Laughing to Death: Joking as Support amid Stigma for Zulu-speaking South Africans Living with HIV

Abstract

Moving beyond the truism that laughter is the best medicine, this article integrates multiple linguistic anthropological models of humor to theorize why joking is so often used to address issues at or beyond the boundaries of what is considered acceptable to talk about in most other contexts. The article analyzes joking about HIV among members of a Zulu gospel choir who are living with HIV/AIDS in South Africa, identifying the precise linguistic features that choir members used in dealing with stigma by means of joking interactions. Three properties of joking about HIV are discussed, and it is suggested that these properties may operate cross-culturally in other genres of humor in which individuals approach topics that are shameful, embarrassing, upsetting or taboo. Utilizing anthropological perspectives on genre, poetics, and play, this article discusses linguistic properties of joking that make it ideal for constructing and contesting support amid stigma. [humor, stigma, verbal art, joking, HIV/AIDS in South Africa]

Humor is a powerful resource for social commentary and transformation, in part because it is often considered to be innocuous—“mere” entertainment. Anthropologists have long recognized the utility of humor and joking in regulating social relationships (e.g., Radcliffe-Brown 1940; Fortes 1953). More recent scholarship recognizes flux and play in the linguistic construction of these seemingly rigid role relationships while, yet, there is an underlying continuity with earlier work in the idea that insults, disrespect, and humor in language contribute to the maintenance and contestation of social organization (e.g., Labov 1972; Schieffelin 1986; Irvine 1996; see Stasch 2002 for a recent take on the joking relationship). Relatively little attention is directed specifically toward joking as a distinct object of linguistic anthropological analysis (but see Basso 1979; Beeman 1981; Sherzer 1985)—which is surprising, considering the often-noted impact of joking on social and psychological processes (e.g., Bergson 1911; Freud 1963; Bateson 2000; Goldstein 2003; Gamliel 2004). Based on fieldwork in 2008 with a Zulu gospel choir in which all group members are living with HIV/AIDS amid stigma, this article expands upon previous anthropological scholarship on joking. The analysis below draws from previous models of play, performance, and humor to examine the communicative properties that make joking, and humor more generally, ideal for addressing issues (such as HIV stigma) at or beyond the boundaries of what is considered acceptable to talk about in most other contexts.

In 2008, somewhere between thirty and forty percent of people living in and around the city of Durban, South Africa were¹ living with HIV/AIDS. Despite this overwhelming infection rate, intense HIV stigma and hlonipha, a traditional respect vocabulary of avoidance, led to gaps in discourse about HIV (Finlayson 2002; Wood and Lambert 2008). Examples of these gaps included the use of alternative terms (or euphemisms), opaque reference to people living with HIV, avoidance of discussion about how infection occurred, and joking of a kind that often demeaned those with HIV. This kind of joking was directed at people living with HIV, with the implicit assumption that the joke-teller neither had HIV/AIDS nor accepted those who did. At the same time, few South Africans were willing to label themselves or other interlocutors as being HIV-positive. Many South Africans did not get tested for HIV, and some of those who were tested would not share their HIV status with even their closest friends and family members. The choir with which I conducted research began as an HIV support group predicated upon group members' knowing each other's HIV-positive status. Because of the choir's support-group function, talk about one's HIV infection during rehearsals was less likely to threaten a choir member's “face” than such talk in other, everyday social situations (Goffman 1959; de Kadt 1998). In choir interactions, joking and singing were central types of performance that addressed group members' HIV statuses. In particular, choir members performed jokes that pointed to oneself as HIV-positive, in direct contrast with the stigmatizing jokes heard around Durban. From the choir members' own perspectives, this joking was “just something that they did with each other.” Within the broader context of HIV stigma in South Africa, however, the choir's jokes were notable and might be considered shocking by outsiders. Understanding how choir members' joking constituted support is a matter of not only theoretical but also practical importance. Social workers and activists' recognition of joking and humor as facets of successful support activities could, potentially, improve future HIV/AIDS treatment and support efforts in South Africa and beyond.

This article analyzes precisely how choir members' joking about HIV confronted dominant ideas about the disease and considers why humor affords human engagement with topics that otherwise would be at or beyond the boundaries of acceptable conversation in a given cultural context. Integrating previous research on ritual insult, joking, play, and performance, my analysis indicates that three properties of the choir's jokes made them conducive to facing stigmatization: (1) joking relied on shared presupposition that took the form of a “secondary text” (Basso 1979:41), (2) joking was improvisationally co-constructed (Sawyer 1996; Duranti and Black 2011), and (3) joking occurred within a play/performance frame (Bateson 2000:180; Goffman 1974) and a speech genre. As I discuss below, choir members' joking about HIV included a presupposition of acceptance, and their co-constructed humor within this speech genre effectively expanded rather than constrained the possibilities for talk about HIV, becoming a key type of performance through which “support” was constituted and community was reinforced. Such social functions of joking have been noted in other cultural contexts in which marginalized communities are maintained in oppositional dialogue with dominant ones (e.g., Basso 1979; Goldstein 2003). I therefore suggest that these three properties might be found across cultural contexts in other forms of joking and humor that confront topics of avoidance, fear or embarrassment. Further, this research supports a model of joking as neither uniformly hegemonic nor transformational, but instead as shifting and ambiguous—and there lies the power as well as the limitation of joking (Goldstein 2003:6–7; see also Samuels 2004:8).

Secondary Texts, Play and Performance

Before turning to an analysis of the above three properties of joking about HIV, it is useful to unravel a theoretical knot that has developed in linguistic anthropological understandings of joking, teasing, and humor. This knot involves the concepts of secondary texts, play and performance. In one of the few ethnographic monographs to discuss the linguistic and communicative dynamics of joking, Keith Basso (1979) analyzes Western Apache joking interpretations of the Whiteman. In “ordinary” (Sacks 1984) Western Apache social life, personal questions, pats on the back, and other prototypically “Anglo” interactional patterns are seen as invasive and offensive. In joking performances, though, these patterns are permitted in the service of making fun of non-Apache interactional styles and strengthening relationships between Western Apache interlocutors who are already quite close. Basso labels the “ordinary” interpretation of offensive and intrusive actions a “primary text” and calls the joking interpretation of these same actions a “secondary text” (Basso 1979:41). For the participants in the joking interaction, the secondary text “is intended to be understood as a facsimile or transcripted copy of the primary text on which it is patterned” (Basso 1979:41). If the joking fails and participants see a joker's prototypical Anglo enactment as primary text—that is, as a Western Apache interlocutor's nonjoking discourse—arguments and fights may ensue.

