The Betsimisaraka Legacy

One of the intriguing features of Tanambao, Ambatobe, and other villages on the northern Masoala Peninsula is their well-established pattern of ethnic mixing. Makoa, Antaimoro, Anjoatsy, and Betsimisaraka all live side-by-side, and some mix of these often forms the heritage of a single individual. However, despite their recognition of other ethnic identities, Betsimisaraka remains the second-most common label people provide when asked their ethnicity (karazaña), and many who cite multiple ethnicities or clans begin with or include Betsimisaraka. Historically, the east coast of the Masoala Peninsula was settled early by members of the Betsimisaraka Confederation. The port at Andrombazaha¹¹

was once a stopping point for European slave ships going on slave raids to the Comoros, which are said to have been manned primarily by Betsimisaraka. Furthermore, fluidity of identity is one of the important features of Betsimisaraka social life, in both the past and the present (Lahady 1979; Mangalaza 1994). Thus, there are features of Betsimisaraka history that might be useful in understanding contemporary inter-group relations on the northeast coast. Clearly, Betsimisaraka identity alone does not sufficiently explain the level of integration between Betsimisaraka and Makoa, as shown by persistent divisions between the two ethnic groups in the eastern village of Fasina studied by Fanony (1975). However, I would argue that the relatively non-centralized nature of Betsimisaraka organization became important in the context of other factors particular to this region: abundant resources and high levels of in-migration of other ethnic groups.

Until the early eighteenth century, there was no group known as the Betsimisaraka. Due to the region's steep hillsides, dense forests, and the many rivers that impede north-south travel, there was limited casual interaction among the several small isolated groups that inhabited the various river valleys (Esoavelomandroso 1979:33–37). Social organization was primarily lineage-based. In times of celebration or warfare, the lineages would unite under the control of a chief (filoha) into clans, primarily for the purpose of cattle raiding (Cabanes 1982a). Each clan was connected by its common ancestry, territory, and taboos (fady). Occasionally, clans united as a larger confederation or tribe under the leadership of a single “big chief” (filohabe), staging raids and counter-raids, sometimes using human captives to bargain for the return of stolen cattle. Less frequently, all of the tribes would unite in raids or wars against inland peoples, usually with the aim of obtaining cattle or human captives. In these cases, the capture of humans was intended to humiliate the opposing group, not to bargain with them, and there was no expectation that the captives would be returned.

Until the early eighteenth century, all the lineages along the northeast coast had mostly unimpeded access to ports and could trade directly with Europeans, who were attracted by the abundance of natural gemstones, cattle, rice, locally produced rum, ebony, and other desirable hardwoods. Additionally, the state of almost perpetual warfare and raiding that existed on the east coast and the well-established practice of enslaving the captives of war helped to provide a steady supply of slaves, which were sought by the Dutch in the seventeenth century, and later by the French and the British to provide labor for their Indian Ocean island colonies.¹²

Initially, a roughly equal balance of power was maintained among the northeastern lineages as they supplied the European trade (Cabanes 1982a:159). However, in the early 1700s, there emerged a single powerful chief, Ratsimilaho, who had the personal charisma, beneficial lineage connections, and political sensibilities to unite the disparate lineages into the Betsimisaraka Confederation. In 1712, he obtained an oath of loyalty from many of the northern lineage heads. It was at this time that they became known as the Betsimisaraka: “the many who will not be separated.” Making strategic marriage alliances, first with the daughter of a powerful lineage chief of the southern Antatsimo tribe and later with a Sakalava princess, Ratsimilaho led the Betsimisaraka to become the most powerful force on the east coast. However, lacking his charisma and having no other basis of power, Ratsimilaho's successors failed to keep the Confederation together after his death in 1750. Divisions within the Confederation were exacerbated by French traders desirous of the slaves that resulted from internal conflict (Rochon 1971). By the end of the century, the Confederation collapsed into lineage and tribal warfare and had lost much of its power (Deschamps 1972:106–7).¹³

Trade returned to a focus on lineage-based acquisition. In 1785, in order to meet the rising demand for slaves by European traders, Betsimisaraka¹⁴ began to conduct external slave raids to the Comoros islands, bringing the slaves back primarily to Nosy Boraha (now Ile Ste. Marie) for sale or trade, with each participating lineage having an equal share in the returns.

