Traditionally the family has been viewed as the nexus of a constellation of social institutions which together comprised Luso-Brazilian society. While a number of studies have appeared describing contemporary Brazilian social organization, not until very recently has more than scant attention been paid to the colonial antecedents of present-day social phenomena. This essay is an effort to fill in that background, using Vila Rica, the capital of colonial Brazil’s goldmining region, as a case study. Three major contentions will be advanced. First, the patriarchal, extended family, so often viewed as the predominant colonial family type, existed for only a handful of people. Second, in preindustrial Vila Rica a wide range of family types existed, with nuclear and matrifocal families predominating. Third, marriage was not a means of integrating the society but served instead to differentiate segments of the population.
Foremost among earlier efforts to describe colonial social organization is the pioneering work of Gilberto Freyre, exemplified by his major works, Casa-grande & senzala and Sobrados e mucambos.1 In these studies, Freyre traced the development of the upper-class patriarchal family, then its displacement from the plantation “big house” to the city. In both settings, the patriarchal extended family was considered the central institution for transmitting culture.
Freyre’s observations had a great impact on other students of social organization. But, whereas he precisely identified the patriarchal, extended family as only one of a range of family types and limited to specified socioeconomic environments,2 others carelessly generalized from his conclusions. This is particularly evident in the works of Fernando de Azevedo, João Camillo de Oliveira Torres, and L. A. Costa Pinto.3 These authors focused their attention exclusively on the patriarchal, extended family, and ignored alternative forms of social organization. The same basic position is found in T. Lynn Smith’s influential Brazil, People and Institutions, wherein the traditional family is defined as the patriarchal family, which “early obtained almost unlimited sway in Brazil.”4
Other social scientists commented in passing on the existence of alternate family types but generally ignored these in their analyses, or allowed their personal social values to interfere with their objective examination of these alternative forms. Of these Antônio Cândido and Thales de Azevedo are typical. Antônio Cândido, in one of his early works, described colonial society as divided into two groups: the familial (or patriarchal family) and the nonfamilial. The latter was composed, he said, “of a nameless mass of the socially degraded, those cast off by family groups or brought up outside of them … [and who] reproduced themselves haphazardly and lived without regular norms of conduct.”5 Antonio Cândido, whose primary concern was the contemporary family, presented the modern conjugal family as having developed from its patriarchal predecessor.6 Azevedo presented a similar analysis, contending that although alternative family types, such as the “partial family” headed by a woman, had existed during the colonial period, the patriarchal extended variety had been the most significant.7 He partially ascribed the development of the nuclear family in the twentieth century to the disintegration of the patriarchal family.8 These studies’ reliance on subjective impressions, rather than on analytical data, for the colonial era is not surprising since they focus primarily on the contemporary period, using a description of the colonial system only to provide general background.
But not all social scientists have accepted this view of colonial society. Emilio Willems, for example, was able to escape the allpervading impact of Freyre’s analysis. Willems’s study of rural society led him to suggest a differentiation of family types along class lines, with the patriarchal, extended family seldom being a feature of lower-class society.9 While concerned with contemporary social phenomena, Willems’s analysis reopened the examination of colonial familial organization, which the general acceptance of Freyre’s conclusions had precluded. Foremost among those who have sought to redefine the nature of the colonial family are Maria Luiza Marcílio and Luis Lisanti. In a number of recent publications dealing with São Paulo they have begun to look beyond Freyre to challenge existing theories on family size and composition.10 The present article will contribute to the same effort.
The urban center used for this study, Vila Rica, was one of the leading centers of eighteenth-century Brazil. Founded after major gold deposits were discovered in the interior province of Minas Gerais in 1695, it rapidly evolved into the major economic center of the mining district, and, after 1720, became its political capital. The peak of the town’s growth was reached in the 1740s, when more than 1,000 houses were registered for the densely packed urban core alone and the town’s total population probably numbered about 15,000 people. The period after 1752 witnessed a decline in the economic fortunes of the region, a decline that became precipitous after 1760. By the end of the century the situation had stabilized.11 The remaining population, hovering around the 7,000 mark, was supported by residual gold deposits, the advantageous position of the town as a trade center, and its administrative prominence as the capital of Minas Gerais. During this phase of its history Vila Rica also became a major cultural center.
It was also during this period of genteel decadence that census taking was initiated. One of these censuses—that of 1804—provides much of the data for this study.12 Because of the critical importance of this material some mention need be made of its nature. The termo, or county, of Vila Rica was divided into parishes. The urban parishes of Antonio Dias and Ouro Prêto (the name by which Vila Rica is known today) were then divided into smaller districts: Antônio Dias into Morro, Alto da Cruz, Padre Faria and Taquaral, and Antonio Dias itself, and Ouro Prêto into Cabeças and Ouro Prêto. The enumeration of the residents of each district was carried out by militia officers.
The census tracts for these six districts are consistent, each district’s being divided into residential units. Each residential unit is then divided into its major component parts: the primary family encompassing the head of the household, his spouse, and their children; secondary families composed of married children or single children who were heads of families;13 agregados, dependents or retainers, including relatives; boarders; and slaves. The census provides the following information for each person: name, marital status, race, title occupation,14 age, and relationship to the head of the residence.
A central problem in the analysis of the census data is the formulation of an operational definition of the family as opposed to the household. Portuguese attitudes evidence extensive confusion between the two. An eighteenth-century Portuguese dictionary defined the family as “the people who make up a house, parents, children, and servants.”15 According to one student of eighteenth-century Portuguese society, this dictionary definition was applied generally.16 It would appear that the family, as a descriptive term, is synonymous with the “houseful,” and thus is similar to the definition used in other parts of Europe.17 In this study, however, “houseful” refers to all those individuals who live on the premises, which includes such structures as slave quarters that are physically separated (but connected by authority) to the nearby main house. The “houseful may contain individuals who do not constitute a family because they are not biologically or juridically related. The household refers to those people sharing the same dwelling. Thus, a houseful contains one or more households.18
In Brazil this definition is applied to a socioeconomic environment that included large numbers of agregados and slaves. Although often mentioned, the agregado has escaped the detailed study he deserves, unlike the slave who has been subjected to intensive examination. Some dependents and slaves were related to the head of the house and can be considered as members of the household—the primary means of differentiation being personal relationship to the head of the houseful and place of dwelling.19 Most, however, were not related to the houseful head. The image of the Brazilian houseful that emerges from the census data is one of concentric circles, with the head and his immediate family at the center surrounded, first by dependents who are related to the head; then in the next circle, by the nonrelated dependents and some of the slaves; and finally, by the majority of the slaves in the circle furthest removed from the center.