A potentially confusing aspect of Basso's analysis is the relationship between secondary texts and play/performance frames. In developing his own perspective, Basso uses Erving Goffman's Frame Analysis, specifically drawing upon the notion of “primary frameworks” providing guidance for interpreting social action (Goffman 1974:21–39; Basso 1979:102). For instance, the same words uttered in a church or on the street corner will be understood very differently because those words are spoken within different primary frameworks guiding their interpretation. For his part, Goffman relies on Gregory Bateson's theory of play (Goffman 1974:7). Bateson's concept is inspired by his observations of play across species (notably of dogs and monkeys play-fighting). He explains that the establishment of a play frame involves an interpretive paradox in which “the playful nip denotes the bite, but it does not denote what would be denoted by the bite” (Bateson 2000:180). Though rooted in the same framing process that occurs in the play of other animals, human play and performance can be much more complex. What Goffman demonstrates in great detail is that there are innumerable ways that humans key discourse as being within specific frames, including but not limited to play, such that participants are directed to look for alternative forms of interpretation of performed actions (Goffman 1974:44–45). Why humans engage in such activities has been a perennial question. Two explanations relevant to this article are that performance allows individuals to “practice” emotions within the conceptually safe space of framed discourse (Beeman 2010), and, similarly, that play opens up space for critique and subversion within the constraints of particular social situations (Limón 1989:478).

Often, speech play is keyed by formulaic phrases that conceptually bracket talk as play or verbal art (Kochman 1983; see also Bauman 1977:21), but all sorts of conventionalized statements, or even a gesture, facial expression, the prosody of one's talk, or the sequential organization of one's turn at talk can mark a moment in the interaction as shifting into the realm of play or performance (Goffman 1974:44). While there are many different frames that challenge interlocutors to understand alternative interpretations of talk (e.g., as theater, as a sermon, as a lecture), the invocation of a “secondary text” is distinctive. A joke, as the term is normally used, is dependent on the secondary text. As a joke proceeds, audience members may or may not be taken in and believe that the performance is serious, involving a primary text (some jokes require this belief while others do not). The “punch line” or humorous ending to a joke depends on the (often dramatic and sudden) recognition by audience members that they share in common the knowledge needed to decode the relationship between the performed secondary text and a presupposed primary text (Bateson 1952). A review of research on other activities described as ritual insult and teasing reveals that participants do not utilize secondary texts in these cases (e.g., Labov 1972; Kochman 1983; Schieffelin 1986; Irvine 1996). As will be shown, it is precisely the use of secondary texts in joking about HIV that allowed participants to invoke rich worlds of presupposition and implicature without explicit discussion and to mediate social differentiation internal to the choir.

Singing about HIV in South Africa

The research presented here is part of a larger ethnographic project centered on the lives of Zulu-speakers living with HIV in and around Durban, South Africa. In 2008, when the bulk of the research was conducted, Durban was a large metropolis—one of the biggest port cities in southern Africa—with a concentrated downtown area surrounded by a checkerboard of wealthy suburbs, impoverished informal settlements (also known as squatter camps), and townships (also known as “locations”) of former apartheid notoriety. As in other parts of the world, structural inequality and HIV infection were closely aligned (Farmer 1999). This structural inequality was the result of political developments prior to independence in 1994. The increasing severity of indirect rule (also known as “decentralized despotism”) imposed by colonial Dutch and British authorities was set into law by a minority-rule government organized by former European colonists in southern Africa beginning in the early 1900s (Mamdani 1996). Apartheid, as this system became known, relied on ethnic distinctions among southern Africans that were a result, in part, of the Zulu army's violent conquest of much of the eastern part of southern Africa throughout the nineteenth century (Thompson 2000:80–87). Apartheid segregated and gave different human/civil rights to South Africans on the basis of four racial groupings: black, coloured, Indian, and white. The apartheid government then further divided black South Africans according to ethnic groupings codified from a relatively fluid precolonial organization (Comaroff and Comaroff 1992). The specifics of ethnicity and tradition, though constructed from precolonial forms, were colonial inventions throughout Africa (Apter 2007:136). Despite this, the continuing relevance of ethnicity, race, and structural inequality in the post-apartheid South Africa of 2008 cannot be overstated, although for a small proportion of new “black elite,” class had in part transcended ethnicity and even race. Structural inequality was a serious constraint for most of the members of the gospel choir with which I conducted research; yet choir members (all black Zulu-speaking South Africans) were able to partially and temporarily overcome that inequality through the recognition that came as a result of their affiliation with an American Christian aid organization. In 2008, seven of fifteen choir members worked for AIDS treatment and prevention programs, hired because of their unique combination of local cultural knowledge and scientific medical model expertise, along with their willingness to speak more freely than most about their HIV statuses. Choir members felt ambivalent about their connections to international aid. The group received medical treatment (free in the choir's early years, low cost in 2008). Also, in my analysis, the choir's association with powerful, high-status foreigners was a counterpoint to Zulu linguistic patterns and conceptual frameworks that marked HIV as foreign in a negative or stigmatizing sense (a topic for future research). At the same time, however, choir members' participation in international aid was limited by structural inequality. The group had solicited donations through choral performance for international audiences, yet did not have control over or direct access to those funds, leading to fears and claims of exploitation.

This study draws on data collected over nine months of ethnographic research in and around Durban, South Africa, in 2008, as well as two months of pilot research in Durban and Pietermaritzburg, South Africa, in 2005. This was a collaborative accomplishment among choir members and me. One of the effects of apartheid and structural inequalities was a skyrocketing violent crime rate, and my ability to carry out even nine months of extended research, spending most of my time in the townships, informal settlements, and rural areas, depended on choir members' help and protection. (As one choir member told me, “Everybody in this neighborhood knows me, Steve. As long as you are with me, they cannot touch you.”) During these nine months, I engaged in extensive participant observation with choir members, other Zulu-speakers living with HIV, and South African university students; I also recorded 50 hours of video of choir rehearsals and audio of interviews with research participants. Working with members of stigmatized communities presents methodological difficulties because many are afraid to be recorded or documented. At the beginning of my project, I was told by South African researchers that I would never get the data I needed, that choir members would not talk to me about their experiences with HIV. The choir members' previous work with aid groups and international performances was one of the key factors that led them to allow me to engage in participant observation; within this context, video-recording was also a collaborative accomplishment.

The choir was a unique community that crosscut the sprawling townships and informal settlements of the Durban area. While several choir members had been neighborhood friends since childhood, the group was drawn together by their common participation in an HIV clinic support group in the late 1990s and early 2000s. As a result of collaboration with a United States-based Christian aid group, choir members had been among the first in South Africa to receive medical treatment and American-style support-group therapy. This was at a time when many doctors not equipped with antiretroviral medication felt that HIV cases were hopeless and turned patients away (Oppenheimer and Bayer 2007). Many of the choir members had been trained and were working as professional HIV counselors. These individuals helped others on a daily basis to navigate the trauma of an HIV-positive diagnosis.

