Emotions and HIV / AIDS in South Africa : A multilingual perspective

This paper argues that in order to gain a more informed perspective on emotions and HIV/AIDS, crosslinguistic differences in emotion language need to be taken into account, particularly in a multilingual context. The paper reviews four published academic articles with the aim of illustrating how more consideration of the crosslinguistic and multilingual aspects of emotion language could have contributed to better theoretical understanding of HIV/AIDS and emotions as well as aiding the development of practical interventions in HIV/AIDS counselling and care. Finally, this paper argues that a lack of engagement with language and multilingualism around HIV/AIDS and emotions will stifle the development of a theoretical account of emotions as multi-semiotic and embodied, as well as the development of locally based, community-driven practical interventions.


Introduction
South Africa is commonly regarded together with India to have the largest number of people living with HIV/AIDS and the largest infection rate.In South Africa approximately six million people are living with HIV/AIDS, thus more than 10% of the population (Lambert and Wood 2005;Dageid and Duckert 2008).A number of reasons have been put forward for the high infection rate, such as badly designed HIV/AIDS prevention campaigns, cultural beliefs and practices, poverty, and stigmatization surrounding HIV/AIDS and sexual practices in general (Crew 2002;Bok 2009.)Robins (2004:653) argues that responses to HIV/AIDS both by the state and citizens in South Africa are tied up with issues of race and identity.Illustratively, Robins (2004:654) makes reference to Thabo Mbeki's flirtation with AIDS dissident scientists, and his public statement about the links between HIV/AIDS and poverty which, he believes, should be read "as the products of historically constructed and politically driven processes embedded in specific histories of colonialism, apartheid and capitalism."Current South African president, Jacob Zuma, has also made controversial statements about http://spilplus.journals.ac.zaHIV/AIDS (e.g. the shower statement during his rape trial in 2005) which have been taken up in debates not only about HIV/AIDS but also about gender and race relations.Thus discussions on HIV/AIDS have often stood proxy for a number of other social issues in South Africa such as poverty, apartheid, race, capitalism and gender.Even though HIV/AIDS has been debated in relation to social and historical structures and processes, language and HIV/AIDS has been a neglected topic (see Anthonissen and Meyer 2008;Deumert 2010;Saal 2011 for exceptions) in South Africa.
The importance of language in healthcare contexts and in particular in HIV/AIDS discourse and communication is however increasingly receiving attention elsewhere (see for example the special issue on language policies and health in Language Policy (2010), and Higgins and Norton's (2010) contribution on language and HIV/AIDS).Language is not only acknowledged to be important in doctor-patient consultations and HIV/AIDS counselling, but also in the distribution of material that gives information about HIV/AIDS prevention and living with HIV/AIDS (see for example Saal 2005Saal , 2011)).In a multilingual context, such as South Africa, the healthcare sector is characterised by diversity in terms of language, culture, race, religion and major socio-economic differences.In communication encounters around HIV/AIDS the doctors, caregivers or counsellors often do not share a first language or culture with their patients (Anthonissen and Meyer 2008).Reliance on English (and other languages such as Afrikaans) as lingua franca and the use of (formal and informal) interpreters are therefore paramount in facilitating understanding.Such linguistic diversity is usually constructed as negative as it can have consequences such as miscommunication, misdiagnosis and lack of access to adequate healthcare (Deumert 2010).The possibility that linguistic diversity and multilingualism can have positive effects in the context of HIV/AIDS communication have, however, not been seriously investigated.
