Compadrio in rural Brazil: structural analysis of a ritual institution

Compadrio in rural Brazil: structural analysis of a ritual institution

Antonio A. Arantes

ABSTRACT

This structural analysis focuses on the compadrio system. The empirical background is provided by observation carried out among sertanejo peasants of Bahia in the late 1960s and by the literature on the Latin American and Southern European variants of this institution. It is mainly concerned with two complementary problems. On the one hand, to draw a model that might represent that institution's elementary structure, virtually present in the variants of this system; on the other, to offer an interpretation of its meaning, by contrasting it with elements of the kinship and marriage systems, and taking in consideration the peasants' religious background. This exercise was inspired by Edmund Leach's Rethinking anthropology and his ideas about the Virgin Birth. Analytical perspectives for further research are suggested.

Keywords: compadrio; Bahia's sertão; ritual kinship; structure; generalization.

RESUMO

Esta análise estrutural focaliza o sistema de compadrio. Sua base empírica deriva de observações realizadas entre camponeses do Alto Sertão baiano no final da década de 1960 e da bibliografia sobre as variantes registradas na América latina e Europa meridional. Ela se ocupa, principalmente, de duas questões complementares. De um lado, desenhar um modelo que possa representar a estrutura elementar virtualmente presente nas variantes desse sistema; de outro, oferecer uma interpretação de seu significado, por meio de contraste com alguns elementos dos sistemas de parentesco e casamento, e levando em conta aspectos da religiosidade desses camponeses. Este exercício foi inspirado pelas reflexões de Edmund Leach em Rethinking anthropology e em Virgin Birth. Sugerem-se novas perspectivas de análise para pesquisas futuras.

Palavras-chave: compadrio; sertão baiano; parentesco ritual; generalização.

SUMMARY

I. Introduction

II. Some theoretical problems

Functions of compadrio

Intra-class and inter-class compadrio

III. Towards a definition of compadrio's elementary structure

Choosing the compadres

Compadrio and the dyadic contract

Compadrio as a system of social relations

Variations

Compadrio of confirmation

Compadrio and marriage

Compadrio of St. John

Compadrio 's elementary structure

IV. Compadrio and the elementary family

Compadrio, filiation and siblingship

Compadrio and affinity

Some ideological bases of compadrio

V. Post scriptum

References

This short essay has a relatively long history. I started studying Brazilian peasants in 1966 when I was a graduate student of anthropology at the University of São Paulo. The late professor Gioconda Mussolini, then my supervisor and a very good friend, helped me in delimiting a geographical area as well as formulating some problems for research among the sertanejos of Bahia.

I spent some weeks in the field in 1967 and a month or so in the beginning of 1968 as research assistant of Dr. E.R. Durham (University of São Paulo) and Dr. D.R. Gross (CUNY, New York) respectively. In June of 1968, I finally went to the field with a research project of my own, sponsored by the Research Foundation of São Paulo (FAPESP) and with the encouragement and theoretical advice of Prof. Gioconda. I stayed in the sertão for nearly 3 months and collected data on social and economic organization of rural neighborhoods, part of which is interpreted in the present paper.

In 1968, sponsored by the State University of Campinas (UNICAMP), I received a scholarship from the FAPESP and spend the next 2 years abroad. In the academic year of 1968-69, I studied with Prof. Yves Gentilhomme at the Centre de Linguistique Apliquée, at the University of Besançon, France. At the end of 1969 I went to Cambridge, where I was a research student working under the supervision of Sir Edmund Leach, at Kings College. He has helped me not only in choosing the subject matter of this dissertation but also, and mainly, in formulating hypotheses and interpreting the data in which it is based. I am deeply indebted to him for his professional assistance and courtesy.

After the death of Prof. Mussolini, Dr. Eunice Durham supervised the final preparation of this paper that was presented in 1970 as a Master's thesis in anthropology at Universidade de São Paulo. Professor Carmen Junqueira de Barros Lima and Dr. Peter Fry made several stimulating criticisms as examiners of this dissertation. Dr. Gross also read the first draft of this paper, as did Dr. Fry and Dr. S. F. Gudeman. P. Scheldrake and his students at Kings have listened to a preliminary version of it in my stuttering English. I would like to express my sincere gratitude to all of them, and particularly to Peter Fry who convinced me to publish this 40 years old academic exercise in totum for the first time.

I. Introduction

The aim of this paper is to formulate some hypotheses for the understanding of compadrio, a ritual kinship system found today mainly in peasant communities of Latin America and Southern Europe. It involves a couple of parents, their child and the child's godparents which I will call in this paper the compadrio group.¹

Most of the works on this topic, published up to the present time, pay little or no attention to the universal aspects of this widespread institution. With the exception of a few really worthwhile writings which have direct implications for a general theory, the great majority can be considered as ethnographical accounts, sometimes perceptive, sometimes naïve, which describe the peculiarities of compadrio in specific communities.

In this paper, I will be a bit more ambitious. I will be mainly concerned with two complementary problems. On the one hand, I will try to introduce more precision and, at the same time, more generalizations into the definition of the compadrio system To do this is, to a certain extent, to construct a model² the anthropologist might use in order to represent its elementary structure, or the system of social relations that is virtually present in its variants. On the other hand, I will examine its meaning, contrasting it with other relevant social institutions.

In the peasant communities that I have visited, the kinship system provides people with a basic chart for classifying their fellow villagers, for defining preferential groups of persons on which they can count for economic or quasi-economic cooperation and trust and, at the same time, the delimiting of exogenous groups and access to landownership. Taking this issue into account, one could ask why people should have built up a ritual kinship system which might be considered, in many respects, a duplication of the ordinary one since it tends to link close kin. I suggest that a clue to a possible answer might be found in the distinctive features that compadrio shows when contrasted to the elementary family as well as in some aspects of the peasants' religious background.

My experience in fieldwork is limited to few months spent in some peasants' communities in the State of Bahia's backlands, locally called sertão. This part of Northeastern Brazil is traditionally occupied by extensive cattle breeding ranches combined with subsistence cultivation areas. Today these activities are side by side with the cultivation of cotton and agave. In the communities where I collected my data (several peasants' neighborhoods in Monte Santo, Jeremoabo, Uauá and Euclides da Cunha municipalities) landownership is based on kinship ties; economic activities are predominantly directed towards subsistence, with a small surplus for exchange and the market. The peasants typically engage in small-scale cultivation of beans, manioc and corn, combined with the breeding of small numbers of cattle, goats, pigs and chickens.

In this paper, I will take the situation found in this area as a paradigm for my argument. Nevertheless, in order to complement these data and to focus the problem from a wider perspective, I will appeal to works already published either on Brazilian or on Spanish American peasant communities.

At the same time, I would like to emphasize here the exploratory character of this paper. This work has no conclusions in the proper sense of the term; it has not been designed to be a demonstration. More systematic fieldwork is necessary in order to verify the validity of my argument.

II. Some theoretical problems.

The evolution of the theories that have been formulated for the explanation of the compadrio system is not relevant from the standpoint of the analysis which I will try to undertake in the following pages. I do not even intend to draw a general outline of their history. Nevertheless, and I believe it will be clear by the end of the present section, a critical review of some of the works on this subject is worthwhile for the formulation of the basic problems focused by my research.

Several anthropological studies have their roots in Tylor's writings; compadrio is no exception. Indeed, it seems that he was the first to note the relevance of these ritual ties for Latin American peasantry in his book about Mexico and the Mexicans (1861: 251). To my knowledge however it was Robert Redfield who first gave the clues for the understanding of of the sociological meaning of this institution.

In his well-known The folk culture of Yucatán (1941: 92-4, 122-4, 212-26) compadrio is only referred to in a few pages as part of his argument about the way that secularization and individualization operate within the folk-urban continuum. Among others, Oscar Lewis and George Foster criticized this theoretical construct. Oscar Lewis (Lewis 1951), who in fact presented a complementary approach to the study of peasant life, showed the necessity for understanding the community within the wider sociological context of the nation; and George Foster (Foster 1953) tried to improve the model of the folk-urban continuum by stressing the distinction between "peasant" and "primitive", which was confused by Redfield who used of one label - "folk" - for both sociological constructs. These criticisms were taken into account in Redfield's later works (Redfield 1953, 1956).

Like Malinowski, however, Redfield's ethnographical accounts suggest more than he explicitly states. If one considers the data and interpretation of life in each of the communities he has studied, one will find at least two very helpful suggestions for the synchronic understanding of the compadrio system.

Functions of compadrio

The first of such suggestions is summarized in the following quotation: "after one marries, and has children one arranges, for each, the padriños of the baptism and of the hetzmek³ and later sees to it that one's son finds the wife. And each of these undertakings to complete the social position of that child for whom one is responsible creates a new tie between oneself and one's wife and the person or the couple chosen to sponsor one's child or with the parents of the child's spouse. Or, at least, it solemnizes and sanctifies a relation of intimacy and trust that has already come into existence through kinship or friendship. Each of these relations is created in ritual and sanctioned by tradition" (Redfield 1941: 124).

Actually, there are two ideas implied in this passage: compadrio as a way of widening someone's network of institutionalized bonds and, at the same time, as a way of "sanctifying" (sic) the already existing social relations. The explanation of these two functions detected by Robert Redfield has been taken further by some of his successors in the study of the social life of Latin American peasants.

