  Pp. 241-272.


Tu Wei-Ming.  1985.  Confucian Thought:  Selfhood as Creative Transformation.  Albany:  SUNY Press.


Tuchman, Barbara.  1978.  A Distant Mirror: The Calamitous 14th Century.  New York:  Ballantine.


Unschuld, Paul.  1986.  Medicine in China:  A History of Ideas.  Berkeley:  University of California Press.


—2003.  The Huang Di Nei Jing Su Wen.  Berkeley:  University of California Press.


Vayda, Andrew P., and Bradley Walters.  1999.  Against Political Ecology. Human Ecology 27:167-179.


Von Glahn, Richard.  2004.  The Sinister Way.  Berkeley:  University of California Press.


Wallace, Anthony.  1970.  Culture and Personality.  2nd edn.  New York:  Random House.


Ward, Barbara.  1955.  “A Hong Kong Fishing Village.”  Journal of Oriental Studies 1:195-214.


—  1958.  “A Hong Kong Fishing Village.”  Geographical Magazine 31:300-303.


—  1959.  “Floating Villages:  Chinese Fishermen in Hong Kong.”  Man 59:44-45.


—  1965.  “Varieties of the Conscious Model:  The Fishermen of South China.”   A.S.A Monographs 1:113-138.  London: Tavistock Publishers.


— 1966.  “Sociological Self-Awareness:  Some Uses of the Conscious Model.”  Man n.s. 1:201-215.


—  1967.  “Chinese Fishermen of Hong Kong:  Their Post-Peasant Economy.”  In Social Organization, ed. by Maurice Freedman.  London:  Cass.  Pp. 271-288.


—  1969.  “Temper Tantrums in Kau Sai.”  In:  Socialization:  The Approach from Social Anthropology, ed. by P. Mayer.  London:  Tavistock.  Pp. 109-125.


Watson, James L.  1975.  Emigration and the Chinese Lineage:  The Mans in Hong Kong and London.  Berkeley:  University of California Press.


Watson, James L.  2004a.  “Fighting with Operas:  Processionals, Politics, and the Specter of Violence in Rural Hong Kong.”  In Village Life in Hong Kong: Politics, Gender, and Ritual in the New Territories, by James L. Watson and Rubie Watson.  Pp. 311-323.


—  2004b.  “From the Common Pot:  Feasting with Equals in Chinese Society.”  In Village Life in Hong Kong: Politics, Gender, and Ritual in the New Territories, by James L. Watson and Rubie Watson.  Hong Kong:  Chinese University Press.  Pp. 105-123.


—  2004c.  “Funeral Specialists in Cantonese Society:  Pollution, Performance, and Social Hierarchy.”  In Village Life in Hong Kong: Politics, Gender, and Ritual in the New Territories, by James L. Watson and Rubie Watson.  Pp. 391-421.


—  2004d.  “Of Flesh and Bones:  The Management of Death Pollution in Cantonese Society.”  In Village Life in Hong Kong: Politics, Gender, and Ritual in the New Territories, by James L. Watson and Rubie Watson.  Pp. 255-389.


—  2004e.  “Standardizing the Gods:  The Promotion of Tian Hou (‘Empress of Heaven’) along the South China Coast, 960-1960.  In Village Life in Hong Kong: Politics, Gender, and Ritual in the New Territories, by James L. Watson and Rubie Watson.  Pp. 269-309.


Watson, James L., and Rubie Watson.  2004.  Village Life in Hong Kong:  Politics, Gender, and Ritual in the New Territories.  Hong Kong:  Chinese University Press.


Rubie Watson.  2004.  “Chinese Bridal Laments:  The Claims of a Dutiful Daughter.”  In Village Life in Hong Kong: Politics, Gender, and Ritual in the New Territories, by James L. Watson and Rubie Watson.  Pp. 221-249.


Wallace, Anthony F. C.  1970.  Culture and Personality.  2nd edn.  New York:  Random House.


Weber, Max.  1950.  From Max Weber.  (Ed. Hans Gerth and C. Wright Mills.)


Weller, Robert.  1987.  Unities and Diversities in Chinese Religion.  Seattle:  University of Washington Press.


Wen Duzhong and David Pimentel.  1986a.  “Seventeenth Century Organic Agriculture in China, Part I:  Cropping Systems in Jiaxing Region.”  Human Ecology 14:1-14.


—  1986b.  “Seventeenth Century Organic Agriculture in China, Part II:  Energy Flows through an Agrosystem in Jiaxing Region.”  Human Ecology 14:15-28.


White, Leslie.  1949.  The Science of Culture.  New York:  Farrar, Straus and Cudahy.


Wierzbicka, Anna.  1999.  Emotions across Languages and Cultures:  Diversity and Universals.  Cambridge:  Cambridge University Press.


Wolf, Arthur.  1974.  “Gods, Ghosts, and Ancestors.”  In Studies in Chinese Society, ed. by Arthur Wolf.  Stanford:  Stanford University Press.  Pp. 131-182.


—   1995.  Sexual Attraction and Childhood Association:  A Chinese Brief for Edward Westermarck.  Stanford:  Stanford University Press.


—  2003.  “Maternal Sentiments:  How Strong Are They?”  Current Anthropology 44:S31-S50.


Wolf, Margery.  1968.  The House of Lim:  A Study of a Chinese Farm Family.  New York:  Appleton-Century-Crofts.


— 1972.  Women and the Family in Rural Taiwan.  Stanford:  Stanford University Press.


Wolpert, Lewis.  1993.  The Unnatural Nature of Science.  Cambridge, MA:  Harvard University Press.


Wu Ch’ing-Tzu (Wu Chingzi).  1957.  The Scholars.  Tr. Hsien-li Yang and Gladys Yang.  Beijing:  Foreign Languages Press.


Wu Yuey-len (Wu Yuelin).  1936a.  “Investigation and Report on Samshui Rivermouth.”  Lingnam University Journal 2:1-53.


—  1936b.  “The Boat People of Shanam:  A Statistical Study of Population and Economic Conditions.”  Nankai Social and Economic Quarterly 9:613-665.


—  1937.  “Life and Culture of the Shanam Boat People.”  Nankai Social and Economic Quarterly 9:807-874.


Yan Yunxiang.  1996.  The Flow of Gifts.  Stanford:  Stanford University Press.


Yang, Mayfair.  1994.  Gifts, Favors, and Banquets:  The Art of Social Relationships in China.  Ithaca, NY:  Cornell University Press.


Ye Xian’en.  1995.  “Notes on the Territorial Connections of the Dan.”  In Down to Earth:  The Territorial Bond in South China, ed. by David Faure and Helen Siu.  Stanford:  Stanford University Press.  Pp. 83-88.


Young, James, and Linda Garro.  1994.  Medical Choice in a Mexican Village.  Boulder: Westview.


Young, John Aubrey.  1974.  Business and Sentiment in a Chinese Market Town.  Taipei:  Orient Cultural Service.


Yu, Anthony.  1977-1983.  The Journey to the West.  (Translation/edition/commentary on Xi Yu Ji by Wu Cheng’en.)  Chicago:  University of Chicago Press.


Zajonc, Robert.  1980.  Feeling and Thinking:  Preferences Need No Inferences.  American Psychologist 35:151-175.






[1] The Yucatec Maya lump certain orioles that nest at the tips of branches, while separating others that look the same but nest in the crowns of palms.  Genetic research has just confirmed that the relationships do indeed go with nesting behavior, not with appearance.  The Yucatec were ahead by about 3,000 years.


[2] The Hong Kong fishermen could have said, with my Maya friends in south Mexico:

“Es la ley de San Garabato,

comprar caro y vender barato.”

“It’s the law of St. Hook:

buy dear and sell cheap.”

St. Hook may not be in the Catholic Encyclopedia, but his law is, sadly, much truer than a great deal of religious or economic teaching.


[3]  Producers’ cooperatives have a long and honorable history; we in California have known Del Monte (fruits and nuts), Sunkist (citrus), and Calavo (avocados) for decades.  Similar coops continue to flourish in many parts of the world, but have become less important with the rise of agribusiness and the cult of the “individual” in economics and economic development.  In the 1970s, microlending appeared on the scene, thanks to Bangladeshi economist Mohamed Yunus, and more or less replaced cooperativization as the popular thing to do in the world of development aid.

An interesting comparison emerged when Marja Anderson and I studied cooperatives in Chinese fishing villages in Malaysia in 1970-71 (Anderson and Anderson 1978).  Here, coops were less successful.  First, the fish buyers there were in fierce competition, which forced them to pay top dollar to the fishermen, extend credit on reasonable terms, and shave their own profit margins to the minimum; therefore, the cooperatives could not improve terms for the fishermen.  Second, the cooperatives were not always free from corruption.  Third, the government used the cooperatives as one way of managing and controlling the fishermen.  These being (like the Hong Kong boat people) a less-than-docile group, they did not take kindly to such control, and thus distrusted the coops.  Withal, coops worked, not spectacularly, but well enough to keep the fishery alive until it succumbed to the same forces that destroyed it in Hong Kong.


[4]  This was, interestingly, no less true of our own children than of the Cantonese; we used hikers’ baby-backpacks, not meitaai, but the real issues were doubtless more complicated.  Marja and I concluded that much of the trauma of weaning—from breast or meitaai—is the loss of contact involved.

Nothing proves the utter irrationality of the human animal better than the disappearance of the meitaai in the last 30 years.  It has gone from universality to extinction in Hong Kong, having been replaced by the baby-stroller.  Part of the reason is probably sheer laziness (babies are heavy), but most of it is clearly emulation of the west.  A custom that combined total convenience for the mother with tremendous psychological benefits for the child has been sacrificed out of abject adulation of all that is foreign.  The pattern reprises Hong Kong’s replacement—fortunately less total—of Chinese food by hamburgers and of Chinese classical music by rap.

Meanwhile, the brief career of baby-backpacks in the west has also run its course.  Today, babies suffer in strollers, unable to touch the mother, or at best are carried in frontpacks, which are as inconvenient as they are hard on the parental back and neck.


[5] Compared to Near East studies, China studies are relatively free of Orientalism in its more obvious forms, but Orientalism at its worst does raise its head.  Regrettably significant is the constant republishing of Arthur Smith’s Chinese Characteristics (1894), one of the most viciously racist books to come out of British imperialism.  It is still in print.  Even Smith had a conflicted, self-contradictory affection for the Chinese (Smith 1899, right from the first page), but in general he was a strong critic.  No books today approach his, but Lloyd Eastman’s Families, Fields, and Ancestors (1988) included a few unguarded remarks about the brutality and callous viciousness of Chinese families.  These were, however, buried deeply in an otherwise well-done survey of Chinese society, and I do not mean to imply that Eastman is in any way close to Smith ideologically.  Elisabeth Croll’s book Endangered Daughters is a still different work, a superb, insightful, deeply moving book about the tragic wastage of girls’ lives throughout Asia.  It tells no more than the sober truth, in a sympathetic and humane tone.  The point here is that all three give us worst-case stories, and perhaps the most important thing my work can do is provide a story that may balance them out a bit.

We never saw at Castle Peak the dragooning behavior and harsh discipline that characterizes “the Chinese family” in western stereotype.  As noted elsewhere, this is clearly not because we were self-deluding romantics; we had no difficulty seeing such behavior—the local people saw it as deviant, but it was quite common—in the Malaysian Chinese communities we studied in 1970-71.  There, our children participated in the general unhappiness, though of course they were shielded from any direct abuse.

Castle Peak was a very, very good place to grow up.  My children were never happier, healthier, or better behaved—to a great extent because they were treated very much better than by the wider community at home.  This was not because we were foreign; quite the reverse; as foreigners, we were subject to prejudice and some hostility.


[6]  For the record, my adopted son recently and quite accidentally located his birth mother, who gave him up for adoption at babyhood in 1967.  The reunion was emotional.  She had never ceased thinking of him, with all that “thinking of” implies in such cases.  We glady incorporated her in the family.


[7] I set off a daunting outbreak of screaming anger in Taipei in 1975 when I addressed in Taiwanese (Hokkien) an elderly man who turned out to be a mainlander.  I believe this would never happen today.  Younger Taiwanese I meet now seem to be the offspring of “mixed” marriages as often as not.







Religion” is a western European concept that is rather unusual in world perspective.  Traditional China recognized no break between the “supernatural” and the “natural”; gods, fate, and good and evil spiritual forces were as real as the wind and the scent of incense.

Indeed, it is difficult to classify such Chinese concepts as hei—the famous “qi” or “Ch’i” of martial arts and medicine—as “supernatural” or “natural,” or to classify fungseui into any of the three classic categories of “Magic, Science and Religion” (Malinowski 1948).  Qi and fungseui seem “supernatural” to a westerner, or to a western-influenced Chinese who does not believe in them, but they are “natural” to a believer.

The Chinese do recognize approximately the same category as the western idea of “worship,” in the verb paai (Putonghua bai).  The root meaning is “to pay respect with folded hands.”  On the whole, the word is equivalent to the western idea of worshiping supernaturals, and thus to the western idea of religious behavior.  Conceptually, the invisible world of beings and forces addressed by and used in paai was a single world, Heaven or Sky (t’in/tian).  However, one can paai venerable living elders, as well as ghosts and spirit-inhabited rocks, so the paai category too is different from western concepts of “religion.”  Moreover, such forces as qi are generally outside of the realm of paai.  However, qi and fungseui are associated with such entities as dragons and rock spirits, which are given paai, and which are in the realm known as “supernatural” in ordinary English.

The western opposition of “religion” and “science” is rather similar to the Chinese contrast of intentional, willed action by a single supernatural agent vs unintentional, automatic action by impersonal forces.  However, qi, as well as such things as ch’e, impersonal evil force (see below), would certainly be seen as “supernatural” by western scientists, and were certainly assimilated by Castle Peak Bay Chinese into the realm of paai and spirit mediumship as well as the world of natural, pragmatic, rationally understood forces.

In fact, “magic, science and religion” are western categories, and are notoriously unclear even to the most self-conscious western writers[7].  They represent one way to parcel out the knowledge space.  They do not represent the only way, and perhaps not the best way.  They are useful for some analytical purposes, but not for others.

This chapter concentrates on the realm of paai—the realm in which nonvisible persons and uncanny forces are contacted through the rituals and ceremonies that are the specific concrete referents of the verb.   That realm is close enough to what anthropologists mean by “religion” that I can simply use the English word without much confusion.  However, we are really dealing with two concepts here—the western concept of religion that I use for comparative purposes, and the boat people’s concept of paai and the realm of beings that people could contact only through paai rituals.  The differences between the two concepts will appear through this chapter.

Since the pioneering book by Rob Weller (1987), China scholars in all disciplines have focused on practice as well as doctrine.  The age-old concern with “the classics”—with religious texts, especially foundational ones—is still very much alive, and going through its own revolutions, but the study of actual religious practice—either through ethnography or through available documents that can be analyzed to reveal behavior—has flourished (see e.g. Dean 1993)  Even text-based historical scholarship now takes serious account of practice (von Glahn 2004).

The boat people displayed a form of Chinese religion that was unusually loose, non-doctrinaire, and otherwise adapted to the world of wandering seafarers.

Belief in supernaturals, and in a fundamental cosmic order sustained by them, is universal among humans, and has at least some biological roots (Atran 2002).  Among these roots is the human tendency to see behavior as consciously and deliberately willed unless proven otherwise.  Intentionality is the default option in human inference about actions and states.  Therefore, without evidence to the contrary, humans everywhere seem to assume that the world and everything in it is (to at least some extent) the product of deliberate actions by conscious beings.  These beings may be gods and goddesses, local spirits, animals and plants, ancestors and other deified humans, disembodied forces, or cosmic flows of energy and power (which may or may not be conscious).  For the Chinese of Castle Peak Bay, they were all of the above.

Another clear biological root is the desperate need of humans to feel in control of their lives, or at least to understand what is happening.  Religion often serves as a back-up in this regard; when all attempts at control and understanding fail, people cling desperately to hopes of a deus ex machina.

