Abstract
Within institutional ethnography (IE), texts—reproducible written and graphic materials—are understood to play a central role in coordinating social relationships. This is the case because of their ubiquity and because texts allow proceedings in one place to influence actions in another. This article considers the role of texts in settings that are less text-saturated than the Western societies whose organization Dorothy Smith first sought to understand through IE. Drawing from insights of a study examining the social organization of maternity care in rural northern Uganda, I discuss the differences in the role of texts in Global South research including the role of illiteracy, the need to read for absences or silences, and the prominent role of international texts. Identifying texts and observing their role in sequences of action are nevertheless essential for institutional ethnographers working to understand social organization in the Global South. This article seeks to contribute to IE methodology in Global South research settings.
Keywords
Introduction
This discussion of institutional ethnography (IE) and texts in Global South research settings draws from research focused on the social organization of maternity care and birth in Amuru subcounty, a rural area recovering from conflict in northern Uganda. The study’s methodological and analytical approach was guided by the insights of IE. IE is an approach to social research developed by Canadian sociologist Dorothy Smith. IEs focus on analyses on the processes of social organization by learning about people’s everyday actions and experiences and how they are coordinated via observable institutional processes. Smith’s insights leading to the development of IE were grounded in the Canadian women’s movement as she worked to bridge the gap between women’s lived social realities as mothers and sociological or institutional paradigms of motherhood (Griffith & Smith, 1987; Smith, 1987, 1997). Since these first inquiries, IE scholars have contributed to other movements for social change. As health sociologists Eric Mykhalovskiy and Liza McCoy (2002) suggest IE’s focus on “how people’s daily lives and troubles are organized socially and institutionally” (p. 20) is conducive to the creation of knowledge supporting equity, representation, and inclusion.
A central focus of this discussion is texts, broadly defined within IE as written or graphic materials that can be reproduced and are used in coordinating translocal activities. That is, texts produced in one place are influential in coordinating activities and knowledge in a different place. Importantly for researchers in rural Global South settings, the cultural and technological shifts that make texts so central have not occurred universally. For texts to extend beyond the place they are created, and thereby be significant to IE, they need to be transmissible and reproducible. Such processes are facilitated by a range of technological innovations, from printing to texting, whose reach has been uneven. How and whether texts are reproduced depends on available technologies, norms within bureaucracies, and other factors that differ between research settings.
In the study discussed here, I mapped the social coordination of maternity care and birth in Amuru subcounty, northern Uganda, starting with the accounts of 45 childbearing women of age over 18 with at least 1 child of age under 2 and extending to interviews with 22 formal and informal health workers and health administrators. This research also included analyses of locale-specific texts, discourses, and policy documents related to maternity care and how they were activated to coordinate maternity care and practices related to birth. In a setting where many people are illiterate and access to education and information is limited by lack of infrastructure and other factors, I argue that the role of texts is different than in the industrialized urban Global North where a majority of IE studies have been undertaken. Drawing on examples from this study, I discuss the impact of illiteracy on relying on texts as data and the role of international texts.
Although a majority of IEs have been undertaken in Global North settings such as Canada, a growing number of institutional ethnographers are turning their attention to the Global South or toward transnational ethnographies. This discussion of how IE approaches to texts may need to be modified for Global South research settings both benefits from and seeks to contribute to the growing body of transnational and Global South IEs. Daniel Grace’s (2013, 2015) transnational IE follows the series of text–work–text interactions through which “best practices” and model laws on human immunodeficiency virus (HIV) funded by the United States are translated into laws criminalizing HIV transmission in West African states with negative repercussions for public health. Marie Campbell and Katherine Teghtsoonian (2010) also track global governance processes in their examination of how aid effectiveness reorganizes approaches to women’s equality and human rights in Kyrgyzstan. Campbell has collaborated with Sonya Jakubec (Jakubec & Campbell, 2003) to examine how Western discourses have come to shape approaches to mental health in the Gambia. And Aaron William and Janet Rankin (2015) grapple with conducting IE in a disaster zone in which conventional texts were seldom present, discussing how they relied on other empirical data within an IE approach. The work of these institutional ethnographers in tracing global texts to local contexts has contributed to this discussion of methodology for IE, texts, and Global South research settings. This article identifies that in low-literacy settings, such as the rural, Ugandan villages where the study discussed was set, texts operate in different ways than in the text-saturated high-income countries in which most IEs have been set, for reasons including lack of access to information, lack of access to communication technology, poor education, illiteracy, and the disproportionate influence of transnational texts originating in the Global North. The discussion of these considerations will be useful to other researchers working to implement IE methodologies in the Global South in ways that are sensitive to local contexts.
