Chapter 3
The Demography of Family and Class

The Data

The written works of the T'ai-ho literati are full of short pieces that fall generally under the category of obituary or necrology—that is, commemorative accounts of the lives of dead friends, relatives, and neighbors and their wives. I have found about five hundred obituaries written on behalf of T'ai-ho people. They come in several genres, most of them related to the funerary cult. Most common was the mu-chih ming , a life account chiseled in stone and placed underground near the grave of the deceased. A bit less common was the mu-piao , a life account, also engraved on stone, that was exhibited in the open air so that people could read it. (Many people were honored with both forms of inscription.) There are also available a number of hsing-chuang , or "draft necrologies," usually very detailed, done by family members or other intimates and given to writers of tomb inscriptions to use as source material. There are several minor genres, like eulogies (ai-tz'u ), appreciations (tsan ), and biographical accounts (chuan ). Many of these accounts state just how the necrology came to be written.^[1] Copies of all these documents were kept by their authors and were later included in their collected literary works (wen-chi ), which makes it possible to read them now. By this means, even a mu-chih ming , interred as stone, became quasi-public in its paper copy and thus available to interested readers.

Chinese necrologies are fairly rigid in their form, and as a rule they provide (1) a list of immediate ancestors; (2) a list of a man's wives and

concubines or the name of a woman's husband; (3) a list of children and grandchildren, usually by name; and (4) date of death and an age reckoning in sui , although many accounts give birth and death dates to the exact day.^[2]

All these sources constitute very interesting data with which to explore social questions—but how good are the data? They are surely useful, although there are errors and pitfalls to watch out for. Some errors and omissions can be corrected easily if, for example, two or more accounts exist for the same person or if several members of the same immediate family each have an account written for him or for her. A serious pitfall that can rarely be corrected for is the common habit of ascribing to the principal wife the children her husband sired by other women. Also, sui -ages are sometimes found to be wrong when compared against exact dates, when exact dates are given; but the sui -age is usually all there is, and one can do nothing but accept it.

There is also a problem with children, in that children who died in infancy or early childhood may not be listed in the epitaph of the mother or father. The word shang , meaning "to die in infancy," sometimes flags a given child in the necrology of its parent; and sometimes a child is noted merely as having "died early," a statement without exact meaning. Usually, however, children listed in their parents' epitaphs were only those who survived at least into early adulthood. Thus one cannot use the epitaphs, except in a few isolated cases, to get any idea about infant or child mortality. (For example, we happen to know that Liang Ch'ien [1366-1418] lost a son aged three and a little daughter who died at an age just short of two, because he wrote touching little epitaphs for them. Yet Liang Ch'ien's own epitaph, composed by Yang Shih-ch'i, omits them completely and lists only the four sons and two daughters who grew to adulthood.)^[3]

A hundred or so T'ai-ho men are credited with having authored literary collections—four of them in the Sung, four in the Yuan, eighty-eight in the Ming, and eleven in the Ch'ing. I have been able to gain access only to nineteen, all of Ming date. Most of the rest are probably now lost. So I have constructed a T'ai-ho population out of what I could find in the writings of the T'ai-ho authors of the Ming, plus everything I could find in the way of obituaries written for T'ai-ho people by non-native authors.^[4] The total comes to 508 men and women.

The appropriate first step is to arrange these 508 people in chronological order. The earliest known was born in 994, and the last known died in 1888; but most of the people lived in the Yuan after 1279 and

Ming (1368-1644). Table 1 places everyone in birth cohorts. Starting with the year 1250, when information starts to get plentiful, the people have been arranged in twenty-five-year groups based upon their year of birth.

Immediately evident from this breakdown is a strong clustering in the returns, with two peaks, a larger one for people born in the years 1350-74 and a smaller for people born a century later, 1450-74. These peaks reflect the active years of the more prolific T'ai-ho writers. I should like first to confine my discussion to those people born in the Yuan and Ming (1279-1644), next to profile a few features of the whole Yuan and Ming population, and last to give special attention to the fourteenth- and fifteenth-century maxima, because these coincide with known changes in bureaucratic recruitment rates and with stages in the development of formally organized lineage systems.

The obituaries give us natural families, not the complex households into which natural families were often enfolded or the fiscal households that appear in the tax-and-service registers. In the population in table 1, there are 353 men and 111 women whose obituaries give information about their children. The progeny of the men ranged in number from zero to thirteen, and that of the women from one to eight. The average father had 4.58 children, and the average mother, 3.93. The difference reflects, of course, the polygyny or remarriage of the fathers.

Sex Rations

The obituaries list sons and daughters. The ratio of sons to daughters is a topic crucial to explore because of its significance for population growth or decline and for social structure. What sort of ratio should one expect to find? Other things being equal, a random distribution of human births should approach 104:100, and since infant survival rates favor girls, an adult population should either favor females or show something close to parity.^[5] The T'ai-ho ratios show nothing of the sort. Counting up all the children of T'ai-ho fathers born 1279-1644 gives 985 boys and 555 girls (177:100). For T'ai-ho mothers, the total is 281 boys and 155 girls (181.100).

These ratios are so lopsided as to appear insane. Was female infanticide responsible? Kiangsi in the Ming and later was notorious for its practice of female infanticide. In 1526, it came to the attention of the court in Peking that "many female babies in Kiangsi are not raised"

(pu chü ), with the result that boys coming of age often failed to find mates and so resorted to litigation or to kidnapping widows. The court ordered the censorial authorities to "make strict the former prohibitions on infanticide."^[6] A casual comment by an outsider, Chu Kuo-chen (1557-1632), began with the phrase "Kiangsi people are fond of drowning girls," and then went on to relate a gruesome anecdote about a commoner who drowned four girls and buried them, only to have to keep reburying them because their arms and legs kept protruding from the soil.^[7] T'ai-ho writers were themselves extremely reticent about infanticide. The only mention of it I have found is in Ou-yang Te's epitaph for his in-law K'ang I-sung (1464-1524), which relates how K'ang's father urged him and his brothers to adhere to Chu Hsi's Family Rituals and not drown daughters, as a result of which seven daughters were saved.^[8] Female infanticide was still enough of a general problem, however, that in 1575 the Chi-an prefect issued a prohibition against it.^[9] It seems safe to conclude that female infanticide was practiced in T'ai-ho, but at what rates? How much of the lopsidedness of the local sex ratio can infanticide have been responsible for?

A number of estimates for sex ratios in pre-1949 south China popu-

lations practicing female infanticide are available. George W. Barclay gives 106:100 (106 males for every 100 females). Michael Marmé suggests 108.7:100. Edwin B. Moise offers 110.7:100.^[10] If one were able to make an actual head count of adults in Ming T'ai-ho, the ratio would, in all likelihood, be more like 110:100 than 176:100.

If the 176:100 ratio does not reflect a countable reality in a straightforward way, it may well reflect a social reality of a different kind. If one counts, say, all the sons and daughters of the sixteen Ming emperors, one finds a total of one hundred sons and seventy-six daughters, for a ratio of 132:100, which is also badly lopsided.^[11] Closer to home, the 1610 salt tax for T'ai-ho County was imposed upon a population of 49,921 males (ages 15-59, jen-ting ) and 26,452 females (fu-nü-k'ou , presumably of the same age range)—a sex ratio of 189:100, which is remarkably close to what the T'ai-ho obituaries show. Whatever was going on, it seems to have infected the Ming imperial house and those in T'ai-ho capable of paying salt tax.

One possibility is, of course, a simple failure to report girls, which would explain the shortage of daughters. How can one tell if that is the case? One way to check is to compare sex ratios to family size to see whether small and large families report their daughters differently. In fact, there is a big difference. When the entire population is considered, the sex ratio in fathers' families with one to five children is 200:100

(535 boys, 244 girls), and in families with six to thirteen children the ratio is 145:100 (499 boys, 343 girls). Simple negligence in reporting should have shown a more even distribution across family sizes. The larger T'ai-ho families had boy-to-girl ratios that approached those of the families of the Ming emperors.

Another way to check is to isolate and compare all families, irrespective of size, that report either all boys and no girls, or all girls and no boys. If girls are merely being omitted by the composers of the obituaries, then this exercise should help detect it. The expectation is that the number of zero-boy families should, other things being equal, be comparable to the number of zero-girl families. However, the data in table z clearly indicate that there are far fewer zero-boy families than zero-girl families: nearly thirteen times as many families report no daughters as report no sons.

