Signs of Mission: Material Semeiosis and Nineteenth-Century Tswana Architecture

Abstract

The missionary encounter between the London Missionary Society and Sotho-Tswana communities of southern Africa has been explored by Jean and John Comaroff as work that took place at the level of both signs and practices. In this article, I consider what a Peircean semeiotic might offer to this narrative. I argue that it provides ways to disrupt the sometimes binary relationship of signs and practices while also providing opportunities for productive interdisciplinary conversations about the affective, material, and processual nature of changes in belief and practice.

Efforts to change belief are often tied to efforts at remaking place. A host of studies of nineteenth-century missionary activity in Africa have shown how this played out in many different parts of the continent (e.g., Showers 1989; Jacobs 1996; Harries 1997; Ranger 1999; Leonardi 2003). As missionaries attempted to introduce Christian beliefs, they also worked to create new ways of dwelling that imported moral ideals from their home countries. Efforts were made to change the houses that people lived in, to reconfigure the structure of families, and to inculcate new modes of clothing and adornment and of work and play. These accounts demonstrate that the beliefs brought to Africa by nineteenth-century missionaries did not float free of the material world. Instead, they were embedded in an entirely different habitus, which was organized around very different landscapes and traditions.¹ Attempts at conversion, therefore, met with resistance both in the discursively formulated responses of the people encountered and in the distribution and organization of the temporal and spatial world.

In their groundbreaking studies of mission in southern Africa, Jean and John Comaroff have shown how the work of changing belief was closely allied to changing the form and manner of living in the world of those that missionaries aimed to convert. In a series of studies, they have explored the changes in “signs and practices” that were entailed by the encounter between missionaries (primarily from the London Missionary Society) and communities now commonly grouped together as “Tswana,” showing how the work of missionaries took place through the material world as much as through language (Comaroff 1985; Comaroff and Comaroff 1986, 1991, 1992, 1997). The Comaroffs’ focus is the “Southern Tswana” Setswana speakers living in present-day South Africa who are part of a broader Sotho-Tswana language community also found in Botswana and Lesotho today. Their research explores the signs and discourse of mission and the practical changes that were enacted on the ground; it has been influential in calling attention to the exported habitus of missionaries as much as their efforts to inculcate Christianity. In exploring the effects of mission on Tswana communities in southern Africa, the Comaroffs trace the ways that people reworked the ideas and practices brought by missionaries while being drawn irrevocably into the developing colonial economy and the emergent political and social formations of modernity. These narratives of the encounter are deeply engaged with the material world of practices and things, and in foregrounding the material entailments of mission the Comaroffs tread the territory of archaeologists as much as historians. In this article, I revisit their and others’ work on nineteenth-century mission among Tswana communities, offering an analysis based in a Peircean semeiotic to think through the material-semiotic dimensions of the encounter.² I am particularly interested in how a Peircean orientation opens up avenues to explore the affective and material qualities of interpretation and to destabilize the analytic binary of signs and practices.

Proportional to its influence on southern African studies, the Comaroffs’ anthropological history has come under strong criticism. One line of critique has been that their account credits too much initiative and agency to Europeans at the expense of local people (Gulbrandsen 1993, 1996) and that they present a generalized ethnographic picture of Tswana practices rather than looking at the specifics of historical change (Landau 1992). They have also been charged with writing a narrative that is overly polarized in terms of the dialectic between Tswana and Europeans (e.g., Landau 1995, xxii; Donham 2001; Elbourne 2003). The word Tswana refers today to a collection of differentiated but closely related Bantu-speaking communities found in Botswana and South Africa. At the time of the London Missionary Society (LMS) mission, the term was not in common usage as a collective noun of identity but rather developed as a product of the colonial encounter (see, e.g., Volz 2003). Although the Comaroffs are clear on this point, and make explicit that they are interested in precisely how the dualisms of Tswana/European, black/white, and heathen/Christian emerged from the encounter, it is possible that their reliance on missionary sources gives the impression of a more homogeneous grouping than was the case (Landau 1999; Feierman 2001). This is despite emphasizing the transformative and dialectical nature of the “long conversation” between Tswana and Europeans, in which no one was left untouched. Jean and John Comaroff have responded to many of these criticisms (Comaroff and Comaroff 1997; Jean Comaroff and John Comaroff 2001), but in a series of works of such scope and richness, there will always remain plenty of opportunities to find things to discuss. This article is not meant as a continuation of these critiques but rather as a thought experiment: an effort to think through the material evidence of nineteenth-century changes in Tswana ways of life, by putting a Peircean perspective into conversation with the Comaroffs’ work and that of other historians and anthropologists of southern Africa.

The encounter is situated in the Comaroffs’ account as a heterogeneous affair, occurring “in multiple registers … in the spiralling flow and counterflow of signs and objects, means and ends, that drew indigenous communities into an expansive imperial economy” (Comaroff and Comaroff 1997, 5). To write this narrative, they describe the “saturated signs” of mission—books and gardens, coins and crosses, mirrors and clocks, all “ingeniously redeployed to bear a host of new meanings” (Comaroff and Comaroff 1992, 5; Comaroff 1991). In thinking about the ways in which Christianity became caught up with local beliefs and practices, the Comaroffs have foregrounded the “signifying practices” through which Tswana communities sought to engage with missionary incursions, “to reestablish the coherence of their lived world and to render controllable its processes of reproduction” (Comaroff 1985, 5). This approach was developed as a way to complicate and reintegrate the long-standing analytic dualism of practice and consciousness, or agency and structure, which still pervades much social analysis. Drawing together the practice theory of Bourdieu (1977), Sahlins’s structuralist-inflected concern with history (1976, 1981), and Stuart Hall’s Gramscian-influenced work on ideology (1977), Jean Comaroff characterized this in 1985 as a problem of symbolic mediation, arguing that signs provide a site where material and semantic orders of determination meet and where the dialectical relationship between them may be explored (1985, 5). Embedding signification in practice provides a way to consider “the meaningful structure inherent in practice and the practical structure inherent in meaning” (6 n. 2).

In many ways, the Comaroffs’ early dialectical account of how signs and practices were folded into one another was a precursor of the current anthropological interest in materiality, which takes as its focus the mediations between people and the material world in an effort to complicate any easy divide between signs and objects (e.g., Graves-Brown 2000; Miller 2005; Knappett and Malafouris 2008). The Comaroffs’ narratives of colonial encounters in southern Africa were among the first to fully explore the imbricated nature of signs and practices. In doing this, their work has examined how attempts to change belief worked at two levels, one of which attempted to “overwhelm … with arguments of images and messages” and the other of which tried to “inculcate … the spatial, linguistic, ritual and political forms—of European culture” (Comaroff and Comaroff 1991, 311). These different dimensions are emphasized through the reiterated pairings of “signs and practices” and “signs and objects,” which both call attention to the mutual constitution of ideological content and hegemonic forms, while also frustratingly reinscribing the divide. It is here that Peirce can perhaps provide a useful point of intervention, a way of further disrupting the foundational tension between the universe of representation and the world of practice and things (see Preucel and Bauer 2001). This is to take “signifying practices” seriously but to turn to a Peircean semiotic theory to conceptualize them and to see what his approach might offer.³

Peirce’s sign relation attends to things as much as concepts, and archaeologists have found that it provides rich opportunities for exploring the ways in which semeiotic processes operate through and within the world (Gardin 1992; Bauer 2002; Preucel 2006). As I explore below, Peirce’s triadic model of semeiosis offers at least two powerful ways of shifting the terrain of analysis. The tripartite nature of his sign relation provides an effective means to cut across and destabilize the binary categories outlined above, and its processual and unfolding vision of semeiosis provides routes to theorize performance—in Robert Preucel’s words, to “embrace the dynamics of mediation” (2006, 249).