Closely related to the structure of the houseful is its size. Perhaps surprisingly, in place of the large houseful usually depicted as prevailing in colonial Brazil, we discover a houseful which, in terms of size, was not unlike that found in early modern Europe. Furthermore, in this characteristic Vila Rica does not appear to differ from other areas of colonial Brazil.20 The average size of the houseful in Vila Rica was only 5.11.
As suggested in Table I, slaves and dependents comprised significant proportions of the population of Vila Rica. Of all housefuls 44.9 percent contained dependents and 41.4 percent slaves. Retainers represented 18.4 percent and slaves 30.0 percent of the total population. It is significant that 602 households, comprising 35.3 percent of the total, were simple or nuclear housefuls without dependents or slaves.
It is also clear that the concept of a patriarchal head of an extended family is not applicable to the vast majority of the households of Vila Rica. Based on the 1804 census returns, 13.5 percent of all housefuls contained relatives other than children, or three or more generations including the head; and, of the 230 housefuls, 91 were headed by men. Thus only 5.3 percent of all housefuls can be described as being composed of patriarchal, extended families. The identifiable racial composition of the heads of extended housefuls was 16.7 percent white, 44.2 percent mulatto, 5.1 percent cabra, and 37.0 percent black.21 The fact that so many of the extended or multiple housefuls were headed by nonwhites suggests that family members may well have served as economic contributors rather than as proof of status—they fulfilled an economic role rather than reflecting conspicuous consumption.
It is usually assumed that the bonds linking the family together rest upon the institution of marriage. Like the concept of the extended family itself, this is not supported by the evidence from Vila Rica. (See Table II.) Less than one-third of the heads of extended family housefuls were recorded by the census taker as legally married. The low marriage rate of heads of all housefuls is even more significant (Table III). Two-thirds of these men and women had apparently never been married and of these the majority were single women. The appearance of so many women as heads of housefuls, 45.0 percent, is startling and will be examined in greater detail below. It is probable that most of the women still single would not marry: 39.4 percent were 50 years old or older, 60.0 percent were 40 or older, 81.6 percent were 30 or over, and 94.1 percent were over the median age of first marriage—22 years.
While it may appear that 32.6 percent is a low marriage ratio, it is clear that this was still higher than the rate for the population at large (see Table IV). Over 83 percent of the population of Vila Rica over the legal age of marriage was single, a figure which seemingly supports August de Saint-Hilaire’s remark that Rrazil was “a country where there existed extensive repugnance to legitimate unions.”22 Clearly marriage was not the bond that held together the family.
The low incidence of marriage existed despite the efforts of church and state to encourage people to marry. The Portuguese government was interested in marriage as a means of imposing law and order on the turbulent miners of Minas Gerais. As early as 1721, King João V ordered the governor of Minas Gerais to urge “the important residents [pessoas principais] and even others” to marry. Furthermore, he sought the advice of the governor on the advisability of restricting public service to married men.23 The Portuguese authorities hoped that once the miners established roots they would be less inclined to threaten royal control—rebellions against royal authority having occurred in 1708-1709 and 1720. In addition, it was assumed that children raised in a married family household would be even more obedient."24 The suggestion was opposed by the governor on the grounds that there was an insufficient number of women in the mining district.25 The primary thrust of royal policy was to tie down the elite with family burdens; should “even others” be convinced to marry, that would be a bonus.
It is probable that the local elite did not need royal pressure to marry, but rather a sufficient number of suitable women. Marriage had become a symbol of status—an indication of social differentiation. A suggestion of this can be found in the accounts of foreign travelers during the nineteenth century; for example, James Wells described a well-to-do black man who made it a point to introduce his wife and informed Wells “proudly” that they “were married in all due form by the priest.”26 The fact that he had been married legally was used as a mark of status. The same conclusion is also implied by the census data. As is indicated by the 1815 report of population summarized in Table V, there was a greater propensity for whites to marry than there was for either mulattoes or blacks.27 Whites represented a greater proportion of the married than of the population as a whole. While the differences are greater for males than for females, those for the latter are still significant.
This conclusion is supported by an examination of people identifiable as being of high social status. Whereas in 1804 only 16.6 percent of the general population over the legal age was married, 37.1 percent of all women listed as donas, or ladies, and 31.5 percent of all titled males were married.28 Thus two indicators of status—race and title— show some degree of correlation with marriage.
This generally low marriage rate must be viewed against a background of an ecclesiastical policy that allowed people to marry at a comparatively young age. Canon law permitted males to marry at the age of 14 and females at 12 or even earlier, if “possessed of discretion and sufficient disposition.”29 This encouragement of early marriage was probably a response to the necessity for maintaining and expanding the population in the face of a high mortality rate, especially that of infants. Probably for economic reasons, men seldom married at the legal age. Women, however, were more likely to do so.30
This suggestion is supported by an analysis of the census data to ascertain the age of the mother when she gave birth to the eldest child still residing in the household and, thereby, to provide an approximate age at marriage (Table VI). The results should be viewed as maximum figures, because there is no way of ascertaining elapsed time between marriage and the birth of that child, nor of determining whether the eldest child listed was the firstborn (the firstborn could have died in infancy or, conversely, could have survived to set up a separate household).31 The overall average age was 24.1 and the median 23. When a minimum period of nine months (for pregnancy) is assumed, the median age is lowered to approximately 22 years. Even this maximum figure is significantly lower than that reported for most parts of Western Europe during this period.32
Conversely, the age differential of spouses was larger than that reported for Western Europe.33 On the average, husbands were 7.1 years older than their wives, and for blacks the differential was 8.1 years. Some of the disparities were extreme, as in the case of Francisco de Crasto, who was 43 years older than his wife.34 In 28 cases, or 7.5 percent, the variation in age was 20 years or more. On the other hand, in only 52 cases, representing 13.9 percent of the total, was the wife older than her husband, and in 26 others (7.0 percent) she was the same age as he.
