Hearing the unspoken in research amid HIV stigma

In 2008, approximately twenty percent of South Africans were living with HIV/AIDS. Some estimates of the infection rate in the province of KwaZulu-Natal—where Durban is located—were as high as forty percent (Human Sciences Research Council 2008; UNAIDS 2008). Like other places in the world plagued by HIV, structural inequalities fueled a drastically uneven distribution of AIDS (Farmer Reference Farmer2003). Colonialism and apartheid played a primary role in shaping these structural inequalities and shaped their impact on health and illness (Packard Reference Packard1989). Despite the existence of an urban middle class and a history of black economic mobility (Hunter Reference Hunter2010), black South Africans were approximately ten times as likely as white South Africans to be HIV positive, with the weight of infection falling disproportionately on impoverished individuals (Squire Reference Squire2007:18). In contemporary South Africa (as in southern African more generally), HIV/AIDS was primarily conceptualized as a heterosexual illness. Women were unfairly and disproportionately associated with the illness, and HIV stigma was pervasive. A key component of stigma was a complicated, multilayered symbolic system of avoidance of talk about HIV rooted in concerns about respect, maintenance of hope for the future, and a more general avoidance of illness (e.g. Selikow Reference Selikow2004; Squire Reference Squire2007; Wood & Lambert Reference Wood and Lambert2008). Before antiretroviral drugs (ARVs) became widely available, the autoimmune disorder's progression was often swift and noticeable, and names like “slim” were used to describe the debilitating effects of the illness on human bodies. As mentioned, in 2004 the South African government began the “rollout” of ARVs to the public, reversing its highly controversial denial of the link between HIV and AIDS (Leclerc-Madlala Reference Leclerc-Madlala2005; Fassin Reference Fassin2007). While this reversal may be leading to shifting attitudes about HIV/AIDS, in the short-term it has made it possible to better “pass for normal” (Goffman Reference Goffman1963) and thereby avoid detection as someone living with HIV. ARVs have allowed people with HIV to manage the ill effects of the virus and live healthy lives. In some cases, then, the availability of ARVs has led to more indirection and avoidance than before, not less.

Conducting linguistic anthropological research in such a context presented special challenges. In South Africa, though many with HIV have participated in questionnaires and in-depth interviews, fewer have been willing to allow researchers (especially those noticeably different in terms of class, ethnicity, race, and/or nationality) into their everyday lives. To do so would be to draw unnecessary attention to oneself in communities where many have not explicitly revealed their HIV statuses to others. My presence as a white, wealthy (relative to research participants), American male attracted commentary and shaped my interactions with research participants. Here it was helpful to be working with a performance organization that had experience with American and European researchers and aid workers. It was also useful to be a semiprofessional jazz saxophonist. I was able to attend choir rehearsals that took place in one group member's garage in a township outside of Durban, South Africa—an informally segregated group of neighborhoods comprised almost entirely of black, working-class, isiZulu-speaking South Africans. This group exemplified what others have touted as creative, vibrant responses to the HIV pandemic (Comaroff Reference Comaroff2007). My presence at the rehearsals could be explained to other community members as that of an American musician learning about the vibrant tradition of Zulu choral music. In other words, I engaged in a complementary process of “passing,” one without personal consequences, as a researcher who was not focused on HIV. Furthermore, choir members could conceptualize video recording of rehearsals as an extension of their performance orientation, allowing me to look “out of the corner of my eye,” metaphorically speaking, at communication about HIV.

Interactions outside of rehearsals were another matter. It took five months of “hanging out” before some group members who had not disclosed their HIV positive statuses to family or community invited me to spend time with them in other contexts. Gender played a crucial role here. Women more than men in the group had disclosed more widely and offered me more of a view into home life. Women were much less likely than men, however, to discuss health issues, sexual practices, and by extension, lived experience with HIV, with me. One glaring absence in the data I collected is reference to or discussion of sexual assault and rape. South Africa has one of the highest sexual assault rates in the world, and it is likely that at least one of the female research participants had experience with rape or coerced sex. This sort of topic was not something that could have been easily breached in mixed gender company (Thetela Reference Thetela2003). In other forthcoming research based on this project, I analyze indirection in women's talk about sex and HIV in interviews I conducted (Black Reference Black2013).

