Navigation – Plan du site
Altérités, inégalités et mobilités au sud-ouest de l’océan Indien

Ritual as a Social Diagnostic and Lens of Comparison in Mayotte and Its Neighbours

Le rituel, caractéristique et point de comparaison des sociétés mahoraises et voisines
Michael Lambek
p. 63-77


Dans les sociétés du canal du Mozambique – sociétés apparentées qui s’interpénètrent – le rôle majeur que jouent les rites dans la constitution de la personne et des relations sociales peut servir de point de comparaison. Ngazidja (Grande Comore), décrite et analysée par Sophie Blanchy, incarne la société du don telle que l’a présentée Marcel Mauss. À Ngazidja, la circulation du don – à savoir des choses qui ont le plus de valeur – s’articule autour du Grand mariage, institution qui à la fois reproduit les personnages (matérialisés par leurs maisons) et produit et reproduit les distinctions entre personnes et entre familles. Ceux qui parlent le kibushi à Mayotte ont pris leurs distances vis-à-vis de la hiérarchie comorienne et ont transformé le Grand mariage en shungu, institution qui met en œuvre toute une série de rites de passage comprenant non seulement le statut acquis par la circoncision et le mariage mais aussi les statuts donnés par la naissance et la mort, de manière à produire plus une égalité de distinction que de l’inégalité. Ici, le don se trouve dans l’acte, c’est-à-dire dans le travail réciproque de reproduction et de reconnaissance des autres. Comparé à Ngazidja et Mayotte, Madagascar met l’accent sur les rites mortuaires. Les actes de reproduction et de témoignage y sont constitutifs, là aussi, de hiérarchie : chez les Sakalava, la production des ancêtres royaux implique la reproduction des distinctions sociales entre, d’une part, les membres du clan royal et, d’autre part, les razan’olo formant une caste de serviteurs royaux et une « espèce particulière » de personnes (karazan’olo).

Haut de page

Texte intégral

  • 1 I am indebted to the Social Sciences and Humanities Research Council of Canada for extensive financ (...)

1Every student of society, whether anthropologist or historian, has to address the task of delimiting the unit of study and then rationalizing t’ose limits1. This takes place along at least three dimensions, temporal, geographical, and sociocultural. Time and space allow arbitrary cut-off points, but it is less easy to delimit boundaries with respect to the social.

2At first glance, one might suppose that each island of the western Indian Ocean contains a discrete culture or society, bounded from the others by its shoreline and available for distinct synchronic analysis, thereby providing the classic prototype or stereotype beloved of 20th century anthropology and especially of its critics. But a closer examination shows immediately that this is not the case. Not only is there intensive migration and travel among the islands today, but there has been a continuous circulation of goods, ideas, and people, and a waxing and waning of military, political, and economic power for a very long time, bringing with it an overlap, imbrication, and fermentation of cultural practices and social formations. This goes back to the very first habitation of the Comoros and Madagascar from Southeast Asia, eastern Africa, and the Arab societies of the Hadramaut and Gulf, and subsequently both there and in the Mascarenes by South Asian, Chinese, Portuguese, Dutch, American, and especially British and French traders, sailors and pirates, slavers and slaves, indentured labourers, colonists and colonizers, missionaries and tourists, development workers and capitalists, albeit in differing combinations and with differing but interrelated histories on the different islands.

3Given this picture, I want to develop a framework capable of both understanding society at particular times and places and comparing social life across times and places but with respect to criteria that are culturally relevant in the region as a whole. That is, I want to find a means to recognize the local and regional at multiple levels of inclusion. Such a means should enable the discernment of both continuities and discontinuities and hence form a basis for relevant comparison.

  • 2 See M. Lambek [2011] for an extensive discussion of ‘the local.’

4I do think it is possible to speak about distinct localities and local forms of life. These local forms of life are characterized by their specific histories of conjunction and articulation of diverse traditions and by the degree of overlap in their respective communities of practice as configured within the specific political and economic circumstances and histories of the place. They are also characterized by the histories of the specific social actions – such as gifts given and received, ancestors acknowledged or disregarded – that happened at each locality. It is insufficient to replace this specificity – indeed, this singularity – with a global concept like ‘hybridity’ or ‘cosmopolitanism,’ despite the attraction and relevance of both these epithets.2

5Locality however does not equate with strict social or cultural boundaries. Although Mayotte is only 375 kilometers square it is impossible to assign it with a specific ‘culture,’ either internally homogenous or readily discernable from that of its neighbours. Inhabitants of Mayotte since at least the early 19th century have included speakers of a number of Comorian and Malagasy dialects and of people who traced their origins to various parts of Madagascar, to each of the other three islands in the Comoro archipelago, to Mozambique and the Swahili coast, and to the Middle East, South Asia, La Réunion and France.

  • 3 The mixture creates a problem for transcription. Here I use ‘u’ for the sound that in French is rep (...)