Sidney Mintz and Eric Wolf, in an article published in 1950, outline the historical antecedents of present day compadrio and analyse it in terms of some of its functional correlates like social stratification, social mobility and isolation. According to them, the emphasis on the function of the extension or of the intensification of the social ties will depend on the specific functional and historical context in which the institution exists (Mintz & Wolf 1950: 355), but whether widening or sanctifying the already existing social relations, its main function is seen as being that of promoting social solidarity. "By imposing automatically and with varying degrees of sanctity, statuses and obligations of a fixed nature on the people who participate", they argue, "it makes the immediate social environment more stable, the participants more interdependent and more secure" (Idem, idem).

Here starts my problem: when they shift the emphasis of their analysis from these two, so to speak, secondary functions to what they consider a more basic one, namely that of promoting social solidarity through the creation of bonds which are "used" as channels through which a system of material exchanges operates. But let me refer to just one more author before developing my argument.

George Foster argues, in his first monograph on Tzintzuntzan (Foster 1948) that the functional importance of the compadrio system is based on the fact that it both strengthens the position of the individual and is a kind of social insurance on the economic level (because compadres exchange labor, lend money to each other etc.). His argument is further developed into articles on dyadic contract (Foster 1961, 1963) where he emphasizes the importance of the economic aspects of that institution. He argues that in non-indigenous Latin American peasant communities, the "dyadic contract is the underlying principle which gives the social system coherence" (Foster 1961: 1173). "Dyadic contracts are based on the principle of, and are validated by, reciprocal obligations expressed in exchange of goods and services" (Foster 1963: 1281). Godparenthood, although creating a set of formal and explicit ritually sanctioned connections, is seen as one of the institutional frameworks through which the informal dyadic contract operates; at the same time, these ritual ties can be reinforced by a series of implicit dyadic contracts. In fact, he writes, "the system contributes to social stability only when the implicit contract follows" (Foster 1961: 1183).

As I will show on the following pages, Foster's hypothesis on the dyadic contract and other aspects of Mintz and Wolf's analysis are, in fact, of basic importance for the understanding of the functions of compadrio. Apart from the documentary value, one merit of the ethnographies published on the subject is that of proving the validity of such hypotheses. At the same time, one of their weaknesses is derived from Foster's and Wolf's model. In answering the question of how compadrio performs what is seen as the main function of social institutions – namely, that of preserving the social order - they do not pay the necessary attention to the specificity of this particular case. The performance of this function is mainly seen in terms of the solidarity it creates, which is expressed by economic cooperation. It is claimed that compadrio ties are extremely important for labor exchange on which depends the production of peasants' subsistence. Compadres cooperate in the system which might be called of restricted exchange. Nevertheless, this is only one level of the analysis whose relevance relies on the fact that it is one of the most important ways - but not the only one – to understand the structure of this institution. I understand that to explain compadrio in terms of the system of material exchanges that operates through it is, by all means, a simplification. This will never allow the analyst to characterize the distinctive features of this particular institution as opposed to others. This kind of explanation takes the analyst somewhat "out" of the institution he is studying, without allowing him to understand its specificity, the richness of its symbolic content, and emphasizes a function which is also performed by other social institutions such as the family, the neighborhood etc. Furthermore, to validate this explanation by showing how important this mechanism is vis-à-vis a vague and general function such as the maintenance of social stability has even worse consequences; besides providing a banal explanation, it does not stimulate further investigations. I think that an anthropologist must have general hypotheses on the basis of the research, but not so general ones that cannot provide an adequate characterization of the specificities of his object.

I think, however, that it is worthwhile looking somewhat beyond this. As I will show later, compadrio does not link people who are not already related in other ways. Compadres tend to be previously linked through kinship, neighborhood, economic bonds or some combination of these. For this reason, another problem emerges: why, being already related to each other, peasants add to this set of social statuses those of compadre, godparent and godchild. What is the meaning of this?

The analyses of R. Ravicz and Julian Pitt-Rivers (Ravicz 1967; Pitt-Rivers 1968) are far more suggestive from this point of view; to a certain extent, they refer back to the second aspect of Redfield's observation mentioned at the beginning of this section. Among other things, these two authors stress the moral or ideological aspects of the institution and, consequently, of the relations involved. According to Raviz, these relations should be understood as "respect relations" and Pitt-Rivers' argument derives from the following pairs of oppositions: compadrio: kinship :: spiritual regeneration: natural regeneration :: sacred : profane.

From my point of view, the uniqueness of compadrio relations can only be determined when one considers them as a kind of moral contract between adults, that can be understood (1) in terms of the religious meaning of the rituals that establish them, (2) the previous statuses of the people involved as determined by the customary rules for choosing compadres (taking into account the kinship system, the social structure of and the values the peasants attached to them) and (3) the reciprocal expectations which they involve.

Before taking this argument further, I would like to make reference to another hypothesis.

Intra-class and inter-class compadrio

Another suggestion made by Redfield is not so immediately apparent. In Dzitas, he writes, "the selection of the parents is often guided by consideration of wealth or influence and upper-class godparents are designed although such persons may be reluctant to act" (Redfield 1941: 222). Comparing the corresponding situation in the other communities taken into account in this analysis, one finds that Dzitas is a particular case, not an example of a general trend. The author does not establish any explicit theoretical correlation between social stratification and the tendency to select compadres in the upper classes. The fact that another distinctive feature of this community vis-à-vis the other ones brought to comparison is that it is socially and culturally more stratified makes one wonder whether such a hypothesis might not be true. Indeed, Mintz and Wolf seem to be the first to demonstrate this hypothesis. According to their cross-cultural analysis compadrio bonds can be either horizontal (intra-class) or vertical (inter-class). They argue that "such patterning will prove to be determined, not haphazard in character, nor determined solely along continuums of homogeneity-to-heterogeneity, greater to lesser isolation. Rather, they will depend on the amount of socio-cultural and economic mobility, real and apparent, available to an individual in a given situation" (Mintz & Wolf 1950: 358. Emphasis added).

This distinction can be further supported and the whole hypothesis refined if one takes into account George Foster's model of the dyadic contract, specifically the distinction between symmetrical and asymmetrical contracts (Foster 1961). While Mintz and Wolf make explicit the social context of each type of compadrio by pointing out their functional correlates, Foster provides us with a model of their content.

"Depending on the relative positions of the partners, on the kinds of things they exchange," he argues "two basic types of dyadic contracts may be recognized. Colleague contracts tie people of equal or approximately equal socioeconomic status, who exchange the same kinds of goods and services. Colleague contracts are phrased horizontally, and they can be thought of as symmetrical, where each partner, in position and obligations, mirrors the other. Patron-client contracts tie people (or people to beings) of significantly different socioeconomic status (or order of power), who will exchange different kinds of goods and services. Patron-client contracts are phrased vertically, and they can be thought of as asymmetrical since each partner is quite different from the other in position and obligations" (Foster 1961: 1281).

I would like to point out here certain reservations to the application of these hypotheses. Although patron/client compadrio occurs in stratified groups, it is quite unlikely that all compadrio relations of that group would conform to this basic pattern. First, because upper-class parents choose their compadres in their own class; second, because the rules for choosing compadres tend to privilege certain kinds of kinsmen: these people are chosen more because of their position in the kinship system than in the class structure. Furthermore, the usage of the notions of colleague and of patron-client compadrio implies a certain degree of simplification. Both may be all right when one is describing implied economic transactions and interests; but they appear to be not sufficient in the analysis of other levels of exchange like the moral one, for example.⁴

In this short review of the main theses on compadrio I do not intend to present and discuss all the problems studied by all specialists on the subject. My aim has been just to point out the basic available theories and the most important authors concerned in order to show that up to the present time:

The analyses have basically characterized how and to what extent

compadrio contributes to the maintenance of social order;

They have been mainly concerned with the economic underpinnings this institution;

And that some fundamental problems have not been fully solved such as: what it is, why it has the characteristics it does and which of these are distinctive vis-à-vis other social institutions.

In the following pages, trying to take some of these comments further, I will elaborate a few hypotheses which might prove useful in solving the problems I have set out so far.

III Towards a definition of compadrio's elementary structure.

The compadrio system may be loosely defined as a system of ritual relations, mainly but not always, built-up through the Catholic life crisis ceremonies of baptism and confirmation.

It was created in the 19th century, when the Roman Catholic Church, formerly prohibiting parents from being godparents to their own children, declared those who sponsor such ceremonies as the child's spiritual parents. The present form of the system of spiritual kinship as defined by the church has persisted since the 16th century and involves the two baptismal godparents, the godparents of confirmation, the child and his parents.⁵ Apart from prohibiting sexual intercourse and marriage between people who are linked by such bonds, canon law assigns to the so-called spiritual parents the task of being a religious and moral advisor to the baptized (or confirmed) child. On the basis of such bonds, a system of socially sanctioned relations is built up, not only recognizing the sacred link between godparents and godchild, but also ritually connecting the parents of the baptized (or confirmed) child and his godparents.

Today, in most Latin American countries, one finds other ways of creating compadrio which may be occasionally sacred but which are mostly of a secular nature. Examples can be found in the literature. Marriage ceremonies and the child's first haircut on nail trimming, as well as the dedication of an altar or a house, a communal celebration and commercial dealings have been mentioned by several authors as ways of creating compadrio (Mintz & Wolf 1950:354; Pitt-Rivers 1968 : 411; Durham 1966 : 77; among others). Gillin, for instance, in his monograph on a Peruvian community (Mintz & Wolf 1950: 354), has observed not less than 14 different kinds of godparents. In Bahia, besides madrinha de carrêgo (or de apresentação)⁶, who is an extra baptismal godmother, I only came across compadre São João (or de fogueira).⁷

From the point of view of Canon Law, the real forms of compadrio are only those created through baptism and confirmation, the others being nothing but fictions. But from the anthropological point of view, the matter is not so simple. As pointed out by Pitt Rivers "if the authority of custom is accepted, then any form of compadrazgo established on a generally recognized pretext is genuine" (Pitt-Rivers 1968: 412).