Support Amid Stigma: the Importance of “Face”

Choir members, like others living with HIV/AIDS in South Africa, had to deal with the threat of stigmatization on a daily basis. From an interactional perspective, stigmatization is about the management of face. Goffman (1963) explains that an individual who has a stigmatized aspect (e.g., physical, social) can at any moment be “discreditable” or “discredited.” HIV is a virus that sometimes does not produce noticeable signs of illness, especially in early stages of infection and when individuals seek and receive adequate medical care. When stigma is focused on a potentially invisible feature such as HIV infection, infected individuals can avoid losing face, and avoid becoming discredited, by effectively managing a given interaction (Goffman 1967). The process of managing interaction amid stigma is a complex one involving indirectness, ambiguity, and the ability to communicate multiple messages through single utterances to conversational participants who comprise a heterogeneous audience (Bogen 1987).

In the Zulu context, face has traditionally been about social relationships in a hierarchical system rather than about individual positive-face wants (cf. Brown and Levinson 1978). Zulu cultural conventions of face are shaped by a system of hierarchical organization, previously also an avoidance register, called hlonipha, often translated as ‘respect’ (Irvine and Gal 2000). Respect is expected on the part of subordinates, whose status is determined, in large part, by age and gender (de Kadt 1998:182). This cultural understanding of face has noticeable implications for the management of stigma. One clear example has been the fact that many HIV-positive individuals do not reveal their HIV status to parents. This nondisclosure is in part because it would be disrespectful both to talk about a sex-related topic and also to implicate elders in such a stigmatized matter (Wood and Lambert 2008). In fact, the choir member in whose garage the group rehearsed had not disclosed his HIV-positive status to his own parents, who lived at the home.

The choir rehearsed in this garage two or three times per month, usually on Sunday afternoons. This secluded location, along with the group's dual functioning as both support group and gospel choir, gave choir members a measure of anonymity even within their local community, which meant a break from the face-work constraints of everyday life due to societal stigmatization of HIV. A common trope among group members was that the choir was a “second family,” one in which group members (who were all within a ten-year age range) considered each other to be brothers and sisters and called each other bhuti‘brother’ and sisi‘sister’. This relationship had implications for joking. Although it did not seem that choir members were invoking joking relationships as such, their understanding of each other as family and as age-mates allowed the group to bypass many of the most constraining aspects of the hlonipha respect system. Furthermore, the activity of choir rehearsal was, by default, playful; it required special effort to shift out of (rather than into) a play frame.

Joking Property #1: Shared (Unspoken) Presupposition

While this article analyzes joking about HIV among choir members specifically, the choir's joking occurred within a larger communicative context in which joking commonly addressed topics of danger and death. It was not unusual for choir members and many other South Africans as well to adopt a cavalier attitude toward danger and death. Among nations that kept accurate crime statistics, South Africa in 2008 was one of the most violent, crime-ridden places in the world (Lemanski 2004; Comaroff and Comaroff 2004). In addition to the dangers of being stabbed or shot for a cell phone while in downtown Durban, it was a well-recognized fact that driving or riding in a minibus taxi in South Africa was a hazardous, life-threatening event. Research participants explained that a cavalier attitude toward death, including joking about dying, was a way of coping, of accepting the possibility of one's death at any moment (cf. Gamliel 2004). A secondary pattern related to joking about danger and death but also central in linguistic processes of hlonipha and stigmatization was nonchoir joking that was directed at people living with HIV (rather than joking with them). For example, one phrase used to describe a person with HIV was uyatreka (“he/she is trekking” or “he/she is trey”). In this phrase, trek is borrowed from the Afrikaans term for traveling over large distances, as in “he gets around.”Trey, slang for three, references amagama amathathu‘a three-letter word’, HIV. A related way to refer to HIV is “zed three” (Z3), a BMW sports car that gets around quickly and is flashy. Such terms index models of promiscuity that denigrate people living with HIV.

Choir members' joking about HIV was a transformation of such denigrating joking and an extension of the broader use of joking to cope with danger, but the choir's jokes were also a way to recognize one's humanity despite being HIV-positive. Viewing oneself as ordinary among group members was synonymous with “accepting” one's HIV-positive status. This was seen as a central aspect of leading a productive life after becoming diagnosed with HIV (Marzano 2009). A key outcome of stigmatization is the classification of a person or people as not normal, as outside the boundaries of conceptual ordinariness (as opposed to the ordinariness of a statistical majority) (Sacks 1984). Several choir members told me that they weren't worried about HIV, that they would probably be killed some other way—in a car accident or a mugging, for instance. By emphasizing the other ways that they could die, choir members contradicted stigmatization and placed themselves back into the purview of what was considered normal (see Bucholtz and Hall 2004 on markedness and normality in identity maintenance). As one choir member explained to me during an interview, “I always tell myself that I don't care about HIV. It won't kill me anyway. Something else will kill me . . . I always take it as a joke. I don't take it serious” (Interview 08-23-2008). While this may seem to be a lack of engagement with a diagnosis of terminal illness, it was in fact a way of reestablishing a care for the self (Heidegger 1962[1927]:370–378) after the trauma of diagnosis. Joking about HIV was an especially poignant method of emphasizing one's normality because it involved a reversal or recontextualization of the stigmatizing humor noted in everyday life.

The choir's particular kind of acceptance, including an assertion of normality, was not often commented upon but was a central presupposition of joking about HIV among choir members. As mentioned above, choir rehearsals, which occurred in a garage in the township outside of Durban, were the primary place where joking about HIV occurred. The sprawling urban/suburban center of Durban, like many globalized cities, was characterized by unevenly accessible spaces with corresponding ideologically constructed public/private distinctions (Gal 2005; Appadurai 1996). For choir members, such spaces became places of disclosure or nondisclosure, depending in large part on the identity of audience, auditors, and potential overhearers (Goffman 1981). Rehearsals in the township garage were a safe space in which all participants knew of each other's HIV-positive statuses. The choir met on Sundays at the home of one choir member's parents. The parents did not know the individual's HIV status, but were rarely home on Sundays because they regularly attended church and did not usually come near the garage. Even if they did, explicit reference to an individual's HIV-positive status was extremely rare and so nonpurposeful disclosure was unlikely. The cinder-block garage with corrugated fiberglass roof housed the choir's equipment, including a PA system, a keyboard, and a dilapidated drum set. Rehearsals were usually set up with the bass player, drummer, and pianist seated with their backs against one wall of the garage, facing the vocalists who arranged themselves in a broad semi-oval.

Excerpt 1 is an example of joking during choir rehearsals. Just before this excerpt began, choir members were rehearsing, not only singing and playing music together but also practicing choreographed dance moves to accompany the music making.