Humans must necessarily relate emotionally and intellectually with radically different others:  the world and its natural forces, the wind, the fish and stars.  This relationship, often intense, is another wellspring of religion, as pointed out by Thomas Csordas (2004).  Religion is, in part, a way to construct socially and culturally such attempts to deal phenomenologically with what Csordas calls “intimate alterity.”

No individualist or psychological account of religion is adequate.  Countless people go to the temple, church, or mosque solely because it is the social thing to do.  Such people outnumber mystics by at least a hundred to one.  Similarly, for every penitent desperate for a thin simulacrum of control from the supernatural realm, there are a hundred seeking more immediate and social validation from their congregations.  Religion differs from individual spirituality in being inescapably communal.  It involves a communion of believers, united by beliefs and rituals.

Therefore, I follow Durkheim (1995, Fr. orig. 1912) in regarding religion as a social relationship with supernatural beings, mediated through rituals and ceremonies, and sustained by a moral system.  Religion in this sense is an emergent phenomenon of society.  It is, in Durkheim’s terms, the collective representation of a social unit.  For people who believe that fish and mountains and winds are or have intelligent spirits, society does not stop at the human border, and, as Csordas points out, this socialization of the truly different is a wellspring of religiosity.  Human society remains the focal, best-known part of one’s society, but one’s society also includes gods, demons, sacred fish, tree spirits, and spirit-inhabited rocks.

Other classic explanations for religion (explanation of origins, failed science, psychological aberration, “opiate of the masses,” etc.) do not explain or adequately account for religious practice, in China or anywhere else.  One problem with them, especially with psychological explanations (which range from “temporal lobe epilepsy” to “awe”—see reviews in Atran 2002 and Csordas 2004), was classically noted by William James (1903):  “the varieties of religious experience.”  As James found, and as common experience confirms, even a small American mainstream Christian congregation typically has a range of types—from rational intellectuals to passionate mystics, from cheerful optimists to guilt-wracked sinners desperate for forgiveness, from hardheaded skeptics (who go because their spouses or parents make them) to devout believers, from seekers to unquestioning loyalists, and from subtle folk-theologians to persons of simple faith.  Claims that “religion is X,” where X is some one psychological thing, must be rephrased:  “For some people, religion appears to be X, as shown by behavior 1 and statement 2—but the same congregation has people for whom religion appears to be W, Y, Z….n.”  The waterfront of Castle Peak Bay had all James’ types of individual-level variation, and more.

At the social level, the commonest source of variety was difference in the gods worshiped.  Each community had its own set.  Individuals might add particular devotion to a particular deity.  Women were often particularly fond of the Goddess of Mercy, while men often selected more martial patrons.

The western world often deploys an argument for religiosity known to Christians as “Pascal’s Wager” and to Jews as “Spinoza’s Wager”:  better to believe, because if you are wrong you lose nothing, and if you’re right you gain Heaven; but if you do not believe, and are right, you gain nothing, while if you are wrong, you get eternity in Hell.  Castle Peak Bay residents I interviewed gave this as by far the commonest reason for worshiping, but they had many more gods to believe or disbelieve in.  They burned incense and sacrificed food to any gods and spirits that were popular or salient, in the hope that at least one might be moved to help them.   Repeated disappointment, conversely, often convinced a worshiper that the god in question was ineffectual or indifferent.  The worshiper would then turn to a different god.  Alternatively, he or she might try other pragmatic methods of getting the god’s attention.  In times of drought, a god’s image might be taken out of the cool, humid temple and stood in the hot sun, to suffer along with the disappointed faithful.


Religion is universally used as a carrier for moral teachings or other basic principles of the society’s order.  Chinese religion has often been analyzed from this perspective (A. Wolf 1974).  It is not adequate to provide simple equations of Chinese society with that of the “other world,” since, at the very least, there is much slippage; the “other world” still reflects Qing imperial reality, not the state organization of  Communism.  Durkheim works at a broader level; people inject not only their image of society, but all their social concerns and social fears, into their beliefs about supernaturals.  Symbols become “multivocal” and complex (Sangren 1987; von Glahn 2004:8-9).  Actual religious practice draws on a range of meanings and ideologies, usually in the service of some immediate practical end.  However, the whole framework—the belief system, the symbols that express it, the practice that uses these symbols in ordinary life—is a social representation system, and cannot be understood any other way.

Traditional spoken Cantonese has no word equivalent to the English word “religion.”  The term “religion” in standard anthropological usage refers to the social construction of reverential belief in supernatural entities:  spirits, gods, ancestors, etc.  Religion is, by definition, a structured cultural system.  It may be loosely structured and individually interpreted, but it does not include purely idiosyncratic spiritual views.

The word kaau (jiao), “teaching,” is often translated as “religion,” but actually it refers to any organized body of doctrine.  This both excludes the folk religion and includes some philosophies that are not religious in the English-language sense.  In Castle Peak Bay Cantonese, the term kaau was often expanded to tsung kaau, “ancestral teachings.”  These were, focally, Confucianism, Taoism, and Buddhism—conveniently combined in the Three Sages Temple, the main temple for land-dwellers on the Castle Peak Bay shore.  It was dedicated to Confucius, Lao Zi, and Buddha, worshiped together.  Confucianism is considered a philosophy, not a religion, in standard academic and elite discourse on China, but at the folk level Confucius was a deity and Confucianism had, at the very least, a religious side, related to ancestor worship.  Christianity was also a kaau.

The religion of the boat people was a variant of Chinese folk religion.  In the past, Chinese folk religion was sometimes mistakenly considered a fusion of Confucianism, Taoism, and Buddhism, but it is now widely recognized as a separate thing in itself.  Archaeology has revealed that its roots are in ancient Chinese religion, which long predates any kaau.  The folk religion was a vast congeries of rites and beliefs.  Eventually, it incorporated much from the three kaau, and overlapped with them in many ways, but it had an independent existence with its own rituals and entities.  The kaau were separate from each other and sometimes competing; by contrast, the folk religion paid reverence to leaders of all three.

Some have questioned whether there is a “Chinese religion” at all, given the diversity and lack of definition.  Stephan Feuchtwang titled an article “A Chinese Religion Exists” (Feuchtwang 1991).  A Buddhist might argue against this; for Buddhists, nothing exists.  Most others will be convinced by Feuchtwang, to whom I leave the issue.  Suffice it to say that the boat people saw a single unified body of worship, sharply demarcated from, say, Christianity.

The verb paai begat the verb-object forms paaipaai (“to perform a worshiping act”) and paai san (bai shen) “to pay reverence to, or worship, deities.”  Theoretically, living elders can be reverenced, but normally the word in Castle Peak Bay Cantonese referred to sacrifice rites performed for deities and spirits.  Such beings are “supernaturals” in English, but both land and water people of Castle Peak Bay explicitly stated on many occasions that such beings were part of the order of the cosmos just as humans and animals were.  The English word “supernatural” can be used only with the understanding that it covers beings that are “above nature” to the average reader of this book, but not to its subjects.  San (shen) are supernaturals, especially gods, but by inclusion any supernaturals.  They are usually invisible and intangible, but real.  (A few—the sacred fish—are visible and tangible, though perhaps the invisible in-dwelling spirit that makes them sacred is what is really intended by the term.)  They are firmly within the natural order.  In fact, they corresponded closely with the human social order—specifically, the Qing Dynasty social hierarchy.  The Jade Emperor ruled Heaven, and had his ministers and local governors, and so it went, on down to the lowest.

Supernaturals were defined, however, by the paai rite.  Supernaturals differed from ordinary mortals in that they could be reached only through special types of rituals.  These opened communication with the other world, the realm of Heaven and Hell and of ghosts and spirits.

Usually, communication was invoked in order to obtain something, such as curing of illness, good weather, or safe fishing on the stormy sea.  The boat people extended the rule for human society:  dealing with a powerful person or agent necessitates building up relations of good feeling and mutual obligation.  One sacrifices to the gods—giving them food and wealth—so that they will reciprocate by dispensing what they control, especially health and luck.

The worshiper hoped to make the gods sifuk—happy or comfortable—so that they would be disposed to help.  (I rather irreverently titled a paper “The Happy Heavenly Bureaucracy” [Anderson 1972].)  Paai rites defined folk religion; it was the set of activities defined by the need of these rites for opening and maintaining communication.   At least a minimal paai rite—burning three sticks of incense and bowing reverently—had to occur if any contact was to be made and any communication effected.  For important transactions, far more elaborate rites were performed, including sacrifices of many pigs, whole logs of sandalwood, and much else.

“Magic, science and religion” came together in the quest for long life, numinous power, mystic enlightenment, paranormal powers such as invulnerability, and, above all, good fortune.  Fungseui and much of folk medicine (the minor traditions, such as herbs to chase ghosts away) were conspicuous examples of this realm.  Some practices were clearly magical by modern western standards; sympathetic magical practices were typical, such as serving lotus seeds (lin ji, Putonghua lianzi) to encourage the coming of abundant descendents (lin ji).  Few people took this sort of magic seriously.  They were more apt to call it “trivial” or “superstitious” (mixin in Putonghua).  However, it graded upward into serious and frightening rituals like the invulnerability cults.  Some practices were rational and empirical, and thus science-like.  Others were frankly paai.

This realm is abundantly attested in older literature.  One historical maestro of it was Ge Hong, whose writings on immortality and transcendence can be analyzed as religion, science, or magic with equal ease and equal inadequacy (see Campany 2002).  Another was Tao Hongjing, the great polymath of the 6th century, who was far more seriously involved in both science and religion, but who also devoted much of his effort to this intermediate realm.

One is reminded of the endless debates over whether Chinese classical thought—the thought of Confucius, Zhuang Zi, and their contemporaries—was “philosophy” or not.  In fact, it partakes somewhat of philosophy, somewhat of political science, somewhat of descriptive psychology.  By modern western definitions, it is not any of the three; yet, at the same time, it is all of them and more.  Chinese categories of knowing, and especially Chinese knowledge of nonordinary and uncanny matters, do not fit well into western categories.  Trying to make them do so has been the bane of sinology.



Religion in the sense of paai, and religious space, were rather sharply defined by incense.  Just as the minimal paai focused on burning three incense sticks, any paai performance was spatially defined by the incense smoke.  The sacred zone was established by the aroma, which was pleasing to the gods and necessary to any interaction with them.  In outdoor rites, incense (often, as noted, whole logs of scented wood) was constantly burning, and any ritual regulations were enforced within the area defined by constant scent of the smoke.  The secular world began at the boundary of the smoke pall.

Religious space could be anywhere.  Temples, necessarily on shore, provided the most secure and defined space (see Anderson 1970:168-171 for a detailed description of temples in the area).  The major temple for the boat people was that of the goddess Tin Hau (T’in Hau, Tian Hou, “Heavenly Queen,” known also as Ma Tsou, Putonghua Ma Zu).  This temple was in a relatively unpopulated area just under Castle Peak.  In 1965, a more venerable image of Tin Hau was brought from China by boat refugees.  This image remained on the large boat of the refugee family until a small temple was built for her; this temple was a second focus of the cult as of 1974-75.  The land-dwellers had other, quite different temples, including the Three Sages Temple noted above.  There was also a Taoist temple of the formal Taoist religion—the Tou kaau (Dao jiao)—near the Bay, but it appealed to an educated, urban clientele, and had virtually nothing to do with Bay social life.

Other defined religious spaces were small and humble.  On shore were small shrines on and near piers, wells, rocks, large trees, and other landmarks.  More important were the tiny but vital religious spaces on the boats themselves.  These consisted of the shrine of the Bow Spirit—the boat’s tutelary deity—in the bow, the Sky God (t’in san, who is in charge of weather) at the foot of the main mast, the shrine of the Sea Surface Spirit at the stern, and the boat’s main family shrine in the cabin.  Of this more will be said later.  In addition to these, at New Year, firecrackers were fired from the main masts, and red paper strips (for luck) were affixed to the stern of the boat.  The entire boat was thus guarded by strategically placed shrines.


In looking over the record of religion in Castle Peak Bay, one is first struck by the sheer extent of it.  About a fourth of my field notes are descriptions of religious beliefs and ceremonies.  Religion took an enormous percentage of time, money, and effort.  Following orthodox traditions of thought in anthropology, I attribute this to two things:  the great uncertaintly of life and wealth, leading to a need to assert control (Malinowski 1948); and the great difficulty in building community, leading to a need to invest heavily in the Durkheimian construction of society through collective representation.

Indeed, religion at Castle Peak Bay fit perfectly with Durkheim’s view of religion as the collective representation of the community (Durkheim 1995, French orig. 1912).  More literally and directly than most cultures’ lists of supernaturals, China’s heavenly bureaucracy mirrors the earthly one.  The major difference is that the heavenly one never faced the depressing and “disenchanting” (cf. Weber 1950) political turmoil of modernity; it retained the beauty, splendor and glory of an idealized Qing Dynasty imperial system.  It had its emperor, officials, and ordinary citizens.  However, it was more perfect, more wondrous, more numinous than anything on earth.  Its beings had special powers, fantastic appearances, mysterious natures.  It worked in strange ways, as befitted a world with more power and more access to knowledge than ours.  Yet, in the end, it was definitely a Chinese imperial order.  It differed from the earthly one primarily in having the special traits it needed to carry out its different mission.

This mission consisted of controlling those matters that we on earth cannot control, such as weather, illness, and fate.  We humans control worship rites and the goods they provide, through paai rites, to the supernaturals:  incense, food, and paper representations of all sorts of worldly material items.  The paper goods, when burned, turn into the real things in the other world.  Our earthly role in religion is to provide these things, in a reverent and structured context, for being who will, in return, be “comfortable” enough to want to provide us with good weather, good health and good luck.

The whole system was structured down to the last detail—a Lévi-Straussian dream.  This was a realm of Castle Peak Bay culture in which structuralism was particularly valuable as a way of looking at a domain.  Yet, as in other realms, structure was dynamic.  It was constantly re-negotiated and re-created by practice.

Everyday worship forms varied little between individuals or households.  For instance, everyone burned incense sticks at morning and night to the protectors of their boats or houses.  The differences became more and more apparent as rites became more spectacular.  No two weddings were alike, because of different personalities involved and differential borrowing of western customs.  No two demon-quelling ceremonies were alike, because the demons differed—or, in outsider view, the demon-haunted people and their circumstances differed.  No two spirit-medium sessions were alike, because the questions differed according to the needs of the petitioners.

The degree of structuration of particular religious rituals varied.  The great Tin Hau ceremonies, which involved tightly coordinated effort by thousands of people, were subject to countless rules, but were never the same across time and place.  At the opposite end of the spectrum of complexity were consultations by individuals with spirit mediums.  These too had tight rules, but many fewer of them.  The dynamics of the people involved, and the nature of the requests, made every consultation a different case.  Consulting with a self-admitted master of evil powers about a bewitchment was by no means similar to consultation with a worship-woman about a son’s failure in school, let alone consulting Tin Hau—in the form of a possessed fisherman at the great Tin Hau Day rite—about ill luck in the family.


The logical place to begin an account of Castle Peak Bay religion is with the most stable and fixed structural order: the classes of the inhabitants of the other world.

The supernaturals recognized in folk religion were many and various.  Proverbially, Chinese folk religion recognizes “gods, ghosts and ancestors” (A. Wolf 1978).  The truth, for the boat people, was more complex.

The term san, often translated “gods” in English, covered beings that were relatively powerful, at least potentially benevolent, and willing to respond to paai by broad classes of people.  At the top of the hierarchy of san was the Jade Emperor, the Emperor of Heaven, equated with the North Star (and thus locally known as pak tai, “North Emperor”; he is called Zhen Wu in some parts of China; see von Glahn 2004, esp. 157-161; his elevation seems to be a Song contribution).  The Jade Emperor was the imperial dynast.  He was worshiped in a huge old temple on Cheung Chau Island, but he was not of much note at Castle Peak.  Under him were various courtiers and officials.

