Expropriated from the hereafter: The fate of the landless in the Southern Highlands of Madagascar

Abstract

During the period following the abolition of slavery by the French colonial government in 1896, the Southern Highlands of Madagascar was settled by ex-slaves. These early settlers constructed a foundation myth of themselves as tompon-tany, or ‘masters of the land’, a discourse not only equating land with tombs, kinship and ancestors, but also coupled with a skilful deployment of ‘Malagasy customs’. In order to exclude later migrants who also wanted to settle, the ‘masters of the land’ attempted to establish control over holdings in the area. To this end, and to reinforce their own legitimacy as landholders, the tompon-tany labelled subsequent migrants andevo (‘slave’ or of ‘slave descent’) who – as a tombless people – have no rights to land. Because they have neither tombs nor ancestors, the landless andevo are socially ostracised and economically marginalised. As an ‘impure people’, they are not entitled to a place in the hereafter.

Notes

1 By the end of the 1980s, popular opposition to President Didier Ratsiraka gathered momentum. Ratsiraka had ruled the country since 1975 under a highly centralised government outwardly committed to the tenets of revolutionary socialism. In 1989, protests were a daily occurrence in the capital city of Antananarivo, and chronic work stoppages by public officers paralysed the country for months on end. However, it took until 1992 for the opposition to sufficiently unite and force presidential elections. In the first round of the 1992 elections, opposition candidate Albert Zafy tolled 45% of the vote to Ratsiraka's 29%. Zafy easily defeated Ratsiraka in the second round of elections which were held in February 1993. President Zafy's short-lived mandate came to an end with his impeachment in 1996. In 1997, new presidential elections were held. With 50.7% of the vote, Didier Ratsiraka was re-elected President. For an overview of the political processes and events in the 1990s see Roubaud [2000].

2 Any data concerning Madagascar should be viewed as estimates, given the difficulty of obtaining correct information. Most studies focus on cash income, which also might lead to misleading figures, particularly with respect to the countryside.

3 Data from a census of the Malagasy population, carried out in August 1993 by the Malagasy National Institute of Statistics (Direction de la Démographie). In 1998, the population was estimated to be 14 million [IMF, 1999: 3].

4 Industry accounts for only 13% of GDP, while services – particularly those linked to tourism (especially eco-tourism) – accounts for 55%.

5 For both chronological and methodological reasons, the research falls into two periods, the first from 1989 until 1992 and the second from 1996 to 1999. Throughout the initial phase ending in 1992, a mainly socio-economic approach was followed. Upon return to the region in 1996, villagers thought that I was sufficiently ‘behaving Betsileo’ (mitondra tena Betsileo ianao) to be allowed to participate in events that were of cultural import. Most relevant in this respect were the funerals and their accompanying rituals.

6 The concept of tompon-tany is known throughout Madagascar. It is translated as ‘master(s) of the land’ and not ‘owner(s) of the land’ because, despite the fact that their claim to land is considered legitimate by all villagers, they do not have registered titles.

7 Kinship groups are organised around the tomb. Communal rituals related to the ancestors perpetually strengthen their family bonds.

8 In reality, this land only became ‘ancestral land’ (tanin-drazana) by virtue of a razana being placed in the tomb. Razana means both ‘ancestor’ and ‘dead person’ or ‘corpse’[Bloch, 1971: 112]. Those calling themselves tompon-tany say that they never relocated an ancestor from elsewhere with the object of making their present holdings ancestral, but simply waited until a family member died. This is noteworthy, as it does not comply with the common practice in the highlands of transferring at least one corpse from the old to the new tomb whenever a new tomb is erected. This also is consistent with the hypothesis that those who currently claim the status of tompon-tany are actually of slave descent and did not have tombs.

9 Research attempts were made to verify whether the migrants who became tompon-tany were actually of free descent. Most of those interviewed were willing to disclose their place of origin. Follow-up inquiries involved visiting seven villages which were claimed to be the ancestral lands of seven tompon-tany families in the region. A certain number of family members indeed lived there, but it was difficult to conclusively determine whether the villages were in fact their ancestral homelands. Often these relatives admitted that they also were migrants, who had themselves only recently become tompon-tany by establishing a family tomb on their newly acquired land. Unsurprisingly, people with tombs were not reluctant to discuss their family tomb. Indeed, the tomb was a source of pride and an opportunity to showcase their social status. In the tantara (‘tale’ or ‘(hi)story’) about their lives, the tomb generally constituted the point of departure. People of slave descent, however, skirted the topic of ancestral lands and tombs, which might reveal either the absence of a tomb or its recent pedigree. Most slaves lost their tombs upon enslavement. Slaves were prohibited from having permanent tombs throughout the period of slavery. They were forced to bury their dead in perishable tombs [Bloch, 1971: 136].

