Keywords

8.1 Introduction: Learning Voice

In spite of strong and widespread beliefs to the contrary, people never learn ‘a language’. They always learn specific and specialized bits of language, sufficient to grant them voice—“the capacity to make oneself understood” by others (Blommaert 2005a, p. 255) . They learn voice by processes of enregistering semiotic forms—putting forms in a kind of order that generates conventionalized indexical meanings—and such processes of enregisterment involve complex and delicate orientations to existing or perceived norms (Agha 2007; Creese and Blackledge 2010; Jörgensen et al 2011; Juffermans and Van der Aa 2011 provide an overview and discussion) . We encounter the ‘glossic’ elements here that are central in Bakhtin’s work: specific and specialized resources that are formed by the social history of its users and can, when deployed, articulate social positions in interaction with other such ‘glossic’ resources . People’s repertoires , therefore, can be seen as an organized (‘ordered’) complex of semiotic traces of power: the semiotic resources they gathered in the course of their life are things they needed in order to be seen by others as a ‘normal’, understandable social being (Blommaert and Backus 2011) .

Such learning processes, as we know, develop in a variety of learning environments and through a variety of learning modes, ranging from the tightly regimented and uniform learning modes that characterize schools and other formal learning environments, to fleeting and ephemeral ‘encounters’ with language in informal learning environments—as when a tourist learns the local word for beer in a foreign country, returning home with a microscopic amount of foreign language along with the other souvenirs of the trip. Increasingly of course, the intensive use of online and mobile communication technologies opens a vast space of opportunities for such forms of informal learning, offering users access to vocabularies, registers, genres and styles, as well as cultural templates for practices (see e.g., Gee 2003; Leppänen and Piirainen-Marsh 2009; Varis and Wang 2011) .

The latter kinds of informal learning will be central to our concern here and there are several reasons for this. The first has already been announced: we see a tremendous expansion of informal learning environments and practices and an increasing number of researchers are directing their attention towards it. The second, the very nature of these modes of learning, prompts us to revisit learning as an activity; the Vygotskian framework in which learning is both socioculturally and historically contextualized and mediated through instruments, objects and worlds of reference, appears to have a second wind. And the third, there is the reflexive dimension, in which our own scholarly modes of learning become more relevant as themes for inquiry than perhaps before. Our own knowledge procedures are, in effect, mostly grounded in informal learning practices, especially when we engage directly with informants in the field (cf Blommaert and Dong 2010; Velghe 2011) , but similarly when we engage with people and their messages in the virtual world. So here is a case for taking informal learning seriously in an attempt to provide a more solid grounding for our own knowledge, and ultimately, our own voice.

In this chapter, we will focus on the way in which a woman we call Linda acquires, maintains and deploys a ‘ supervernacular’ (Blommaert 2011) , and how she does this in conditions of extreme marginalization. The supervernacular in question is a variety of ‘textspeak’ , a mobile phone texting code used in the Wesbank township near Cape Town, South Africa. As a variety of textspeak, the code used by Linda bears the usual characteristics of abbreviations, homophonic writing, emoticons and so forth; it is one of these extremely dynamic codes that characterize today’s new communicative environments. Linda, however, faces major problems: the macro-contextual circumstances of poverty, unemployment and social marginalization turn various forms of literacy into rare commodities; and to complicate things, her capacity for writing and reading is minimal; she is in all likelihood dyslexic. Notwithstanding these tremendous constraints, Linda uses textspeak intensively, drawing on an intricate scaffolding system for literacy usage she developed herself. In discussing the case of Linda’s use of textspeak, we will also have to consider the way in which the ethnographer’s own learning practices encountered Linda’s, and how this led to a new understanding of what textspeak is, and what it means in communities such as the one we investigate. Linda’s case, thus, compelled us towards reflexivity.

Let us start by preparing the canvas and provide some background about the research on which we draw here and on some of our conceptual tools. We will first look into the contextual factors that define Linda’s life: the township where she lives, her own background and the importance of mobile phones in her community. After that, we will turn to Linda’s own practices of learning and using textspeak.

8.2 Texting in the Township

8.2.1 The Field

Wesbank is situated on the dry and sandy Cape Flats, the so-called dumping grounds of apartheid , 27 kilometres out of the centre of Cape Town and surrounded by many other apartheid townships such as Khayelitsha, Nyanga, Crossroads and Delft. Wesbank is by all standards a very peripheral community (Blommaert et al. 2005; Newton 2008) , secluded and bordered by a highway, two very busy municipal roads and a wetland nature reserve, and located 12 kilometres away from the closest job opportunities. Although officially recognized and named as Wesbank, the name of the community is nowhere to be found, neither in local roadmaps, on traffic signs nor on the World Wide Web.