One way of dealing with this diversity of ways of creating compadrio bonds would be to try to explain (1) why this redundancy exists, and (2) what remains the same in all these forms in such a way that the social actor can identify them as belonging to the same class of "things", calling all of them by the same name.

I will not discuss the first issue in detail because it would take me somewhat away from my argument. In relation to it, I would only to mention that, according to Mintz and Wolf, this is an evidence of the high social and secular plasticity of the institution (Mintz & Wolf 1950: 357).

The second argument concerns more directly the scope of this paper. To discuss it is, in a way, to think about what I shall call the elementary structure of compadrio and, conversely, to define compadrio's elementary structure is to show that all these forms are variations on the same basic pattern.

I would like to make another remark on these general considerations. Since, as I shall point out below, baptismal compadrio is the most complete form that I have observed in field work or found in the literature, since it is historically the original form of creating such bonds and occupies the highest rank in the peasants' ideology, it will serve as the central focus of attention in this paper.

Choosing the compadres

In the peasant communities that I have visited, parents choose their child's godparents almost immediately after his birth and baptism is performed either in the chapel of the nearest village or preferably of the county seat, as soon as they can afford its costs.

In Brazilian peasant communities (Mello e Souza 1964: 198; Castaldi 1967: 50; Willems 1961: 65; etc) there are customary rules instructing the parents on how many godparents that child should have for each ritual and how to choose them. For the baptism, one godfather, one godmother and one godmother de carrego are needed. After the choice, besides being a friend of the child's parents, the prospective compadre must fulfill to basic requirements, one based on the kinship system and the other stressing, perhaps for practical reasons, the physical proximityof the compadres' households. While some informants say that they choose people living nearby, others say that they usually "give a firstborn to the [paternal] grandparents; the second one, either to [their] brothers-in-law, to [their] brothers or to the mothers' parents".

Although widespread in Brazil, this preference for paternal grandparents to act as godparents of the firstborn is often not strictly obeyed and, instead, either the child's maternal grandparents, an older and married sibling of one of the parents involved, or one of their marriage godparents can be chosen. But in all these instances, they tend to be close relatives, older than the parents, and to have established their own independent household, which means they tend to be full members of the community, both from the material and moral points of view. Two questions arise from this: why should a firstborn have preferential godparents and why should his godparents have the characteristics that they do? I will consider these two points later.

As to the other children - and of course statistically this is the most common case since these families tend to be quite large - there are no preferential godparents. Consequently, these sponsors are taken from outside the group of close kin, and chosen according to degrees of social proximity, of kinship and friendship bonds with the child's parents as well as according to their financial situation. Another informant told me, for instance, that when he kept animals, many people invited him to be their compadre; "I could give their children a lamb or goat", he explained, "but now that I depend entirely on farming, people only asked me to be compadre de São João". It seems impossible, from the data presently available, to determine the relative importance of each of these criteria, but it is certain that the parents' final decision is the result of a complex evaluation of each of the prospective compadres in terms of these elements. The fact that people often choose one of the members of a married couple without necessarily choosing the other corroborates this point of view. Once, asking a man about this, the answer was "simply because we're closer friends to him" [the husband]. So, it seems reasonable to say that each compadre is chosen individually, through careful evaluation made by both the father and mother, based on a combination of social rules, personal preferences and circumstance.

From the 50 cases that I have fully documented, two criteria seem to be particularly relevant: the fact of living relatively near and of being a relative. These cases can be distributed as in Table 1 below.

Six out of the 32 people who are both compadres and relatives have been referred to by my informants just as relatives, without specification of the kinship bond; the others belong to the categories listed in Table 2:

The villages which I visited are typical of one kind of settlement where co- residents are, at the same time, co-heirs of the land on which they live, none of them holding legal title to the portion he occupies and exploits.⁸ For this reason, my data alone do not allow me to make general statements about rural Brazil as a whole. Perhaps the number of compadres which are each other's relatives is probably higher here than elsewhere. For this area, however, it seems that the compadres tend to be chosen mainly from one's own generation and from within the group with close kin and neighbors (60% of the compadres are both neighbors and relatives, whereas 18% come from outside such groups), which means, in this specific case, that they tend to belong to the same social class as the child's parents. Other anthropologists, studying other communities in Brazil, noted this tendency towards choosing close relatives as compadres, mainly in the lower class, but their evidence in this respect is not sufficiently explored.

For other contexts, another important criterion would have to be added to these. In Contestado, in Itá and in the plantation area of the Northeast, for instance, compadres are also selected either from families with a relatively superior economic and political status to the parents (Pereira de Queiroz 1957: 55-66; Wagley 1964: 150-9; Hutchinson 1957: 130-47; etc.).

To certain extent, the Brazilian data support Mintz and Wolf's hypothesis (1950: 364) that "in the cases where the community is a self-contained class, (. ..) compadrazgo is prevailingly horizontal (intra-class) in character. In cases where the community contains several interacting classes, compadrazgo will structure such relationships vertically (inter-class)".

Before discussing this matter in more detail, I would like to characterize the compadre relations as I observed them.

Compadrio and the dyadic contract.

The stress on godparents as individuals has its counterpart in the performance of the rite of baptism where each participant has a specific part to play. Sometimes, the parents themselves do not attend the baptism; they give their child to the godmother de carrêgo. In such cases, she carries the child in her arms from the parents' house to the chapel, accompanied by the other two godparents. This instance can be seen as symbolizing one of the basic functions of compadrio, namely that of providing the child with ritual substitute parents. The godfather is expected to pay the baptism fee and the godmother provides the child's christening gown; at the church, each of these two have their own parts to play. The godmother de carrêgo carries the child back home. A small celebration may take place, in which case the godfather usually invites the father for a drink.

These two instances, namely, the stress on each co-parent as an individual at the moment of choosing, and the existence of specific tasks for each in the ritual, suggest that – although forming a more complex social system - baptism links people together in dyads. This insight is reinforced by the fact that social rules and reciprocal expectations apply to each dyad within the system: father/godfather, father/godmother, mother/godfather, mother/godmother, child/godfather and child/godmother.⁹

Nevertheless, being both ritually and socially attached to the same person (or couple), the various godparents of someone's children do not have any special links between themselves and never act as a corporate group. In fact, this institution does not provide the members of the compadrio group with any specific patterns of behavior or social expectations towards other people also related to them, either through kinship and affinity or through compadrio. I am not suggesting that there are not any differences between the reciprocal behavior of people through which one could trace this kind of indirect bond, and that of people not related at all; people the first group certainly have more opportunities to interact than others, but there are not any specifically institutionalized ways of interacting. One's compadre's spouse children other than the ones which are ritually related to that person remain simply one's compadre's spouse and children; the same applies for one's compadre's compadre or comadre. Two conclusions can be drawn from this. First, as suggested by George Foster, compadrio relations seem to be built up according to the dyadic contract model; second, that there is no transitivity in this system of social relations, strictly speaking.

In his essay on Catulé, a peasant community in Minas Gerais, Castaldi notes that married people inherit "each other's godparents and compadres" (1957: 51-2). To my knowledge, this evidence has not been reported by other ethnographers. Even so, this instance could suggest that compadrio ties are not stricto sensu dyadic, since they could be extended to individuals belonging to the compadre's family. It seems to me that, if this is not a metaphorical usage of the term, this occurrence may suggest (1) how closely compadrio relations are associated with marriage: just as one spouse becomes related to the other's blood relatives, so he/she will be with his/her ritual kin; (2) the relevance of the affinal tie for a group whose basic patterning of the social relations depend, to a large extent, on kinship ties.

Compadrio as a system of social relations

The set of nine ritual connections created by baptism can be graphically represented as in Figure 1. Through these relations, a number of exchanges will take place through time.

These relations are not identical to one another. They vary not only according to the statuses involved but also according to the sexes of the individuals; obviously the relation between godparent and godchild is different from the relation between compadres; the co-fathers' relation is different from the co-mothers' one. This is easily observed at the economic level. Compadres help each other in their agricultural and pastoral work; since the division of labor relies to a large extent on sexual differences, the specific content of cooperation between men will be different from that between women. But even so, there are many instances where compadres ask for the help of their comadres and vice versa. Nevertheless, although the compadres of the same sex a generally more intimate towards one another than with those of a different sex, the formality in the reciprocal behavior, the general expectations of respect and solidarity do not vary. The same is valid for the relation between grandparents and godchildren.

It seems, then, quite acceptable to consider these relations – for the sake of model building – as forming two sets. The F/GF, F/GM and F/GMC relations shown in Figure 1(i) will be considered equivalent to the M/GF, M/GM and M/GMC relations and I will refer generically to them in Figure 1(ii) as the parent/godparent relation (P/GP); similarly the GF/GC, GM/GC and GMC/GC relations will be referred to as the godparent/godchild relation (GP/GC). Let us look more closely to this set of social relations, particularly focusing its moral and economic facets.

Godparent/godchild relation

As I already pointed out, godparent and godchild usually belong to different and, in most cases, successive generations. The elder is ideally expected to provide for the younger's religious education, to be an example to be followed in everyday life and to advice in moral, religious and practical matters. To the godchild's obedience and affection corresponds the godparent's authority. These expectations imply a kind of intimacy of the older towards the younger, which is reciprocated with formality and respect.