Still, I became especially close with the family of two sisters in the choir, visiting their township home often and employing one of the sisters as a research/ transcription assistant. Inevitably, rumors circulated in their neighborhood that I was the lover of one of the sisters, or that I was a special white doctor who was visiting them at home to treat them for HIV. These women had already disclosed to many in their neighborhood, though, and enjoyed an especially strong support network of friends and family. My research was greatly aided when, after seven months in Durban, my own partner came to stay with me. In an instant, women in the group gravitated towards her and my social presence was legitimated in a new way.

Avoidance of HIV talk made it difficult to record everyday conversations in which non-choir members discussed the illness. Furthermore, I simply could not record in situations where others did not know group members' HIV positive statuses, for fear of drawing unwanted attention to them and jeopardizing my own version of “passing for normal” as a musician (and not someone studying talk about HIV). I was still able to learn important information about communication and HIV, however. A prevalent general strategy for approaching a taboo topic is to couch the topic as talk about the way others talk about the taboo subject—in other words, to couch one's discussion of a taboo topic in metalinguistic/metapragmatic discourse (Irvine Reference Irvine2011). Much of the data analyzed below come from discourse about the language of HIV collected in interviews and in choir rehearsals. In my analysis, I triangulate multiple sources of data including interviews, song lyrics, participant observation, and popular media to get some idea of how people may have referred to HIV. This provides a necessary counterpoint to choir members' creative ways of facing stigma that I discuss in the second part of the article.

Ethnographic context and research methods

Durban, South Africa, is the largest city in the province of KwaZulu-Natal, the third-largest city in South Africa, and one of the largest port cities in Africa. The concentrated downtown area is flanked on all sides by sprawling informal settlements (areas that included shacks and also cinderblock homes with running water and electricity) and townships (communities built by the apartheid government to be segregated residences for black South Africans working for white employers in the city). As far as I could tell, there were few formal businesses in these surrounding areas, such that employment (when it could be found) usually meant travel by minibus taxi to the downtown area, to wealthier suburbs, or to various factories and shipyards. Most research participants lived in the townships and informal settlements. One consequence of this geographic division was that it was possible for many to avoid disclosure of their HIV-positive status in home communities but then to disclose as HIV positive in support groups, at hospitals, and in the case of the choir, at performances for international audiences in Downtown Durban. This was possible because it was unlikely that one would see one's neighbors in these other, more anonymous contexts.

When I worked closely with the choir in 2008, the group consisted of a core group of five men and six women, with other group members attending rehearsals occasionally. Rehearsals occurred three to four times per month on Sunday afternoons. One limitation that led to the variation in rehearsal attendance was the cost (approximately two US dollars) of traveling to and from choir rehearsals. All group members were multilingual, speaking isiZulu fluently with variable degrees of competence in English, Afrikaans, and Xhosa, among other South African languages. Most group members were in their early to mid thirties, with one younger male member who had recently joined the group. Five group members had grown up in the same neighborhood of the township. Though choir rehearsals were a central focus of my research, I also spent a great deal of time with group members and their various acquaintances in the neighborhoods in which they lived and I traveled with choir members to their family homes in rural areas. I conducted fieldwork from February to November 2008, supplemented by two summers of pilot research and two years of study of isiZulu language and culture. Here I draw upon extensive participant observation, about ten hours of recorded interviews, and approximately forty hours of video recordings of choir rehearsals, performances, and other social events such as weddings, funerals, and birthday parties. This is, notably, a limited recorded data set from which to draw. My goal in this article, however, is not so much to make a generalization about how South Africans engaged in stigmatizing linguistic practices, but rather to understand how the linguistic constitution of stigma was understood by this singular group of people living with HIV and how the choir addressed or faced this experience of stigma through language and music.

Talk about how “they” refer to HIV

Early in my fieldwork in 2008 I interviewed two female University of KwaZulu-Natal students with the goal of learning more about the ways that South Africans who were not HIV positive talked about and conceptualized the illness. While this interview also does not provide enough evidence for generalizations, the overlap between these two students joking and discussion of talk about HIV, on the one hand, and choir members' characterizations of stigmatizing talk about HIV, on the other, does suggest at least some shared linguistic features among two groups of people who rarely, if ever, came into contact with one another. One of the university-student interviewees later became a research assistant who helped to create annotated transcriptions of audio and video recordings. Example (1) is an excerpt taken from my interview with these two students. Prior to the start of this excerpt, I had asked the two students whether they talk about HIV with their friends mostly in English or isiZulu, to which they responded, “Obviously it's better expressed in English… ‘cause all the words are English (laughing).” On further inspection, though, the students recalled that there were isiZulu terms for many things related to HIV. They decided that one possible reason for using English was that isiZulu words were too long. They also explained that they might use isiZulu if they were to speak to family or friends from the rural area that had limited English proficiency. Some ensuing discussion, including their introduction of a few terms in English and isiZulu, led to their mention of what they called “metaphors” for AIDS.

