The secretary of a small village in Murrupula district in northern Mozambique received me and my research assistant with a concerned expression on his face when we visited the village for a second time. Following our first visit, the police had arrested four people from the area and incarcerated them for six days. During our first stay, we had conducted extensive interviews with former members of the community-initiated Naparama militia, which was disbanded at the end of the war. At the time of my fieldwork, some militia units were trying to lobby for recognition of their war effort to receive demobilization benefits. The village secretary had made a connection between the imprisonments of the four residents and our first visit, since the four were arrested while helping to register former Naparama members (and other militiamen as well). The registration served the purpose of counting all former militia members in the area to pressure the government to formally demobilize (and pay) them. The Naparama leader of Nampula province who had introduced us to the Naparama militia in Murrupula district organized the registration. After the provincial Naparama leader had collected names and fees from about 250 militiamen and left, the police charged the local Naparama leadership who had helped with the registration with betrayal and arrested them. According to the police, it was unlawful to register militiamen and collect money from them. The arrested men were released after the provincial Naparama leader paid a significant fine to the municipality. Afterwards, people asked the local Naparama leaders to find out what happened to the money they used to pay their registration fees.

This story from my fieldwork in rural Mozambique in 2011–12 demonstrates the ways in which fieldwork in the aftermath of war can have unintended consequences and can create ethical and methodological dilemmas for the research process. The researcher’s activities may provide a backdrop for social mobilization and opportunities for personal enrichment for interlocutors, who decide to take advantage of people’s hopes of future benefits. Nampula’s Naparama leader had not visited the local Naparama community in Murrupula since the general elections in 1994. Only when I asked him to introduce me to that community and we went there together did he reestablish contact with the former militia unit. In a way, I had encouraged the reestablishment of that contact, which the provincial Naparama leader abused for his personal benefit. That benefit had both monetary and political ramifications. During our conversations, he had tried to establish himself as the primary Naparama leader during the war, a fact that many other sources contest. It is likely that through this registration process, he was trying to mobilize Naparama to bolster his claim of being the one and only Naparama leader. As with other Naparama members (and also former members of the armed forces), he was disappointed that the government had neither recognized him as a war veteran nor provided him with demobilization payments. In fact, a considerable number of military members were not recognized as demobilized soldiers as part of the peace agreement signed in Rome in 1992 and were thus ineligible for demobilization benefits. These included members of the armed forces who were demobilized before the end of the war, members of Frelimo’s auxiliary forces such as the Naparama, and the “popular militias.”

These unintended consequences are linked to how legacies of war – social, economic, and political polarization and historical marginalization – influence how communities make sense of researchers’ activities in their midst. As Sluka reminds us, research participants “are naturally going to try to figure out what you are doing here,” and previous experiences with strangers in the community provide categories such as “spy, journalist, policeman, tax collector, and missionary” that may be mistakenly applied to the researcher (Sluka Reference Sluka, Nordstrom and Robben1995, 283). Experiences from the war in Mozambique continue to impact daily lives, and contemporary concerns about the distribution of social, economic, and political benefits all contribute to the perception of the researcher as a powerful and ambiguous figure who can influence people’s lives in positive as well as negative ways. Although some community residents may feel disempowered by the researcher’s presence, others may attempt to manipulate the researcher’s work for the purpose of their own economic and political empowerment.

This chapter reflects my attempt to navigate the polarized political landscape in Mozambique’s society. Though I encountered many challenges along the way, I collected more than 10,000 pages of documents in government archives and conducted more than 250 interviews and oral histories with community members, former militia members, former rebel combatants, former soldiers, (former) government officials, politicians, and academics in five districts and the capital. I worked together with two Mozambican research assistants who spoke all the necessary local languages and had experience in data collection for international projects. They helped me with arranging interviews, translating from local languages into Portuguese, and explaining cultural particularities. As Mozambicans from the regions we worked in, but long-term residents of the city of Nampula and Quelimane, respectively, my assistants were sufficiently knowledgeable about the provinces we worked in (and their languages), but outsider enough to not be identified with a certain political position. In fact, as I outline below, in Nampula, my research assistant was perceived as a stranger just like me, which made our access to some respondents problematic.