Not only is HIV/AIDS communication in South Africa complicated by linguistic diversity, communication about the disease is also usually emotion-laden, as is often the case in situations with life-threatening or chronic diseases (Cleirigh, Ironson, Antoni, Fletcher, McGuffey, Balbin, Schneiderman and Solomon 2003).It has been shown that emotionality influences understanding and interpretation of verbal and non-verbal messages, adherence to medication, attitudes to the disease (Murphy, Johnston, Roberts, Hoffman, Molina, and Lu 2003), and long-term emotional coping (Cleirigh et al. 2003:225).In a paradigm where the construal of emotions is seen as universal, misinterpretations of emotions due to linguistic or cultural differences will be rare.However, increased attention to language and diversity and the revival of the once-denounced Whorfian hypothesis have focused attention on the possible ways in which differences in language can lead to differences in conceptual processes.Colour perception, motion event cognition, the processing of spatial relations and the construal and interpretation of emotions are examples of domains that have all received attention in this neo-Whorfian paradigm.The findings from these studies point to the complex interplay between language and conceptualisation (Cook and Bassetti 2011;Pavlenko 2011).The focus of enquiry in this research paradigm has moved away from providing evidence for or against the position that language determines thought, instead the focus is on the ways in which and the extent to which language might sculpt or influence cognition.There is some evidence already that language does have an influence on cognition.Empirical investigation on colour perception for example shows that "speakers of a language with a single term to refer to blue and green will judge blue and green stimuli as more similar than speakers of a language with a lexical blue/green distinction" (Athanasopoulos 2011:30).Naturally, this does not mean that the speakers of the language with no lexical blue/green distinction cannot perceive the difference between the two colours, but that they attach less importance to that distinction.
With increased attention to multilingualism, research has also shown that these cognitive processes (e.g. the perception of colour), do not stay stable through life but might be altered by the addition of another language.A growing body of research shows that languages across the world vary considerably not only in their linguistic construal of emotions, but also in the way that their speakers understand, interpret, and react to emotion words and emotion-laden events (Wierzbicka 1992;Mesquita and Frijda 1992;Pavlenko 2005Pavlenko , 2008)).Furthermore, recent research has shown that multilingual individuals may exhibit different conceptualizations of emotions in comparison to monolingual speakers (Dewaele 2004a(Dewaele , 2004b;;Pavlenko 2008).These findings are apparently of utmost importance for research on HIV/AIDS and emotions in a multilingual context such as South Africa, as they show not only possible crosslinguistic differences in the way emotions are expressed, experienced, practiced or understood, but also differences between monolinguals and multilinguals.A multilingual understanding of emotions in HIV/AIDS discourse has various theoretical, methodological and practical implications.Through the study of HIV/AIDS and emotions in multilingual contexts, we can possibly increase knowledge about how language and emotion, and in particular emotions and multilingualism, might be interlinked and dependent on each other.A multilingual approach also provides an opportunity to critically reflect on the methodologies that are currently used in studies on HIV/AIDS and emotions.These theoretical and methodological implications will, of course, affect practices surrounding HIV/AIDS such as counselling, consultation, the creation of prevention material, and reporting on the lived experiences of people living with HIV/AIDS in South Africa.An understanding of the specific lived experiences of people infected with and communities and individuals affected by HIV/AIDS, informed by linguistic diversity and multilingualism, could make for localised and community-driven intervention strategies, not only in terms of HIV/AIDS prevention, but also in terms of how people who are infected and affected by the disease express voice and agency.Previous studies have shown that the use of "locally situated intervention strategies" is extremely effective, such as the efforts of the Treatment Action Campaign (TAC) to make antiretroviral medication (ARVs) available in state-run South African healthcare facilities (Robins and von Lieres 2004:580).A crosslinguistic and multilingual perspective of emotions would clearly insert local meaning and understanding of HIV/AIDS and emotions into current research and discourses.
Consequently, the aim of this paper is to draw attention to the possible implications of crosslinguistic and multilingual perspectives of language for research on the expression and interpretation of emotions in an HIV/AIDS context.We will review studies conducted on emotions and HIV/AIDS generally and in South Africa, with the intention of showing how a multilingual perspective could possibly have served to enhance these studies or to problematize their findings.The scope of the paper is circumscribed to the linguistic and cognitive aspects of verbal and non-verbal expression of emotion-related events and behaviour.In our discussion we will draw extensively on Pavlenko's (2008) work on multilingualism and emotions. http://spilplus.journals.ac.za

Critical review of previous studies: Emotions and HIV/AIDS
This section will review four papers (Demmer 2006;Thomas 2007;Bar-Lev 2008, andDageid andDuckert 2008) in which the expression of emotions by people infected with and affected by HIV/AIDS was either the central focus, or one of the explanatory variables in the analysis.The review is restricted to four papers only since there is currently a dearth of studies on HIV/AIDS and emotions in general and even more so in the South African context.The four papers differ somewhat with respect to the contexts studied: Two of the studies were conducted in South Africa (Demmer 2006;Dageid and Duckert 2008), whereas Thomas's (2007) investigation was conducted in Namibia.Bar-Lev's (2008) study, in contrast, focused on a virtual community and is therefore not specific to a particular geographical region.Firstly, general information about the studies is given and secondly the findings of the studies are framed around current research on crosslinguistic and multilingual perspectives of emotions and what such a perspective might add to knowledge about HIV/AIDS and emotions.