If women could and did marry while relatively young, it is not surprising that marriage engagements could legally be made when the participants were as youthful as seven years of age. Engagements were called desposorios de futuro, or promises of future marriage.35 The ecclesiastical records of the Archdiocese of Mariana, in whose district Vila Rica was located, attest to the use and abuse of this provision. A bizarre example involves the successful efforts of Fernando Dias Leite to have the Bishop of Mariana dissolve his promise of marriage to a very young girl so that Leite would be free to marry the girl’s recently widowed mother. Leite freely acknowledged that his sudden shift in affection was due to the widow’s sizeable inheritance.36 Abuses of engagement contracts were legion. Often they were used to entice a young woman to engage in premarital sexual activities, with the prospective groom disappearing after learning of the girl’s pregnancy, or shortly before the wedding day.37
While the church authorized early engagements and encouraged marriages, it also created certain barriers to matrimony that made marriage an ideal attainable only by some. Among these obstacles were such traditional impediments to marriage as affinity within the fourth degree, impotency, consanguinity, and the use of force or false pretenses.38 In order to ascertain whether these or other restrictions existed, engaged couples were required to have their marriage bans read in their parish churches on three successive holy days, so that anyone having derogatory information could report it secretly. If problems were uncovered, as they often were, the couple could then petition the bishop for a waiver of the troublesome provision. The petition process, however, cost money and time, and this probably served to prevent some couples from formalizing their relationships.
Far more important as a barrier to marriage than the church’s traditional restrictions was the great inconvenience resulting from the church’s insistence that positive evidence be presented that the engaged couples were single. For those born and residing in and around Vila Rica the problems which this entailed were minimal. But, for those born elsewhere, the perplexities escalated in proportion to the red tape involved. Thus, for example, the wedding of Domingos Ferreira and Anna Rodrigues da Fonseca was delayed by the insistence of the church that the bans be read in Oporto where the prospective groom was born. Affidavits were required of the parish priest in Oporto stating that the bans had been read and that no negative evidence was presented.39
On other occasions, pending receipt of positive proof of marital status, the wedding ceremony was performed with the proviso that the marriage not be consummated until the evidence was received. An illustration is the case of Manuel da Rocha Perreira and Anna Maria de Jesus. They were married on January 26, 1764, with the stipulation that “under pain of excommunication … they are not to cohabit until they present certification of the bans from the district of Sabará where the groom had resided, [this requirement] having been waived prior to the wedding ceremony by the Illustrious and Most Reverend Canon, Capitular of this Diocese.”40 The restrictions were lifted in such situations, and the couples were allowed to live together only after certification had been received that the bans had been read.41 When there was substantial evidence supporting the contention of an individual that he was single, provision was made for the couple to post a bond with the ecclesiastical authorities, to ensure the satisfactory fulfillment of the provisions of canon law.42 These ecclesiastical restrictions not only functioned to delay many weddings but, because of the costs involved, probably dissuaded others from seeking the sacrament of marriage.
Such provisions were not applied exclusively to the free population. There is substantial evidence that the same requirements for evidence of singleness were applied to slaves also. This was an impossible demand for those slaves born in Africa—a problem resolved by the ecclesiastical authorities by requiring proof of marital status only for the period after Catholic baptism. Thus slaves, either crioulo or African-born blacks, were required to have their marriage bans read in the areas where they had resided. João Batista and Maria, both African-born slaves, were required to have their marriage bans read in Rio de Janeiro and also in Salvador, where they formerly had resided.43 Clearly this requirement was a complication to slaves and ex-slaves in their efforts to marry, and its existence helps to explain the especially low marriage rate among slaves.
While the church hierarchy often insisted upon the fulfillment of all the provisions of church law, local church authorities not infrequently made exceptions for influential residents. Consequently, the requirement that the bans be read publicly prior to the wedding ceremony was waived occasionally. In this regard the church was inclined to exercise a wide range of powers, even to the point of disregarding the wishes of the parents. An example of this exercise of power was the suspension of the public reading of the bans of Dr. Manuel Rodrigues Pacheco de Morais and Clara Maria do Pilar until after the wedding. This ruling was made out of deference to Dr. Morais’s fear that the family of the bride would react violently to the news, perhaps even preventing the wedding from taking place.44 On still another occasion bans were not read publicly because the husband, a Portuguese-born militia officer, wished to marry his slave without notoriety.45 Thus if the petitioners were sufficiently influential, the church was willing to allow exceptions.
The activity of the church authorities in waiving impediments to matrimony was not limited to dispensing with the public reading of the bans. The church often exercised its power to legitimize situations caused by such illegal activities as kidnapping and seduction. Both of these actions were means of presenting the authorities—parental, civil, and ecclesiastical—with a fait accompli. This also was a way around many of the difficulties inherent in the process of getting married and probably absolved the father from paying a dowry. The church had no alternative but to authorize the marriage, if petitioned, since to do otherwise would have been to punish the female victim of these acts. A clear statement on the ramifications of the act of seduction was made by Antonia Martins Cerqueira: “The petitioner finds herself dishonored by her co-petitioner, for which reason if she does not marry him, she will be villified, with neither anyone wishing to marry her and thus being without redemption [desempenhada], nor being able to support herself because she is abjectly poor.”46 Kidnapping had the same impact, since the news soon became “public and notorious gossip,” thereby preventing the victim from finding a husband “of her station in life.”47 These examples demonstrate how crucial was the “purity” of the woman; once having been “dishonored,” the woman was forced to marry her abductor or suffer loss of status and the opportunity to marry a social peer. Furthermore, it points to the strength of the traditional norm that a bride had to be a virgin at the time of marriage: the woman either married her abductor or lived a single life. No such restrictions, of course, were placed on the male. The same ethical considerations often forced the church to waive impediments arising from blood relationships within the fourth degree or spiritual relationships emanating from godparentage (it was assumed that godparents of the same child held a bond as strong as that imposed by blood relationships). Marriages between relatives, familial or spiritual, were common, and many of the petitions made to the bishop of Mariana fell into this category.
Beyond the aspects of ecclesiastical policy that served to restrict marriage, the traditional importance of the dowry played a similar, albeit harder to define, role. The custom of providing a dowry was common in eighteenth-century Minas Gerais, the amount depending on the social position of the family and the status of the child—legitimate or illegitimate.48 Many individuals included in their wills provisions for giving money to orphans to be used as dowries.49 Furthermore, this pressure was so great that some of Vila Rica’s poorer residents found themselves compelled to solicit the municipal council for assistance. For example, one “abjectly poor” father asked the council for a small sum to be used as a dowry for his daughter who was “so poor and miserable [that she] did not have clothes in which to appear [in public] or other necessities for her personal use.”50 The efforts that were made to obtain governmental assistance imply that marriage was still a norm of the society. It was, however, a standard attainable only by some, due to the limitations imposed both by ecclesiastical restrictions and social customs.