10 To protect the identity of the villagers, place names and names of individuals are pseudonyms.

11 The claim by these migrants to be Betsileo is accepted by the tompon-tany at face value, whereas their social origin within the Betsileo group is subject to question, requiring a further inquiry into whether their ancestral lands and family tombs are really in the Southern Highlands.

12 Marovato is a migrant village and used here as a case. 36.2% of its population has settled since 1970. They are called mpiavy (‘migrants’) because they do not have family tombs in the region. Subsequently, six more migrant villages in the Marovato region were encountered. The Marovato region includes these and other villages located within a 15 km radius of Marovato.

13 The tompon-tany village council functions next to the formal fokon-tany council. Marovato has one representative in this assembly, who with four members from two neighbouring villages, constitutes the council. The Marovato representative is appointed by the tompon-tany council. Generally, tompon-tany choose one of their relatives. The fokon-tany delegates represent the national government at the local level. They do not interfere with the tompon-tany settlement policy. In practice, fokon-tany councils were found to be passive entities. Its members only held meetings a few times per year.

14 Labels such as tompon-tany or andevo do not necessarily refer to historical realities. ‘Free descent’ is legitimised through acquisition of a tomb, however recent. In the same manner, one is ‘andevo’ by virtue of being tombless. Nothing physically distinguishes the andevo from the other villagers.

15 Bloch [1980: 120] writes on this issue: ‘Without an ancestral homeland one was a non-person.’ Over the last few decades, the tompon-tany have only allowed migrants presumed to be of slave origin to settle in Marovato if they agree to live in the Western periphery. Migrants of free origin may locate in the Eastern part of the village. This would appear to be a general phenomenon in the Marovato region, six Betsileo villages other than Marovato having similar East–West configurations. Kottak [1980: 137–8] also refers to this practice in his study on Betsileo villages around Ambalavao. Dwelling in Western Marovato automatically entails being deemed by other villagers to be a member of an inferior group. Generally, Betsileo ideology considers the West to be the least favourable ritual location [Kottak, 1980: 137–41]. In Marovato, land in the Western quarter is openly seen as impure. No Betsileo of free descent would consider living in the Western periphery. Those who dwell in Western Marovato are referred to as andevo. People with tombs designate them as olona maloto (‘impure’ or ‘dirty people’), whereas they call themselves olona madio (‘pure’ or ‘clean people’).

16 Commonly slave descent of one of the marriage partners was only established after the union and the birth of children (who are seen as ‘impure’). Tompon-tany would force the marriage to be dissolved after such a discovery.

17 The agrarian history of Madagascar is the focus of work by Stephen Ellis [1985] and, more recently, Gwyn Campbell [2005].

18 When compared to other villagers, landless andevo exhibit a number of different behavioural traits. For example, unlike other villagers, none of them ever shook hands. They also seemed to suffer from ‘amnesia’ when asked about their place of origin.

19 In his book on Betsileo villages near Ambalavao, Kottak observes that andevo there are also prohibited from buying land. Tompon-tany only allow them to rent land. Usufruct rights are passed on from father to son but the tompon-tany can reclaim the land whenever he feels like it. This is also the case in Marovato. In addition, Kottak concludes that andevo generally cultivate 0.45 hectare per holding whereas tompon-tany exploit more than twice that (1.2 hectares per holding). Kottak also states that this quantity of land – no more than 0. 45 of a hectare – is insufficient to support the andevo[Kottak, 1980: 135, 163].

20 ‘Ireo razana dia antoka ho ahy fa izao no fomba gasy. Koa raha maty koa aho dia antoka ho an'ireo zanako hiaina ny fomba gasy ihany koa.’

21 In 1996, Marovato had 631 inhabitants, of whom 314 classified themselves as tompon-tany, 151 as migrants (of whom 31 were referred to as andevo), and 166 as labourers or mpikarama (for a further discussion of the mpikarama see also Evers [2002]).

22 This also is the case with those mpikarama who are not actually related to the tompon-tany family head. ‘Kinship metaphors’ can be found throughout Madagascar [Bloch, 1973, 1989; Middleton, 1988].