Wesbank was built in 1999 as part of the Reconstruction and Development Programme (RDP ), a South African socio-economic policy framework which the first democratic government in South Africa designed and implemented after the abolition of apartheid in 1994 in order to tackle the economic, spatial and racial legacy of the apartheid era and to improve government services and basic living conditions for the poor. The housing of the RDP aimed to provide 1 million subsidized houses before the year 2000, as a response to an ever-growing crisis in housing due to internal migrations from rural areas and homelands into the cities. The building of the Wesbank community was the first post-apartheid housing project in the area of Cape Town that was not segregated along racial lines but was intended to give home to deprived people, irrespective of colour and descent. This first so-called rainbow community had to give a home to 29,000 residents in 5,149 fully subsidized houses, reallocating people who had never owned a house before or who had been living in informal settlements for most of their lives. The actual number of residents in Wesbank is estimated to be much higher, as extended families live together on one plot, and people have built shacks in the backyards of the houses. Due to the socio-economic instead of racial criteria in the selection of the inhabitants, the population in Wesbank is very diverse (Blommaert et al. 2005) . An estimated 73 % of the population is “coloured” and Afrikaans speaking, 25 % is Xhosa and the remaining 2 % are Whites, Asians and foreigners coming from other African countries such as Zimbabwe, Congo and Somalia (Dyers 2008) . The houses have an average size of 25 square metres, are built with brick walls and corrugated iron roofs and are not insulated. Every house has a living room, a very small bathroom with a toilet and a washing table and one bedroom.

Recent unemployment rates for Wesbank are not available. The latest report dates from 2001 and mentions 60 % of unemployment amongst the economically active population. This figure even increases when considering women (70.4 %) and Black people (76 %) (Nina and Lomofsky 2001) . Although more and more people have found their way into informal sector employment and welfare systems, such as child support, disability support, support for the elderly, etc., one estimates that nowadays the unemployment rates are even higher. According to Newton, 77 % of those living in Wesbank have to survive on a monthly income of R400 (approximately € 40) or less (Newton 2008) .

Basic service delivery is minimal. Although two were planned, there is only one high school in Wesbank, insufficient for the number of teenagers in the area. There are three primary schools, although according to the official South African norm there should be five. For 3 years, Wesbank has had its own day clinic, but the clinic is only open for babies, children and patients with tuberculosis (TB) and human immunodeficiency virus/acquired immunodeficiency syndrome (HIV/AIDS). For the last 2 years, Wesbank has had its own taxi (private minibus) rank and a multipurpose centre. Apart from shebeens (illegal bars) and many informal small shops, there is only one (relatively expensive) supermarket in the community. Gangsterism and crime rates are very high. Nevertheless, the police station responsible for the area is about 8 km away.

8.2.2 Methodology and Data

The data for this chapter have been collected during two separate ethnographic fieldwork periods in the community of Wesbank, from January to May 2011 and November 2011 to March 2012, with a special focus on cell phone use and cell phone literacy amongst middle-aged womenFootnote 1. The study included face-to-face interviews with 20 different women, all lasting more than an hour and held in the houses of the people interviewed. Other data were collected by handing out questionnaires in the high school and one primary school in Wesbank. The questionnaire consisted of two parts, one part to be filled in by the learners and the other part by the (grand)mothers. Eighty out of 160 questionnaires were returned. Six interviewees kept a mobile phone diary , in which they noted all the text message and phone calls they made and received during the course of one week. Text and chat messages from residents, daily observations, informal conversations and fieldwork notes form the rest of the data used for this chapter.

Most of the data used here however, are instant chat messages between the second author and Linda, a 25-year-old female inhabitant of Wesbank, complemented by written words and sentences, collected during two writing and reading sessions the second author held with Linda. All chat conversations were held on MXIT, a very popular mobile phone instant messaging programme (for more information on MXIT, see Chigona et al. 2009; Chigona and Chigona 2008; also Deumert and Masinyana 2008) . A long interview with her mother and with Linda herself and observation of her texting behaviour also form part of the data.