Surely, some of this lopsidedness must be attributed to the simple neglect of the writers to list in the daughters (some epitaphs state that a man had no sons, but no epitaph ever states that he had no daughters). However, some of it must indicate that there were families that actually

had no daughters, or none they wished to report. One is tempted to throw out all these single-sex cases as flawed evidence, but the sex ratio derived from the gabelle leads one to suspect that many, if not all of them, must in some sense be genuine. The seven families without sons are probably reported correctly; in all but one case, the father in question adopted a male kinsman as his heir. I see no compelling reason to subtract all children from single-sex families from the grand total in table 1, so perhaps the evidence is best left just as it is, on the grounds that much of it may indeed represent something real, if not strictly demographic.

Unquestionably, there is a major problem of unaccounted for daughters here, but the practice of infanticide can explain only about 10 percent of it. If we accept the proposal that an actual Yuan-Ming head count, reflecting female infanticide, would show a real male:female sex ratio of 110:100, then to match 985 boys we need to find some 886 girls. But only 555 girls are reported. That leaves 331 missing daughters. Some, but not all, must have been omitted inadvertently. What can have happened to the rest? Plain-language evidence may offer some guidance.

Wives, Concubines, And Maids In T'al-Ho

There are available 114 epitaphs written for T'ai-ho women. All were written for principal wives, usually the husband's first wife. (Law and custom permitted only one wife at a time. A man could marry a sec-

ond wife only if the first died or was divorced. Concubines and maids ranked below wives and could be acquired at any time.)

The evidence from the epitaphs indicates that, while wives were expected to bear sons, it was even more important that they possess managerial and interpersonal skills and altruism and that they put these characteristics to use within the circle of the family.

It was hard going for a wife without sons, but such a wife could keep her status if, like Ch'en Tou (1369-1409), who lost her only son in infancy, her cooking and tailoring and service to her husband's parents were superior and if she helped buy a concubine for her husband so that his line might continue.^[12] Two of Ch'en Tou's brothers were imperial officials. Her common-descent group, the Willow Creek Ch'en, later prided itself on the excellence of its daughters. "Our family," boasted Ch'en Ch'ang-chi in the sixteenth century, "has produced many notable girls, who have married the great talents in the best families, and have had outstanding men as sons."^[13]

A good wife did some or all of the following things: gave her sons their first lessons in the Four Books; gave aid and counsel to her husband and sons; managed the family budget, including debts and loans; extended the fields and gardens and fixed and enlarged the housing; managed the household bondservants with efficiency; conciliated her sisters-in-law; managed banquets and funerals; arranged nursing for motherless infants; arranged marriages; supervised property division among heirs; and, if her husband died young, protected the property rights of her sons by resisting pressures to remarry.

Just what the better class of T'ai-ho men expected of their principal wives was spelled out by Wang Chih (1379-1462). Wang Chih edited a continuation (now lost) of a well-regarded reader for women, for which he wrote the following interesting preface:

The rise and fall of families is intimately connected to the presence or absence of virtue in the women [the men marry]. That is why we must have concern for how women are taught.

Chu Hsi [1130-1200] established the method for elementary texts when he collected examples of good words and fine actions of the great men of the past. In the Yuan era, Hsu Hsien-ch'en [Hsi-tsai] selected out of the classics and histories good examples for women, and he edited these into a text called the Book of Female Teachings . That text gives a good picture of how to be a proper daughter, wife, and mother. Wu Ch'eng [1249-1331] said that it deserved to rank with the elementary texts [edited by Chu Hsi] and should circulate right along with them.

So Hsu did a good job, but I have been impressed by the Book of Changes ,

where it says that superior men develop virtue by widening their knowledge of the words and deeds of the past. Superior men, as we know, are people who put prime value on such knowledge, but they have the advantage of teachers and classmates and intellectual exchange. Obviously, you can't develop virtue if you're isolated and knowledge-poor. Yet the women's quarters are isolated; what is said inside doesn't leak outside, and what is said outside doesn't get inside. Unless there is a text available, it is hard for women to broaden their horizons and increase their knowledge so that they can establish their virtue and refine their actions.

In moments when I have been free of official duties, I have copied extracts and brought together new material not to be found in Chu Hsi or Hsu Hsi-tsai. Of this I've made a book.

Scoffing at my effort, a friend said: "Women are supposed to be soft and yielding, not hard and assertive. It is a bad woman who makes a slave of her husband, defies her in-laws, and fights her neighbors. She will bring on disaster for sure. You can't let her live with you. Yet here you take the hexagram k'un [female] from the Book of Changes and you gloss it to mean 'hard and square.' Surely that's going too far."

"K'un is the opposite of ch'ien [male]," I replied. "It is the mother of all things. What is female virtue to model itself on if not the hexagram k'un ? Softness and yielding are correct as the 'substance,' but hardness and squareness define how that substance materializes as 'function.' Herein lies the key to handling things both in normal times and in times of crisis. With hardness, [women] can hold firm in the face of threat. With squareness, they can fix their determination and never waver. This is the fulfillment of the 'substance' of softness and yielding, because those words really mean that [women's] minds should be free of violence, that they should not inflict their anger on others, that they should be courteous and forbearing, and never vicious or abusive. Women must have the virtue of hardness and squareness, because when they are weak and swayable, others can control them. The ruin of men and the fall of families results when [women] go along with [bad schemes], or do nothing.

"You have the wrong idea of hardness in mind. The hardness I have in mind is a good hardness. Wise people know how to choose good and avoid evil; everyone is endowed with [the rudiments of] hardness and softness and resistance and compliance because everyone has received the same matter-energy of Heaven and Earth. So when I teach women these things, I'm just guiding them in the light of what they possess already. How can you say I'm going too far?"

My friend couldn't argue with that, so I have restated my point here, to serve as a preface to the book.^[14]

Wang Chih's text-writing project provides an important clue to the problem of the missing females. Whether Wang's educational effort was a success or not, it is clear that the role of principal wife was a demanding one. It was a challenging responsibility that not every daugh-

ter was prepared either by nature or by training to fulfill, no more than every son was equal to the demands of a career in bureaucracy. It is quite possible that many of the missing daughters were girls whom their parents or relatives, for one reason or another, were unable or unwilling to match to a suitable husband as his principal wife.

Indeed, when daughters are accounted for in the epitaphs, it was usual for the writers to give their husbands' name, or to say that they were promised to someone, or that they had married into "famous line-ages," or that they were young and still at home. That obituaries seem to list married or marriageable daughters Only and omit mention of any others is another clue.

Marrying off a daughter as a principal wife could prove difficult and expensive. For example, there was the case of Yang Tzu-p'ei (1391-1458), who had four sons and one (listed) daughter. A local man, though poor, had presented gifts with a view to marrying that daughter, and a marriage agreement had been made. Then later, for some reason, the suitor changed his mind and backed out. Outraged, Yang charged the fellow with breach of contract and took him before the magistrate. Yang won his case, but he ended up having to pay all the marriage costs himself.^[15]

There was also the case of Madame Tseng (1341-1422), who had a similar problem with one of her four daughters—a suitor who could not afford the customary bridal gifts. She had to give the fellow a subvention so that the marriage could go through in proper style.^[16] In another example, when Tseng Yü-hung (1521-88, no relation to Madame Tseng) was a rising star on the examination track, an unnamed powerful person tried to engage his daughter' to Tseng; however, Tseng was already engaged, and it was remembered as an excellent mark in the dossier of his life that he indignantly refused to break the earlier engagement. Later, he himself undertook to find mates for the sons and daughters of a deceased brother, fearing that the job would be too hard for his father to handle.^[17] In fact, a major family responsibility, often mentioned in the accounts, was finding marriage partners for orphaned relatives, male and female alike.

Although the above evidence could be extended by further examples, the point can simply be made here that a first-class marriage was not easily arranged, and the evidence shows how, generation after generation, pools of unmarried or unmarriageable women might have formed. But because no one wrote epitaphs for such women, or even discussed the issue in a general way, one has to imagine an invisible but

by no means small class of luckless T'ai-ho females spared the awful fate of infanticide but unable to become first-class wives because of the poverty of their parents, the inability of their parents to find mates for them, the death of their parents, broken engagements, or their own unsuitability for the role. What can have become of such women?

