Medical pluralism, boundary making, and tuberculosis in Lambaréné, Gabon

Scholars of medical pluralism are interested in how healers position themselves and their healing practices within a therapeutic landscape, and how patients navigate an array of therapeutic traditions. Based on fieldwork in Lambaréné, Gabon, this article examines the discursive practices of tuberculosis patients and healers, finding that therapeutic traditions were kept separate. Examining a national programme that fosters traditional medicine, I show how the Gabonese government engages in practices of boundary making by reinforcing traditional healers’ position within the Gabonese therapeutic landscape. This research confirms popular paradigms of boundary making within the medical pluralism debate, wherein boundaries are produced and crossed to contrast, strengthen, purify, and divide the therapeutic landscape. Additionally, formal state-sponsored discursive practices refer to a merging of traditional medicine and biomedical medicine. This stands in contrast with patients’ and healers’ discursive practices, and their wariness of fluid or adaptive boundary-making processes. To explain this, I introduce the concept of ‘conventional boundary making’ and then analyse it in the context of tuberculosis and in relation to theories of state power, Gabonese therapeutic identity politics, and structural violence.

Jadoungou had brought his wife, Adeline, to a hospital in Lambaréné, Gabon, because she had felt very ill for some months. There she was diagnosed with extrapulmonary tuberculosis (TB). Adeline was hospitalized for several weeks and put on first-line antituberculous drugs, but did not recover. During my visits at the hospital, Adeline was too ill to converse, so I mostly talked to Jadoungou. He continuously expressed his troubles regarding the care of his wife. He feared that her TB was magical, that someone had bewitched her, but he also struggled with the taboo of using different therapeutic traditions simultaneously. Their church leader had approved their seeking help at the hospital, but condemned consulting a traditional healer, as such healing practices were associated with work of the devil. However, in the case of bewitchment, hospital treatment could not cure his wife. During several of my visits, Jadoungou confided that he was planning on visiting a traditional healer, one time explaining: 'If my wife is still ill after some time, we might go to the traditional healer. We Protestants are not supposed to go there. The church doesn't like it if two different religions are combined. But you have to do something to get healed'. Jadoungou made clear that he could only justify switching to another therapeutic tradition if it was certain that the biomedical treatment would provide no cure.
Adeline died after four weeks, leaving Jadoungou devastated and with mixed feelings about his own care-taking choices. On the one hand, he regretted that he had not done everything to save Adeline's life. But on the other hand he described feeling relieved at not having sought out another therapeutic tradition at the same time. Jadoungou's careful care for his wife and his decisions raise the question of how medical pluralism is shaped and organized in Lambaréné, Gabon. In this article, I focus on TB patients and healers in Lambaréné, both of whom resort to discursive practices to explain how they navigate a plural medical field while acting on TB. With 'discursive practices', I refer to discourse that not merely reflects the existence of a reality but also defines the social rules that construct this reality (Foucault 1972, 48). Moving away from literal actions or behaviour, this study analyses the idea of boundary-making processes that are principally constructed within speech.
TB is one of the deadliest infectious diseases worldwide (Bourzac 2014) and one of Gabon's major health concerns (WHO 2015a;Mvé 2010), despite the existence of antibiotic treatments that can completely cure (most) patients. A study conducted in Gabon's capital city Libreville found that only one-third of TB patients recovered, and that almost half of the patients abandoned their treatment (Mvé 2010). About two decades ago, Paul Farmer (1996a) emphasized the need for a critical anthropology of infectious diseases, such as TB, to address patients' realities within and across particular contexts. Various ethnographies have assessed TB patients' health care-seeking behaviour and TB programmes in different parts of the world (Greene 2004;Harper 2006;Koch 2013;Gerrish, Naisby, and Ismail 2013). Some scholars argue that too much attention has been paid to the role of culture as obstacle to treatment, and that socioeconomic barriers have been downplayed (Peltzer, Mngqundaniso, and Petros 2006). TB is often described as a 'disease of the poor' (Spence et al. 1993;Schwartzman and Menzies 1999) and linked to theories of structural violence (Farmer 1996b;Farmer 2004). The disease mainly affects the most marginalized and vulnerable people with a low socioeconomic position, due to social structures of inequality and poverty. This article aims at understanding the cultural, socioeconomic, and political contexts of TB patients, as they navigate the therapeutic landscape of Gabon.
Within the field of medical anthropology, encounters with different therapeutic traditions have often been discussed in the light of boundary-making processes. The idea that there are fixed, static boundaries between different therapeutic traditions has been critiqued, and scholars have shifted to theorizing how various illness perceptions and therapeutic traditions intermingle and interact in people's practices while they search for a cure (Last 1981;Pool 1994). The reproduction and the crossing of boundaries are mutually constitutive acts, as no boundaries are crossed if there are none produced and the other way around (Luedke and West 2006, 5, 6). These boundary-making practices between therapeutic traditions are often presented as being fluid and continually reconfigured (Bernstein 2001;Luedke and West 2006;Krause 2012, 8).
To explain the functioning of medical pluralism in societies and the associated practices of boundary making, various authors have referred to the role of the state, both in Eastern countries (Hyma and Ramesh 1994;Zhan 2001) and Western countries (Cant and Sharma 1999;Broom and Tovey 2007). States may impose a discourse that foregrounds biomedical explanations and practices (Finkler 2004), but may additionally focus on the professionalization and strengthening of the position of traditional medicine within society. These efforts may prompt processes of negotiation and the redefinition of both traditional and modern medicines (Zhan 2001). Luedke and West (2006) describe how boundaries are created to define and delineate traditional healing, but at the same time can be crossed, challenged, and shifted.