Conflict researchers have recognized the ethical and practical challenges that research on violence entails (Nordstrom and Robben Reference Nordstrom1995; Wood Reference Wood2006, Reference Wood, Joseph, Mahler and Auyero2007; Sriram et al. Reference Sriram, King, Mertus, Martin-Ortega and Herman2009; Fujii Reference Fujii2012; Mazurana, Jacobsen, and Gale Reference Mazurana, Jacobsen and Gale2013; Parkinson and Wood Reference Parkinson and Wood2015; Cronin-Furman and Lake Reference Cronin-Furman and Lake2018). However, as Malejacq and Mukhopadhyay (Reference Malejacq2016) note, there is still little transparency and debate on how researchers form and manage relationships in the field and what kinds of ethical compromises and methodological adaptations they have to accept in order to collect the necessary data for their projects. Researchers in political science have learned from their colleagues in anthropology (and geography) for whom the position and impact of the researcher on the local community has become a central concern for how to “do” anthropology (Clifford and Marcus Reference Clifford and Marcus1986; England Reference England1994; Sirnate Reference Sirnate2014). However, what is often obscured rather than openly addressed are the ways in which the researcher becomes a political actor capable of reinforcing existing power structures and, by disempowering or empowering local actors, influencing social realities in communities under study. This is significant, as the autonomy of not only the researched but also the researcher may be jeopardized, and there is a risk that local actors may manipulate researcher’s presence and work. This is true not only for research in today’s volatile conflict zones (Malejacq and Mukhopadhyay Reference Malejacq2016), but also for research in (postwar) polarized societies in which political conflicts linger on and reinforce economic, social, and political inequalities (Gerharz Reference Gerharz2009, 2).Footnote ¹

The limited understanding of the workings of power, and by extension the limits of researcher neutrality, is often due to the fact that field researchers are typically more concerned about the data that they extract from the field site and how to mitigate systematic bias than about what happens to the field site as such. This means that challenges of access to research participants or the “subtext” or “meta-data” (Fujii Reference Fujii2010) from conversations, such as lies, silences, and evasions, are considered “obstacles” rather than “a source of knowledge for ethnographers” (Wedeen Reference Wedeen2010, 256). In fact, researchers may alter the field site and the data in ways that are difficult to account for and “reverse” during data analysis. As Goodhand argues, such intervention in conflict settings is not only a methodological challenge but also an ethical issue, as it “may affect the incentive systems and structures driving violent conflict or impact upon the coping strategies and safety of communities” (Goodhand Reference Goodhand2000, 12).

It was puzzling to me that the social, economic, and political legacies of the war affected my interactions with rural communities in Mozambique, as the country is often hailed as a successful example of postwar peace building and reconciliation (Boutros-Ghali Reference Boutros-Ghali1995). One could expect that (unofficial) reconciliation processes, national reconstruction, and the passing of time would have helped create confidence in people’s futures (Honwana Reference Honwana2002; Igreja, Dias‐Lambranca, and Richters Reference Igreja, Dias‐Lambranca and Richters2008). However, the country remains polarized even decades after the end of the war (Weinstein Reference Weinstein2002; Darch Reference Darch2015). Fear of renewed violence still influences political and social life in rural Mozambican communities – for good reason, as the resurgence of violence in the center of the country in 2013–14 demonstrates (Darch Reference Darch2015; Pearce Reference Pearce2020). Moreover, the spoils of recent discoveries of natural resources have not (yet) reached the ordinary citizen, leading to increases in already high levels of inequality (International Monetary Fund 2016).

In a society seeking to overcome its violent past and advance economic development, the ways in which communities tried to make sense of my as well as my research assistant’s presence had two major consequences for the (perceived) autonomy of research participants and of my own work. The first was related to a narrative of suspicion and mistrust about me and my work that stemmed from people feeling severely disempowered with respect to their control over their own well-being. My presence was threatening to some community residents because I reminded them of other white foreigners who had mingled in their community’s affairs throughout history. The second narrative was related to whether and how participants could manipulate my presence and my work in a ways that benefited them economically or politically. Some research participants saw my presence as an opportunity to escape from the uncertainties of their own lives regarding jobs, livelihoods, and political projects. In the remainder of this chapter, I analyze these two responses and what they meant for the perceived autonomy of research participants and my own work. I provide specific examples from my fieldwork to highlight the implications of residents’ ambiguous responses toward neutrality and power during fieldwork in polarized societies.