6I conducted fieldwork in a pair of villages of Malagasy speakers in Mayotte for 14 months in 1975-1976 followed by some 9 additional visits of varying duration between 1980 and 2009, some of which also included visits to villagers who were residing in La Réunion. Significantly, they refer to their language as Kibushi, a term that itself is constituted through linguistic mixture.3 As to their collective identity, when pushed, they preferred to refer to themselves as Silamu, Muslims, an appellation that precisely did not distinguish them from the majority of other people in their milieu. At other times they called themselves ulu maskin, poor people, again a term that did not emphasize cultural or ethnic difference. Many speakers of Shimaore and other Comorian languages had married into the villages; most Kibushi speakers were bilingual in Shimaore, and language distinctions were not particularly salient boundary markers. In later years the villagers said, zahay favazaha; we have become French.

7Moreover, despite the surface appearance of homogeneity there were differences in origin among the Kibushi speakers themselves and a tacit hierarchy based on these assumed and attributed origins, as well as stereotypes about the character of people and society in their respective places of origin. I happened to live among people some of whose ancestors had been brought as late as the early 20th century from the coast of East Africa, from Mrima, and who were sometimes referred to as Makua. Some, but not all of the people whose ancestors had come from Madagascar considered themselves a kind of elite. These distinctions were reinforced by tacit patterns of marriage, relative skill at fishing and seafaring, initial interest in French schooling, and a few other matters, but they were no sharp boundaries. The other villages of Mayotte could be characterized along similar differences of origin, yet they were all knit together internally and externally as communities, primarily through the ceremonial exchanges that I will shortly describe.

8People in Mayotte did refer to others from where they came, especially Adzudzu (from Grande Comore) and Ndzwani (from Anjouan). Famously, the inhabitants of Mayotte rejected a political unification with the other three islands of the Comoro archipelago in favour of a continuing relationship with France. The issues were primarily political and economic rather than ethnic or cultural. Inhabitants of Mayotte were afraid of being dominated by the elites of the other islands and swamped by their land-hungry larger populations. Nevertheless, they did not discriminate against individual inhabitants with origins in these other islands and they shared Muslim celebrations and spirit possession activities, frequently intermarried, and recognized kinship ties.

9Although recent developments in Mayotte have created sharp political borders and economic divisions, overall the region has not been characterized by a politics of exclusion. Socially, the boundaries remain very porous. Indeed in Mayotte it is no secret that today immigrants arrive regularly through both formal and informal channels. Many migrants discover old kinship ties with citizens of Mayotte or establish new ones through marriage, procreation, labour, etc. At the same time, and as another refraction of the same economic and political forces, many citizens of Mayotte travelled to other islands in the region, especially to La Réunion (Larunyon), primarily to acquire money, in the form of state benefits, and to Mahajanga, primarily to spend money cheaply. Cultural practices move with people and it is more useful to speak of the distribution of a variety of traditions and communities of practice, such as Sunni Islam, the music of specific Sufi tarikas, or possession by particular kinds of spirits, than of either ‘culture’ in the abstract or distinct societies, per se.

  • 4 At the same time, one should not exaggerate this openness – I keep in mind the massacre and expulsi (...)

10In northwest Madagascar society is even more open.4 Here the collective epithets by which people identify themselves and one another draw from no single set of criteria and do not produce internally homogenous and externally bounded commensurate units either. Indeed, given the bilateral kinship system and exogamous marriage practices, it is striking how many collective attributes any given individual can draw upon. In the port city of Mahajanga it is not unusual to find people whose grandparents have ancestors from four different continents, let alone from different parts of Madagascar itself. As I have come to understand it, ‘Sakalava’ is primarily a political term, indicating some kind of loyalty or subservience to members of a royal clan and the historical kingdoms they founded. But even there such loyalty and service is never exclusive. Conversely, many people whose primary ethnonym is Tsimihety or Merina participate in Sakalava royal service despite the fact that Tsimihety were originally constituted as people who refused to show obeisance to the Sakalava monarch, whereas Merina were political opponents and conquerors. Whatever the pragmatics of use, the various words do not describe homogenous bounded units or mutually excusive identities discriminated by a consistent set of criteria. There is no clearly nested set of social categories of increasing inclusion. Indeed, members of one of the most popular lineages of Sakalava royal spirits, from Betsioko, are identified by means of Merina cultural traits.

11In sum, the places I know best – northwest Madagascar and Kibushi-speaking villages in Mayotte – are characterized by inclusiveness, a receptiveness to new people and ideas, and little inclination to assert strict boundaries, divisions, or orthodoxies [Lambek 1995]. There are families in northern Madagascar who consign one child to Christianity, another to Islam, and a third to ancestral practice. In complementary fashion there are people like the Antankaraña monarch Isa Alexandre Tsimanamboholahy whose very name is constructed from these three traditions that in other parts of the world might see themselves as mutually exclusive rather than mutually incorporative. In Mayotte, adherence to Islam has been stricter, but the Mahorais saw no contradiction in making the practical decision to ask France to incorporate them within the French state. For me, this openness is one of the great attractions of these places, albeit precluding the kind of Dumontian structural anthropology that looks so exciting when practiced elsewhere.