Similarly, in Gabon, boundary-making practices within the Gabonese therapeutic landscape have changed and shifted in response to colonialism and nationalistic political programmes. TB patients must negotiate treatment decisions in a context where the national political programme fosters Bwiti, a religion and associated set of traditional therapeutic practices, which the Gabonese government claims represents Gabonese tradition. Scholars describe Bwiti as a political project to reunify the various ethnic groups of Gabon, by defining what is authentically Gabonese (Swiderski 1988, 128-29;Samorini 1993). The state thus may produce boundary-making processes by acknowledging the position of traditional healers in the therapeutic landscape, and it may break down boundaries by encouraging collaborations among healers. Such a therapeutic discourse of boundary making and breaking can achieve local and national relevance, informing the medically plural landscape and the position of its actors.
Within this context of medical pluralism and boundary making, I would like to add a new concept: 'conventional boundary making'. This concept is derived from my respondents' discursive practices, which referenced a separateness between therapeutic traditions. Both healers and patients explained that boundaries were crossed at times, a practice that is extensively discussed in the literature, but, they maintained, boundaries were not shifted or reconfigured. Jadoungou, for example, expressed a preference to engage in therapeutic traditions serially, not simultaneously. Similarly, Orr (2012, 518) briefly notes that, in contrast to contemporary theories, he hardly encountered any blurring of boundaries between therapeutic traditions in southern Peru.
To elucidate this notion of conventional boundary making, I draw on five months of fieldwork in Lambaréné, Gabon. This research was part of a larger study on TB in Lambaréné, and is why the focus of this fieldwork was narrowed down to TB patients and their healers. In this article, I discuss some of literature concerning medical pluralism, with a focus on conceptualizations of therapeutic traditions. This involves a dive into history, as medical pluralism has already been the subject of study for more than four decades. I then introduce my fieldwork and the main therapeutic traditions encountered in Lambaréné. I draw on the notion of conventional boundary making to present TB patients' and healers' discursive practices in response to their experiences of navigating the therapeutic landscape. I do not claim that the notion of boundary making is new, nor that medical pluralism is absent. Rather, I describe a variant of boundary-making processes that manifests discursively, one in which a reluctance towards boundary reconfiguration is embedded. In other words, boundaries are said to be crossed but not challenged and shifted. Finally, I assess why such discursive boundary-making practices occur, and describe whether and how medical pluralism intersects with concepts such as state power (Werbner 2001), national therapeutic identity politics in Gabon (Swiderski 1988;Samorini 1993;Ngolet 2000), and structural violence (Farmer 1996b;Farmer 2004).

Medical pluralism and processes of boundary making
In the 1970s, Leslie (1975) represented the variety of biomedical, traditional, and alternative health care in India and China as 'medical systems', thereby acknowledging their welldeveloped and sophisticated nature. This idea sharply contrasted with previous studies that considered traditional medicine merely a part of traditional religions. The representation of equal medical systems strengthened traditional healers' positions and recognized their ability to compete with biomedical systems (Johannessen 2006;Krause 2012). Kleinman (1980) further highlighted the ubiquity of biomedical interpretations by differentiating between the biomedical definition of diseases and patients' experience of illness. With the idea of individual illness experiences, scholarly attention was increasingly given to the influence of specific sociocultural, political, and economic contexts, leading to a wide range of models to explain illnesses and health care (Krause 2012).
The notion of 'multiple explanatory models' (Kleinman 1980) provoked discussion about patients' responses to environments in which a variety of therapeutic traditions were present. In the 1980s, some scholars posited that parallel medical systems were fixed, with sharp boundaries, and used by patients either sequentially or simultaneously (Last 1981;Janzen 1979;Kleinman 1980). Such representations of medical systems as closed -with a homogeneous group of practitioners and patients, and structured treatment regimens -were criticized, as they did not match ethnographic complexities found during fieldwork (Hörbst, Gerrets, and Schirripa 2017). Last (1981) described a spectrum of unsystemized and scattered medical practices from which pragmatic patients chose, when it was appropriate to do so; other anthropologists also described patients' pragmatism whilst navigating political and economic constraints (Janzen 1979;Whyte 1997). Several anthropologists noted that making decisions via 'trial and error' was common: failed treatment in one therapeutic tradition would trigger the patient to seek health care from another (Last 1981;Muela et al. 1998).
In earlier studies of medical pluralism, anthropologists often awarded biomedicine a dominant position, though this hierarchy became widely rejected (Last 1981;Pool 1994). Initial worries that traditional medicine would be supplanted by biomedicine proved unfounded, as cultural maintenance and interplay remained a prominent feature within and between healing practices (Arnold 1993;Hörbst, Gerrets, and Schirripa 2017). African therapeutic traditions did not seem to be destroyed by colonialism, but appeared adaptive and resilient (Vaughan 1991). Conversely, Europeans likewise adopted and internalized elements of African therapeutic traditions, such as herbal remedies (Comaroff and Comaroff 1997, 364). Following these ideas, later approaches presented biomedical and traditional therapeutic traditions in a more symmetrical fashion, leading to a reconsideration of the latter and its importance for patients. Moving beyond the idea of systems within certain hierarchies, various authors have offered looser frameworks, such as the term 'therapeutic traditions' (Feiermann and Janzen 2002).