I asked the students, “Well what would you guys- when you guys would talk about AIDS, no you just say AIDS?” This resulted to a fairly raucous listing of terms for HIV. Example (1) begins in the middle of this discussion.Footnote ¹

1. (1) Transcript of interview 2-28-2008, 32min 34sec

   Participants include Lebo and Sthabiso (black female, South African University of KwaZulu-Natal students), SB (Steven Black, ethnographer), and a friend of the students, a study-abroad student from Holland who is observing the interview.

   

Sthabiso and LeboFootnote ² laughed and overlapped with one another as they named different terms for HIV that they knew and taught each other new terms. Sthabiso overlapped with Lebo (lines 1–3) when she recognized the term “Hi vee” (Jefferson Reference Jefferson, D'Urso and Leonardi1984). Hi vee is a sound play on the acronym H-I-V, conventionally pronounced using the sounds of each individual letter. Later (line 4), Sthabiso mentioned a term that Lebo did not know, iHlengiwe Ivy Vilakazi. Hlengiwe is a common female isiZulu first name and Vilakazi is a common isiZulu last name. Note the female gendering of HIV in this term. Ivy is an English name, following a widespread though changing tradition of using both isiZulu and English names (Suzman Reference Suzman1994; Koopman Reference Koopman2002). As Sthabiso indicated (line 9), this is a form of humor rooted in giving the virus a human name. The /i/ prefix before Hlengiwe is a noun class 5 marker, a subject to which I return below. Though Lebo did not know the term iHlengiwe Ivy Vilakazi, she quickly learned it and engaged in overlap (lines 8–9).

By the time the students’ brief discussion concluded, they were using the third person plural pronoun “they” (shown in lines 9, 11, and 15) and the general address term “people” (in line 15) as in, “people avoid saying HIV, they would use, you know, other words. Slang.” This deictic shift might have been related to concerns that the activity of interviewing should result in generalized conclusions about what “people” say (Briggs Reference Briggs1984). However, it may also be that the use of the third person “they” functioned to distance the women from HIV/AIDS and/or from linguistic practices that could be construed as overtly stigmatizing (Irvine Reference Irvine2011). It is unclear if this distancing was gendered. That is, it may also be that because many South Africans incorrectly viewed HIV as something that was transmitted primarily by women (Leclerc-Madlala Reference Leclerc-Madlala2001), these female students had a special interest in distancing themselves from the virus.

The laughter, overlap, and general excitement at the discovery of unknown words for HIV/AIDS and the mention of known ones, along with use of the word “slang” to describe these terms for HIV (line 15), indexed those who would use the terms as fun and clever though engaged in inappropriate behavior. Indirect indexing also accompanied these constructions of hypothetical speaker identity—play and joking negatively sanctioned those actually living with HIV (see Ochs Reference Ochs, Duranti and Goodwin1992:342; Hill Reference Hill1998). The students' discussion indicated that at least some South Africans used speech play to refer to HIV/AIDS. Furthermore, the students themselves used humor to comment upon such forms of reference. In general, previous research indicates that humor is commonly employed to address topics at or near the threshold of conversational acceptability (Sacks Reference Sacks, Jefferson and Schegloff1992; Norrick Reference Norrick2005). In other work, I have analyzed HIV humor as a genre of stigmatization and also demonstrated how the choir I worked with transformed HIV humor into a mode of constructing and reinforcing community as well as confronting the existential traumas of infection (Black Reference Black2012; Black Reference Black2013).

Critiquing how “people” talk about HIV

Members of the choir with which I worked provided another example of talk about how others talk about HIV. Group members distinguished themselves in part through their ways of talking about HIV. On multiple occasions group members compared how others talked about HIV to how they talked about the virus. The distinction was made between terms that gendered HIV as feminine, indicated it as being foreign, and/or placed it within the realm of witchcraft, on the one hand, and scientific medical frameworks, on the other hand, with which group members had become familiar as a result of their interactions with social workers, hospital staff, and international aid groups. Previous research indicates that discourses of rights and responsibilities associated with HIV activism, postapartheid democracy, and international aid opened up spaces for alternative gendering of the HIV/AIDS epidemic (Robins Reference Robins2008:ch. 7; Susser Reference Susser2010). This was also the case among members of the choir with which I worked.