To undertake this task, I start by reviewing the rich literature on nineteenth-century Christian mission among Tswana communities in southern Africa. Apart from the work by the Comaroffs, a number of historical accounts have given the production of place a primary focus in their analyses (e.g., Grove 1989; Landau 1995; Morton 2004; Gulbrandsen 2007). These accounts can be juxtaposed with the careful archaeological work that has been carried out at key mission sites among the northwestern Tswana in present-day Botswana (Börjeson and Lane 1996; Reid et al. 1997; Lane 1999; Sekgarametso 2001) and in South Africa (Hall 1997), as well as ethnographic and ethnoarchaeological research into the use of space among Tswana communities (Frescura 1981; Larsson and Larsson 1984; Schapera 1984; Fewster 2006). The scholarship around the missions to the Tswana provides an example of the ways in which historical, archaeological, and ethnographic strands of research can work together productively to explore how changes in routine activities and practical understandings of place were semeiotically grounded. Here, I focus on the changes in habitus that are visible architecturally, in order to think through the affective and dynamic dimensions of material semeiosis. My discussion is partitioned into three sections. After beginning with a sketch of Tswana dwelling practices and the impact of mission on them, I turn to look at the question of the relationship between signs and practices in more detail. In this second section, I lay out Peirce’s sign relation and explore its relevance for the missionary encounter. In the third part of the article, I draw on ethnographic, historical, and archaeological accounts to consider the development and persistence of architectural forms during the latter half of the nineteenth century. I focus on two Tswana polities, both in present-day Botswana, starting with the Ngwato (ruled by Khama III from 1875 to 1923) and then turning to the Kwena polity, united under Sechele I in the second half of the nineteenth century. In exploring the semeiotics of architectural change, I consider what this offers for understanding the changing political processes and gender relations in the later nineteenth century (see fig. 1 for a map of locations mentioned in this article).

Figure 1. Map of southern Africa showing locations mentioned in the text. Present-day country boundaries and capital cities are shown in gray.

Tswana Ways of Dwelling and Missionary Impact

Missionaries traveled widely in southern Africa and focused their efforts on the densely settled conurbations, or metse, that were found across the region. These urban centers had a hierarchical political organization with a paramount chief, or kgosi, at the head, providing an intuitive point of contact for Europeans interested in inserting themselves into Tswana communities. Despite variation within and between the different Tswana polities, missionaries predominantly interpreted them in terms of the nucleated settlements they knew from Britain, translating the concept of motse (pl. metse) as ‘town’. What this obscured was the way in which metse were inextricably wrapped up with the authority of chieftainship. The kgosi’s central position was echoed in the idealized concentric layout of the settlement. At the heart of the motse lay the chief’s ward, composed of a roughly circular arrangement of house compounds with the chiefly court (kgotla, pl. dikgotla) and the chief’s cattle at its center. The kgotla and the chief’s ward were surrounded physically and conceptually by the other wards of the settlement; these in turn were ringed by agricultural fields, then pastureland, and finally the wild uncultivated land of the bush (Maggs 1976, 277; Comaroff and Comaroff 1991, 128–30).

Ethnographic research carried out in the twentieth century has been important for conceptualizing Tswana communities in the nineteenth century. Isaac Schapera described Tswana wards in the early to mid-twentieth century as the most basic unit in the administrative system. The number of wards in a settlement depended on its size, with the smallest villages having only one (Schapera 1984, 46–47). Each ward was broadly kin related and administered by its own headman and his senior associates drawn from the male family heads. All wards were organized along similar lines to the kgosi’s, with houses in a circular arrangement around the open space of the kgotla (Schapera 1935). Wards had their own agricultural lands and local political organization, but the headman also answered to the central chief of the motse. This gave rise to political and economic strains in the life of the motse that Christopher Morton has characterized as a tension between centripetal forces converging on the central meeting place (kgotla) and centrifugal forces pulling people out to the edges of settlement following their cattle and fields (2004, 348). Architectural forms reproduced the circular form of the ward, with circular-plan houses clustered in compounds. Found in front of the house was a semipublic walled courtyard (lolwapa, pl. malapa) that acted as “the point of articulation between the domestic unit and the encompassing structures of the polity” (Comaroff 1985, 58). Behind the houses were more private yards (segotlo), associated with women and where household grain was stored. The central meeting space of the kgotla was associated with men and with their cattle, which were corralled there when not at the cattle posts outside the town. In the nineteenth century, women were not included in political meetings at the kgotla, but they did have a role in religious activities there (Kinsman 1983, 50–51). They also oversaw the cultivation, threshing, and storage of crops. When not at their fields, women’s productive activities revolved around the house and its courtyard; they were also responsible for house building. Although in some cases men erected the exterior wooden roof supports and door, women always built the walls of the house, using a mixture of clay and dung, sometimes with stone foundations (Hall 1998, 247; Sekgarametso 2001, 30–31; Comaroff et al. 2007, 75–81).

Gary Okihiro (2000) has criticized the use of twentieth-century ethnography to reconstruct nineteenth-century Tswana ways of life, arguing that it masks historical change and fails to take account of the variation in Tswana experiences. For example, he emphasizes the role of immigrants in the formation of the Kwena polity in the late eighteenth and early nineteenth century, also suggesting that women’s association with grain cultivation was established during this period (86–89). Equally, Paul Landau has detailed changes in the consumption of sorghum beer over the eighteenth and nineteenth century that were tied to transformations in politics and gender relations (1995, 83–95). Missionary accounts describe how during the nineteenth century women’s produce was drawn upon by men to maintain their political position, both in the form of beer and through agricultural surplus, which was used with cattle to build and maintain alliances and pay bridewealth (Gaitskill 1990; Comaroff and Comaroff 1991, 135–38). This shifted with the introduction of Christianity. Landau shows how the kgosi of the Ngwato polity, Khama III, banned the consumption of sorghum beer as a way of controlling political meetings that could pose a challenge to his rule.