23 In effect, less tompon-tany land lay fallow after the entry of the mpikarama in Marovato. In 1992, the tompon-tany held 143.6 hectares of farm land, of which 32.8 hectares was leased out and 41 hectares lay fallow. In 1996, the area of the tompon-tany land had remained the same but now they leased out less, namely 28.3 hectares, while only 18 hectares lay fallow. The fact that in 1996 they cultivated more land the tompon-tany attributed to the arrival of the mpikarama. Under the new system, tompon-tany produced more manioc, but because of difficulties in finding buyers for this crop, their income did not increase significantly.

24 For details about the andevo who left the village see Evers 1997, 2002.

25 ‘mitondra tena Betsileo ianao’ translates as ‘behaving Betsileo’, but was phrased in various manners, and often included the qualifiers tahaka or toy, meaning behaving ‘like’ a Betsileo.

26 For the constituent elements of culture, see Keesing [1981: 364–5, 509].

27 The architecture of the tombs is uniform. The materials used in the cube-shaped, top portion of the tombs, visible from above-ground, are stones which are hauled from rock quarries and cut into forms sufficiently flat and small for the walls of the structure. A larger square slab of stone, supported by four columns, serves as the roof. Horns of zebus sacrificed to the ancestors during the funerals are placed on top of the tombs. The portals of the tombs are usually constructed of hardwood (often from merana or nato trees). Some tombs use the more traditional rafeta, a massive flat stone, as the entrance door. The tombs measure approximately 2½ m wide, 4 m in length and 2 m high. The tombs are generally divided into three parts: a vazohon-kady (small entrance and gallery) which leads inside the tomb, the underground hady (burial vault where the corpses are placed), and the aloalo, or upper cube-shaped construction. For more detailed information on the physical construction of Betsileo tombs, see Rajaonarimanana [1979: 181–93] and Dubois [1938: 666–79].

28 The first stage is the funeral itself (fandevenana), where the spirit (ambiroa) of the deceased should enter the tomb. Second comes the ritual of separation (toets'ambiroa), through which villagers lead the ambiroa further on its way to the ancestors. Last comes the ritual fiefana, a feast at the end of the mourning period, which marks the final integration of the ambiroa of the deceased into the general category of the ancestors (razana). The concept of ambiroa has received relatively little attention from those who study Madagascar. Abinal [1885: 211–12] and Dubois [1938: 729] trace the etymology of the word ambiroa to meanings such as ‘being two’, ‘the body double’ and ‘the surplus of two’. The ambiroa leaves the body upon death.

29 For comparative material on Highland funerals, see Bloch [1999: 211–31], Decary [1962], Dubois [1938], Edholm [1971], Kottak [1980], Rahamefy [1997], Rajaonarimanana [1979: 181–93], and Razafintsalama [1983: 193–209].

30 Southall [1986: 414] translates hasina as ‘sacred ritual potency’. He considers it to be a central tenet held by all groups in Madagascar: ‘Here is one of those pervasive themes which justify emphasis on the essential unity of all Malagasy culture, despite its apparent regional contrasts.’ Bloch [1989: 65] also writes that the notion of hasina is the ‘kernel of Malagasy thought’. For eschatological aspects and the ontology of the Malagasy: Bloch [1971], Cole [2001], Dubois [1978], Heurtebize [1997], Lambek [1993], Mangalaza [1998], Molet [1979], Ottino [1998], Rahamefy [1997], Rainihifina [1958–61], Rakotomalala [2001], Thomas [1996; 1997].

31 The concept of hery is highly ambiguous for the Malagasy. Hery commonly means strength and is neither positive nor negative but neutral. It is regarded as a good thing only if one can control it [Bloch, 1989: 65]. Delivré, however, describes hery more in negative terms, as informants in the Marovato area do. He writes [Delivré, 1967: 188]: ‘La signification du hasina dépend donc essentiellement de l'utilisation qu'on en fait: cette puissance n'est bénéfique que lorsque les rapports hiérarchiques entre certains être … sont soigneusement sauvegardés. Mais si ces rapports sont inversés et que le hasina est détourné de sa fin propre, il devient accidentellement une force du mal, et on emploie de préférence le terme “hery” pour le qualifier.’