The next excerpt comes from an interview that a research assistant (Amahle—a choir member) conducted with a former support group member. Though in one sense this excerpt is divorced from the exigencies of everyday conversation, it nonetheless provides a window into the themes and concepts that Amahle (and other group members) felt were significant in terms of conceptualizing and discussing HIV infection. The excerpt also showcases some of the ways that scientific medical terms were incorporated into isiZulu discourse.Footnote ³

1. (2) Transcript of interview 9-19-2008, 45min 50sec. A = Amahle (research assistant), R = Rachel (interviewee).

   

This excerpt began with Amahle distinguishing the concept of poisoning (by witchcraft) from the scientific medical concept of opportunistic infection (line 2). She codemixed isiZulu and English as she discussed the link between TB and HIV. Amahle used isiZulu noun class 5 prefix (discussed below), “i-,” to mark the English phrase “HIV and AIDS” and the acronym “TB,” as well as utilizing the English past tense verb “related.” Here, the distinction between a scientific medical explanatory framework and a traditional Zulu explanatory framework was represented by the distinction between English terms and isiZulu terms (on explanations of illness, see Kleinman Reference Kleinman1980 and Garro Reference Garro2000). This codemixing may also be an indication that there is, in actuality, no term that is an adequate scientific medical equivalent for HIV or TB in isiZulu despite the existence of complementary terms (more on this below).

Where Amahle's question indicated a clear distinction between scientific medical and traditional Zulu explanations of illness, Rachel's answer mixed the two together conceptually and linguistically. In discussing the impact of TB on the body, Rachel said (in isiZulu), “when it is that TB has eaten a person a lot,” (line 6) and then indicated that HIV might not be diagnosed if the “TB poison” in one's body is too much (line 7). The notion of TB “eating away” at a person may be used colloquially in English discourse but does not represent a scientific medical perspective on the illness. Note here that Rachel used the English term for poison in line 7. The use of “poison” rather than the isiZulu idliso may have been a way for Rachel to indicate that her understanding of TB was distinct from that of the anonymous “others” that Amahle had critiqued as adopting the traditional Zulu framework of idliso. This shift into English was a significant indicator of such a distinction even though the notion of TB as poison significantly reproduces that very same traditional explanatory framework. This interview data is an initial indication of the conceptual difficulties of navigating isiZulu and English terminology for HIV and related concepts and of translation of scientific medical terms into isiZulu. Incorporation of English terms into isiZulu discourse was not a simple matter. It is for this reason that below I consider codemixing to be a creative act, despite the fact that Amahle, Rachel, and others had learned scientific medical terms in support group meetings and training sessions with doctors and other health practitioners.

Other choir members shared Amahle's critique of what they perceived as hegemonic isiZulu discourse about HIV/AIDS. For instance, the term idliso ‘poisoning by witchcraft,’ also appeared in a song composed by choir members that was rehearsed regularly. The title of the song was Into enjani lena, which could roughly be translated as ‘this thing that is like what.’ The song title referred to HIV/AIDS. Such a title and first lines of the song immediately invoked stigma and the avoidance of direct discussion of HIV.

1. (3) Transcript of Into Enjani Lena lyrics

   1. 1 Into enjani lena, into enjani lena

      ‘The thing that is like what, the thing that is like what’

   2. 2 Okukhulunywa ngayo

      ‘That is talked about’[repeated 2x]

   3. 3 Abanye bahlekisa ngayo

      ‘Some joke about it’

   4. 4 bayibiza ngeqhoks ((iqhoks = ‘high heels’))

      ‘They call it iqhoks ((some are teasing about it))’

   5. 5 Bayibiza ngamagama amathathu

      ‘They call it “a three letter word” ((HIV))’

      [repeated 2x]

   6. 6 Abanye bayibiza nge-TB

      ‘Some call it TB’

   7. 7 Abanye bayibiza ngedliso

      ‘Some call it poisoning ((by witchcraft))’

      [repeated 2x]

   8. 8 Waze wangiyala, wangiyala wengculazi

      ‘How you have taught me through pain, you, AIDS, have taught me through pain’

   9. 9 Ngibengisathe ngibathe kukhulunywa ngami

      ‘Whatever I do, they are talking about me’

      [repeated 2x]

In addition to the avoidance of naming HIV, the first verse of the song is notable for its absence of a grammatical agent. The verb -khuluma ‘talk’ takes a relative prefix and passive suffix (resulting in palatalization characteristic of this phonological combination in isiZulu). The English translation of this phrase would be, ‘that is talked about.’