General information about the studies reviewed
To begin with, we wish to point out that all of the studies selected for review are conducted within disciplines outside of discourse analysis, applied linguistics or communication studies.In other words, these studies did not specifically set out to investigate emotions in connection to language or communication.Thomas (2007) investigated the use of diary-based approaches in eliciting emotions around HIV/AIDS.Her study is situated in the discipline of human geography.The study was conducted in the Caprivi region of Namibia and had as participants both people living with HIV/AIDS and those who act as caregivers to people living with the disease.The participants were asked to record their emotions by using verbal and photo diaries.The diaries were then analysed thematically.Bar-Lev (2008) investigated the performance of emotions in an online HIV/AIDS support group (HOPE) in a longitudinal study stretching over two years.Bar-Lev was interested in the social relationships created by HOPE from a therapeutic as well as a moral point of view.She specifically argues that sociologists need to acknowledge that discussion forums serve a moral function.Bar-Lev (2008:512) downloaded discussion threads from this forum, and focused on the theme of anger.She used narrative analysis while also focusing on discrete characteristics of the particular narratives "such as the use of metaphors, repetition of phrases, and use of active versus passive speech" (Bar-Lev 2008:513).Demmer (2006) examined how HIV/AIDS caregivers handle the emotions which result from working with death and bereavement on an almost daily basis.The study was conducted in KwaZulu-Natal, South Africa.Interviews were used as the primary means of collecting data.Dageid and Duckert (2008) had as focus the coping strategies of black rural women living with HIV/AIDS in the Limpopo province in South Africa.Data was collected by means of indepth interviews and focus group discussions.Half of the women in this study spoke English, while the other half used Sepedi in the interviews.The interviews in Sepedi were translated into English by one of the volunteers at the hospital.
The studies either investigated emotions in a broad sense, or focused on a specific emotion.The studies are valuable for a number of reasons.Firstly, they attempt to investigate this under-researched area of HIV/AIDS and emotions, contributing knowledge about the emotional impact of the disease as well as the experiences of people living with HIV/AIDS.These studies also provide researchers working within the area of emotions and HIV/AIDS with an indication of the types of methods of data collection and analysis available.It is our view that although the papers reviewed did not set out to do language-or communicationrelated studies on the expression and construal of emotions, they could have benefited from research on crosslinguistic and multilingual aspects of emotion construal.We have identified three themes which current studies on the expression and construal of emotions have shown to be important and which were not attended to in much depth in the studies reviewed.The goal of this section is not to criticize the studies for these apparent shortcomings, but rather to discuss how taking these themes into account could have problematized the findings or shed light on the expression of emotions in HIV/AIDS communication in multilingual contexts.

Translatability of emotion concepts
The reviewed studies seem to take a particular view of emotion words and emotion concepts.Thomas (2007:75) seems to situate herself within a paradigm that acknowledges that certain aspects of emotion experiences are not universal by citing Bennett's (2004) view that emotions must be seen "within the context of the changing embodied relations, discourses and socio-cultural environments in which they are located".This is an acknowledgement that cultural and contextual differences exist in how emotions are expressed.However, even with this awareness of the role of cultural and contextual differences, Thomas (2007) does not explicitly mention how differences in language might affect the expression and construal of emotions.Instead, language seems to be treated simply as equal to culture and not as a variable on its own which might have an effect on the encoding of emotions.This is particularly evident in the method of analysis used in the study.All participants could write in their language(s) of choice, but the diaries were translated into and analysed in English; not in the language it was originally written.The translator, although believed by the author to be fluent in all six languages of Namibia's Caprivi region, was not professionally trained, nor was this person's proficiency level in each of these six languages reported.A similar method is used by Dageid and Duckert (2008).The persons who carried out the translation from Sepedi to English were not professionally trained and, as with the previous case, the translated transcripts were analysed, not the original transcripts.Current research in crosslinguistic variation in emotion language and in emotions has shown that there is substantial variation in size among the emotion lexicons of individual languages.For instance, whereas German is said to have approximately 250 emotion words (Gehm and Scherer, 1988), the number reported for English amounts to 2000 (Wallace and Carson, 1973).In concrete terms, this means that a certain emotion concept that is central to a speech community may be either peripheral or completely absent in another.For example, in Japanese, the concept amae expresses a kind of indulgent dependency, rooted in the bond between mother and child, and claims have been made that amae is distinctive to Japanese culture and child-rearing in particular (Doi 1973).Whereas the question remains unanswered as to whether the emotion concept of amae 1 is found exclusively in Japanese culture, it should be noted that many other languages (such as the Germanic and Romance languages) do not have this concept.Another important point to be made, in relation to the topic of crosslinguistic variation in emotion concept repertoires, concerns the relationship between a given concept and the experience denoted by that concept.The absence of an emotion concept such as amae in English does not http://spilplus.journals.ac.za necessarily mean that speakers of English are not able to feel "indulgent dependency".Rather, what these differences in conceptual repertoires entail is that in some languages and cultures certain concepts are made more salient than others.Such circumstances may, in turn, result in those concepts having a higher cognitive prominence in the mind of the speaker2 .