Reciprocal Ritual Action

  • 5 Of course there have been many theoretical challenges to objectification, notably the structuralist (...)

12How then are we as students of the western Indian Ocean to delineate and discriminate among our units of study, to describe the larger culture, society, or social field of which our particular studies form instances? How are we to recognize differences and similarities, continuities and discontinuities, among them? What kinds of theoretical language are most appropriate and least distorting? Do we return to a language of diffusion and the clinal distribution of traits? Do we turn to an analysis of cultural structures and transformations or to a political-economic language of modes of production and flows of capital and governance? And a related question, if collective ascriptions and broad social categories are not particularly salient, what are the cultural means by which social discernment and distinction are carried out? Given that descent is largely bilateral across the region (albeit with emphasis on matrilaterality and even matrilineality in some traditions and locations and patrilaterality in others) and that marriage rules are not prescriptive, neither of these traditional means of anthropological comparison appear adequate. It is striking that anthropologists generally work with nouns; we assume we are studying social or cultural entities and their component objects; this tends to collude with our own local ideologies and mystifications.5 I propose shifting to a model that emphasizes forms of social action rather than global material forces, ideational structures, cultural objectifications, or instrumental and individual practice.

13More specifically, the questions to which I wish to respond are two. First, on what bases are local forms of life culturally constituted, socially reproduced, and historically transformed? Second, along what lines can we both describe continuities between localities and generations and also discriminate between them, in order to both articulate differences and make comparisons, as well as to see how more recent migrants and longer term residents themselves build bridges or maintain separations? My answer to both is inspired by Marcel Mauss, drawing on the subjects that inform two of his famous essays, at least the two most famous in English translation. Thus on the one hand, I consider the extent to which we can see these as societies of the gift (le don) and compare the respective ways in which the gift is constituted and circulates. On the other hand, the question concerns the constitution and reproduction of particular kinds of social persons [1985 (1938)]. Mauss is usually invoked with respect to either the gift or the person, but I understand these to be intrinsically connected: persons are constituted through exchanges and exchanges depend on the right sort of persons. The exchanges in question happen primarily at socalled life cycle rituals, rites de passage, as well as at other collective rituals, and flow between generations as well as within them.

  • 6 For further discussion of acts see M. Lambek [2013].

14Rituals are enacted; hence I focus primarily on kinds of social acts rather than kinds of social groups. I believe that action is also the way to look at the gift. Although gifts are materially manifest in objects, like houses or money, what circulates primarily is not the objects themselves but the acts of giving, receiving, and returning, that is, the acts of producing social persons and acts of witnessing and acknowledging the production of persons by participating in the feasts and prayers associated with social reproduction. In other words, the gift qua material object may be less important than the act of giving and in fact material objects may be largely irrelevant. The cycles of exchange which characterize ritual performances are ones that entail material goods but also various kinds of acts and work – pronouncing blessings, cooking, eating, making music, etc. Most of all they entail the affirmation that comes from bodily presence. What people give of themselves is primarily their acknowledgment of one another.6

  • 7 Elsewhere I have referred to a reciprocity scenario [Lambek 2000].

15In Mayotte the primary object of exchange is the act of praying on someone’s behalf. Acknowledgment of acts of prayer is signified, materialized, and immediately reciprocated in token gifts of cash and cakes, but these are far less significant than the blessings, protections, and other performatives enacted in prayer. These acts are reciprocal and hence social reproduction is mutual.7 However exchange takes place along two axes. Along what one could call the vertical axis, people of adjacent generations produce specific acts for one another – parents marry off their children and adult children bury and commemorate their deceased parents. Along what one could call the horizontal axis, people directly exchange services with one another, neighbours performing the blessings and acting as producers, consumers, and witnesses at the enactment of the rites de passage of each other’s children and parents. Along both axes the gift is located in the act and in the reciprocal work of reproducing and recognizing others and, in so doing, reproducing and recognizing oneself as a social person.

16My focus on acts is informed by the ethnography. When I first went to Mayotte, there seemed to be rituals happening all of the time, and when they weren’t enacting them, people were planning and saving up for them. Thirty-five years later much has changed: now there are paved roads and schools, electricity, running water, and larger, more solid and more ostentatious houses. Subsistence rice cultivation has entirely collapsed and cash cropping radically declined. Many people hold white or blue-collar jobs or have migrated temporarily to receive state benefits at higher levels in La Réunion or metropolitan France. But rituals are still conducted frequently and their number and expense, as well as their performance as an explicit form of exchange between families and communities remain impressive. Their intensity is emphasized by the fact that I return in the dry season, which is the old post-harvest season and now the time students and workers can and do return home en vacances. What people do en vacances is hold rituals and they prefer to hold them back home where they can draw on local garden produce and the labour of kin, but more importantly draw in fellow community members as participants and demonstrate their achievements and receive acknowledgment in the public arena that matters most. The terms and contents of certain rituals have changed somewhat but interest in their performance remains very high.