While the Betsimisaraka were raiding one another and the Comoros, the Merina empire, based in the central highlands, continued to expand. They increased their control of the slave trade, capturing Malagasy in the highlands or in western Madagascar and also taking slaves from East Africa across the northern part of the island and selling them in markets on the northeast coast (Campbell 1993:140). However, by 1825, the Merina were actually attempting to prevent the export of slaves from Madagascar because of the labor needs of their own empire, and they shut down the trade out of east coast ports.¹⁵

The hostile reaction by the French and by Creole planters on Mauritius and the Comoros led eventually to violent Merina retaliation against coastal populations that supported the French against the Merina. It is estimated that between 1828 and 1840, over 200,000 people were enslaved as a result of Merina aggression (Campbell 1981). Those who were not enslaved were subjected to abuse by Merina administrators and soldiers posted throughout Merina-controlled regions of Madagascar. During this time, thousands of Betsimisaraka refugees fled Merina domination, settling on the French-controlled island of Nosy Boraha (now Ile Ste. Marie), ultimately being recruited into contract labor on Mauritius (Campbell 1981).

From this history of the Betsimisaraka and their relations with one another, with European traders, and with the Merina, an overall impression emerges of a group that developed neither a strong sense of solidarity nor a well-formed, ideologically supported social hierarchy. Their “royalty” was not associated with rituals uniting its followers and legitimating their leadership. Trading and raiding were conducted by multiple groups with an entrepreneurial spirit rather than through the guidance and control of a single centrally organized political leader. Ultimately, their reaction to Merina domination was dispersal rather than unified resistance. In this past, we see precursors to many of the patterns we observe today among people, many claiming Betsimisaraka identity, residing in northeastern Madagascar. They practice open forms of kinship, allowing marriage with all other ethnic groups except the Merina. Their relations of authority are primarily lineage-based, with few matters requiring broader coalitions. And, just as members of early Betsimisaraka groups set a precedent in the northeast by conducting long-distance raids and collecting in local forests to collect captives and goods for trade with foreigners, there is no stigma in the region today in surviving by a mixture of subsistence farming and market transactions.

Population Flows in the Nineteenth and Twentieth Centuries

To talk about history with contemporary villagers in northeastern Madagascar is to talk about movement. People came to the Masoala Peninsula and the surrounding area for a variety of reasons. Most villagers say their grandparents and great-grandparents, if not enslaved, came in search of wealth (mitady). For some, trade with Europeans and Creoles was their primary incentive for migration. One villager's grandparents made money by entering the forest to gather nuts, from which they then extracted oils to sell to foreigners (vazaha), while another villager reports that early settlers used to earn a living doing “what ‘true Malagasy’ did”: they collected rubber and ebony for sale to Europeans. Other people moved to the Peninsula to work for wages on foreign-owned concessions and agricultural plantations.

In the late nineteenth century, Merina prime minister Rainilaiarivony granted several forest concessions to European and Creole settlers on the Masoala Peninsula (Esoavelomandroso 1979; Petit and Jacob 1965; Valette 1966). Among these concessionaires was M. D. Maigrot, a Creole from Mauritius, who received 6000 hectares on the Masoala Peninsula, the largest concession in the region and encompassing the area around Andrombazaha. In 1898, with Madagascar under French rule, an observer noted that Maigrot employed “a certain number of Antaimoros”¹⁶

to collect ebony and rosewood, and that the area around Andrombazaha had been denuded by Malagasy people working for Maigrot and clearing fields for themselves (Chapotte 1898:885–86). Thus, this appears to be one site to which migrants were attracted in their search for money and opportunity.