Some must surely have been sold into bondage as maidservants or concubines. Again, no one wrote obituaries for such women. And while maids and concubines, as did all women, kept the surnames of their fathers, no T'ai-ho maid or concubine can be traced to any specific family, common-descent group, or lineage. A few stories concerning the concubines of T'ai-ho men can be found, however. Yang Shih-ch'i's wife Yen Hsiu died in 1425 at the age of forty-seven, and her epitaph, composed by Yang himself, ascribes all his children, four sons and four daughters, to her (a fifth daughter, who died in infancy, is ignored by Yang here). While Yang was on duty in Peking as grand secretary, Mme Yen spent most of her time looking after things in T'ai-ho, and rumor had it that the third son, Yang Tao (later vice minister of the Court of Imperial Sacrifices, d. 1483), was actually the son of Yang Shih-ch'i's housemaid, a woman surnamed Kuo. Gossip has it that housemaid Kuo was an abused and beaten-looking creature:

When Yang Shih-ch'i was grand secretary, his wife had already died. He had but one maid to handle his sash and comb. One day the imperial palace women invited the wives of the high officials to come to court. When the empress heard that Yang had no regular wife, she ordered her servants to summon his maid. When she saw how lowly and ragged the maid was, she had her combed and rouged, dressed her in jewelry and robes, and sent her off, with the laughing remark: "Master Yang won't recognize her now!" The next day she had the authorities confer a title upon her, as per regulation [for the regular wife of a grand secretary].^[18]

Grand Secretary Ch'en Hsun's (1385-1463) principal wife was Tseng Chang (1387-1431). She gave birth to two boys, who, unfortunately, died as infants. Ch'en was a Han-lin compiler in Nanking at the time. Madame Tseng returned to T'ai-ho, and a few months later, in the fall of 1422, Ch'en acquired a concubine surnamed Mo, who was not a native of T'ai-ho but a young woman left stranded in Nanking by the sudden death there of her brother, an assistant instructor. Ch'en soon lathered one son by concubine Mo (and, immediately thereafter, another by wife Tseng). Reportedly, concubine Mo and wife Tseng got along well together. When wife Tseng died in 1431, however, Ch'en Hsun did not promote concubine Mo to the status of principal wife.

Instead, he kept her as concubine and married as second wife a T'ai-ho woman, Kuo Miao-chih (1413-49, of Kuan-ch'ao, township 31). She produced only a daughter, who died before she could be married. Thus even having produced a first son did little for concubine Mo.^[19] Ch'en Hsun never raised her to the rank of principal wife. In fact, aside from the empress's joke on Yang Shih-ch'i, there is not a single example of a concubine being promoted to wife in the available T'ai-ho data.

Ch'en Hsun is rumored to have had some strong feelings about wife-concubine relations. There is a vivid anecdote about that:

Kao Wen-i [Kao Ku, later a grand secretary] had no sons, and so he took on a concubine. But his wife was jealous, and she got between them and wouldn't let him approach her.

One day Academician Ch'en Hsun paid a visit, and while he and Kao were discussing this problem, Kao's wife, who was listening in behind a screen, suddenly burst forth, screaming and cursing. Feigning anger, Ch'en lifted the table and rose. He hit Kao's wife with a stick until she fell to the floor and couldn't get up, and he kept on hitting her, until finally Kao Ku interceded. Ch'en then admonished her: "You don't have a son, and you can be legally divorced. But Kao hasn't divorced you. He's set up a concubine. And you're cutting off his line of descent by getting between them. If you don't stop, I'll report this in a memorial to the court, where you'll be prosecuted without mercy!"

After that, the wife's jealousy abated. Thanks to Ch'en's anger, there was eventually born Kao Huan, who later became a secretariat drafter.^[20]

To come back to the problem at hand, it appears that T'ai-ho men of means and influence acquired maids and concubines, at least in some cases, in order to sire male heirs by them. In this connection, the suspicion arises that some proportion of the daughters unaccounted for in the parental obituaries—specifically those daughters who lived to become adults but never underwent formal marriage ceremonies, and so "disappeared" into the anonymity of the maid and concubine class—reappear in the obituaries of the men of the next generation, now listed among their consorts because they were the bearer of one or more of the men's children. (Some epitaphs make a clear distinction of rank between a man's wives and his concubines and carefully point out which women produced which of his children. Many epitaphs, however, simply list all the consorts without indicating rank and without assigning any of the children to any mother in particular, so that one is never certain whether a given extra consort is a second wife, a concubine, or an exconcubine promoted to wife.)

To conclude, then, social-class distinctions imposed upon each gen-

eration of T'ai-ho daughters must be considered an unquantifiable but major force, which acted in some powerful way to help produce T'ai-ho's terribly lopsided sex ratios. Of all the T'ai-ho women born into respectable station and spared infanticide, one may estimate that only some 60 percent ended up as principal wives, so managing to preserve their social station. Upwards of 30 percent could not stay in the station into which they were born and were pushed out, or socially downward, into concubinage or servitude.

The Problem Of Upper-Class Population Growth

At this point, I should like to make the assumption that any man or woman in T'ai-ho whose relatives or descendants went to all the trouble and expense of securing a written obituary to be engraved on stone from one or another high ranking scholar-official must in some sense have been socially respectable and, as such, a member of what I propose to define as T'ai-ho's upper class. This group included rich and poor people, people whose lives were serene, people whose lives were filled with distress, people with good educations, and people with only minimal educations; but they all shared some recognized genealogical identity, and they are all described as having performed effectively at some time in their lives some sort of management role—if not as officials then at least as landowners, teachers, family managers, or as parents of successful children. Not one was a servant, a clerk, a maid, a bondsman, a craftsman, a small trader, a tenant farmer, a porter, a hired worker, or anything that smacked of the proletarian, the menial, or the servile.

The first thing to be noted of this upper class, so defined, in connection with the question of its population growth, is that the men often had more than one consort, because they remarried on the death of a principal wife or because they bought concubines or both. (The sex ratio of upper-class parents in the Yuan and Ming was 68:100.)

Was it the case that the more consorts an upper-class male acquired, the more children he had? Surprisingly, perhaps, this is not quite what a study of the issue shows. The results are interesting. For 336 fathers and 1,540 children born in the Yuan and Ming periods, the proposition that having more consorts leads to having more children holds true—to a point (see table 3). However, the T'ai-ho data also bear out Gary Becker's axiom that the more wives a man has, the fewer children each

wife will produce.^[21] But there is more. It may be a quirk caused by the limitations of the T'ai-ho data, but there is a remarkable illustration here of the mutual behavior of not one but three productivity curves. The average productivity curve steadily declines as more consorts are added. The total productivity curve rises until it reaches the second consort, where completed family size achieves its maximum, and then it too declines. The marginal productivity curve (which rates each new consort's contribution to the completed family) shows the maximum impact of the first, the lesser impact of a second, and the increasingly negative contributions of the third and a fourth (the first consort contributes

4.15 children, the second 1.26, the third takes away .55, and the fourth takes away a further .19).^[22]

One might also ask whether, judging from the numbers and sex ratios of their children as supplied by their obituaries, T'ai-ho's upper class was capable of reproducing itself over time, and if so, at what rate? In other words, if everyone born in the Yuan and Ming (1279-1644) is considered and the absence of upper-class in-migration is assumed, did T'ai-ho's quality population expand in size? It is an important question.

To help answer it, I have modified the inputs to a simple equation used by Nathan Keyfitz to answer the related question of how one determines the annual rate of increase in a population with a given average family size.^[23] The procedure is to find the mean age of women at childbirth and use that as the nth root of that proportion of the aver-

age family represented by females who grow up to bear children. The result gives an annual average growth rate. As was noted earlier, however, there is a shortage of obituaries for T'ai-ho women (there are three times as many epitaphs for men as there are for women), and the obituaries suffer from the common custom of ascribing all children to the principal wife, whether they were really hers or not. Therefore it is preferable to identify families from the men's epitaphs, not the women's, and to calculate the reproduction rates of the men rather than of the women. The rephrased question, then, is whether annual reproduction rates, taken from the men's point of view, led to an expanding, a shrinking, or a stable upper class in T'ai-ho.^[24]

The great problem in the average father-centered T'ai-ho family was its large surplus of listed boys relative to listed girls (985 boys and 555 girls of the Yuan and Ming). Eighty-nine of the above boys are reported as having "died early." The exact meaning of this phrase is unclear from the sources, but let us assume arbitrarily that half (approximately 45) of them died before marriage, or before having children. That leaves 940 boys. Thirteen girls "died early." Making the same assumption for them leaves 548 girls.