Texts and Their Role in IE
Texts are important as a bridge between local contexts and broader, or extralocal, contexts. They are one means through which decisions made in one place are brought into being through the actions of people in another place. As Smith (2006) writes: The magical character of replicable texts from the point of view of institutional ethnographic interest is that they are read, seen, heard, watched in particular local and observable settings while at the same time hooking up an individual’s consciousness into relations that are translocal. (p. 66)
Texts can be produced through basic or technologically advanced means. Replication is a means via which texts can be influential translocally; therefore, texts that can be and are reproduced are most relevant to IE. Smith and Turner (2014) write that, texts are material objects that carry messages—stone carvings, sand sculptures, writing or pictures on walls, paintings on canvas, writings on cloth, parchment, paper or computer screens, music recorded on records, CDS, or on tape, images on film, television and so on. The texts that are of particular relevance to institutional ethnography are those that are or can be reproduced many times, so that different people can read the same text in different places or at different times; it is their replicability that is central. (p. 5)
Texts and Contexts
Within IE, the importance of texts is often situated in historical contexts that, while broad, are particular to the industrialized Global North. For example, Smith (2002) situates her discussion of texts in European history and North American politics and concludes a study of how texts facilitate ruling relations with the statement: “Advanced contemporary industrialized societies are pervasively organized by textually mediated forms of ruling” (p. 212). Campbell and Gregor (2002) refer to how literate people put texts into action in everyday life (p. 32), while Smith (2006) labels texts as “those most commonplace objects of our contemporary world, so much present that we take their ubiquity entirely for granted” (p. 26). Grace (2013) asserts, “in late (post) modernity institutional knowledge is text-mediated” (p. 38). These examples demonstrate that the social shifts making texts so central in organizing knowledge and activities have not occurred universally. Major industrial, technological, and information shifts have taken place in ways that are patchy rather than even in terms of their reach and their social impact. What has not yet been theorized is the role of texts in societies that are not primarily characterized by industrialization or the ubiquity of media.
Low saturation of media (including Internet access, print media, and even radio) and low rates of literacy are part of what make Amuru subcounty a very different place from most of the Global North where the majority of IE studies have been situated. Amuru is a rural community in subSaharan Africa and is not best described by phrases such as “late (post) modernity” or “advanced contemporary industrialized society.” For example, all farming is typically done with hand tools in Amuru; while motor vehicles facilitate travel between centers, private vehicle ownership is rare, and walking and biking are the most prominent means of transportation. Activities such as checking e-mail, browsing the news, or reading and writing for work that shape or are interspersed through our days in the industrialized west are unusual in rural Amuru, in contrast to Uganda’s cities and towns. The forays of information technology are uneven but by no means absent: even without electricity infrastructure in the subcounty, cell phone ownership and use are relatively widespread with people relying on solar panel fueled charging stations and a knowledge of areas where reception is possible. Rural residents live in hand-built grass-roofed huts. During research, I came to see my own life as highly dependent on textual literacy—I read to fall asleep—in contrast to the limited textual literacy I encountered among rural residents. Importantly, lack of textual literacy is not to be confused with lack of skills, knowledge, or creativity (Abu-Lughod, 1991/2006). During fieldwork, I began to pay attention to how the contexts of low literacy and low media consumption influenced the circulation and operationalizing of texts.