Under French rule, many foreigners received concessions to develop agricultural plantations. For those Malagasy seeking money to pay taxes, work on these plantations provided needed income. For example, in the nearby Andapa region in the northeast, vanilla plantations attracted migrants seeking work (Cabanes 1982b; Laney 1999). In the area immediately around Ambatobe and Tanambao, current villagers report that a foreigner named Pastelaire developed his concession into the first large coffee plantation in the vicinity and hired workers to farm both coffee and rice.¹⁷

Some contemporary villagers descend from these workers. One such woman reports that when her uncle began his military service, he sent his children to work for Pastelaire, who, he said “protects his workers. Murderers and thieves can't get to them there.” Later, when coffee, cloves, and vanilla became available for villagers to plant for themselves, they readily adopted them and created their own small cash crop plantations. Since Madagascar's independence in 1960, the pattern of migration in search of money and opportunity in the northeast has continued. In 1967, Les Grands Moulins de Dakar received seven parcels of land in the sub-prefecture of Antalaha for timber extraction.¹⁸ This brought better roads and jobs to the region. It also brought people. In May of 1970, that concession was revoked, but since the 1990s, other exporters operating out of Antalaha have attracted workers willing to collect ebony and rosewood from peninsular forests.

In sum, the inhabitants of this region have been people who were willing and in many cases forced by circumstances to work for wages, to exploit forest products for sale, and to grow their own cash crops. Perhaps because there were so many migrants and so few long-term settlers, settlement patterns in this region have not followed the pattern set in other agricultural regions of Madagascar where being a “first-comer” continues to carry special status. In those regions, working for wages and working with cash crops is often seen as a sign of submission to an external, polluting force (Feeley-Harnik 1991; Sharp 1993). This is not the case in the northeast. Nor is there much rationale for a general hierarchy of first-comer over later arrivals. Here, the history of movement means that one of the primary connections contemporary villagers have to their past is the necessity to be flexible and the ability to move. Many people did not migrate here with the intention of returning to their homeland. As one of the current residents of the villages says, “Once people settled here, they were just from here. We're all just from here.”¹⁹

And once they were here, they continued to be on the move, due in part to the physical environment and in part to the social and political environment.

When asked about the relationship between Tanambao and other nearby villages, villagers today recount a history of back-and-forth movement due to epidemics and floods. Four villages were founded by families fleeing a flu epidemic in Tanambao in the early years of the twentieth century. One of those villages was later deserted when people escaping floodwaters migrated back towards Tanambao and established Ambatobe. Politically, from the times of Merina rule to the present, the region has drawn the attention of government officials interested in controlling both forest resources and the human population. For example, the French colonial government targeted “slash-and-burn” agriculture as one of the destructive practices that needed to be eliminated, and they produced numerous laws forbidding the burning of forests by farmers (Althabe 1969; Jarosz 1993). Additional attempts to conserve the forest led to the creation of Réserve Naturelle No. 2 on the Masoala Peninsula in 1927. These policies served to channel migrants to denser settlements on the coastal plains, which also allowed the government to more efficiently monitor their activities. There, local people could farm irrigated rice fields, but these were neither numerous nor large enough to sustain the population that had been pushed onto them. The shortfall in rice production further fueled the drive to make money through cash crops and wage labor. Local people today refer to this time as one when they or their ancestors temporarily moved away in search of land and work.²⁰

In 1964, the reserve was declassified and the forest reopened to public use, drawing a new wave of settlement.