Taking into account this research on translatability of emotion concepts, the method of analysis utilised by both Dageid and Duckert (2008) and Thomas (2007) is problematic.Firstly, there are problems with the translation process itself.According to Pavlenko (2005:241), there is evidence showing that untrained interpreters and translators commit a variety of errors when translating emotionally charged conversations including "addition, substitution, omission, condensation, paraphrasing, and misinterpretation of information".Secondly, emotions are generally difficult to translate due to language-and culture-specific factors, whether the translators are trained or not.Both Thomas's (2007) study and that of Dageid and Duckert (2008) make the implicit assumption that all emotion concepts have translation equivalents in English.This method of analysis raises a concern articulated by Wierzbicka (2009) that English emotion words are regarded as universal by many researchers.This dominance of English in emotion research is expressed by Wierzbicka (2009:4) who says that "it is common practice for scholars to write about human emotions using English emotion terms as if these English words could give us an accurate, objective and culture-independent perspective on human emotional experience in general."It is clear that both Thomas (2007) and Dageid and Duckert (2008) assume and use English emotion terms as universal which, as current research shows, is misinformed.A comparison of emotion concepts across languages shows that three possible types of relationships might occur between two languages.Two emotion concepts might be similar or identical in all cases, one language might not have a concept which is present in the other, and lastly there might be partial but not a complete overlap between emotion concepts (Pavlenko 2008:150).The strategies used by the translators when difficulties occurred in the translation of emotion words or the linguistic encoding of emotion words were also not reported on.Thomas (2007:78-79) subsequently classifies a number of emotions that were identified as common themes in the analysis of the diaries, namely "fear and confusion, frustration and loss of purpose, optimism and sense of purpose".Once again, this classification is problematic as it is not clear whether the languages in which the diaries were originally written linguistically encode these emotions, and if these emotions are conceptualised in the same way across the languages.For example, Pavlenko (2008:151) reports on how Elena Koreneva, a Russian actress, had difficulty translating the English word "frustration" into a single word in Russian.A further question for consideration is how emotion concepts relate to the physiological reactions that are associated with a given concept.Pavlenko notes that research to date has not dealt extensively with this topic, but quotes Panayiotou's (2004) study on the Greek emotion concept stenahoria ('discomfort/sadness/frustration'), which is typically accompanied by a feeling of suffocation and claustrophobia (p.152).In the case of both Taylor (2007) and Dageid and Duckert (2008), insights might have been gained into whether the expression of emotions differs for speakers of languages other than English if the diaries were analysed in the languages in which they were originally written, or if more was known about the translation strategies.In the studies reviewed, emotion research is largely based on universal (English) accounts of emotion language and emotion concepts.

2.3
Linguistic profiles of participants A further concern is not only that the translatability of emotion terms, and subsequently emotion concepts, were not taken into account, but also that no information was given about the linguistic profiles of the participants in all of the studies reviewed here.For example, in Demmer's (2006:99) study, it is reported that six out of eight participants spoke Zulu; no further information is provided as to whether it is Zulu as first (L1) or second language (L2).Similarly, Dageid and Duckert (2008:184) reveal that half of the women in their study used Sepedi, while the other half spoke English, thus omitting any information about their participants' linguistic trajectories and proficiency levels.The studies by Thomas (2007) and Bar-Lev (2008) provide no information regarding the linguistic profiles of their respective participants.