The result is an excess of 392 boys, who could not have found quality mates in T'ai-ho (where almost all marriages were contracted) and must therefore have been unable to reproduce themselves, at least at the same social level into which they were born. (Elite genealogical identity and pride, and good family management, required upper-class wives; and, as stated earlier, there is no evidence that any concubine was ever promoted to the rank of wife.) Thus only 548 of the 940 boys can have married "good" wives, if they married at all. (Meanwhile, the proportion of the 548 girls who married "good" husbands and bore children must have been close to, if not exactly, 100 percent). For 336 fathers of the Yuan and Ming, the average family size was 4.58. Adapting Keyfitz's procedure, then, we should multiply 4.58 by -37 (the estimated proportion of all the children of the average family which consists of boys able to marry first-class wives). The result is 1.69.

The next step is to calculate a mean age of women at childbirth and use that as the nth root of 1.69. How old were T'ai-ho women at the time of the birth(s) of their children? The answer to that has to be painfully figured out by matching cases where epitaphs of mothers, fathers, and their children are available. I found fifty-five instances where a woman's age at childbirth was known; these ages range from 16 to 44, and the average age comes out to be 26.5. The 26.5th root of 1.69 is

1.0200, which means that, over the very long term (1279-1644), the elite, upper class of T'ai-ho people could have expanded at an annual rate of z percent, which is very high. It implies that the upper class was capable of doubling its size every seventy years at the least.

Given the patriarchal family order of T'ai-ho, it may be preferable to consider the growth rate from the men's point of view. The ages of 106 men are known as of the time one or more of their children were born. These ages range from 16 to 49, and the average is 30.9. That still gives 1.0171, meaning a growth rate of over 1.7 percent a year. (These rates are only somewhat less than recent total population growth rates in parts of the present-day developing world).^[25]

While capable of such growth rates, T'ai-ho's upper class could not have exploded at a rate of nearly z percent a year for very long. It is absurd even to suppose that a county of limited size and resources like T'ai-ho could have supported such growth. One remedy was, indeed, emigration, a topic I should like to address in a moment. But the function of two other remedies—infanticide and downward social mobility—should now be quite apparent. Every generation, the socially pedigreed class of T'ai-ho people engaged in behavior that had the effect of reducing excessive numbers, so that it did not, in fact, grow at a rate approaching 2. percent a year, even though in theory it was capable of doing so. For purposes of social reproduction at the elite level, the average elite family size of 4.58 was an illusion.

What was the purge rate? In spite of the risk of giving a false impression of numerical exactitude, the issue invites numerical exploration. Let us go back to the population of 940 boys and 548 girls, defined as those who survived long enough to marry and bear children. The sex ratio of this population is 172:100, which, as explained earlier, cannot reflect demographic reality in the strict sense. If, given female infanticide, the "real" sex ratio was 110:100, then the "real" population must have consisted of 940 boys and 855 girls. However, only 548 girls are accounted for, so the remainder (307) of them must have disappeared into the maid or concubine class, where their contribution of children to their masters' families was minimal, as indicated earlier.

The effect of this was to create a sizable population of unmarriageable males. Of the 940 males, 392. cannot have found first-class brides.^[26] As social detritus, these 392 males were joined by 307 luckless females, creating a total cast-off population of 699. If we divide that figure by the number of fathers, 336, the result is a "family" whose average number of children is 2.08.

It is, of course, a "ghost" family. It is the elite family that could have been but never was. It is impossible to say to what extent such ghost families were real at some lower social level. Some of the girls were re-absorbed into elite families as maids or concubines (however, only 33 of 289 extra consorts are specifically noted as having been concubines, and nothing at all definite is known of working maids). Some of the males must surely have joined the large army of Kiangsi sojourners and émigrés who plied their crafts and trades in many different parts of China in the Ming.^[27] At all events, the sex ratio of this notional family is 128:100, much more favorable to females than the elite sex ratio.

The ghost family is an offscouring of the elite families. Its creation, generation after generation, is one major reason why the T'ai-ho upper class did not, in fact, expand at a z percent annual rate through the Yuan and Ming periods, a rate that in theory it was quite capable of achieving.

Because of the drag it exerted, the latent demographic power of the ghost family should somehow be subtracted from that of its elite parent. One possibility might be to subtract its potential growth rate. The total of cast-off (or émigré) children comes to 699; .78 of the boys "marry" all the girls (multiplying 2.08 by .78 gives 1.62); and, again using 30.9 as the nth root, the annual rate of increase comes to 1.0157. This "ghostly" rate, subtracted from the elite rate of 1.0171 obtained above, leaves a remainder of .0014, or .14 percent. Indeed, something that low may well have represented the real rate of increase of the pedigreed class of T'ai-ho over the very long run, that is, for the whole period 1279-1644.

The Demographic Behavior Of The T'al-Ho Elite In The Ming

To return once again to table 1, there are two periods a century apart where the obituaries bunch up. One is the quarter century 1350-1374. The other is the quarter century 1450-1474.

These two cohorts lived and reproduced in periods of Ming history that were different both in the constraints they imposed and in the opportunities they provided. The first cohort helped rebuild and repopulate T'ai-ho after its devastation in the civil wars of the 1350s and 1360s. A few of its children reached the very highest positions in Ming imperial government. Hundreds more availed themselves of unusually favorable opportunities to become officials and government teachers at

every level. (As the next chapter will note, this cohort was also heavily involved in researching and compiling the genealogies that became the founding documents, so to speak, of T'ai-ho's pedigreed common-descent groups.)

The second cohort faced new challenges. The rate of bureaucratic recruitment declined by half. Opportunities for local estate-building, available earlier, shrank as well. This cohort's children contributed to the social anomie that people said infected T'ai-ho, as well as Kiangsi generally, in the Hung-chih period (1488-1505) and later. People emigrated in some number. They also spearheaded the founding of organized, property-holding lineages, which were an evolutionary step beyond the more loosely ordered common-descent groups of a century earlier. Did these cohorts share the same "demography," as outlined above? Or did demographic change accompany changing general conditions?

I would propose to enlarge both groups by adding birth cohorts on either side, thus creating seventy-five-year groups, 1325-1399 and 1425-1499 (see table 4). These correspond roughly to early Ming and mid-Ming. Men whose obituaries provide information about children number 179 for the first group, and 82 for the second. The numbers in parentheses in table 4 are those said to have "died early."

Subtracting half the "died early" cases as before and calculating in the same way as was done above for the whole Yuan and Ming population, we find that there were indeed measurable differences between these historically distinct cohorts. While the early Ming cohort, born in a time of opportunity, was nominally capable of expanding at a yearly rate of 2.2 percent, the later cohort, born in a time of reduced opportunity, had a nominal rate of growth of only .7 percent, a dramatic decline.

A comparison of the sex ratios of the children of the two cohorts

(154:100 for the early cohort and 219:100 for the second) shows that while both groups pruned excess children, the elite families of the fifteenth century acted in such a way as to push their children down the social ladder, or encourage them to emigrate, at a significantly higher rate.

In the early Ming, female reproduction rates were comparatively high; on average, each consort produced 3.51 accounted-for sons and daughters who lived to become adults. The fifteenth century witnessed a dramatic decline, as each consort produced only 2.63. Meanwhile the fathers acquired consorts at about the same rate in the fifteenth century as they had in the fourteenth (1.42 in the earlier period, 1.52 in the later).

The above computations relating to elite family formation and its inevitable by-product, social downloading, help explain a number of instances of family behavior discussed in plain language in the obituaries. Whether elite society was expanding, as in the fourteenth century, or shrinking its size, as in the fifteenth, it still extruded, or was forced to extrude, great numbers of its children. The experience was not pleasant. In the fourteenth century, good people sometimes intervened to try to prevent its happening.

For example, there was Liu O (1295-1352), of the Chu-lin Liu of township 38. It was said that a malignant ghost was spreading disease. The parents of one of the Chu-lin Liu families died of the disease, leaving three orphans and unpaid taxes. The tax collectors decided that proceeds from the sale of the three orphans into bondage should be used to clear the account. Liu O went before the T'ai-ho magistrate and made a dramatic plea:

Long ago, our ancestors held the highest positions at court. But now some descendants have fallen into difficulties. They have no food or shelter. They are destitute. They've become slaves to others. This distresses me no end.

