Social Representations Theory and Young Africans’ Creative Narratives about HIV/AIDS, 1997–2014

Abstract

Research on social representations (SRs) has often focused more on categorical than narrative-based representations. However, narratives are considered to play a key role in the organization of social representations. This paper describes an empirical study of some 2,000 creative narratives about HIV written by young Africans from five countries between 1997 and 2014 and examines the theoretical, methodological and applied relevance of SRT for this study and the implications of the study for the intersection between narrative and SRT. The study is unusual within the SR paradigm: it is temporal and cross-national; addresses a subject whose science has evolved over time; and uses creative narratives as its data source. A narrative perspective foregrounds holistic understandings of SRs as systems of thought. Creative narratives fit well within an SR framework. Our triangulating methodologies foreground categorical or narrative dimensions depending on the objectives of specific sub-studies. Central Core Theory provides a framework to articulate stability and change within narrative representations. In creative narrative, objectification also happens at the level of plot and characters, such that dominant cultural narratives can be viewed as a form of hegemonic SR. We link with health communication and embrace more critical streams within SR research.

Introduction

The past decade and a half have seen unprecedented scientific, technical and programmatic advances in the prevention and treatment of HIV in sub-Saharan Africa. However, persistent stigma continues to have negative effects across the HIV prevention and treatment continuum, creating a barrier to testing, disclosure, prevention strategies, treatment, adherence, and retention in care (Deacon, Uys, & Mohlahlane, 2009; Mahajan et al., 2008). Sociocultural dimensions of HIV, like stigma, may differ from one country to another and over time. Understanding how HIV and other health-related phenomena are socially and symbolically constructed and reconstructed in the collective lay imagination has the potential to inform health outcomes at the individual and population levels and is crucial to understanding and influencing processes of sociocultural change.

An essential part of the process by which meaning is produced and exchanged between members of a culture, the practices of representation have been described as “the embodying of concepts, ideas and emotions in symbolic form which can be transmitted and meaningfully interpreted” (Hall, 1997; 10). Representation has thus been seen as the intersubjective link between “culture-in-the-world” and “culture-in-the-mind” (Shore, 1996).

Social Representations Theory (SRT) focuses on the complex symbolic, emotive and social aspects of everyday lay meaning-making and on the dynamic quality of representation formation and transformation. Coined by Moscovici in 1961 (2008), social representation refers to the process whereby social knowledge is constructed and a shared system of meaning elaborated within and across social groups through processes of communication. The term also refers to the product of that process: the shared imagery, metaphors, values and practices that allow us to make sense of, navigate, and position ourselves within the social world. Social representations tend to be image- and emotion-rich and foreground the symbolic, iconic, and metaphorical aspects of everyday thinking (Flick, 1995; Joffe, 1999). Created and recreated in everyday social interaction and spread through interpersonal and media communication, social representations are amenable to influence through communication, hence their applied importance.

SRT has predominantly focused more on categorical (e.g. discrete imagery, concepts, stereotypes) than narrative-based representations. However, the importance of narrative to social representations research from a theoretical, methodological, and applied perspective has been highlighted by Jovchelovitch (2002, 2007, 2012), László (1997, 2008), Murray (2000; 2002; 2008), and Flick (1994, 1995).

Narrative theory, particularly as articulated in Jerome Bruner’s narrative-based cultural psychology, aligns with SRT in a number of respects and the common genealogy of the two approaches has been described (László, 2008; Murray, 2002). Per Bruner, narratives are “a culture’s coin and currency” (Bruner 2002: 15); as such, like social representations, they underlie social knowledge. Inherently dialogical, narratives draw on the lived experience of both teller and audience as well as on their shared cultural models: storytelling is a testament to the dialectical relationship between the subjectivity of personal experience and the creative imagination, on the one hand, and our reliance on cultural norms and systems of representation to create and share meaning, on the other. Social representations, as described by Marková (2003), are similarly the product of the dialectical interaction between individual cognition and the social world. In a memorable metaphor for this dialectical relationship, Wagner et al. propose (1999; 96) “Instead of imagining representations within minds it is better to imagine them across minds, resembling a canopy being woven by people’s concerted talk and actions”. As Jovchelovitch (2002) indicates, both narrative and social representations also serve to make the unfamiliar familiar, linking the ordinary and the extraordinary; to organize experience, making it understandable and less threatening; and to articulate identity and create community.