Contrasts such as those described above in access to information technology have been conceptualized as a “digital divide” (Fuchs & Horak, 2008). For many scholars who undertake research in rural Global South settings, communication barriers are prominent. For example, Omona and Aduo’s (2013) discussion of gender issues in post-conflict Nwoya, a northern Uganda district bordering Amuru, found that the marginalization of women was exacerbated by “the low level of information available to women due to poor communication channels” (p. 129). In my own field, the study of maternity care and childbirth, communication challenges—including lack of access to health information, illiteracy, or lack of facility in the language in which health services are provided and lack of mobile phone infrastructure or access—are considered, alongside transportation challenges, poverty, and inadequate services, to be among the nonmedical factors contributing to high rates of maternal mortality. As well as shaping the environment in which women live and seek care, these communication challenges shape the research environment.
Of course, rural sub-Saharan Africa and urban Europe are not simply different and distant places; they are interrelated through transnational contexts. Health care in Amuru, and other locations within sub-Saharan Africa, is governed partially by decisions made in Europe and transnationally. For example, international development priorities are identified by Millennium Development Goals (MDGs) created by the United Nations (UN) in 2000 to identify development priorities and by the Sustainable Development Goals, which replaced the MDGs when they expired in 2015. Another example of the impact of transnational texts on health care in individual African states is the rollout of model laws criminalizing HIV, based on best practices as identified by United States Agency for International Development (USAID) (Grace, 2015). These laws can extend to making HIV testing mandatory and allowing health workers to disclose patients’ HIV status, impacting patient rights and access to care, as was the case with Uganda’s 2014 HIV Prevention and Control Act (Rudrum, Oliffe, & Brown, 2015). The MDGs and other documents identifying development priorities or best practices and thus shaping health-care delivery are texts that, through being reproduced and through the status accorded them by members of the international community, extend to lives beyond the setting in which they were produced. As Grace (2015) argues, “Institutional ethnographic research offers conceptual and methodological tools to help understand how local public health action is coordinated translocally” (p. 12).
For Stacey Langwick (2012), such global development documents are one of the directors of “the choreography of global subjection” (p. 31). Langwick contrasts the mobility of paper with that of people who cannot be in multiple places at once: In contrast to people, paper can span multiple distances simultaneously and relatively cheaply. People cannot always move, and one person cannot be in multiple places at once. Limited resources and busy schedules, among other things, restrict the mobility of people at all levels. Paper, therefore, is a particularly powerful and critical medium for global relations. (p. 44)
Because of such inequities in access to shaping policy and health-care agendas, transnational texts may act as “boss texts” (Smith & Turner, 2014). Boss texts are a text or set of texts that supplies the context for what we can see, hear, and know. There are subsidiary documents that come into being and are organized under these texts, which are positioned at the top of a hierarchy of texts. (Bisaillon, 2012, p. 610).
Cards, Kits, Letters, and Goals: Texts at Work in Amuru
Without reviewing all study findings, I will briefly discuss some key texts that were activated in the social organization of maternity care and birth in Amuru, identifying their role and considering how this is related to the research setting. These include the following examples: the antenatal care (ANC) card, the Local Council I (LC I) letter, and the MDGs. I also consider whether the role of “absent texts” in low-literacy setting and times when turning away from texts is necessary for a faithful description of the organization of social relations.
ANC Cards
One of the important texts coordinating maternity care and birth in Amuru is the ANC card. This is a printed card that is filled in by a health worker to record the patient’s basic personal and health information as it pertains to pregnancy, including the number and outcome of previous pregnancies and HIV status. Having been completed, the card is given to the pregnant woman, who is to present it at future health center visits, including, importantly, the time of delivery. Without the ANC card, participants feared they would be turned away, since attending ANC and testing for HIV were presented as requirement for delivering at a health center. They sometimes referred to attending ANC as “catching the card.” Significantly, however, the recipients of the ANC card were not often literate and were not all English speakers.