Excerpt 1

	03-02-2008, choir rehearsal, tape 2, 12min 8sec
	B = Bongiwe (female), L = Lethu (male), N = Ndumiso (male), T = Thulani (male).
	Other participants included Amahle, Celokuhle, Zethu (females), Philani, and Siphesihle
	(males); for transcription conventions, see Appendix 1.
01	B; (????????????) uyabona uLethu ungikhahlelile =
	“(???????????) Do you see (that) Lethu kicked me?”
02	T; = ((laugh//ing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .]
03	L;	[ mina?
		[“me?”
04	B;	[ uzotheleleka ngempela futhi umuntu ozo]sondela la
		[“a person who comes close here will truly be infected again”
05	T; ((higher pitched start, laugh//ing))
06	L;	[ ukhahlelwe imina?
		[“you were kicked by me?”
06	B; awuboni ukuthi igazi leli?
	“don't you see that this is blood?”
07	T?; seriously?
	“seriously?”
08	B; [ (nginenkinga)
	[“(I have a problem)”
09	L; [ ngikukhahlele ngikephi?
	[“Where did I kick you?”
10	B; ngeteku // ((laughing))
	“by my shoe // ((laughing))”
11	L;	[ ((laughing))
12	T; eish eish- uzosuka pho njengoba uhleli lapho phansi wena
	“[expression of disbelief]- you will leave as you are sitting down there, you”
13	B; (?????????)

Excerpt 1, like other instances of self- and other- directed joking about HIV that I witnessed, was setting-specific (Handelman and Kapferer 1972), created in response to the emergent dynamics of the local communicative context (Goodwin 1979; Goodwin 1981). In Excerpt 1, two choir members, Bongiwe and Thulani, collaboratively constructed a play frame (lines 01-05: see sections on improvisation and play frames; Bateson 2000:177–193; see also Goffman 1974). Within the play frame, Bongiwe offered at least two subjects for evaluation. First, she noted Lethu's clumsiness by pointing out that she had been kicked by Lethu (line 01); second, she commented on the possibility that her open wound could re-infect another choir member (line 04).

In Excerpt 1, just as there were two subjects of joking, there were two sets of shared presuppositions. While all interaction progresses based on the metapragmatic implications of presuppositions/entailments of just-previous statements (Silverstein 1993), joking is especially tied to presupposition. As noted above, joking is humorous in part due to the unspoken recognition by multiple participants that they share the knowledge needed to decode presuppositions and recognize the secondary text of the joke (Bateson 1952; Basso 1979). The first set of presuppositions in Excerpt 1 involved the fact that a semiprofessional singer/dancer should be proficient enough not to step on other members of the group. Lethu was, in fact, the youngest and most novice member of the choir as well as the only male member without some sort of special role (the other men all played instruments, except for Thulani, who directed the singers). These things singled him out as someone who could be toyed with at times—though no group members were immune from a quick laugh at their expense.

The second set of shared presuppositions involved acceptance of the fact that all of the group members were HIV-positive; Bongiwe's statement that a person who got close to her would be infected again indexed the shared knowledge that she and the others were already infected. In addition, there was possibly an element of bravado involved. Choir members were unusually knowledgeable, in comparison with many other South Africans, about the scientific medical model of HIV. They explained to me at a later time that becoming infected with other individuals' strains of the virus could greatly reduce one's lifespan. The virus mutates quickly within each individual, meaning that a re-infected individual would become a doubly infected individual. Joking about such a serious consequence of contact with another's bodily fluids may have been a way for Bongiwe to showcase her acceptance of her HIV status.

A final key aspect of this joking was the ambiguity of the “butt” of the joking. In contrast to research that clearly differentiates (1) teasing from (2) joking about an absent other and (3) self-denigrating joking (Boxer and Cortés-Conde 1997:279), the joking discussed here is ambiguously about self, present others, and absent others. Bongiwe called attention both to herself and to the other group members as being HIV-positive in her statement, “a person who comes close here will truly be infected again” (line 04). She drew attention to herself as the implicit agent of infection, both in line 04 and in line 06, “don't you see that this is blood?” (The blood was her own blood, which would re-infect the others.) She also drew attention to the others in the group through her use of the Zulu word futhi, translated here as ‘again’ (line 04), indicating that the other choir members were already infected. By relying on choir members' knowledge of the scientific medical model of HIV and including all the choir members as potential objects of re-infection, Bongiwe's joking reinforced the choir community at the same time that it distinguished the group members from an out-group: namely, nonchoir-member Zulu-speaking South Africans (cf. Kivy 2003). The primary text of Excerpt 1 had Lethu kicking Bongiwe, drawing blood and thereby presenting the danger that anyone who came close would be infected (again); a secondary text portrayed Lethu as clumsy for a semi-professional singer and dancer. Acceptance was a key aspect of this humor's effectiveness in constituting support: participation in the joking implied not only an understanding of one's HIV-positive status, but also a willingness to play with it and not dwell on the inevitability of one's death.

Joking Property #2: Improvisational Co-construction

The second property of joking about HIV, its improvisational co-construction, was significant because it demonstrated the collaborative nature of acceptance of one's own and others' HIV-positive statuses. As noted, stigma is relational, constituted between individuals rather than an inherent aspect of a person or a particular attribute (Goffman 1963). It follows that an effective speech genre for facing stigma might be especially dependent for its success upon input from multiple participants. Previous research suggests that performance “allows humans to practice the display and reception of emotional states and the social transformation of individuals in a protected ‘framed’ environment” (Beeman 2010:129, emphasis in original; more on framing in the discussion of joking property #3, below). Through joking, choir members were able to practice (or rehearse) an inversion of the relational construction of stigma, emphasizing collaborative acceptance within the “protection” of two performance frames, namely the joking itself and choir rehearsals more broadly.

The improvisational co-construction of the choir's joking meant that the content of joking responded to the emergent dynamics of choir rehearsals rather than being “canned” or prefabricated (Sawyer 1996). Previous research indicates a basic similarity in the semantic structure of “canned” and improvised joking (Norrick 2003). Both involve a statement or situation that can be interpreted in two ways, according to a primary and/ or secondary text (as discussed above). In terms of participation structures (Philips 1972), however, collaboratively improvised joking is quite distinct from “canned” joking. In collaborative improvised joking, active semantic input by co-participants is required to sustain the joking frame (Silverstein 1993; Sawyer 2001; Duranti and Black 2011). In this way, mutual understanding of the shared presuppositions of the joking is displayed, and “intersubjective understanding” is maintained (Heritage 1984:259). This is sometimes considered one among many types or levels of intersubjectivity available to human interactants, built upon a baseline intersubjectivity that involves recognition of the Other as a distinct thinking subject (Duranti 2010). Through such moments of intersubjective understanding, the improvisational co-construction of choir members' joking was a key property that contributed to the reinforcement of community membership.

Excerpt 2 is an example of how choir members improvisationally built off of each other to construct jokes about HIV.² This excerpt comes from before the start of a choir rehearsal, when group members and I were cleaning up the garage in which the choir rehearsed, moving things around to make space and setting up the keyboard, bass, and drum set.