The impact of these social and ecclesiastical requirements was to make marriage difficult for many people. But by themselves they probably would not account for the low marriage rate, for the demographic situation was favorable for marriage. There was a major sex imbalance with 79 males for every 100 females. Thus eligible females were available for marriage. It seems very probable that the economic recession, which was into its second half-century made the responsibilities of marriage too onerous for many men. This probably served to restrict marriage only to those who could afford it. Perhaps because of these difficulties, marriage formed the bonds holding together only about one-third of the housefuls of Vila Rica.
Housefuls varied significantly in size. The greatest range is found among female-headed housefuls. Widows headed housefuls that averaged 7.56 persons, while those headed by single women averaged only 4.13—reflecting generally higher economic status. While the typical houseful headed by a male was larger than that headed by a female, women who were or had been married headed larger housefuls than males who were or had been married—7.45, as compared to 6.75. (See Table VIII.) The difference is due, in part, to the large number of slaves owned by married houseful heads—reflecting generally higher economic status—and the number of children still residing in the household. The average size of the married male’s houseful was 6.84, while that of the female’s houseful was only 4.69.
The occupational characteristics and age of single and married male heads are set out in Tables IX and X. There was a greater tendency for miners, farmers, soldiers, and haulers to be married than for other men, although only in the last category was the difference appreciable. The employment rate among married men was slightly higher than for single men, 85 percent versus 81 percent, again emphasizing the correlation between economic factors and marriage. In terms of age, the two groups are roughly similar, except that single heads are evenly distributed from 30 to 69 years of age, whereas the ages of married males evidences a bell curve, with the peak coming between 40 and 49 years. The rapid decline of the number of married males 50 years or older is a reflection of high mortality rates, as suggested by the dramatic increase of the number of widowers aged 50 or older.
The traditional view of the patriarchal family obviously applies to these housefuls headed by males. Custom and law gave the male head extensive powers over all the members of the household. By law he could punish “his servant, or apprentice [discípulo], or his wife, or his child, or his slave.”51 More importantly, however, the dominance of the male head was rooted in custom. It is reflected in the contemporary literature, such as Antonil’s comparison of the father to a herdsman who must keep his cattle and horses penned to prevent straying.52 It also is reflected in many facets of everday life. A married woman could not join the status-conveying brotherhood of Saint Francis of Assisi without her husband’s approval and signature, the same restriction applying to dependent children.53 Time after time a husband who named his wife as executor of his estate or tutor of his minor children felt compelled to justify his decision in terms of her special competency.54
But this discussion of the institution of marriage and of households dominated by married males should not obscure the fact that ecclesiastical bonds of matrimony were a reality for only 16.6 percent of the population over the legal age and for only 32.6 percent of the heads of housefuls. Marriage, while existing as the norm of the society, was not relevant to the lives of the great majority of people.55 The internal dynamics of a society whose economic infrastructure, in the course of one century, had undergone tremendous fluctuations had created a different pattern of social organization.
Numerically the most significant of the family units was the matrifocal family; women headed 45.0 percent of all housefuls.56 Of these housefuls, 83.1 percent were headed by women who, as far as the census taker recorded, had never been married. Should this phenomenon be found to apply generally for colonial Brazil, a reevaluation of the basic nature of colonial society will be in order.57
In terms of size, the matrifocal houseful is comparable to that headed by males, except for the number of slaves. Whereas only 34.8 percent of all slaves resided in matrifocal housefuls, 47.4 percent of all dependent children and 52.0 percent of all agregados did so. Looked at in another way, the matrifocal houseful averaged 1.33 children, while the patrifocal averaged 1.21; for agregados the figures are 1.20 and 0.91. However, there were only 1.17 slaves per matrifocal houseful, as opposed to 1.92 per patrifocal houseful. This difference, plus the presence of the spouses of the married males, results in a mean houseful size that is significantly larger for patrifocal than for matrifocal housefuls—5.46 compared with 4.69.
It is not surprising that this structure evolved. The barriers to marriage were great enough to turn this sacrament into a symbol of status differentiation, despite a sex ratio, as noted earlier, of 79 males to every 100 females. Long gone were the days when, as in the early eighteenth century, the low ratio of married couples could be attributed to a shortage of women.58 The ensuing surplus of women probably encouraged promiscuity and created the conditions leading to the evolution of numerous matrifocal households. The process was apparently encouraged by the relative lack of jobs for men, while domestic employment for women was more readily available. Furthermore, as the economy entered a period of recession, it is probable that spatial mobility increased, with men moving on to other gold fields or to the cattle and coffee areas of southern Minas Gerais. Women remained as the stable elements of the population.
The racial structure of the female heads of housefuls is significant. Of those whose race is known, only 10.5 percent were white, with mulattoes and blacks accounting for the remainder. Even when all those whose race is unknown are considered as white, the total of whites is still only slightly over one-third. Thus, there is no question that the vast majority of the women who headed housefuls were nonwhite. In support of the hypothesis that marriage served as a means of status differentiation, it is also important to note the disproportionate share of whites who were, or had been, married.
The implication that the large number of female heads generally represented a social grouping characterized by lower-class standing is supported by the very low number of female houseful heads identified as ladies or donas. Only 40 of these women were so labeled, comprising only 5.2 percent of the total. Fifteen of these were, or had been, married. Thus, only 25 women who were titled and single were acting as heads of housefuls.
The housefuls headed by women show significant variations in size. These variations follow the racial identity of the heads, with the largest housefuls headed by whites and the lowest by African-born blacks. To a large degree this was obviously a reflection of the differences in economic status demonstrated by the ownership of slaves. When the slaves are removed from the data, housefuls headed by whites are surpassed in size by those headed by mulattoes and crioulos, Brazilian-born blacks.