Until the last set of lyrics, this song is about how others talk about HIV. There is no “you,” “I,” or “us,” only a “they.” “Some [people]” joke about HIV (line 3), and name HIV using terms of avoidance and gendered stigmatization—iqhoks ‘high heels’ and amagama amathathu ‘a three letter word’ (lines 4–5). iQhoks is an onomatopoetic term mimicking the sound of high-heeled shoes. In some Zulu frameworks, this may become a substitute for HIV because it is associated with femininity, in particular a kind of “Western,” Euro-American, or White femininity. In the next set of repeated lyrics, “some (people)” name “it” as TB and idliso ‘poisoning by witchcraft.’ In these lyrics, HIV/AIDS, which is “the thing that is like what”—the object being referred to by these stigmatizing terms—remains unspecified.

When the song reaches the last set of lyrics, it finally addresses HIV/AIDS directly, stating “How you have taught me through pain, you, AIDS, have taught me through pain.” Here ingculazi was the word I found most often used in isiZulu public discourse about HIV/AIDS (e.g. in medical brochures, magazines, and song lyrics). This term was reportedly not often used in daily conversation, however. For instance, when I was conducting pilot research for this project, an isiZulu-speaking doctor said that his children spoke about HIV in English because words like ingculazi felt like bad words. The song ends with the phrase, “whatever I do, they are talking about me.” This mirrors the beginning of the song, where it is “the thing that is like what” (HIV) that is talked about. Through this poetic parallelism, connections are indirectly made between HIV, stigma, and the singer or songwriter's experiences as HIV positive.

Avoidance and scientific medical terminology

The above excerpts do not give a clear picture of how most isiZulu-speakers actually talked about HIV/AIDS in everyday discourse, but they do provide data for understanding how some characterized discourse about HIV and for modeling the explanatory frameworks that were available to choir members. Research participants took pains to distinguish scientific medical terms for HIV/AIDS and related concepts from other Zulu explanatory frameworks of health and illness. For instance, I went with one choir member on multiple occasions when she visited a Christian faith healer. In healing sessions, that healer blessed large jugs of water. Patients were then directed to do one of several things with the water—to ingest it, to steam it and breath it in, or to ingest it and then force oneself to vomit it up. This choir member had told the faith healer, however, that she could not partake in any healing sessions that involved vomiting since this would limit the effectiveness of antiretroviral medication she was taking. A male choir member, for his part, was going to an inyanga, a Zulu traditional healer. He told me, however, that when he reached the point that he was supposed to start taking ARVs he would stop going to the inyanga because he knew it was bad to mix ARVs with other untested medications.

The most notable Zulu explanatory framework that was linked to linguistic avoidance of HIV was hlonipha ‘respect,’ which consisted of historically situated practices of linguistic and bodily avoidance rooted in the notions of respect and pollution. Traditionally, respect was an orienting principle of a social life that was oriented towards social duties and responsibilities (de Kadt Reference de Kadt1998). Fifty years ago, respect was communicated through a linguistic register that included specific respect words. For instance, socially subordinate speakers—especially but not exclusively young brides entering their husbands' homesteads—substituted respect words for many ordinary terms to avoid saying the names of people who were hierarchically superordinate, such as in-laws, elders, chiefs, and the Zulu royal family (Herbert Reference Herbert1990). In the early twenty-first century, these respect words and the linguistic process for creating them were all but forgotten by many isiZulu speakers. However, general respect practices based on comportment and some less specific kinds of linguistic avoidance in the presence of superordinate individuals were still prevalent (Finlayson Reference Finlayson1984, Reference Finlayson and Mesthrie2002).