3.1 Disempowerment and Research Participant Autonomy

One evening in Mecubúri district in Nampula province, a local government representative told me and my research assistant that people had been talking and wondering what we were up to. In the days before, we had walked through some of the neighborhoods of the district town and conducted interviews with residents and local leaders. The government officer reported that some people were afraid we were spreading diseases, as a number of residents had recently suffered from diarrhea. Others thought that our presence meant another war was on the horizon. As the officer elaborated, it became clear that these fears had been triggered by several events that had occurred in the area, in the province, and abroad. A few days before our arrival in Mecubúri, in October 2011, Libyan head of state Muammar Gaddafi had been killed by rebels, and the youth leader of the African National Congress in South Africa, Julius Malema, had made divisive speeches (for which he later got expelled from the party).Footnote ² Mozambicans follow the news of both countries closely, and in their eyes, the instability in Libya and South Africa was concerning.

In addition, in the officer’s view, some events closer to home had exacerbated people’s wariness about our presence. A theatrical performance attempting to explain that China would build and sell 5,000 houses to the community was understood to mean that 5,000 Chinese would “invade” and settle throughout the province. People were also concerned about news that, a month before, in September 2011, one British and four Americans with heavy weapons in their luggage were briefly held at Nampula airport.Footnote ³ The men claimed that they had come to rescue a boat from Somali pirates. Over the course of our conversation that evening in Mecubúri, we learned that we were not the only strangers who were treated with suspicion. NGO workers of a US-funded project seeking to improve access to safe water regularly distribute “certeza,” a chlorine-based water-purifying liquid to prevent cholera outbreaks. However, whenever cholera breaks out, Mozambicans believe these workers brought the disease (Serra Reference Serra2003).Footnote ⁴

Our presence in the district seemed to fit into this ill-boding sequence of events whose origins and consequences remained uncertain.Footnote ⁵ As Gerharz (Reference Gerharz2009) confirms, people’s suspicions about researchers’ motives are often triggered by their memories of past violence. In highly polarized settings such as the civil war in Sri Lanka, residents of Colombo quickly accused Gerharz of being a Liberation Tigers of Tamil Eelam (LTTE) sympathizer when she discussed the humanitarian situation in LTTE strongholds (Gerharz Reference Gerharz2009, 5–6). Since Gerharz’ comments reminded people in the South of their suffering from seemingly unpredictable episodes of violence, it was difficult for her to highlight the suffering of the other side and at the same time claim neutrality. Similarly, in Mozambique, people did not want a return to war and wondered about the true meaning of my work since it focused on the history of the war.

Such suspicion created a situation that prevented me from establishing trust with and gaining access to some members of the community, crucial preconditions for any successful fieldwork. Drawing on research experience in Northern Ireland, Knox shows that gaining access in politically contested environments is often problematic because communities suspect that the real research objective is “unlikely to be viewed by local actors as neutral or altruistic” (Knox Reference Knox2001, 211). In the highly contested political environment of Northern Ireland, “There was immediate suspicion about the ulterior motives of this research, which had the potential to block access at worst or severely curtail data gathering” (Knox Reference Knox2001, 211). This is also true for the context of my research. The officer we met that evening in Mecubúri was nowhere to be found when we tried to meet with him for an interview the following day.

All these concerns were troubling, as, without realizing it, I had become part of a social and political context in which people feared that, as a consequence of interacting with me, they would further lose control over their health and well-being. The more I (or people like me) entered their lives, the less they felt in charge. At the same time as people overestimated my power, they underestimated their own. People’s responses to my presence in their communities had a similar meaning as their resistance against the distribution of chlorine, which Serra (Reference Serra2003) interprets as an expression of severe disempowerment. As Serra’s analysis reveals, resistance against outsiders in the form of suspicion and mistrust is an expression of people’s distrust in state institutions, which they perceive as being absent and failing to deliver promised services.