The large number of households headed by women also suggest the need to reevaluate the role and status of women in traditional Brazilian society. The traditional view was fostered by many of the travelers who crisscrossed Minas Gerais in the nineteenth century. Time after time they depicted the women of the house scurrying away in embarrassment, or spying and giggling at the strange customs of the traveler.59 Occasions when these travelers were able to meet, dine, or talk with women were exceptional.60 But these travelers generally visited representatives of the local aristocracy; seldom were they forced to reside with members of the lower strata, and even more rarely did they seek to describe the social configuration of these households. Clearly their view of mineiro society was one-sided. The stereotype of women as mere shadows of their husbands, involved only with children, cooking, church, and the supervision of domestic slaves is only partially true.
Customs that had governed the behavior of women in Portugal were modified in Brazil. This modification is illustrated by a confrontation between the Portuguese Overseas Council and a bishop of Rio de Janeiro who, in the early eighteenth century, prohibited women from going outdoors after dark. The bishop was chastized by the Council as well-meaning but badly informed about the situation in Brazil. The Council contended that it was necessary for poor women to do their shopping and water-carrying after dark. More importantly, the bishop was urged not to equate immorality and sinfulness with going out after dark.61 Women in Brazil could not be forced to remain in the seclusion to which their Portuguese sisters were accustomed.
Many women in Vila Rica worked to help sustain their households. Much of their work was domestic, such as weaving, sewing, and washing clothes, but a significant number of women were shopkeepers and gold prospectors. Women owned about 40 percent of all the shops in Vila Rica and comprised at least one quarter of the prospectors and more substantial miners. But the majority of women who headed housefuls either had no occupation listed or had marginal jobs, implying that resident children and agregados were important economic supports of the household.
The implications of the widespread existence of the matrifocal household are far reaching. What was the impact this structure had upon the identity formation patterns of children? Many were raised without the presence of a father (34.6 percent of the total children) while fathers of others were not present because of abandonment or death (12.7 percent).62 One possible implication is suggested by the work of Roger Burton and John Whiting into the process of identity formation in matriarchal societies. They suggest that in societies where the child sleeps with or near the mother, the mother becomes the identity model since, by controlling the resources wanted by the infant, she becomes the subject of envy and then of emulation. Girls raised in this environment, Burton and Whiting contend, follow the lifestyle of their mothers. Boys become overly dependent, infantile, and adopt characteristics socially defined as feminine, until they come under the influence of older boys and men, at which time they seek to obliterate the primary female identity by rejecting all feminine influence and behaving in an exaggerated masculine manner.63 This hypothesis is highly suggestive for an understanding of the dynamics of identity formation in colonial Vila Rica. In a society where a significant number of all children were raised in matrifocal homes, presumably with little opportunity to identify with males until old enough to participate in the male-dominated world beyond the home, the pyschological impact of the absence of the father must have been important. The conflict between a primary female identity and the patriarchal environment might help explain the exaggerated masculinity known as machismo.
Another effect that the large number of these households had on the social system was to create a class of people who were technically illegitimate. About 43 percent of the total number of dependent children resided in households where the parent of record had never been married. The existence of so many illegitimate children necessitated the evolution of a system flexible enough to blur the differences between the legitimate and the illegitimate and to integrate the latter into the society without creating major dysfunctions. Accommodations were, in fact, made on both local and imperial levels. Charters legitimatizing children were available for a price. These were probably of greater importance to the elite, since through them the children were allowed to inherit the title and entailed estate of the parent.64 Certainly the charters were of little importance to the majority of the population. For the majority of the illegitimate children questions of inheritance were resolved by the wills of their parents. Many wills contained provisions for “natural” children.65 In other wills parents recognized the paternity of children living in matrifocal households.66 But the society confronted by so many illegitimate children had been forced to modify other rules. Even the priesthood was not closed to illegitimate children, since this restriction could be waived by the proper authorities.67
In addition to those children whose parents were not married, however, there were a number of children whose parents were unknown. Called enjeitados, these were children who had been abandoned by their parents. Responsibility for them rested with the municipal council, which often paid foster parents to raise the waifs. Because of the large number of enjeitados, this had become a considerable expense for the council.68 The 1804 census listed 146 enjeitados. Enumerated among the agregados, they comprised 8.7 percent of all agregados and 6.1 percent of dependent children.
The council had, on occasion, reacted to this situation, going so far as to call the mothers of the enjeitados prostitutes.69 Presumably included among these so-called prostitutes were the mistresses maintained by the well-to-do; some mothers were forced to abandon newly born children by the exigencies of their status.70 It seems probable that the word “prostitute” was used by the town fathers to deal with a social structure—the matrifocal family—which they perceived as immoral. To limit the practice of abandoning children, the council appointed sworn inspectors who were responsible for ensuring that pregnant women in their districts did not desert their children. The inspectors were enjoined to be especially watchful of mulatto and black women, since they were viewed as more likely to abandon their children.71 The abandonment of children was viewed as being connected with a specific social class—the poor and nonwhite.
The church also was active in trying to curb this practice by curtailing concubinage. The most effective tool in its arsenal was the ecclesiastical visitation, from which judicial action (including exile) could arise.72 The visitation was an effort to impose on the parish priest the values of the church hierarchy. Some visitations were meticulous and forceful, especially during the first half of the eighteenth-century. For example, one of these ordered that mistresses, “even if useful for running the household,” were to be ejected from the home and all relationships with mistresses outside of the home were to cease. To enforce these orders, priests were instructed not to confess those men supporting mistresses nor accept as parishioners couples who could not prove that they were married.73 The policies of the church paralleled those of the state: the church tried to root out extramarital relationships (while not really changing the prerequisites for marriage), and the state, attempted to avoid the expenses resulting from these relationships (without attacking the institution of concubinage itself).
The family structure in the region of Vila Rica during the late colonial period is more complex and diverse than that usually portrayed for Brazil. The traditional patriarchal, extended family, based on coresidential patterns, is found in relatively few housefuls. Far commoner were nuclear and matrifocal families, with the large number of the latter emphasizing the status-differentiating function of marriage. The role of women emerges from this analysis as having been very significant, both in numerical and social terms. If the conclusions reached for Vila Rica are, in fact, generalizable, then the hypothesis that the extended Brazilian family subsequently disintegrated under the twin onslaughts of industrialization and urbanization is not valid. The extended family as a common residential unit has not disappeared—it never existed. It may be more correct to state that modernization has led to the increase in the number of nuclear families and a decrease in mono-focal families.