Various concepts came into use, such as circulation (Vaughan 1991) and indigenization (Kleinman 1980), to describe how actors crossed, shifted, and reproduced boundaries by adopting each other's therapeutic elements. Anthropologists have repeatedly analysed the entanglement of different therapeutic traditions and explanatory models, describing a range of processes, including bricolage, hybridism, syncretism, assemblage, eclecticism, selection, and adaptation (Krause 2012). In the last decade, yet another perspective has gained popularity, which emphasizes the fluid and continually changing nature of boundaries between therapeutic traditions (Bernstein 2001;Luedke and West 2006;Krause 2012, 8). The reconfiguration of people, objects, and entities, and the idea of dealing with change features centrally in work of Hsu (2017), who studies medical pluralism in Chinese clinics in East Africa. Another newly introduced concept is mobility, used to describe how diseases and health care-seeking behaviour may travel across borders whilst being negotiated and transformed (Hörbst, Gerrets, and Schirripa 2017;Raffaetà et al. 2017).
The paradigm of medical pluralism in combination with the idea of boundary making is closely related to debates about power, as boundaries are generally considered arbitrary yet always powerful (Nader 1996). The maintenance and crossing of boundaries within the medical landscape are theorized as ways to empower therapeutic traditions (West and Luedke 2006, 4). For example, biomedical professionals may aim to maintain sharp boundaries with alternative therapeutic traditions in order to prevent competition and to confirm their superiority (Marsland 2007). Similarly, faith healers may create a boundary by stressing how their healing approaches diverge from other approaches (Krause 2008). Many ethnographers discuss how actors (mainly from the traditional or religious realm) transgress imposed boundaries, using this action as a tool of resistance against taken-for-granted hierarchies in the therapeutic realm (Simmons 2006;Luedke 2006;Marsland 2007;Hampshire and Owusu 2013). Marsland (2007) describes how traditional and faith healer categories seem to mix and entangle, leading to healers representing a variety of healing traditions and combining healing techniques. Some traditional healers transgress boundaries by incorporating biomedical practices (Luedke 2006;Hampshire and Owusu 2013) or creating institutions (Marsland 2007). Boundary transgression may reinforce medical expertise, authority, and legitimacy.
Boundary-making processes can similarly be enacted by states, who can thus influence the development of a medically pluralist society and the maintenance or crossing of symbolic boundaries between therapeutic traditions. Most studies of medical pluralism that focus on the role of the state explore the coexistence and intertwinement of therapeutic traditions. The literature describes the integration of traditional medicine into national health systems (Hyma and Ramesh 1994;Cant and Sharma 1999;Broom and Tovey 2007). Zhan (2001) describes how the state aimed to make Chinese medicine more scientific, in order to enable competition with biomedical science. As a consequence of the success of traditional medicine, biomedicine and traditional therapeutic traditions were woven together, intermingling, interchanging, and adapting. This brief overview of medical pluralism, focusing on processes of boundary making and the role of the state therein, scratches only the surface of the available literature. However, it provides a framework for how I approached my research, and for understanding the findings I present below.

Methods and setting
In 2012, I conducted five months of ethnographic fieldwork in and around Lambaréné, Gabon (Cremers et al. 2013), with assistance from Grace Bikene, a local researcher. Aiming to explore how TB patients navigated a medically plural landscape, I approached thirty patients who were undertaking TB treatment either at the Albert Schweitzer Hospital (ASH), the regional government hospital, or at the government's ambulatory health care centre for human immunodeficiency virus (HIV) and TB (GAHC). Recruitment took place with the help of researchers from a TB epidemiology cohort study (PanEpi) at the ASH in Lambaréné. We visited patients multiple times at their homes and interviewed them and their family members in French about their experiences with TB and the associated (medically plural) care they engaged in. Interviews lasted one to two hours, were in-depth and semi-structured, and were complemented by informal conversation. Additionally, we approached faith healers (N=8), biomedical healers (N=10), and traditional healers (N=5) to understand how they positioned themselves within the medically plural landscape of this particular setting. I also approached one traditional healer outside of the research area, in the capital Libreville, because he was well-known to most respondents as a healer who combined biomedical and traditional healing practices. Two focus group discussions were organized, with eight TB patients each, one at the ASH and one at the GAHC, to discuss perceptions of TB and health care-seeking behaviour. I conducted extensive participant observation at patients' homes, in their villages, and at biomedical hospitals; during nocturnal traditional healing practices and during exorcist sessions conducted by faith healers in various churches (Pentecost, Protestant, Catholic, Christiansme Celeste, Christianisme du Réveil).
In order to understand the position of the state regarding the therapeutic landscape and traditional healing, I visited the government's Institute of Pharmacology and Traditional Medicine (IPHAMETRA) in Libreville. I conducted participant observation and several indepth interviews with its director and traditional healers (N=5) to discuss the role of traditional healing in Gabon. As stated earlier, Bwiti is promoted as the Gabonese traditional identity (Swiderski 1988, 128, 129), but there are many other traditional societies and healing practices such as the Ndokwe, Mekum, and Ndjembe. These have received less political attention compared to Bwiti (Ngolet 2000, 66). The IPHAMETRA website, however, referred to the broad term 'traditional healing' and did not particularly specify Bwiti.