The level and degree of multilingualism have been shown to have an influence on the way that emotion words and concepts are used, related to, and understood by multilinguals.Pavlenko (2005Pavlenko ( :2008) ) lists no less than seven different outcomes that might occur when multilinguals deal with emotion concepts that are not exact equivalents.First, two languagespecific concepts may coexist in the multilingual mind in the sense that their representations are similar to those of monolingual speakers of the respective languages.This outcome may be expected among bicultural bilinguals, particularly if they acquired both languages early in life.The second outcome is so-called L1-based conceptual transfer, in other words, the strategy by which the bilingual relies on concepts specific to his/her L1 to interpret L2 emotion concepts.Jarvis and Pavlenko (2008) note that this process is most commonly manifested among less advanced L2 learners who have had little opportunity for L2 socialization.A third possibility is internalization of new concepts, which refers to the process by which a learner, through the L2, acquires a novel emotion concept that does not form part of his/her L1 repertoire.This L2 concept is then usually used in the L1 as a loanword.The fourth outcome identified by Pavlenko is conceptual convergence, which occurs when two partially overlapping L1 and L2 concepts are merged to form a unitary concept in the speaker's mental lexicon.Even though conceptual convergence of emotion terms is a theoretical possibility, the current scarcity of empirical evidence suggests that it is not a common process.The three remaining outcomes concern changes that a given L1 emotion concept may undergo under the influence of an L2: Conceptual restructuring refers to the process whereby the L1 concept is slightly remodelled to approximate an L2 concept; conceptual shift may be defined as conceptual restructuring taken one step further, that is, the originally language-specific and partially overlapping L1 concept has been fully adapted in meaning to an L2 concept; conceptual attrition is the outcome where a given L1 concept or way of encoding a concept has been lost.These latter three processes are common among speakers who have resided for a long period of time in an L2 environment with a low degree of contact with the L1.The classification used by Pavlenko (2008) is particularly useful for research contexts where bilingualism is the result of immigration.The South African context is interesting in that the type of multilingual profiles found here are less common in Western contexts (Banda 2009).By paying attention to the linguistic profiles of participants in emotion studies in South Africa, it might be possible to find evidence that some of these possibilities identified by Pavlenko work differently in different contexts of multilingualism.
As is also shown in studies on bi-and multilinguals, multilingual individuals often perceive and express themselves differently in the different languages they know (Besmeres 2004;Pavlenko 2005).In an early study, Ervin (1955) showed that bilinguals who were asked to report on the same stimuli, in both of the languages they knew, often reported different or conflicting information.In the reviewed studies that use interviews as the method of data collection, no information is provided as to whether the interviews used to elicit data were conducted in the interviewee's L1 or L2.A point of consideration for future studies looking at emotions in HIV/AIDS discourse could be to establish a linguistic profile of their participants, and where the individuals are bilingual, conduct interviews in both languages or carefully choose the language of the interview.In the case of Bar-Lev's (2008) study, the lack of information regarding the linguistic profiles of the participants can partly be ascribed to the method used for data collection.The reliance on data gathered on an internet forum makes it difficult to establish a biography of the participants, except the information that they offer in their discussion with others.Here too, ethical considerations would prevent the researcher from giving biographical information that can identify the participants.Nevertheless, information on which languages were used in the postings would have been valuable, and would have possibly illuminated the findings.Bar-Lev (2008:510) finds that emotion talk is prevalent in the posts.It is also clear that this online community created their own discursive norms around emotions, including which emotions are acceptable and which not, or which are seen as more useful than others.For example, anger is seen by this online community as acceptable only in as far as it provides a venting vehicle, but is not further viewed as a useful or legitimate emotion.Bar-Lev does not give any information about which language(s) the postings in the forum were made.One can only guess that the community uses largely (or, perhaps, exclusively) English and that Bar-Lev assumed that everybody is aware of this.The absence of this information is surprising as Bar-Lev (2008:518) concluded that this community has a "purely linguistic existence".If this community is purely linguistic in nature, surely information about which language(s) are used by this community is paramount.