Previous scholarly and journalistic work discusses how respect and avoidance impacted South Africans' ways of talking about HIV (Squire Reference Squire2007; Steinberg Reference Steinberg2008). Some work also suggests that older cultural practices are not to blame for the ineffectiveness of a biomedical response to HIV (Niehaus Reference Niehaus2007). Based on previous research and my own fieldwork experiences, I suggest that while these cultural (and linguistic) practices may not be a hindrance to treatment, they are a part of the ongoing process of stigmatization. In particular, avoidance of talk about an acquaintance's HIV infection may be tied to respect for that family's privacy in matters of health and illness (Wood & Lambert Reference Wood and Lambert2008). Here, it is significant that the parent-child relationship is a key dyad around which respect is organized. In my fieldwork experience, this made it very difficult for individuals to breach the topic of HIV with their parents or other elders, severely limiting possibilities for disclosure of one's HIV positive status in the family sphere. This situation is similar to other scholarly models of abjection. For example, in Tonga transgender individuals' possibilities for voicing their gender identity are limited by Tongan norms of respect and a prototypical brother-sister relationship (Besnier Reference Besnier1997, Reference Besnier2004). In Tonga as in South Africa, individuals negotiated exclusion through linguistic borrowing and association with the foreign, a topic to which I return below (Besnier Reference Besnier2004, Reference Besnier2011).

In addition to respect, fears about pollution also sometimes motivated isiZulu speakers' linguistic avoidance. Now-classic ethnographic work on Zulu health and illness indicate that when one was sick or exhibited other telltale signs associated with pollution, one stayed subdued and apart from others. Others, for their part, practiced physical and linguistic avoidance to mitigate the threat of spiritual pollution (Ngubane Reference Ngubane1977; see also Berglund Reference Berglund1976). A telling anecdote recounts resistance fighter, hero, president, and Nobel peace prize recipient Nelson Mandela's visit to a township outside of Johannesburg. On that visit, he noticed a house that the crowd surrounding him was conspicuously avoiding. He asked about that house and was told that two children living with HIV were living there. Upset, Mandela went into the home and spent time interacting with the children. When he exited the house, the crowd refused to get close to Mandela and he eventually retired to his car (Ashforth Reference Ashforth2005:154–56).

In their talk about HIV, Choir members' often voiced critiques about avoidance practices (seen in (2) and (3)) that relied on these alternative explanatory frameworks of HIV. Group members thought that terms discussed above like idliso were superstitious and words like iqhoks, iHlengiwe Ivy Vilakazi (seen in (1)), and other joking terms were demeaning. The alternative, which they developed through extensive training in support groups and health education courses led by dedicated social workers, health practitioners, and international aid groups, was English-language scientific, medical terminology. This alternative set of terms was especially significant because isiZulu translations of medical terminology are infused with additional meaning beyond the biomedical. For instance, TB is commonly translated into isiZulu as isifo sofuba, lit. ‘sickness of the lungs.’ Within traditional Zulu explanatory frameworks, however, isifo applies “to disease that is manifested by somatic symptoms, to various forms of misfortune, and also to a state of vulnerability to misfortune and disease” (Ngubane Reference Ngubane1977:22). The choice to codemix and use English medical terminology in isiZulu discourse made sense for a group of people who were activists hoping to guide others towards scientific medical understandings of HIV/AIDS.

Marking the foreign

My analysis of the isiZulu noun classes associated with choir members' terms for HIV/AIDS and the university students' discussed terms (example (1)) indicates that something else was also at stake though—namely a sense of inclusion or exclusion. While all languages have various conventions for categorizing people and other nouns, some (including isiZulu) have well-developed grammatical systems for doing so. isiZulu noun-class prefixes become significant in a number of different grammatical processes—for instance, in indicating the subject and direct object in verbal morphology (Gowlett Reference Gowlett, Nurse and Philippson2003:620). Noun classes, as they are often known, “become the representative points of mutual opposition… the fundamental ‘kinds of things’ for the speakers of a language, who map such total partition structures onto many other realms of experience as a language- and culture-specific set of ‘metaphors’ by analogy” (Silverstein Reference Silverstein1986:504). Nominal classification systems have been discussed in terms of the “ontological commitments” made by a particular system (Stephen Levinson, cited in Lucy Reference Lucy and Senft2000:335). In other words, examination of the nominal classification system of a given language has the potential to reveal how speakers may differentiate different types of entities. Noun-class systems, however codified, are nonetheless always potentially productive (Lucy Reference Lucy and Senft2000:333).

While there are fourteen distinct isiZulu noun classes, most of these are not easily associated with particular concepts or topics. As long ago as 1927, Bantu language specialist Clemente Doke (Reference Doke1927/1930) noted, “It is probable that in Proto-Bantu each class of nouns had a definite significance. With certain classes in modern Zulu this significance is still recognizable; with others it is lost. Class 1, for instance, is distinctly the ‘Personal class’” (Doke Reference Doke1927/1930:37). Indeed, in my experience with isiZulu speakers in 2008, noun classes 1 and 2 were still often used for human referents, especially for personal names. Furthermore, noun classes 5 and 6 were semantically productive for marking foreign-language lexical borrowings (Demuth Reference Demuth and Senft2000). That is, many bilingual English/isiZulu speakers that I spoke with used noun classes 5 and 6 whenever they code-mixed to incorporate an English noun they wished to use into isiZulu discourse. Note that noun class 6 is the plural of noun class 5.