Excerpt 2

	05-11-2008, choir rehearsal, tape 1, 11min 10sec
	B = Bongiwe (female), N = Ndumiso (male), other participants included Dumisile
	(female), ethnographer (Steven Black) (transcription conventions in appendix 1)
01	B; ((sweeping)) sima lapha vele?
	“is it supposed to be here?” ((“it” refers to a small bench))
02	N; yah sima lapha. noma uzosidonsa uhambe naso.
	“Yeah, it's supposed to be here. or you'll drag it and go (home) with it.”
03	B; hhayibo. ngiyaphuquza mina.
	“Hey (no). Me, I am making dust.”
04	N; sizoba ne TB
	“We will get TB”
05	B; kade ngingashaneli benginiphathisa ngethi- ehh. nge TB.
	“I wasn't sweeping I was just infecting you all with- ehh. with TB.”
06	N; mmm.

In Excerpt 2, speakers Ndumiso and Bongiwe built each next statement upon the previous, co-constructing a play frame and then a joke about tuberculosis infection. The excerpt began with Bongiwe sweeping and asking about where to position a small bench upon which choir members sometimes sat (line 01). In line 02, Ndumiso responded “yeah,” and then established a play frame with his exaggerated claim that if Bongiwe did not leave the bench where it was, she would “drag it and go home with it” (line 02). Recognizing this play frame, Bongiwe then explained that she was “making dust”—the opposite of what one should do when sweeping (line 03). Ndumiso next expanded on Bongiwe's statement, perhaps perceiving an implicit indexical entailing of TB (tuberculosis) through the verb phuquza 'to make dust'. Ndumiso said outright, “we will get TB” (line 04). After this, Bongiwe made the link between making dust and TB explicit, saying that she “wasn't sweeping” but instead was infecting nearby individuals with TB (morpheme-labeling conventions are explained in Appendix 2):

05	B; kade ngingashaneli be- ngi- ni- phath- isa ngethi- ehh. nge TB.
	ContPst- 1s- o2p- carry- CAU
	“I wasn't sweeping I was just infecting you all with- ehh. with TB.”

In this statement, Bongiwe made the “butt” of the joking ambiguous, rather than specifically directed at Ndumiso, through her use of the second-person plural object marker ni- on the verb phath-‘infect’ (lit. ‘to carry’).

TB is an airborne disease that infects an individual through the lungs. It is closely associated with HIV in South Africa, in part because there are not enough resources to effectively combat the opportunistic infection even though it is easily curable with careful treatment (Farmer 1999; Packard 1989). At least two research and treatment organizations where choir members were employed focused their medical efforts on the intersection of HIV and TB infection. In fact, the association is such that many South Africans would not admit to being infected with TB because it would mark them as HIV-positive. One significant aspect of this excerpt was that both Bongiwe and Ndumiso borrowed the English medical term TB, rather than the Zulu isifiso sofuba. Borrowing of English medical terms in reference to HIV/AIDS was common among choir members. In doing so they indexed their association with international and local aid.

In Excerpt 2, Bongiwe and Ndumiso deftly played with their knowledge of the epidemiology of TB and the well-known association of TB with HIV, suggesting that by sweeping dust into the air they would potentially expose their lungs to pathogens. Here as in Excerpt 1, the presupposition upon which the joking is built is the fact that the conversational participants are HIV-positive. In Excerpt 2, there is the added link that being HIV-positive made choir members extremely vulnerable to TB infection. Again, there is a kind of bravado or cavalier attitude involved, since choir members know that they are much more vulnerable to TB infection than individuals who are not HIV-positive. The improvisational co-construction of joking about HIV meant that group members displayed a form of acceptance of their HIV-positive statuses that was inherently relational in its composition, in parallel with the relational construction of HIV stigma (Goffman 1963; Samuels 2004:68).

Joking Property #3: A Play Frame/Speech Genre

As mentioned at the beginning of this article, secondary texts are interpretable as such within a play frame. I suggest that, due to the sensitive nature of the topic, joking about HIV was possible in part because it was a performance of a secondary text—that is, it was contextualized within a play frame and a genre of joking. Erving Goffman's influential work on stigma emphasizes both the presentation of self and the possibility of passing for normal as important constraints on “the management of spoiled identity” (Goffman 1963: ch. 2; see also Goffman 1959). As mentioned, Goffman begins with the notion that stigma centers on “an attribute that is deeply discrediting” within a local set of relationships (Goffman 1963: 3; see also Link and Phelan 2001). In South Africa in 2008, HIV had the potential to be “deeply discrediting” partly because of local explanatory frameworks of spiritual pollution that prescribed avoidance, both physical and verbal (Ashforth 2005: ch. 6; see also Douglas 1966; Steiner 1956). The significance of this for the current discussion is that verbal avoidance of talk about HIV was not just a matter of passing for normal, but also part of embodied, habitual patterns of avoidance that resulted in pervasive silence, euphemism, and a confusion of multiple models of HIV infection.

Correspondingly, in the choir's joking about HIV, acceptance of one's HIV-positive status was doubly indirect. As discussed in detail above, the first form of indirectness was use of presupposition to indicate acceptance. The second form of indirectness was the performance of the secondary text. The concept of the secondary text was introduced at the beginning of the article. Here I go into more detail about the linguistic practices of joking about HIV that produced the secondary text, namely (1) the invocation of a play frame, and (2) the locally (semi)defined speech genre that made joking about HIV recognizable as such.

Play Frames. As indicated above, choir rehearsals were generally performance-oriented, and most talk during rehearsals was interpreted within a default play frame. As is the case in many rehearsals for performance, a play frame made it possible for group members to creatively manipulate aspects of the music and suggest changes in a forgiving context. In joking about HIV, choir members reinforced and reconstituted this play frame, especially through prosody, facial expression, and exaggeration. For instance, in Excerpt 1, involving Bongiwe's accusation of Lethu kicking her, semiotic resources (Goodwin 2000) for maintaining a play frame included: sequential organization, in Bongiwe's delayed response to Lethu's questions in favor of continued commentary about Lethu's footwork ineptitude; body orientation, in Bongiwe's maintaining an orientation and eye gaze toward the rest of the group even when responding to Lethu; and facial expression, in Bongiwe's incipient smile that then was interactionally coordinated to coincide with Lethu's laugh.