In return for what are practically moral responses from the child to the godparents,¹⁰ the latter is expected to give not only moral "things" but also presents, mainly at birthdays and Christmas, to grant hospitality, to advance small sums of money and, if possible, to help the godchild in starting out in life.

Parents must care for their children and if, for any reason such as poverty, illness or death, they cannot do so, the godparent is expected to act as their substitute. Sometimes godparents may adopt a young orphaned godchild and, mainly if they have no children of their own, they may have the grown up godchild living with them for a certain period of time to help in the domestic tasks. When adult, the landless godchild may ask the godparent's permission to live on his land as an agregado.¹¹ At marriage, if the groom's father is missing, the godfather may ask for the godson the bride's hand to her father (or godfather, or older brother).

This asymmetry can also be seen in the terms of address. A godchild never addresses his godparent by name, while the parent does so; and while the former addresses the latter by the formal phrase "padrinho, o senhor…" ("godfather, sir…" or any variations of this), the latter uses the name and the informal você towards him. As to the terms of reference, in the first case it is "my padrinho…" and the second "my afilhado".

Respect, authority and responsibility, as components of compadrio relations, appear in practically all essays on the subject. The relevant point about this is that each of these elements changes its degree of importance at two moments in the peasant's lifecycle. When still young, he is expected to respect and obey his parents and godparents; they, in their turn, are responsible for him both materially and morally; and at the same time, their behavior towards him is based on authority (cf. Figure 2 below). As soon as the godchild reaches adulthood and establishes his or her own household, the parents and godparents are no longer responsible for him or her. On the contrary, as time passes, the elderly parent becomes increasingly dependent, in material terms, on the adult son or daughter. The elements of authority and obedience, still present in the relation, become much more ideal and real and the godchild is expected to ritually show respect and deference towards godparents as peer adults. In Figure 2 it is worth noting the difference between the first kind of respect, asymmetrical, and the second, which tends to symmetry.¹²

Parent/godparent relation

Compadres are expected to respect one another, to give moral and effective support in political disputes or in family feuds, as well as to help each other materially. As in the godparent/godchild relation, compadres and comadres use specific terms of address and of reference. Even if they are close as brothers and sisters, but not if they are parents and children, the term compadre or comadre should be added before each other's name.

Mutual respect is the basic ingredient of this relationship. In the words of one informant "if I give my child in baptism to someone, it is because I respect that person." "If they're respectable people", said another, "a compadre is forever". Indeed, as far as they can respect one another a compadre is essentially an ally, even when, for any reason, material exchanges do not take place between them. Such exchanges do strengthen the relation, but it does not follow from this that the absence of material benefits would destroy the social and ritual link. As Durham notes, migration, for instance, keeps relatives and ritual kin apart sometimes for a number of years and sometimes forever; but these bonds can be reactivated after being latent for a long time (Durham 1966: 78).

These instances could suggestthat the compadres' relation tends to be symmetric. In fact, in homogeneous groups like the ones I visited, where compadres belonged to the same social class, to the same community, often to the same generation and being close relatives (either siblings, in-laws, or first cousins)¹³, symmetry tends to be far more frequent than asymmetry. This can also be observed that the economic level through the fact that they exchange labor in the institutionalized forms of batalhão and troca dias. In the bottleneck periods of agricultural work, as well as when special care with the animals is needed, people ask for the help of their neighbors and kin; in such cases the support provided by compadres is highly expected. They can either form a group of about 20 or 30 men and, at the end of the work, share a meal and drinks offered by the person who has joined the team, or ask for the help of one or two close friends (mostly compadres or siblings). In the first case, which is called batalhão (or mutirão), the one who forms the group becomes indebted to all the participants and is expected to help them whenever they need; in the second case, called troca dias, the one who was helped owes the others the same amount of labor that was received (half a day, one whole day, etc.).

Nevertheless, symmetry is not a universal characteristic of the compadres' relation, not even when only one part of the country is considered, as I'm considering rural Brazil in this paper. As I mentioned earlier, there is much evidence according to which compadres are chosen, in some rural communities, from richer and more powerful social groups. In those cases, as George Foster has pointed out, the relation is assimilated to the patron/client bond. Although the same patterns of mutual respect and help exist ideally, compadrio relations between people from different classes like the landowner and the landless agricultural worker are not, in the words of Charles Wagley, as close and as intimate as those between people of the same economic and social position"(1964: 157). Respect and loyalty (mainly political, in the broader sense of the term) are much greater from the peasants towards the landowner than vice-versa; and the opposite happens at the economic level. "Filleuls et compères», writes Pereira de Queiroz, «apportaient au parrain leur aide pour abattre les adversaires, grossissant le rang de ses électeurs de ses défenseurs. Le parrain, de son côté, devait protéger le filleul, luis trouver occupation ou emploi, lui donner l'aide économique, lui permettre de cultiver ses terres, le faire sortir de prison si cela lui arrivait pour quelque faute" (1957: 59).

In this context, it is relevant to note that, according to Catholic dogma - and probably to peasant ideology as well - the statuses of parent and godparent, although complementary to each other, are slightly asymmetrical. The parent is responsible for the birth of the child and for the material aspects of his earthly life, while the godparents – ideally at least - are responsible for his religious (and moral) life and, consequently, for the salvation of his soul, in other words for heavenly life. Although the inculcation of religious values is expected to be provided by both, parents and godparents, theoretically at least, if the godparent - who is the one ritually invested with this task - fails to perform it, his sin will be seen as greater than that of a neglectful parent. R. Ravicz' argument about how the element of respect flows between compadres corroborates this view. According to him, also in some contexts there is "a slight imbalance favoring the padriño" (1967:241).

These issues make me wonder whether the characterization of the nature of compadrio bonds in terms of the nature of the exchanges that operate through them is completely satisfactory. In fact, this level of analysis seems limited because it can only demonstrate the different ways through which the institution operates, not its structure. Analyses like that of George Foster (1961, 1963) mainly show that compadrio relations tend to assimilate the form of the crucial structural relations, in the sense that Raymond Firth uses the concept (1963[1951]: 31).

I am persuaded that in order to solve this puzzle one has to appeal to evidence which is being considered as an exception. In fact, there are cases where exceptions can be more helpful than the statistical norm. In other words, I think that the existence of preferential godparents to the firstborn throws new light on the understanding of the meaning of this institution and my argument is as follows.

The firstborn is, at least for two reasons, different from his siblings. The first-born completes the socially expected structure of a new household group (a married couple with their children), adding to the adult couple the roles of father and mother. Probably for this reason, this birth has the special aura of an inaugural event. Besides it is an evidence of the fact that the couple is fertile and do not avoid procreation, which although often practiced, is not openly approved.

The sponsor of this child at what is not only his but also, in this particular case, the parents' rite de passage into the moral society, must surely, by the same token, be someone special. In a society traditionally grounded on the authority of the male and of the elder, this choice often falls in the father's father and, if not him, someone who, like him, is socially "older" than the child's parents, respected by them and by their fellow villagers.

At this level, considerations of a utilitarian character are clearly not sufficient. If one starts from the premise that the social function of godparenthood is to provide both parents and children with an informal "social security" network, it might look quite odd to choose, particularly for the firstborn, someone who, by virtue of his age, will himself become - or actually is - dependent on someone else's help as an elderly person.

What I'm getting at is that when one considers compadrio as a moral contract which has the baptism of the firstborn as a kind of prototype or matrix, then, I suppose, the puzzle is partly solved.

In the baptism of the firstborn, compadrio links an adult couple to people who occupy a relatively higher social position whether in egalitarian groups or not. In stratified rural communities, a lower class person is allowed to have an upper-class compadre by giving him his child in baptism; but the reciprocal is not true: a landless peasant will rarely be chosen as godparent to a landowner's child. In more egalitarian groups, if one can choose one's parent or older sibling to be one's child's godparent, a father never chooses his own child and when a younger sibling is chosen this is seen as exceptional and a great honor for the youngest.

For these reasons, I think that compadrio relations should be basically understood as moral bonds through which an individual, by giving his child in baptism to somebody else shows, with the community's recognition, his loyalty and respect towards that person.

The fact of being asymmetrical at the symbolic level does not prevent compadrio to potentially conform to the basic pattern of the social relations that define the social structure of the group considered. Hence, the system of material exchanges, whose limits are established by those relations, may be either symmetrical or asymmetrical.

Variations

As already mentioned, apart from baptismal compadrio one finds in rural Brazil two other variants: the compadrio de São João and of confirmation. At the same time, the rite of marriage is sometimes referred to in the literature as a rite creating compadrio bonds.

Compadrio of confirmation

Since in the case of confirmation there is only one parent involved (either male or female depending on the sex of the child), three ritual connections are created instead of the nine created by baptism. In Figure 3 below, GP being the confirmation godparent to GC, and F, M the father and mother respectively, the set of ritual connections will be, in this case, constituted by the set formed by F/GP, M/GP and C/GP relations.