8.2.3 Introducing Linda

Linda is a 25-year-old coloured and Afrikaans-speaking resident of Wesbank. She lives together with her 3-year-old daughter, her mother and her brother and little sister. Linda went to the high school in Wesbank, as she was 12 years old when her family moved to their RDP house in the community. As early as the first year of primary school, teachers expressed their concerns about Linda’s writing and reading skills. Those concerns, however, never caused specialized and individual follow-up, either at home or at school. Once in high school, Linda’s literacy level started causing serious problems. Linda could not absorb the graphic word images she was taught, and she could not read written texts. She did give evidence of having memorized certain word images and thus getting it right occasionally. If tested, there is little doubt that Linda would be diagnosed with a severe form of dyslexia.

During our fieldwork interaction with Linda, we got evidence of this. Linda’s literacy practices outside textspeak seemed to be even more restricted than her scaffolded textspeak repertoire. Asked if she could write down the days of the week, she could only, very laboriously, spell three of the 7 days ‘correctly’. During a job interview, she did not manage to simply copy words with pen and paper from a blackboard, and consequently missed out on the job opportunity. During the dictation exercise we did with Linda we observed several times that words that she could not get ‘right’ with pen and paper were typed out correctly in textspeak on her mobile phone. Textspeak offered Linda a way to circumvent her struggle with the sequence of vowels in a word, since textspeak abbreviations tend to mainly leave out the vowels. Môre (tomorrow) was spelled as mroe on paper and as mre with the keypad for example. Warm was written down as wram but typed out as wrm.

Linda’s spoken repertoire was, by contrast, relatively well developed. She could display multi-generic fluency in the local variety of Afrikaans—her ‘native’ language—and local vernacular English was also attested in her speech. We did note, however, that Linda’s fluency in spoken ‘monoglot’ English, was remarkably low, compared to other subjects in Linda’s age category in the township. Linda had problems sustaining an ‘English-only’ conversation, for instance, and felt compelled to switch back and forth between her variety of Afrikaans and vernacular English in such events. Linda’s mother, at that time involved in the organization of adult education in the community, forced her daughter to follow extra literacy classes after school hours. For 4 months, Linda did follow the extra classes, and both she and her mother had the feeling that she was benefiting from the extra attention. Her marks at school improved and she struggled less with reading and writing. When her mother had to stay at home because of pregnancy, however, Linda lost interest and stopped attending the courses.

At school, Linda tried to manage and keep up with the help of her friends, who would read things out for her and correct her writing as much as possible. Frustrations and a loss of motivation, however, made her drop out of high school before reaching the final matriculation year. Linda has not had a real job since then. She sometimes gets interim jobs for a couple of days, but she never manages to keep these jobs for a long time. Lately she decided to follow her mother’s example, applying for a home caring course, and she is now waiting for an answer to her application. Currently she sometimes replaces her mother in her home caring duties, taking care of patients her mother used to take care of.

8.2.4 Mobile Phones in Wesbank

Now that Linda has been introduced, let us have a look at the general patterns of use of mobile phones in an area such as Wesbank. Mobile phone penetration in South Africa is the highest on the African continent, standing well over 100 mobile phones per 100 inhabitants (ITU 2010). This high uptake rate has been confirmed during fieldwork in Wesbank. Of the people who filled in the questionnaires, 83 % possessed a mobile phone; only a small 3.4 % of those questioned had never possessed a mobile phone. The remaining 13.6 % had had a mobile phone before, but it had got lost, stolen or broken, while financial circumstances precluded buying a new phone. Only one out of the 20 interviewees still had a (prepaid) landline connection in the house, and was using it in combination with a mobile phone. Others had never had a landline or had cancelled the connection as soon as they got a cell phone.

The high uptake of cell phones is an example of how even in impoverished areas like Wesbank, people with modest means manage to take part in the new communication environment. Thanks to the marketing of very basic but cheap mobile phones, the introduction of prepaid non-subscription plans, the caller party pays system (Minges 1999) and the possibility of purchasing airtime in very small amounts, nowadays even the people at the bottom of the income pyramid have and use mobile phones. For the first time in history, they can take part in the telecommunications society, which, according to Castells, Fernandez-Ardevol, Linchuan Qiu and Sey (2006), dramatically changes the ways people communicate.