Locality spirits, t’ou tei (tudi, “earth locality [spirits],” often miscalled “earth gods”), corresponded to governors of particular places, and patron deities corresponded to heads of particular communities. T’ou tei kung  (Lord of the Earth) was the “earth god”—the god of the whole world as a location.  Under him were local t’ou tei (kung), who ranged from gods of provinces and districts down to the little locality spirits worshiped in tiny roadside shrines.  Often these shrines contained phallic stones (natural or roughly carved), which were once described to me as the “airports” of the t’ou tei.  He would enter and inhabit the local stone when visiting his territory.  Large, strangely-shaped rocks and trees were homes of powerful t’ou tei, showing it by their odd forms, created—the believers said—by roiling and powerful flows of qi.  Thus these natural features were worshiped with incense morning and night and frequent offerings of tea and wine.

There were also gods of particular communities and occupations.  Most important among these, to the fishermen, was Tin Hau.  A separate and self-contained hierarchy was comprised of san for particular locations.  Tin Hau was not only patroness of fishermen in general; she was the specific patron of Castle Peak’s boat community.  Hau Wong Ye (“Marquis King Father,” the deified uncle of the last boy-emperors of Sung) was the parton of Tai O, where—according to highly dubious local tradition—he had fled from the Mongols with the last fated lad.  Hung Sing (“Great Sage”) was worshiped in some places, and reportedly was the principal patron of Shanam, up the estuary (see Burkhardt 1:19-23).  A god named Taam Kung received occasional reverence, being useful in sending good weather.

The boat people also revered the deities more popular among local land-dwellers, especially Kun Yam (Kuan Yin, the Buddhist “Goddess of Mercy”).  She was the most popular san, overall, in the area, addressed—especially by women—when any mercy, compassion, or aid was needed.  Boat people went far afield to her temples, especially the huge one in Shau Kei Wan on Hong Kong Island.  She often possessed spirit mediums.

Next most commonly seen was Kuan Kung, the former general under Lau Pei (Putonghua Liu Bei) in the 3rd Century.  Kuan Kung is one of the best-known Chinese san; his loyalty made him a symbol of fidelity, and thus he evolved into a patron of businesses, which—ideally!—rely on honesty and trust.  (It also made him a patron of soldiers, but, pace a classic mistake in western literature, he is not the “God of War.”)  Essentially all businesses, no matter how tiny, had a shrine with his image in the back of the shop.  Usually, a red bulb illuminated his ruddy features and scarlet and gold robes.  No san is more conspicuously associated with the color of blood, life, strength, and activity.  He was sometimes shown with his comrades Lau Pei and Cheng Fei (Jiang Fei); they are the Three Loyal Worthies.

There were innumerable other gods, but very few of them were important. Yama, the (originally Buddhist) King of Hell, was among the important ones, since he controlled the destiny of the sinful.  He was wined and dined at his festival, in hopes that he would relent and release one’s family’s black sheep from his punitive clutches.  The three Door Guards were pasted anew—in paper form—on every set of house doors, at the New Year.  Na Ja (Nazha), the Firewheel Prince, a possessing spirit of the most wild and disturbed spirit mediums, was, like Kuan Yin, a Buddhist borrowing.  The Kitchen God was of little note to the boat people, who had no real kitchens, but was worshiped in the stove (or equivalent) by land-dwellers.  The heroes of the famous Buddhist tale told in the novel The Journey to the West (Yu 1977-1983), Monkey and his merry crew, got attention at major festivals.  Countless other san existed in Chinese worship, but the boat people rarely considered them.

The Sky Spirit or Sky God, Sea Surface God (Hoi Hau, Putonghua Hai Ho, “sea mouth”), star gods, and other more strictly Chinese rulers of natural phenomena were also important.  Boats were protected by a Bow Spirit, who was called into the bow of the boat when it was launched; they burned incense to him in the bow.  He was the boat equivalent of the door spirit worshiped by land-dwellers at the right-hand side (as one enters) of the house door.  The Sea Surface God was worshiped at the side of the boat, the Sky Spirit at the mast foot, or, in a mastless boat, the inside wall of the cabin.  The Sky Spirit watches over the real sky, not Heaven.  These were, conceptually, very different places, though the word t’in was used for both.

Boats also had shrines in protected places—deep in the cabin, usually—where Tin Hau, Kun Yam, and other deities were worshiped.  Confucius, Lao Zi, the Buddha, and—for Christian boat people—Jesus, Mary, and God the Father were often among those present.  The few Christians simply added these last to their collections of san, and were quite mystified when Western missionaries objected. Christianity had come to Castle Peak, but the few missionaries in the area were not liked or trusted; they were considered an impoverished and foolish segment of the hated colonial rule.  They were sometimes even said to be people exiled from their homelands.  Christian converts were few, and those we talked to were more interested in free medical care and other aid than in doctrine.  Indeed, they kept images of ancestors and gods on their boats, adding images of Jesus and Mary.  When the missionaries objected, the boat-dwellers replied:  “But, they are all good people, aren’t they?”

At the stern of the boat were gold-flecked red papers, which warded away bad spirits, and sometimes a shrine as well.   The boat was thus defended on all sides from evil.

Some san had always been san, including the Jade Emperor and the t’ou tei spirits.  Others were deified humans.  Tin Hau was one of these; she had been a Sung Dynasty fisherman’s daughter who saved her father and brothers, and later other people, from storms, eventually dying in one of these salvific missions.  (At least that was one common story; several variants could be heard; see Burckhardt 1953-1958, von Glahn 2004:174).  Her festival day was the 23rd of the 3rd lunar month.  This rivalled Chinese New Year as the major festival at Castle Peak.  At least 100,000 persons were in the bay for the 1966 Tin Hau festival (see below).  Her cult is extremely popular throughout southeast China—she is the most popular deity in Taiwan, for instance.  Kuan Kung and Hau Wong Ye were other former humans.  The omnipresent door gods, posted on the two leaves of the front door and on the back door, had been the bodyguards of the great second emperor of Tang, who killed his father and brothers in a palace coup and then lived in terrror of their ghosts.  His bodyguards protected him, and came thus to be the guards of houses forever after.  Many other such saintly figures were known in local stories.

Two quite different types of patron deity reveal themselves here.  The t’ou tei are patrons of an actual area—a geographic entity.  By contrast, guardian gods like Tin Hau and Hau Wong Ye are patrons of a social group—in these cases, the boat people of Castle Peak Bay and of Tai O.  The land people there had different patrons (the Three Sages and the god Kuan Tai, respectively).   These latter have helpers; the Kuan Tai temple at Tai O had a Tin Hau image in a side alcove, showing, among other things, the unusual social closeness of land and boat people in that harmonious village.

Second in importance after san were tsou sin or tsou tsung:  ancestors.  These differed from san in that they had been ordinary people, and thus were worshiped only by their families.  For the land people, who had huge and ramified lineages, all ancestors in the lineage were worshiped collectively and sometimes separately.  These ancestors were worshiped in the form of tablets bearing the name (and if possible the photograph) of the deceased.  Ideally, these would be stored in the great ancestral halls that formed the centers of the old lineage villages of the New Territories.  Lineages without the money and space to build lineage halls rented rooms or corners in local temples.

The boat people, on the other hand, worshiped their immediate family ancestors for about five generations—in practice, until no one could remember anyone who had known them in life.  They were then “sent to the sky,” by ritual burning of their images or tablets.  Among the land people, individual families within lineages did the same with their household tablets, but then their ancestors were still available for reverence, as tablets in the lineage hall or local temple.  The boat people, lacking lineages and halls, forgot their sky-sent dead forever.  Among fishermen in Malaya, we found the same contrast:  those who maintained lineage ties could travel to a lineage hall, but those without ties sent their ancestors to the sky after four or five generations.

Being almost all nonliterate, the boat people usually had small images rather than tablets.  Particularly characteristic were the images of children who died young; the boys rode tigers, the girls rode white cranes.  (If the family uses ancestor tablets instead of images, the children’s faces were painted on the tablets, instead of names.) The land people had no such commemoration of young children, since such children contributed nothing to lineage continuity.  They found the boat images uncanny and a bit disturbing.

The ancestors worshiped in these rites were the more ethereal, heaven-related components of the human soul, the uan (hun).  The soul has several components, which cluster into the ethereal uan and the earthly p’ak.  These soul-terms have been aptly translated as “cloudsoul” and “whitesoul” respectively (Bokenkamp 1997); the words are homonyms of, and probably cognates of, words for “cloudy” and “white.”  Cloudsouls and whitesouls had subdivisions or parts, about which no one was clear.  In practice, they were treated as units.  The cloudsoul goes to Heaven at death, where it is judged.  If reasonably good, it becomes a citizen of Heaven.  If bad, it is sent to Yama’s care for a purgatory period, or even for all time.  The whitesoul stays with the body, and slowly decays with it.  As the uan was worshiped on the house altar (and, for land people, in the ancestral hall), so the p’ak received food, clothing, and graveside rites, especially on the third day of the third lunar month, the Clear and Bright Day.  On that festal spring occasion, those who could do so would visit ancestral graves and perform major sacrifice rites there.  They would then hold a picnic to consume the earthly remains of the sacrificed foods and liquors.

Components of the souls can get displaced from the body during life, thus causing blackouts, “blanking” on exams, and the like.  A spirit medium may need to invoke a spirit to recover such a missing soul-piece.  Many individuals knew only that a deceased person became an ancestor but also stayed in the grave or became a ghost.

The term kuai (kuei) included both ghosts and demons.  Ghosts were the p’ak of humans who had died unnaturally or violently, or whose bodies had not been properly buried.  (The equivalent shades of those who died peacefully stayed with the body.  There was some disagreement as to just which components of the overall soul became the ghost.)   Of particular concern to the boat community were the ghosts of those lost at sea.

During 1965-66, a mentally ill young fisherman leaped overboard and was lost.  His father, who should have been responsible for him (and who was also somewhat deranged), became more and more depressed, and eventually threw himself overboard at the same spot.  This father’s elder brother (the next up the chain of responsibility) then held a major and very expensive ritual, with Taoist officiants, to placate the ghosts, since he felt them pulling him down in turn.  These ghosts were particularly fearful since they were suicides, and drowned, and lost—any one of which three conditions would have rendered them extremely dangerous, especially to family members in positions that made them ritually responsible.

Similarly, a young wife tormented by her mother-in-law could threaten (and, very rarely, actually commit) suicide; such a suicide was a “wronged ghost,” and would follow the mother-in-law and hound her to madness and death.  This was taken very seriously, and was a major sanction.  Not only did mothers-in-law genuinely fear this, but for a wife to make such a serious threat would alert the whole extended family—and, inevitably, the whole neighborhood—to her problems.  This led to pressure on the mother-in-law to be more merciful.  This dynamic is known throughout much of China.

Many demons were of obscure origin, and may or may not have been ghosts; some were ghosts or spirits of nonhuman presences.  Animals and even inanimate objects could have ghosts.

In addition to these “gods, ghosts and ancestors,” there were many spirit beings that were harder to classify.  Dragons of many species had a role in controlling rain and storms, and thus were feared.  There were also several species of san yu, “sacred fish” (see below; also Anderson 1970, 1972).

These various dragons and fish were not focal san, because they were not conscious human-like beings.  They exercised their powers naturally, not by conscious will.  The status of the indwelling spirits of trees and rocks was more ambiguous.  Such spirits had some power, but, again, were not usually thought to have the full human-like will and agency of san.  The dragons and tigers that dwell in hills, according to fungseui (fengshui), had this same character.  They worked not through deliberate agency but through natural if subtle qualities; how much actual intention they had was debatable.

There were also completely disembodied supernatural forces.  One was ling “numinous force,” which inhered in spirit-medium equipment, temples, and other religiously charged inanimate objects, and also in natural forces.  Another was disembodied evil (ch’e or ch’ei in Cantonese).  This was a force that could be controlled by witches.  There were suspected to be other, less-known forces.  Hei, the famous qi, “breath, energy, vital force,” which seems supernatural to the Westerner, was seen as a purely natural force, and thus a different sort of energy.

Finally, as a common proverb held, “even the gods are subject to fate.”  Fate, meng (ming), was the ultimate determiner of all things.  There were various types of fate. Individuals had a particular fate assigned to them; some people, for instance, were fated to have a “hard” (ngang), bitter, aggressive nature, while others were fated to be gentle.   Major natural forces, from typhoons to the cosmos itself, were ultimately in the hands of fate.  And the proverb meant what it said:  the gods, even the Jade Emperor, could go only so far in overriding set fates.  Fate had a course of its own, incomprehensible and mysterious.  One could always seek to know it, via spirit mediums or predictive texts such as the Chinese almanac, but knowledge was always cloudy and imperfect.  It required interpretation by experts, and everyone knew they were often wrong.

Interestingly, the above structure is almost exactly the same as that described by Richard von Glahn (2004, esp.164-168) as originating in a major reorganization of religious cults the Song Dynasty.  Many paragraphs of von Glahn’s account of the new Song organization could be simply carried over without change as descriptions of reality in the 1960s and 1970s at Castle Peak Bay.  The relationships of Confucianism with Daoism and Buddhism, the hierarchic organization of tudi spirits, and the relations of demon-quellers and their demons are particularly noteworthy examples.



A religious rite required three kinds of people:  those who wanted to communicate with the gods or spirits; the gods or spirits themselves; and the religious functionaries who managed the communication process.  These last were a diverse lot.

Most prestigious, but least involved in the lives of ordinary people, were the priests of the major Taoist and Buddhist temple complexes of the area.  These highly respected people dealt largely with urban elites who visited the temples.  The boat people hardly ever saw them.

Much more visible and relevant to ordinary lives were the caretakers of the local temples, notably the Three Sages Temple on the main waterfront and the Tin Hau temple across the bay from it.  In the late 1960s, a new Tin Hau temple was built near the Three Sages Temple, providing three main temples.  There were also several minor shrines without permanent staff.

A temple caretaker’s main job was to guard, clean, and maintain the temple.  Often, he transformed it into a small local art museum, arranging the various paintings and religious sculptures that were donated from time to time by worshipers.  This was common in urban areas, but rare in isolated and impoverished Castle Peak.  He maintained the ancestor tablets of land people’s lineages that had no ancestor halls and used the temple as a virtual hall.

He also cast lots.  This involved throwing two comma-shaped blocks of wood.  When they came up favorable, he would shake, in a cup, several slips of wood with fortunes written on them.  Eventually one would work its way up, projecting above the others.  This would convey the fortune of the client.  The fortunes on the slips were written in cryptic and ambiguous style.  The caretaker’s job was to interpret.  He did this very freely, using his knowledge of the client’s actual problem and situation.  Thus, in reality, the slip merely prompted the caretaker to give his own advice.

Another major category comprised the naam mo yan, “chanting people.”  Their name comes from the Buddhist mantra nam mo o mi to fat—though in fact none of the nam mo yan were Buddhist, nor did they normally chant Buddhist mantras.  (See J. Watson 2004c; he uses the slangier term naam mo lou, “chanting guys.”)  These came in two sorts.  One was the local folk-level Taoist priesthood, the tou si (dao shi, “Taoist teachers”).  They were members of more or less organized Taoist orders (“red hat” or “black hat,” usually the latter).  They held spectacular ceremonies, usually to exorcise kuai.  Their robes were vividly colored: red and black with gold and green designs.  They performed very loudly on gongs, woodblocks, drums, cymbals, and sona (a shawm, the “Chinese oboe”), alternating this with loud chanting.  Some of the ceremonies involved dramatic rites, such as slitting a rooster’s comb, whirling the unfortunate bird around to drive away demons with its powerful blood-spray, and then sacrificing it.   These Daoists were apparently members of one of the more obscure and irregular branches of the Mao Shan tradition; they had little in common with the elite Daoists of the Castle Peak temple.  Daoism, like Buddhism, is a highly diverse religion, with many sects and levels (see Bokenkamp 1997).