As speech play, group members' talk about HIV was subject to an evaluation of one's ability to contribute creatively to this local genre of verbal art (Bauman 1977; Sherzer 2002). In rehearsals, framing talk as speech play meant focusing on the interpretation of poetics, including a focus on the form of the message: a focus on the poetic function of language (Jakobson 1960:356). As Jakobson explains, “The supremacy of poetic function over referential function does not obliterate the reference but makes it ambiguous” (Jakobson 1960:371). An increased emphasis on ambiguity and interpretation means that the scope of presupposition and entailment becomes greatly expanded. In formal logic, presupposition has sometimes been discussed in terms of truth-value (e.g., Russell 1905; Russell 1957; Strawson 1950), but within such dialogues, it is important to keep in mind that in talk, truth-value is contextualized within frames. A frame defines a presuppositional context against which the truth-value of utterances within the frame will conventionally be evaluated (Tyler 1978:307). A play frame allows for a certain leeway in interpretation of truth-value, a leeway rooted in the simple presupposition of nonaggression (Bateson 2000[1972]). The implications of entailment are similarly altered when talk occurs within a frame (Tyler 1978:303). This expansion of the scope of presupposition and entailment allows for a plethora of interpretive moves by performer and audience. Texts can be recontextualized through the use of devices such as metaphor, parallelism, and the heteroglossic use of prosody or other genre-specific features to index associations among texts and among ideas without explicitly stating these associations (Bauman and Briggs 1990). Play thus becomes a potential locus for expressing the otherwise inexpressible, and therefore for navigating stigmatization. Of particular relevance here, given the choir's subordinate structural position relative to both local hospitals and international aid groups, are Marxist perspectives that assert that play allows for critique and subversion of dominant ideals within the constraints of particular social situations (Limón 1989). As discussed below, the choir's jokes conceptually inverted stigmatizing jokes about HIV that, choir members said, permeated daily life in South Africa.

Speech genres. As locally recognizable forms of talk, speech genres are a central aspect of framing discourse in ways that suggest possible interpretations, often according to implicit, habitualized ways of understanding and reacting to genre-d speech (Hanks 1987:670). While the term “genre” can be used to refer to patterned ways of speaking that are more or less formally defined within a given cultural context (Hymes 1974:441–442), recent work recognizes the difficulties of differentiating one genre from another based on formal properties, opting instead for research-participant identification of genres (Briggs and Bauman 1992; Duranti 1994:167). For choir members, joking about HIV was a semi-conventionalized speech genre. I suggest that by placing talk about HIV within this speech genre, choir members diffused individual responsibility for the social implications of that talk.

This understanding of genre and individual responsibility, significantly developed in previous research (see Hill and Irvine 1992), draws upon literary theorist Mikhail Bakhtin's notion of heteroglossia (e.g., Bakhtin 1981b). Bakhtin explains that the novel, as a literary invention, incorporates heteroglossia not only in the writing of a narrator and characters but also in a single character's speech (Bakhtin 1981a). Bakhtin (1984) writes, “Ultimate semantic authority—the author's intention—is realized not in his direct discourse but with the help of other people's words, created and distributed specifically as the words of others” (p. 188). Bakhtin also formulated a theory of speech genres based on this concept, explaining that, in some sense, the words that we speak are not our own, and that our speech is shaped in key ways by sociocultural norms or genres of expression (Bakhtin 1981c). At the same time, however, speech genres are the “sclerotic deposits” of intentional action in the world (Bakhtin 1981a:292). It is within this dialectic of self and society in historical trajectory that genres gain form, value and authority (Bauman 1992:128).

As mentioned earlier, joking about HIV was an understood genre within the choir community, an extension of joking about death and danger more generally and a transformation of stigmatizing jokes. It is no surprise that choir members engaged in self- and other-directed joking about HIV, given the habitualized and internalized nature of joking about death, danger, and HIV more generally (Bourdieu 1977:78–79; Hanks 1987; Hanks 2005). Choir members utilized joking about death creatively, though, inverting others' stigmatizing jokes about HIV. As one choir member wrote: “MY MESSAGE TO THE WHOLE WORLD IS: DON'T LAUGH AT US; RATHER, LAUGH WITH US” (emphasis in original; this line of text was part of a personal life story, written for publication in a book of choir members' life stories—a book that was never finished). A central identifying feature of the choir's transformation of joking was that participants self-identified as HIV-positive and pointed to other interlocutors as similarly infected. This was a marked shift from stigmatizing joking about anonymous HIV-positive people in which none of the interlocutors would be labeled as infected.

Within the heteroglossic confines of a speech genre, the distribution of responsibility for talk about HIV was effectively dispersed throughout the choir. Any particular group member could successfully be seen to be instantiating only a generalized formula that all members shared in common (Chafe 1992:73; Du Bois 1987). As Irvine's (1992) study of ritual insult poetry indicates, diffusion of responsibility can occur by “dissociating speaker from author, and dissociating addressee from target” (p. 128). Similarly, the conventionalization of a genre of joking about HIV had the effect of dissociating speaker from author, while the ambiguity of the “butt” of the joking dissociated addressee from target (Goffman 1981:144). Joking about HIV was a creative, improvisational endeavor for confronting stigma. The fact that this speech genre was shared by choir members helped to diffuse responsibility for confrontations with stigma.

In sum, in the local cultural setting the contextualization of co-constructed talk about HIV within a play frame and within a conventionalized speech genre expanded, rather than constrained, the scope of possibilities (cf. Black 2008) for approaching stigmatization. Rooted in the presupposition of acceptance, improvised self- and other-directed joking about HIV became an integral part of support by emphasizing shared community membership.

Gendered Social Differentiation Through Joking

In Excerpts 1 and 2, a key function of joking about HIV was to reinforce membership in the choir while separating the choir from nonchoir members. Sometimes, though, joking functioned to link only some choir members together while separating those individuals from other group members. The processual model of differentiation through joking was the same, whether choir members were distinguishing themselves from nonchoir members or one group of choir members were distinguishing themselves from other choir members (Irvine and Gal 2000). Excerpts 3 and 4 showcase the divisive capacity of joking and other related forms of humor such as insult.

The following excerpts demonstrate a division of the choir based on gender. Gender was the most commonly invoked social distinction within the choir, and this division was also strongly evident among Zulu-speakers in 2008 (Appalraju and de Kadt 2002; Ige and de Kadt 2002; Meintjes 2004). Gendered patterns were a crucial aspect of Zulu-speakers' responses to the HIV epidemic (Clark 2006; Leclerc-Madlala 2001). Women were sometimes constrained in their discussions of sexuality and HIV due to gendered Zulu notions of hlonipha‘respect’ (Finlayson 2002; Thetela 2003), yet often seemed to be handling the epidemic more successfully, getting tested and taking their anti-retroviral medicine. Zulu-speakers' conceptualizations of stoic warrior heterosexual masculinity (Meintjes 2004) were also a central piece of the puzzle.

Excerpts 3 and 4 come from the same rehearsal as Excerpt 1. At the end of that rehearsal, one of the group leaders called everyone into a huddle to discuss the possibility of appearing on a local TV show. Even though the choir had gone on tour to the United States and England, some of the choir members had not disclosed their HIV status to their parents, close friends and/or family. The issue that needed to be discussed was whether or not choir members were prepared for neighbors and family members to see them on television discussing their HIV-positive status. For individuals who had not disclosed their status, appearing on the TV show meant risking disclosure. Excerpt 3 begins after a group leader, Thulani, had introduced this topic.