The general impoverishment of Wesbank seems not to be an obstacle to having and using a phone. Asked for negative consequences of having a cell phone, only 10 % of the interviewees and people who filled in the questionnaires mentioned the extra financial burden a cell phone generates. Impoverishment and financial constraints do, however, very clearly influence the use of the phone. People try to limit and control their cell phone costs in all kinds of ways. Many residents have more than one subscriber identity module (SIM) card from different providers; depending on whom they are calling and what time of the day or the week it is, they switch SIM cards to cut costs. ‘Please call me’ messages, a free service offered by all mobile phone carriers, allows sending a short message service (SMS) text message when you run out of airtime to any other cell phone number, with a request to call back. Those free messages, often a daily limited amount of them, read ‘please call me’ and feature the cell phone number of the sender. For many interviewees, sending a ‘please call me’ message was the only thing they could do on their phones.

People top up for very small amounts of money. Top-up cards for R5 and R10 (€ 0.5 and € 1, respectively) are the most commonly purchased vouchers. Airtime is only purchased when there is money available and often lasts for only one or two short calls. Between two top-ups, residents seek recourse to the ‘please call me’ messages and the free messages or call minutes that mobile phone carriers sometimes offer. Exploration of features on the phone that are not free of charge is very limited. Voicemail is hardly used, as no one wants to run the risk of calling someone in vain and having to pay for it. Mobile Internet is hardly explored, out of both financial constraints and Internet illiteracy . Although most of the interviewees have never ‘seen’ the Internet, mobile or not, or have even never used a computer , they all have high but very unspecific expectations about what the Internet can bring them with regard to information, help, job and other opportunities.

Mere device illiteracy among the middle-aged women interviewed and questioned is very high. Eight out of the 20 interviewees did not know how to send a text message and one-third of the youngsters questioned said they would like to teach their parents how to send a text message. The interest in ‘cell phone courses’ to learn to work with basic features such as sending text messages was high among most of the residents interviewed.

8.3 Learning a Supervernacular as a ‘Substitute’ Language

Now that we know the setting and the scenery, we can take a closer look at the ways in which Linda acquired textspeak, and how she uses it with her friends.

8.3.1 Linda ‘can’t read and write’

Sitting bored and jobless at home, Linda spends most of her time on MXIT, chatting with friends from inside and outside the community. Friends of hers introduced her to the instant messaging programme in 2011, after a friend gave her a mobile phone as a present. Her mother complains that Linda is literally day and night on MXIT, ignoring the domestic tasks that she is in charge of and forgetting about the fact that she is supposed to look for a job.

Due to negative press coverage over the years and to the urban legends spawned by this, MXIT has a bad reputation, especially among parents. People stigmatizing MXIT connect the programme to potential substance addiction, abuse, antisocial behaviour, adultery and exuberant sexual behaviour. They regard it as a free zone for unsafe behaviour, rudeness and pornography, and express fears that the textspeak used on these communication platforms will ‘pollute’ the youngsters’ capabilities to read and write in ‘standard’ varieties of their language. As MXIT is the cheapest way of communicating over the phone (chatting on MXIT is significantly cheaper than SMS messaging, respectively 0.01 ZAR and 0.80 ZAR per message), the instant messaging service has become the most important means of digital communication for most youngsters (and adults as well) in South Africa and one of the most important and time-consuming leisure activities. By October 2011, there were more than 10 million active MXIT users in the country (World Wide Worx and Fuseware 2011).

Linda’s friends introduced her to MXIT; they assisted her in downloading the application on her phone and are still assisting Linda when it comes to her reading and writing on the instant messaging programme. We observed that Linda’s textspeak is done in a local vernacular variety of Afrikaans with frequent shifts into English and an abundant use of emoticons. During the first weeks of chatting on MXIT, Linda constantly carried a piece of paper with her on which her cousin wrote down the most common abbreviations, emoticons, contractions and number homophones used in textspeak.

Linda’s use of literacy on MXIT is obviously scaffolded, and we will return to this in greater detail below. At this point, we can observe that Linda was not an ‘autonomous learner’ of textspeak . Her learning trajectory was collective and proceeded with the vigorous support and intervention of several friends. Other people ‘gave’ knowledge and skills to Linda, for her to use in specific ways. We are observing here a Bakhtinian phenomenon in its purest form: the development of a ‘glottic’ resource through social and collective forces. The process of learning we observe here is also a Vygotskian phenomenon in its purest form. Textspeak itself is not an isolated object of learning, but in Linda’s case, its acquisition went hand in hand with the further development of pen-and-paper writing and reading. This point will be developed further below, and it is interesting for several reasons.