Thus, we see that since their beginnings these villages have been continually in flux—laws change, and boundaries shift. People are pushed and pulled to test new crops and farming practices. Entire villages relocate. These processes continue today. Violent cyclones in March and April of 2000 caused Tanambao to rebuild on higher ground, distancing itself geographically and politically from Ambatobe residents who chose not to move. Vanilla and clove prices have soared, inspiring farmers to expand existing plantations and convert rice fields to cash crops. A new national park was created on the peninsula in 1996, bringing with it new regulations governing the use of forest resources. Under these conditions, the logic of a social hierarchy based in longevity on the land, so common in other regions of Madagascar, makes little sense. Here, strength comes from the ability to take advantage of new opportunities, to negotiate effectively with neighbors, and to be willing to move to new ground when necessary. These capabilities are not particular to any one ethnic group or to any one ancestry. Under these conditions, even someone of slave descent can participate fully in community affairs, and sometimes even become a leader.

Behoraka's Story²¹

This is not his real name. Life histories and other interview data were tape recorded in the local Betsimisaraka dialect and later transcribed. Extended interviews were conducted with seven elders regarding village history. Their stories were very similar to one another, with Behoraka's being the most detailed.

Behoraka is one of the oldest members of the village of Tanambao. His version of village and regional history stresses the complex relationship between slave and free since abolition in 1896:

Later in our discussion, Behoraka discusses his grandfather's reasons for first moving away from Andrombazaha to the eventual site of Tanambao. At this point, he reveals that his great-grandfather was himself a slave owner who used slave labor to construct rice fields on land concessions he received from Merina rulers.²⁴

These were the fields to which Behoraka's grandfather moved near Tanambao. While Behoraka's story my overstate the grandeur of his own background, its broad details are not disputed by other villagers and it serves as a guide to what is considered a socially acceptable historical account of the origins of these villages. Furthermore, his story raises issues, such as the nature of social identity in the past, that can be pursued further for what they might reveal about contemporary practices and perceptions.

Behoraka's story introduces two features that are central to understanding the open relationship enjoyed between people of slave descent and people of free descent in these villages today. First, ethnic identity is expressed at a very personal level, largely through adherence to taboos. Everyone, including Makoa, has taboos that are particular to their clan or ethnic group. There is no broad social pressure to follow the taboos of a single dominant ethnic group or to otherwise try to attain membership in a particular group. I will argue that this type of identity expression is a reflection of the non-hierarchical, highly decentralized nature of regional social structure. Second, the relationship between slave and free historically bears the mark of client-patron relations, similar to the type of relationships described by Kopytoff and Miers in their essay discussing slavery as an institution of “marginality.” This, I will argue, is the significance of the lohateny relationship repeatedly mentioned by Behoraka as an important factor in his grandfather's sense of obligation to the Makoa. I address each of these issues—the absence of social hierarchy in the present and the presence of patron-client relations in the past—in the following sections.

The Personalization of Ethnic Identity

In Behoraka's discussion of history and in his description of contemporary village life, he reveals that he married a woman of Makoa origin, a granddaughter of one of the ex-slaves who settled here upon emancipation. Currently, he refers to her as being, along with him, one of the leaders of the village, and others in the village agree with that assessment. Her position as a village leader of Makoa origin is not uncommon. At the time of my research, such a man was considered the primary mpijoro, or ancestral prayer leader, for Tanambao. Furthermore, the unexceptional nature of marriage between people of free and slave descent provides evidence of the lack of stigmatization of slave heritage in these villages. Examination of my household data reveals that nearly three-fourths of all marriages in which Makoa is an identity of at least one spouse involve marriage between someone of Makoa descent and someone of other descent, usually identified as Betsimisaraka, Betanimena, or Antaimoro.²⁵

As an additional measure of their local status, it is notable that members of these households possess land in comparable quantity and quality to members of non-Makoa households.