Landau is critical of the “diagrammatic vision of tradition” that is often presented for nineteenth-century social and spatial organization (1995, xxi), and similar criticisms have been made by archaeologists of attempts to use twentieth-century ethnography as a model to understand the deep past of the region. The layout of towns and villages in southern Africa has been remarkably consistent over time, and archaeologist Thomas Huffman argues for the material persistence of a “Central Cattle Pattern” as evidence of an accompanying worldview that goes back as far as a millennium (Huffman 1982, 1986, 2001). While this model has been a powerful tool for interpreting past southern African settlements (papers in Hall et al. 1984; Pistorious 1992; Huffman 2007), it has also come under sustained criticism for its static and homogeneous view of past communities, which leaves little room to understand how the model itself emerged and changed over time (see, e.g., Hall 1984, 1986; Lane 1994–95, 2004; Maggs 1994-95; Beach 1998; Landau 2010, 45–47). More recently, research on the eighteenth and nineteenth centuries has integrated historical texts, oral histories, and archaeology, in order to look more carefully at local dynamics and the particular constellations of any particular area and period (Swanepoel et al. 2008; Delius and Marks 2012). Work carried out at the abandoned Tlokwa capital of Marothodi has, for example, drawn upon oral histories, archaeology, and historical texts to write a history of the town that is embedded in its particular local context (Anderson 2009; Boeyens and Hall 2009). Archaeologists are more alive to the nuances of history (Reid and Lane 2004; Behrens and Swanepoel 2008), and historians have argued forcefully for the importance of material evidence in the composition of historical narratives (Morton 2004). The work of John and Jean Comaroff has been vitally important in shifting the ground of historical studies to open up the possibility of more discussion among ethnographers, archaeologists, and historians.

As missionaries arrived into the dynamic and changing landscape of southern Africa, they were perturbed by Tswana ways and forms of dwelling. Although individual missionaries made some adjustments to the radically different milieu of the mission field (as shown, e.g., by Hovland 2007), Jean and John Comaroff have argued that a common feature of the missionary encounter in southern Africa was the rejection of the landscapes and landscape practices that missionaries encountered as sinful and immoral (Comaroff and Comaroff 1997, 279–86; see also Grove 1989). Efforts were made by missionaries to reorganize dwelling practices along lines that echoed or transformed those that they knew from home (Comaroff and Comaroff 1991, 205; 1997, 288–96). The need to relocate converts into Christian space was common to missionary efforts throughout southern Africa and to missionaries of different denominations. Although early travelers commented on the tidy, well-swept forecourts of Tswana houses and were not overly critical in their descriptions (Comaroff and Comaroff 1997, 279–80; Anderson 2009, 5–21), a distinction quickly emerged in the discourse of missionaries between heathen “huts” and Christian “homes” (Reid et al. 1997, 385–86). The work of the missionary was seen to extend into reworking the very space that people inhabited: “how arduous is the work set before native women in changing that heathen kraal into a Christian home,” wrote an Anglican correspondent in the Bloemfontein Mission Quarterly Paper for October 1875, echoing the sentiments of LMS and Methodist missionaries (quoted in Labode 1993, 126). One of the ways in which the Comaroffs argue that this was expressed was in attempts by missionaries to encourage people to live in rectangular-plan houses. Square buildings embodied British values of privacy and the divided and gendered taskscape of the domestic home.

It was not just houses that had to be made rectilinear in a bid to order what often appeared to missionaries as chaotic and immoral. Jean and John Comaroff show that this concern extended to field boundaries, roads, and markings across the wider landscape. The seasonal movement of women from village to fields and the quotidian movement of men inward toward the court and kraals at the center of settlement, where political decisions were made and exercised, expressed the important symbolic connection between center and periphery in Tswana communities. This movement was described by missionaries as “most unnatural” (Comaroff and Comaroff 1991, 137). Focusing on built form, they aspired to replace and rework local architectural forms and practices of inhabitation, seeing “the concentric arcs of Tswana circular settlements … as material impediments to ‘healthy individualistic competition’” (Comaroff and Comaroff 1986, 13). These disordered spatial arrangements were associated by missionaries with filth, disease, and sloth; local houses and villages became viewed as dirty and unhealthy places to live (Comaroff and Comaroff 1992, 279–80). This response seems to have grown in tandem with later nineteenth-century notions of progress, which located Christian civilization as the idealized end point of an evolutionary trajectory of social and political development. As missionary presence grew over the nineteenth century, changes in habitus were inculcated through the creation of mission stations, houses, and boarding schools that taught young people the appropriate domestic tasks necessary to create a Christian home (Labode 1993). In establishing “homes,” missionaries hoped also to establish familiar and morally ordered gendered behavior (Gaitskill 1990; cf. Skeie 1999).

Clearly, missionaries did not operate in isolation from other European actors, and, as Okihiro emphasizes (2000), Tswana towns were neither homogeneous nor insulated from other southern African communities. In the earlier nineteenth century, Tswana settlements were affected by social unrest and political upheavals that troubled the entire region. This period, known as the difaqane (or mfecane), saw increased violence and raiding, as well as the displacement and movement of people across the region (Omer-Cooper 1966; Cobbing 1988; Hamilton 1995; Etherington 2001). For the purposes of this article, however, I focus narrowly on the effect of the European and missionary encounter on Tswana townscapes, attending particularly to the conjunctures of the second half of the nineteenth century, rather than tracing the changing relationships to emerging structures of colonial governance (see, e.g., Porter 2004; papers in Chima and Njoku 2007) or to the wider social and political landscape of southern Africa. My focus is also mostly on Tswana architectural changes rather than on those of the missionaries. However, although I do not explore these issues here, I do want to note that there were a variety of different orientations toward place among missionaries, as there were with other European incomers (e.g., Pratt 1985; Morton 2004, 360). These differences were important in how the various missions were established and developed. Indeed, it should be remembered that not all missionaries were European and that native evangelists played an important part in the spread of Christianity (Hodgson 1982; Elbourne 2003, 19–20; see also Peel 1995). Within the community of Europeans, missionary interpretations of place depended as much on individual positioning within relations of gender, class, and nationality as on personality and political agenda (Beidelman 1974). All of these were expressed and altered in interaction with the people and places they encountered (e.g., Crossland 2006; Hovland 2007).⁴ Along these lines, John D. Y. Peel cautions against overplaying the degree to which Christian belief and nineteenth-century capitalism were aligned (Peel 2000, 5–6; see also Elbourne 2002). With these caveats in place, I return now to the question of signs and practices in Tswana encounters with mission.

Material Signs

Coming back to Peirce’s semeiotic, I begin here by outlining his sign relation, before moving on to consider what it can offer to understanding Tswana reworkings of place at the end of the nineteenth century. Peirce’s sign relation is complicated, awkward to grasp, and nowhere published in full and complete form.⁵ His “labyrinthine reasoning and runaway terminology” (Watts 2008, 187) has perhaps contributed to the relative unpopularity of a Peircean approach in anthropology, at least in comparison to the semiology of Saussure and his descendants. Peirce was insistent that his sign relation was irreducibly triadic, composed of sign, object, and interpretant. He paid careful attention to the relationships among these three elements, detailing, for example, three modalities through which a sign can relate to its corresponding referent, or “object.” These are through similarity (icon), relation (index), and convention (symbol), discussed in more detail below. This dimension of Peirce’s analysis is perhaps the best known aspect of his semeiotic. It has been drawn upon by a number of archaeologists to think about the constitution of past material signs (Graves-Brown 1995; Knappett 2002; Jones 2007; Liebmann 2008). Here I am more concerned with the material aspects of the third dimension of Peirce’s sign relation, the interpretant (see, too, Bauer 2002; Lele 2006; Watts 2008). The interpretant may be understood as another more developed sign, a tendency toward habit, or, ultimately, a habit change that emerges from and interprets the sign-object relation. For Peirce, semeiosis only takes place when an interpretant acknowledges or is elicited by the relationship between sign and object.