Linda grew up in circumstances and areas that are hardly stimulating for literacy practices. In areas such as Wesbank, it is very rare to find reading material in the house or for people to read in their free time. Asked if they read or write in their leisure time, 66 % of the women who returned the questionnaires claimed to read sometimes, but more than half of those only read the Bible every now and then, mostly in Afrikaans. Newspapers and magazines were not commonly read, and ‘Die Son’, a sensational tabloid with a lot of pictures, was the most popular newspaper among the middle-aged women interviewed. Of the youngsters who filled in the questionnaires, 70 % said they “sometimes read in their free time” but only 10 % of them claimed to have read something the day before they filled in the questionnaire.

Compared to reading and writing on paper, a lot more reading and writing is done on the mobile phone, especially amongst youngsters. In spite of the moral panics and public anxieties (see Vosloo 2007 for a critical approach on the effects of texting on literacy) , people in the new communicative environment shaped by information and communications technology (ICT) are reading and writing more than ever before. Text messages, instant messaging, chatting, blogging, tweeting, Facebook, etc. all form platforms of literacy and literacy acquisition, although research has shown that most do not think of their electronic or digital communications as ‘real’ writing or reading (see Lenhart et al. 2008) . The answer in one of the questionnaires “I don’t read or write, I’m always on MXIT” demonstrates this traditional printed and pen-and-paper-centred view of literacy practices. If one spends the whole day on MXIT, one is actually reading and writing non-stop, immersing oneself in a literacy environment that probably would have been much more limited without the existence of the mobile phone. According to Banda (cited in Deumert and Masinyana 2008) , SMS writing constitutes an important form of everyday literacy in South Africa, especially in the metropolitan areas.

This is the local world of language, knowledge and meaning in which Linda’s practices develop and make sense. We now begin to see her as a special case. In Linda’s practices, we see that textspeak is not separated from pen-and-paper writing, but that both forms of literacy development proceed in parallel. Since she started being active on MXIT, her pen-and-paper writing also increased and acquired specific functions, thus creating a more complex, intertwined and layered literacy learning environment in which pen-and-paper literacy is a critical support infrastructure for textspeak. Textspeak, consequently, is not something that is harmful to her ‘ordinary’ writing skills; it actually stimulates and expands them within the limits of her capacity. For someone who was qualified as near illiterate due to her disability, textspeak proved to be an instrument for considerable progress and self-development. Whatever she possessed in the way of reading and writing skills was mobilized for it, and it was developed and made more specialized in it and through the collective efforts of a group of peers. The informal learning environment provided by MXIT, thus, appeared to provide motivation to learn and relearn, as well as an efficacy of learning practices, which Linda had never encountered at school.

8.3.2 Linda’s Learning and Scaffolding Practices

We noted above that what we observe here is a purely Bakhtinian and a purely Vygotskian phenomenon. Let us dig somewhat deeper into this issue. Linda claims to never read or write in her free time and to never having done it, except for the things she currently writes down in relation to her MXIT activities. Since the first day she was on MXIT she started to write with pen and paper as well, copying words and sentences from her chat partners and when asking writing advice from her friends. All over the house, papers and notebooks can be found, on and in which Linda has taken ‘textspeak notes’, writing down status names and sentences she might use in the future. In this way, Linda has collected a ‘corpus’, so to speak, of copied words, expressions and phrases both drawn from MXIT and prepared for use on MXIT. Given the important place of MXIT in her social life, this corpus is the main instrument by means of which she is capable of sustaining relationships with people in her network—it is crucial social capital for her; lacking it would result in a strongly reduced social life for Linda.

Observe what happens here. Linda copies the visual images of words and expressions on MXIT, and later copies these visual images back onto MXIT. The copying is a graphic enterprise, in which Linda attempts to provide a precise visual replica of the forms she intends to copy. The meaning of those forms was very often explained by her friends who read them out to her. Having established the sense of these forms, Linda applies herself to copying them as visual signs into her ‘corpus’ of usable MXIT signs. She has to remember what these signs stand for, because the usage of the signs in MXIT interaction has to obey pragmatic rules of appropriateness. Linda has thus managed to construct some level of communicative competence, enregistering certain forms as meaning this-or-that and using them more or less appropriately in interactions.

More or less, we say, because Linda does not always get it right. As we will see below, she sometimes scrambles the visual image of MXIT expressions from her corpus, and she adopts particular tactics of pragmatic deployment when she runs into communicative trouble. The literacy skills she has developed through and around the use of MXIT are therefore fragile and elementary, and they compensate her constraints on reading and writing only to a certain extent.