32 The traditional healers (ombiasy) and their working methods are worthy of a separate study and beyond the scope of this article. It is nevertheless significant that the tompon-tany depend so heavily upon their ombiasy, who they believe have superior hasina to all other living beings. The ombiasy also interpret the wishes of the ancestors and ensure that rituals are executed according to ‘Malagasy customs’. They act as intermediaries between the living and the dead. Ombiasy have the knowledge to manipulate hasina and hery to perform sorcery (called ‘Malagasy medicine(s)’, fanafody gasy). ‘Malagasy medicines’ are usually amulets (ody) composed of herbs, roots, grains and leaves which are infused with ancestral hery by the ombiasy. Tompon-tany rely on their ombiasy to use sorcery as an instrument of power or to avenge wrongs.

33 The fact that kinship groups [Bloch, 1989: 66] possess differing degrees of hasina‘is not the result of their achievement but is given in their nature. The concern of the hasina holders should be to preserve it; creating hasina is out of the question.’ In other words, hasina is reproduced through endogamy.

34 Rasolomanana [1997: 333], working on the Northern Betsileo, also observes that andevo are not allowed to participate in funerals of free descent people. He writes: ‘une femme “tsy madio” (ayant du sang andevo) qui voulait à tout prix s'intégrer dans le groupe des descendants s'acheminant vers le tombeau ancestral, a été enlevée par un tourbillon subit, pour ne plus être retrouvée!’ Unfortunately, he provides no further details on this issue [Rasoamampionona, 2000: 369–75].

35 Bloch also discusses hasina as a source of legitimate or traditional authority. He postulates [1989: 64–5] that each deme (local kin group) of Merina society possesses hasina which corresponds to their rank in the socio-political hierarchy. Bloch makes a division between what he calls hasina mark I and mark II. By hasina mark I, he refers to the supernatural virtue that is possessed in differing degrees by all living beings and whose ‘innate religious superiority’ was concentrated in the Merina king. The hasina and authority of the king was presumably unchallenged, bestowed and sacred. Hasina mark II refers to the recognition of that hasina, i.e. the practice of inferiors rendering homage to superiors. Hasina mark I is a supernatural essence, an innate religious state of superiority, which flows in the form of fertility from the superior to the inferiors, whereas hasina mark II is a natural action, manifested by the giving of gifts, respect and honour, flowing upwards [Bloch, 1989: 67–8]. Several authors have drawn parallels between the political component of hasina and the concept of ‘honour’ in English. Superiors might possess honour but only provided they continue to be honoured by inferiors [Bloch, 1989: 66; Kottak, 1980: 70). Historically, the Merina king infused his subjects with hasina through the ritual of the royal bath [see Bloch 1986: 43–7]. In turn, his subjects rendered homage to the king by giving him the symbolic gift of an uncut silver coin, also called hasina [Callet, 1908: 663–5; Delivré, 1967: 186, Molet, 1956). The replication of this interaction between king and subjects was performed within each kin group through the ritual of tsodrano. This means the ‘blowing of the water’. In this ritual the elder sprays junior family members with water from a saucer he holds before his lips [Bloch, 1989: 68]. In the Marovato region, this practice is still engaged in by ‘free descent’ fathers for their children. As a father's hasina exceeded that of his sons, the rite was performed by elders for their juniors as a form of blessing for fertility and success. The juniors in turn respected and honoured the elders in order not to be deprived of hasina.

36 Even young unmarried girls of tompon-tany families receive a proper burial in accordance with the tompon-tany burial practice.

37 Villagers often compared the impurity of an andevo to that of a dog. This metaphor is common to other Malagasy groups as, for example, the Merina [Ramamonjisoa, 1984: 66] and the Sakalava [Goedefroit, 1998: 117]. The Antemoro refer to the Antevolo, the lowest group in their social hierarchy, in similar terms [Beaujard, 1983: 160; Deschamps/Vianès, 1959: 49].

38 This kind of statistical data is virtually impossible to gather, due to the andevo reluctance to even be identified as a group.

39 Even hasina of the ancestors can change into hery. This is said to occur when people of free descent violate ‘Malagasy customs’.

40 As Graeber [1996: 373] states: ‘The beauty of the tombs is thus visible evidence of incessant ritual activity, which has transformed the hill into a shining display of the newfound power and prosperity of Betafo's mainty. Certainly the andriana themselves have nothing like this. Besides them, even the tombs in the very center of Betafo, the most prestigious noble tombs, looked shabby and unkempt.’