17Ritual performances in Mayotte include Muslim ceremonies and spirit possession feasts but even more prominent are performances specifically associated with social reproduction – especially rituals that take place at the circumcision of boys, the virgin marriage of girls, and at death and after death to recognize and bless deceased relatives, especially parents. These events have been the focus of a great deal of energy and interest, they are labour intensive and costly in material terms, and each performance engages large numbers of people. Moreover they generally include lengthy performances of Sufi or Sufi-inspired song and dance such as the daira or the mulidi or versions of the maulida, and these performances also have lives of their own, being the means for, or substance of, personal, village and island-wide events as well as occasional events off the island. All these events, ranging in scale from the domestic to the village and island-wide, are marked by elaborate rules of participation for sponsors, producers, principals (on whose behalf the ceremony occurs), performers, witnesses, and guests. They are of great social importance to and for the major protagonists and they serve to constitute complex and salient communities of practice and of mutuality.

18I understand the performance of such rituals to be constitutive of persons and relations in Mayotte. But one way that Mayotte may stand out from many societies elsewhere in the world, or where my analysis may stand out from other accounts of so-called life crisis rituals is that the weight and salience of achievement is placed less on the initiate or principal – the person who is being circumcised, married, buried, or commemorated – than on the people sponsoring the ritual and enacting it on behalf of the principal. That is, the critical people are the parents of circumcised boys and virgin brides and the adult children of deceased parents. In other words, the most salient aspects of adult personhood entail the production and completion of these rituals on behalf of people in adjacent generations. To be a complete citizen, fully respected and self-respecting, to have gone through life well, both men and women need to have sponsored at least the circumcision of one boy, the wedding of one virgin girl, and the commemoration of their parents. Moreover, these rituals implicate and engage the entire community.

19When they first heard I came from a city of two million inhabitants, villagers were dumfounded, saying I must spend all my time at funerals. The point here is the way that social and ethical personhood is constituted through expected forms of collective and reciprocal action, assisting in producing, performing, and witnessing these ceremonies for one another’s kin and thereby constituting communities and localities, as places where people have and are engaged in mutual reproduction. Furthermore, I argue that kinship and community citizenship themselves are constituted through such performative acts.

Comparison with Ngazidja

20Since I did fieldwork first and longest in Mayotte, this has become my benchmark for comparison with neighbouring places and the source for my more ambitious forms of theorizing. However, while I have long considered the importance of reciprocity in the production of such rituals in Mayotte [Lambek 1983 and 2000], it was only upon reading Sophie Blanchy’s [2010; see also Boina 2011] rich account of the Grand mariage in Ngazidja – the Grande Comore – that the relevance of Mauss was fully brought home to me. In effect, Blanchy shows that Ngazidja forms a perfect exemplification of the so-called archaic society of the gift as described and analyzed by Mauss [1990 (1925)].

21I will not elaborate on this here, and Sophie Blanchy may demur, but life in Ngazidja focuses upon these events of a lifetime in which huge amounts of resources are harnessed and dispensed and in which social life is anchored around the houses that are built or transmitted at the Grand mariage. These are equivalent to the so-called ‘stone houses’ of Swahili towns and I expect that one could readily use these events to draw comparisons with life in places like Zanzibar and Lamu as well as Mayotte. The circulation of the gift – those things of greatest value in Ngazidja – is articulated through the Grand mariage, which both reproduces distinctive personnages (as materialized in houses) and produces and reproduces distinctions among persons and families.

22I draw only a few brief points from Blanchy’s complex work. First, the gift functions in Ngazidja to reproduce a sharp social hierarchy. Not everyone inherits or can build such a house or can amass the resources to celebrate such a transformation in the acceptable manner. Hence the Grands mariages correspond to only a small percentage of the total number of weddings in Ngazidja. Second, while circumcisions and funerals also take place, of course, the Grands mariages are far more salient and elaborate than other life cycle events; indeed the latter hardly figure in Blanchy’s account. Third, completion of the Grand mariage makes you a full citizen of the community.

23The situation in Mayotte is broadly comparative to Ngazidja. In both places production, exchange and consumption are focused around life cycle rituals and the building of houses, and in both places these events approximate what Mauss called ‘total social facts,’ insofar as they conjoin kinship with political, economic, aesthetic, and religious means and ends – or rather do not fully distinguish among what in other societies come to be disembedded as such. In neither place is alliance the main subject of marriage rituals. In both places the production of life cycle rituals is understood to produce full adult citizens. However, whereas Sophie Blanchy [2010] neatly shows the division between, as her title says, Maisons des femmes, cités des hommes, in Ngazidja, in Mayotte houses are mainly owned by women, but women and men equally become citizens of their communities through the production of rituals. The gendered division between civic and domestic space is much weaker in Mayotte and women play an active role in political life [Lambek 1983, 1990, 1995, 2000 and 2004; Blanchy 1990].