For my analysis of these ethnographic data I used a grounded theory approach, which led to the identification of five themes of conventional boundary making: a dichotomy between the medicines of the Blacks and Whites, a taboo against combining therapeutic traditions, the justification for changing therapeutic traditions, the danger of combining two therapeutic traditions, and therapeutic traditions in their essence being the same.

Medical pluralism in Lambaréné
In Lambaréné, a vast array of therapeutic traditions were available. For the sake of convenience, I loosely divide this medically plural landscape into three categories: traditional healers, faith healers, and biomedical healers.1 The three categories are based on a combination of the literature from Gabon, my observations during ethnographic fieldwork, and the framing of healers and patients.
In order to understand the construction of 'the traditional healer' in Gabon and the sociopolitical processes involved, traditional healing needs to be situated within the history of Bwiti. Around 1890, the Fang ethnic group incorporated the masculine initiation cult Bwiti and its similarly named religion Bwiti, as practiced by the ethnic groups Mitsogo and Apindji. With the arrival of colonist and missionary influences in the same era, Bwiti developed into a syncretic religion in which God, ancestors, and nature spirits were equally recognized. Many subcategories of Bwiti arose, having different functional specializations (le Bwiti Disumba, le Misoko, le Ngonde, le Myobe), varying rituals (le Ndea), or schismatic innovations (le Sengedya) (Bonhomme 2003). Samorini (1993) refers to interethnic marriage, travel, and local migration in order to explain how a wide variety of Bwiti interpretations was spread and practiced by a majority of ethnic groups. Ngolet (2000, 67) argues that this variety occurred because of different responses to the economic and social problems encountered in Gabon.
Christians generally perceived the Bwiti religion, its traditional healers, and its followers as diabolic, because of its history of human sacrifices, anthropophagy, sorcery, and communication with the deceased. With support from the French colonial government, Christians aimed to eradicate it by killing Bwiti leaders and destroying their temples. This persecution and weakening of traditional knowledge led Bwiti to become a collective tool of 1 In order to present a more symmetrical conceptualization of these traditions, I refer to each of them as 'healers' in this section. However, later on, these terms are used interchangeably with the terms used by my respondents. Interestingly, respondents often referred to biomedical healers as 'doctors' or 'medical doctors'. However, one could argue that all categories can be considered as medical and all healers as doctors. Therefore, I have chosen to nuance this category with the term 'biomedical'.
resistance against colonial rulers and to enhance social cohesion among Gabonese. In 1948, several spiritual Bwiti leaders started a movement to unify the different (Bwiti) cults and to strive for recognition by the Gabonese government (Swiderski 1988, 126). In 1960, Gabon became independent, and its first president Léon Mba was a Bwiti initiate. After years of persecution, Bwiti came to symbolize a national and anticolonial sentiment (Samorini 1995). Nowadays, Bwiti is a fully recognized religion in Gabon and an important aspect of national identity and traditional values. Its therapeutic traditional values are promoted by IPHAMETRA, though the organization includes only herbal traditional healing, and excludes spiritual traditional healing.
Faith healers are not part of IPHAMETRA's assortment of traditional specialists, despite the fact that they have a prominent role in Gabonese society. Both faith and traditional healers included in this study explained TB as the involvement of evil spirits and were seen as authorities on witchcraft and sorcery, but faith healers responded defensively when this similarity was pointed out. They often argued that no linkages existed between them and traditional healers, because they condemned traditional healers' work as practices of the devil. The proliferation of different churches, such as Christianisme Celeste, paralleled the emergence of new healing practices (Ngolet 2000, 61-65). New sects raised social and political issues and advocated for an Africanization of the church. They criticized church policy for its focus on moralistic and religious issues, which they linked to its state sponsorship (Ngolet 2000, 61-65).
Biomedical healing was introduced in Gabon during the colonial era and with the arrival of missionaries. In 1913, Albert Schweitzer founded the first hospital in Lambaréné (Schweitzer 1933). The formerly colonial hospital -with foreign sponsorship and foreign biomedical healers and nurses -has evolved into a partially state-sponsored biomedical service with almost all the staff being Gabonese, and is one of the three main biomedical facilities in this area. Biomedical healing has become more acceptable and accessible in Lambaréné, but is not universally trusted nor considered appropriate for all illnesses or conditions.

Medical pluralism and tuberculosis in Gabon
Although I have used 'tuberculosis' and 'TB' throughout this article, there are many diagnostic labels for the disease. The traditional healers who participated in this study used different names for TB but recognized its biomedical synonym. They treated TB patients with herbs; some additionally used spiritual healing. Traditional healers were scattered, not comprising a group as such, perhaps because their knowledge and skills can be obtained from a variety of sources (for example, through a dream, via training by a spiritual father). Yet, there were some similarities as all healers were initiated in a traditional cult (three were Bwiti initiates), carried the name 'nganga' (meaning healer, in Bantu languages), and used the hallucinogenic plant iboga as a central element in their (TB) healing practices. Iboga, or what respondents referred to as 'le Bois Sacré' (holy wood), was firstly used by the Pygmees, who are seen as the original inhabitants of Gabon (Samorini 1995;Bonhomme 2003, 9) and is a symbol for Bwiti healing practices (Samorini 1995). Moreover, these traditional healers were connected by a state campaign to professionalize traditional healing, and each obtained an IPHAMETRA-issued certificate to confirm their expertise in traditional healing.