The sources of such feelings of disempowerment and loss of autonomy in the central and northern provinces of Mozambique are varied. First, the history of the central and northern provinces is one of political marginalization by the government in Maputo in the south of the country (Chichava Reference De Bruin2007; Do Rosário Reference Do Rosário2009). Frelimo, the liberation movement and party in power since independence in 1975, has been perceived as a southern movement; the independence movement’s penetration of both provinces during the liberation struggle was slow and ineffectual or, in the case of Nampula province, completely absent (Legrand Reference Legrand1993, 88); and the peasant population opposed Frelimo’s policies after independence. In Nampula, it sparked popular discontent when the Frelimo government constructed communal villages and abolished traditional authorities (Geffray Reference Geffray1990). In Zambézia, it was the disrespect for traditional values more generally that in turn provoked opposition (Ranger Reference Ranger1985, 189; O’Laughlin Reference O’Laughlin1992, 115). As a result, the region was and remains a Renamo stronghold.

Second, the particular character of the civil war, a typical guerrilla war, contributed to suspicion toward strangers in rural communities. Community residents responded to my presence in ways that reflect Sluka’s (Reference Sluka, Nordstrom and Robben1995, 283) observation that people misapply preexisting categories (e.g., being a spy) to strangers who enter their communities, which was a common concern during the war in Mozambique and in many other wars (Vlassenroot Reference Vlassenroot2006). Many people referred to the war as a “war between brothers.” In contrast to the anticolonial struggle, members of either side could not be identified easily, as they all belonged to the same community. The enemy could therefore always be lurking somewhere in the community. Moreover, the Renamo rebels were actively supported by Rhodesia and Apartheid South Africa. White South African advisors were regularly flown into Renamo bases. Community residents linked that experience to my presence and wondered whether I had anything to do with the war, since I was so eager to speak to them about that time period. At the end of an interview with an older male community resident, I was asked whether the war would return once I left the village. When I worked in an area in Murrupula district, Nampula, where one of the main Renamo bases was located during the war, the chief of staff of the local administration told us that there had never been a delegation with a white person staying overnight. He urged the community police chief to inform residents so that they would not think something was wrong, as his village had been “an area of the enemy.”

Moreover, although Mozambique has received much development aid and has also recently discovered more natural resources, people feel they have yet to benefit from economic development. Serra’s (Reference Serra2003) analysis points to the arrogance and distance of NGO workers that creates discontent among community residents. In different regions of Mozambique, residents have been displaced by foreign companies, such as coal mining in Tete province or the Brazilian large-scale agribusiness project ProSAVANA, which adds to the impression that strangers meddle with people’s affairs, to the detriment of their livelihoods (Abelvik-Lawson Reference Abelvik-Lawson2014; Lillywhite, Kemp, and Sturman Reference Lillywhite, Kemp and Sturman2015; Chichava and Durán Reference Chichava2016).Footnote ⁶

Finally, much of the hesitance in talking to us was connected to current party politics and reflects the fact that the Frelimo party never lost its dominance in Mozambican politics despite having introduced multiparty politics in its new constitution in 1990 (Sumich and Honwana Reference Sumich2007). Some former government officials declined to be interviewed since they did not feel qualified, which suggests that they did not feel authorized to speak and were afraid of violating the official party line. In other cases, these officials made sure that I had respected the administrative hierarchy and attained permission from their (former) supervisors.

This past and contemporary experience of marginalization contributed to the perception of my research assistant and I as “intruders.” I dealt with this situation in several ways to establish “research legitimacy” (Knox Reference Knox2001). I always respected the social and administrative hierarchy when coming into a district I had never been to, and I introduced myself and my work to local leaders to receive “‘approval’ from key stakeholders” (Knox Reference Knox2001, 212). In the districts I visited after Mecubúri, I asked for an elder who was respected in the community as a guide who could introduce me to people, or asked for referrals from research participants, a strategy commonly referred to as “snowball sampling” (Sluka Reference Sluka, Nordstrom and Robben1995, 284; Knox Reference Knox2001, 212; Cohen and Arieli Reference Cohen and Arieli2011; Romano Reference Romano2006). Mistrust between Frelimo and Renamo elites implied that I was to pursue relationships with several types of “gatekeepers” (Campbell et al. Reference Campbell, Gray, Meletis, Abbott and Silver2006): with Frelimo party and state structures and, separately, with Renamo party structures. I also respected people’s wishes to not being interviewed alone; when I was interviewing men, their wives often sat next to them to listen in on the conversation. I also tried to visit communities several times to establish rapport (Browne and McBride Reference Browne and McBride2015; Norman Reference Norman, Sriram, Kind, Mertus, Martin-Ortega and Herman2009).