This became clear when we did a dictation exercise with Linda. Asked to write the word ‘week’, she wrote ‘weender’; this happened three consecutive times. She then read what she had written as ‘week’. When asked where the ‘k’ sound was in ‘weender’, she could not answer and was in fact surprised to see that what she pronounced (‘week’) did not at all correspond to what she had written (‘weender’). Other tasks included writing down ‘unknown’ language bits (from French); there too we observed severe problems in converting sounds into symbols, and in converting the symbols into sounds afterwards. Figure 8.1 shows a page from the dictation notes.

Fig. 8.1
figure 1

Linda’s dictation notes

Full size image

Linda’s literacy skills are thus not generative. She has no control over the logic of orthography and over the functions of literacy in relation to spoken language; she cannot improvise and innovate in writing. Her writing resources form a tightly closed package of copied forms, the meaning and function of which have been memorized.

When Linda engages in MXIT interactions, consequently, she copies standard “passe partout” phrases and expressions and sends them off. She asks standard questions such as “wat maak jij?” (‘what are you doing?’) and is able to reply to such predictable and ‘phatic’ questions by means of routine answers. This can go on for quite a while, and it satisfies the requirements of interaction in many instances. Her illiteracy is masked rather than compensated by her scaffolding practices, but by masking it, she can and does appear to be a competent user of MXIT.

8.3.3 G2G (‘Got to go’)

The possibilities and limitations of Linda’s literacy repertoire became clear during an episode in which she chatted with Fie Velghe. On MXIT, like on other social media platforms, members make a status profile, often a slogan or motto. Linda changed her status profile daily (another sign of her desire to be perceived as a competent user). That day, her status update read:

WU RUN THE WORLD GALZ… WU FOK THE GALZ BOYZ… LMJ NW HOE NOW::op = csclol = @

We see that Linda has clearly, and accurately, copied part of this phrase, using vernacular and heterographic English code: “wu run the world galz … wu fok the galz boyz” (‘who runs the world? Girls … who fucks the girls? Boys’). What follows after that, however, is considerably less clear and looks rather like a random compilation of signs: “LMJ NW HOE NOW::op = csclol = @”. An ethnographer is always a learner of the practices s/he observes, and Fie has been deeply immersed in informal learning processes of Wesbank textspeak (see Velghe 2011). Here as well, Fie inquires into the meaning of what she received. Let us have a look at the MXIT interaction of that day. The exchange is in the local Afrikaans-based code; English translations are provided between brackets; untranslatable and erratic items are italicized.

figure a

Several intriguing things happen in this short interaction. Fie opens with a general question (how are you?) and receives a standard answer. Her next question, however, probes into the cryptic meaning of the status message, followed by a repeated routine question (everything all right?). After 2 minutes have passed (a marked pause in an instant messaging environment), Fie repeats her probing question (wmj?). The reply she receives, however, answers her previous routine question, and it contains a cryptic form siu (probably ‘sure’). Fie spends two turns providing information about her whereabouts and the day’s programme, and then inquires about the meaning of siu. Linda eventually responds with wt, ‘what?’. Fie reiterates the nature of her inquiry and gets an ‘oh, ok’ back, but this is followed by another standard question from Linda: ‘what are you going to do today?’. Fie questions another expression (csclol) and answers the question from Linda, upon which Linda finally replies to the probing questions: “x wieti” (I don’t know). This prompts Fie to again clarify the nature of her questions—‘I am curious’—but this provokes a status message ‘Linda is now busy’, followed by Linda writing ‘g2g’ and effectively going offline.

The general pattern is clear. Linda is quite fluent in asking and answering routine questions (how are you? Where are you? What are you doing?), but quickly reaches the limits of her literacy resources when different questions are being asked, requiring creative and non-routine answers. In this case, the questions were directed at features of Linda’s MXIT writing, and they were asked in a fieldwork mode, in an attempt by Fie to learn what might be new features of the cryptic Wesbank textspeak. Linda’s responses express no comprehension either of the object of the questions (her own erratic writing samples) or of their purpose. Answers are bland and superficial: ‘ohk’, ‘ja’—the sort of answers that can be given to almost any question. Fie’s insistence, however, causes her to withdraw from the interaction. These questions extend far beyond the limits of Linda’s resources, and withdrawal is the only available instrument to avoid exposure of her limited skills.

8.3.4 Two Modes of Learning and Usage

There is a clear dimension of inequality in the interaction between Fie and Linda—the first one having the fully developed and generative literacy skills that are missing from the latter’s repertoire. It is, consequently, easy to overlook an important dimension of the exchange: the fact that it is an exchange between two ‘non-fluent’ users of the textspeak code. The nature of these two forms of non-fluency is, however, fundamentally different.