According to Catholic dogma, the rite of confirmation does not have the same degree of coerciveness as baptism in the sense that the salvation of the soul depends much more on the latter than the former. Correspondingly, as C. Castaldi pointed out (1957: 51), Brazilian peasants do not attach the same importance to both. I do not have figures to demonstrate this, but in peasant communities one quite frequently finds people never confirmed, while unbaptized people are extremely rare. Confirmation binds people together according to the same basic pattern as baptism. Similarities will be found in material exchanges, in the terms of address and reference (although when talking about them, people may add the expressions de crisma or de batismo in order to differentiate them) and in the expected patterns of behavior. At the symbolic level, they have similar meanings. in terms of our model, it seems acceptable to consider the F/GP and M/GP relations as equivalent to each other and adopt the transformation shown in Figure 3 indicating the structural similarities between baptism and confirmation compadrio.

Compadrio and marriage

At marriage, the bride and groom usually have two matrimonial sponsors each; these are called padrinhos and madrinhas following the baptism pattern. Nevertheless, this cannot properly be considered as a ritual creating compadrio because, at least in Brazil, it does not create the compadre relation.

The marriage godparents are chosen preferably from older people, by the bride and groom themselves. Although this relation is more formal and, since all the persons involved are adults, slightly less asymmetric than the baptismal godparent/godchild relationship, similar expectations are implicit in both. The godparent is expected to advise and help his godchild whenever he needs it; and the godchild, in his turn, should give respect and affection. In this case, then, only four ritual links are created, the ones between the two spouses with the two godparents to each. As the four relations do not present significant differences in terms of the reciprocal behavior expectations (although the relation between people of the same sex tend to be more close and intimate than those of different sexes), the model can be transformed as in Figure 4. The second part of this transformation is an abstraction from the first one; actually, it represents a pattern which is repeated four times in the system of social interactions, showing itself as a simplified version of those previously presented for compadrio of baptism and of confirmation.

Compadrio de São João

In contrast to baptism or confirmation compadrio, the status of compadre de São João (also called compadre de fogueira) is not created through children's rites de passage. Two friends, either of the same of different sexes can become compadres, or a youngster and an adult can become padrinho and afilhado, by jumping the bonfire together at the St. John's Eve celebration. Although this form of compadrio is not as binding and does not have the same religious meaning as the variants presented previously, it implies the same kind of expectations: mutual respect and economic assistance. Other peculiarities of this form are that it is usually symmetrical (to my knowledge there is no inter-class compadrio de São João) and only two people at one time, either as compadres, as I have observed in Bahia, or as godparent/godchild as reported by Charles Wagley (1964: 153-4).

Compadrio 's elementary structure

From my argument it follows that if the elementary structure of the compadrio is to be defined, by which I mean a universal basic pattern that can be found in its variants, then two sorts of ritual connections have to be taken into account: the relations between compadres and between godparents and godchildren. These two relations have been considered in the literature (mostly implicitly) as coexisting inside the same institution. My next question is whether a more precise description would not be achieved if one considered them not only as coexistent relations, but also as forming an ordered pair.

It seems that these relations are not independent from one another. In the more complete forms of compadrio (that is, in compadrio through baptism and confirmation), the link between parent and godparent presupposes a sponsored infant and the recognition of both as "co-parents" of the same child. In this sense, it seems plausible to say that the compadres' bond depends on the godparent/godchild one and to infer that the relation between compadres would be a relationship between adults established through the recognition of a symbolically common child.

I the ritual of becoming compadres de São João the mediating element is replaced by St. John's will, so to say. While jumping the bonfire together, the prospective compadres say: "St. Peter slept, St. John woke up. So and so is my compadre cause St. John ordered it."¹⁴ Here, it seems that the formula "St. John will" plays the role of the third structural element, i.e., the sponsored child in the case of baptism or confirmation. My hypothesis is that, if a list of all the variants of compadrio were examined, it would be possible always to find an element linking symbolically the other two, in such a way that they would form a paradigmatic series.

From this, it follows that at this level of analysis, the above mentioned ordered pair would be such as the godparent/godchild relation would take precedence over the compadres' one, which is coherent with the fact that the religious rituals attach far greater importance to the former than to the latter. If instead, the level of actual behavior is taken into account, a significant difference emerges: while the northern-Spanish and Italian variants seem to confirm this order of precedence, it seems to be inverted in Latin America. In the New World, society puts a far greater emphasis on the parent/godparent relation and the effectiveness and persistence of godparent/godchild one will highly depend on its strength and continuity, at least as far as no into individual involvement of affections is considered.

Compadrio's structure may be defined, then, as a set of size N of ordered pairs of dyadic relations linking three elements: a sponsored child, parents and godparents. Through these bonds, material and moral exchanges take place. These exchanges, when performed by adults, tend to follow the basic patterns of the social structure, considering both social stratification and kinship. In the case of the godparent/godchild relations, when the sponsored is still young, he reciprocates the godparents' gifts mainly with affection and respect, material responsibilities being transferred to his parents. As he grows up, he will be expected to help the godparent in an increasingly more effective way; then, the basic pattern of the exchanges will depend on relevant aspects of the social structure (class structure, neighborhood, kinship etc), as in the compadres' relation. At the symbolic level, both connections are ideally asymmetric, the status of godparent being slightly higher than the others.

The size of this set varies according to which ritual is being taken into account and according to its socio-cultural context. In the case of baptism, in the peasant communities of the sertão there will be usually three; in the more urbanized regions where the godmother de carrêgo does not occur, it is reduced to two. In confirmation, it is one. In the case of marriage and of St. John's compadrio, the model is only partially accomplished. In the first case, the parent/godparent relation does not occur, in the second each ritual creates only one bond at a time and there is a possibility of the model to be accomplished through time. The absence of the compadres' bond in marriage shows that if it is true that the existence of the parent/godparent relationship depends on the godparents/godchild one, this bond, in its turn, does not necessarily imply the other; this implication seem to be unilateral.

Hence, the elementary structure potentially found in the variants of compadrio can be defined as being an ordered pair of dyadic relations - ritually created and sanctioned by the catholic religion - represented by the following model C={Gp/Gc, P/G} where one (but not both) of the constitutive elements can be nonexistent, C being the general form of the institution (compadrio), Gp/Gc and P/Gp standing for the generic godparent/godchild and parent/godparent relations respectively.

Compadrio then, being canonically defined by the Catholic Church, is redefined by custom and takes various shapes in particular contexts. From the ethnographic point of view, it can be understood as a multi-channeled system of exchange built on the basis of ordered pairs of dyadic ritual relations. Such exchanges are performed in terms of the social system, economic organization, language, ideology and ritual, following the general principle of reciprocity.

In this section, and following the literature on the subject, I specially referred to one aspect of such exchanges system, focusing the economic level. However, a more detailed analysis is needed in order to confirm the accurateness of such a characterization. A quantitative evaluation of the goods and services exchanged through these channels could be worthwhile for a more substantive description of the symmetry and asymmetry of such exchanges. And the introduction of such elements in above mentioned Figures 1, 3 and 4 would make their heuristic value more explicit, as a representation of how this system of communication really operates. But the data presently available do not allow me to go any further in this direction. On the other hand, other features implicit in my analysis can be developed through the data I have. So, in the next section I will proceed in my interpretation of this system, pointing out some contrasts with the elementary family and putting forward some ideological implications of compadrio.

IV. Compadrio and the elementary family

Most writings on this subject consider compadrio and kinship relations as some kind of syntactical synonyms at the level of social life. It is as if given a certain context - labor exchange, for instance - one or the other could be used interchangeably, with no sociological consequences: both are socially accepted criteria for defining preferential areas of exchange. There are instances in the literature where they are alternates (Ravicz 1967: 251) and others, which are more frequent, where both are used simultaneously (Mello e Souza 1964: 100-1 and many others).

Other authors suggest more than a functional equivalence between them. Among these, R. Ravicz, Julian Pitt-Rivers and S.J. Gudeman are of special interest for my argument. According to Ravicz (1967: 241-2) padrinazgo (which I have called the godparent/godchild relationship) "reproduces the nuclear family in terminology and, partially, in function." Compadrazgo (or the parent/godparent relation), he argues, "bears no resemblance to the family or other kin group in organization, terminology or behavior. (…) Terminology differentiates only by sex; generation differences are collapsed, accompanied by an emphasis on horizontal extension and the near disappearance of lineality and differences of authority." In contrast to padrinazgo, "the compadre bond affects interfamilial relationships providing the community with a model for social interaction at all levels".

Julian Pitt-Rivers (1968: 412) emphasizes another aspect of the problem. According to him, compadrio is opposed to cognatic relationships. First, it is what the other aspires to but cannot be: immutable and free of ambivalence; second, it does not have origin in the past or projection into the future; and third, it is a kind of spiritual kinship as opposed to the profane ordinary one.

S. F. Gudeman's argument (1969) is linked to these two in various ways. In contrast to Ravicz, he maintains that both the compadrio group and the elementary family are based on the same underlying model of a three-member group. And taking further Ravicz's and Pitt-Rivers' suggestions he argues that both systems are logically opposed; the first are lasting bonds of a sacred nature and the second ones are breakable, of a physical nature and based on material exchanges. His analysis seems to illustrate Durkheim's bipolar distinction between sacred and profane beings or things in which improper contact with one of these things or beings - which he calls contamination or sacrilege - would have the final effect of destroying its original characteristic. In fact, Gudeman argues that in the former, material exchanges are "inhibited" because of its sacred nature, and to keep it out of the general circuit of economic transactions (which is profane) would be a way of keeping the ritual bonds unpolluted. At the same time, ritual and ordinary kinship are complementary to each other in the sense that, if the family provides the model for the internal organization within households, compadrio provides the pattern for the relation between households.

My disagreement with the arguments outlined here can be reduced to a few basic points.