It is difficult to criticize Bar-Lev for a lack of focus on the possible linguistic differences in performance of emotions due to the focus of the study and the type of data utilised.Nevertheless, a number of insights can be gained by conceptualizing a similar study that does take these factors into account.For example, Pavlenko (2008:151) points out that crosslinguistic variation is seen in emotion-causing events and consequences, as well as in appraisals of emotions themselves.She provides the example of "envy": among Englishspeaking Westerners, envy is typically seen as a negative emotion, whereas in a Chinese context, its translation equivalent (xian mu) denotes a "feeling of admiration of a person who has something you want".As such, xian mu is likely to be appraised as more pleasant than envy.Another example can be found in Fant's (1995) analysis of the appraisal of anger expressions in children in Scandinavian and Spanish cultures.Whereas a child's manifestation of anger may be perceived of as both a lack of self-control and poor education among Scandinavian families, anger expression is encouraged to a greater extent in Spanish families as it is regarded as strengthening the child's character and will.
A study which takes the linguistic profiles of participants into account in the context of Bar-Lev's study, can draw conclusions on how the performance of anger is linked to language, and on how useful anger is seen as a resource in counselling on HIV/AIDS across languages.From a bilingual perspective, a similar study, which has both monolinguals and bilinguals as participants, can provide further insight into if/how bilingual and monolingual individuals perform and judge anger in different ways.Do bilinguals conform to how people who share the same L1 with them perform anger?Or are their expressions, conceptualisations and performances of anger similar to monolingual speakers of their L2?Or is it different from monolinguals in either language?The absence of information about linguistic biographies in emotion studies points to a general disregard of the role that language plays in the discursive construction of emotions.

2.4
Benefits offered by multilingualism for eliciting of emotions So far, we have focused on how linguistic and cultural diversity in emotion scripts may constitute challenges when people with different backgrounds talk about, translate, and interpret emotions.
In Demmer's (2006) study, one of the central themes expressed by the professional caregivers who acted as participants is the difficulty in getting their clients to talk about death and grief.A number of reasons are offered for this by the participants.One reason was that clients are focused more on tangible consequences of death, such as funeral arrangements, financial difficulties etc.Another reason was that there are cultural differences in how death and bereavement is viewed.The participants in the study stated, for example, that in the Zulu culture, people are not encouraged to talk about their feelings, that grieving is confined to the funeral and that after that the grieving process is seen as completed.Similar to Demmer (2006), Dageid and Duckert (2008) found that the participants in their study typically make use of avoidance-coping strategies to deal with their condition, and that cultural differences are put forward as a reason for avoidance of emotion.Sepedi culture is described as one that "encourages restrained expression of emotions, and do[es] not see them as separate subjects of discussion" (Dageid and Duckert 2008: 191).These perceived cultural differences are handled with varying strategies by the caregivers in Demmer's (2006) study.For example, one of the participants in the study refers to the use of "natural empathy" (Demmer 2006: 101) to get people to be more forthcoming with expressing their emotions.Empathy has been shown by Gibson and Zhong (2005) to be an important part of intercultural communicative competence.They cite Benett (1979) in defining empathy as "intellectual and emotional participation in another person's experience".
Empathy and intercultural communicative competences have also been found to be enhanced by bilingualism.This is found even in cases where counsellors or caregivers do not share an L1 with their patients; these bilingual caregivers are still perceived as more empathetic than their monolingual counterparts (Costa 2010).Information is not provided on how many of the caregivers in Demmer's (2006) study are bilingual.It is our suggestion that future studies need to investigate how bilingualism might affect communicative competence and empathy in HIV/AIDS communication.If Costa's (2010) findings that bilingual caregivers are perceived as more empathetic than monolingual caregivers are supported by future studies, one can argue that in HIV/AIDS related encounters, the training of multilingual caregivers and healthcare practitioners should be encouraged, as people infected with HIV/AIDS might be more prone to discuss and share their experiences with practitioners whom they find more empathetic.