It has been argued that narratives are involved in the genesis of social representations (Jovchelovitch, 2002), with the latter being constructed through the social process of narration and storytelling (Murray, 2002, 2004). As a result, it has been claimed that social representations have an underlying narrative structure that has been overlooked by many researchers (László, 1997, 2008) and that narratives plot social representations into a meaningful architecture (Jovchelovitch, 2007, 2012).

Two processes are central to the formation and transformation of social representations: anchoring, whereby new concepts are integrated within existing knowledge; and objectification, whereby an image, symbol or metaphor comes to stand for the new phenomenon or idea, making it more concrete and intelligible. László (1997) argues that these processes are intimately related to the stories prevalent within social groups.

It follows from the role narratives play in the formation, structuring and transformation of social representations that narrative-based research methodologies would be useful to their study. László (1997) argued that the best way to identify how people relate to shared systems of meaning is to explore how they locate these categories in a narrative context. The four scholars who have contributed most to our understanding of the intersection of SRT and narratives have used in-depth interviews (“pseudo-narratives” per László (1997)), in some cases combined with other methods. Flick (1992) introduced the episodic interview, which seeks to access “retrospective anchoring” by asking research participants to recount past and recent encounters with a phenomenon thereby generating a thick, narrative description.

Although examination of the intersections between SRT and related theoretical constructs have been welcomed and pursued as a means to foster theoretical enhancement (Voelklein & Howarth, 2005), the links between SRT and narrative are still underdeveloped (Murray, 2008). Murray and Flick have addressed social representations of health and illness, Jovchelovitch the historical narrative of Brazilian national identity and origins, while László adopted a more literary perspective. Given the diversity of narrative types and of methodological approaches addressed in narrative-based social representations research, it is helpful to revisit the question of the intersection of narrative and SRT, not least with reference to a distinctive empirical study. In this paper, I situate a study of creative narratives about HIV written by young Africans from several countries between 1997 and 2014 in relation to SRT. I seek to examine the theoretical, methodological and applied relevance of SRT for this study of narratives and the implications of the study for the intersection between narrative and SRT. Specifically, I focus on questions of a theoretical order that have emerged in the course of our empirical analyses and that I feel merit debate:

1. What are the strengths and weakness of our study design and methodologies from the perspective of SRT?

2. Based on our empirical study, how do we understand the interrelation of social representations and narratives and what are the implications of that understanding?

3. What can SRT contribute to our understanding of commonalities and differences in our narrative data across time, place, and demographics?

4. How does the intersection of SRT and narrative inform our applied aims?

Background to the study

In the context of a vast multidisciplinary literature on narrative, definitions abound, with Barthes’ (1975) – referring to an infinite variety of forms of narrative – being perhaps the most all-encompassing. In light of our distinctive data source of creative narratives about HIV, I rely here on: a widespread understanding that narrative at its most basic constitutes the representation of a series or sequence of events (Rudrum, 2005); on perspectives related to the study of literary narrative texts focusing on plot, characterization, and hermeneutics; and on Bruner’s (1991) understandings of the narrative construction of reality. In linking the creative narratives that are our data source with the sociogenetic perspectives that guide our analysis of them, I use the term “cultural narrative” to refer to culturally dominant narrative-based ways of thinking about HIV in terms of certain kinds of characters, actions and outcomes (“stories”). This is in contrast to the more generalized use of the term to refer to universalizing hegemonic discourses as in “grand narratives” (Lyotard, 1984). I invoke the dichotomy of the categorical vs. the narrative-based employed by others who have written on the subject of narrative and social representations (László, 1997; Murray, 2008). However, I recognize that– depending on the definition of narrative applied – some may consider the categorical and narrative to be poles, with a spectrum of thematic qualitative approaches lying in between.

Researchers have embraced differing conceptualizations and interpretations of social representations, and adopted a range of philosophical and methodologicalapproaches that are positivist/interpretivist, categorical/narrative, and quantitative/qualitative to varying degrees. These range from the structuralist school of Aix-en-Provence (Abric, 2001) to the anthropological and primarily qualitative approaches of Herzlich (1969) and Jodelet (1991) (yet even Herzlich’s seminal qualitative study has been critiqued for neglecting the narrative qualities of her interviews and focusing instead on categorical anchoring and objectification (László, 1997)).