It is up to the magistrates to ensure that abuses are stopped, that social custom is restrained, and that compassion is extended. As we know, righteousness means suppressing abuse, decorum means restraining custom, and benevolence means extending compassion. All three are fundamental principles of imperial administration, and those who rule the people cannot ignore those principles. We common folk are no more than baby birds trying to fly without feathers when the magistrates don't give us protection!

Liu O then burst into tears, and the magistrate was moved to compassion. He remanded the three enslaved orphans to Liu O's care, and Liu O raised them and arranged marriages for them.^[28]

Things did not turn out this well for Ch'en Meng-hsing of the Willow Creek Ch'en of T'ai-ho city. A one-time sheng-yuan (county student) expelled for a rules violation, Ch'en died a depressed and broken man in Nanking in 1390, when he was thirty-four. His only son, orphaned young, could not study. He became a small tradesman of some kind and eventually died without heir.^[29] Such for him was the cost of downward social mobility.

Servitude was certainly a possible fate for the downwardly mobile. Escape from servitude is the theme of the celebrated story of P'eng Hsu (1381-1430) and his mother, Liu Ling (1365-1432). She was from neighboring Wan-an County, and her husband's family were Moon Pond P'engs of township 56 in T'ai-ho.

Pieced together from several sources, the story goes like this. Liu Ling's husband died in 1390. He had been a li-chia (tax community) chief for his part of T'ai-ho, and he died in serious arrears with his tax payments, for which his kinsmen became responsible. Accordingly, little P'eng Hsu's uncle, a man who "ran after the properties of widows and orphans," seized the child's inherited property and, at least for a time, placed him in bondage as a household slave (p'u-li ). At length Liu Ling fled with her little son to her father's house, and there she personally supervised the child's primary education. Later she sent him to live with his paternal cousin and her husband, Hsiao Yung-tao (1359-1412), a minor official and an expert in the Book of Documents . P'eng Hsu went on to have a minor career of his own. He entered government through the recommendation channel and ended up as an archivist (9B) in the northern imperial college. P'eng Hsu's gratitude to his mother knew no bounds. Eventually, he moved her back to the P'eng settlement in T'ai-ho and saw to it that she was honored and well cared for in her old age.

It is interesting to see in this example that, while some family-centered institutions and values speeded downward mobility, a countervailing set of institutions and values were sometimes activated to create a rescue net. Here, the indebtedness and early death of P'eng Hsu's father set the conditions in motion that propelled the little fellow right down the social slide. Liu Ling could perhaps have made things easier for herself if she had simply abandoned her son and remarried. Instead,

she chose to protect her son. She wove and sold cloth to help pay off her husband's debts, while she and her child resigned themselves to "living in a bare room, with a ragged coverlet and a tile lamp [as their only possessions]." Her father's family, the Liu of Wan-an County, played an important role in the rescue, as they put up or "boarded" (kuan ) P'eng Hsu and later hired him as a primary tutor. It was said that P'eng Hsu could have earned more as a tradesman or accountant than as a tutor, but his mother made him study and teach the Confucian classics. She knew the Analects and the Classic of Filial Piety and taught these orally to her son.^[30]

It may be noted that the Moon Pond P'eng, who tried to enslave one of their own, were not a poor and obscure bunch but one of the most prominent common-descent groups of T'ai-ho. One can also see the value in one's marrying not just any girl, but a girl of quality, partly as insurance against the ever present possibility that social disaster might engulf one's sons.

However, to accept boarding (kuan-ku ) at someone else's expense, as P'eng Hsu did in his mother's family, was to accept a status of dependency, and that was not necessarily a good situation for a boy to be in. For example, when Ch'en Hsun was fifteen years old, he declined the kind offer of an uncle to accompany him as a student to the home of a certain K'ang family in the T'ai-ho countryside, where the uncle had accepted a teaching position. Ch'en Hsun declined because the K'ang would have had to feed him, and he did not want to be "treated lightly by people," as he would have been had he accepted the dole.^[31] Even hired tutors, like Ch'en's uncle, were in a socially delicate position because of a tendency on the part of their employers to treat them as though they were laborers or servants. As highly regarded a local teacher as he was, Hsiao Ch'i (1325-96) experienced this problem. When he taught in the Lo family school in T'ai-ho city, a matriarch, Madame Liu (1324-1400) had to warn her sons not to ignore or humiliate him. She gave Hsiao Ch'i some face by seeing personally, morning and night, to the preparation of his meals.^[32]

It is evident that T'ai-ho's patriarchal families were often torn between a practical desire to rid themselves of their weaker members and a moral obligation to go to all lengths to provide for their welfare. Yang Shih-ch'i and Wang Chih and the other epitaph writers were intent upon praising those few who acted on their moral duty to prevent the downward mobility that the statistics show to have been so very common.

Some of the writers, indeed, had personally experienced the threat of social extinction in their childhoods. Yang Shih-ch'i is a case in point.^[33] Another is Wang Chih, who gives interesting and detailed testimony about how his step-grandmother, Madame Li, intervened to save him (and his sister and brother) from the predatory designs of his male kinsmen, members of the socially prominent Ao-chieh (Hollow Street) Wangs of T'ai-ho's western suburb:

Madame Li certainly helped us Wang. She raised me and my brother long enough to let us continue the sacrifices to our ancestors, and thanks to her, our ancestors buried below can rest in peace. What she did for us can hardly be repaid. I cannot forget that. That is why I must now write about her, so our posterity will know.

Madame Ch'en was the first wife of my grandfather [Wang Chu-t'ing, 1317-83]. After she died, he married Madame Li. My mother, Madame Ou-yang [1349-84], served Madame Li faithfully, and they worked together harmoniously to manage the family, so that my grandfather and my father [Wang Po-chen, 1342-1416] were not burdened with domestic duties and could spend all their time in literary pursuits.

Then my grandfather died, and a year later my mother died. There were no maids in the house. My brother and I were young [Wang Chih was five], and Madame Li was the only person we could rely on to stay alive. She looked after us carefully, as though we were her own children. She fed and clothed us and kept us from doing foolish or dangerous things.

When I was five or six sui , she had me study under tutor Tseng Chung-chang. I used to shirk my studies, and she would cry angry tears and say: "The job of you Wangs has always been to study, but you shirk! Do you want to become a small man?" Every evening she would remind tutor Tseng to come by at daybreak to make sure we got to school. It was as though she felt she couldn't do enough to show her care and her worry for our futures.

Careful as she was of us, we still hadn't encountered serious troubles.

My father [at the time a bureau secretary (6A) in the Ministry of Works], after mourning leave, returned late to duty, and for that infraction he was ordered exiled to An-ch'ing [in present-day Anhwei Province, where he remained, ca. 1388-98]. A Wang kinsman then got the idea to seize his personal property, and suddenly Madame Li was faced with more trouble than it appeared she could handle. First the kinsman tried cajoling her.' she could stay and be fed in his house and live with his mother. But Madame Li replied: "I'll just stay in this poor house and raise my young grandchildren. I won't go, even though you say you'll take care of me." That was the end of that ploy.

Now it happened that there was a widow in our common-descent group who fell for that kinsman's false promises. He got hold of her dead husband's deed and sold all the property [she was holding in trust]. Finally, she died homeless, and her grandson fell victim to hunger and cold. People then came to see that my grandmother was extraordinarily farsighted.

When his first ploy failed, the kinsman seized the old home of my grandfather's younger brother and sold it. He also claimed that a certain garden of ours was partly his possession. He divided it up. Then he invited a bunch of bad people to buy cheaply [the topsoil rights?], and he had them bring their tools and come work it. Someone objected to all this, and he struck and killed him. But Madame Li spoke up and said: "I'm just an old Wang widow, but when your uncle and older brother were alive, you never said anything about that property, and yet now you and those slaves [nu-pei ] just seize it. I'm going to be buried there when I die, so if you want that land, you'd better kill me first. As long as I'm alive, I won't let you have it!" When they all heard that, they put out their tongues in shock, and in the end, they backed off. So the garden was saved.

Our ninth generation ancestor was buried in a safe place. Everybody considered the site geomantically good. The P'eng, a powerful common-descent group of our county, were trying to find a place to bury a father of theirs, and they were offering a high price for a site good enough to bring wealth and status to the dead man's descendants. At that time, my father was still in exile. My brother and I were in prison in the provincial capital, because of an accusation made against us. Our kinsmen, eager for the money, wanted to exhume [our ancestor] and sell the tomb site to the P'eng; but they were afraid of Madame Li, and so they made the argument that the tomb site was in fact no good, given that none of the Wang families was doing well, with three men presently in exile in army garrisons, and two grandsons under indictment, with who knows what disaster in store for them. But that argument did not impress our step-grandmother, who forbade them to sell the tomb site.