24There are some further interesting differences. Here I speak not of Mayotte in its entirety but only of villages of the kind in which I worked and which were distinguished from those inhabited by an older Shimaore elite [Blanchy 1990; Breslar n.d.] and now from the influx of economic migrants. During the first decade in which I visited them and for some period before, the ceremonies of social reproduction articulated a strongly egalitarian ideology and were performed explicitly to ensure equality of expenditure between ceremonies of the same kind, and hence what Marshall Sahlins, shifting the language of Claude Lévi-Strauss, referred to as direct reciprocity. To be clear: this was not understood as equality between wife givers and wife takers – categories that were not recognized as such – nor as reciprocity between two distinct affinal groups engaged in marrying their offspring to one another. The point was not a wife-for-a-wife between equivalent kin groups, but rather a feast (or rather, a series of specific, named food exchanges) of a certain size at this wedding for a feast of exactly the same size and contents at other weddings, understood as an exchange among all the adult persons of the community.

25The obligation of strict equality has been dropped since the mid 1980’s, as I will describe shortly, replaced by competitive inflation, but the need and desire to produce such ceremonies remains. In neither period in Mayotte did the system serve to distinguish between those who have the inherited right to perform ceremonies of a certain kind or degree of expenditure from those who do not. All full residents of any given community could and should participate. Most adults do complete the ceremonies and confirm their citizenship. Thus, a major difference between Ngazidja and Mayotte is the use of ceremonies of social reproduction to affirm and reproduce social hierarchy in Ngazidja and social equality in Mayotte. And this was quite explicit on both sides.

  • 8 Circumcision was also compared to the stronger pain of childbirth.

26Another major difference is that while weddings are of great importance and may entail greater expenditures than other kinship ceremonies, in Mayotte other life cycle rituals carry an equivalent importance; not everything hangs on weddings. Thus, there is an explicit comparison between the circumcision of sons and the marrying off of virgin daughters – indeed, based also on an analogy made between circumcision and defloration as ethically salient yet painful acts of irreversible transition.8 Moreover, persons who themselves did not undergo a full marriage could still produce felicitous weddings for their daughters. The production of circumcisions and weddings for ones children, and blessing ceremonies for ones deceased parents are listed together as the three main social expectations of adult life. More specifically, every adult is expected to be the primary sponsor of at least one of each of these kinds of ceremonies – biologically childless adults will take on the first two for nephews and nieces – and they should receive ceremonies after their death as well. Completion of the three ceremonies has been the main goal and sign of achievement of socially responsible adulthood.

27A third difference is that in Ngazidja since it takes many years to acquire the resources for the Grand mariage, the central protagonists, particularly the men but sometimes the women, are often already in middle age. Hence the focus is less on adults producing a wedding for their youthful children than on adults producing a wedding for themselves – possibly long after they have been already married to each other (or to different partners) by less spectacular means. In Mayotte the elaborate wedding celebration can take place only for a wedding in which the bride is a virgin; her virginity is one of the gifts that she herself gives, and especially to her parents, in return for which she receives a house from them. The man may be somewhat older, but over the decades that I have been visiting, the age difference between bride and groom has been decreasing and companionate marriage is becoming the norm. This is related to changes in education and employment. However this also means that the lavish celebration, which is associated with the gifts brought by the groom to furnish the house (as well as jewellery and clothing for the bride) may be deferred and can now take place some months or possibly even years after the Muslim and civil ceremonies and the defloration itself. Today also some adults now build houses for themselves.


28In the work of Mauss and in many of the analyses that follow from him, the gift is understood to circulate in a horizontal and relatively synchronic fashion. It moves between equals or competitors (in which case inflation may be a feature) and what is significant are either the pure social functions of reciprocity and circulation or the political ones of rivalry, exhibition, and shame. While this is true of exchange in the region, what is equally evident is the salience of what Pierre Bourdieu but also Meyer Fortes and Victor Turner describe as the temporal dimension of practice.

29Temporality is relevant in at least four ways. Phenomenologically, it concerns hoping, planning, scheduling, worrying, and disappointment or relief. Structurally, it has a strong intergenerational component, constituted as the articulation of exchanges between the generations over time. Rituals are not simply about people of one generation moving through life, nor simply about inheritance and succession; performance of the rituals is an exchange of services between the generations; what parents enact on behalf of their children is reciprocated by the offspring’s enactment on behalf of their parents after death, although each act is also ethically sufficient in itself and would never be described as direct reciprocity. Moreover each performance has simultaneously marked social effects for the principals and producers of each generation. Thus the ceremonies articulate the simultaneous movement through the life cycle of adjacent generations and reproduce relations of mutuality between them. The transitions and achievements of each generation are dependent upon and inextricable from the transitions and achievements of adjacent generations. One could possibly deduce from this the qualities of dividuality [Strathern 1988] or mutuality [Sahlins 2011] sometimes attributed to kinship.

30The third temporal dimension is that ceremonies and exchanges take place in changing historical and economic contexts – including such historical events and processes as commoditisation of the materials required to produce the ceremonies (and hence fluctuating prices), the mass migration of Grande Comorians and, in a different fashion, people of Mayotte, to France, and changing modes of production. The final temporal dimension is what I would call narrative or chronotopic, the way that the differential weight placed on different moments of the life cycle gives life a characteristic kind of plot. Returning to the phenomenological dimension, it is clear that individual experience is shaped with respect to both kinship generation and historical cohort as well as to the anticipated plot of life.