Another way in which multilingualism might be beneficial relates to the fact that in the multilingual individual, different languages often have different emotional loadings.Apart from the different conceptual outcomes that arise as a function of crosslinguistic variation in emotion concepts, research has also shown that the languages of the multilingual may have different emotional values attached to them.Pavlenko (2008) discusses this phenomenon in terms of emotionality which is commonly defined as an individual's emotional reactivity http://spilplus.journals.ac.za (both behavioural and physiological) to a given stimulus.Findings from skin conductivity response investigations have shown that emotionality (measured as a response to emotionladen words) is often stronger in the L1 than in the L2. 3 For example, curse-and taboo words are often perceived as less offensive in the L2 than in the L1.(e.g.Harris 2004).Moreover, clinical case studies on bilingual patients provide further evidence of language-specific emotionality, showing that if therapy is carried out in the L2, a distance may be created between the patient and the traumatic experience being discussed.However, if the language of therapy switches from L2 to L1 (made possible through bilingual therapists), emotional outbursts or breakthroughs may be triggered (Pavlenko 2008:156).
Based on their findings, Dageid and Duckert (2008), for example, propose a number of interventions which might aid women in the Limpopo province and assist them in dealing with HIV/AIDS more effectively.These include a need for interventions that are culturally sensitive and relevant to the context at an individual and collective level; a holistic, community-driven approach, and the utilization of the local languages.However, this intervention can go even further in investigating how emotions are linguistically expressed, understood and interpreted in Sepedi.A focus on the local languages is essential, but the usefulness of multilingual resources should not be discounted, as they might enhance empathy, intercultural communicative competence, and different ways of talking about HIV/AIDS.The detachment that an L2 might allow could make it a vehicle for the expression of painful emotions.
The degree to which emotionality is language-specific in the bilingual individual is influenced by language acquisition history and language dominance, such that the earlier acquired language (or the dominant language) exhibits the highest emotionality index.An approach that takes a balanced account of multilingualism and the expression of emotions might be able to offer alternatives to dealing with emotions in communities that are linguistically and culturally diverse.Crosslinguistic differences can therefore potentially aid discussion on emotion-laden events as multilinguals have access to multi-semiotic resources which can be utilised in HIV/AIDS communicative encounters.

Implications of a multilingual perspective of emotions in HIV/AIDS contexts
All the studies reviewed here make important contributions to our knowledge of HIV/AIDS and emotions.However these contributions are limited by assuming a universal (English) way of understanding and performing emotions.These studies could benefit significantly from paying attention to work that has been done in bilingualism and crosslinguistic studies of emotions.This lack of attention to studies on linguistic encoding of emotions and its relation to the conceptualisation of emotions also filters through to the way in which data are collected and analysed.These methods of data collection and analysis clearly show a disregard for the importance of language in the investigation of this phenomenon.One cannot criticize these studies for what they were never intended to be, i.e. linguistic studies.However, in a world where multilingualism is the norm rather than the exception (which is particularly true in the South African context), more attention needs to be paid to language.Language is an important tool for the prevention strategies, treatment and counselling of the disease (Peräkylä 2010), and the role that language plays in the verbal expression and conceptualisation of emotions in multilingual contexts should therefore be more seriously investigated.
It is also important that HIV/AIDS and emotions are more actively investigated as there is currently very little research on the topic, especially in the South African context.The links between language, the expression and construal of emotions, citizenship, and voice in HIV/AIDS communication is in need of further research (see Luphondo and Stroud in this issue).How people express emotions or react to emotion-laden events can be linked to the linguistic resources they have at their disposal.In a multilingual context one would want to investigate how the linguistic and other semiotic resources people have access to in order to express their emotions around HIV/AIDS aid or hampers access to counselling, adequate care and treatment.One would also need to investigate to what extent these resources used in expression of emotions can make for changes in policies and practices surrounding the disease.
The studies reviewed have indicated which methodologies are already available to investigate emotions.However, to progress further, ethical use of corpus linguistic approaches, psycholinguistic methods of data collection and analysis, together with sociolinguistic approaches such as interviews, diaries and autobiographical studies are needed.These different methods of data collection and analysis will give a richer account of how people infected with and affected by HIV/AIDS use both local and global linguistic and semiotic resources to express and experience emotions.Such an account not only offers various practical implications such as contributing toward a more holistic view of treatment and counselling around the disease, but also theoretical implications.Theories of communication as an embodied and multi-semiotic practice will be enhanced by investigating the role of emotions in communication about such an emotion-laden subject as HIV/AIDS.