More common were the naam mo yan who worked in the large barnlike establishments known as paper-goods stores.  The main business of these shops was selling religious goods:  incense sticks, red candles, paper representations of money and wealth, and the ever-present firecrackers that made even the smallest ritual audible all the way to Heaven.  However, the paper-shop men (all employees were male) doubled as chanters and spirit mediums.  As chanters, they were the ordinary staff of weddings, funerals, boat-launching rituals, shop-opening celebrations, and all the other daily events that brought color and excitement to the waterfront.  Red was the dominant color.  I often reflected that Stendhal’s color system was reversed here:  “the red” was the religious sphere, “the black” was the ordinary one of the black cotton clothes and the dark wood tones of boats and buildings.  There were three shops on the waterfront, and several more not far away.

The chanting men had to know a vast range of folk chants, spells, invocations, and other texts.   Some of these came from the Daoist canon; others, probably most, were unrecorded and are probably now lost forever.  The chanters knew how to make the huge and complex paper and foil constructions that were obligatory for all major ceremonies.  They also had to put on really artistic performances; otherwise the gods would not be sifuk.

Their activities bring us to spirit mediumship.  Spirit mediums are well known in China (especially full accounts are found in de Groot 1892-1910 and Elliott 1955; see also Potter 1978, Weller 1987, and almost any older ethnography of China).  Spirit mediumship is a worldwide phenomenon, much studied by anthropologists (there are literally thousands of references; see e.g. Sharp 1993.)   As elsewhere, a medium’s role is to put themselves into a trance state, in which he or she receives a spirit from the Other World.  This spirit speaks through the medium’s mouth, writes through his or her hands, dances and sings and acts with the medium’s body.  The medium’s own person is absent or cryptic during this séance, and the medium reports no memory of the event after the spirit leaves.  At Castle Peak Bay, trance was achieved through ritual and self-hypnotic techniques:  dancing, hyperventilating, gazing fixedly at a point of light, chanting, and the like.

The paper-goods stores each had a spirit medium.  These men had very modest power, and used their abilities to call up rather low-level spirits for minor purposes.  I watched as a woman called up her dead grandmother to ask for a cure for a sick child.  The medium quickly went into apparent trance, spoke with an old woman’s voice, and produced a perfectly good—and conventional—symptomatic remedy.  I was once present when a medium was possessed by an English-speaking spirit.  I was asked to translate, but the man was producing pure glossolalia.  I explained that “it was a different dialect of English from mine,” and everyone was satisfied.

More serious spirit work involved involved sing kung (“awakening dukes” or “star dukes”; usually male) and paai san p’o (“worshiping women,” the female counterpart).  These were older men and women calling up fairly minor spirits for small fees.  One older lady would dance and do Qi Gong in a dynamic, light, graceful manner, sustaining it for many minutes; she claimed she could not do this much forceful activity when not possessed, but she was in fact very light and active at all times.  She called up ancestors, returned soul-components scared out of child minds by hard schoolwork, told fortunes, and did minor curing.  Sing kung and paai san p’o usually served in such minor curing roles, finding lost objects, and locating strayed persons.  They did not practice black magic or officiate at ceremonies.

The really serious spirit work was done by mou si (wu shi, “shaman masters”)[7].  Very few of these individuals operated in or near Castle Peak, and those few were viewed with fear and awe.  The most prominent of these was an older fisherman who became possessed by Tin Hau every year on her festival day.  He went into a trance at dawn and stayed in it till night, speaking in a feminine voice and acting like an elite woman of olden times.  During this time he had the needle-sharp point of a long magical iron stuck through his cheek; sometimes he had one through each cheek.  The iron was heavy, and was carried by several men when the mou si moved.  The mou si spent most of the day slitting his tongue and writing charms with the blood on yellow paper; this was to be burned and the ashes drunk with herbal medical tea.  Hundreds of people lined up for this curing service. The mou si allegedly felt no pain and showed no scarring.  He did, however, visibly wince when the iron was moved.

Other mou si included two—a man and a woman, not related—who were known for black magic, and made no secret of their belief in their ability to curse people.  They had started falling into trances spontaneously in childhood.  Both appeared to us to be frankly paranoid schizophrenic; they were unable to live normal lives, and were cared for by their families.  They spent their time communing with dark forces, including some highly idiosyncratic spirits.  The woman had as one familiar the “Heavenly Police Commissioner,” worshiped in the form of a small wooden sculpture of a British policeman in shorts and pith helmet.

Mou si were, in general, supposed to be “hard” (ngang) persons, able to control ch’e (evil force; see above).  Their power depended on the san preserving the mou si’s sin kuat (“immortal bone” or “fairy bone”; see Potter 1978).  If this is lost, the mou si would die.  Mou si are apt to die early in any case.  Mou si were regarded as extremely powerful, being able to take soul journeys (like Tungus and Mongol shamans), make themselves and others invulnerable, send good or evil onto others, control the behavior of women and children, and make themselves invisible.  However, they were not genuinely terrifying, nor were they involved in seriously problematic behavior, as they sometimes are in Penang, Malaysia (Anderson and Anderson 1978; DeBernardi 2004), and elsewhere.  They were rather few and weak at Castle Peak Bay, with little of the power they have in the Chinese communities of Malaysia and of Singapore (Elliott 1955).

Spirit mediums have long been regarded as uncanny and disturbing characters in China.  Donald Sutton (2004) has recently reviewed a large literature on this topic, especially the decline of spirit mediums in status after Yuan Dynasty times.  He finds exactly what we found at Castle Peak:  spirit mediums were regarded as “wild,” unregulated, outside the bounds of ordinary hierarchic society.  He points out that a few elite scholars really doubted their abilities, but most saw their abilities as real enough; the problem was that they were free, wild spirits, not subject to the tight control of the elites and of elite ritual and religious culture.  All his conclusions fit perfectly with what we heard from relatively high-status individuals on the waterfront.  Ordinary people, however, respected the spirit mediums, feared them, and saw them as necessary.  In spite of elite opposition, spirit mediums flourished in Hong Kong, as in Malaysia and Singapore and most other well-studied Chinese areas.

Thus, the traditional religious practitioners in the area were diverse, but classified into a neat order, ranked in prestige (temple priests first) and in power (mou si first).  They formed a texture of control, management, and soliciting of powerful unseen forces.



People must supplicate, wine, and dine the officials of the other world, and give them paper imitations of money, clothes, houses, and other wealth goods.  These paper items are burned and thus sent to the Sky; the food and wine are real, but the san and kuai consume only the spirit, while the worshipers consume the material remainder.

Minor magic attempted to insure against fate.  Boats often grew medicinal herbs in the stern area, and some of these plants were there to keep away kuai.  Children wore minature life-preservers, old copper cash, sea horse eyes, and other charms.  Everyone who could afford a jade bracelet wore one, and those who could not afford one often wore a glass or soft-stone copy.  Everyone knew stories of people who had fallen and broken the jade bracelet instead of an arm bone.  Our old friend Leung Taai had experienced this himself.

Humans and supernaturals (to use the western term) are complementary.  Humans can supply food and wealth goods, and entertain the supernaturals with festivals including dances, music, and Chinese operas.  Supernaturals have control over the things that humans do not control:  luck, weather, getting sick, getting a good haul of fish.  Humans must maintain an ongoing warm relationship; the gods should stay “contented.”  If they are not, they will not send good, and may even send evil.  Life’s dramatically unexpected bad fates, such as losing a man overboard in a typhoon, result from this.  However, just as people do not always completely control food and paper wealth, so the gods do not fully control fate, as we have seen.  Moreover, people, fate, and gods all have some share in managing the world.  People can live healthy lifestyles, and thus resist or partially resist any illness that fate or the gods throw at them.  Readers of imperial Chinese texts will know that Chinese officials were supposed, at least by the “common people,” to have some power over winds, waters, crocodiles, and other phenomena.  Imperial officials were expected to read edicts to locusts or rain-dragons to bring successful protection to crops.

Thus the supernaturals were far from all-powerful.  The boat people thus viewed them with less than total awe or reverence.  Humans, for their part, had some power over things usually in the spirits’ realm.  Religion was rather pragmatic—serious, but not abject or terrifying.  People had to be self-reliant; they could not entrust themselves to an Almighty.  The gods might shower favor—the man who made $30,000 HK in one net-cast went to the Tin Hau temple to thank her—but the world was always an unpredictable place.  (On fatalism and its limits in Chinese thought, see Eberhard 1966.)

Fishermen everywhere engage in magic and religion, because of the danger of their occupation.  Malinowski (1948) noted that lagoon fishing, which is safe, had less magic associated with it than the dangerous open-sea fishery.  The boat people, fishing rough and treacherous waters, secured themselves as best they could.  They were, however, a hard-headed and pragmatic group.  They were certainly no more religious than people I have fished with in other seas, including California’s.



All this knowledge came to a focus in the actual paai rites—the practice of religion.  Rites established, maintained, or re-established good relations with the supernaturals.  They also kept away evil beings.

Simplest and most universal was a small rite, universally practiced at dawn and dusk.  The female head of household, or designated replacement (usually a daughter or daughter-in-law), burned incense at each household shrine.  She made a circuit; at each ritual point, she lit three sticks of incense, bowed with them toward the shrine, and placed them below it to provide fragrant smoke[7].

A typical circuit of the boat might begin with the bow, move to the sides and then the mast foot, and end at the household shrine in the cabin.  This corresponded with the land people’s frequent circuit from door guard shrine to main household shrine.

Common for the land people, but much less so for the boat people, was burning a trio of sticks at the local t’ou tei shrine—tree, rock, or small structure.  These often received three tiny cups of tea or Chinese wine in addition to the universal trio of incense sticks.

Above this level were more elaborate offerings, including those that might periodically occur at temples, to insure luck or give thanks for it.  A minimum sacrifice rite—for an ancestor’s birthday, a blessing received from Tin Hau, or the annual day of a minor san—included not only incense, tea, and wine, but minor food offerings.  Oranges were always present, and other citrus might occur, for these native south Chinese fruits have been sacred and ritually necessary for thousands of years in the area.  The bright red-gold color of oranges, and especially of tangerines, symbolizes religion, luck, and good things in Chinese culture, even where traditional culture has faded[7].  Nuts (generally peanuts in the period under consideration, and also often melon seeds) symbolized descendents, especially sons, and thus one tried to provide an abundance of them to the gods and ancestors, in hopes of the implied return gift.  Red jujubes often accompanied these.

Larger rites included rites of passage—marriage and death ceremonies and such lesser rites as birthdays.  Still more impressive, being at community level, were calendric rituals:  New Year, gods’ days, and other festivals.  Another class of rite was apotropaic:  rites to drive away kuai or to atone for mistakes.  This was rare above the family level on the waterfront, but nearby land communities had rituals to sacrifice to the san to get their favor back if a village leader had done wrong, or sacred sites had been cultivated or profaned, or a san had been neglected.  Normally, stages were set up in front of temple doors, facing in, so the main images had a clear line of sight.  Once, on Tin Hau’s day, the opera had inadvertently been located where Tin Hau’s image in the local temple could not see it.  Three leaders of the community died soon afterward—so said local tradition, at least.  The community had to perform a whole new ceremony.

Rites of passage may be said to start with birth, but birth was little noticed.  The baby was too apt to die, or poverty might make the family give it up.  Sheer pain had led China’s ordinary folk to the consoling belief that the soul was not fully assembled until the 49th day—seven times seven days after birth.  Before that, a baby was not really human yet, and infanticide before that time has been called “postnatal abortion” in some of the relevant literature.  At Day 49, the soul was completely formed.  A small family sacrifice rite took place, and the baby was welcomed as a member of the family and the human community.  Thereafter, birthdays were virtually unnoticed, New Year being taken as “everybody’s birthday.”

The most complex and dramatic life-passage rite, by far, was marriage.  Marriages reflected economic situation.  Many occurred in early 1966, a good time following some hard years.  Few took place on our second visit; times had grown thin once more.

No two weddings were alike; families tried to make each one unique and memorable.  However, rules tightly constrained the variation.  A go-between of some sort theoretically arranged the match.  The boat people lacked money or opportunity to hire a professional, and usually used a friend or relative with high status.  Theoretically, matches were arranged by the men in consultation with the go-between.  Actually, the women of a given neighborhood formed a tight network, constantly discussing eligible young people.  They frequently did the real arranging.  Actual matchmaking usually involved much prior discussion by all concerned, climaxing with the mother (or other older female relative) of one party seeking out counterparts in the other party’s family.  They usually arranged a suitable match, involving compatible young persons.  Moreover, the young people themselves had a say in the matter.  Very often, a couple fell in love, or even developed a sexual relationship, and then asked their parents to make the arrangements.  Usually, parents respected this.  The sight-unseen marriages and child marriages that are stock themes in foreign writings on China were rare indeed on the water.

A rapidly increasing series of events, visits, and exchanges preceded the wedding.  Ritual gifts flowed back and forth (see Yan 1996 on gift-giving in a north Chinese community).  Among the boat people, the groom’s family was expected to contribute as much as, or more than, the bride’s.  This seems to be a genuine boat-dwellers’ cultural trait; it was noted by Wu for the Guangzhou boat community in the 1920s and 1930s (Wu 1937).  It stood in sharp contrast to the land people, for whom bridal dowry was all-important, with the result that girls were a drain on the family, and therefore disfavored (see Croll 2000).  Complicated rules governed the disposal of the gifts; the couple got much of it, their families got some.  In practice, this was a matter for protracted negotiations.  At least part of the reason for the difference between boat and land customs was the fact that the bride was not leaving her lineage and family forever; first, she had no lineage to leave, and, second, she was rarely moving more than a few boats away.  She would be back for visits—sometimes daily ones.  Even moving to a different community did not cut her off, since the boat people were constantly plying their craft back and forth.

Arrangements being done, a date was set, in consultation with the almanac and the temple experts.  The almanac may provide more pretext or excuse than real advice (Eberhard 1963).  More serious was the need to set the date in fall or winter, when fish are running, the weather cool, and typhoons unlikely.

Senior men of the families officiated at a wedding, along with their wives and the chanters from a local paper-goods store.  The bride and groom received attention, but focus of the actual ceremony went to the senior relatives, because marriage cemented two families—not just two persons.

A wedding took three days.  (Sometimes, the poor shortened it to two.)  The first day was spent in preparations.  The groom and his family readied the ceremonial arrangements and gifts for the bride’s family.  The latter readied their “counter-prestations” (Mauss 1990) and devoted as much attention to the bride’s dress and makeup as families do elsewhere in the world.  Finally, on the first night, the chanters would come.  The second day began with gifts from the groom’s family to the bride’s, brought by the groom’s brothers and friends..  The bride spent the morning with her family, saying goodbye.  Traditionally—as at Tai O in 1965-66—she, her mother, and other close women relatives sang long improvised chants to each other.  These were extremely emotional and moving.  Long-suppressed feelings came out in full force.  The performances were musically superb and were heart-wringing.  The songs concerned family closeness, griefs, hurts long hidden in silence, and above all close familial love.  These sessions certainly destroy any lingering credibility in the classic colonialist stereotype of the “unfeeling” or “impassive” Chinese.  (On these laments, see the excellent studies by Blake 1978 and R. Watson 2004.)

At some point in all this, the gifts would arrive from the groom’s family.  Usually some ritual foods such as peanuts and oranges were included, as well as more valuable goods.  Preparation started for the main event, the second-night celebration.