Finally they sent a dull-witted fellow to go to Madame Li, to tell her that everyone's mind was made up, that she had no sons or grandsons and had no business stopping the sale.

At that time, my son Wang Tzu [1397-1458] was born. Madame Li held him on her knee as she replied: "Four big Wang families [fang ] descend from that ancestor. Three of them want to sell. But that leaves us. If our sons and grandsons never return, there is this great-grandson here to serve as master of the tomb, and he'll never let it go. You people are violating the principle of Heaven!"

She placed the child on the floor, grabbed a stick, and drove the fellow away. When the P'eng heard about the dispute, they decided it would be wrong to continue to try to buy the site, and that is how the site continues to be ours to the present day. We Wang had declined in those days, and except for Madame Li, we could not have survived our troubles.^[34]

Wang Chih went on to make the very same point, about how important k was for women to be "hard and square," that he made in the preface to his sourcebook for women.

Wang did not mention here that when he was twelve years old, he was sent away to Lu-ling County to live with a maternal uncle, a pro-

fessional tutor, who "fed and taught" him for two years. Nor did he mention that when he was five and had just lost his mother, Madame Li needed help at once; and so a child-bride (Madame Ch'en, 1377-1458, of the Willow Creek Ch'en, and a relative of Wang's dead grandmother) was brought in nominally as a wife for Wang Chih but principally for her skills in weaving and tailoring.^[35]

Wang Chih spent most of his adult life as a high official far away in Peking, but even from that distance, he tried to redirect Wang family strategy—away from predation on the weak and toward a new goal of social welfare for all. His oldest son, Wang Tzu, grew up to become a government Confucian instructor, but Wang Chih had four other sons, who took turns living with him in Peking. Using those sons as agents, he was able to keep a hand in family management back in T'ai-ho. By 1450, that family was a large one; Wang guessed that it contained "over a hundred" people of high and low status.^[36]

In Peking, Wang lived with concubine Ch'iu. His wife Madame Ch'en stayed in T'ai-ho until her sons grew up. At one point, there were so many dependents that some of them were going to have to move out, but Madame Ch'en saved the day for them. In her epitaph, Wang Chih wrote that she "sold all her own jewelry and utensils to buy paddy at the west wall [of T'ai-ho], and then day and night she allotted farming tasks to the male and female slaves or bondservants [t'ung-pei ] and to the older sons who were capable of farm work, and with Heaven's help there was enough food [to keep the family together]. Thanks to her direction, my sons to this day provide for themselves by working at farming."

Another case from the early Ming era centers upon Liang Hun (1370-1434) and his wife, Madame Liu (1369-1432). In their case, cooperation among intermarried common-descent groups created a social safety net for as many as five weak orphans, who might otherwise not have survived. It must be emphasized that examples of this kind (there are many other early Ming instances of it) must have been exceptional and thus worthy of comment by the local literati. In view of the implications of the statistics, such cases cannot have been the rule. Yet they-serve to illustrate very dramatically the relentlessness of the pressures that created downward mobility in T'ai-ho.

In 1393 or 1394, Investigating Censor (7A) Ch'en Chung-shu, a Willow Creek Ch'en, died, unexpectedly and destitute of personal prop-

erty, in his early 40s in Kwangtung. His wife had already died several years before. Thus his son Ch'en Shang (aged fifteen or sixteen) and his daughter became orphans. His younger brother Ch'en Chung-heng (1354-1413) refused to concede them a share in his personal estate. He owned a nice house and substantial farm property near T'ai-ho city, but his own four sons wanted it all for themselves. So censor Ch'en's grieving orphans were placed "in a condition so wretched that few people could have borne it"—and it is left to the imagination whether that meant beggary or enslavement or what. Then fortune smiled. Ch'en Shun-chih (1344-1426) was the older sister of Ch'en Chung-shu and Ch'en Chung-heng, and she was Liang Hun's mother. She imposed it as a moral obligation upon her son to step in and help raise her brother's orphans.

The Liang also owned farm property near T'ai-ho city, so there was a resource base to help support their social work. (Liang Hun was no stingy manager. He rented some of that land to unnamed in-laws, who suffered some untimely deaths and so fell behind in their rent payments; but he forgave the debts, and even signed over some of the land because the renters were a Confucian [ju ] family that had fallen temporarily on hard times.)

Liang Hun was able to raise Ch'en Chung-shu's orphan daughter and eventually to marry her to a good boy of the Yen family. The other orphan, Ch'en Shang (1378-1413), soon proved a brilliant success. In 1:411 he achieved his chin-shih degree and had just begun his official career when he, like his father before him, died young, leaving behind, just as his father had done, two young orphans, a boy (Ch'en I) and a girl, and no property. Again, they and their mother faced beggary. Someone made an appeal on their behalf to investigating censors, who interrogated Ch'en Chung-heng about the matter and forced him to allot them "a tiny plot where they could stay." At some point Ch'en Shun-chih also stepped in and again got her son Liang Hun to help take care of them. Later she also arranged the orphans' marriages. She married Ch'en I to her own granddaughter, née Liang. So here were two generations of Ch'en orphans that Liang Hun and his wife and mother helped raise.

Liang Hun also took in and helped raise and teach an orphaned paternal cousin, Liang Chiung (1384-1429). In fact, Liang Hun's wife, Madame Liu, did most of the nurturing, and she also took care of her aging mother-in-law, Ch'en Shun-chih. She had a staff of maidservants (nu-pei ), whom she was recalled as having managed with skill. It was

her third daughter who was married to orphan Ch'en I. They all lived in T'ai-ho. As the male orphans grew old enough, the women sent them for their education to wherever Liang Hun happened to be at the time as an instructor in county schools in various far-flung parts of China.

In addition to Madame Liu, Liang Hun had another woman, probably a concubine, who lived with him when he was away from T'ai-ho. Between them, Liang Hun fathered four boys and four girls. The girls were given good marriages (although the youngest was married to Grand Secretary Yang Shih-ch'i's oldest son, a real estate entrepreneur who died in prison on charges of extortion and murder in 1444). One of Liang Hun's sons, Liang Li, was taken by recommendation into government teaching and ended his career as instructor second class in the imperial college. All the orphans did quite well. Liang Chiung went on to achieve a chin-shih degree and was a bureau director (5A) in the Ministry of Justice at the time of his death. Respectable marriages were arranged for all the girls. Ch'en I's son Ch'en Yen became a county Confucian instructor. Altogether, this was a very successful social rescue operation.^[37]

If the figures derived from obituaries of people born in the fifteenth century provide anything like an accurate view into elite social processes in T'ai-ho County, they show that while the rates of reproduction went down, the rates of downward or outward mobility went up. Moreover, the stories of social rescue that one often encounters in the epitaphs of people born in the fourteenth century are vanishingly scarce in the epitaphs of people born in the fifteenth. It was men born in the fifteenth century who took the lead in the founding of the organized lineages that became prominent features of the social landscape of T'ai-ho in the sixteenth century and later; and while lineages had welfare functions, it is impossible to tell whether they were more effective, or less, as inhibitors of social disaster than the informal interfamily networks that they seem to have replaced.

It is interesting that the few social rescue stories of this later era should center upon travelers and émigrés—upon people down on their luck and far from their native county.