31The historical dimension can be most sharply captured with the closure of the shungu system in Mayotte. The shungu system was patterned after or drew from the same underlying model as the Grand mariage but was designed explicitly to be open to everyone and to produce and reinforce equality rather than hierarchy and privilege. It enabled what I would call an equality of distinction. I have discussed the shungu and its demise at length elsewhere [Lambek 1990 and 2004]. In brief, the shungu was a formal system of reciprocity that people entered through their age groups by beginning to attend one another’s wedding feasts. Once one began to eat at the feasts of others, one was expected to hold equivalent feasts of ones own, first simply at and for the age group and then at and for the entire community. Each feast of a certain kind had to be offered to exactly the same people (or their designated proxy) at whose feasts one had eaten or expected to eat. Moreover, the feasts of each kind had to be of identical size (irrespective of the number of participants), and indeed the raw ingredients were publicly measured. The people who participated in the shungu were the residents of a given village although entry into the system was by choice and some residents opted out. By participating everyone served as witnesses to each other’s moments of achievement and transformation and thus the lives of everyone were implicated in one another. Once one had completed the production of one’s last shungu obligation and eaten at the last shungu feast to which one had the right of invitation one was replete with citizenship, as it were.

32As villages grew in size and as subsistence cultivation declined, replaced by the commoditization of food and hence fluctuating prices, contradictions emerged between the ideology and practice of equal reciprocity. Eventually, village-by-village, people decided to close their shungu. This meant that the new age cohort did not enter it; people already in the system had to continue to play until the very last feast was reciprocated. That is to say, they stuck by the principles to the end. Change was orderly and in the hands or agency of those undergoing it. While the shungu is gone the expectation of lavish feasts at rites de passage has continued. These are now subject to both inflation and changing fashion. But more important, with the exception of funerals and certain prayers, participation is now by invitation and hence each community has lost, by degree, its internal cohesiveness. People are less fully embedded in one another and kinship and friendship have taken precedence over citizenship.

33In sum I have argued that society in these places is constituted in ritual performances and the cycles of reciprocity they demand and engender. They may express and produce stratification as in Ngazidja or egalitarianism as they once explicitly did in Mayotte and they may be more or less comprehensive and rule-based or selective and individualized. Such acts, cycles, or orders of performance form appropriate axes of comparison and analysis, both historically and between localities. A distinguishing feature here has to do not only with the emphasis on specific performances but also with the kind of liturgical order of which the individual ritual performances form a part.

34If we look at the entire life cycle we see that relative to other moments or other events Grande Comorians place a lot of emphasis on marriage, especially on the Grands mariages. Virgin marriages have and remain very important in Mayotte but they are understood to be in some ways equivalent to circumcisions – if more expensive. In addition, both virgin marriages and circumcisions are compared to the mandeving, the rituals of blessing supplied to deceased parents by their offspring at some point – months or years – after the funeral. If these events are relevant in Ngazidja they are less salient there. They are not discussed in Sophie Blanchy’s book, whereas it would be inconceivable to write a book about life in Mayotte that did not also consider these other events. Coming finally to the third locus of comparison, in Madagascar pride of place goes, of course, to mortuary rituals.

Northwest Madagascar

35I close by considering very briefly whether this model is useful for extending comparison to northwest Madagascar. Here my field experience is based on nine research trips, first briefly among Antankaraña in 1991 and 1992, and later among Sakalava and other people associated with spirit possession and the royal ancestral shrine in Majunga [Lambek 2001 and 2002; compare Feeley-Harnik 1991; Ballarin 2000; Kneitz 2004].

36Most of this work is urban-based and hence I cannot make a strict comparison with the villages in Mayotte. However, a central difference is that the most salient collective ceremonies to produce a kind of ethical and political citizenship are those focused on royalty. Life cycle rituals among ordinary people occur, of course, but in my limited experience they are less directly relevant for collective citizenship. In Majunga there are neighbourhood associations (fikambañana) whose members assist one another at such events but membership is voluntary and without strong social repercussions. Social identity is founded in less formal kinds of practice and is singled out through interpellation by specific ancestors. Collective identity is political insofar as it distinguishes particular rights and obligations with respect to the ancestral polity, including in some instances the right or privilege not to be obligated. The distinction is marked by the difference between the word asa (work) used for rituals in Mayotte and fanompoa (service) for royal rituals in Majunga. One could say that this expresses a difference between citizens and subjects as these are differentially produced through acts of ritual.

37The rights and obligations vis-à-vis Sakalava royalty and other powers are inherited bilaterally and hence kinship serves to reproduce them. However, because everyone has multiple ancestors, they also have multiple collective identities, with overlapping, competing, or conflicting rights and obligations that need to be worked out, demonstrated, and legitimated in practice and over the life course. A chief means for doing so is by practicing or ignoring particular taboos and by bodily responses to their violation. Ambiguity is finally resolved only with the form and means of burial; this is one reason why funerals and subsequent attention to tombs and graveyards are so critical, and indeed more salient than weddings or circumcisions.