This feast varied from strictly religious observances to huge feasts with fifteen courses (many of them sea food) and hired bands.  No two second-night celebrations were alike; creativity reigned.  The bride wore a red dress, symbolic of fertility and life.  A few well-to-do families were beginning to use western-style white lacy gowns for part of the ceremony, soon changing back to the safe red to make sure of good fortune.  (Gowns of white, which is the color of death in China, were still a matter of intense discomfort for older guests.)  The groom dressed in best-quality black clothing of traditional boat style.  At this point or earlier, he took a new name; this was a religious necessity.  Feasts typically included symbolic foods, such as long noodles for long life.  Foods of ritual sacrifice, notably chicken, duck, and roast pig, were requisite.  Lotus rhizomes were boiled to develop their sticky, glutinous character; then, during the feast, someone would break a rhizome and pull the two pieces apart, to show how they remained connected by the sticky bands, “just as the couple will stick together no matter how much they may be pulled apart!”  Many foods concerned sons, or descendents in general.  Thus one might serve duck eggs for fertility, lotus seeds for abundant descendents, taro leaves for the same reason (many grow from one root), big tubers so the sons would grow big,  pig intestines for cleverness.  At this feast or at some other point on the second day, guests, relatives, and friends were expected to contribute “red packets,” red envelopes containing presents of money.  The money had to be in multiples of two (two dollars or twenty or two hundred were especially favored).  Such packets were a feature of all major ceremonial occasions, from birthdays to New Year.

On the third day, the bridge goes to her new home, atfter a long morning of farewell festivities at her own parents’.  Her final gifts and dowry come with her or close behind her.  Then comes the final feast.  Ceremonies end in time for food in late afternoon.  As usual, the women and children eat first, the senior men after them.  Women joke about the advantages of this way of showing respect; it gets them the best food when it is hot.  The senior men get cold leavings.  Women would sometimes allege that this was the intent of the ancients who founded the custom.  The bride hands round tea and wine, getting a small present from each person as she does so.  When everyone is full and happy from alcohol (or at least tea), the entertainment begins.  The heart of this is “playing with the bride.”  She has to sing, perform, answer toasts, and withstand teasing.  This last is usually good-natured, but can be barbed and even vicious.  (This was very well described in Eileen Chang’s novel The Rice-Sprout Song [1955], in the description of a Hong Kong land-dweller wedding).  At a typical boat-dweller wedding, the teasing is thoroughly good-natured, and soon merges into a long night of singing, drinking, talking, and quiet enjoyment.

Two or three days afterward, the couple would return to the bridge’s home with a few last gifts and thank-yous.  The bride would return every few days thereafter for a quick breakfast or friendly visit.  By this time she was thoroughly married, and expected to stick with her husband.  (Calling off the marriage just after it occurred was not unknown, and was accepted.)  But she did not sever ties as thoroughly as land people often did; families remained close.  After all, the marriage had been contracted, more often than not, with an eye to cementing family relationships.  Even a love match created real family ties.

Relations between the bride and her husband’s family were generally good among the boat people, who do not have the lineage-based tensions of the land-dwellers.  However, the husband’s mother inevitably has to adjust to a new situation involving some competition for the man’s support and sympathy (M. Wolf 1968, 1972).  A range of adjustments could be found, from idyllic to seriously dysfunctional (Anderson 1992).

The other major rite of passage was the funeral.  The deceased was washed in water in which pomelo skin had been soaked.  (This was the water one had to “buy”; see Chapter 1.  For this and the whole formal funeral process, far more elaborate than the boat version, see J. Watson 2004d.)   The corpse was then dressed, pillowed in the coffin, covered with a cloth, and combed.  A woman in particular had to be carefully combed and arranged, or she would become a kuai.  Then the descendents worshiped gods and ancestors, with the eldest son officiating, and the chanters from the paper shops providing ritual texts and music.  The coffin lid was nailed on with four nails.  The deceased may lie in state for varying amounts of time, but among the boat people the coffin was very soon borne to the grave.  Among the land people, reburial in a more auspicious site would often follow after some years.  Indeed, the dead might be temporarily buried, then dug up and kept in a large jar until an auspicious site could be found.  The boat people rarely had the opportunity to perform this filial act.

The family dressed in mourning:  unbleached and undyed cloth (off-white) for the closest kin, white for the next, blue and green for collaterals.  They offered food and drink at the grave.  Animals were killed later, and a solemn and often fearful feast took place.

Seven days later, paper clothes and money burned on the grave.  At this time the soul of the dead was being judged in the Court of the Sky, by the sober judge who is shown in every Chinese temple (usually in the process of watching dancing girls—possibly not the most sombre of occupations).  Again the chanting men officiated; the ceremony required seven of them, which meant that two or three shops pooled their manpower.  They held up a red banner with the name of the deceased.  They directed their attention to the “filial piety idea hall,” a tiny shrine set up near the ancestors’ shrine in the boat hold.  Pigs’ heads, indicative of sacrifice, lay before it.  At this time the women began their laments.  These were wild keening, expressive and powerful.  The women improvised the words from traditional formulas that asked the dead why he or she had left, and praying for the soul to return.  These laments came at subsequent commemorations as well.  A major part of boat folk song, they were extremely moving and powerful (see Anderson 1975), with rapid heavy rhythms alternating with long-drawn-out vocal lines.  Like lamentations in many cultures, they were a formalization of weeping sounds and rhythms.  Relatives and friends came with candles and paper goods for offering.  Red packets were not appropriate.

Smaller but similar ceremonies followed this one for six more weeks, until a final sendoff on the 49th day.  As the baby’s soul required seven weeks to assemble, so the soul of the deceased required seven to disperse.  In cool weather, the corpse was not buried until this time, since revival before it was always possible.  A ten-day cycle (memorials at the tenth day up to ten times ten) also existed, but no one seemed to celebrate it in our immediate neighborhood.  After the 49th (or, alternatively, the hundredth) day, mourning went into a lighter grade, but immediate family continued to mourn and hold special memorial ceremonies until the beginning of the third year after death—carrying on a classic Chinese pattern.

A child, especially one dying before the 49th day, was sometimes simply wrapped up and abandoned in a basket on “Little Ghost Island,” a small island in the bay.  This island was carefully avoided by everyone.  Very small children were seen as too young to bury.  Their kuai would haunt the living, however, unless their families properly cared for them.  There they remained, therefore, on the small lovely island, looking out over the bay they had inhabited for their short lives.

Once dead, a family member is an ancestor.  Memorials then shift to the san kung, a family festival held every few years to renew ancestral connections.  This was much more important to land people, with their lineages and lineage halls, than to the boat-dwellers.  The latter had only their family members to maintain, and these received a small, unobtrusive ceremony.

In one full ceremony we observed, chanters and a minor spirit medium came.  Chanting wen ton all night.  A virgin cock was brought, its comb cut by the head chanter, and its blood scattered about periodically to drive off ghosts.  The head chanter also blew consecrated water from his mouth.  The chanters summoned the spirits, then set up a paper model of Yama, and images of Kun Yam and T’in Hau.  The usual paper goods flamed before them.  One less usual paper representation was a horse-and-rider figure, a messenger summoning Yama and his cohorts.  A pole hung with red-painted tin cans guided them to the boat.

The eldest son held a censer for the chanters all night.  The second son lit candles and incense.  The wife of the family head gave a cloth, representing her own family and its contributions.  The chanters read a paper with a list of ancestors; images or tablets of these were touched with consecrated water, to brighten their way in the dark afterworld.  Then the family read out the names of the living.  The spirit medium prayed over a range of feast dishes; these went to Yama and his legions, as well as the ancestors and the chanters’ and medium’s own spirit guides, before being enjoyed by the living.  Finally, at dawn, the suffering rooster was killed, and the ancestors and spirits sent away to the Sky.



Calendric rites began and climaxed with the great New Year celebration, a ritual that, in those days, united virtually everyone of East Asian background throughout the world.  New Year normally occurred on the second new moon after the winter solstice.

Two to three weeks before the day, the boats stopped fishing.  Not only was this a holiday; there was danger of piracy.  All debts had to be paid by New Year, and desperate debtors resorted to crime, making the last two weeks of the year a dicey time to be alone.  So the boats jammed the harbor, a solid wall of weathered brown timber.  Navigation became difficult.  Shops also began to close; every business was shuttered for the first two or three days of the new year.

Weddings and other celebrations went on everywhere.  Particularly in 1966, the sounds of firecrackers and ceremonial music came every night from some part of the water.  People bought new clothes to wear at the new year.  The new year itself began at 11:00 p.m. with a spectacular firing of firecrackers.  Chains of these, hundreds of feet long, hung from masts and tall buildings.  The noise continued uninterrupted, on all sides, for the rest of the night, and sporadically through the day.  People wandered around, visited, enjoyed themselves, and above all feasted.  Tangerines, the red-gold “lucky” fruit, piled everywhere in vast arrays.  Peach flowers graced the houses and boats of those affluent enough to afford them.  The richest land people had narcissus in pots, as well.  Afternoon was spent nibbling sweets—the new year had to be sweet—and seeds, symbols of abundant descendents.  These formed merely an appetizer for the huge feasts to come, involving whole roast pigs with gold-red sugar glaze, masses of pink buns, chickens and ducks glazed or dyed red, and other foods of the fortunate colors.  These and all the minor accompaniments were washed down with seas of beer and Chinese liquors.  People had to lay down a food reserve, since they could light no fires lit until the end of the third day; food was cold leftovers till then.  (This was all right at Castle Peak Bay—Hong Kong has perhaps the coldest winters of any tropical coastal area in the world—but our experiences in Malaya with long-leftover food were harrowing.)   On the 15th day of the year, a small festival took place, but with little attention.

The next serious festivity, then, was the day of the t’ou tei, on or around the 20th of the first month.  It could involve full-scale Cantonese opera performances, as well as the inevitable firecrackers, incense, and feasting.

This and other sans’-day fairs were largely organized by religious societies, ui (hui “society”).   Von Glahn (2004:168) found cases where the word hui was generalized to mean the god’s day itself; I did not hear this particular metonymy at Castle Peak Bay, but sometimes heard pao so used.  Normally at Castle Peak the days were simply called “gods’ days” or “temple fairs.”  Secular ui were savings-and-loan associations, similar to one found all over the world (e.g. the tontines of Europe and credit rings of Mexico).  Almost all ui, however, were religious.  Religious ones saved and invested to acccumulate money for the big day.  Sometimes this was done by straight savings—a monthly sum pledged and donated.  Other ui loaned out money in rotation, as secular ui did, but charged interest of a few percent, the proceeds going to the fair.  Religious ui often integrated both boat and land people, and thus served as the only significant corporate social bodies that cut across that boundary.  The ui organized the lion dances, operas, major sacrifices, and other performances (I provided much more detail on ui in Anderson 1970b:81-83).

Next on the calendar was a different sort of festival, ch’eng meng (Qing Ming).  This ancient spring festival, on the third day of the third lunar month, commemorated the dead.  Supposedly, it was once a fertility festival, but was later toned down and regulated into a simple visiting of tombs (Eberhard 1958).  Families visited graves and picnicked there, sharing food and the usual accompanying goods to the dead.  The waiting lines at the Kowloon bus stations were a mile long on the day morning, as virtually every urbanite tried to get to family graves.  However, the occasion was not as all-important as in Malaysia, where grief over racist oppression led Chinese to develop a real cult of the dead (Anderson and Anderson 1978).

Then, on the 23rd, came the boat people’s own festival:  the day of T’in Hau.  Castle Peak Bay, with a major temple in 1965 and two after that, was a focus second only to the great temples on Hong Kong Island and at Macau (whose name is A Ma Kau, Ma Tsu’s Temple).  In early 1966 an old fisherman in a huge, majestic, decades-old sailing trawler escaped from Communist China, bringing with him a T’in Hau image over two centuries old, which he spirited away from an ancient temple near Guangzhou.  This image was on his boat for the great festival of 1966, making it—and the whole bay—an unprecedented attraction.  Worshipers soon built a temple for this image, across the bay from the other T’in Hau temple; both survive now, many decades later.  It was on the boat with the venerable image that the mou si described above spent most of the day, writing charms; for the afternoon temple festivities, he threw ritual rice about him, and then moved, with pain and difficulty, to the temple on shore.  Here he stayed, in the dense, blinding incense smoke, writing more charms.  (James Watson [2004e:296] reports that T’in Hau cults in the old lineage villages of the western New Territories lack self-torture and spirit mediumship.  This reveals another major difference from the boat world.)

The day passed with firecrackers, lion dances, and a great deal of miscellaneous “hot noise” (or “heat and noise”; the Chinese term for festive, exciting, noisy activity).  Almost everyone visited the temple and the image boat.  The pier and waterfront were spectacularly ornamented with huge red-and-gold signboards and decorations.  Cantonese operas in a huge temporary opera hall, in front of the T’in Hau temple, blared loud enough to be heard across the bay through the firecracker sound[7].

For days, the paper shops had been working overtime, constructing large frameworks of bamboo and covering them with red paper, foil of all hues, and brilliant decorations in the most vibrant colors.  These they hung with metal rings, papier-mache sculptures, bangles, and live ginger roots.  Finally, they ornamented these with figures of mythic divinities, especially Monkey and his friends.  These objects were ritual creations known as p’au, “cannons” or “rockets.”

The festival climaxed in the afternoon.  A vast crowd gathered at the temple.  Huge logs of sandalwood were burning in great bonfires, scenting the whole shore and thus making it sacred territory for the time.  A tall bamboo platform stood in the middle of the great courtyard before the temple.  Leaders of the temple organization shot rockets from this platform—one after another to the number of sixty.  As these fell to earth, the strongest young men from the T’in Hau uis leaped up to vie for the spent rocket-heads.  Whoever got the first one to fall got the best p’au sculpture, and with it the best of T’in Hau’s blessing and good fortune for himself and his whole ui for the coming year.  He was then disqualified, and the remaining uis’ men leaped up for the next.  A good deal of fighting developed, and police had to intervene several times.  So it went, until the last, most miserably unfortunate ui got its feeble final rocket.  The p’au sculptures—magnificent works of art, built to last but a day—were disassembled at the uis’ subsequent feasts and celebrations.  Parts were auctioned off.  Getting a jade(-like) ring or a ginger root guaranteed that one would have a son that year. Everyone knew of exceptions, but there were always ready explanations for these:  the man wasn’t faithful enough to T’in Hau, the jade was really glass, the p’au was the 58th instead of the first….

The money earned this way—and it was substantial, especially if several couples wanted sons badly—went to fund the next year’s activities.

After the last rocket, as dusk began to fall, everyone dispersed to an evening of feasting and drinking.

The next major occasion on the bay was Taam Kung Day, a major occasion at Taam Kung’s temple on Hong Kong Island, but minor at Castle Peak (see Burkhardt 1953-1958, esp. vol. 1: 19-24 and vol. 3: 155-156).  At a variable time in the fourth month came the Bun Festival of Cheung Chau Island, nearby; this provided huge mountains of buns for the unfortunate dead—lost to pirates or in war.  It drew many visitors from Castle Peak, Cheung Chau being another boat center where many people had relatives.

Then came the Dragon Boat Festival, the fifth day of the fifth lunar month.  This centers around boat races in light, thin boats propelled by paddles and energized by drum-beating.  The festival is formally said to commemorate the ancient poet Qu Yuan, who drowned himself to protest bad government, but the people of Castle Peak Bay denied all knowledge of this worthy, and saw the day and its races as purely a matter of placating the dragons and water ghosts.  (This is clearly the real origin of the festival, which once involved human sacrifice; see Eberhard 1958.)  Naturally, the boat people generally win the boat races, which in the main Hong Kong areas involved all ethnic groups—even the Gurkha soldiers and the colonial British.  Waterfront wisdom held that the British usually came to grief, upsetting their craft.

Hau Wong Ye, noted above as the patron of Tai O, was celebrated there on the sixth of the sixth, with a festival that was a much smaller and milder version of T’in Hau’s.

The seventh of the seventh brought unmarried girls out to celebrate the yearly meeting of the Weaving Maid (the star Vega) and the Cowherd Boy (Altair).  These stars are separated by the Sky River (the Milky Way), but every year on this date the Birds of Good Luck (hei tseuk, magpies) fly to the Sky to build a bridge with their wings, and the lovers can meet.  This festival was almost invisible at Castle Peak.  More serious was the Feast of Hungry Ghosts on the 14th of the seventh.  This Buddhist ceremony involved placating the hungry ghosts and other unfortunate spirits—the homeless, the wronged, the drowned, the wandering.  Families and religious groups provided food and goods for them.