1
The card organized care by acting as a ticket to delivery care, meaning that attending ANC and “catching the card” were prioritized by women who sought a health center delivery. Participants feared that without a properly completed ANC card, they might be turned away when they arrived at the health center in labor. A compulsory approach to couples HIV testing contributed to this fear (Rudrum et al., 2015). For example, Betty,
2
a 32-year-old mother, stated that when women arrived at ANC without a husband to test for HIV with, women’s HIV test results were withheld with a threat that future care would be denied. She said, They give cards but without (HIV test) results. Each time you went they would remind you to go together (with your husband) to receive your results. They actually scare you that you won’t deliver at hospital if you didn’t come for results. They can’t avoid, because they know that when they come in labor, or in case of any problem, they need antenatal care, they have to have the card. All of them in the village know that without antenatal care it is not easy to go for delivery when they are in labor.”
As well as shaping childbearing women’s approaches to care, the ANC card coordinated the work of nurses and midwives, particularly with regard to delivery care for patients whose card indicated a positive HIV status; to attend such patients, health workers wore protective gear such as masks in addition to gloves and worked to position the laboring woman in ways that would minimize contact with blood and other fluids. Also significantly, traditional birth attendants (TBAs), those who provided labor support in home environments, were often unable to read the card, in particular the HIV status. This prevented TBAs from making decisions about when, whether, and how to assist at births based on HIV status (Rudrum, 2015). This contributed to the relegation of TBAs to the sidelines, a marginalization that is sanctioned by global health bodies such as the World Health Organization (WHO), but that can make birthing at home more risky for both patients and providers when providers are prevented from accessing adequate information and safe delivery supplies.
The ways in which the ANC card coordinated both patients’ and health workers’ activities in Amuru were different than it might have been in a highly literate, technologically sophisticated, text-saturated environment. The card was less likely to coordinate information seeking, as it might if patients could read it. Patients were unlikely, for example, to point out an error or ask a question about why parts of her medical history are relevant. For patients, the card was important for its symbolic role as a key or ticket to health center delivery, rather than for its content. In a different setting, patients might be more likely to question the compulsion to present the card or the dependence of future care on husbands’ participation. As well as tracing the observable operation of texts in coordinating social organizations, I found it was important to track where texts “stopped” and which text-reader conversations could not happen because of illiteracy or other contextual factors.
The LC I Letter
Also related to HIV status as a factor around which care was planned was the LC I letter. In Amuru, HIV testing during ANC was offered to the husbands of pregnant women. In areas with high HIV prevalence, this practice, known as couples testing, is used to identify women who test negative but who have HIV-positive partners. The goal is to identify HIV-positive women during pregnancy so that they can be treated, thus preventing mother-to-child transmission of the virus. Couples’ testing in Amuru was offered in such a way that it was enacted as compulsory, though this ran counter to policy guidelines (Rudrum et al., 2015). An important way this compulsory approach to couples testing was enacted was through requiring a document from women whose husbands did not test, whether because their husband was dead, absent, or refused to test or there simply was no husband. This letter was to be written by a LC I, a political representative at the parish level. Although these were individual letters, and not texts that were reproduced, they were essentially form letters. The letters tended to state that the husband was dead or in jail, although in fact men were often simply reluctant to test, for a variety of reasons examined in more detail in an earlier article (Rudrum et al., 2015). This letter was a health center requirement; however, it ran counter to a testing policy that, at the time, stated that testing was not mandatory and that individuals could opt out. The role of the letter in coordinating maternity care was facilitated by power dynamics including gendered relational inequities and the exertion of professional and social status. The letter’s role was also leveraged in part by the illiteracy and poor information access of childbearing women. In another setting, such a requirement might be “exposed”: Patients would learn of the official testing policy and refuse to participate. In Amuru, it would be difficult for a woman to state that she wanted to continue with ANC without a husband’s participation or a letter. The overall approach to HIV testing was facilitated partly by an environment in which texts were not ubiquitous. This diminished the possibility for a “competing” text that would undermine the role of the LC I letter and increased the potential for the LC I letter to play a powerful role. To receive the letter, women would have to walk to visit the LC I, plead their case, and coordinate the printing of a letter signed and stamped by the LC I. The difficulty women encountered in obtaining such a document and the relative rarity of presenting a document in this setting helped to position the letter as a significant text coordinating women’s efforts to organize safe delivery care.