38If you hear a noisy procession in Mayotte chances are it is the celebration of a wedding; in Mahajanga it is far more likely to be a funeral. In all the months I have spent in Majunga, whether by accident or not I have never attended a wedding. For non-Muslims, virginity is not an issue – I often heard the remark that virgins simply don’t exist (or at least the passage of virginity is unremarked), as sexual experimentation begins from an early age. For the Muslims I knew it was simply too expensive to hold a virgin marriage. In the course of some twenty years, two of the brothers in the Muslim family with which I live in Majunga did get married with relatively large weddings, as did one of the daughters. The sons were probably in their 30’s at the time. Most other weddings were simply marked by a small invitation and the guests served soft drinks. Likewise, while I heard much talk about royal funerals, any talk about royal weddings was likely to refer to the United Kingdom.

39A social focus on weddings or on the completion of an individual’s shungu obligations certainly produces a different kind of temporality from a focus on funerals. Once a person is buried in northwest Madagascar they become an ancestor of a certain kind, hence with particular social repercussions when they intervene in the life of a descendant. As with weddings in Ngazidja, funerals and accompanying mortuary rites in Madagascar are acts of completion, of reproduction and witness, and are frequently constitutive of social hierarchy. In particular, the production of royal ancestors entails the reproduction of social distinctions between royalty and ‘ancestor people’ (razan’olo) or ‘kinded’ people (karazan’olo) who are expected to service the corpse. There remains talk about how people of one kind (Jingo) might once have even been expected to offer a young person of their own to serve as a human sacrifice and some kinded people serve as permanent guardians in the tomb villages. Spirit mediums are not specifically kinded but they too serve royal ancestors.

40Regardless of their importance for commoners, funerals are absolutely critical for articulating royal kinship and politics. In recent years there has been an intense dispute over royal succession in Majunga that hinges in large part over where the previous monarch was buried and whether the obligations and privileges of the various ancestor people, tomb guardians, royal ancestors, and spirit mediums were satisfied. In sum, the focus on funerals and the deceased, much remarked by other students of Madagascar such as Rita Astuti, Maurice Bloch, Jennifer Cole, Gillian Feeley-Harnik, Sophie Goedefroit and Jacques Lombard, Richard Huntington, and Karen Middleton, helps constitute particular kinds of social order, life trajectories, transitions, and historicity.


41How do we think about society as action rather than structure? Or rather, how do we transcend the boundary between structure and action (or structure and history) that permeated the social structural models that appeared so effective in the anatomy of kinship based societies?

42Following somewhat in the footsteps of Arnold Van Gennep, Marcel Mauss, Victor Turner, Roy Rappaport, and perhaps Meyer Fortes, Pierre Bourdieu, and Claude Meillassoux, as well as colleagues like Sophie Blanchy and Maurice Bloch, I have turned instead to the way rituals qua specific social acts produce persons and society (that is, contract, sameness, difference, equality, hierarchy, continuity, departure, mutuality, individuality, reciprocity, etc.) and also the way people thus formed are enabled to act further (ethically, with respect to goals, exercising judgments, taking initiatives, as citizens or political subjects, etc.).

  • 9 They are also differently materialized, Comorians attending more to production of houses and urban (...)

43It is not my intention to create a static typology based on the kind or quality of rituals (weddings versus funerals) but rather to pay attention to the production, performance and effects of ritual acts. In each setting one can examine and compare society through the reproduction of social persons and the kinds of acts and exchanges entailed in such work. The different weight placed on different life transitions in relation to each other further produces distinctive social formations. These are characterized by different temporalities and different understandings of social persons as dividuals and as public citizens or subjects.9

44If we look further for a language in which to phrase this we might do worse than turn to literary theory and think about social life with respect to distinctive chronotopes, characters or personnages, and modes of polyphony, but we would want to do so in such a way that preserves the essential quality, forces, and effects of action.

Haut de page


Ballarin, Marie-Pierre — 2000, Les reliques royales à Madagascar: source de légitimation et enjeu de pouvoir (XVIIIe-XXe siècle). Paris, Karthala.

Blanchy, Sophie — 1990, La vie quotidienne à Mayotte (Archipel des Comores). Paris, L’Harmattan, ‘Repères pour Madagascar et l’océan Indien’. — 2010, Maisons des femmes, cités des hommes. Filiation, âge et pouvoir à Ngazidja (Comores). Nanterre, Société d’ethnologie.

Boina, Aboubakari — 2011, ‘The Impact of the Diaspora on the Society of the Grand Comoro’, in T. Curtis ed., Island as Crossroads. Sustaining Cultural Diversity in Small Island Developing States. Paris, UNESCO Publishing: 159-176.