Gilberto Freyre, Casa-grande & senzala: formação brasileira sob o regime de economia patriarchal, 14th ed. (Rio de Janeiro, 1969) 2 vols. Gilberto Freyre, Sobrados e mucambos: decadência do patriarcado rural e desenoolvimento do urbano, 4th ed. (Rio de Janeiro, 1968) 2 vols.
Freyre, Sobrados e mucambos, I, xlvii.
Fernando de Azevedo, A cultura brasileira, 4th ed. (São Paulo, 1964), p. 513; João Camillo de Oliveira Torres, Estratificação social no Brasil (São Paulo, 1965), p. 35; João Camillo de Oliveira Torres, Historia de Minas Gerais (Belo Horizonte, n.d.), II, pp. 487, 519; L. A. Costa Pinto, Lutas de familias no Brasil: introdução ao seu estudo (São Paulo, 1949), pp. 46-50.
T. Lynn Smith, Brazil, People and Institutions, rev. ed. (Baton Rouge, 1963), p. 459.
Antônio Cândido, “The Brazilian Family,” in Brazil: Portrait of Half of a Continent, ed. T. Lynn Smith and Alexander Marchant (1951; rpt. Westport, Conn., 1972), pp. 303-304.
Ibid., p. 292.
Thales de Azevedo, Social Change in Brazil (Gainesville, Fla., 1963), pp. 17-18.
Ibid., p. 17.
Emilio Willems, “Structure of the Brazilian Family,” Social Forces, 31 (1953), 339-346.
Luis Lisanti and Maria Luiza Marcilio, “Estrutura demográfica, social, e econômica da vila de Lajes, 1798—1808, Estados Historíeos, 8 (1969), 9–52. Maria Luiza Marcílio, “Tendances et structures des ménages dans la capitainerie de São Paulo (1765-1868) selon les listes nominatives d’habitants,” in L’Histoire quantitative du Brésil de 1800 a 1930, ed. Frédéric Mauro, Colloques Internationaux du Centre Nacional de La Recherche Scientifique (Paris, 1973), pp. 157-165; Maria Luiza Marcílio La ville de São Paulo: peuplement et population, 1750-1850 (Rouen, 1968).
While the economic recession that struck the mining district probably resulted in major demographic movements within the captaincy as a whole, these do not appear to have resulted in a substantial change in the population of Vila Rica. The population reports for the parish of Ouro Prêto during the period 1796-1799, for example, show the population remaining constant (5639, 5635, 5635) and the free male population actually increasing slightly (1569, 1577, and 1586), Mappa da população, Ouro Prêto, 1796-1799, Arquivo Público Mineiro (cited hereinafter as APM), Belo Horizonte, Planilhas 20364 and 20357. It appears that the primary effect of the later coffee boom was to drain slave labor away from Vila Rica; the free population increased significantly during the first quarter of the nineteenth century, from 6385 in 1804, 6637 in 1815, 6867 in 1818, to 7762 in 1823, Mappas da população, Tenno de Vila Rica, 1815 and 1818, in APM, Maço da População; Mappa Estadístico do Termo da Imperial Cidade do Ouro Prêto, 1823, Arquivo Nacional, Rio de Janeiro (cited hereinafter as ANRJ), Cód. 808.
This census has been transcribed and published by Herculano Gomes Mathias, Um recenseamento na capitanía de Minas Gerais: Vila Rica—1804 (Rio de Janeiro, 1969). This article is based on an analysis of all those enumerated by the census takers.
Discussions of household typologies often focus on the importance of the “conjugal link” in defining the family. The use of this link may be relevant for European societies where the majority of the heads of households were or had been married. Peter Laslett, “Mean Household Size in England since the Sixteenth Century,” in Household and Family in Past Time, ed. Peter Laslett (Cambridge, Eng., 1972), p. 147. It is not, however, as relevant to a study of Vila Rica, where only about one-third of household heads were married. Because of this low marriage rate, I have included within the definition of the family, groupings of a single parent and children. Marriage and/or paternity are the basis of my working definition of the basic family unit.
Only in the district of Ouro Prêto were records of occupations consistently provided for slaves.
Raphael Bluteau, Vocabulário portuguez e latino, 8 vols., (Coimbra, 1712-1721) IV 28. João Antonio Andreoni (Andre João Antonil), Cultura e opulência do Brasil (São Paulo, 1967) p. 142. Antonil, in lumping together “family, sons, and slaves,” suggests that this definition was used in Biazil.
Suzanne Chantal, A vida quotidiana em Portugal ao tempo do terremoto, trans. Alvaro Simões (Lisbon, n.d.), p. 105. Chantal defines the family as all those belonging to the house, whatever their status. The very term used to designate servants was significant. They were called criados, meaning people raised in the family and comprising a part of it.”…
Phelippe Ariès, in his study of the family and childhood in France, defined the family as including the conjugal unit, servants, friends, and protégés: Phelippe Ariès, Centuries of Childhood: A Social History of Family Life, trans. Robert Baldick (New York, 1962), p. 393.
Peter Laslett, “Introduction,” in Household and Family, pp. 23-41.
The English traveler James Wells noted that a fazendeiro dined surrounded by his family and countryfolk, undoubtedly agregados, who “according to their position, sit above or below the salt, as in old feudal days.” James Wells, Exploring and Travelling Three Thousand Miles through Brazil from Rio de Janeiro to Maranhao (London, 1886), I, 202-203.
Laslett’s data for 100 English communities from 1574 to 1821 show a simple mean household size (hereafter MHS) of 4.841, with a range of 3.63 to 7.22; when the Vila Rica data are recomputed using the household rather than the houseful, the simple MHS are almost identical—4.841 for England and 4.888 for Vila Rica, Laslett, Household and Family, p. 133. While there are some variations reflecting the rural-urban dichotomy, the houseful size in Vila Rica was not unusually low for Minas Gerais. An examination of 45 different localities within Minas Gerais during the first one-third of the nineteenth century reveals a range of houseful size of 3.30 to 9.84, with the median being 6.15. Termo de Marianna, 1831; Mappas da População, Termos de Sabará, 1815, Queluz, 1831, and Itaverava, 1831, APM, Maço da População; Mappa Estadístico do Termo de … Sabará, 1823; and Mappa Estadístico do Termo de … São Carlos de Jacuí, 1823, ANRJ, Cód. 808. Neither this range nor the mean houseful size (hereafter MHFS) of Vila Rica itself appears unusual in comparison with other areas of Brazil. The MHFS in the province of Santa Catharina in 1830, for example, varied from 6.26 for the capital to 3.94 for the other districts; the average for the entire province was 5.14, Mappa da População da Provincia da Santa Catharina, 1830, ANRJ, Cód. 808. The 15 districts comprising the capital of Brazil and the province of Rio de Janeiro exhibited an average houseful size of 8.63, with a range of 6.94 to 10.32. Mappa… da Cidade e Provincia do Rio de Janeiro, 1821, ANRJ, Cód. 808. Houseful size in the province of São Paulo was very similar to that of Minas Gerais, fluctuating from 5.9 to 6.6 during the years 1798-1836. Marcílio, “Tendances et structures des ménages,” p. 159.