Wang Ch'iu (1445-1507) was a grandson of Wang Chih, and his second wife was Jen Lien-chen (1454-1526), youngest daughter of Jen Heng (a provincial degree holder from T'ai-ho city who ended his career as a county magistrate in Fukien). Around 1500, Wang Ch'iu was

serving as an office director (6B) in the Court of Imperial Entertainments in Nanking, and Jen Lien-chen stretched his small salary to help pay for entertaining visitors from back home in T'ai-ho; for a tutor for their five sons; for board and school supplies for some other boys who came to study with her sons; and for charity to locals in straits. She helped poor Hsiao Jou. "Hsiao Jou had come [from T'ai-ho] to the Ministry [of Revenue in Nanking] accompanying a tax shipment," her epitaph reads. "But his comrades had abandoned him because he was sick, and so he came crawling [to Wang's house in Nanking]. Wang was away in Peking at the time, and the gatekeeper wouldn't let him in. But Madame Jen, when she was told, frowned and said: 'He's sick and so far from home, and he has nowhere else to go.' At once she had him housed nearby, and fed and medicated him until he got well and then gave him money enough to return home. People admired her for that."^[38]

There was also Lo Fu (1438-1507), a T'ai-ho man who also lived in Nanking, where he was registered with a naval guards unit and served as something of a one-man social welfare agency for failing elites from back home. The early deaths of his two brothers left him the task of supporting their widows and one orphan daughter. He also supported his sister and her husband, who lived with him. Plus there came indigent kin from back in T'ai-ho. People said his kitchen was always busy, with maids and servers constantly running about under his personal direction. His T'ai-ho neighbors in Nanking reported that he took in two generations of orphaned cousins, plus a distantly related orphan boy from back home, plus a destitute friend and his little daughter. The friend soon died, and Lo made it his responsibility to marry the daughter off. His neighbors also reported that a certain Hsiao Hsin owed him money that he could not repay by any means other than selling him his daughter; but Lo Fu indignantly refused and forgave the debt altogether.

Lo Fu was obviously well off. Where his income came from is a complete mystery because his grandfather had been exiled to Manchuria early in the Ming for the offense of "discussing national affairs as a mere commoner" and had returned home to T'ai-ho in his old age. His father moved to Nanking under military registry, but because he was a noted tutor in the Book of Changes , his commander excused him from duty. Lo Fu himself was no more than a sheng-yuan who had failed five times to get his provincial degree. Eventually his son, Lo

Feng, achieved his chin-shih degree in 1496 and went on to have a moderately successful official career.^[39]

Emigration

There is no doubt that emigration drew off some substantial percentage of the excess population of T'ai-ho. The numbers cannot be counted. It is possible, however, to catch certain glimpses of the process.

There are signs that T'ai-ho began to send out émigrés in early times, when it was still attracting new settlers. There were, for example, successive generations of Wus who moved from T'ai-ho to Kwangsi in the thirteenth century.^[40] In Kwangtung Province, five of forty Hakka surnames claimed ancestors who were T'ai-ho people in Sung, Yuan, or Ming times, and these include émigrés from such well-known common-descent groups as the Hsiao of Nan-ch'i and the Lo of Ch'ueh-ch'eng (which in T'ai-ho are never flagged as being Hakka).^[41]

Early in the Ming, three brothers of the Liu of K'an-ch'i Ward (township 63) moved away to make new homes for themselves in Yen-t'ing County, Szechwan.^[42] Wang Chih's distant kinsman, Wang Tsai (1314-90), was a government instructor in Kan-chou in southern Kiangsi, and he decided to settle there permanently. Wang Chih's maternal uncle, Ou-yang Huai (1357-1444), moved from T'ai-ho to the market town of Yung-ho across the border in Lu-ling County, where he was attracted by an offer to board as a live-in son-in-law to a local resident. Ou-yang made a career as a private tutor in classics and calligraphy.^[43] Yang Yun-wen died in office as magistrate of Ch'ang-yuan County, Honan, in the 1370s. His young son, rather than returning home, took up registry there, and two descendants of these ex-T'ai-ho Yang later became local government students in Ch'ang-yuan.^[44] Sometime during the Ming, various Chiangs, Ch'ens, and P'engs emigrated to Hunan Province, where they became founding ancestors of lineages that by Ch'ing times (1644-1912) had achieved national prominence.^[45] A study done long ago shows that, in fact, at least some fifty-four émigrés from T'ai-ho moved west to Hunan, as part of a large outflow of Kiangsi people, mainly in the early part of the Ming period.^[46]

Military exile created a certain amount of early Ming emigration. A few ragged original records remain of twenty-two T'ai-ho convicts who were exiled to garrisons in Manchuria in 1392; twenty-one of them died there.^[47] But some exiles found permanent homes in Manchuria—

the Liu at the Fu-chou guards, and the Hsiao at the T'ieh-ling guards, for example—and they are regularly referred to, even in the late Ming, as T'ai-ho people registered in Fu-chou or T'ieh-ling. Their T'ai-ho connections were never completely cut.^[48] Whether or not the rate of emigration increased in the middle years of the Ming dynasty is hard to say. In any event, it surely continued. The T'ai-ho County gazetteer, in its list of native sons who qualified themselves for official position, begins to note from the Ch'eng-hua era (1465-87) onwards significant (and increasing) numbers of one-time natives who achieved their provincial (chü-jen ) degrees in provinces other than Kiangsi, or who entered the imperial colleges as sui-kung (annual tribute) from county schools in various parts of China.

It is also clear that the T'ai-ho émigrés of the fifteenth century and later found two new and undeveloped frontier regions in China to settle in and exploit. One such region was the far south of Kiangsi where, earlier on, Wang Tsai had gone. An official report of around 1530 describes southern Kiangsi (Nan-Kan) as having long acted as a magnet for people from Chi-an Prefecture, because paddy and mountain land were available, and newcomers might reap profits from rice, indigo, timber, and charcoal production.^[49] "Kan-chou has a lot of high-grade fields and idle soil," wrote Ch'en Ch'ang-chi, probably sometime in the 1540s. "The taxes there are light, and the people are few; whereas in Chi-an the fields are low-grade and the imposts are heavy, and the population is thick, but production is thin. As a result, many poor people take their families and kin and find work in Kan-chou, and they never return."^[50]

From 1563 to 1564, Hai Jui (at the time magistrate of Hsing-kuo County, T'ai-ho's neighbor to the east) also noted that the population of Chi-an Prefecture was dense, that land and commercial opportunity were scarce, and that therefore many people were leaving the province altogether. He argued that more people would migrate within the province if it were not so hard to do so legally.^[51] Despite the difficulty, there were at least twenty-eight T'ai-ho men who emigrated to southern Kiangsi, entered county schools in that region as sheng-yuan , and qualified themselves as officials in the Hung-chih, Cheng-te, and Chia-ching eras (1488-1566). Around 1568, an official complaint was voiced that Hsin-feng County, about a hundred miles south of T'ai-ho, had in effect been seized and occupied by emigrants from Wan-an and T'ai-ho, even though they could not legally register there. It was asserted that most of the shopkeepers were outsiders and that 70-80 percent of the

land was in the hands of outsiders as well.^[52] Some Hsin-feng earnings were repatriated to T'ai-ho: the Kuo lineage temple in Ch'e-t'ien Ward (township 60) was funded in part from the proceeds of fields owned by kinsmen down in that county.^[53]

The other region that attracted T'ai-ho émigrés was extreme northwestern Hupei, about five hundred miles away, a turbulent and violent frontier zone in Ming times.^[54] A massive emigration of Kiangsi people into all parts of Hupei was underway in the fifteenth and sixteenth centuries.^[55] T'ai-ho émigrés concentrated in a few counties near the Shensi and Honan borders. Representatives of at least twenty-one T'ai-ho lineages settled in there, and soon at least fifty men from these émigré lineages achieved chü-jen degrees or entered the imperial colleges. By the 1570s, Liang Ju-hsiao and his son had organized in Yun-hsi County, near the Shensi border, a replica of the Liang lineage back in T'ai-ho, with its own endowed ancestral temple and school. Fourteen of these transplanted Liangs became sheng-yuan and eventually Ming officials.^[56] South of Yun-hsi, in Chu-shan County, Ou-yang Jen made similar kinds of arrangements, and the Ou-yang lineage there, like its parent back in T'ai-ho, came to be regarded as an unusually large one. Ou-yang Te (1496-1554) certainly considered these émigré Ou-yangs as kin; he once wrote them a letter congratulating eight of them for qualifying to take provincial examinations and scolding the others for their squabbling.^[57]

Bondage

While it can be stated with certainty that the sale of persons took place in T'ai-ho and that the upper class placed certain of its children in bondage, there is no way to tell whether these unfortunate males and females were then exported from their native county, or whether they were kept in family service somewhere in T'ai-ho itself. Was T'ai-ho's servile population a population of strangers? It is impossible to know. Judging from the few known surnames of individual bondsmen (Lin; Yang, meaning "sheep"), it appears that at least some were strangers to the county.

It is also unclear to what extent people in bondage reproduced. Nor is it clear whether bondage was hereditary, life-long, or for a stipulated term of years. It was surely a persistent institution, at least in a mild form, as witnessed by the "hereditary servants" of eighteenth-century T'ai-ho and the occasional field slaves observed there in the 1930s.^[58] In

Ming times, bondage was often mentioned in the obituaries of men and women of T'ai-ho's upper class as a troublesome issue in family management.