Breslar, Jon Haskell — undated [ca 1979], ‘L’habitat mahorais: une perspective ethnologique’. Étude à la demande de la Direction de l’équipement de Mayotte. Paris, AGG Print.

Feeley-Harnik, Gillian — 1991, A Green Estate. Washington DC, Smithsonian Institution Press.

Kneitz, Peter — 2004, Die ‘Kirche der Sakalava’ und die vier heiligen Brüder Andriamisara. Köln, WiKu-Verlag für Wissenschaft und Kultur.

Lambek, Michael — 1983, ‘Virgin Marriage and the Autonomy of Women in Mayotte’, Journal of Women in Culture and Society 9 (2): 264-281. — 1990, ‘Exchange, Time, and Person in Mayotte. The Structure and Destructuring of a Cultural System’, American Anthropologist 92 (3): 647-661. — 1995, ‘Choking on the Qur’an and Other Consuming Parables from the Western Indian Ocean Front’, in W. James ed., The Pursuit of Certainty. London, Routledge, ASA Monographs: 252-275. — 2000, ‘Localizing Islamic Performances in Mayotte’ in D. Parkin and S. Headley eds., Inside and Outside the Mosque. The Anthropology of Muslim Prayer Across the Indian Ocean. Oxford, Curzon: 63-97. — 2001, ‘The Value of Coins in a Sakalava Polity. Money, Death, and Historicity in Mahajanga, Madagascar’, Comparative Studies in Society and History 43 (4): 735-762. — 2002, The Weight of the Past. Living with History in Mahajanga, Madagascar. New York, Palgrave-Macmillan. — 2004, ‘The Saint, the Sea Monster, and an Invitation to a Dîner-Dansant. Ethnographic Reflections on the Edgy Passage – and the Double Edge – of Modernity, Mayotte 1975-2001’, Anthropologica 46 (1): 57-68. — 2011, ‘Catching the Local’, Anthropological Theory 11 (2): 197-221. — 2013, ‘The Value of (Performative) Acts’, Hau 3 (2): 141-160.

Mauss, Marcel — 1985 (1938), ‘A Category of the Human Mind: The Notion of Person, the Notion of Self’, in M. Carrithers, S. Collins and S. Lukes eds., The Notion of the Person. Anthropology, Philosophy, History. Cambridge, Cambridge University Press. — 1990 (1925), The Gift (Essai sur le don). London, Norton.

Sahlins, Marshall — 2011, ‘What Kinship Is (Part One)’, Journal of The Royal Anthropological Institute 17: 2-19.

Strathern, Marilyn — 1988, The Gender of the Gift. Problems with Women and Problems with Society in Melanesia. Berkeley, University of California Press.

Yengoyan, Aram A. ed. — 2006, Modes of Comparison: Theory and Practice. Ann Arbor, University of Michigan Press.

Haut de page


1 I am indebted to the Social Sciences and Humanities Research Council of Canada for extensive financial support, to agreeable research subjects in Mayotte and Madagascar, and to Sophie Blanchy and Laurent Berger for their feedback.

2 See M. Lambek [2011] for an extensive discussion of ‘the local.’

3 The mixture creates a problem for transcription. Here I use ‘u’ for the sound that in French is represented by ‘ou’ and in standard Malagasy by ‘o.’

4 At the same time, one should not exaggerate this openness – I keep in mind the massacre and expulsions of Comorians from Majunga in 1976 (although that was not instigated by northwesterners). The prejudices that operate between citizens of the various islands in the Comorian archipelago or between people of northwestern and highland origin in Madagascar are also evident. Nevertheless, my interest is less in the politics of nationalism, ethnicity, or class than in social ascription and social relations at a more basic level. I am also less concerned with geopolitics or with the material realities and consequences of French colonization, rule, or postcolonial interference and influence, than with local forms, ideas and practices of social reproduction.

5 Of course there have been many theoretical challenges to objectification, notably the structuralist revolution that saw the object as structures of relations and transformations, processual analysis, practice theory, Marxist dialectics, and phenomenological affect. For a very useful overview of the problem of comparison see A. Yengoyan [2006].

6 For further discussion of acts see M. Lambek [2013].

7 Elsewhere I have referred to a reciprocity scenario [Lambek 2000].

8 Circumcision was also compared to the stronger pain of childbirth.

9 They are also differently materialized, Comorians attending more to production of houses and urban plazas and Malagasy to tombs and tomb enclosures.

Haut de page

Pour citer cet article

Référence papier

Michael Lambek, « Ritual as a Social Diagnostic and Lens of Comparison in Mayotte and Its Neighbours », Études rurales, 194 | 2014, 63-77.

Référence électronique

Michael Lambek, « Ritual as a Social Diagnostic and Lens of Comparison in Mayotte and Its Neighbours », Études rurales [En ligne], 194 | 2014, mis en ligne le 01 janvier 2014, consulté le 18 septembre 2019. URL : ; DOI : 10.4000/etudesrurales.10107

Haut de page

Droits d’auteur

© Tous droits réservés

Haut de page
  • Logo Éditions de l’EHESS
  • OpenEdition Journals