Overall, I avoided talking about politically sensitive topics (Sluka Reference Sluka, Nordstrom and Robben1995, 283) and avoided mentioning “politics.” In the process of trying to make sense of my presence in their communities, residents wanted to make sure that I do not have anything to do with “politics,” which has negative connotations in Mozambique, as in many parts of Africa, because people believe politicians lie and enrich themselves (Ekeh Reference Ekeh1975). A businessman and veteran of the pre- and postindependence wars in Nicoadala in Zambézia province invited me over to his house for lunch to finally “forget about politics” and “just chat.” He could not understand why I was willing to “suffer” and study political history rather than do business, as Mozambique was “the place to do business.” Religious community residents were concerned about my political intentions. In Murrupula, the first question of a sheikh was which party I was affiliated with.Footnote ⁷ In Nicoadala, a pastor only agreed to meet with me once I assured him I would not talk politics under the roof of his church.Footnote ⁸ When I was confronted with these concerns, I emphasized my status as a student who is independent of party politics (Knox Reference Knox2001, 212).

But as many field researchers have recognized before me, neutrality is difficult to achieve, and sometimes not even desirable (Nash Reference Nash, Rynkiewich and Spradley1976; Sluka Reference Sluka1990, Reference Sluka, Nordstrom and Robben1995; Gerharz Reference Gerharz2009; Malejacq and Mukhopadhyay Reference Malejacq2016). The strategies I adopted mitigated many of these concerns, but posed some new methodological and ethical dilemmas. For example, it was important to take into account the ways in which people introduced me to certain communities and to consider whether the presence of certain people during interviews influenced and changed the conversations. As Campbell et al. (Reference Campbell, Gray, Meletis, Abbott and Silver2006, 115–16) argue, rather than trying to be “neutral” in general, it is important to emphasize your independence from gatekeepers. It also meant that some community residents might have felt compelled to talk to me because an authority figure told them to, and not because they themselves had volunteered. It was also important to consider “gatekeeper bias” (Cohen and Arieli Reference Cohen and Arieli2011), and in particular the issue of sampling bias (Groger, Mayberry, and Straker Reference Groger, Mayberry and Straker1999). These dilemmas required me to provide detailed explanations and to be transparent about my activities to ensure that people were comfortable talking to me, but I also had to consider the emerging methodological limits during analysis.

However, fieldwork challenges did not always arise out of people’s concern about their own disempowerment and the limits to their autonomy, but also out of their hopes for political, social, or economic empowerment, as I discuss in the next section.

3.2 Empowerment and Researcher Autonomy

While the reports of mistrust and suspicion in northern Mozambique were troubling, the way in which they were communicated to my research assistant and me appeared to be for political currency. The local government representative who warned us about the concerns within the community in Mecubúri apparently used these stories to pursue his own agenda and fight a political battle against the district administrator. My research assistant found out that, for unclear reasons, the administrator was not well liked among local government employees. The officer we talked to was wary of the fact that the district administrator had given us permission to work without a guide accompanying us to interviews with community members. It seemed likely that he felt his position within the local administration was not taken seriously. As someone who was in constant contact with the local police and other local leaders, he used his monopoly on information to manipulate us for his political interests and divert attention from the fact that he himself mistrusted us. As mentioned earlier, throughout our time in that district, the officer avoided being interviewed, although he had agreed to do so earlier.

This politicization of mistrust and suspicion has a long history in Mozambique. During our conversation in Mecubúri, I learned that members of the “opposition” sprinkle “chlorine” (actually they use flour), which supposedly spreads cholera, on some people’s doorsteps, implying that if the residents touch it, they will be contaminated. Thus, while the initial narrative about how cholera spreads expresses disempowerment and distrust of state institutions, this counter-narrative puts blame on the “opposition,” a diffuse group of people who oppose the Frelimo government and may be sympathetic to Renamo. Overall, such suspicions reinforce political cleavages, which are understood in many parts of Mozambique as existential threats rather than aspects of democratic politics.