Fie’s non-fluency is an effect of her learning trajectory as an ethnographer (cf. Blommaert and Dong 2010) . She has been involved in learning processes in the field, of skills and resources that must enable her to conduct fieldwork on language and through language in Wesbank. This learning trajectory involves learning the local vernacular varieties of Afrikaans and building a degree of spoken and written fluency in it. It also involves acquiring a degree of fluency in the more specialized textspeak vernacular derived from this variety of Afrikaans—a matter of learning a highly peculiar form of literacy, in other words. This complex learning trajectory is made easier because of the similarities between Fie’s native language—Dutch—and Afrikaans. Fie can ‘mask’ her apprentice status in Afrikaans by drawing on vernacularized forms of Dutch, and we see traces of that in Fie’s turns in the MXIT exchange. The use of ‘universiteit’, for instance, is understandable to interlocutors but would, in their register, more likely become ‘varsity’. Fie’s competence in Dutch, thus, provides a degree of elasticity in the levels of understandability of what she produces. Even if her expressions are Dutch rather than Afrikaans (and thus locally dispreferred or marked), they would still be understandable to most interlocutors.

We also see a degree of elasticity in Linda’s writing. Some of her writing errors pass as understandable because their visual image is close enough to the correct form to be understandable and occurs in a preferred interactional slot (as when ‘x wieti’ is understood as ‘I don’t know’). The difference is of course in Linda’s repertoire, which does not contain the generative skills that would enable her to construct a potentially infinite range of new forms producing new meanings, and to decode and understand a potentially infinite range of incoming utterances by others. The elasticity in her writing, consequently, operates on an infinitely smaller range of signs and forms than in the case of Fie. Her repertoire , thus, enables her to perform a restricted set of communicative practices both productively and receptively, and within that narrow bandwidth she can appear as a fully fluent user. It is on the basis of this restricted set of practices and the perceived fluency she enacts in them that Linda is capable of building and sustaining a large network of friends, and to engage in intensive forms of interaction with them. The very small repertoire she possesses, thus, offers her voice .

However, note that the learning trajectories of both Linda and Fie are collectively scaffolded ones—Bakhtinian and Vygotskian, as we noted earlier. Linda’s scaffolding practices have been discussed above. As an ethnographer, Fie equally had to rely on informal collective modes of acquiring and sharing the knowledge and skills she put into her MXIT communications (see Velghe 2011 for an elaborate case study) . We see how an ethnographer gradually builds ethnographic knowledge through situated activities in which others point things out for her, in which she tries to apply them, is corrected or rejected, tries again and so forth—a very Vygotskian epistemic trajectory of deeply situated and mediated learning in which perpetual adjustments need to be performed to the local economy and ecology of meaning. And this learning process is, like that of Linda, targeted at voice, at making oneself understood as an ethnographer in relation to other people; as someone who, step by step, is able to translate, so to speak, the differences between social positions, backgrounds of knowledge and repertoires. We see the practical epistemology of ethnography in full detail here, the sometimes conflicting meeting of different habituses (cf Bourdieu 2000; Wacquant 2004; cf Blommaert 2005b for a discussion) ; and this practical epistemology is, of course, in its very nature heteroglossic.

8.4 Learning Voice

In a superb and moving paper, Charles Goodwin (2004) described the case of a man who due to a stroke had lost almost all of his capacity to speak. Chil—the name of Goodwin’s subject in this study—had a repertoire reduced to just a handful of crude signs: sounds and groans. These signs he would, however, deploy actively and (as Goodwin showed) masterfully in interactions with others, so that this extremely restricted repertoire made him understandable to people accustomed to interacting with him, and turned him into a ‘competent speaker’. Chil had ‘ordered’ his small set of resources in such a way that they made sense to others. Thereby he defied a strong cultural assumption used in everyday life as well as research, described by Goodwin as that of

an actor, such as the prototypical competent speaker, fully endowed with all abilities required to engage in the processes under study. Such assumptions both marginalize the theoretical relevance of any actors who enter the scene with profound disabilities and reaffirm the basic Western prejudice toward locating theoretically interesting linguistic, cultural and moral phenomena within a framework that has the cognitive life of the individual as its primary focus (Goodwin 2004, p. 151)

Goodwin (following Bakhtin, Voloshinov and Vygotsky) makes the powerful point that language and interaction are often collaborative, with people drawing on others’ linguistic and cultural repertoires to make sense. Chil is just an extreme case of a general category, that of

a speaker and (…) a human social actor whose competence does not reside within himself alone but is deeply embedded within the actions of others as they build lives together” (Goodwin 2004, p. 167) .