The possibilities offered by the interpretant open up ways to address some of the critical commentary on the Comaroffs’ work. Sherry Ortner, for example, has suggested that their focus on African agency concentrates on questions of power at the expense of intentionality. She asks after the projects, purposes, and desires of different Tswana individuals and groups as located within a framework of Tswana terms rather than one imported by missionaries (Ortner 2001). Similarly, Akhil Gupta asks, “under what conditions of colonial syncretism and transgression does the encounter not produce starkly binary identities, or even binary ascriptions of identity?” (2001, 45). What space is there, in other words, for finding projects and concerns that were relevant to Tswana people but perhaps invisible to or downplayed by missionaries? In what ways might Tswana projects have cut across the categories of “Tswana” and “European”? Paul Landau pursues this issue in his authoritative account of mission among the Ngwato polity under Khama III and subsequent rulers (1995). He explores how Khama imagined a new form of polity, a form of “ecclesiastic statehood,” from the ideas and practices imported by missionaries. In doing this, Landau traces how the Christian Word, or thuto, was inserted into the space of the kgotla, physically and conceptually, in the process allowing new possibilities for men and women to rework political power and claim new forms of authority. Like the Comaroffs, Landau also considers how thuto was incorporated into Ngwato ways of life “consumed in teas and soaps, and enacted in other behaviors” (xvii). Gupta suggests that it is here that one might find the voices of the colonized, not in sentences and words but “in styles of dress, practices of agriculture etc.” He wonders how we might incorporate these “voices” through “senses other than the auditory” (2001, 45). Here, I explore how a Peircean alternative not only extends the scope of semiosis to bring words and things within its compass (see also Preucel and Bauer 2001; Cipolla 2008; Watts 2008) but also shifts the terrain of interpretation, providing space to investigate archaeological and architectural traces as forms of interpretation alongside feelings, practices, and words.

Missionary concern with the organization of the house provides an excellent example of how a Peircean orientation can redirect the focus of analytical attention. First, we can consider the sign of the house as it appears in experience and may be recognized as conforming to expectation. Second, when it differs from expectation, the house thrusts itself into consciousness, forcing an energetic physical response. In so doing, it acts on those dwelling within it and challenges habitual practices and beliefs. Third, the house works semeiotically on inhabitants through the sign relations that are formed and reproduced within the context of these affective and energetic responses. These divisions are not arbitrary but are grounded in Peirce’s identification of three “modes of being” that he argued comprise all experience (Peirce 1868a). These may be described as “feeling” (first), “relation” or “reaction” (second), and “habit” or “mediation” (third). Whether or not one accepts Peirce’s metaphysical claim for the reality of his categories, they provide a useful point of entry for considering the effects of mission among the Tswana. Peirce’s conception of experience foregrounds the prejudices and expectations that are brought to any understanding of the world. These have their basis in established habits that are developed recursively in relation to the experienced world. Peirce suggests that any element of the world is first given to experience prereflexively as “feeling.”⁶ His discussion of the dynamic ability of the material world to change and alter representations (“relation”/“reaction”) anticipates much of the recent interest in materiality (as Watts [2008] has explored). Finally, his discussion of “habit” or “mediation” has some similarities with Bourdieu’s conception of habitus that I will explore below. Space precludes any detailed elaboration of Peirce’s categories here, but it should be noted that the categories are refracted in Peirce’s description of the sign relation.⁷ They provide a useful way to think through the different aspects of semeiosis, in particular, the question of how the sign-object relation invites particular affective, energetic, and habitual responses and how the material instantiation of signs and objects may be manipulated to precisely this end.

Given the emphasis in the literature on missionary efforts to rework place, I start here by using mission architecture as an example to sketch out Peirce’s sign relation, before turning to look in more detail at ways of dwelling among Tswana communities of the nineteenth century. The remains of David Livingstone’s house at Mobatsa (where he was based before moving to live with the Kwena) reveal the internal divisions of space that were so important to British missionaries in defining interior space and structuring dwelling practices (Comaroff and Comaroff 1992, 280–81; 1997, 288–92). The rectilinear and divided space of the ruined house shows how private, hidden places were created and maintained within the house. Neither bedroom nor library could be entered from the exterior (the house at Mobatsa is sketched by Livingstone and reproduced in Schapera 1959, 105). The Comaroffs suggest that missionaries worked along two axes to rework space: they made efforts to create an exemplary model for dwelling practices and also “sought, by their own deeds, actively to intrude their designs onto the local terrain and into local habits and habitations” (Comaroff and Comaroff 1997, 288). In Peircean terms, these actions can be understood as signs with different relationships to a similar object, one iconic and the other indexical. Mission stations attempted to reproduce the spatial patterning of British towns and villages, with enclosures and gardens and separate places of burial. In this sense, they were in an iconic relationship with remembered houses and villages back home and, as the Comaroffs have shown, provided a model for living.

Michael Taussig (1993) has explored how such a model gains its potency from the original that it copies, taking on some of the force and power of its progenitor. However, this was a model that could only be fully appreciated by those who could identify the iconic relation and already had some familiarity with the idealized British original. The morality that was inscribed in the house was clear to the British missionaries, and it was especially clear at moments when it was transgressed through inappropriate actions. Richard Parmentier has shown that the patterned social relations that are inscribed through performances in place are not simply iconic descriptions of a cultural ideal (as Landau critiques) but that the instantiation itself is meaningful. He argues that such spatial models act as “instruments of limited self-representation” (Parmentier 1987, 125–27). If the model was evident to the missionaries, it was opaque to the people who encountered it in southern Africa, at least at first. Such a model needed to be experienced to be properly understood. Within the cluster of iconic sign relations that the house offered, indexical signs were needed to indicate how the space should be negotiated. Such sign relations were distributed throughout the house and the performances that took place within it, for example, the proffering of a chair or a teacup to drink—not only might these objects be gestured toward, indicating the proper place to sit or to sup, but the affordances of the objects themselves acted as indexical signs of the appropriate way to sit or to grasp a cup and drink (cf. Knappett 2004; Kockelman 2005, 2006). A closed doorway that blocked entrance to the private and sequestered parts of the building can also be understood as an indexical sign whose object was to restrict the movement of people. Such indexical signs establish a connection with their objects, asserting some kind of real or existing relationship. They are also located in a particular place and time, making indexical challenges to Tswana dwelling practices more restricted in scope and effects. People had to actually experience the interventions of missionaries on a case-by-case basis to understand their indexical power to demonstrate new ways of dwelling.

Peirce also outlined a third mode, of symbolic relation, in which the object of a sign is defined through habitual association. A biblical verse above the mantelpiece relied on conventions of language and writing in order for its object to be understood. Different forms of dress for men and women were established conventionally and through training and upbringing. Symbolic relations rely on familiarity with historically and culturally specific traditions of practice to be understood, and, as Parmentier has explored, they are oriented toward the future “in that this semiotic relation is essentially a processual regularity” (1985, 840). Symbolic relations resonate with expectation and anticipation, while also being unreadable to those who are not familiar with the conventions under which they are specified. As these examples indicate, the three modes through which the sign specifies its object rarely operate in isolation. Both iconic and indexical sign relations have conventional dimensions, for example, and it is the way in which these conventions are assumed or unrecognized that gives iconic and indexical signs their apparent force, on the one hand, and room to be misunderstood, on the other. Indeed, the semiotic richness of the house as a model for living comes from the dense interplay of shifting sign relations. The recognition by missionaries of the powerful semeiotic effects of the house as both a model and a demonstration may be seen in their efforts to bring Tswana individuals into mission spaces as a step toward their conversion.