Of the 230 heads of housefuls, race was given for 138. It is important to note that the census identifies only resident relatives and provides no information on relatives residing in nearby but separate housefuls. Nor can the census data identify those families that had passed through an extended family experience. The data provide only a static view of a dynamic situation.
August de Saint-Hilaire, Viagem pelas provincias de Rio de Janeiro e Minas Gerais, trans. Ciado Ribeiro de Lessa (São Paulo, 1938), I, 160.
Royal Letter to Count of Assumar, Mar. 22, 1721, in Elmar G. Quiroga, “O valor sociológico de um documento,” Revista do Instituto Histórico e Geográfico de Minas Gerais 2(1946), 144, and APM, Cód. 23(SG), fol. 6.
Ibid.
Assumar’s response appears in Feu de Carvalho, Primeiras aulas e escolas de Minas Gerais,” Revista do Arquivo Público Mineiro, 24(1933), 350-351.
Wells, Exploring, I, 84-85.
The 1815 population summary is used, rather than the 1804 census, because it racially defined the entire population.
Forty women were listed as “donas” and, of these, fifteen were married. There were 214 titled males and, of these, 85 were or had been married. Priests were excluded from the data.
Constituições primeiras do arcebispado da Bahia (Lisbon, 1765), Livro 1, Titulo 64, 117.
This is supported by the initial results of a long-range family reconstruction the author presently has underway. Of a small sample from the third quarter of the eighteenth century, 21 cases yielded precise information on the age of one or both spouses. The age of first marriage was calculated for 16 women, who ranged in age from 12 to 27. The mean age at marriage was 17.4, and the median was 17. The age at marriage for 6 husbands ranged from 19 to 30, with the mean being 23.3 and the median, 22. While this is too small a sample upon which to draw conclusions, it is indicative of the youthfulness of women at the time of marriage. The sources used for the family reconstruction include the registers of baptisms and marriages of the parish of Antonio Dias.
On the other hand, there is no way to account for those cases in which a woman married a widower, and no notation was made differentiating natural from stepchildren.
Pierre Goubert has shown that in eighteenth century France the average age of mothers was 26-27 at the time of giving birth to a first child. Laslett has described English women in the seventeenth and eighteenth centuries as marrying at an average age of 24. Vilaricano women, therefore, married at least 2-5 years before their European counterparts. Of special significance is Goubert’s view that people under 25 could exist with security only within the family. In Vila Rica, 8.1 percent of all heads of housefuls and 20.1 percent of wives were 25 or under. Pierre Goubert, “Legitimate Fecundity and Infant Mortality in France during the Eighteenth Century: A Comparison,” Daedalus, 98 (Spring, 1968), 593-603; and Peter Laslett, The World We Have Lost (New York, 1965), pp. 82-83. In Laslett’s study the median age of women at marriage was 22 years, 9 months.
Goubert contends that in France men married at an age that averaged 2-4 years older than women, Goubert, “Legitimate Fecundity,” p. 594. The age differential in seventeenth-century England was 3 years, Laslett, The World We Have Lost, pp. 82-83.
Mathias, Um recenseamento, p. 191.
Constituições primeiras, Livro 1, Titulo 63, 115.
Petition of Fernando Dias Leite, 1765, Arquivo da Curia de Mariana, Mariana, Minas Gerais (cited hereinafter as ACM), Casamentos, No. 495. The petition was approved.
The petition of Isabel Antónia Maciel, July 16, 1748, ACM, Casamento Avulso, is a good example of the abuses that frequently accompanied engagements.
Constituições primeiras, Livro 1, Titulo 67, 124-126.
Marriage Petition of Domingos Ferreira, Mar. 18, 1795, Arquivo da Paroquia de Ouro Prêta, Ouro Prêta, Minas Gerais (cited hereinafter as APOP), Avulso No. 95.
Marriage Registration of Manuel da Rocha Perreira and Anna Maria de Jesus, Jan. 26, 1764, Arquivo da Paroquia de António Dias, Ouro Prêto, Minas Gerais (cited hereafter as APAD), Livro de Casamentos, fol. 218.
Marriage Registration of João Ferreira da Silva and Teresa Jacinta de Jesus, June 22 and July 9, 1759, APAD, Livro de Casamentos, fols. 189 and 190.
Petition of Manuel Carlos de Abreu Lima, Apr. 20, 1773, APOP, Avulso No. 97.
Petition of Joao Batista and Maria Obu, Jan. 8, 1728, ACM, Casamentos, No. 42.
Petition of Dr. Manuel Rodrigues Pacheco de Morais, 1763, ACM, Casamentos, No. 466.
See Marriage Registration of Captain Manuel Correa de Paiva and Jacinta de Barros, Aug. 4, 1737, in Geraldo Dutra de Morais, Historia de Conceição do Mato Dentro (Belo Horizonte, 1942), p. 215.
Petition of Manuel Ferreira Marques and Antônia Martins Cerqueira, Sept. 20, 1749, ACM, Casamentos, No. 337.
Dispensation granted Manuel Francisco Dias and Josefa Maria de Jesus, Mar. 15, 1749, ACM, Cód. 12, fols. 136v-137.
See will of Salvador Fernandes Furtado, May 24, 1725, in Felix Guisard Filho, “Bandeirantismo Taubateano: testamentos e inventarios,” Revista do Arquivo Municipal de São Paulo, 17(1935), 41-46.
E.g., will of Antonio Fernandes Braziella, 1742, Arquivo do Patrimonio Histórico e Artístico Nacional—Ouro Preto (cited hereinafter as APHAN), Cód. 93, No. 1198.
Petition of Hieronimo da Cunha, July 2, 1749, APM, Documento Avulso.