Most, though not all, of the terms that research participants listed as words used to talk about HIV/AIDS were accompanied by noun classes 5 and 6 prefixes. Example (4) is a list of the words I recorded (uttered in isiZulu) with an indication of the noun-class prefix used.

1. (4) Words for HIV/AIDS and related concepts

   1. a. ama- gama amathathu = ‘a three letter word’ (refers to HIV acronym's three letters)

      nc6- word

   2. b. i- Hlengiwe Ivy Vilakazi = HIV (Common isiZulu name made from the acronym)

      nc5- Hlengiwe

   3. c. i- Slim = ‘Slim’ (HIV, referring to the wasting away of the body)

      nc5- Slim

   4. d. i- qhoks = ‘high heels’ (HIV/AIDS)

      nc5- [sound of high heels]

   5. e. in-  gculazi = ‘HIV/AIDS’

      nc9- AIDS

   6. f. isi-  fo sofuba = ‘Tuberculosis’ (literally sickness of the lungs)

      nc7- sickness

   7. g. i- dliso = ‘poisoning (by witchcraft)’

      nc5- poisoning

It seems unlikely that it was random that five of seven terms for HIV/AIDS that I recorded were prefixed with noun classes 5 and 6, especially since the two that did not use these prefixes were regarded as the medical translations of HIV/AIDS and tuberculosis. Given the general notion that HIV/AIDS was a European, American, or otherwise “Western” illness, especially, it is significant that the noun classes used in these avoidance terms were also the noun classes used when uttering any English noun in an isiZulu phrase.

The use of noun classes 5 and 6 for reference to HIV/AIDS was a utilization of a system of linguistic markedness to link HIV/AIDS with the foreign (Trubetzkoy Reference Trubetzkoy and Baltaxe1939/1969; Jakobson Reference Jakobson, Waugh and Monville-Burston1990; see also Caton Reference Caton1987). In past research, morphological noun-class systems have been discussed as systems of linguistic markedness and have been linked to the production of language ideologies (Silverstein Reference Silverstein1986). The concept of “marked” and “unmarked” linguistic forms refers to explicitly present or absent affixes or sounds. For instance, waiter, actor, and steward would be considered unmarked forms in English. Waitress, actress, and stewardess, which include an additional -ess affix to indicate the feminine form of the word, would be considered linguistic marked forms in English. Recent scholarship extends such understandings of markedness to a broader constellation of lexical and nonlexical features that index social identities, including prosody, word choice, accent, and even clothing and body posture (Bucholtz & Hall Reference Bucholtz, Hall and Duranti2004). For instance, it has been suggested that in many social situations in the United States, typically white, middle class, heterosexual male ways of speaking are “unmarked” (Bucholtz Reference Bucholtz2001). Applying such a model to the examples above, the ways that isiZulu terms were used to avoid of reference to HIV effectively marked the HIV as other and as foreign in particular. This fits with the other research that discusses stigma as a general social process of marking and of othering, including discrediting and distancing (Jenkins & Carpenter-Song Reference Jenkins and Carpenter-Song2008, Reference Jenkins and Carpenter-Song2009).

The semantic productivity of noun classes 1/2 for human referents and noun classes 5/6 for foreign borrowings also indicates the possibility of a concurrent overlapping system of markedness that dehumanizes those living with HIV. As an example of the opposition between classes 1/2 and 5/6, remember that in example (1) an interviewee mentioned a term for naming HIV with a common isiZulu name. There, she said, “iHlengiwe,” using noun class 5 (see example (2)). If she had been stating a person's name, she would have said, “uHlengiwe.” In using noun class 5, Sthabiso was uttering a set of Common isiZulu first and last names but using the foreign word-borrowing class 5 prefix “i.”

1. (5) Binary opposition between noun class 1 and 5

   i- Hlengiwe vs. u- Hlengiwe

   nc5- Hlengiwe nc1- Hlengiwe

In this opposition of noun classes 1/2 and 5/6, since classes one and two commonly identify a referent as human, the use of noun classes 5 and 6 for a human referent may be experienced as dehumanizing.