MDG 5: Absent but Present
DeVault and McCoy (2006) write about the role of “text-mediated discourses that frame issues, establish terms and concepts” (p. 34). Discourses are overarching narratives that have the power to shape social realities. At the same time, people are active in reproducing discourses and can also modify or disrupt them. Drawing on Foucault, Smith writes (2005) that: “The functions of ‘knowledge, judgement, and will’ have become built into a specialized complex of objectified forms of organization and consciousness that organize and coordinate people’s everyday lives” (p. 18). Discourses operate at various levels of specificity and influence, with some becoming part of a larger ideology. As Smith (2005) explains, “ideological discourses are generalized and generalizing discourses operating at a metalevel to control other discourses” (p. 224). In an IE approach, discourse is always understood in relation to social actors and activities as well as to language.
Through texts including MDG 5, “maternal mortality” not only refers to the everyday/everynight reality of women dying in childbirth but, in the context of sub-Saharan Africa, is also an ideological discourse that influences what can be said and done in relation to reproductive health. The MDGs are eight goals for international development adopted by world leaders at the UN in the year 2000. The goals set targets in each of eight focus areas to be achieved by 2015. MDG 5 is “to improve maternal health” and has two targets: 5.A, “to reduce by three quarters the maternal mortality ratio,” and 5.B, to “achieve universal access to reproductive health.” Target 5.B of the goal is a more recent (2005) addition and has been less prioritized than 5.A (Yamin & Boulanger, 2013). Two indicators measure progress on target 5.A: the maternal mortality ratio itself and “the proportion of deliveries attended by skilled health personnel” (WHO). The narrow focus on maternal mortality, rather than a broader focus on sexual and reproductive health rights, is one way in which the ideologies of conservative politics shaped MDG 5 (Yamin & Boulanger, 2013). Further, the focus on measurable success of the goals can detract from a focus on quality of care alongside quantity of care (Richard et al., 2011; Spangler, 2012). The focus on birth attended by skilled personnel in order to prevent maternal mortality supersedes other possible concerns, including a focus on access to contraception, the right to compassionate care, or any of myriad concerns that also impact women’s sexual and reproductive health and rights. Through the MDGs, those aspects of health-labeled development priorities eclipse aspects of health that have not been explicitly prioritized as development.
Among childbearing women, the discourses framing childbirth and maternity care were mediated very locally, and the role of texts was at times indirect and difficult to trace. For example, MDG 5 was not part of childbearing women’s talk, and neither were related concepts such as “barriers to access” and “maternal mortality rates.” However, participants did speak about barriers without using such terminology, as well as in some cases expressing that they had feared for their lives during pregnancy or delivery. The MDGs and institutionalized concepts around maternal health were drawn on by health-care providers and administrators and thus played a role in how health care was organized. For example, health workers referred obliquely to the MDGs. When a midwife explained that HIV testing practices were “based on the government program where by 2015 no babies will be born with HIV infection,” this was partly an iteration of MDG 6, the goal to “combat HIV/AIDS, malaria, and other diseases.” However, rather than referring to the MDGs as the text health center approaches were based on, she referred to a “national policy.” This exemplifies a text–work–text sequence (Grace, 2015; Smith, 2005), as through work, the text of the MDGs is reprocessed into texts governing national policies or guidelines, which, through work, are then adapted to texts or practices at health facilities.
Blystad et al. (2010) write that “To facilitate the ambition of global relevance, international policy documents or global texts […] appear as ‘existing beyond culture and time’” (p. 2 of 3). However, international policy documents tend to be a product of Global North cultures and tend not to adequately engage with the particularities of place where they are to be implemented. In Blystad et al.’s research, international prevention of mother-to-child transmission of HIV guidelines related to breastfeeding “lack reference to people’s lives” (p. 2) and “do not directly address the immense constraints imposed by poverty” (p. 2); overall, they “are deployed within—but do not acknowledge explicitly—a global landscape characterized by gross inequity” (p. 2). These omissions and oversights are typical of policy that is created to have global reach but without regard to local particularities. Such international texts create a “broken telephone” effect in which they shape policies at all lower levels, often in unanticipated and problematic ways. This was the case with the model HIV laws Grace (2015) examines, wherein the so-called best practices had no evidentiary basis yet were nevertheless replicated in laws criminalizing HIV transmission throughout West Africa with negative impacts for human rights and public health. The LC I letter can be interpreted as in conversation with the MDGs; tracing a text–work–text sequence, one can see that MDG 6 on HIV influences national policies on HIV; exhorted to prevent the vertical or mother-to-child transmission of HIV by a particular date, health administrators innovate locally. The letter can be seen as the end of a text–work–text sequence that began with the MDGs, which were created by the UN. While global bodies including the UN would likely decry the coercion and gendered operation of power evident in the letter and its uses, the letter exists as a way of complying with targets created by those same bodies.