Of the numerous Christian churches in Lambaréné, many engaged in faith healing of patients, using a power derived from the Holy Spirit. TB healing was conducted in many different forms, ranging from prayers to individual or communal exorcism sessions, during which patients were relieved of evil spirits. Some faith healers additionally made use of herbal medication, holy water, or fasting. Some faith healers informed their patients of the option to seek additional treatment at the hospital, but most condemned seeking help from traditional healers.
Biomedical tuberculosis treatment at the hospitals studied consisted of a first-line four-drug therapy regimen that can cure patients in six months. According to the World Health Organization (2019), two-thirds of TB patients who do not take biomedical treatment will die. The Gabonese National Tuberculosis Programme (NTP) provided treatment for free, but hospitals faced repeated drug stock-outs (Cremers et al. 2013;Belard et al. 2015), which sometimes persisted for several months. At the time of this study, there were no diagnostic facilities to diagnose resistance to first-line drugs. Patients not responding to first-line drugs were unable to receive alternative biomedical care, as second-line drugs were not available. Gabon did follow the WHO guidelines and signed the International Covenant on Economic, Social and Cultural Rights of the United Nations in 1983 (WHO 2015b; UN 2001), but was unable to meet international standards regarding essential medications.
Gabon's rich history of medical pluralism, with its wide spectrum of traditional, faith, and biomedical therapies, raises the question about what discursive practices are used by TB patients and healers while navigating this plural therapeutic landscape. In Lambaréné, conventional boundary making was reflected in my respondents' descriptions of how therapeutic traditions were kept separate.

Dichotomy between medicine of 'the Blacks' and 'the Whites'
My research assistant's mother, Joessabe, was a traditional healer in one of the villages outside Lambaréné. On a daily basis, patients visited her house, often accompanied by their family members. Joessabe's position as traditional healer within this village and the importance of traditional healing for the people living in the area was emphasized during our various discussions during my many visits. While conversing about traditional healing practices, village residents placed emphasis on terms such as 'tradition' and 'real Gabonese'. Most framed it as 'this is what we have done for generations' or 'what we have is the medicine of the Blacks'. Joessabe and the other traditional healers explained that their practices did not contain any medical elements, but relied solely upon traditional Gabonese resources, such as herbs, fetishes, and connection with ancestors. These discursive practices made a boundary between 'the medicine of the Blacks' and 'the medicine of the Whites'.2 The hospital, described as a facility of the Whites, was considered 'un-African' referring to the fact that doctors did not know their patients and family members could only visit during strict visiting hours. Moreover, they explained that the traditional way of healing was considered 'normal' for this country and its people, 'the Blacks'. This dichotomy between the hospital and traditional healers, between 'the Whites' and 'the Blacks', seemed to play a central role in most of my conversations with both TB patients and traditional healers. Sometimes this topic was closely intertwined with stories of distrust towards doctors, including accusations of stealing blood for witchcraft and keeping Black people dependent on medication by not instantly curing their TB (and HIV) despite the availability of medication. Moreover, the hospital was perceived by many as a place where people died and therefore some people were reluctant to seek care in this place.
In general, traditional healers were the first point of care for many Gabonese people. The majority of Gabonese people were initiated in a traditional cult, marking their transition into woman-or manhood, and protecting against evil spirits that might bring diseases. Udagudu, one of Joessabe's TB patients, explained: '[If you are not initiated] you are not protected against evil spirits and more importantly, you are not part of the group'. Members of this group were expected to seek help from their traditional healer first and foremost, and the hospital was largely not an accepted place to seek healing.
At the governmental institute IPHAMETRA, the presentation of traditional medicine formed a sharp contrast with those accounts described above. Instead of a Black/White dichotomy, discursive practices reflected a merging of therapeutic traditions. Biomedical terms, such as 'consult' and 'prescription', and written instructions were combined with traditional terms. This mixing was also visible when I went to visit the IPHAMETRA centre in what appeared to be a big, sterile, white building with traditional healers wearing white coats, where there was both a laboratory and a pharmacy with herbal medications. These biomedical elements gave the traditional healing institute the impression of a hospital. Additionally, the employees of IPHAMETRA underlined the need for traditional healers to professionalize and to collaborate with medical doctors.

The taboo of combining therapeutic traditions
The state's Bwiti campaign and IPHAMETRA's efforts to professionalize or enhance cooperation between traditional and medical healers seemed to have little impact in Lambaréné. On hospital territory, biomedical traditions dominated formal discourse. Biomedical doctors explained to me that the majority of patients attended the hospital only when they had developed a very advanced stage of TB, because they had previously sought help from other healers. At times patients arrived who were close to dying. In general, patients were very reluctant to discuss their visits to faith healers or traditional healers with their medical doctors. Doctor Jean explained why this was problematic: 'It is so important to know what healing they have done before coming to us. So I say: "Tell me, because I can see you are traditional".3 But even then patients will deny they have done any traditional healing practices'.
During an in-depth interview at the home of TB patient Omelia, we discussed the topic of secrecy regarding traditional healing in hospitals. Omelia and Bikene, a local researcher, explained that TB patients and hospital-based health workers were reluctant to admit using other healing modalities, for fear that doctors would scold or ridicule them, and make them feel ashamed. Moreover, patients were aware that doctors wield considerable power and were afraid that this might negatively influence their TB treatment. These boundary-making processes between biomedical healing and traditional or faith healing were enforced by biomedical healers, because they often viewed the latter as problematic, due to the number of TB patients who delayed seeking biomedical health care. Patients additionally engaged in boundary-making processes via attitudes of secrecy, as they understood that reliance on traditional and faith healing was disapproved of within the hospital setting. Camilla, one of the nurses, expressed firmly, 'You can't combine different sides. That is not good. That is a taboo. You can go to the hospital or to the traditional healer'.