The case of Linda, a young woman who lives in a marginalized community and manifestly suffers from a severe case of dyslexia, supports Goodwin’s crucial observation. Linda depends on a very small range of signs in a ‘truncated’ repertoire; this enables her to perform a limited set of interactional practices, in which she tries to be perceived as a competent user by copying standard phrases, interspersed with often erratic and non-routinized writing. The fact that this works well for her—MXIT chatting is one of the most important activities in her life—testifies to the fundamentally collaborative nature of social interaction. As we have seen, many of her friends are fully aware of the grave literacy limitations experienced by Linda—they effectively scaffold her interactional work by providing her with texts to copy, read them for her and grant her writing the degree of elasticity in understandability we discussed earlier. Linda thus draws on the repertoires of others to achieve her social goals, and what is usually called ‘communication’ may better be called ‘conviviality’ here.

Within her network of friends, Linda is accepted as a ‘fully competent’ member. The reason is that her messages are less seen as linguistic objects than as indexical ones, not as carriers of intricate denotational meanings but as phatic messages that support Linda’s role as a group member and define her relations with her peers as agreeable and friendly. It is only when an outsider comes across—an ethnographer inquiring into the nature of Linda’s textspeak here—that Linda’s messages become linguistic again: they are now suddenly measured not by the standards of the indexical order of conviviality, but by the standards of language and orthography. We have seen how rapidly this volte-face meets the limits of Linda’s competence, and how quickly it triggers silence.

Two conclusions can be drawn from this. The first one is about the nature of the sign system that Linda deploys. As said, it is best not to see this sign system in terms of linguistic structure and functions, but in terms of indexical ones. Linda’s use of textspeak is not primarily a use of ‘language’, it is a deployment of voice—of a sign system that opens channels of peer-group communication and conviviality , and establishes and confirms Linda’s place in her network of friends (see Blommaert 2008 for elaborate examples of grassroots writing having similar functions) . Linda invests tremendous amounts of time and effort in learning and using these signs, not because these signs enable her to express denotational meaning (we have seen the limits of her generative writing and reading skills) but because they are a crucial and essential social instrument for her, one of the few very valuable instruments she possesses to make herself recognized and respected as a human being, a “human social actor” in Goodwin’s words. Linda’s voice is shaped, both in its affordances and its constraints, by the social conditions for voice to which she is exposed, at all levels—from the extreme poverty in her community to the severe dyslexia she has to cope with individually.

The second conclusion is one of methodology. It would be good if researchers would pay attention to the indexical functions of sign systems such as the ones discussed here, and approach them primarily as instruments for voice rather than as phenomena of language and literacy. Fie’s apprentice habitus in the field directed her towards the linguistic and orthographic features of textspeak, as we have seen. The failure of this line of inquiry was a case of fieldwork serendipity which provoked a sudden clash of habituses and hence a change of perspective, forcing Fie to reflect on entirely different dimensions and functions of Linda’s textspeak. A ‘mistake’, so to speak, in ethnographic inquiry proved to be a very productive line of inquiry in its own right (for a similar and inspiring case, see Fabian 1990) . Encountering the limits of Linda’s linguistic and literacy skills raised issues of what this restricted code represented for Linda. The answer was voice.

This line of inquiry may be of critical importance for understanding the vast (and expanding) array of new sign systems that emerge in the field of literacy in the context of globalization and the use of new media, including the supervernaculars of textspeak and chat codes (Blommaert 2011) , and in affluent and technology-saturated environments as well as in marginal and technology-poor environments such as Wesbank. People often enter the new social arenas of today’s world of communication with very limited resources—in fact, it is safe to see this as a rule. These resources may or may not develop into full-fledged normative code varieties; various specific factors will determine such processes, and their development needs in any event to be seen and understood in relation to the local sociolinguistic economy. The resources are ‘glossic’ in the purest Bakhtinian sense. However, in every stage of their existence, they will be deployed in attempts to provide voice to their users, to make their users identifiable and recognizable as such-and-such a person, and so establish them as members of communities and networks. Sign systems are always emerging and rarely ever fully stable; their function as instruments for voice, however, remains a constant throughout the rapid processes of change and development in their repertoires.