41 The act of witchcraft is called mosavy. The word mpamosavy refers to the person, a witch.

42 Kottak [1980: 175] states that: ‘Within a primarily agnatic village there is a socially recognized difference between the children of sons and those of daughters. Agnates express their hostility to nonagnates, most commonly by leveling accusations of witchcraft against them’. Bloch [1971: 67] goes even further and claims that among the Merina most accusations of witchcraft are typically between unrelated people. In the Marovato region also involve accusations of witchcraft that are principally, but not exclusively, between unrelated people.

43 Ramamonjisoa [1984: 51] stresses that, among the Merina, the mainty (literally ‘black people’, the term connotes slave descent) are more closely associated with witchcraft: ‘Ces pratiques en fait sont communes aux deux groupes, mais à des degrés différents: exceptionnelles chez les hova [‘people of commoner descent’], elles sont plus généralisées chez les mainty.’

44 Villagers think that witches might harm others unintentionally whereas sorcerers are seen as conscious agents (for parallels with Africa, see Evans-Pritchard [1987], Middleton and Winter [1963], and Douglas [1970]).

45 Ramamonjisoa [1984: 51] postulates that free descent people believe the mainty are unable to resist the forces of the occult: ‘inférieurs sociaux que l'on dit ne pas avoir resistéà la force des charmes qu'ils manipulent et qui les domineraient.’ Here, it seems that the occult power that some mainty are supposed to manipulate is derived from charms or amulets (ody). This is not the case for the andevo who have no access to these ody as they are not allowed to see the ‘traditional healers’ (ombiasy).

46 People with tombs can also become possessed by the ambiroa of deceased royalty. This form of possession, called tromba, is seen as a positive phenomenon. When somebody is possessed, he or she makes gestures which remind people of the late king or prince who is manifesting himself through the possessed person. Possession endows the possessed with healing powers. Villagers say that it is the hasina in the ambiroa of the nobleman that is at the disposal of the possessed for the duration of the possession. They believe that the possessed person only has to touch ill people in order to cure them [Estrade, 1977; Jaovelo-Dzao, 1996; Sharp, 1993]. No such possessions were observed in Marovato. Conceptually, witchcraft might be viewed as the counterpart of tromba. In witchcraft, the ambiroa of a deceased person also possesses somebody, but instead of being filled with hasina it is filled with hery.

47 On Malagasy witchcraft, see Bloch [1971: 65–8], Danielli [1974: 261–76], Dubois [1938: 907–11], Feeley-Harnik [1978: 410], Gardenier [1986: 337–52], Graeber [1996: 416–39], Grandidier [1920: 476–8], Kottak [1980: 212, 215–7], and Lambek [1993].

48 The witch is ‘l'indésirable par excellence’[Dubois, 1938: 908], a position similar to that of the andevo. And just like the andevo who managed to get married to people with tombs, witches also are not allowed to enter the tomb [Callet, 1908: 273; Dubois, 1938: 679–80].

49 The discourse about witches as described by Kottak [1980: 212] for the Betsileo in the Ambalavao region is very similar to what was observed in Marovato. He states that people view witchcraft as the dark side of hasina (‘beneficial occult energy’), and as such can be seen as the systematic inversion of Betsileo ceremonial life: ‘However, the Betsileo also must deal with the dark side of the force, the occult powers controlled by sorcerers and witches, people who mock the social order by using its most potent symbols (e.g. food and tomb, major symbols of kinship) in inappropriate contexts.’

50 Because villagers believe that witches are possessed by hery, they fear witches can spread illness or death by a mere touch or look. Tompon-tany often stated that they possess sufficient hasina to repel hery attacks, but prefer the path of prudence by keeping at a safe distance from witches. For children, even those of tompon-tany families, witches present a bigger danger, because their hasina is undeveloped. Most parents provided their children with amulets filled with ancestral hasina, which were fashioned by a traditional healer. When a person is believed to be the victim of witchcraft, the ombiasy would also prepare an amulet to neutralise the hery afflicting the victim.

51 This raises the question as to whether ‘andevoness’ exists in other parts of Madagascar, on which see Evers [2002] and Rakoto [1997; 2000].

Additional information

Notes on contributors

Sandra Evers

Financial support for the fieldwork was provided by the Research School CNWS (Leiden University), the Netherlands Organisation for the Advancement of Tropical Research (WOTRO), the Netherlands Organisation for Scientific Research (NWO) and the Amsterdam School for Social Science Research (ASSR, University of Amsterdam). Thanks are extended to Tom Brass for valuable editorial advice which facilitated the revision of an earlier draft.