Justifying a change of therapeutic tradition
Patients regularly mentioned the need to put their faith into one therapeutic tradition at a time. The case of Cedric is a good example: I met him at the Albert Schweitzer Hospital, after he had received his first package for TB treatment, and I offered him a ride home. We needed to drive for two and a half hours, take a canoe to cross a river, and walk for half an hour to reach his home. In contrast to his initial claim that everyone in his village goes straight to the hospital when they fall ill, he later elaborated on the role of different therapeutic traditions within his village: 'When I fell ill my uncle tried to heal me with medical herbs from the jungle. In every family there is someone with herbal knowledge, so that is the first thing to do. When this didn't work, I went to the pharmacy. I kept coughing and then I turned to four different traditional healers. If a traditional healer can't heal you, you try another one, and so forth'. Cedric recalled that '[the traditional healers] gave me the diagnosis "Tuberculosis of the Hospital". That meant they could not do anything for me anymore', and he was advised to go to the hospital.
Cedric explained that he had lost confidence in the effectiveness of traditional treatment, which spurred him to try out the medicine of 'the Whites'. In his description about shifting therapeutic traditions, he did not refer to the long distance between his village and the hospital, which could have posed an obstacle to accessing biomedical care. In the end, Cedric did not engage in multiple therapeutic traditions simultaneously, maintaining the boundaries between therapeutic traditions.
The choice to follow only one healing practice at a time can be linked to the often-heard phrase that it is important to have faith in your treatment. TB patient Fussala emphasized why: 'If you believe [in] it, you will get cured, but if you don't believe in it, it won't work.
That's how it is. For the traditional healer and for the hospital too. For all different healing methods'. She explained that a patient turning to different explanations for TB at the same time would hinder their own healing. Both patients and healers expressed that believing in one treatment at a time not only stimulated treatment compliance but also enhanced the healing process on a spiritual level.

The danger of combining therapeutic traditions
Besides faith or spiritual considerations, people also gave physical reasons for engaging in a single treatment at a time. I was introduced to Jack, a nurse from the hospital, whose sister was following TB treatment at his hospital. His sister lived next door with her husband who was a traditional healer. Jack, and with him many other biomedical health workers, had to navigate a pluralistic therapeutic landscape within their own family. Jack talked with respect about the work of his brother-in-law and described how his sister followed traditional treatment for eight months using herbs and plants from the jungle, but was not getting better. When she decided to go to the hospital she also stopped taking traditional treatment. Jack explained that the two treatments should not be combined, because the body was not strong enough to deal with two powerful treatments at the same time. This resonates with the statements of biomedical doctors who advised patients to not combine treatments because of possible negative drug interactions. However, they argued that patients should only take biomedical treatment.
There appeared to be a fine line between traditional and biomedical therapeutic traditions in the case of a biomedical doctor in Libreville, who was at the same time a traditional healer. However, instead of an ostensible intertwinement of therapeutic traditions, the doctor stated that healing traditions were strictly separated: 'You should not combine the different healing practices. So we have different consults for them'. The therapeutic traditions he engaged in were carefully separated in space, practice, and in speech. It was not considered 'good' to combine two healing traditions, as these were considered too powerful to be used simultaneously. Yet, it was deemed safe to first do a biomedical consult with him in his biomedical consulting room and then continue with a traditional approach in his traditional consulting room.
These boundary-making processes -keeping treatments separate by time and respecting the danger of combining powerful traditions -were similarly presented during interviews with faith healers. Despite the fact that the church's doctrine condemned traditional healing and prohibited communication with ancestors, most faith healers accepted the position of traditional healers within the therapeutic landscape. Aruna explained that even though it was 'wrong' to combine faith and traditional healing, everyone in Gabon would 'follow his roots' and make use of traditional healing. He explained: 'We are in Gabon, in Africa. It is part of where we are and who we are. Everyone attends the traditional healer, you can never stop that. And some are able to heal TB. But you should never combine two big forces. Then it is better to choose [between traditional or faith healing]'.

Therapeutic traditions in their essence the same
Patients and traditional healers frequently mentioned that healing from the Whites and from the Blacks were in their essence the same. Dala explained how this worked: 'The Africans, the plants they use from the jungle, those are used by the Whites as well. The Whites are in a way using the knowledge of the Blacks. They only compress it in tiny, balanced pills. In contrast, the Blacks cook a big pot full of medication and tell you that you should drink everything. But in fact, it is the same thing, exactly the same medication. So it wouldn't make sense to use both. Patients choose one healing'.
The idea that traditional and biomedical healing are in their essence the same was repeated by several biomedical doctors who explained that their pills consisted of plants from the jungle. Yet, they did stress the importance of laboratory-made pills and the careful balancing of ingredients (again referring to overdosed patients who had followed a treatment from traditional healers). This idea reflects a blurring of boundaries between therapeutic traditions, but triggers boundary-making processes when patients ought to make a choice between therapeutic traditions. Joessabe emphasized, 'You either go to the hospital or you go to the traditional healer. It is your own choice'. This would indicate that choice is then no longer based upon the quality of treatment, but on your culture and religion, accessibility, or personal preference for a type of health care.