First, if useful analogies and contrasts are to be drawn, one has to take into account not each of the relations separately but the two systems as wholes. Neither can filiation be fully understood if one does not take affinity into account, nor siblingship without affinity and filiation. Similarly, as I have already argued, the creation of the parent/godparent bond depends on the existence of the godparent/godchild one, and the explanation of one relation depends on the explanation of the other.

The second point refers to the use of concepts. I do not agree that padrinazgo reproduces the nuclear family (as Ravicz puts it) since the former is defined as a relation between a godchild and his godparents while the latter is a system of three bonds linking four elements two by two (siblings, parents to child, husband to wife); from this point of view, the only possible inference is that the godparent/godchild relation have aspects in common with the parents/child one. Furthermore, Gudeman's assertion that both are three members groups is not acceptable because he does not take siblingship into account.

The last one may be a matter of empirical data. According to the Brazilian case, the sacredness of the compadrio does not inhibit economic cooperation among the ritual kin. On the contrary, there are several instances in the literature showing that, in rural Brazil and elsewhere, these bonds activate economic transactions (Mello e Souza 1964: 100-1; Durham 1966: 76; Wagley 1964: 155-6; Mintz & Wolf 1950: 356; Pitt-Rivers 1954: 108 and many others). Furthermore, the interrelation of households seems to be far more dependent on whether or not the social group is stratified or on the way in which the household group participates in the economic system, than on moral precepts for social behavior. Whether or not the peasant neighborhoods that I have studied and Gudeman's community in Panama correspond to two different points in a continuum defined by the secularization process, is a matter for further investigation.

My argument is that, on the one hand, being a logical negation of marriage and physical procreation, compadrio system as a whole affirms symbolically, and in different instances, the ideologically "proper" elementary family: that of Joseph, Mary and Jesus. On the other hand, it provides structural complement to affinity as ways of establishing institutionalized connections between villagers. But let me start from the beginning.

Compadrio, filiation and siblingship.

Analogies between the godparent/godchild relation and filiation can be found in several contexts. In the first place, let us examine the level of language. The mutual terms of reference used by parents and children are paralleled by the terms used by godparents and godchildren. The former are: pai, mãe and filho, filha; the latter are padriño, madriña and afilhado, afilhada. In both cases they are asymmetrical terms, denoting differences by sex and generation, and they are differentiated from one another through the monemes -iño(a) and a-,-ado(a) respectively. The first of these denotes diminutive and indicates intimacy and affection; the other two indicate the quality of being similar to, or of the nature of, or tending to; -o and -a mark, respectively, masculine and feminine.

Apart from the fact that authority and obedience are distributed in the same ways in both situations, similarities would also be found at the level of economic life. From early adolescence, until marriage and the formation of a household, a sertanejo child is not merely a dependant individual. In fact, young girls look after younger siblings, clean the house, cook meals, wash clothes and carry water from reservoirs to the house; young boys, in their turn, take meals to their parents and older siblings at their working places, look after the families' livestock, deliver messages etc. By so doing, young adolescents free their parents and older siblings for more productive tasks. The older they are, the more fully they participate in the sexual division of labor. Boys and girls share women's tasks in agriculture; they help to clean gardens before burning the weeds and old roots, they seed, harvest and clean the beans etc. Their contribution to the household economy is neither voluntary, nor occasional, nor dispensable; in fact, it is as essential condition of the production of the domestic group's subsistence (Wolf 1966: 66; Durham 1966: 68-9 etc.).

Most times youngsters work with parents, but it is not usual to find particularly girls living with old or childless godparents, uncles or grandparents. One of my informants, for example, an old widow, lived with a bachelor son, a 14-year-old granddaughter and a slightly older goddaughter. The girls help her in all domestic tasks. The goddaughter's parents lived across the river, in the same village, and their daughter used to visit them quite often. I noted several such cases during my fieldwork. Another example is a childless couple who had a three-year-old niece living with them. Children were never said to be working for their godparents but living with them. As an old woman explained, this is seen as a very good opportunity for the child and for the parents, because old people know everything about bringing up children and parents do not have to worry about feeding them.

I would like to emphasize here that I am not suggesting that there are not differences between economic cooperation among family members and ritual kinsfolk. Within the elementary family this is obligatory, while within the compadrio group it depends on "favor and benevolence", as Pitt-Rivers puts it (1968: 412). What I am trying to show is that, in both cases, the practical results are equivalent and equally important in terms of subsistence production.

Apart from showing that the godparent/godchild relation can be, in a certain level of material exchanges, analogous to the parent/child one, these instances suggest another functional similarity between compadrio and kinship. Both seem to be ties through which a fosterage system operates; and the relevance of such a system lies in the fact that it provides a way of equalizing the composition the household groups and balancing its size in a social context marked by subsistence economy based on the elementary family's work.

Analogies between the parents/godparent relation and that of adult siblings or in-laws can also be drawn. The terms of address and of reference between compadre/comadre and irmão, irmã are symmetric; they differentiate status by sex, and refer them, in the first case, down to the next generation and in the second, up to the one before; compadres are, in a way, parents of the same child, while siblings are children of the same parents. From the point of view of economic transactions, the cooperation between compadres parallel the siblings' one (and that of siblings-in-law). Although neighbors participate in a system of interpersonal exchange of goods and services of various kinds which, to a certain extent, is a way of defining the limits of these neighborhood groups, it from the compadres, from the siblings and from the in-laws that peasants expect most.

These analogies are further reinforced by the fact that sexual intercourse and marriage are prohibited for people linked by filiation and siblingship, as well as by compadrio bonds. I do think it is necessary to emphasize the relevance of this prohibition in terms of the patterning of kinfolk's behavior. As to its strength vis-à-vis the compadrio relationships, I think it is sufficient to evoke that compadres are also sometimes chosen in such a way that latent or possible extra conjugal liaisons are put to an end (Wagley 1964: 151; Willems 1961: 65; Mintz & Wolf 1950:356; etc). Harry Hutchinson mentions an interesting evidence of such prohibition. "Usually", he writes, "the couple who breaks this taboo is highly censured by the rest of the community, for the wife, if she finds out, tells her friends about it. In one case that occurred in Vila Recôncavo during the authors stay, the woman was so ashamed and frightened by the sin she had committed that she killed herself by drinking rat poison" (1957: 147).

Comparison also shows that there are fundamental distinctions between them. While kinship relations are social links through which property and status are transferred from generation to generation, in both lines of descent, there are rare usages - if at all - of compadrio in inheritance, as noted by Mintz and Wolf (1950: 355). From the legal and customary points of view, in Brazil, an individual has no right to property owned by his godparents or compadres, although properties can pass along these lines through gift or testament. There are records in the literature about godchildren or compadres having the permission to cultivate their own subsistence in their godparents' or compadres' lands, but without having property rights over it.

Compadrio and affinity

So far, I have mainly paralleled compadrio and kinship relations; but there are also quite apparent analogies between ritual kinship and affinity. The similarity of the cooperation between compadres and between in-laws derives not only from the nature of the exchanges performed but also from the fact that in both cases they are involve adults belonging to different households, whose link is based in ritual contracts. Besides this both are similarly opposed to consanguineous kinship in the sense that, if the latter derives from birth, within certain limits, affinity and compadrio are based on choice and established through the mediation of a woman, in the first case, and of a child in the second.

But to try to find similarities between compadrio and affinity is to miss a rather crucial point which derives exactly from the way that they are opposed. In order to take this argument further, let me consider some aspects of marriage as observed during my fieldwork, especially the way in which the sertanejo chooses the wife.

Despite the ideal pattern of marrying kin, which is justified through assertions such as "one never knows how outsiders behave", my data suggests two basic tendencies.

First, there is an increasing tendency towards marrying outside what is called "the close kinship group", i.e., the group which extends itself from the nuclear family to first-cousins. Taking the last three generations into account one notes in Table 3 below that 26% of marriages took place inside such group while 74% are with "non-relatives".

Second, locality seems to be a relevant criterion of choice since 60% of the fully recorded marriages were between people living in the same neighborhood. All the others but one were between men and outside women; only one woman married to an outsider lived in her parents' neighborhood. This issue indicates a tendency towards virilocality which is reinforced by the fact that, in the case of in-group marriage, the new couple usually build their house near to the husband´s parents.

In terms of keeping a minimal amount of land sufficient for the production of subsistence, the tendency towards marrying outside the immediate kinship group is balanced by the fact that, in most cases, marriage occurs among people who are neighbors. But most neighbors are kinfolks as well. Indeed, with only one exception,¹⁵ the various kin groups living in each neighborhood focused in my fieldwork were related at the fourth or fifth ascending generation, all its members being heirs of the same fazenda.¹⁶

What really matters most for my argument is that the exchange of labour, of women and compadrio occur most times among people belonging to the same group, the one defined in the sertão trough land co-ownership. This is the context where the contrast between compadrio and marriage seems to be particularly significant.

Before giving a child in baptism to other adults, a man will give his sisters to - and get a wife from – men who do not belong to the same kinship group as himself; or, if from relatives, from people who are, structurally speaking, as far as possible and not closer than a first-cousin. When choosing a compadre, the tendency will be the reverse of this and the closest relatives will be preferred to non-relatives.

What I am suggesting is that, in terms of linking together people from the same generation and belonging to the same village, compadrio and marriage tend to be more relevant at extremes of a scale which ranges people between the two opposing categories of close kin (parents and siblings) and not-relatives, compadrio belonging to the former category and marriage to the latter. For this reason I argue that they are at the same time opposite and complementary sorts of contracts between adults.¹⁷ Furthermore, I would like to emphasize that one will see this opposition in other instances. Marriage implies transmission of property, sexual relations and socially accepted procreation; compadrio not only doesn't it also prohibits these.