Even in the absence of a coherent noun-class system, the use of borrowed words to refer to HIV would have certain weight of its own, pointing to associations with foreignness. Noun classes, however, provide an additional level of categorization from which people can draw (Silverstein Reference Silverstein1986). The semantic productivity of noun classes 5/6 provides a formal analogue for use in the functional differentiation of HIV/AIDS as something foreign, something nonhuman, and something to be avoided.

Multiple ideologies of the foreign

The above discussion indicates that HIV stigma in South Africa may centrally involve an ideology of exclusion maintained through the interrelation of linguistic structure, language use, and metapragmatics (Silverstein Reference Silverstein and Clyne1979; Kroskrity Reference Kroskrity and Duranti2004). In other words, foreignness was marked through linguistic structure (use of noun classes 5 and 6), avoidance of overt reference to HIV/AIDS was practiced in language use, and metapragmatics (talk about how others talked about HIV) sometimes involved joking (e.g. “some joke about it”—choir lyrics) and other ways of indicating discomfort. Such an interpretation can be linked to a historically robust pattern of linguistic avoidance practices in which Khoi San languages, ideologically viewed as “icons of the foreign,” were incorporated into isiZulu hlonipha respect terminology (Irvine Reference Irvine, Schieffelin, Woolard and Kroskrity1998; Irvine & Gal Reference Irvine, Gal and Kroskrity2000). Similarly, I suggest that borrowings from English and the use of noun classes 5/6 marked HIV/AIDS as foreign and were central in a language ideology of exclusion.

Theorization of language ideologies, however, includes the notion that ideologies are often contested and there are always multiple language ideologies in play (Gal Reference Gal, Schieffelin, Woolard and Kroskrity1998). In the choir that I worked with, association with the foreign that was central in the constitution of exclusion was turned on its head. Choir members found support through contact with foreigners and white South Africans who were seen as outside the structure of Zulu society; and in facing stigma, they relied heavily on foreign language borrowings that they learned through a socialization process interacting with medical authorities at a local clinic (cf. Robins Reference Robins2006). The above examples also demonstrate the choir's alternative language ideology in which group members used English medical terms and concepts with one another, extending a support-group framework beyond the confines of the clinic and rehearsal space.

Choir members utilized scientific medical terms for referring to HIV in interactions with one another. Significantly, these terms were rooted in the same noun-class structures discussed above in stigmatizing reference to HIV. Group members, however, generally insisted on a scientific medical model of HIV, viewing infection with the virus as an illness that could be treated and understood through scientific medicine. Example (6), an excerpt taken from another interview conducted by Amahle, this time with neighbors—some of whom had disclosed only to Amahle—is an indication of how group members and others sometimes used scientific medical terminology when talking to each other about HIV/AIDS. Prior to the start of this excerpt, Amahle had asked Zandile about whether or not Zandile's “partner” knew about her HIV positive status. Zandile had responded with a story about her partner's initial resistance to the diagnosis. After this, Amahle asked the following question:

1. (6) Transcript of interview/conversation 9-11-2008, 3min 34sec.

   A = Amahle, Z = Zandile; ethnographer and two other female community members also present. Code-mixing indicated in bold.

   

In example (6), Amahle used the English term “HIV positive” to describe Zandile and her partner's diagnosis (line 1). In her response, Zandile used the common borrowing ikliniki (lines 3–4) and also used the English medical terms for ARVs (antiretroviral medication) and CD4 count (a measure of the number of white blood cells in a person's blood stream). One might argue that Amahle was “leading the witness,” in the sense that her linguistic choices led Zandile to respond in kind. Whether this is true or not, though, Zandile's seamless use of this medical terminology would not have been possible without a prior familiarity with these terms and ways of using them. Each of these instances of codemixing includes the use of noun classes 5/6: i- HIV positive (nc5), i- kliniki (nc5), ama ARVs (nc6), and i- CD4 count (nc5).

Example (6), alongside examples (2) and (3), provide examples of how choir members' and other' linguistic choices built upon the wide use of noun classes 5/6 for reference to HIV, but imbued such linguistic codemixing with alternative indexical valences. Group members constructed an ideology in which association with the foreign was part and parcel of support and activism—in particular, association with medical discourse and with people and organizations that were resource-rich in comparison to choir members and many other black South Africans. In this way, the exclusion of being associated with the foreign was reframed, at least in some social situations, as a powerful alterity—as association with the Otherness of economic and medical power. Choir members' past work with international aid groups provided the grounding for this ideological shift.