The field of interpretation that grows out of the sign-object relation was explored by Peirce in terms of a new, more developed sign that he called the “interpretant.” Peirce was quite clear that the interpretant is not restricted to cognitive entities.⁸ Altering Peirce’s terminology slightly, the interpretant may be glossed as an unexamined affect or feeling (an “affective interpretant”), a material or physical response (an “energetic interpretant”), or a cognition or habitual disposition (a “habitual interpretant”).⁹ Peirce viewed the tendency to habit as the most developed form of interpretant, itself embedded within affective and energetic responses (Peirce 1868b, 1868c). While Peirce’s notion of habit has much in common with Bourdieu’s description of habitus, it subsumes Bourdieu’s concept, in that habit for Peirce is not tied to the realm of the human. Instead, it refers to a tendency toward generality and mediation that may be found within the world as a whole. In this sense, the inclination of humans to establish habits of thought and action is simply one form of habit among many. Human semeiosis differs from other forms of biosemeiosis, as E. Valentine Daniel has pointed out, in that the human predilection for habits also includes the special habit of habit change, of reflecting on one’s dispositions and of altering them (Daniel 1996, 190–99). Peirce’s conception of habit, therefore, incorporates a more dynamic sense of an unfolding process than Bourdieu’s habitus, which can seem somewhat stable and resistant to change without the intervention of outside influence, as Jean Comaroff has discussed (1985, 5). Equally, it allows for some exploration of how reflexive, “discursive” activities might operate within “practical” consciousness, as I explore below.

The unexamined feeling of recognition (or affective interpretant) of a space and of the practices that are ordered within it depends on a habitual and embodied sense that emerges from past experience. This practical understanding of how to “go on” that Pierre Bourdieu (1977, 1990) and Anthony Giddens (1984) have explored is usually unchallenged and unreflexive until it is disrupted or something calls attention to it. Bourdieu has outlined how the organizing principles and dispositions that comprise habitus are learned through a lifetime’s experience and are produced from particular material conditions of existence (1977, 72). These dispositions are drawn upon creatively in order to act effectively, as people adjust to the semeiotic worlds they encounter. The embodied understandings of place that missionaries brought with them derived from a universe of undisputed and implicit principles (19) that seemed intuitively right, proper, and therefore moral. The house that David Livingstone grew up in, in Blantyre, Scotland, would have been experienced by him and accepted unreflexively for the most part. The missionary encounter, therefore, has the potential to dislocate both missionaries and the people they proselytize, as they discover new ways of living and acting within the world that do not conform with those they know. Peirce suggests that is only when some unexpected element of experience surprises us into acting or doing something differently that we actively notice elements of the world (e.g., CP 1.358–405, ca. 1890): for example, a curved wall where a corner might be expected, a body buried in the yard outside a house instead of in a graveyard.¹⁰ Here the “brute fact” of existence is brought into relation with the experiencing body and forces itself on it, eliciting unanticipated responses. Peirce said of this, “We are continually bumping up against hard fact. We expected one thing, or passively took it for granted, and had the image of it in our minds, but experience forces that idea into the background, and compels us to think quite differently” (CP 1.324, 1903). The knowledge of how to act, learned and drawn upon from childhood onward, is necessarily challenged through the encounter with foreign people and landscapes, but it is also adjusted continually in the course of day-to-day encounters with the dynamic habits of the material world, including those of other people.

As missionaries brought local people into relation with the architecture of the house, embodied energetic interpretants were elicited, which while reinscribing practices of movement and timekeeping for British occupants, also encouraged the formation of new habits of dwelling in those who visited. In this, the indexical and relational qualities of the house—the way in which a wall or a door encouraged or prevented movement for different people and at different times of day—were perhaps more powerfully felt than any iconic model of an invisible ideal. Equally, symbolic and linguistic signs would be more difficult to grasp without knowing the conventions under which they were formed, but in any modality the shock of the unexpected could bring these new sign relations into discursive awareness and encourage reflection and debate, or perhaps laughter and ridicule. Conflict could arise when the divergent habitus of missionary and potential convert created from different ways of living within the world “cause one group to experience as natural or reasonable practices or aspirations which another group finds unthinkable or scandalous” (Bourdieu 1977, 78). Practices could also be misrecognized with consequences for the success of the mission. David Chidester outlines the risk of translating Christian concepts into Setswana equivalents, observing how Tswana “comparativists” recognized a relation of similarity between reading the Bible and their work of divination, a move that was unacceptable to missionaries.¹¹ Chidester describes how in instances where Africans “found no equivalence between the religion of the mission and local tradition” missionary claims could provoke shocked laughter and advice to refrain from repeating such statements (Chidester 1996, 195). Missionary reactions to living and building practices could also elicit laughter. Denbow and Thebe quote LMS missionary Robert Moffat’s advice to women preparing to thatch a newly built house. “They ought to get their husbands to do that part of the work,” he suggested, observing that “this set them all into a roar of laughter” (2006, 95). Here “facts” that did not fit with expectations were dismissed or held aside rather than seriously evaluated. The missionary encounter was characterized by attempts to make sense of these semeiotic shocks. We see these attempts in letters and journals sent home by missionaries, but where might we look for similar attempts among local people? How might a Peircean perspective provide ways to explore how both representation and practice were enrolled in ongoing and reflexively monitored habitual semeiotic processes?

Tswana Projects

Despite the upheavals caused by the introduction of Christianity and colonialism, there were remarkable continuities in local forms and practices of dwelling. As John Comaroff has observed, both politically and architecturally Tswana communities have been “remarkably persistent in the face of dramatic external change” (appendix to Schapera 1984, 71; see also Comaroff and Comaroff 1991, 203–4). This continuity constitutes another strand in the history of hybrid forms that emerged from the encounter. The tenacity of local architectural techniques, and the persistence of the preexisting spatial layout of Tswana villages and towns, points to the stability of local ways of dwelling, but it also suggests some ways in which changes unfolded. This was not the missionaries’ hoped-for inculcation of Christian beliefs and practices and the habitus that went along with it. Instead, Christian beliefs and practices were incorporated in a variety of ways into local ways of dwelling, depending on the particular circumstances of the encounter.