The rules of the social game, then, as I see them would be such that given the two opposite groups of people, the ones one can marry and the ones one cannot marry, what this logics suggests is that one should have a wife from (or give a sister to) the first and a compadre (or give a child to) from the second. Because marriage admits transitivity, with the wife one acquires also ones in-laws; because compadrio does not admit transitivity, one establishes as many compadrio bonds as one can.

Some ideological bases of compadrio

From my argument it might follow that, to a certain extent, ritual kinship is redundant vis-à-vis the kinship ties. Instead of accepting this point of view I would rather formulate a last problem. Why should the peasants create compadrio bonds were previously existed kinship ones? What would be the meaning of giving a child to be baptized by someone else, chosen so carefully? I am very much inclined to suppose that this issue is related to the basic distinction the peasants make between spiritual and natural things, or between the sacred and the profane.

Going back to the rite of baptism, which has been considered in this paper as the prototype of the rites which create compadrio, I would suggest that it symbolically denies natural kinship. Being immersed in blessed water and emerging afterwards, the child ritually dies and is re-born by the hands of a priest and with the sponsorship of the godparents, purified from original sin (Eliade 1959:130-1) Only from this moment on is the child is able to attend religious services, to marry in church, to be buried in a cemetery and to acquires the possibility of having his soul saved: he is no longer pagan, but a Christian, a member of the moral religious community, a son of God.

Besides that, the symbolic denial of the biological fact of birth by ritual kinship becomes apparent from contrasting filiation to the godparent/godchild relation and marriage to the compadres one. In both cases, at the bottom of the ritual relationships there lies the social recognition of a kind of filiation and co-paternity thought of as being morally superior to the ordinary bonds: social parents do not have to be present at the ritual, to godparents the religious life of the child is entrusted and the effectiveness of the compadrio system is more dependent on moral bonds and choice than kinship.

Edmund Leach argues that "the structure which is symbolized in ritual is the system of socially proved 'proper' relations between individuals or groups" (Leach 1954: 15). Indeed, what seems to being symbolized here is the 'proper' elementary family: that of God himself. Mary is married to Joseph; both are human beings. Mary gives birth to a child, the son of God. Christians make a careful distinction "between Jesus legal status as a man and his essential nature as a God. As a man he is the legal son of Joseph husband of Mary, and in this legal sense he belongs of the lineage of David. In contrast, his divine essence derives from the fact that the male component of his conception was 'the Holy Spirit'."( Leach 1969: 97) The bond between Mary and the Holy Ghost is, in the myth, opposed and complementary to her marriage, from the point of view of the statuses of her son, which are dissociated: he is, at the same time, a man and a God. Jesus has neither brothers nor sisters, but this is possibly due to the uniqueness of his being.

An ordinary child, in his turn, having social parents, is given sacred parents (godparents) at the time of his ritual rebirth. The first ones are responsible for his physical birth; they transmit original sin and must look after him materially. The second ones provide for his spiritual rebirth, the cleaning of original sin and are responsible for his religious life on which depends is eternal life. Parents and godparents have, then, complementary and opposed duties towards the child who is thought not merely as a human being, which must be given food and cared for, but also as having a soul which must be "looked after". Consequently, being coincident with kinship bonds, the compadrio give them the religious backing which they lack. This is one way of interpreting the moral asymmetry between the statuses of parent and godparent. Kinship and compadrio, then, can be thought as being two sides of the same coin, one profane and the other sacred, one being the negation and the counterpart of the other.

Usually, a paper ends with a summary of the argument and with reference to how the main partial conclusions can be put together and how they fit into the general statement of the introduction, in a Q.E.D. kind of explanation. The exploratory nature of this paper however not only allows me not to close my argument circularly, but also to end up suggesting a further implication of it. This arises from a point made by Leach in his "Rethinking Anthropology". Grounded on ethnographical evidence from societies with kinship systems based on unilineal descent, he argues that "in any system of kinship and marriage there is a fundamental ideological opposition between the relations who endow the individual with membership of a "we group" of some kind (relations of incorporation), and those other relations which link "we group" to other groups of like kind (relations of alliance), so that, in this dichotomy, relations of incorporation are distinguished symbolically as relations of common substance, while relations of alliance are viewed as metaphysical influence." (Leach 1961: 21. Emphasis added)

In the unilineal societies that he takes into account, when such relations of common substance are associated with the mother/child kind of filiation (in other words, in matrilineal systems), metaphysical influence will be associated with the agnatic bonds; and vice versa. But since this assertion is supposed to be true for any system of kinship and marriage a question immediately arises. What would be the case if cognatic systems were considered?

In a society with a cognatic kinship system like the one I have considered in this paper statuses are attributed not only through the social recognition of both lines of descent, but also physical appearance. Children are quite often said to resemble one parent or the other, and sometimes both. It is quite common to hear that a child has inherited physical appearance from one of the parents and the personality from the other, or that he partly resembles physically one of the them and partly the other. A child, then, is very much thought of as being a kind of mechanical synthesis of the characteristics of both parents, both in looks and personality. Taking this into account, the following Figure might be added to figure 3 of the article just mentioned (Leach 1961:20):

If the Brazilian evidence is taken in consideration, it would be an evidence where the ratio z=p/q would be nearly 1, p being filiation with the mother, q being filiation with the father, and p being almost equal to q.

So far, so good. But where could one possibly find the distinction between the relations of incorporation and the relations of metaphysical influence? Here, I think, compadrio defined vis-à-vis consanguinity as two opposed yet complementary sorts of kinship, one sacred and the other profane, seem to have a certain amount of relevance.

If, instead of taking into account the patterns of attributing physical resemblance or witchcraft, we characterize the relation of mystical influence in a positive way (from the point of view of the group involved) - through the participation in one's religious life, by the sponsoring a rite that marks the passage from a state of pagan to that of Christian, the moral distance and the symbolic asymmetry among the participants of the institution - then it would follow that probably compadrio relations could be interpreted as analogous to the second element of Leach's dichotomy.

The fact that compadrio links mainly consanguineous relatives, does not seem contradictory to this hypothesis. On the one hand, this analogy is possible because the contents of these relations are not antithetic; on the contrary, I have already argued that they have many aspects in common. Indeed, many pages have been written by different authors in order to show that such mutual relations are nothing but images of kinship. On the other hand, when a kinsman becomes ritual kin, the previous bond is, so to speak, denied or redefined. One evidence of this is that the new bond is superimposed over the old one in such a way that in address and reference, for example, the new terminology is used and the new kind of relationship is thought of as having precedence over the kinship relations. Furthermore, the system is preformed in that same way: among the someone's nephews, for instance, better and more expensive presents will be given to the child who is also a godchild, and vice versa, among someone's uncles and aunts, the one who is also a godparent should be more loved and respected. In the case of the compadres' relation, the situation is analogous.

Let us return to the problem of choosing one of the grandfathers as preferential godfather to the firstborn. Analyzing this issue in terms of Leach's dichotomy, it is not only possible to emphasize its raison d'être, but also to understand better its meaning in terms of the ideology of the group considered, as well as the assertion that the creation of the first compadrio bond is part of the young married couple's rite de passage.

According to this perspective, the transformation of filiation (or of a relation equivalent to this one the parent being substituted by an uncle, aunt or older sibling) into a godparent/godchild relation implies the following consequences. Being initially a relationship based on the principle of authority and submission, on the idea of the unilateral material dependency of the child on the father, and symbolically marked as a relation of common substance, when a parent becomes someone's compadre or comadre this relationship is redefined and becomes, following Leach's model, one of metaphysical influence. What are the implications of this?

The first implication is that the relation of authority is transformed into a relation of interpersonal and reciprocal respect, given the religious meaning of the rituals that bring about compadrio. The relevance of this kind of relationship from the point of view of social life in peasant communities, whether homogeneous or stratified, does not have to be further emphasized in this essay.

The second, which is a consequence of this, the ideological interpretation of the economic aspects of family life (in which the parents are basically conceived as providers) is redefined in terms of the generalized cooperation between equals. In other words, they are more akin to the ideal relations of cooperation within the community (as opposed to those within the family). The effectiveness and strength of the bond, as well as the moral superiority of the elder, is assured by the sacred nature of the relationship.

V. Post scriptum

I would like to raise two points in connection to the present text. The first has to do with editorial matters, since this version of my essay includes a final passage (pages 38-41) that was omitted or referred to only indirectly in previous publications (Arantes 1975, 1982, 1993). In these closing remarks I suggest that Leach's argument developed in his Malinowski memorial lecture about social incorporation and mystical influence as two contrasting ingredients of kinship and alliance (Leach 1961) can be applied to compadrio. I furthermore argue that this extension of his thesis contributes to a clearer understanding of the symbolic dimensions of that institution. My insight is that, if considered side by side with the other instances that form the set of structural variations analyzed in Rethinking anthropology, (Leach 1971[1961]: 8-27) this case would indicate that metaphysical influence is a variable that can not only alternate between kinship and affinity bonds, but also be, so to say, displaced from the elementary family structure to ritual kinship.