Politicization can occur on several levels. Another example of how current political developments affected my research was how Renamo leaders in the provincial capital of Quelimane reacted to my request for permission to interview former Renamo combatants in the province. Since at the time of my request, national Renamo leader Afonso Dhlakama had threatened to stage a (peaceful) overthrow of the government on December 25, 2011, provincial leaders of the party did not consider this a suitable time to allow such interviews.Footnote ⁹ When I tried again in February 2012, the provincial party leaders in coordination with Dhlakama himself granted me permission, as the political situation had since calmed.

But my work was not only politicized with respect to its potential negative consequences. Others played with my work’s potential positive consequences, as I outlined in the beginning of the chapter. In a way, the Naparama leader who had organized the registration of Naparama after our departure from Murrupula was manipulating people’s hopes of future benefits, which I had (unwillingly) raised in the first place. Former Naparama members were not only surprised but also humbled by the fact that someone wanted to talk specifically to them so long after the war had ended. I was able to conduct many interviews in Mecubúri and Murrupula with Naparama who walked many miles to meet with me. Naparama members hoped that my questions would precipitate their being registered to eventually receive demobilization benefits or at least funds for “projects.”

This demand for recognition and “projects” had meaning for both the individual and Naparama as a group. “You can’t talk to Naparama individually,” a former commander of the militia told my research assistant and me one morning in Nicoadala. We had just introduced ourselves and our project during a meeting with the group’s leadership. The commander informed us that the (former) Naparama high command could give us all the information we needed, and that the remaining former combatants would speak to us as a group. He claimed that individual Naparama were not mentally capable of talking properly about the Naparama, which would result in contradicting stories. They sought to restrict my access and allow me to interview only former members that they could “control” or combatants of high rank, while discrediting other members as not telling “the truth.”

The Naparama commander clearly sought to control what version of the history of the community-initiated militia would be told. He did not want my research project to jeopardize Naparama’s ongoing struggle to receive recognition from the government and compensation for the group’s wartime efforts. This concern was not completely unfounded. As the commander later explained, he had been taken once to the Mozambican intelligence agency’s office and charged for not providing a certain document that the agency had received from other sources. The commander was afraid that I would pass along information that the combatants told me to the intelligence agency, which would in turn interrogate the Naparama leadership for not having disclosed this information previously.

My research assistant and I emphasized that I was an independent student writing a thesis and that I was unaffiliated with parties or the government. But by emphasizing my student status the leaders concluded that the Naparama militia would not benefit from the study, and they therefore suddenly ceased to cooperate with us. In a last effort to solve what at that point seemed to be an insurmountable hurdle, I explained why I found my study important: most histories of the war had focused on Frelimo and Renamo while ignoring the important contribution of the Naparama. Since the militia leaders had been demanding recognition from the government for a long time, they appreciated that I highlighted the value of their contribution and thus agreed that all the leaders could be interviewed individually.

These examples of individuals and groups attempting to manipulate my research project represent instances of the aforementioned gatekeeper bias, but in more intentional and manipulative forms, which is common in fieldwork with marginalized or high-risk communities who have certain grievances that they want addressed. Access is traded for a certain version of representation that benefits research participants and the political groups to which they belong. Gerharz (Reference Gerharz2009), for example, mentions how the rebel group LTTE in Sri Lanka attempted to make use of the many researchers in order to polish its own image. In the eastern provinces of the Democratic Republic of Congo (DRC), Vlassenroot (Reference Vlassenroot2006, 197), working among armed groups, experienced how his “writings were used as proofs that [respondents’] claims or grievances were justifiable.” Researchers thus can be used to improve a group’s or an individual’s reputation. Malejacq and Mukhopadhyay (Reference Malejacq2016, 1014) report how a handshake of one of the authors with an Afghan governor was broadcast on TV to counter “the governor’s reputation as an uneducated countryman by exhibiting his connection to a foreign university professor.”

While I was careful in all conversations to avoid making the impression that talking to me would result in political or monetary benefits, this hope was difficult to dispel. Part of the problem was that local leaders who helped to connect us with former combatants were insensitive to the ways in which they might create false expectations. In a rural area in Murrupula district, the secretary of the locality had called all demobilized soldiers for a meeting. When we started our interviews with some of the demobilized soldiers and explained what we were doing, they were disappointed because they had hoped to finally receive the benefits that they had been anticipating for such a long time. At times, it seemed as if local leaders had deliberately misrepresented the purpose of such meetings, because they realized that people would not show up if they knew the purpose of my research.