Behoraka's story provides a glimpse of another widespread feature of identity in these villages. When he and his grandfather are identified as Antimoroño, the claim is supported by providing a list of the taboos they follow. It is common in many parts of Madagascar for ethnic identity to be understood as something an individual achieves through performance.²⁶

Taboos play a particularly powerful role in this process, as adopting them can be a way of signaling one's membership in a community of fellow adherents to ancestral or village proscriptions (Brown 1999:248–62; n.d.; Lambek 1992; Lambek and Walsh 1999; Sharp 1993:57–58; Walsh 2002). Behoraka's way of describing his grandfather's identity is mirrored in many other discussions I had with villagers about their own identities. Frequently, a villager would immediately follow an ethnic label with a list of the taboos she follows. Likewise, when asked the differences between various ethnic groups, villagers always began with taboos, followed by place of origin (usually north or south), and then appearance. Taboos, however, are not seen as an indicator of marital compatibility or social status.

If two people with different taboos marry, their children should follow both sets of rules. All taboos are not followed equally strictly, but decisions about which to follow are made on a practical basis. For example, when abandoning a taboo work day, a common justification was, “once my spouse and I combined our work taboos, we didn't have enough time to get our work done.” And when one man started eating sea turtle, breaching an ancestral taboo, he explained, “They are abundant here and easy to catch,” unlike in the region of his clan's origins. In more stratified communities, taboos can be used to indicate social rank. For example, in the northwest, a single ethnic identity is often understood to correlate with the category of tompon-tany, or first settlers, of a village or town, and tompon-tany exercise considerable control over land and labor access. Therefore, newcomers may attempt to assume the identity of the dominant group through marriage (and subsequent childbirth), adoption, or even spirit-mediumship (cf. Sharp 1993) which can justify adherence to the dominant group's taboos.

While taboos are important throughout Madagascar, in some regions there are additional means of performing identity and establishing one's social status. For example, among the Merina, the Sakalava, and the Antankarana, performing royal rituals is one way group members mark their identity as separate from, and in many cases superior to, non-members (Bloch 1986; 1987; Feeley-Harnik 1984; 1991; Lambek and Walsh 1999). Alternatively, not performing certain tasks can also be seen as a marker of status. Both Lesley Sharp (1993) and Gillian Feeley-Harnik (1991) document that among the Sakalava in northwestern Madagascar, avoidance of wage labor is perceived to be an important indicator of the higher status that Sakalava indigenes enjoy over more recent non-Sakalava immigrants, some of whom moved to the region primarily in search of wages.

Ranked societies in Madagascar exhibit particularly clear discrimination against people of slave descent. Among the Betsileo, with their clearly demarcated hierarchy of nobles, senior commoners, junior commoners, and slaves, meaningful social distinctions continue to be made on the basis of free and slave ancestry. Descendants of slaves depend on people of free descent for land access, and they also suffer social stigma. They are not allowed to intermarry with nobles or commoners. “The commoner who marries a slave loses the right to be buried in his family tomb” (Kottak 1980:103). In everyday life and on ritual occasions, andevo are given tasks that emphasize their lower status. If possible, people of slave descent attempt to hide their heritage from outsiders (Kottak 1980:20; Evers 2000). The neighboring Merina possess an even stronger hierarchical ideology than the Betsileo, one that is reinforced through norms of deme-endogamous marriage (Bloch 1971). The Merina are also noted as a group among whom knowledge of slave descent is a nearly insurmountable burden, one that is marked by the use of the terms fotsy, or white, to refer to people of free descent, and mainty, or black, to refer to people of slave descent (Bloch 1971; Graeber 1996; 1999). Thus, it seems that one commonality among at least some of the Malagasy groups who continue to place a strong stigma on slave ancestry is a well-established social hierarchy, with no accompanying ideology that allows for a slave to move up the ranks.²⁷