Excerpt 3

	03-02-2008, choir rehearsal, tape 3, 21min 20s
	A = Amahle (female), C = Celokuhle (female), D = Dumisile (female),
	T = Thulani (male), Z = Zethu (female); other participants include Bongiwe (female),
	Siphesihle (male), Philani (male), Lethu (male)
01	Z; for i disclosure, uqondise kwi disclosure?
	“For disclosure, are you directing it [the conversation] to disclosure?”
02	T; ya
	“yeah”
03	?; kade
	“a long time ago”
04	A; engani sazidisclosa
	“but we disclosed (already)”
05	T; yah, kodwa phela . . . noma kwi this case seku uyabona // (???)
	“yeah, but truly . . . or in this case it is you see”
06	A;	[ (ayifuthi) kuzothiwa amantombazane,
		[“not again it will be said (that) girls,”
07	>abafana ba negative abafana amantombazane a positive< usholo lokho?
	“the boys are negative the girls are positive, are you talking about that?”
08	T; sengisholo ukuthi // in this case-
	“I'm saying that // in this case-”
09	C;	[ ((laughs, slaps thigh))
10	T; [ (sengisholo ukuthi)
	[“I'm saying that”
11	D; [ abafana ba negative ((laughing voice))
	[“The boys are negative ((laughing voice))!!”

In Excerpt 3, Zethu incredulously asked Thulani if he was directing the conversation towards the topic of disclosure of HIV status (line 01). After Thulani's reply, “yeah” (line 02), Amahle protested, “but we disclosed already” (line 04). Thulani started to respond, but Amahle immediately launched into a performative criticism: “not again will it be said that . . . the boys are negative the girls are positive, are you talking about that?” (lines 06-07). By speeding up the pace of her talk and utilizing a repeated singsong prosodic contour, Amahle invoked a play frame.

Amahle's question in lines 06-07 referred to an incredible violation of trust perpetrated earlier in the group's history by the men of the choir. At that time, the men of the group told local community members that the women of the choir were HIV-positive but the men were not. The men said that they were helping the women because they were so much more knowledgeable about music than the women. This lie violated the trust and solidarity of the choir. Amahle brought up this issue in response to the notion that it was necessary for the group to discuss disclosure even though, according to many group members, they were supposed to have fully disclosed long ago.

After Amahle's performed utterance of lines 06-07, Thulani again tried to respond (lines 08 and 10) but was quickly overtaken by other women's laughter and participation in Amahle's joke (lines 09 and 11). In line 11, Dumisile's statement “the boys are negative” was infused with breathy laughter, indicating the absurdity of that statement within the shared framework of the choir's acceptance of their HIV-positive statuses. In this excerpt, the women of the group co-constructed a joke at the expense of the men of the group, effectively criticizing the men's inability to fully disclose their HIV statuses and to be local community activists.

As the conversation continued, social differentiation between the men and women of the choir became focused on a single member of the group, Siphesihle, who had still not disclosed his HIV status to his parents. This was a well-known but unspoken fact. Amahle later explained to me that this group member was afraid his mother could not handle the news that he was HIV-positive. During the conversation about disclosure and the local TV show appearance, though, none of this sensitivity emerged. In Excerpt 4 there was no secondary text, only a primary text centered on the insulting of Siphesihle's inability to disclose his HIV status:

Excerpt 4

	03-02-2008, choir rehearsal, tape 3, 22min 30sec
	A = Amahle (female), C = Celokuhle (female), D = Dumisile (female),
	T = Thulani (male), Z = Zethu (female); other participants include Bongiwe (female),
	Siphesihle (male), Philani (male), Lethu (male)
01	Z; oh kay ph//ela akasho individually
	“Okay truly he/she must say individually”
02	C;	[ (????????????) individually uzozake naleyo nkingake =
		[“(?????????) individually he/she will come (then) with that problem”=
03	Z; = ehheh
	=“Uh huh”
04	C; but the rest of the group ayikho inkinga uma uloga kwi internet ufuna uZethu Ngubane
	“But the rest of the group, there is not a problem. If you log into the internet looking for
	Zethu Ngubane”
05	B?; umthole
	“you will find her.”
06	C; uyangena, umthole kuthiwe unengculazi wayithola ngelanga elithize . . .
	“You enter, you will find her it will say she has AIDS what day she got it . . .
07	akusiyona nje imfihlo (4.0)
	it is not just a secret”
08	C; ° wathula kangaka nje Siphesihle ((whispered))
	“Why are you so quiet Siphesihle”
09	C?; ushawa uvalo?
	“Are you scared?”
10	((women laughing))
11	S?; (??????//??)
12	?;	[ as “hamben” ma ehke
	[“Let's go, then, eh girls”

In lines 01-02 of Excerpt 4, Zethu and Celokuhle clearly articulated the notion that a problem about disclosure was an individual problem. While the group might have tackled such issues in the past, Celokuhle, assisted by Bongiwe, made it clear that she thought the choir had moved on to become a publicly available activist-type organization. Celokuhle asserted that “if you log in to the internet looking for Zethu Ngubane . . . you will find her, it will say she has AIDS, what day she got it,” and then concluded, “it's just not a secret.” (lines 04-07). When four seconds of silence ensued, Celokuhle finally singled Siphesihle out, whispering, “Why are you so quiet . . . are you scared?” One of the women ended the joking in solidarity, saying, “let's go, then, eh girls” (line 12). Here, gender continued to be explicitly invoked as a dividing line within the choir. Women indexed themselves as disclosers and the men as non-disclosers. This was a potentially threatening commentary, because any of the choir members could have revealed Siphesihle's HIV status at any time. In reality, both women and men in the choir sometimes withheld disclosure based on the social situation and participants involved.

The social distinction of Excerpts 3 and 4 reveals how a focus on acceptance of one's HIV status despite stigmatization could sometimes lead to social differentiation in addition to community reinforcement. Joking about HIV was considered by choir members to be a part of thinking positive: as one song's lyrics assert, “If you're negative, stay negative; if you're positive, think positive.” In cases where choir members were unable or unwilling to display acceptance, joking about HIV quickly lost its secondary text and became a biting, teasing social commentary.