This misrepresentation created an ethical dilemma, as I depended on others to introduce me to community residents who had been involved in the war, but I did not have complete control over how people represented the purpose of my work. When researchers depend on core contacts and gatekeepers who can manipulate their activities and writings, this dependence inverts the power relationship between researcher and researched and constrains the autonomy of the researcher and her project (Vlassenroot Reference Vlassenroot2006). Such power asymmetries in favor of research participants are especially pronounced in dangerous settings in which researchers depend on certain elites for their personal protection (Adams Reference Adams1999; Kovats-Bernat Reference Kovats-Bernat2002; Malejacq and Mukhopadhyay Reference Malejacq2016, 1013). But they find similar expression in polarized societies in which researchers depend on certain individuals to gain access and trust.

3.3 Conclusion

These narratives from my fieldwork in central and northern Mozambique demonstrate how the autonomy of the researcher and the researched are closely interlinked. The people I asked to interview thought I had particular powers, both positive and negative. Some believed my mere presence and/or the subject of my work threatened their livelihoods and well-being. For others, my presence provided an opportunity for them to receive support for their visions of politics so that they could reach their political, social, and economic goals.

In such contexts, the researcher becomes a political actor within the field site and fieldwork becomes “a form of intervention” (Malejacq and Mukhopadhyay Reference Malejacq2016), which curtails community residents’ autonomy over their lives and well-being. As a consequence of the researcher’s presence, the field site experiences a qualitative change, which is difficult to “factor out” of the resulting data during analysis. At the same time, community residents become “actors” in the research project, which may constrain the autonomy of the researcher (Vlassenroot Reference Vlassenroot2006) and contribute to her “relative powerlessness,” restricting her role to that of a “mascot researcher” (Adams Reference Adams1999). For example, the Naparama commander of Nicoadala that I mentioned previously attempted to influence the research design by limiting access to certain individuals, thereby becoming an author of the study rather than its subject. A researcher therefore needs to consider how any negotiation of her position within the field site as well as potential biases due to gender and other characteristics of the researcher affect data analysis. Instances of empowerment and disempowerment (and their consequences) can only be recognized when discursive strategies such as rumors about the researcher, inventions, denials, evasions, and silences are treated as “meta-data” of fieldwork (Fujii Reference Fujii2010).

Rather than conceiving of ourselves as external observers, analysts, and critics of disempowerment, researchers need to consider the ways in which we may, inadvertently, contribute to empowering some and disempowering others. Even if (or especially when) researchers try to be neutral and retain distance from community life, they unwittingly become actors in local or national conflicts (Sluka Reference Sluka1990, Reference Sluka, Nordstrom and Robben1995; Gerharz Reference Gerharz2009). Some researchers have embraced the impossibility of remaining as neutral and impartial observers, in particular in violent settings. Malejacq and Mukhopadhyay (Reference Malejacq2016), for example, discuss the ways in which they have “intervened” in their respective field sites and engaged in “tribal politics” during their work in Afghanistan and Somalia, creating informal networks of informants who provided access and protection. Security concerns make research difficult in violent settings, but researchers in highly polarized postwar contexts can also intervene in local politics even though they are “only” observers.

Overall, the two narratives demonstrate that community residents react to strangers in deeply ambiguous ways. First, I argue they do not trust strangers who intrude in their communities with short-term projects that are contingent, conditional, and subject to review. Conversely, some projects provide jobs and seed money, which could improve people’s lives. This again confirms Serra’s (Reference Serra2003) notion that what people ask for is not a complete absence of the state and its services, but when external agencies intervene, they need to be more accountable and reliable.

By extension, what communities ask for is not that researchers stay away from them, but that they are aware of the political nature of their work. Researchers become part of a community and shape social realities in ways that may be neither anticipated nor intended, creating opportunities for both empowerment and disempowerment. Such reflection remains important, both for research transparency and research ethics (Parkinson and Wood Reference Parkinson and Wood2015).

Chapter 4 introduces the reader to the historical context of the war in Mozambique and explores the historical roots of social, political, and economic polarization that shaped my fieldwork.