Ordenações Filipinas, Livro 5, Titulo 35, Para. 1 in Pinto, Lufas de familia, p. 186.
Antonil, Cultura e opulencia, p. 166.
Institutos da Irmandade de São Francisco, Livro 26, Capitulo 26. Arquivo da Irmandade de São Francisco, Ouro Prêto.
E.g., Will of Amaro de Araújo, Nov. 5, 1752, APHAN, Cód. 300, No. 6481.
It is possible that some part of the society recognized alternative forms of marriage beyond the church ceremony. However, this would not appear to be very general for male-headed housefuls. There were 534 cases in which the head was a single or widowed male. Of these, only 64—12.0 percent—had residing children; and 183—34.3 percent—resided completely alone, or only with resident children. Clearly in these cases no consort was present. In 124 cases—23.2 percent—the male resided with slaves and/or resident children. Given the nature of the society, it is unlikely that long-term, socially accepted relationships existed between master and slave. In fact, wills and testaments suggest that the more common practice was to free the female slave and provide her with clothing, jewels, a house, slaves, etc. More likely was the establishment of non-jural ties with persons listed as agregados. This was possible in 102 cases where one or more eligible women were present. This is a maximum figure as it includes a number of instances where there is a low probability of the existence of an accepted conjugal unit. Even so, non-jural conjugal units were possible for only 18.7 percent of all eligible male houseful heads and 6.0 percent of all housefuls.
Matrifocal is used to refer to a “women-centered” houseful; that is, a houseful whose head was recognized by the census taker to be a woman. This definition does not preclude the presence of a male serving in the role of “husband” and/or father, but it assumes that the census taker was reflecting the community’s view of the role of the female head. Obviously, had the enumerator assumed that only a male could head a houseful then the female would have been listed as an agregado, and the houseful characterized at patrifocal. That the census taker listed so many women as houseful heads precludes coincidence as a general explanation.
Significantly, parishes as diverse as rural Casa Branca in Minas Gerais and urban São Pedro in Salvador, Bahia, show similar patterns: The proportion of women heading housefuls was 37.3 percent and 41.5 percent, respectively, Relasam dos Moradores da Freguezia de Santo António de Casa Branca do Termo de Vila Rica da Capitanía de Minas Gerais, Aug. 12, 1804, AN, Lata 130, Pasta 2. Avelino de Jesus da Costa, “População da cidade de Baía em 1775,” in V Coloquio Internacional de Estudos Luso-Brasileiros, Actas (Coimbra, 1963), Vol. I, pp. 191-286.
Feu de Carvalho, Ementário da história de Minas, Felipe dos Santos Freire na sedição de Vila Rica—1720 (Belo Horizonte, 1930?), p. 21; and especially see Lourenço de Almeida to João V, undated, in Feu de Carvalho, “Primeiras aulas,” 351.
E.g., Saint-Hilaire, Viagem, I, 14, 142; Wells, Exploring, I, 302, and II, 2-4, 14, 27, 66, and 254; Richard F. Burton, Explorations of the Highlands of the Brazil with a Full Account of the Gold and Diamond Mines (London, 1869), I, 248.
Wells, Exploring, II, 20; Saint-Hilaire, Viagem, I, 248; John Mawe, “Viagens ao interior do Brasil particularmente aos districtos do ouro e do diamante, em 1809-1810,” in Collectanea de scientistas extrangeiros, ed. and trans. Rudolfo Jacob, (Belo Horizonte, 1930), I, p. 20.
See Consulta of the Overseas Council, Sept. 4, 1703, in Documentos históricos, pp. 93, 158-159.
It is possible that for some of these matrifocal housefuls, the father was listed as an agregado. It would appear, however, that the existence of many such situations is improbable since the structure of the census was based on authority rather than on economic status.
Roger V. Burton and John W. M. Whiting, The Absent Father and Cross-Sex Identity,” Merrill-Palmer Quarterly of Behavior and Development, 7 (April, 1961), 85-95.
Royal Carta de Legitimação, May 31, 1740, APM, Colonial Section, Documento Avulso.
E.g., will of Antonio Gonçalves de Araújo, Jan. 28, 1756, APAD, Livro de Obitos, No. 3, fols. 262v-263.
Will of Joao Barbosa de Amorim, Dec. 3, 1771, APHAN, Cód. 335, No. 7041.
For example, António Leite Esquerdo petitioned the bishop of Mariana in 1779 to have two obstacles to his ordination waived. The impediments were “mulatismo e ilegitimidade.” The petition was approved. Petition of A. L. Esquerdo, Aug. 21, 1779, ACM, Habilitações, No. 199.
During the 1758 correição conducted by a royal judge, the council complained of “the excessive number of enjeitados, whites as well as mulattoes and Negroes. …” Correição, Dec. 31, 1758, APM, Cód. 22 (Câmara Municipal do Ouro Prêto, cited herein after as CMOP), fol. 139.
The council asked all citizens to report cases where known "public prostitutes … resorted to the abominable practice of abandoning their children.” Council Ordinance, Mar. 2, 1763, APM, Cód. 77 (CMOP), fols. 240v-241.
An example of this is the accusation made against a priest charging that he had lived with a woman for years and had “children by her, sending her out of the hamlet when it was time to give birth, and [then] abandoning the children in different parts.” The priest was found innocent, but the charges were not viewed as outlandish and were investigated. Sentence of Padre António Vieira de Matos, Apr. 15, 1750, ACM, Livro que ha de servir para registrar todos os mandados sentenças e Ordens pertencentes ao foro contenciozo, fol. 10.
Correição, Oct. 24, 1761, APM, Cód. 22 (CMOP), fols. 142v-143.
A mulatto freedwoman was exiled for one year after being accused of being a “public and scandalous prostitute” who took part in “lascivious dances and diabolical batuques,” and of having become the mistress of a resident. Registration of the Condemnation and Exile of the Defendant Rita de Oliveira, Dec. 1, 1751, ACM, Cód. 10, fols. 50v-51.
Visitation, Apr. 26, 1727 in Morais, Historia de Conceição, pp. 92-93.
Author notes
The author is Assistant Professor of History, The Cleveland State University. He wishes to express his appreciation to Harry Langworthy HI, Roger Manning, and Robert Wheeler for their helpful comments. Much of the research used in this essay was conducted under the auspices of the Foreign Area Fellowship Program.