Such instances point to the fluid, and sometimes ambiguous, nature of the ideal dichotomies of disclosure/nondisclosure and acceptance/lack of acceptance (a topic that I take up in other work currently in progress). Most often, though, a positive outlook was maintained in choir rehearsals, and this positive outlook extended far beyond choir rehearsals, helping choir members to live healthy and fruitful lives at a time when others living with HIV were faltering. This was driven home in a most forceful way on at least two tragic occasions. The first occasion was the death of the HIV-positive brother of two of the choir members. After the brother died, the two sisters found a shoebox full of unused anti-retroviral medication under his bed. He had been going to collect the medicine, probably to assuage his sisters, but not taking it. The second was the attempted suicide of the HIV-positive husband of one of the choir members. The husband was overwhelmed with guilt after infecting his wife with the virus.

Conclusions

Humor is sometimes overlooked in the anthropological analysis of language and communication, possibly for the same reasons that humor can be profoundly significant in linguistic practices central to social transformation: because humor is complicated, indirect, performative and often pleasurable. The analysis in this article has synthesized previous research on speech play and discussed specific properties of a genre of joking that made it useful for facing HIV stigma in South Africa. The article has discussed linguistic practices of community maintenance and social differentiation in the midst of stigma through an analysis of how a local genre of joking functioned to distinguish groups of people from one another and to construct key facets of HIV support. I have shown how joking about HIV performed these functions through the following properties: (a) reliance on shared presupposition, namely the presupposition of acceptance of group members' HIV-positive statuses, allowing group members to index a stigmatized aspect of their identity without discussing it outright; (b) improvisational co-construction—diffusing responsibility and implicating multiple choir members in a multiply intersubjective performance; and (c) invocation of a play frame and a genre—thereby emphasizing multiple interpretations and valuations of talk based not only on content but on performative aesthetic appeal (often a pleasurable activity), while simultaneously further reducing personal responsibility for the words being spoken. In addition, by recontextualizing a noted pattern of stigmatizing joking about HIV that was present in the larger sociocultural context, choir members were contributing to a local social transformation that was linked to choir members' current and future roles as AIDS activists (Briggs and Bauman 1992). The question of how the shift in choir members' roles and identities might relate to broader social change in and around Durban, South Africa remains to be seen.

The analysis of this article has a broader theoretical and practical significance as well. This article contributes to the small body of literature on the linguistic and discursive components of stigma and, unlike previous questionnaire and survey data, expands knowledge of the actual communicative practices of people living with HIV/AIDS in South Africa. On a practical level, understanding the interactional negotiation of stigma is critical to designing effective activist interventions in response to the South African AIDS epidemic and in contexts of stigmatization more generally. On a theoretical level, I suggest that the linguistic properties of joking about HIV that are identified above might be more general properties of humor across cultural contexts, utilized by people who are attempting to talk about issues at or beyond the boundaries of what is considered acceptable to talk about in most other contexts (Norrick 2005). Recently, scholars have noted that the central interrelationship between speech, on the one hand, and the experience of speaking, on the other, has been largely overlooked in linguistic anthropological scholarship (Ochs 2010). An underlying premise of my focus on joking about HIV has been that the experience of speaking about one's HIV-positive status—and indeed, of playing with one's status—is significant in constituting support. I suggest that in addition to analyzing what people do with genres or discourses, linguistic anthropologists might also discuss what genres or discourses do for people—that is, how the experience of participating in particular genres or discourses may affect individuals and groups over time.

Acknowledgments

Funding this research was provided by the National Science Foundation and the University of California, Los Angeles Department of Anthropology. I thank the woman known in my work as Amahle Ngubane, as well as her amazing and welcoming family. I also thank the other choir members for their generosity of spirit; and I thank Talent Mbeje for her work as a transcription assistant as well as for her friendship. Special thanks to Elizabeth Falconi for fieldwork assistance and guidance throughout my writing. Special thanks also to Alessandro Duranti, Linda Garro, and Paul Kroskrity for their generosity in suggesting changes to this work. Thanks to Robin Conley, Cre Engelke, Heather Loyd, Merav Shohet, members of the UCLA Discourse Laboratory, members of the UCLA Mind, Medicine, and Culture meeting group, members of the University of Michigan LingLab, Joe Henkins, Rupert Stasch, and Kit Woolard for their scholarly guidance and comments on this and other, related analyses. Finally, thanks to the anonymous JLA reviewers for their excellent guidance, and to Paul Garrett for his patient editorial advice and scholarly suggestions. Any mistakes in the article are, of course, my own.

Notes

• 1

  Recognizing the multitude of critiques of the use of the ethnographic present (e.g., Fabian 1983; Clifford 1988), I write in the ethnographic past tense, rather than the present, when describing fieldwork events, linguistic practices and generalizations about South Africa (see Gluckman 1968:xxii). By doing so, I hope to emphasize the constantly shifting dynamics of linguistic and cultural patterns. I move into the theoretical present to describe anthropological models of human processes that I see as potentially being recognizable cross-culturally and/or over time.

• 2

  This excerpt also discussed in Duranti and Black 2011.

Appendices

Appendix 1: Transcription Conventions

Transcripts are documents shaped by theory within a particular discipline, and as such, theoretically informed choices must be made as to what aspects of an interchange to include in a transcript (Duranti 2006; Ochs 1979). In the excerpts for this paper, descriptive details of each transcript corresponds to the analysis being put forward with the transcript. Here are the conventions used in the transcripts of this paper (from Goodwin 1990; Sacks, et al. 1974).

• 1) 
  Comments: (( xxx )) Double parentheses encapsulate anything not spoken by the participants, especially a description of gesture, facial expression, body orientation, or of the way in which the talk occurred (e.g., singing versus speaking).
• 2) 
  Sound lengthening ::: Colons show where the sound just prior to the colons has been considerably lengthened.
• 3) 
  Intonation (.) a period indicates a falling intonation. (,) a comma indicates a falling-rising intonation. (?) a question mark indicates a rising intonation
• 4) 
  Latching = equal signs indicate when there is no break between the end of one participant's talk and the beginning of another's.
• 5) 
  Silence (0.8) Numbers in single parentheses mark silence in seconds and tenths of seconds
• 6) 
  Problematic hearing (????) talk in parentheses indicates that the transcriber was unsure of the veracity of his/her hearing of the material, question marks indicates the transcriber could not make an educated guess as to the material in question.
• 7) 
  Cut-off—A dash indicates a sudden stop in an individual's talk (a glottal stop).
• 8) 
  Emphasisaa underlining indicates emphasis in an individual's talk.
• 9) 
  Overlap // [ Double slashes indicate the place where another individual begins to talk at the same time, while a forward bracket indicates the overlap in the second speaker's talk.

Appendix 2: Morpheme Labels

Morpheme labels were produced in consultation with several contemporary and classic Zulu grammar books (Doke 1930; Mbeje 2005; Poulos and Bosch 1997; Poulos and Msimang 1998). Below is a list of morpheme labels utilized in this article.

Verbal subject markers

1s = 1st person singular subject marker

Verbal object markers

o2p = 2nd person plural object marker

Tense

ContPst = Continuous past tense

Verbal extensions

CAU = Causative extension