One reason for publishing this idea, that has remained unpublished because of being – from my point of view - not sufficiently developed, is that it can suggest new perspectives of interpretation to those who may be interested in doing further research on this topic. Besides this, the Editors of Vibrant and I have pondered that there is also a matter of some historical relevance in the comments that Professor Leach, then my supervisor, made about this topic in his answer to my consultation about critical remarks made when this paper was submitted as a Master's dissertation at USP. In a letter written shortly after my viva I directly asked if the argument developed in Rethinking anthropology would be applicable to all types of kinship and marriage systems; besides this, I wondered if my interpretation of the transformation of filiation into a compadrio bond seemed adequate to him. Leach's reply was that:

"My own feeling is that your extension of my idea (that, with unilineal descent and affinal alliance, the relationship with the affines tends to be seen as 'metaphysical influence') to the context of Compadrio is a very interesting one which seems to fit the facts in a rather convincing way.

As you know, I am convinced that there is no simple unitary explanation for institutions of the Compadrio type even though they may all have a common Christian origin. In different parts of the world this institution of "Godparenthood" has been adapted to the locally existing kinship pattern in many different ways. One of my pupils has just reported a new variant from a Roman Catholic area of Western Ceylon! This makes it all the more important, it seems to me, that the kind of analytical ethnography which your little essay represents should be published, though you should not imagine that this is going to be the last word on the subject!"

The above transcribed comment brings forth an important idea about generalization in Leach's theory which was crucial in constructing the object of the present analysis. As argued in that well-known essay, "generalization is inductive; it consists in perceiving possible general laws in the circumstances of special cases (…) If we are to generalize, a small cluster of interconnected facts must be treated as an isolate expressing a particular principle of social mechanism." (Leach 1971[1961]: 5, 12)

The second point that I would like to raise has to do with a bias that only became apparent to me in revising this piece. It is really surprising that the gender bias present in this essay - as well as in the literature consulted - had not yet been spotted, at least to my knowledge. Gender differences within the compadrio system do not only occur in connection to linguistic variations of terms of address and of reference. Despite the fact that the actors usually focused by ethnographies are male – or referred to as such - relationships between comadres, between comadre and compadre and between godchild and godmother have important specificities, in peasant as well as in urbanized environments, in Brazil and probably in other countries. One relevant domain of social life where gender-specific elements arise is certainly childbirth since, for example, local midwives often have a great number of godchildren, and prospective comadres frequently play some part as helper in the birth of their godchildren. Another particularly interesting area is social control, since comadres typically share intimacy in kinship and neighborhood networks. The metaphorical meaning of the term comadre as a gossiper has a direct implication with this issue.

It is not clear to what extent the consideration of such gender-based realities would change the general hypotheses and tentative generalizations put forward in the bibliography or by this essay. But gender is surely an absent perspective in compadrio studies; and I would suppose that it is a very rich one that deserves to be more systematically explored by future anthropological research.

• Arantes, A. A sagrada família: uma análise estrutural do compadrio. Boletins do IFCH, 5. Campinas: Unicamp, 1975
• Arantes, A. "Pais, padrinhos e o espírito santo" in Arantes, A. et al. Colcha de retalhos: estudos sobre a família no Brasil. São Paulo: Editora Brasiliense, 1982
• Arantes, A. "Pais, padrinhos e o espírito santo" in Arantes, A. et al. Colcha de retalhos: estudos sobre a família no Brasil. Campinas: Editora da Unicamp, 1993
• Bunge, M. "Les concepts de modèle" in L'age de la Science, 3. Paris: Dunot, 1968.
• Castaldi, C. et al. Estudos de sociologia e história. São Paulo: Anhembi, 1957.
• Deshon, S. "Compadrazgo in a henequén hacienda in Yucatán: a structural evaluation." American anthropologist, 65, 1963
• Durham, E. R. Migração, trabalho e família. MS. 1966
• Eliade, M. The sacred and the profane. New York: Harcourt and World, 1959
• Firth, R. Elements of social organization. London: Watts & Co., 1963 [1951]
• Foster, G. Empire's children. Mexico: Smithsonian Institution and Imprenta Nuevo Mundo, 1948
• Foster, G. "What is folk-culture?" American anthropologist, 63, 1961
• Foster, G. "The dyadic contract. A model for the social structure of a Mexican peasant village". American anthropologist, 65, 1963
• Gudeman, S. F. Household, family and compadrazgo in a Panamanian community. PhD Dissertation. Cambridge, 1969.
• Hammel, E. Alternative social structures and ritual relations in the Balkans. New Jersey: Prentice Hall, 1968
• Hutchinson, H. Village and plantation life in Northeastern Brazil. Seattle: The American Ethnological Society, 1957
• Leach, E. R. Political systems of highland Burma. London: Bell & Sons, 1954
• Leach, E. R. "Rethinking anthropology" in Rethinking anthropology. London: The Athlone Press, 1971 [1961]
• Leach, E. R. "The virgin birth" in Genesis as myth. London: Cape Ed., 1969
• Lewis, O. Life in a Mexican village: Tepoztlán re-studied. 1951
• Melo e souza, A. C. Os parceiros do Rio Bonito. Rio de Janeiro: José Olympio, 1964.
• Mintz, S. & Wolf, E. "An analysis of ritual co-parenthood". Southwestern Journal of Anthropology. 6,4. 1950
• Pereira de queiroz, M. I. La guèrre sainte au Brésil: Le mouvement messianique du Contestado. São Paulo: FFCL-USP, 1957
• Pitt-rivers, J. The people of the Sierra, 1954
• Pitt-rivers, J. "Pseudo-kinship". Sills, D. (Ed) International Encyclopedia of the Social Sciences. Glencoe: The Free Press, 1968
• Ravicz, R. "Compadrinazgo". Wanchope, R. (Ed) Handbook of Middle-American Indians, vol. VI. Austin: The university of Texas Press, 1967
• Redfield, R. The folk culture of Yucatán. Chicago: The University of Chicago Press, 1941
• Redfield, R. Primitive world and its transformations. Chicago: The University of Chicago Press, 1953
• Redfield, R. Peasant society and culture. Chicago: The University of Chicago Press, 1956
• Tylor, E. B. Anahuac, or Mexico and the Mexicans. London: Languans, 1861
• Wagley, C. Amazon town. New York: Borzei Books, 1964
• Willems, E. Uma vila brasileira. São Paulo: Difusão Européia do Livro, 1961
• Willems, E. "On Portuguese family structure" in Morgley, J. (Ed) Family and marriage. Leiden, 1963
• Willems, E, & Mussolini, G. Buzios Island, 1952
• Wolf, E. Peasants. New Jersey: Prentice Hall, 1966

1

. In Brazil, parents and godparents of the same child address each other by the reciprocal terms

compadre and

comadre, the first being used for men and the second for women. Godchildren address their godfather as

padrinho and their godmother as

madrinha. A godson is called

afilhado by his godparents and a goddaughter

afilhada. The term

compadrio is used in this essay with reference to the set of ritual relations between parents and godparents and between godparents and godchildren.

2

."…pour saisir le réel on commence par écarter de l'information. Puis l'on ajoute des éléments imaginaires (ou plutôt hypothétiques) mais dans une intension réaliste. On constitue, ainsi, un objet modèle plus ou moins schématique et qui pour donner des fruits, devra être greffé sur une théorie susceptible d'être confrontée aux faits. » «…l'objet modèle (…) represent toute une classe de choses (ou de faits) regardées comme équivalents bien qu'elles diffèrent entre elles. La relation entre modèle et objet concret est donc une relation multivoque. » (Bunge 1968: 168-9)

3

.Hetzmek is « the ceremony wherein an infant is for the first time carried astride the hip and the sponsor puts in its hands articles symbolizing the sought for sound development of the child's abilities." (Redfield 1941: 391)

4

. This argument will be developed in Section III.

5

. For the historical antecedents of present day baptismal

compadrio see Mintz & Wolf 1950.

6

. See P.16 below.

7

. This variation has not a Canonical basis. See P.25-26 below.

8

. This condition is described by the word

posseiro, meaning consensual owner of the land.

9

. These relationships will be more fully described in the next section.

10

. As a matter of fact, the godchild gives from time to time gifts to his godparents. D.Gross observed in Vilanova, Bahia, that boys and girls take a piece of soap or a perfume to their godparents on Good Friday and they use to give them, in return, small amounts of money. (Personal communication)

11

. The

agregado is someone who has the permission to live in somebody else's house or to gain subsistence from his land without paying any kind of rent. Examples of these usages of

compadrio bonds are reported by Willems & Mussolini 1952: 77-8, by Castaldi 1957: 49 and others

12

. This distinction between two kinds of respect is useful also for the understanding of the relation between

compadres, where respect can be either symmetrical (in intra-class

compadrio) or asymmetrical (in patron-client

compadrio).

13

. See Table I, in P.14 of this essay

14

."São Pedro

dormiu, São João acordou. Fulano é meu compadre, que São João mandou." Bahia, 1968.

15

. These are descendants from slaves of a former owner of one of the fazendas included in my survey. They were entitled by the testament of a previous owner to live in his descendants' property for up to 5 generations.

16

.

Fazenda in this context refers to large portions of land occupied by co-heirs of former landowners whose legal title to land is often not fully documented. The effective right to use part of these estates for subsistence or for commercial purposes is regulated by custom and by social memory, although documents in local archives may sometimes prove a particular family´s right of property.

17

. Analyzing kumstvo in Yugoslav unilineal descent groups, E.Hammel shows that "the ties of ritual, consanguineal and affinal kinship are generally mutually exclusive." (Hammel 1968: 85).

Publication Dates

• Publication in this collection
  14 May 2012
• Date of issue
  Dec 2011