The Moon Festival on the 15th day of the eighth lunar month was celebrated with feasts that focused on moon cakes, those round moon-like pastries stuffed with eggs, seeds, sweets, and other fertility symbols.  No one seemed to like them much[7].

Minor details of festivals differ, but much carries over from one to another.  The festivals are great religious structures, constructed not only from visible materials but also from scent, sound, timing, juxtaposition, and flow.  They manipulate and direct emotions, so that they become, at a second remove, huge constructions of emotionality—Durkheim’s intense rituals on a vast scale.  They are assembled from rather simple blocks.

The most basic of these is the paai rite, and the minimal sacrifice, building up from incense to tea to wine to fruit, peanuts, paper goods, and chicken.  In the biggest rites, worshipers offer whole roast pigs, then divide and eat them.  They are covered with a brown sugar glaze that makes them turn red-gold when roasted:  kam jyu yok, “golden pork.”  In all ceremonies, red and gold are stressed, and other brilliant colors complement them.

Silk banners and paper-and-bamboo constructions dominate the festal skyline, and range from small family-level decorations to huge “flower boards” 50 or 60 feet high.

The soundscape is dominated by endless firecrackers, but loud music dominated by gongs and drums drives the lion dances, temple ceremonies, spirit-medium performances, and ghost-expelling rites.  Scent of incense—sometimes whole sandalwood logs burning—defines sacred space and wafts through the whole landscape.  Intense flavors of wine, beer, and festive foods eclipse the bland, crisp flavors of everyday rice and vegetables.  Succulent meat and fat are the foods of a “red time.”

Activities from fighting over p’au to summoning spirits and driving off ghosts are common to most of the bigger festivals of particular gods.  In fact, spirit mediumship seems blind to the particular spirit involved, except for the minor mannerisms of the summoner.  The same questions are asked, the same charms written, the same behaviors invoked.

Each festival has its special traits, and some—such as Ch’ing Ming and “double seven”—are truly distinctive, with little shared above the basic level.  However, in most festivals, a recognizable general flowchart plays itself out with only enough variation to identify the particular san involved.

One festival that sums everything up, and brings everything together, is a world renewal.  Every sixty years, a village must throw a huge festival to renew the cycle—to begin its small world anew.  No such thing happened in Castle Peak, which had no traditional villages dominated by single lineages.   In 1974, Ha Chuen, a small village north of Castle Peak, put on such a fair.  (Actually, Ha Chuen does it every ten years, though I was told this was a special sixty-year rite.)  James Watson was studying Ha Chuen at the time, and gives a wonderful account of this 1974 fair, which I too attended (J. Watson 2004a).  It lasted over a week, after a full year of work to prepare the village and the offerings.  Huge red constructions, visible for a mile around, dominated the skyline.  Lion dances, spirit medium performances, and sacrifice rites went on day after day.  Mounds of pink-dyed buns reached high above the heads of the worshipers.  Huge groups from other villages came to worship Buddhas and Chinese gods in a vast temporary shelter set up at the village gate.  Men cooked huge caldrons of sik pun, a meat stew; cooking the proper ritual stew for such an occasion is a men’s prerogative (J. Watson 2004b).  There was a not-very-friendly rivalry between the old New Territories villages for the honor of putting on the biggest and most expensive world renewal.  The larger village of Kam Tin swore it would outdo Ha Chuen when the time came.



In addition to these relatively standard Chinese festivals and patterns of action, the boat people had a number of specific ideas of their own.

Most important and interesting were the sacred or divine fish (san yu).  These were fish that were large and powerful, and that usually seemed to be more than just fish.  They fit perfectly Mary Douglas’ concept of the anomalous (Anderson 1970; Douglas 1966, 1970).  Douglas found that animals tabooed in religion usually turn out to be those that do not fit comfortably into major, basic categories in the relevant culture’s classification system.  This was true of the sacred fish.  There were several such fish.  One the sturgeon (Acipenser sinensis, locally known as ch’am lung, “sinking dragon”), a rare and weird-looking creature thought to be both a fish and a dragon.  Another group consisted of whales and porpoises (including a mysterious “linked fish”—chun yu—that makes a pilgrimage to the Hung Sing temple up the river near Canton, and that once nodded its head when a girl offered it tea).  The boat people recognize that cetaceans, though they look like fish, behave like mammals, and are like mammals internally (“the meat and organs are like a pig’s,” as some fishermen told me).  Another sacred fish is the sawfish (ku yu, Pristiophorus sp.), which not only has its bizarre tooth-fringed snout but which also was seen as the most powerful and dominant of fish.  Finally, sea turtles were sacred fish, explicitly because “they are both turtles and fish” (in one man’s words), and because they are highly revered in Buddhist and Taoist thought.

A standard explanation for loss of a boat at sea was that the family had caught a san yu and had not treated it properly.

Fishermen tried to avoid catching such creatures.  If they did, by mistake, they immediately released them, or, if the fish had died, they took it to the Tin Hau temple and left it there as an offering.  Ancient, long-desiccated sawfish and sturgeons decorated most local Tin Hau temples.

Some other anomalous water creatures were not san yu but still had uncanny powers.  Both boat and land people shared a fear of the walking catfish or snakehead, known as seng yu (“living fish”) because it could survive out of water (it breathes air with a modified, lung-like swim bladder).  Made into soup, it could cure cancer and other degenerative diseases, but one had to be sure to cook it with pork; if the pork disappeared, the fish was not a snakehead but a “bone-transforming dragon,” and anyone who ate it would disappear, even to the bones.  Another supernaturally curative animal was a small parasite that lived in the gills of large groupers.  When the grouper was caught and died, its qi passed into the parasite, which thus concentrated all the power of the huge fish, thus becoming powerful medicine.  Conversely, the boat people did not use cormorants in fishing, because some thought the rings put around the birds’ necks to prevent them from swallowing fish would also constrict the fishermen’s flows of wealth.

Less visible to the uninitiated, but far more important, were the dragons who controlled rain and storm.  Ordinary beneficial rain was in the claws of ordinary dragons, lung.  Violent storms and typhoons were often blamed on the more savage dock-tailed dragon or dragons (kuat mei lung), whose tails the gods cut off as punishment for bad behavior.

In addition to these, ghosts sometimes took the form of fish, disappearing suddenly.



As we have seen, the boat people saw religion as primarily a matter of providing the gods with feasts and wealth in hopes of making them favorably disposed enough to send exchange goods:  luck, health, birth of sons, good weather, runs of fish.  The idea was exactly and explicitly the same as the plan for dealing with ordinary humans in positions of power.  One tries to create good feelings through doing favors (this is the classic concept of creating ganqing through guanxi, as described in Yan 1996 and Yang 1994), and one tries to use the proper social forms.  Some relations with the other world were much more mechanical and formulaic.  Others were more antagonistic, especially the great exorcistic rites.  Still others were minor and whimsical customs, like the play with lotus rhizomes at the wedding feasts.   Some festivals were deadly serious and almost universally believed to be desperately important; marriages, funerals, and exorcisms came into this category.  Others suffered general disregard and skepticism.

Religion was a social matter, and was invoked for direct and immediate pragmatic reasons.  In spite of a genuine and strong ethical component, popular worship often involved behavior toward the supernaturals that the boat people did not tolerate in the secular world.  The supernaturals were subjected to bribery, hypocrisy, favor-currying, and manipulation.  This was a pragmatic game.  It was judged on its successes.  A god who consistently failed to deliver good fortune, even after proper sacrifice rites, was a god who soon found himself or herself ignored.

On the other hand, “even the gods are subject to fate,” and so occasional failure was expected.  Moreover, in the words of Ge Hong, “Because the size [of this divine administration] is large and its network loose, they may not always respond automatically with precision” (Campany 2002:50).  This view, which seems so wonderfully off-the-wall to a western believer, was echoed almost word for word on the Castle Peak Bay waterfront, more than 1700 years after Ge Hong’s time.  Like human powers, divine powers were far from omnipotent or infallible.  Sometimes they even failed to hear a plea, or were lazy or corrupt, and one had to deal with that.

When all is said and done, it is hard to manipulate fate, hard to change the world.  The popularity of almanacks was due to the determinateness of fate.  A man or woman is born with an allotted life course, and the gods can change it only a bit—less, many said, than one can change it oneself by hard work and honest behavior!

As usual in boat life, any individual was free to manage worship and belief as she saw fit, as part of a detailed, practical, personal way of handling the world.

No one was foolish enough to rely on prayer and worship alone, when human action could affect the outcome.  Fishermen prayed to Tin Hau, but kept their sails and nets in good repair, and followed the government’s weather forecasts assiduously.

Conspicuously absent in most of this were the sense of awe, the mystical or transcendent experience, the ineffable peace and bliss, and the other fabled psychological roots of religion.  Some individuals did have such feelings; most gave no evidence of them.  Feelings were irrelevant to the rituals and ceremonies, which were done in a cut-and-dried, routine manner, though usually with a great deal of drinking, eating, and fun.

However, the great festivals aroused strong emotions.  By and large, ordinary enjoyment was paramount.  Deeper emotions arose from more serious rites.  Being in the presence of a spirit medium in possession trance could not fail to move even the most skeptical and sober viewer.  Funerals and other rituals for the dead aroused intense and unfeigned grief.  Religion served not so much to inspire unusual emotions as to synchronize and socially invoke ordinary, but deep and serious, ones.

















In a brilliant and insightful book, Judith Farquhar (1994) described Chinese medicine as “knowing practice.”  Farquhar worked with practitioners of Traditional Chinese Medicine—current forms of China’s classical elite tradition.  She describes in detail the ways in which doctors invoked a wide and deep knowledge of traditional texts and methods in dealing with particular cases.  The doctors did not mindlessly plug in this knowledge; they considered each case at length, tried approaches they thought would work, and let empirical findings drive their ways of using their skills.  They saw each case as to some extent unique, requiring its own variant of  “practice.”  Traditional medicine was not a set of rigid rules or cookbook recipes, but a vast storehouse of possibilities.

The boat people knew far less of the traditional medical system than did Farquhar’s Chinese medical teachers, but the boat people still knew a great deal of medical lore, and their application of it was “knowing practice” in exactly the same sense (cf. Anderson 1988, 1996; Anderson and Anderson 1975).  They too did the best to apply what they had, with a strong awareness of each case and of the need to use medical knowledge pragmatically in a given setting[7].

The boat people had led hard lives, with little access to medical care.  As a result, they suffered seriously from malnutrition, and even starvation.  On the other hand, in relation to the land people, the boat people got more protein and iron, but much less vitamin C, leading to characteristic patterns of deficiency.  The hard life on the water led to high rates of exposure and accidents.  The boat people had high rates of tuberculosis, and suffered from frequent bouts of measles, diarrheas, leptospirosis, and other contagious diseases.  Within living memory, smallpox and malaria had been terrible scourges, though these were mercifully eliminated by 1965.  Minor respiratory and digestive ailments, and parasite infestations, were constant companions.

The boat people relied for all specialist medical services on shore-dwellers.  They did not have a full or complete medical system of their own.  They rarely went for medical care until they were almost dead.  The shore was a dangerous environment, and medical care there was expensive by boat-people standards.  Thus, the boat people used home remedies when possible.  They tended to deny illness as long as possible, and then “tough it out” until absolutely forced to act.  This led to much chronic illness.

The boat people’s explanatory model—to use Arthur Kleinman’s (1980) term—held that, ultimately, all illness was due to fate.  One suffered according to one’s destiny.  However, this highly abstract concept broke down in practice into two possible types of direct causation: personalistic and naturalistic (sensu Foster 1998).  By far the more important of these was naturalistic.

It makes sense to dispose of personalistic afflictions first, since they were less important.  If an illness could not be successfully treated or alleviated by ordinary medicne, supernatural agency seemed likely.  This could be confirmed by going to a spirit medium, who would enter trance and diagnose the condition.  In supernaturally caused illnesses, control was outside the power of ordinary mortals, but could be gained through ceremonies.  In witchcraft-caused illness, magic could be invoked to counter the curse.

Personalistic causation could be categorized according to the agent inflicting the illness.  Gods (san) and ancestors could punish people for sin, usually the sin of neglecting proper sacrifices to the god in question.  In particular, if one caught a sacred fish and did not offer it up immediately to Tin Hau, one would suffer disaster.  Ghosts (kuai) sent sickness or other affliction to those who had wronged them.

Problems ranged from a simple cold, which might necessitate a temple prayer, to genuine crises—persistent attacks by ghosts, or real outrage of a neglected or inadvertently wronged god—that required full-scale services by Taoist priests.  These involved chanting, instrumental music, sacrifices of animals, vast amounts of incense, and much more, and cost thousands of dollars.

Normally, people assumed illnesses were natural.  Even mental illness was so considered if it was either inborn (in which case it was simply one’s fate) or explained by reference to obvious causes, as in the case of major depression following the death of a loved one.  More often, minor mental illness followed from fear, which could disrupt one’s hei or drive a part of the soul out of the body.  The boat people explained shock the same way.   Chronic neglect or abuse of children could make them insane.  Mentally abnormal people were often shunned, but sympathetic people thought (I think correctly) that this at least sometimes led to more serious mental illness.

Accidents—from getting wrecked in a typhoon to tangling with a shark—were usually just bad luck.  However, they were sometimes sent by an offended god, or caused by bad fengshui.  Clearly physical conditions, like broken bones and sprains, were usually natural accidents in everyone’s view, and were treated by physical methods.  Doctors set bones.  Sprained ankles rested.  Excellent massage was available for sore muscles.

The vast majority of illnesses were purely natural, and treated naturalistically.  In Cantonese medicine, ordinary illnesses were imbalances of heat, cold, wetness and dryness—the classic four qualities of Hippocratic-Galenic medicine.   The Hippocratic-Galenic tradition reached China early (perhaps in the 4th or 5th centuries AD, certanly by the 14th, when texts mention Galen by name).  It fused with China’s indigenous yang/yin theory, but yang and yin were rarely (if ever) invoked by the boat people in talking about illness.  Wetness was invoked rarely, and dryness even more rarely.  As in most of the world (Foster 1993), these two qualities had sunk to a minor role.

Other, more strictly Chinese, factors filled out the Hippocratic picture.  Sheer weakness—depletion of energy or hei—was a common, serious problem.  This came from overwork, exposure, and serious illness, but most of all it came from childbirth (see below).  Blocked, disarranged, disordered, or stagnant hei was vaguely recognized as problematic, but consideration of such subtleties was left to professional Chinese doctors, who were deeply concerned with such issues.  In addition, actual pathological forces were also mentioned:  Poison, dirt, wind, fright, contagion, and occasionally other qualities (Anderson and Anderson 1975; Martin 1978).  “Poison” (tuk/du) meant two or three different things, carefully distinguished by knowledgeable consultants.  First, there was actual poison, as in puffer-fish liver and sea snake venom.  Second, there was a certain mysterious quality in male poultry (and perhaps some other foods, but no one was sure) that could exacerbate cancer and similar illnesses; this was referred to as “poison,” though strictly speaking it merely potentiated or activated preexisting conditions.  Yet another poison was the womb-poison introduced by the pregnancy and birth process; it came out in measles, viewed as inevitable and necessary (Topley 1970, 1974, 1975).  This latter belief led to some opposition to measles vaccination (Topley 1970, 1975).  Wind—one speaks here not of ordinary wind, but of a wind that strikes inside the body—was a problem in itself and because it introduced cold, or cold and wet; rheumatism was due to such drafts.  Dirtiness inside the body was produced by various stresses and illnesses, and had to be cleaned away by herbal teas.