Absent Texts
In addition to the observable texts discussed above, for the institutional ethnographer, it is worthwhile to pay attention to absent texts or texts that are not operationalized because of a low-literacy setting, lack of access to information technologies, or other factors. Most IE studies draw attention to administrative texts such as policies and forms; such “top-down” texts are likely in most settings to play an influential role in organizing the social and can constitute an exertion of power. But patient-generated texts also exist: Lists of questions to ask, letters of complaints, online reviews, or letters to the editor are all commonplace. The context facilitating such patient-generated texts is particular to developed, highly literate societies. The paucity of such texts, or the conditions for their possibility, likely contributes to the nature of health worker–patient relationships in rural sub-Saharan Africa or similar settings, which, unlike North America or similar settings (Nettleton, 2013), have not shifted away from an expert–novice dynamic. Instead of writing back to the bureaucratic channels, protest takes other forms, which may or may not be recorded in text form. A protest that occurred during my fieldwork had women lying down in the road to block the exit of a health official, in order to demand a meeting with him to protest poor health facilities. This wasn’t recorded in the media, so far away people who didn’t witness it had no way of hearing about it (until now). In contrast, when women in Amuru stripped down to protest land grabs, the event was deemed newsworthy, and images and analysis of their protest have circulated (Martiniello, 2015). Health information was often transmitted verbally, in health talks that took place while patients were in the waiting room prior to seeing a care provider or on outreach trips to villages. Such talks drew from yet did not constitute replicable texts. They took the place of the waiting area full of health information brochures that is evident in many Western health-care settings. In such a context, seeing and listening take on an increased relative importance, alongside reading and writing.
As I have described, texts from the MDGs to the LC I letter were central to how birth was socially organized. At the same time, it was also important for me as a researcher to listen to women speak about how the material conditions of their lives contributed to organization processes and rely on their expert knowledge, even when texts as a means of coordinating some aspect of care were absent or not visible. Williams and Rankin (2015) discuss that their investigation of post-tsunami reconstruction in Thailand “did not unearth very many analytically useful texts” (p. 80). While this was not the case in my study (I relied on quite a few analytically useful texts), Williams and Rankin’s insights and dilemmas resonated with me, as they turned to other data that were more evident in the context of rebuilding in the study region. Billboards, public notices, and built infrastructure were features they used as data that I also drew on in my study. They cite a conversation with Dorothy Smith, which reassured them that Texts are not necessary to the discovery of ruling relations. […] while using textual data is one of the important contributions of the IE method, they are not absolutely essential to develop an IE analysis. Other data embedded in people’s talk and activities can be used to link people’s activities across time and geography. (p. 88 citing a personal communication with Smith).
Conclusion
IE studies draw on shared key features of analysis, including specific approaches to institutions, work, participants, and discourse, alongside the IE approach to texts discussed herein. IE, however, should not be viewed as a fixed set of practices; rather IE “is developing new ethnographic techniques as the different terrains and forms of organization demand new approaches” (Smith & Turner, 2014, p. 7). As the reach of IE extends globally including in the Global South, IE methodological approaches will require modifications. This article contributes a consideration of how differences in social setting between the Global South and the Global North impact IE researchers’ work in understanding social organization through tracing texts and how they are activated. Understanding the interaction between setting and methodology facilitates sensitive and rigorous research in the Global South.
Footnotes
Acknowledgment
I wish to sincerely thank Daniel Grace, as well as two anonymous reviewers, for thoughtful feedback on this article.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.