Discussion
TB patients' and healers' discursive practices regarding their navigation of therapeutic traditions reflect processes of conventional boundary making. This is distinct from the concept of boundary work, in which boundaries are framed as arbitrary, permeable, and flexible. In anthropological literature in general (Gieryn 1983;Lamont and Molnár 2002), and in medical pluralism literature specifically (Kleinman 1980;Vaughan 1991;Bernstein 2001;Luedke and West 2006;Krause 2012), boundaries are often described as being maintained, crossed, changed, shifted, and reproduced. With the notion of conventional boundary making, I also refer to the maintaining and crossing of boundaries, but additionally call attention to people's reluctance to remake boundaries. This notion of boundary making resonates with theories of multiculturalism, in which people may resist the dissolution or shifting of old boundaries in order to protect the purity of their cultural identity (Harrison 1999). Along the same lines, anthropologists are concerned with the notion of boundaries to describe how ethnic groups use typification systems to define who they are and who they are not (Tajfel and Turner 1985, 16-17).
I do not want to imply the existence of timeless, coherent, isolated medical traditions. Rather I have described how conventional boundary-making processes are reflected in TB patients' and their healers' discursive practices. These involve an interplay of personal choices for health care, relational processes between therapeutic traditions and healers, and notions of purification, distinction, credibility, and authority. The way these processes influence connections, separations, and alignments within the therapeutic landscape affects how therapeutic traditions are defined and positioned in Gabonese society. Given increased scholarly interest in the shifting and reconfiguration of boundaries between therapeutic traditions, however, these themes may take on new meaning. Moreover, some themes may shed some new light on how and why patients navigate therapeutic traditions and may provide additional perspectives for medical pluralism debates. Moreover, I have placed these findings in the context of TB in the larger society, looking at state power, Gabonese therapeutic identity politics, and structural violence.
The findings of this study suggest that boundary-making processes exist in various forms and carry disparate values and meanings. Respondents emphasized the differences between therapeutic traditions and their associated medication with mention of the dichotomy of 'Medicine of the Blacks' and 'Medicine of the Whites'. This dichotomy roughly splits the therapeutic landscape and its actors, even though each entity represents a wide variety of therapeutic traditions.
This boundary-making process has spatial and discursive dimensions, as it influences what topics can be discussed in which areas. For instance, it is taboo to discuss traditional healing on hospital territory, and it is not considered appropriate for patients who are initiated into a specific cult to (first) consult a biomedical doctor. The latter condition was sometimes enforced by discursive practices including conspiracy theories in which Whites oppress Blacks by not providing them an instant cure for their TB and by stealing their blood for witchcraft. These stories reflect concerns and anxieties regarding biomedical health workers and hospitals, places that are often conceived of as places where people die. Likewise, Saethre and Stadler (2013) describe tales of white malice in the field of public health that exist throughout sub-Saharan Africa. Such distrust may be strengthened by the biomedical TB programme's lack of medications and diagnostic facilities, and its resultant inability to cure all TB patients or prevent them from dying. Muela (1998) similarly describes how malaria patients deal with unexplained events related to biomedical treatment, for example, when symptoms persist or diagnoses do not give results, and how patients consequently seek an explanation and care elsewhere. Moreover, biomedical facilities are for many TB patients not accessible, because of the distance or travel costs, as illustrated by the case of Cedric who had to travel for almost a whole day to reach the hospital. Such infrastructural barriers making biomedical care inaccessible are similarly described in Greene's (2004) ethnography about TB patients in Bolivia. Delaying care also lowers the chance of recovery, adding to distrust about the adequacy of biomedical care. Beyond this Black-versus-White dichotomy, faith healers' discursive practices similarly reflected taboos informing boundary-making processes, as they condemned patients who went to traditional healers, accusing them of working with the devil.
Making boundaries can be linked to underlying power strategies. Both healers and patients' discursive practices reflected clear distinctions between therapeutic traditions, leaving no space for processes of syncretism, bricolage, or eclecticism. The various healers mostly presented themselves as being purely biomedical, traditional/Gabonese, or Christian, in an attempt to pursue an authentic position within Gabonese society. Through such boundarymaking processes, healers may strive for greater political accountability and a strengthened position within the therapeutic landscape. Discursive practices define what healers are and are not, even though traditional and faith healing have included a wide variety of healers and healing practices. Calhoun (1991, 108) similarly describes how healers feel part of a symbolic (therapeutic) community despite important internal differences.
Additional discursive practices illuminated temporal boundaries, which compelled patients to negotiate therapeutic traditions in a certain way: patients were not supposed to engage in different therapeutic traditions simultaneously, but could make use of a variety of traditions one after another. If patients exhausted all means of one therapeutic tradition and consequently lost faith, it was socially acceptable to move to another therapeutic tradition.
The choice of tradition depended heavily on the patient's religion and associated faith, but patients were allowed to change and put faith in a different therapeutic tradition over time.
Similarly, the use of multiple therapeutic traditions simultaneously was thought to have damaging (physical or spiritual) consequences for patients. Crossing this boundary by mixing therapeutic traditions would not cure patients, yet choosing one after the other might result in healing. In this boundary-making process, each therapeutic tradition was constructed as very powerful, allowing each to claim its own space and legitimization within the therapeutic landscape.