The LMS mission at Phalatswe (Old Palapye), occupied from 1889 to 1902, shows how Christian belief and practices were incorporated architecturally at the Ngwato capital of chief Khama III. Khama converted to Christianity and, after his baptism, rejected many of the practices of his people such as the rainmaking ritual and the consumption of alcohol (Parsons 1998, 46–47). Despite Khama III’s apparently full conversion, Paul Landau argues that he took great care to control the influence of the LMS mission (Landau 1995, 34), drawing on Christianity and literacy in order to extend his authority into regions that previously had been relatively independent (xxvii). In this way, the structure and organization of the churches controlled by Khama were incorporated into political strategies of authentication and authority. As Reid et al. point out (1997, 386), the missionaries themselves and the buildings they constructed and furnished were a material resource to be drawn upon in the reinvigoration and continuing effectiveness of chiefly authority. Khama claimed transformation in building for himself and co-opted it at the heart of the motse. He directed the construction of new churches, often personally sponsoring and opening them himself (Landau 1995, 179). His architectural investment in the center of the town meant that the motse took on a more stable and permanent material presence. One effect of this was to make it more difficult for settlements to be relocated when deemed necessary, as outlined by Christopher Morton (2004). Morton explains that the growing need for fresh grazing land enticed the population away and placed pressure on Khama to move to a new site. Despite his investment in place, the chief risked losing his support if he failed to comply (353–54). The fractal-like form of the town and its political organization provided opportunities for wards to detach themselves and relocate under the control of their headman or in alliance with another. This made the risk of the town fragmenting a very real possibility to be negotiated by the kgosi. The decision to move the settlement from Shoshong to Phalatswe was financially draining for both Khama and the missionaries, as the stone church had to be taken down and moved over eighty miles to the new town. Cattle, sheep, and grain were sold in order to move the church, demonstrating how it had become a focal point for displays of surplus wealth and the maintenance of political ties and allegiance (Landau 1995, 179; Morton 2004, 351). The subsequent movement of the motse from Phalatswe to Serowe took place despite great disruption to the mission and the colonial administration. Serowe proved to be the final move, however. In this, the recalcitrant materiality of its stone buildings contributed to the motse’s sedimentation in place as much as the demands of missionaries and colonial authorities.

Morton observes that this tension between “competing material and social forces” is important to understanding the pressures on Khama and the way that settlement change played out (2004, 353). Morton’s careful attention to these different dimensions shows clearly how Europeans mischaracterized Tswana towns as sedentary settlements, as well as demonstrating the importance of attending to the material dimensions of nineteenth-century Tswana politics. However, in placing the material resistance of the town on one side of the equation and the social pressure to maintain supporters on the other, this analysis plays down the imbricated nature of these different forces. Below, I consider how a semeiotic approach might destabilize these categories of the social and the material and in the process capture something of what Bruno Latour (e.g., 1993, 1999, 2005) has framed as the creation of hybrid networks composed of heterogeneous elements.

The sensuous model that the town presented was a complicated set of sign relations with many dimensions. As people continued to live and work in the wards around Khama’s central ward, the activities of daily life worked as habitual interpretants that acknowledged his innovatory and creative chiefly power. These practices recognized Khama’s primary role at the center, which he made visible both indexically and iconically through his innovations in building. However, there was a tension between the indexical sign of Khama’s power (through his investment in place) and the iconic sign of his position (through his location in the center of his people). The pressure on Khama to move the town and the difficulties that this posed show how indexical signs can be awkward to manipulate because of the ways in which they are associated with or linked to their objects. In affirming the power and importance of the center through building elaboration, Khama also allowed himself less semeiotic agency over where and when such an indexical sign could be deployed (see Kockelman 2007, 380).

However, if indexically grounded signs allow less control over where and how they are deployed, then the ability to manipulate them despite these constraints itself becomes an indexical sign of power and control. Although the investment in place made through the building of churches and European buildings meant that the movement of towns became more difficult, it also meant that the maintenance of the habit of moving the town, despite the material difficulties of doing so, acted as a powerful interpretant of Khama’s spiritual force and political authority. Large hilltop towns and abandoned stone ruins of older settlements were a striking feature of the southern African landscape, dating back to the seventeenth century (Huffman 1986; Hall 1995), suggesting that the tension between investment in place and the need to relocate had a long history. The relatively frequent movement of metse in the precolonial period reinvigorated the spiritual authority and efficacy of the chief, in the process reconceptualizing and remaking the landscape around the new town and locus of power (Lane 2004, 289). The successful move and the reinscription of the model of the town in a new location co-opted the whole people into the sign relations specified by the chief, in a communal energetic interpretant of his place and of Christianity at the center of the town and of social life. Equally, if a ward were to detach itself to establish a new settlement elsewhere, this would be a demonstration (an energetic interpretant) of the kgosi’s diminished power and an attempt to establish a pattern of habitual interpretants around a new center. Here the quasi-fractal nature of Tswana townscapes beautifully illustrates Peirce’s conception of semeiosis, which is itself often described as fractal-like (e.g., Brent 1998, 333). As each interpretant buds off from the previous sign-object relationship, it in turn becomes the starting point for a potential new sign relation to form. Thus, the detaching of a ward and the constitution of a new town may be seen not only in terms of signifying practices but as itself a semeiotic process that incorporates people and architecture in one dense but evolving semeiotic cluster or bundle (cf. Keane 2005).

In interpreting Khama’s position through the movement of the entire settlement, the people and town re-created an established habitual interpretant and anticipated the future continuation of these relations. Khama did not allow Europeans to build freely at his capital and carefully controlled the construction of missionary buildings, churches, and other European-sponsored structures (Reid et al. 1997, 385). His son Sekgoma also prohibited the adoption of European-style housing among his people (385), showing that the potential to transform the ways in which houses were built was not equally available to all members of the population and depended to a large degree on the precedent set by the kgosi. Paul Lane (1999, 162) observes that, consistent with this concern to restrict new forms of dwelling, there is little archaeological evidence of change in house construction at Phalatswe (although house traces are not well preserved at this site). This suggests that there were few opportunities for innovations in house form by the population at Phalatswe (Reid et al. 1997, 381). At least at first, the missionaries had less success in changing the ways in which people built their houses among Khama’s people, despite his commitment to their desired form of Christianity. Khama seems to have wanted to fold Christianity into his central position as king of the new “ecclesiastic state” that Landau has delineated. The material claiming by Khama of European forms of dwelling affirmed the continuing relevance of the motse and its wards and houses as a sensuous model of social relations.

With the move to Serowe, however, Khama seems to have changed his semeiotic strategy toward the European mission. At this point the mission buildings were spatially marginalized and located well away from the chief’s ward. Even the “native” church was set apart from the town at a distance of some kilometers (Reid et al. 1997, 383). By this date, circumstances had changed and the political autonomy of the Ngwato was under attack; the habitual interpretants that were emerging from both mission and colonial interventions could no longer be managed and appropriated so easily by Khama. The attempt to displace the signs of mission outside the town was in itself a sign of his weakened semeiotic control of the encounter and his effort to reclaim the central and sacred ground of the motse for himself and his people.¹²

The actions of Khama III and his people at Phalatswe can be contrasted with the ways in which European building techniques were drawn upon at the Kwena towns of Kolobeng and Ntsweng, in present-day Botswana. In 1847, apparently encouraged by David Livingstone, kgosi Sechele I decided to move his capital from Chonwane to Kolobeng (occupied from 1847 to 1852). The mission station moved at the same time. Despite this, it was not incorporated into the new motse. Instead, the mission buildings were kept separate and distinct, with church, mission house, and graveyard located outside the Kwena settlement (Lane 1999, 156). This maintained the spatial and moral order of the motse, while allowing Livingstone to lay out his station according to spatial principles that he recognized as appropriate and morally correct. Yet, while maintaining the spatial separation between town and mission at Kolobeng, Sechele I also drew on the use of space promoted by the missionaries in building a rectangular house for himself. At a later capital, Ntsweng (occupied from 1865 to 1937), the mission station was sited even further away, about two miles outside the settlement, while Sechele’s new house retained the new rectangular building form and was furnished with expensive European goods “dominated by a large crystal chandelier” (Volz 2001, 20). This strategy allowed the kgosi more room to seize the semeiotic terms on which European features were incorporated into the motse. An example is provided by the construction on Sechele’s initiative of a chapel close to his compound when he moved his capital to Ntsweng, even before a missionary had been installed at the town (Lane 1999, 158).