In contrast to more hierarchically-organized communities, the villages of Ambatobe and Tanambao are noteworthy for the wide variety of identities they are free to express, primarily through taboos. There is no identity that emerges as more desirable than the others, no identity that, for status reasons, needs to be suppressed. People of Makoa, Betanimena, Antaimoro, and Anjoatsy origins can freely acknowledge any one or all of their identities, without having their choices determined by community politics. In Ambatobe and Tanambao, people who freely admit slave descent and openly practice their ethnic identities enjoy landownership, marriage with non-slaves, and the assumption of village leadership roles. They have been able to overcome their marginality in both kinship and societal terms (cf., Kopytoff and Miers 1977:16). In the context of a stratified social structure with an entrenched social hierarchy based in group origins and ethnicity, an individual can only overcome the stigmatization of slave descent by denying the past and hiding their identity. Where such hierarchy is absent, people of slave ancestry have opportunities that more closely match those of people of free descent.²⁸

Behoraka's marriage to his Makoa wife and the eventual inheritance of his land by his children of mixed Antimoroño and Makoa descent are extensions of the type of open system discussed in Kopytoff and Miers (1977), Reid (1983), and Watson (1980). In the context of an open kinship system, exemplified here by the absence of strict marriage rules and little felt need to preserve group boundaries, slaves have greater opportunities for incorporation into the lives of the free.

While open kinship and lack of hierarchy provide opportunities for freed slaves, we need further information to understand why and how such opportunities developed in this region and not in others. One of the factors we should explore is the type of relationship established between slave and master. As mentioned earlier, there is evidence in northeastern Madagascar of a type of patron-client relationship between ex-slaves and their former owners. This must be understood in the context of the broader political and economic relationships between slaves and their owners, and among the different categories of freemen that co-existed during the period of slavery.²⁹

To examine these issues further, I return now to the concept of lohateny.

Slavery as an Institution of Dependency

Much has been written about slavery and paternalism in Africa.³⁰

To suggest that a slave owner behaves paternally towards his slave is not to deny the basic facts that to be enslaved is to be removed from one's home and to be exploited for the benefit of others. However, beyond those facts, there is much variability in the type of relationship slaves have historically had with their masters. At one extreme, slaves could be thought of as useful primarily for their economic productivity and treated as merely replaceable laborers. At the other extreme, slaves could be considered useful in the ways they helped to expand a family unit and, thus, they could be treated almost like kin. There is considerable evidence to suggest that slavery in northeastern Madagascar more nearly approached the “kinsman” end of that range than the pure laborer end of the range. Yet, to be kin here, as in much of the world, does not imply equality. Rather, relationships among kin are characterized by a strong sense of age-based hierarchy, and family members have a strong sense of their “belonging to” as well as “belonging in” their paternal and maternal kin groups (cf., Kopytoff and Miers 1977). Thus, it is entirely plausible that Malagasy slave owners could develop a hierarchical relationship with their slaves while also displaying towards them a paternalistic sense of responsibility for their well-being. Lohateny may be one survival of those early slave-owner relations.

Contemporary villagers describe lohateny as a set of ritual relationships that are a remnant of slavery.³¹

In recent times, when one meets a lohateny partner, one is supposed to provide for him or her whatever is asked, whether it be cattle, money, assistance in curing an ill, or food. Furthermore, it is a reciprocal relationship in which the request may be made by either partner. Lohateny, as practiced in this area after the end of slavery, may have been a means to maintain peaceful and practical relationships while masking power inequalities.³² Although on the surface, the relationship appears equal, when Behoraka talks about the lohateny relationship between Antimoroño and Makoa, he uses a language of ownership and of dependency. His Antimoroño grandfather, Sily, had to take care of “his” Makoa; the freed slaves came to Behoraka's grandfather seeking land and assistance.

Why would the Makoa have pursued such an arrangement after having been granted their freedom, particularly since Sily and his father had themselves owned slaves and used slave labor to create their rice fields? Here, I examine three possible reasons: (1) the autonomy granted to slaves by their owners; (2) the relationship between coastal Malagasy and Merina administrators; and (3) the set of opportunities available to freed slaves. To discuss the first two reasons, I begin by sharing the observations of Francisque Coignet, a French mining engineer who visited northeastern Madagascar in 1863.