It was felt to be best when family servants were kept under benevolent but very strict control. It was stated, for example, of Hsiao Ch'uan (fl. 1540s, of Nan-ch'i, township 59) that "he didn't flog his slaves when they made sly excuses for their misbehavior; he would have them flogged for serious offenses, but then he would dismiss such incidents from his mind and forget that they had ever occurred."^[59] Yang Yin-ch'iu's wife, Madame Liang (1548-95), "carefully kept [the family quarters] locked and bolted. When she collected woven hemp cloth from the female slaves [nü-nu ], she would issue them raw hemp, grain, and vegetables. When [doling them food] from the jar, she would give them orders and examine them. 'Who in the world eats without working?' " she would ask. As manager of a large family in T'ai-ho (her husband, an official, was absent), Madame Liang "forbade the slaves to overstep the bounds and flout the family rules, but she would reward the hardworking ones and cherish those who showed a willingness to reform. Whenever the other wives thrashed the slaves too hard, she would weep and scream in an effort to stop the violence, on the grounds that slaves were, after all, human beings."^[60]

There existed a certain sense of moral responsibility for the welfare of faithful bondservants. For example, there came a point in the life of Liu Sui (1455-1533, of Liu-chia-kang, township 61) when he became poor. Yet despite his reduced circumstances, he "could not bear to abandon his slaves and have them go into the service of others"; and his servants were, reportedly, grateful for his determination to keep looking after them.^[61] Hu Chih's mother, poor as she was, kept two bonds-people, a man and a woman, both of whom were old and thin and had no other place to go.^[62]

Slaves in T'ai-ho were ordinarily assigned menial work as house servants and as field hands, but sometimes they, like eunuchs in the palace of the later Ming emperors, won the confidence of their masters and by degrees encroached upon matters of family government and made themselves "powerful" (hao ) in the process. The T'ai-ho gazetteer of 1579 issued a general warning about this matter. "Villainous slaves of the hereditary families are swallowing their masters and mistreating their helpless heirs," it stated. "This has created an atmosphere of growing insubordination and rebellion that numerous prosecutions and punishments have been unable to stop."^[63]

Indeed, sixteenth-century obituaries yield examples of what appears to have been an increasing aggressiveness on the part of T'ai-ho's bonds-men. In Kuan-ch'ao Ward (township 31), Kuo Ch'i-shih (1496-1573) was nearly ruined in some unspecified litigation in which "powerful slaves" brought "false" testimony against him.^[64]

For reasons unknown, one or more "powerful slaves" nearly beat to death the elder brother of Chang T'ing, a sixteenth-century T'ai-ho geomancer and physician. Chang T'ing was forced to sell personal property in order to finance legal prosecution of the offenders.^[65] On Hollow Street in T'ai-ho's western suburb lived Wang Jui, a sixteenth-century descendant of Wang Chih. Wang Jui was a merchant and part-time litigator. Among the component families of the Wang lineage, a "powerful slave" was bullying his young master. The young master complained to the lineage head (tsu-chang ), and when the head did nothing about it, Wang Jui took it upon himself to arrest the slave, flog him before the Wang ancestral temple, and charge him at the county yamen. "There is no need to thank me," said Wang Jui to the grateful young master. "My thought was to prevent the downfall of the Wang lineage, and that I succeeded in doing."^[66]

In Moon Hill Ward (township 32), Tseng I-ch'ing (1538-87) perceived that local custom had degenerated, with "sons selling their fathers' estates, slaves stealing their masters' properties, and the strong eating the flesh of the weak." He further noted that "in some of the old families of [T'ai-ho] city, the clothing of the slaves is as fine as that of the kinsmen themselves." Tseng thought the fault for that lay with the kinsmen. He enforced the appropriate status differences on his own estate and disciplined the menials, so that "none of them dared wield power, or disorder the family rules."^[67]

In the mid-seventeenth century, with the collapse of Ming rule in China, the upper class faced widespread and serious retaliation for the downward social mobility that its demographic behavior had for so long encouraged. For a long time, the elite people of T'ai-ho had exacerbated the situation by extending to their servile population a dangerously contradictory combination of privilege and repression, of empowerment and punishment. The same or worse was the case elsewhere in Chi-an Prefecture. In Yung-feng County, both slaves (nu-p'u ) and tenant farmers (tien-hu ) organized themselves into groups called "small compacts" (hsiao-yueh ) and revolted against their masters.^[68] In the period from 1645 to 1648, "tenants and powerful slaves" (tien-k'o hao-nu ) in Lu-ling, An-fu, and Yung-hsin Counties went on rampage under

the leadership of "leveler kings" (ch'an-p'ing wang ) who threatened to "level master and slave, noble and base, and rich and poor." They forced their way into the great residences, looted the storehouses, and put on finery. They tied and flogged the owners and then made servants of them. "We are all men," they reportedly said. "Why do you call us slaves? We'll reverse that from now on."^[69]

T'ai-ho County suffered the destruction of a murderous outlaw by the name of Liu Chin (or Ching), described as a former slave (p'u ) of an unnamed great local family. He was captured late in the Ming, and the magistrate was going to execute him; however, certain T'ai-ho men holding official status (shen-shih ) took bribes from Liu Chin and effectively interceded on his behalf. Some local people protested. The magistrate's answer to them was that Liu's followers would exact fearful revenge: "If he dies, there will be a rising and certain disaster for you T'ai-ho people. We don't want war out in there in the cantons."

In 1645, in the civil wars of the Ming-Ch'ing transition, "poor people" in the eastern cantons conducted food riots, and "powerful slaves [hao-nu han-p'u ] resisted their masters, formed gangs, and did bad things." In 1649, exslave Liu Chin and his forces seized and occupied the city of T'ai-ho for several months, until Ch'ing troops drove him out. In 1653, finally, Liu Chin was captured in the mountains along the Hunan border and executed.^[70]

It is time for a brief recapitulation. Just as the "economy" of T'ai-ho County in the Ming period is unknowable, its "demography" cannot be known either. What can be discussed are the changing styles of resource management by local elites and the demographic processes that affected those same elites. It is impossible to talk about T'ai-ho County in the Ming without becoming constantly aware of the presence of a robust two-class system there: an upper class of "managers" of various sorts who bore genealogical credentials and a lower class of menials, maids, concubines, bondservants, field families, laborers, small traders and the like who were without genealogical credentials. In the Ming, upward mobility from the lower class was unheard of. "Powerful slaves" were still slaves. (The lower-class revolts of the late Ming may have loosened things a little—they did elsewhere in China—but little is known of T'ai-ho County in the Ch'ing.) Downward mobility from the elite into the lower class was, however, a perennial possibility.

One may ask whether the demographic profile of the T'ai-ho elite can

be understood as a local example of a much wider phenomenon. It has recently been shown that the results of studies of Chinese microde-mography can in some respects be generalized to the region and nation and can lay groundwork for revising accepted notions of the larger trends of China's population history.^[71] Almost all of the work that has been done centers on the late Ming and Ch'ing, however, and uses line-age-kept genealogical records rather than county epitaphs as source material. In an earlier paper, I showed that T'ai-ho mortality rates (lower in the fifteenth century, higher in the sixteenth) mirror those given for south China generally in a 1973 study by Michel Cartier.^[72] But a great deal more work needs to be done in demographic study for the Sung, Yuan, and early Ming periods, and there is as yet little available on the regional or national scene to which to link the microdemography done in this chapter.

What can be emphasized here are connections between the demographic behavior of the upper class and other developments inside T'ai-ho over the course of the Ming. The expansion of the upper class (as shown in the male cohort born 1325-1399) coincides with a loving appreciation of the local landscape, the building or rebuilding of landed estates, the formation or reconstitution of common-descent groups, unusually high rates of recruitment into Ming government, and, as will be noted in part 3, a strong sense of county patriotism among T'ai-ho men inside Ming government itself. The contraction of the upper class (as seen in the cohort born 1425-1499) through increased rates of downward mobility and emigration coincides with a loss of interest in local landscape, the creation of powerful lineage institutions in place of the looser common-descent groups of the earlier period, an abrupt decline in the rates of bureaucratic recruitment, and, for the literati, the adoption of new intellectual orientations in which pride in one's native locality had no place.