Normally, though, heating and cooling hei explained any ordinary illness.  Actual hot or cold weather, especially too much sun or too much cold and wet air, was a risk factor, but more often the problem was heating and cooling foods.  I have dealt with this system extensively elsewhere (Anderson 1987, 1988, 1996) and need not go into it here.  Suffice it to say that foods were heating (it or yit, Putonghua re) if they were high-calorie (thus literally high in heat energy); if they had been exposed to high heat in preparation; if they felt hot, like chile and ginger; or if they were hot colors—red or orange.  Foods were cooling (leung/liang—foods were very rarely hon/han, “cold”) if they were low-calorie, if they had been cooked at low heat in a lot of water (water is cooling), if they were cold-looking (icy white or pale green).  The perfectly balanced food, exactly in the middle, was, of course, cooked rice (faan); throughout the world, Hippocratic-Galenic traditions brand as “neutral” the major staple.  Other foods similar to it—medium calorie content, dead-white color—were also neutral:  noodles, white-fleshed fish, potatoes, and so forth.  Sugar was heating if brown (“red”), slightly heating if ordinary granulated white sugar, and cooling if “ice” sugar (rock sugar, which does look like ice).  Alcohol was normally heating, because it gave an instant feeling of warmth on cold days, but beer was considered cooling. Not only did it cool the body; it also was seen to be the only thing that the British consumed in enough quantity to offset their heating diet of bread and meat.  The boat people jokingly called it “Westerners’ cooling tea.”

Excess of heat caused illnesses that resembled burns in some way, such as reddening, rashes, dry skin, sores, sore throat, irritation, fever.  Also, any drying or constriction, such as constipation, was a hot condition.  This meant that “hot” diseases included what we in English call “colds.”  (English doctors once used the same Hippocratic tradition, but attributed the origin to cold draughts, and focused on the flow of mucus, rather than on the fever and sore throat.)  More focally “hot” was scurvy, the “hot” condition par excellence.  Vitamin C deficiency leads to rashes, pain, sores, dry flaky skin, constipation, and other burn-like conditions.  Another classic “hot” condition was the aftereffects of a big feast: hangover, bloating, feeling of heat, and so on.  In all these cases, it was observed that eating plain boiled vegetables and drinking herbal teas was curative.  This validated the system.

Excess of coolness led to lowered body temperature, pallor, weakness, and other signs resembling the effects of hypothermia.  The classic “cool” ailment was anemia.  In this case, eating gently warming foods—basically, lean meat, organ meats, red beans, ginger, Chinese wine, black vinegar, and a few other standards—was curative.  Again the system was validated; the iron in the meats, perhaps made more available by the wine and vinegar, cured the anemia.

Venereal diseases were considered hot and wet (sap/xi), and thus hot-wet foods like shrimp and marine catfish were avoided.  No diseases were due to dryness (kon/gan), but eating drying foods like dry-roasted peanuts led to a dry, scratchy throat.

Thus, therapy for ordinary mild illness was food.  Minor hot ailments were cured by eating cooling foods, and vice versa.  More serious or chronic cases required herbal teas.  Cleaning or purifying the body, eliminating poison, and the like required medicinal herbs, often cooked along with special soothing foods like Job’s-tears and barley.

Especially important was building strength after major crises.  Most typical and important of these was childbirth. Given the chronic poverty of the boat people, pregnant women rarely got enough good food, and childbirth left them terribly weak and malnourished; anemia in particular was essentially universal.  They thus were careful to “do the month,” as the Chinese say (Pillsbury 1976): spend at least a month (and often 40 days) lying in bed, keeping warm, and eating strengthening foods.  These were foods that were warming, but also strenghtening and repairing; pou (bu) was the term used.  Commonest of these were pork, pig’s liver, pig’s feet, chicken, and similar foods very high in protein and iron. These were usually cooked with ginger and Chinese “wine,” which are warming, and often with black vinegar, which was believed to be strengthening and nourishing.  Also important was a small red berry known as kau kei ji (koujizi, the fruit of Lycium chinense).  This fruit is literally a multivitamin pill; it is so rich in all the vitamins and minerals that a handful of the berries is roughly equivalent to a modern multivitamin supplement.  Herbal medicines such as astragalus and angelica could be added by the affluent.

An important point, frequently made, was that diet therapy had to be tailored to the individual.  People differ in their response to food.  Modern bioscience explains this as being due to allergies, personality, personal history, and metabolic or biochemical differences due to genetics or early experience.  The boat people had no such explanations, but they realized that people were different.  Often, a food that was “cool” for most people was experienced as “heating” by some; this was usually due to allergies, since allergy reactions are often classic “heating” symptoms such as rash, stuffed sinuses, and itchiness (see Anderson 1987 and references therein).  Individual experience over time, and immediate environmental context, both affected nutritional therapy.  Any caregiver would take these factors into account.

The eldest woman in the household was normally the chief caregiver; others acted under her orders.  She would often defer to, or discuss with, her daughter or daughter-in-law when these women were treating their own children, but, as grandmother, the eldest woman still was the final arbiter.  In more serious cases, she was expected to consult with the whole family—not just her household, but the extended family, often scattered over a number of boats or other residential units.  The only exception to this rule occurred when someone else in the family had special medical training or experience.   Sometimes the senior male would intrude.  He might be listened to, but just as often he was ignored, or even told to stay out of the scene.  Theoretically the leader of the family, a senior male had almost no authority in matters such as household health care.  Sometimes, women “saved the man’s face” by telling him that he need not consider himself with such lowly household matters.

When household knowledge and skills were inadequate, the second recourse was local herb-sellers.  These ranged from grocery store owners and sidewalk peddlers to trained traditional pharmacists and herbal doctors, working with Li Shijen’s Bencao Gangmu and related or derivative texts.  High-quality herbs and the services of formally trained pharmacists lay outside the economic range of most boat people.

Western medicine was a common recourse, becoming rapidly more available, more successful, and more important.  In 1965, western medicine at Castle Peak Bay was represented by a few practitioners who gave shots (sometimes mere normal saline, given for effect), antibiotic pills (usually chloromycetin, because it was the cheapest), aspirin, and sometimes worm medicine.  More elaborate Western care required travel outside the immediate waterfront area.  By 1966 this condition was changing, and by 1975 there was quite good medical care available from a number of local hospitals and clinics.  Most of these were free or reasonably priced.  This had led the boat people to use them when food and simple home herbal remedies failed.  Traditional Chinese medicine was neglected, though by no means gone.  There was little knowledge of acupuncture and other specialized systems.  I never heard of acupuncture being practiced.  Moxibustion was practiced fairly often, however, most often using arm and head points.



Individuals and family units took control of their lives.  The family unit was the normal locus of control.  People did not meekly place oneself under care of a healer.  Instead, they immediately took charge of the situation, and did everything they knew to alleviate it.

When I worked to explain Hippocratic medicine in the 1980s (Anderson 1987), I was somewhat aware of the importance of this aggressive stance in the success of the tradition, but I was more impressed by the other virtues of the system.  Most obviously, humoral medicine actually worked at Castle Peak Bay.  It made people eat vegetables when they were short of vitamin C, eat liver and meat when short of iron, drink liquids when sick, and take countless other perfectly sound empirical steps.

On the other hand, it was obviously very far from perfect, so a full explanation had to take other matters into account.  Further research in Mexico has shown me that one of the most solid pillars in the foundation of Hippocratic medicine and its success is its no-nonsense, take-charge attitude.  Of all the medical traditions in the world, it is the one that most clearly directs the patient and her household to act, now.  It works primarily through regimen:  modification of ordinary life activities—eating, drinking, exercising, resting.  It works through common sense, supplemented with a framework that is simple, straightforward, and easily learned.  Even the most humble and least educated can modify diet according to heating or cooling properties inferred in the foods.  Anyone, anywhere, can modify activity.  One does not give up control to an alien, impersonal medical establishment.

Humans need to feel in control of their situation; this is an actual physical need, and can be a life-or-death matter in cases of illness (see Anderson 1996).  The value of a take-control medical belief system is evident.  The boat people were close to the bottom of traditional China’s vast social hierarchy.   They needed to assert control whenever they could, and they did so, in no uncertain terms.

Their explanatory models and medical recourse systems were designed to maximize this.  As they were forced by experience to give up control, they moved down a predictable scale: from home care to herbalists to Chinese doctors to ritualists to full-scale rituals.  This was also a hierarchy of ease of understanding.  Nonexperts did not know the herbal formulas, but they understood herbal medicine as merely a more sophisticated form of food therapy.  There was no line separating foods (sik pan, Putonghua shi pin) from medicines.  All foods had health-relevant humoral properties.  Some foods were particularly medicinal, such as ginger.  Some foods were almost solely medicinal, such as gaugeiji, black vinegar, and foxnuts (Euryale ferox).  These merged imperceptibly into items that were considered purely medicines, such as astragalus and ginseng.  It was impossible to tell whether snow fungi and birds’ nests were “medicines” or “foods.”  The medical formulations used on the waterfront routinely mixed such medicinal foods with items considered solely medicines.

Elite medicine and religious ritual were progressively more difficult to understand and less easy to control.  They drew on explanatory models that were poorly known to the boat people, and that minimized the chance of control by laypersons.  The complex logic behind traditional Chinese medicine, involving links to cosmology, vessel theory, qi theory, yin-yang and five-phases philosophy, and much else, was beyond them.  (It still defies full analysis by scholars.)  The logic of ritual, drawing as it did on Durkheimian collective representations, was more easy to understand, but supernaturals were notoriously difficult to handle.

By 1975, Western medicine had largely replaced the latter three steps, but only in so far as it demonstrably worked better.  Its logic of germs, contagion, immunization, antibiotics, and surgery was beginning to penetrate the local medical system, but was still very poorly understood.  To local people, “shots” were more a magical practice than a scientific measure.

All students of ordinary popular Chinese medicine have found rather similar hierarchies of recourse, no matter where they were working (see Kleinman 1980).  The boat people were probably even more committed to home treatment, and even less prone to put themselves under care of a physician, than most Chinese described in medical studies.  This was partly because of poverty, partly because they wanted to retain full control.

Practice, in Farquhar’s sense, is an assertion of control, and is “knowing” because the practitioners realize that more knowledge allows more control.  Explanatory models are as logical and commonsensical as possible; only when the simpler ones fail does one proceed to the more arcane.

Thus the boat people and their world, seen in wider perspective, point up this issue of control.   The ill desperately need to have a sense of control of their situation; they need to understand it and to assert some agency within it (Langer 1983).  Ellen Langer showed that having control over one’s environment actually improves health.  For people like those of Castle Peak Bay, control is a life-and-death matter (cf. Anderson 1996).  They were at the bottom of the social scale, and needed all the sense of efficacy (Bandura 1986) that they could get.  They met the challenge; as in other aspects of their lives, they took charge assertively.  When they could do nothing else—as was all too often the case—they could at least make the patient feel that family, neighbors, and community cared and were acting to help.



In 1974-75 we carried out a focused study of cancer among the boat people.  Cancer is a hard disease to conceptualize, in any culture.  The real problem is that it is not one disease.  It is a congeries of conditions that have different causes, different courses, and different effects.  The only thing they have in common is that they take the form of uncontrolled multiplication by cells that have lost the natural restraints, checks, and balances that keep the body functioning smoothly.  This loss of restraint can be due to tobacco smoke, hard radiation, viral infection, random mutation, toxic or mutagenic chemicals, and many other environmental stressors.  Often, a given stressor causes cancer in a given site or set of sites, while leaving all others undamaged.

This conceptualization of cancer is of very recent origin.  Until cell physiology and mutation were well understood, and epidemiology had discovered many chemical stressors, no one could conceive of cancer that way.  The modern definition is a child of the cellular-biology revolution of the mid-20th century[7].

Traditional China had no concept of “cancer” equivalent to or similar to the modern English concept.  Visible cancerous growths were called by various terms simply meaning “lump” or “swelling.”  I have no idea what people would have made of things like leukemia or lymphoma or pancreatic cancer; they had no way of knowing what kind of pathology these involved.

We cooperated with Dr. John Ho in a small project on the epidemiology of cancer of the nasopharynx (Anderson, Anderson and Ho 1978).  This cancer was 100 times as common among Hong Kong boat people as among the world population at large.  Cantonese in general had a high rate, but Cantonese raised in the United States had the same low rate as Americans and world citizens in general.  Dr. Ho directed it; Cecilia Choi worked with us.  We were largely working to check out a theory of Dr. Ho, to the effect that nasopharyngeal cancer causation involved nitrosamines created by decay processes in poorly salted fish.  Nitrosamines were suspect of causing cancer when inhaled.   Dr. Ho theorized that antioxidant vitamins in fruits and vegetables would destroy these chemicals.  The boat people were typically very short on such vitamins.

This hypothesis checked out, but—to make a long story short—turned out to be part of a longer causal chain.  Many boat people were more genetically susceptible to nasopharyngeal cancer than most humans are; an allomorph of the HLA-2 antibody gene caused this.  Males were more susceptible than females, in Hong Kong.  Smoking was sometimes involved, but not with our population, few of whom smoked, and none of whom smoked much.  Infection with the Epstein-Barr virus was a major risk factor, but in those days in Hong Kong everyone, or virtually everyone, got that virus early in life, so it was not a very meaningful variable.  Living on salt fish and rice, without many vegetables or fruits, was indeed typical of all our cases except one, and that one individual had had a malabsorption syndrome that prevented her from absorbing nutrients well.  We also observed that the “cancer personality,” retiring and shy and depressive, was a major risk factor, but were maddeningly unable to get enough control data to prove it.

We found that cancer was considered to be a condition somehow involving poison (du).  Cancer sufferers avoided poultry unless they saw it killed and cut up, because, according to Cantonese folk medicine, the hei of male poultry somehow potentiates poisons in the system, making them worse.  Eating a rooster or drake would make the cancer flare up and probably become fatal.

Unfortunately, our interviewees were young and getting westernized, and if they had more knowledge about cancer than that, it was international biomedicine that they knew.  Very likely cancer would prove to have a folk etiology involving hot and cold—as it does among the Maya, who believe that a sudden cold shock to an overheated body can cause it.  (Maya beliefs about hot and cold are strikingly parallel to Chinese; both had a pre-existing medical belief system involving hot and cold that was then impacted by Galenic humoral theories.)   Certainly, disturbed or contaminated qi was involved in cancer.  I think that in earlier times a growing and pathological swelling would be considered the result of blocked or stagnating qi.

In both traditional Chinese and traditional western medicine, then, cancer was anomalous and hard to classify, since no one could see or understand what was really going on.   It did not fit well in either paradigm.  Western medicine developed from a number of practical fields, notably military medicine and sports medicine.  So it had a lot to say about bodily integrity, injuries, and surgery, to say nothing of devils attacking people.  This led to the famous (or notorious) military language that is still with us:  the body “defends” itself against “invading” germs that are “the enemy” of good health, etc.

In addition, early western medicine was much like Chinese in that it stressed the balance of heat, cold, and other qualities or influences.  From ideas of heat, cold, wetness, dryness, and the necessity of balancing these in the body, Galen and his intellectual relatives constructed humoral theory.  The Chinese fused it with earlier yin-yang ideas and with Indian medical theories brought along with Buddhism in the early Common Era.

Contagion became more and more obvious to people in both civilizations, over the centuries, and eventually led to theories of insect transmission; the Chinese began to understand the role of mosquitoes and other pests in premodern times (Elvin 2004), and, eventually, so did westerners. But cancer was not a wound or injury, was not contagious, was not related to insects, was not obviously related to diet, and did not fit any humoral paradigm.  No one could figure out what to do with it, or even how exactly to characterize it, until modern times.



All the facts in the history of medicine argue for a very specific view of how people develop medical theories.  First, people are good observers; they recognize symptoms and syndromes, and can see the more obvious causation links.   For example, lack of green vegetables produced “hot disease,” what we now call vitamin C deficiency.  Second, people are scared by diseases, and the less they understand these the more scared they are.  Third, this puts people under great pressure to come up with explanatory theories—“explanatory models” in Kleinman’s terms.  Fourth, they look for the most plausible, reasonable explanation possible.  In China, that was often a nutritional one, since the links between poor nutrition and poor health were so tragically obvious to all.  Fifth, failing a good sensible theory, metaphor would always do.  Warfare was the obvious source of these for the military medicos of the west.  In the east, flows of breath and energy in nature and the body became the metaphor field.  Chinese medicine could pile metaphors, layer on layer:  wind moves the air, breath is vital to people, so the earth too much have veins of hei energy running through it, the whole human body also has hei channels, hei can get blocked and stagnant just like air, and then make you sick just as stagnant foul air does….