Coignet's mission was to explore the region between Cap d'Ambre and the Bay of Antongil on Madagascar's northeast coast. When he was greeted at Andrombazaha by the Merina customs officer and his soldiers, Coignet misrepresented his mission to them, fearing that acknowledging his ties to the French government would prevent him from gaining access to the interior of the Peninsula. Claiming he was the son of a Mauritian merchant, he was given access to some “Ova” (Merina) guides and allowed to explore some of the region. He noted a considerable Merina presence in the area. On a journey to the north of the river along which Ambatobe and Tanambao are now located, he observed several settlements. One fortress was described as comprising 700 to 800 people, “ova officers, some natives, slaves, women and children.” The army there consisted of 158 men, 80 soldiers and 78 officers. The soldiers were armed with spears and guns (Coignet 1867a:266–67). As Coignet continued northward, he described a “pasture where there are magnificent herds of cattle, and nearby are several houses inhabited by slaves” (1867a:268). He observed more villages of slaves elsewhere in the region, and reported that both non-Merina Malagasy and Merina possessed slaves. Some were assigned to domestic work and lived in the village of their owner, while others were spread across the countryside guarding their owners' cattle and working their land. Local people estimated the number of slaves at two per every one free person (1867b:364).

Coignet enumerated four ways local slaves were obtained. The largest number were inherited. He describes these slaves as being treated much like family, and constituting the greatest wealth most native families possessed.³³

A second way of falling into slavery was to be convicted of committing a crime. A family could purchase such a man's freedom back, but if they failed to do so, he and his children were forever slaves. Coignet reported that most slaves of this type were owned by “ova,” who purchased them from commanders in the provinces and speculated in their resale. A third category of slave was brought from the African coast and was valued much more highly than slaves of Malagasy origin. The fourth way of obtaining slaves was through warfare waged by the Merina, as defeated populations were placed into slavery (1867b:364–65).³⁴

In addition to his discussion of slavery, Coignet provides ample evidence of the harsh treatment to which local free Malagasy were subjected at the hands of the Merina. While it is clear that one of Coignet's goals was to discredit the Merina administration and offer justification for an increased French presence in Madagascar, his account of local relations is consistent with other historical accounts of the period (cf., Esoavelomandroso 1979). As required under Queen Ranavalona I, native people had been forced by the Merina to pay tribute and taxes and to perform corvée labor extracting gum copal and rubber. The commanders of the provinces were all Merina, and they abused their administrative powers, acting as judge and jury in cases of crimes committed by members of the local population. Many local Merina administrators got rich by fabricating criminal charges against members of wealthy native families. If a man could not pay the fine levied against him, he, along with his wife and children, could be sold into slavery (Coignet 1867b:340–41). Because of the harsh treatment by the “hova,” many coastal Malagasy fled the area, moving to the northwest to align themselves with the Sakalava, or heading south to the Fenerive region, where they worked at the ports or cultivated rice for sale to European traders (1867a:281–82).

Coignet's observations of slavery and of the relationships between Merina and local Malagasy people provide critical evidence regarding the relations between slave and free coastal peoples as well as between different free populations. He observed that slaves of indigenous peoples seemed to live independently in villages near the sites where they pastured their masters' cattle. The evidence of dispersed slave settlements, and of slaves being assigned tasks such as cattle herding that allowed for some independence, suggest that slave masters were more reliant on voluntary compliance on the part of their slaves than on regulating them through force. Such compliance could result from characteristics of slaves themselves that would make cooperation among them difficult, such as those summarized by Klein as posing barriers to coordination in French West Africa: “ethnic diversity, the lack of kinship ties, and divisions between old and new slaves” (Klein 1993b:183). However, slave compliance could also result from the achievement of a kind of patron-client relationship between masters and slaves, in which masters were expected to provide for their slaves' needs in return for slaves' provision of labor and loyalty. Thus, it is possible that Coignet was not glossing the truth when he asserted that the native (non-Merina) slaveholders treated their slaves “much like family.” To understand how this could be the case, we have to examine the incentives and pressures as experienced both by both slave masters and slaves.