Discursive practices about 'all healing being in their essence the same' resonate with a study that found antituberculous activity within medical plants in Cameroon (Nkenfou et al. 2015). Such statements seem at first to contrast with the boundary-making processes described above, because if they were found to be equally reliable it would indicate that boundaries had become blurred. But this blurring of boundaries was not related to a transgression of boundaries by traditional healers who copied biomedical aspects (Simmons 2006). Respondents simply explained that the treatment of distinct medical traditions originated from the same plants in the jungle. Interestingly, this blurring of boundaries does trigger boundary-making processes on an individual level for patients, as they must choose one or another tradition, because it would be redundant to do both. This can be related to 'having faith in one therapeutic tradition at a time'. However, individual boundary-making processes informed by the idea that therapeutic traditions are essentially the same omitted a temporal division, as a change of therapeutic tradition later in time was difficult to justify. Therefore, a change of therapeutic tradition is often justified by a change in TB aetiology and consequently the need for a new therapeutic approach.
On the governmental level, discursive practices did reflect processes of negotiation, interchange, and redefinition of traditional medicine shifting and recreating boundaries. IPHAMETRA's aim to professionalize traditional healing was reflected in its laboratories, biomedical devices, and certification of professional traditional healers. IPHAMETRA not only transgressed boundaries between traditional and biomedical therapeutic traditions but also created new boundaries by presenting traditional healing as merely an herbal and not a spiritual practice. Meanwhile, they strove for collaboration between traditional and biomedical healers, breaking down boundaries between therapeutic traditions.
Despite the potential of a state to influence boundary-making processes between therapeutic traditions (Zhan 2001), IPHAMETRA's reconfigurations and reproductions of boundaries were not reflected in the discursive practices of healers and TB patients. In contrast, discursive practices showed processes of boundary making in which the intertwinement of medical practices and therapeutic traditions was considered taboo. Both patients and healers deployed discursive practices that clearly separated the therapeutic traditions and stressed that mixing and combining therapeutic traditions was not appropriate. Herzfeld (1997) likewise describes an incongruence between formal, state-sponsored versus local discourses about identity and boundaries.
Looking at literature about Gabon's therapeutic landscape, the state's discourse regarding therapeutic traditions provides an extra layer of context while analysing boundary-making processes, showing them to be part of a larger politics. In Gabon, state-led identity politics and the government's aim to institutionalize traditional medicine played a role in dividing the therapeutic landscape (Swiderski 1988;Samorini 1993;Ngolet 2000). This is particularly interesting within a postcolonial context wherein social orders have been contested (Werbner 2001).
Explaining the underlying reasons for national identity politics and the focus on Bwiti in Gabon, scholars have made contrasting arguments. Samorini (1993) and Swiderski (1988) describe how the imagined ecumenism of Bwiti society has been supported by the Gabonese state and functioned to empower the Gabonese people and traditional medicine against colonial powers. Ngolet (2000, 68) explains these national identity politics and the emphasis on Gabon d'abord (Gabon first), referring to the state's desire to convince the Gabonese population of their perfect leadership in order to stay in power. Meanwhile, most of Gabonese society remains poor and holds a marginalized position in society. Ngolet explains that this political strategy elicited a counter-reaction among disillusioned Gabonese people, against the state's usurpation of Bwiti (and other emerging traditional societies and religious sects) for its own instrument. In parallel, these processes provided a new space for a wide variety of therapeutic traditions.
Following Ngolet's line of reasoning, I ask whether these patients' and healers' discursive practices can be seen as part of the strategy to rebel against the postcolonial government.
Resorting to the purity of therapeutic traditions might be considered a remedy to deal with 'associated societal problems like political and economic marginalization, disturbed social order, and an associated sentiment of frustration' (Ngolet 2000, 59). This is particularly interesting in the context of TB, a disease that mainly hits the most marginalized and vulnerable people. The notion of conventional boundary making thus needs to be nuanced with attention to structural violence that may also play an important role.
The concept of conventional boundary-making captures how my respondents' discursive practices reflected their attempts to keep different therapeutic traditions separate. This neither entailed rethinking local illness explanatory models nor creating a new syncretic model. Boundary-making processes were intentionally enforced to contrast, strengthen, purify, and to divide the therapeutic landscape in time, space, and speech. Meanwhile, these processes guided patients in their navigation of this space and their search for cure and could confirm healers' place and authority in the therapeutic landscape. This study suggests that scholars studying TB treatment must pay attention to structures of inequality and poverty; in this case, relevant factors include a national TB programme that does not meet international requirements, a national identity politics aiming for reinforcement of traditional healing, and a desire to purify therapeutic traditions in response to the dominant position of the postcolonial government. Still, as this study has focused on participants' discursive practices, no conclusions can be drawn regarding their behaviours. Future ethnographic research might further elaborate on TB patients' and healers' daily activities and how they navigate therapeutic traditions in Gabon in certain ways and not others.
About the author Dr Lianne (A. L.) Cremers is a visual medical anthropologist from the Amsterdam University Medical Centre and a lecturer at the University of Amsterdam, the Netherlands. She conducts research on patients' experience with tuberculosis and health care in sub-Saharan Africa. She has additionally worked as an anthropologist for the KNCV Tuberculosis Foundation, the Amsterdam Institute for Global Health and Development, the European Centre for Disease Prevention and Control, and the AIDS Foundation. Her areas of expertise are ethnography, visual anthropology, infectious disease, community-based care, and interdisciplinary research.