Once a plausible explanatory model is in place, only overwhelming counterevidence combined with a better model can dislodge it (Kuhn 1962).  Thus, a long history of observations that did not fit dominant theories could be dismissed in both east and west, until the mid-19th century brought international biomedicine into sudden meteoric rise.  Then biomedical paradigms in turn became overextended; infection by “germs” seemed to answer all questions, until lifestyle diseases became overwhelmingly obvious after 1950.

The countertheory to my common-sense theory of medical explanations is shared by most positivists and most postmodernists:  people just pick any arbitrary idea that enters their heads.  The positivists say that this is true of all medical notions except international biomedicine; the postmodernists say it is true even of that.  This countertheory is destroyed by even the slightest acquaintance with traditional medical literature, in context, especially if one works in a society still using those traditional  ideas.  One soon learns that such people have a logical system, and that it is based on shrewd and correct observation.  The problem is that the observation is limited by lack of modern laboratory equipment. In the absence of the ability to observe bacteria and the like, one must infer spirits or winds or other invisible agents.  Once that is done, plausible logical extension does the rest—usually bringing people farther and farther from what we would now think is the “real” story.

Western views of Chinese medicine have been diverse, to say the least.  (A particularly thoughtful, insightful review of the more moderate views is found in Sivin 2000).  Many westerners have simply dismissed Chinese medicine, and indeed all Chinese thought, as a fantastic farrago of nonsense and superstition (most recently, Wolpert 1993).  Others have adulated it, seeing it as the key to the future and the great hope of the world.  Various forms or derivatives of Chinese medical traditions are extremely popular throughout the world today, comprising a large proportion of the “alternative” medicine of the United States and other countries (Bowen 1992).  Such extreme views are even held by reputable scholars, though rarely if ever by scholars of Chinese medicine[7].

The latter have, however, taken strong positions of their own.  Some have even doubted if there is such a thing as “Chinese medicine.”  At least in private debate (I find no published source that is this extreme), some have seen “medicine” as a word covering a strictly western concept, so different from yao that there can be no meeting ground.  The implication is that the word “medicine” should properly refer only to international biomedical science, which is, indeed, different from Chinese yao.  However, such restriction clearly flies in the face of ordinary English usage.  “Chinese medicine,” under that name, is widely known in the west now, along with faith-healing, naturopathy, Native American herbalism, and countless other forms of “alternative medicine.”  And restricting the term “medicine” to contemporary bioscience would force us to rename the Greek and Roman traditions that were the original “medicine”!  “Medicine” and “yao” both refer to a wide spectrum of treatment modalities, and are close enough in denotation and connotation to be good translations for each other—not perfect, but neither is any other translation-gloss.

Others have argued that the various Chinese medical traditions are so different from one another that one must speak of Chinese medicines.  As we have seen, demon-quelling, witchcraft, orthodox religious ceremonies, herbalism, acupuncture, regimen, and other modalities are all part of the healing arts in China.  This has been the case for at least 2500 years (Uschuld 1986), and presumably was true long before that.  One can only agree that there are  many different traditions in Chinese practice.  However, on the other hand, they are often combined by one individual in treating self or others.  Moreover, once again we must face English usage:  “medicine” in standard English means all healing traditions, taken collectively.

Still other scholars follow, to varying degrees, the modern practice (especially typical of China’s government today) of defining a Traditional Chinese Medicine—often capitalized and acronymized “TCM”—to separate it from the traditional therapies not included in the concept.  This modern restriction of Chinese practice consists of acupuncture, moxibustion, and much of the herbal therapy, and the formal or elite theories behind these.  It relies heavily on certain classic texts, such as the Yellow Emperor’s Classic (Unschuld 2003) and the Shang Han Lun.  It excludes folk practices and religious or magical medicine.  These latter were widely dismissed in 20th century Chinese sources as superstition (Putonghua mixin—the word is a recent coinage, probably originally Japanese).

It is possibly useful to define such an entity as “TCM,” but it is a modern concept.  Pre-20th-century Chinese were self-conscious to varying degrees about the differences of elite and folk traditions and of the differences between naturalistic and religious treatments, but did not separate them strongly at any period (see e.g. Harper 1998, Hinrichs 1998; cf. Sivin 2000; Unschuld 1986, 2003).  They certainly did not have a widely-shared, tightly-defined subset of medicine corresponding to contemporary TCM.  The Castle Peak Bay waterfront was genuinely traditional—not Traditional!—in this regard; no one except the few elite doctors separated TCM as a practice.

The Castle Peak Bay waterfront may well be reasonably typical of ordinary Chinese society for the last 2000 years or so, in regard to the broadest outlines of a sense of “medicine.”   As we have seen, they followed a hierarchy of recourse.  All the treatment modalities were seen as ways to deal with, and hopefully cure, illness.  The word yao usually covered only the fraction of this world that was specifically medical in the western sense:  herbalism, acupuncture, and the like.  Regimen was too much a part of ordinary life to be marked as yao.  Religious healing came under categories such as paai san (“worshiping supernaturals”) and the like.  Healing practice was defined by verb-noun constructions: “treating illness” and the like.

However, all these practices graded into each other.  We have seen how food graded into herbal medicine.  Herbal medicine graded into religious medicine too, because the religious officiants often provided herbal prescriptions.  Even the gods did this; spirit mediums routinely wrote herbal prescriptions while in possession-trance.  Usually, the possessing god, acting through the medium, would write a charm to be burned and the ashes drunk in herbal tea; prescriptions for the tea were a part of the process.  Healing by spirit mediums may actually have been more effective than most of the more naturalistic Chinese medicine.  All were about equally dubious in biomedical terms.  The spirit mediums’ spectacular performances evoked strong placebo effects—especially since the illnesses they treated were often neurotic, depressive, or hysterical conditions, eminently susceptible to the power of suggestion (at least in the short run).  Naturalistic medicine was less psychologically compelling; moreover, the herbal remedies often had side effects.  (The myth that herbs do not have side effects is one of the stranger bits of western medical folklore.)

Significantly, though, herbal therapy was the great mediator—the common ground, the center around which other forms clustered, and the modality that overlapped most widely with other modalities.  Western observers often think that acupuncture or hei management or some other system is the core of Chinese medicine.  Even Chinese observers may dismiss herbal practice as mere empiricism of no interest to students of theory (see e.g. Hsu 2001a:8).  Yet, throughout history (e.g. in the Mawangdui manuscripts; Harper 1998), herbalism has occupied the central place in practice that it occupies today.

From the point of view of an anthropologist talking about ordinary people trying to keep healthy, “medicine” in the singular is the more sensible usage.  When a mother is trying to save her child, or a son is caring for his aging father, they think systematically and comprehensively.  They consider all possible recourses, and decide on the basis of expected benefits and costs (see Young and Garro 1994 for a classic account of medical decision-making in a traditional society.)  They plan, in short, as if all the medicines were one.


Another tension among western observers of Chinese medicine has been over whether to evaluate Chinese medicine in terms of its actual curative value as tested under modern conditions, or to study it as a purely intellectual system.  The late Joseph Needham agreed with almost all modern Chinese medical researchers in preferring the former view (Needham and Lu 2000; Sivin 2000).  These parties differed in the nuances:  Needham saw the history of Chinese medicine as part of the history of world medical science, and tended to evaluate it in terms of its contribution to modern international biomedicine, whereas modern Chinese researchers tend to view it as a separate tradition, but one that is valuable in and of itself and also as a source of specific drugs and techniques for the world.

Excessive eagerness to mesh eastern and western medicine has led in the past to embarrassing mistakes.  Hsu (2001b) records two confidently-rendered and completely contradictory diagnoses (one by Needham) of an illness recorded in an early case study.  The case study is fragmentary, and both diagnoses clearly go far beyond the evidence.  This sort of guesswork has discredited the biological approach in many circles.

Thus, the vast majority of contemporary historians of medicine prefer to consider Chinese medicine purely as an intellectual creation (Hinrichs 1998; Sivin 2000).  It is a thing to study, not a thing to validate.  It is a tradition to understand as one understands a religion—not a part of world science.

This view is especially popular with those (such as Hinrichs) who study the religious side of medicine.  By definition, supernatural healing cannot be tested and verified, and cannot be explained in this-world terms.  If one believes it is a deeply basic part of Chinese medicine, as Hinrichs does, then one tends to write off the whole system.  It is seen as an intellectual curiosity, and explained purely by recourse to bookish traditions, cosmologic parallels, religious belief, or arbitrary cultural construction.  It is not seen as having any relationship with reality.

Studying boat people’s medicine allows one to take a more comprehensive view, in which both of the above approaches are useful but neither is sufficient.

Above, I have described the system in its own terms.  In general, any discussion of a medical system must begin with such a description, if one is to understand it and properly evaluate parts of it.

On the other hand, Chinese medicine had an appreciable success rate.  Many of the drugs are now internationally used in biomedicine.  Artemisin, from the herb qinghaosu (Artemisia spp.), is now the drug of choice for malaria.  Chinese massage is widely used as well.  Acupuncture shows promise in some areas, though it remains controversial.  Chinese medicine has indeed been influencing modern practice, and has influenced western medicine for a long time (Needham and Lu 2000; many of Needham’s claims have been discounted, but a large residue remains).  Chinese medicine will remain a source of practical wisdom for a long time to come.

There is, moreover, a third way to look at Chinese medicine.  It is the “synthesis” to the above thesis and antithesis.  It involves explaining the actual theories and ideas of Chinese medicine.  This involves, first, understanding the system on its own terms.  But, then, one must understand which of its components actually work—in the sense of producing effective treatment over and above the placebo effect.  Many herbs and techniques do not work well enough to compete with modern biomedical drugs, but do work well enough to have been the best things available in earlier centuries.  The most striking example is chaulmoogra oil, which for centuries was not only the best Chinese drug available for leprosy, but was the best in the world, and the drug of choice in west and east; only recently has biomedical research developed a better drug.  Some other drugs work for a few specific conditions but were vastly oversold; “dragon bones” (fossil bones) were valuable sources of calcium, but they did not have the panacea values ascribed to them in some Chinese herbals.

Once one has some sense of what actual physical effects Chinese medicine has or may have, one can go farther toward explaining it.  I have argued above, and shown at greater length elsewhere (Anderson 1987, 1996), that the heating/cooling (yin/yang) theory of foods succeeded because it was genuinely valuable in practice.  It was not scientifically correct as a theory, but it led to actual practice that did work; this validated the system and kept it alive.

The same sort of analysis, applied to vessel, correspondence, and hei theories, would surely make them more comprehensible.  These theories are “wrong” in international biomedical terms, but just as certainly they were “confirmed” in some sense by the practice of thousands of Chinese doctors, hundreds of whom recorded their findings.  To be sure, these records reflect the selective bias inevitable in true believers who do not use strict case/control experimental methods.  But there must have been a great deal of correct, well-observed medical observation that seemed to confirm the classic theories.  Needham’s attempts to explain them in these terms were premature; he knew neither the details of the Chinese theories nor the biomedical systems to which they might correspond.  Today we might succeed.

Even religious practice has some value.  Not only does it provide ideal conditions for the placebo effect; it also, in many cases, involves long, intimate, personal conversations or consultations between patient and religious healer.  These are comparable to the “talking cure” in western psychotherapy, and seem effective as such.  (My experience in this is confirmed by more trained observers:  Bowen 1992; Arthur Kleinman, personal communication over many years.  Kleinman, a highly expert psychotherapist, was struck by this in his field research.)  Many of the chronic conditions brought to religious healers are, in fact, psychological problems, ranging (at Castle Peak Bay) from “blanking” on exams to major depression over loss of loved ones.  (See also Kleinman 1980, 1986.)  The “talking cure,” as delivered by a sensitive and empathetic spirit medium, is convincing to the patient, and presumably beneficial.22222222223

As Thomas Kuhn (1962) pointed out, a scientific theory is usually underdetermined by facts.  Traditional medicine, in any culture, tends to be an extreme case of this.  People are desperate for cures, and they are desperate to understand illness.  Theorists are thus under enormous pressure to come up with the best possible systematization of observations.  On the other hand, there are facts underlying all theories, and the facts guide thinking even if they do not wholly determine it.

In a culture without organic chemistry laboratories, humoral theories of foods—notably including the Chinese humoral theory—are probably the very best combination of simplicity and utility that the human brain can devise.  Similarly, in a culture without neuropsychology labs but with a widely-believed religious system, the success of psychotherapeutic counseling can best be explained supernaturally.

Many matters are so obvious that they require no leaps of explanatory imagination.  Broken bones are hard to miss, nor is the method of dealing with them.  Arrowhead wounds and dog bites are of obvious causation and straightforward treatment.

On the other hand, infectious disease defies all attempts at explanation unless one has not only microscopes and laboratories but a few hundred years of scientific observation using same.  Leeuwenhoek first saw bacteria, but not until Koch and Pasteur did anyone seriously maintain that they were the agents of infectious disease, and not until the 20th century did further investigation turn up viruses and other tiny agents.

It is to the credit of the Chinese that they did not simply write off infectious disease as supernatural, but tried their best to come up with credible explanations for it.  If those explanations froze early and became fossilized in the literature—and they did—this was not so much a matter of “Chinese conservatism” as of the well-known fact that only a better theory can replace a widely-held theory (Kuhn 1962).  No better theory came forth until the 19th century.  So China continued to reprint the Yellow Emperor’s Classic.

This leaves us with countless questions.  Why, for instance, did Chinese theories so completely silence anatomy?  China developed a science of anatomy and forensic medicine (see e.g. McKnight 1981), but orthodox doctors took little account of the findings.  The standard explanations are either “Chinese conservatism”—a myth summarily disposed of by Hsu and others (Hsu 2001a, 2001b)—or the more reasonable point that a dynamic flow theory needs no static structural base in anatomy.  There is probably more to be said.  Why, also, did Middle Eastern medicine have rather little effect on China?  It was imported in literally encyclopedic proportions under the Mongols (Kong 1996—a modern edition of a Mongol Empire work).  Western Asian veterinary medicine came at about the same time, and almost totally replaced China’s indigenous veterinary tradition.  Why did medicine for humans not have a similarly transformative effect?  Part of the answer is that some of it was already there; another part is that medicine for humans was presumably not as deficient—in the eyes of Yuan and Ming practitioners—as veterinary practice.

The greatest problem facing our analysis is that we know too little of “practice” in past times.  We cannot observe doctors at work.  However, recent historical research often focuses (to varying degrees) on reconstructing practice, rather than solely on analyzing texts (see e.g. Furth 1999, as well as the above-cited works).  Another promising source that has only begun to be studied is the body of literature that China produced over time.  Novels such as the Hung Lou Meng record vast amounts of realistic description of actual medical practice.








[8]  From what I have been able to glean from limited personal observation and from only slightly fuller observations by Ho-fung Hung and others, the boat people of China have indeed maintained a running low-key resistance to Communist collectivism and dragooning, and they have, in the end, been successful.  Many, probably most, have followed those of Hong Kong onto shore and merged with the general population, but some still live quite independent lives on the rivers and seas.


[9]  One exceptionally brilliant and experienced ethnographer of China has frequently commented to me that she cannot deeply value any of the art, literature or music the Chinese created, because the society was so cruelly abusive to women.  This seems to me a sad self-limitation.  Yes, the evils were real, but so were the achievements.  Many of the writers and artists were women, after all.  Many others were champions of women, treating them as equals or reasonably close; one thinks of Mei Yaochen and Yuan Mei.  I fear that if one condemns a society’s arts, or anything else, because of that society’s worst features, no one will ever be able to enjoy anything from any society.


Leave a Reply