Sechele incorporated Christian belief and practices selectively and strategically, integrating Christianity with Tswana practices (Gulbrandsen 1993; Volz 2001). Indeed, Sechele’s idiosyncratic version of Christianity led to his later alienation from the LMS. Despite this, Stephen Volz (2001, 2–3) argues cogently that Sechele’s conversion was genuine and heartfelt, “motivated by spiritual and moral concerns.” Volz suggests that to place too much emphasis on Sechele’s political motivations is to miss other, perhaps less utilitarian, possibilities and to elide the ways in which Christianity may have worked against his interests as ruler. Certainly, to consider Sechele’s politics narrowly as a means to the end of sustaining rule seems an impoverished way to think about what took place. However, insofar as politics could not be divorced from a broader field of cultural practice, then it perhaps serves as an opening to understand some of what went on. We can see how Sechele worked to control the terms on which European semiotic forms and processes were incorporated into Tswana ways of dwelling. Like Khama, his incorporation of European architectural forms in the heart of the capital acted as a sign of his ability to take advantage of the semeiotic shocks that had been delivered by the missionaries and others. In this way the challenges posed to local habitus were themselves seized and reworked to reconstruct and affirm his political and spiritual influence.

John and Jean Comaroff argue that Tswana forms of personhood at the time of the colonial encounter can be understood in terms of an ongoing process of becoming. Personhood was made and known through relationships and through practical activities in which the work of building and making was valorized (John Comaroff and Jean Comaroff 2001, 268–73). Tswana understood the self to be “ranged over sociophysical space-time occupied by the sum total of its relations, presences, enterprises—anything that acted on its traces might affect it for good or ill” (275). This can be considered in relation to changes in domestic space. Archaeological research at Ntsweng has shown that despite changes in the residences of the elite, the majority of the population maintained an attachment to the circular-plan houses and compounds that were the basic building block of Tswana villages (Reid et al. 1997, 376–79). As at Phalatswe, the organization of the town affirmed the relationships between people and remade them daily. However, in contrast to Phalatswe, the archaeological evidence shows that some modifications were made to the usual house form at Ntsweng, with the incorporation of linear elements into the walls of yards outside houses and the inclusion of some straight-wall construction in house walls themselves, producing semicircular rather than circular houses. Just as the sensuous model of the town worked as an interpretant of chiefly authority, so the houses themselves may be understood as interpretants (and hence signs). Paul Kockelman’s work on semiotic agency (2007) suggests a consideration of how in building a house people interpreted a whole set of relationships—whether between the mud and the particular purchase and function that it offered or of particular gender relations and relations of hierarchy and subjugation. To change and experiment with house form was to push and test the relationships within which people were situated and to create new material interpretants of these changing relations.

The alterations made to the design of the house indicate an incorporation of rectilinear space into local architecture that appears to have taken place very much on local terms, rather than the transformation of the space of the house along European lines that missionaries hoped for (cf. Hall 1997). The introduction of straightened walls in Kwena houses and yards at Ntsweng suggests an experimental and reflexive demeanor among those who were building. The changes in house construction indicate that some people (perhaps particularly women) were actively exploring the potential that rectilinear space offered and, in doing so, imagining different possibilities for the future. Local people may have been trying on new habits promoted by missionaries, but only insofar as they worked within a broader field of recognition growing from preexisting practices. As Landau has observed, the innovations introduced by missionaries and others had to resonate for them to be picked up and reproduced locally (1995, xxv). It is clear that, with these changes to domestic space, the embodied sense of how to move within and around the house and compound was maintained, and the shock of the new was incorporated into existing habitus.

Peirce argued that thought (and hence belief) takes place in and through signs, and these signs are inescapably embedded in the world of experience: thought is not a purely internal mental process (e.g., Peirce 1868b). From this perspective, the self (to use a term admittedly freighted with individualist and rationalist associations) may be viewed as a shared and ongoing semeiotic process that emerges as a set of habitual relationships with the experienced world (Peirce 1868c). As Milton Singer explained, for Peirce the self is distributed, social, and public, its locus “found in the sign processes themselves” (1984, 56–57) rather than in any purely mental construct. These ongoing sign relationships shared in, confirmed, and reinscribed practical habits through which people ordered and made sense of the world. In this view, the house is as much a part of the self as any other more usually recognized bodily extensions such as clothing or jewelry (cf. Csikszentmihalyi and Rochberg-Halton 1981, 15–16). In building their houses in wards that remade and pointed to the center, people affirmed the position of the kgosi, both in the work of building and in the homesteads they created. The appearance of linear elements within the walls of houses and yards suggests a form of recognition of Sechele’s innovations as meaningful for people locally. This is interesting in that it contrasts with the apparent failure of Christianity to spread among his people. Stephen Volz notes that, in ten years’ work among the Tswana, Sechele was the only convert made by David Livingstone (Volz 2001, 1) and that many of Sechele’s headmen were made uncomfortable by his European-style house and furnishings (40). Yet some elements seem to have been picked up by the broader population.

At Mabotse, in the former Transvaal, South Africa, Simon Hall has examined similar changes during 1872–80 (1997). He argues that the incorporation of linear elements into the space of the village drew just as much on preexisting local valuations of these elements as it did on European introductions (214). Hall suggests that these changes be understood in terms of the increasing appropriation and control of women’s labor by Tswana men. He observes that the incorporation of straight-line walls tended to be used in the creation of rectangular or subrectangular semipublic enclosures used by men. Houses, in contrast, constructed by and associated with women, remained circular, although, as at Ntsweng, the walls around house yards at Mabotse became more angular. Hall argues that these shifts seem to have followed a trend toward increasing gender hierarchy and the circumscription of choices and possibilities for women within Tswana communities. This can be seen particularly in shifts in agricultural practice during the same period. Margaret Kinsman emphasizes the “web of mutual dependence” that was woven “between husband and wife” (1983, 42) through their respective roles in food production and distribution. A woman gained access to the agricultural land that she tended through her father or husband, and although she could be given cattle upon marriage, the animals were subject to control by her husband (42–43). The introduction of the plow by Methodist missionaries to the Southern Tswana had dramatic effects on women’s control over their productive activities. Jean and John Comaroff observe that contact with the highly valued cattle that were used with the plow was restricted for women and that, as a result, their role in planting and distributing grain was diminished (1986, 13). Instead women were left to care for and harvest the crops—work that was arduous but of lesser cultural value. Whereas previously, using hoes to till the fields on the edges of the settlement, women had a good deal of control over their labor and the food they produced, men moved into the sphere of plow cultivation and took more control of the distribution of crops (Okihiro 2000, 91–98). The same transition from hoe to plow agriculture took place at Shoshong between the 1850s and the 1880s (Fosbrooke 1971, 182–84) and in Khama’s kingdom. Paul Landau notes that around Phalatswe women’s work was increased at harvest time as the result of the plow because larger areas could be brought into production (1995, 74, 104).