In the context of the Masoala Peninsula, two factors emerge as potential incentives for masters to develop a more kin-like relationship with their slaves. One concerns the nature of the local physical environment, the other the nature of the local political and social environment. The physical environment of the Masoala Peninsula in the nineteenth century was densely forested and fragmented by many hills and rivers. An escaped slave would be difficult to track if he or she disappeared into the heart of the forest. Regarding the local social and political environment, there is no reason to believe that there was a great deal of social solidarity among slave owners on the Masoala Peninsula in the nineteenth century. The most obvious cleavage was between the indigenous coastal peoples and the Merina, who would most likely keep any escaped slaves they recovered. Furthermore, the migration history of this region, as discussed earlier in this paper, involves the convergence of many different clans and ethnic groups into this geographic space. This means that someone's nearest non-kin neighbor might have been performing unfamiliar religious rituals, speaking an unfamiliar dialect, and committed to unfamiliar ancestors. Under such socially fragmented circumstances, assistance from one's neighbors in recovering an escaped slave was less likely than under circumstances when slave owners had a stronger sense of collective identity. For these two reasons, slave masters would have been more likely to prefer incentives to sanctions as a way of maintaining their slaves' compliance and to avoid losing them through escape.

The next question to ask is why the slaves themselves would not have taken advantage of their opportunities to flee. First, we have to consider the fundamental fact of slavery: slaves were outsiders, with no local allies. As Kopytoff and Miers usefully point out regarding many African societies, to be free is not to be independent; rather it is to belong, and belonging entails both rights and obligations (1977:17). Not only social identity, but also physical survival, are built upon one's ability to draw on the commitments inherent in long-term relationships of interdependence. This is also true in Madagascar. The “lost people” of Madagascar had no means to return to their homelands and their webs of dependency. Therefore, to leave their masters was to risk being alone in the world. As long as their masters were providing for their needs, such a risk may have been too great to bear.

One of the needs slaves had was to be protected from the Merina. A slave who considered escaping also had to consider that he might be recaptured by a Merina administrator and forced to do more arduous labor, with fewer rights, than under his current owner. Or he might be sold and taken off the island. As Coignet's report shows, the Merina did not view their non-Merina counterparts as fully human. In a very real sense, the slaves and the coastal Malagasy may have seen themselves as united in their hatred and fear of the Merina. Thus, when slaves considered the range of possible outcomes of escape, staying with a master may have seemed the best option.

Behoraka's story regarding how slaves responded to emancipation by coming to his grandfather suggests that, even in freedom, slaves found some value in staying near those upon whom they had previously depended and to whom they had been subservient. Observations by foreigners at the time of emancipation support this assertion that some freed slaves sought to maintain relationships with their previous masters.³⁵

André, a Frenchman, writes that some slaves viewed freedom with apprehension, since gaining free status carried its own burdens with no accompanying provision of social support (André 1899: 80). André suggests that many slaves preferred to hang onto a less-than-free relationship that might protect them from the full force of taxes, forced labor, and military service.³⁶ Makoa, in particular, exhibited a consistent pattern of obtaining land and forming their own settlements where they had been slaves (Rantoandro 1997).³⁷ And if they settled in Betsimisaraka villages, former slaves and their descendents did not remain separate from their “host” populations, but were allowed to inter-marry with the various clans among whom they lived (André 1899).

The fact that some slaves stayed near and depended upon assistance from former slave owners does not, however, address the question of why those former slaves were not stigmatized. Why did their continuing state of dependency not lead to their designation as lower-status and exploited residents? Indeed, in other regions of Madagascar this appears to be precisely what happened. I would argue that one key reason this did not occur in the northeast was the relative abundance of resources in this region. In the following section, I develop this argument through a comparison of two regions that followed distinct trajectories in slave-free relations.