These renegotiations in the relations of production are consistent with Hall’s argument that at Mabotse the kgosi and the male elite were working to maintain gender distinctions, while consolidating and extending their authority and power. Such changes were made on terms that were recognizable and meaningful to locals, but innovations brought by missionaries were also assessed for their effect on local practice. In the context of the longer-term shifts in male power outlined by Hall, these changes in agriculture may be seen not simply as the unexpected consequences of the encounter. Rather, they suggest ways in which gender relations were being reworked, while appearing to maintain habits of practice and association in the face of new farming techniques. Along these lines, Paul Lane suggests that there is a visible shift of emphasis in architectural organization at many sites in the region from the eighteenth century, with an intensified concern with boundedness and control of access to settlement and houses (Lane 1998, 194). He argues that the historic period sees an increased differentiation of domestic space, transforming the house and its yards into “a contested domain in which men and women negotiated for position” (200). The construction of angular yards outside houses at Mabotse seems to be part of these longer-term negotiations taking place over space and over the different domains of men and women.

The reordering of the domestic space of women at Ntsweng took place at the same time that the matrilocal arrangement of grain storage was under threat. However, in contrast to Hall’s interpretation of changes to the yards of houses at Mabotse, which he suggests is a reworking of women’s space under pressure from male relatives, I read the traces at Ntsweng a little differently. Here not only the semipublic and negotiated space of the yard but also the houses themselves were subject to changes. If a woman’s sense of self and of her relationships with her extended family was embedded and remade through the space of the house and the associated yard, then the appearance of linear elements can also be understood as an effort by women to make sense of the semeiotic shocks that were assaulting them from all sides. This was an affective and energetic trying on of new architectural elements but within the familiar space of the house, a search to establish recognizable and sustainable habits of practice in the context of dramatic changes to women’s social and productive roles. In this context, the experimentation with linear elements within the walls of yards and houses at Ntsweng seems to assert an iconic linkage between homesteads and the central position of the chief. Looking at the archaeological remains today, they evoke a feeling of connection and integration within the town, which one can infer was also felt by those building and decorating the walls of the house and yard. The reconfiguring of space suggests one way in which women worked creatively to assert control over their sphere of influence and to maintain their sense of self within rapidly changing conditions and massive disruptions to their traditional activities and status.

Conclusion

In this review of some of the literature on mission among Setswana-speaking people, I have sought to explore what Peirce can offer to thinking about the relationship between practice and representation. Sometimes it is hard to see how to integrate archaeological accounts that focus on a material palimpsest of routinized practices, and documentary accounts that more precisely delineate representational changes in belief wrought by missionaries. The work of Jean and John Comaroff has provided powerful insights into the ways in which changes in belief are embedded in efforts to change living practices, opening up possibilities for productive interdisciplinary exchange. The integration of archaeological evidence with historical documentation provides great potential for future anthropological research, if we can find ways to bring the different orders of evidence into conversation (Reid and Lane 2004, 18–19). A Peircean semeiotic approach provides an avenue for productive interdisciplinary rapprochement. In its redistribution of semeiosis across the boundaries of “mind” and “material,” it creates alternate ways to bring representation and material traces into the same analytical frame. In the context of nineteenth-century Africa, where the signs through which the encounter between Europeans and Tswana peoples was negotiated were constantly under review, Peirce’s complicated model of semeiosis captures some of the complex and shifting nature of changes in belief and practice.

Peirce’s semeiotic draws no a priori distinction between habits of mind and of practice, prompting the question of whether it is only through language that semeiosis can take place reflexively. Looking at the experimental reconfiguring of houses at Ntsweng, Pierce’s semeiotic provides a way to cut across distinctions between (for example) practical and discursive consciousness (Giddens 1984), allowing us to imagine a “discursive” consciousness at the level of practice. The example of apprenticeship suggests how this might work. In learning to build a house, the learning process is reflexive and recursive, and while language certainly intervenes, much of the semeiotic process is affective and energetic, leading to an embodied sense of how to shape and mold the house form that cannot be completely articulated in language but which is certainly reflexively monitored and understood. Practical experimentation with materials and forms of practice provides an example of how new habits might be tried on for size and then either rejected or incorporated into the world of habitus. This creates ways to think about stability that are reducible neither to an unchanging mental map of the world that persists over hundreds of years nor to the resistance of the material world. Instead, Peirce allows us to articulate this continuity in terms of a semeiotic project to maintain the relational sense of self, a sense that is embedded in place and in language—“in thought,” as Peirce might say.

In the case of the missions to the Tswana, the archaeological evidence often seems to have the effect of highlighting continuity in living practices, while historical accounts based on mission documentation emphasize the dislocating rupture prompted by the changes and social upheavals during and after the encounter (Lane 2001, 158). I hope to have shown here how a semeiotic approach can point to how architectural changes were caught up in local projects, particularly the negotiation of changing gender relations, cutting across the interests of missionaries and their narratives of the encounter. Kathryn Fewster notes that the picture of Tswana communities that emerges from twentieth-century ethnographies is one of “a people who are constantly reinventing their identities out of a mixture of ideas: those offered by the increasing influence of Westernism in the country, and the rapidly fragmenting ‘traditional’ rules of their parents and grandparents” (2006, 70). This experimental and dynamic reinvention is one with a long history. Archaeological fieldwork shows that the nineteenth-century Tswana metse described by missionaries had emerged from a landscape of more dispersed settlement sometime in the early eighteenth century (Huffman 1986; Hall 1995; see also Parsons 1995). Attention to this history provides a context for understanding how changes played out later in the nineteenth century (cf. Kinsman 1983, 40). As Jeffrey Fleisher (2004) has argued for the site of Kilwa Kisiwani on the southern coast of present-day Tanzania, by looking to the deeper past another perspective is given on the local politics and regional history at the time when Europeans arrived. In this way a better sense of local projects and intentionality can be worked out.

Many authors have demonstrated how in a range of African societies the organization of architecture and village and the practices carried out around them create a model of the social group, including differentiated relations of gender and political status (e.g., Bohannan 1958; Huffman 1986; Moore 1986; Donham 1999). The dispositions that comprise habitus, and that people draw on creatively in order to act effectively within the world, are learned through a lifetime’s experience within these material and social conditions, as Bourdieu explored. Shifting to a Peircean perspective, we can make slightly different claim: beliefs—whether religious beliefs about God or practical beliefs about gender, status, or anything else—are established sign relationships that are constituted from crosscutting constellations of landscapes, people, things, feelings, actions, and cognitions. As Peirce put it, “The feeling of believing is a more or less sure indication of there being established in our nature some habit” that will guide action (1877, 5). The semeiotic concept of habit thus inhabits the same territory as the Comaroffs’ discussion of signifying practice, but it provides a different lens for thinking about historical change that brings practice fully into the realm of semeiosis. It is through these habitual sign relations that the model of the world is constituted, challenged, and can change and grow. In attending closely to the visible material dimensions of these relationships, it is possible to start to explore changing perceptions of self and other, landscape and domestic space, in ways that complement accounts based on written texts and oral histories.