In light of the philosophical and methodological diversity of the field, Wagner (1996) has proposed that it constitutes less a consistent and well-defined theory than an approach encompassing a number of distinct theories that are not always entirely mutually compatible. Caillaud et al. (2019) argue, however, that methodological discrepancies have come to be viewed as a way to deepen understanding, rather than a threat to scientific validity. They argue that SRT is strengthened by embracing a systematic triangulation that combines the various epistemological, theoretical, and methodological backgrounds of different methods. Batel and Castro (2018) contend that the convergence between SRT and discursive psychology is increasingly acknowledged and has mutual benefit for understanding social change. SRT was intentionally conceived as a flexible, evolving, catholic and pragmatic theory. It aligns well with an increasingly prevalent view whereby micro and macro theoretical perspectives are grounded in empirical data and valued for their contributions to a social change agenda. We embrace such pragmatics, triangulation, convergence and applied theoretical relevance in our study.

While our study conforms to an SRT paradigm by dint of its focus on lay sense-making about HIV, it is unusual as a study within the social representations tradition. It is both longitudinal and cross-national; it addresses a subject (HIV) whose science has evolved dramatically over time; and our data are creative narratives whose referential status is unresolved.

SRT was originally concerned with how new scientific developments, starting with psychoanalysis in Moscovici’s seminal study (2008), were received and appropriated by the lay public. The focus on new phenomena allows processes of anchoring and objectification to be observed in action, hence the prominence accorded to the theme of HIV in social representations research in the 1980s and 1990s (Joffe, 1996; Markova & Wilkie, 1987). Social representations studies have, however, also addressed existing social phenomena. There have been calls to track the evolution in social representations over time (Bauer & Gaskell, 1999), however few such diachronic studies have been conducted (Brondi, Sarrica, Cibin, Neresini, & Contarello, 2012; Gaymard & Cazenave, 2018). With rare exceptions (e.g., Flick, 2000; Joffe, 1999), there are also few cross-national comparative studies of social representations.

The longitudinal study of HIV is of particular interest given the dramatic evolution in HIV-related science and programmatic response. Cross-national comparative perspectives, meanwhile, can have potential policy and programmatic implications. There is, however, a dearth of longitudinal and cross-national quantitative data that can help us compare understandings of HIV across time and place, while most qualitative studies are conducted in a small number of often highly circumscribed settings and without a clear point of comparison outside the study site, limiting the potential for systematic comparison.

Distinctive qualitative data offer an opportunity to examine how young Africans are making sense of evolving developments in HIV prevention and treatment. More than 150,000 young people from across sub-Saharan Africa took part in HIV-themed scriptwriting contests held at 8 discrete time points between 1997 and 2014, creating over 75,000 narratives. Over recent years colleagues and I have engaged in the analysis of a stratified random sample of close to 2,000 narratives from five or six (depending on the study) epidemiologically and culturally diverse countries (Beres, Winskell, Neri, Mbakwem, & Obyerodhyambo, 2013; Singleton, Winskell, McLeod, et al., 2018; Singleton, Winskell, Nkambule-Vilakati, & Sabben, 2018; Winskell, Beres, Hill, Obyerodhyambo, & Mbakwem, 2011; Winskell, Brown, Patterson, & Mbakwem, 2013; Winskell, Hill, & Obyerodhyambo, 2011; Winskell et al., 2015; Winskell, Kus, et al., 2018; Winskell, Obyerodhyambo, & Stephenson, 2011; Winskell, Sabben, et al., 2020; Winskell, Singleton, et al., 2020).

The scriptwriting competitions (Winskell & Enger, 2014) were coordinated internationally by the non-profit organization Global Dialogues (www.globaldialogues.org). I was one of the initiators, alongside my colleague Daniel Enger, who leads the process, and our African colleagues. The young participants in the Global Dialogues contests (known as Scenarios from Africa before going global in 2012) are mobilized by nongovernmental and community-based organizations and local, national, and international media across sub-Saharan Africa. A leaflet provides young people up to the age of 24 with instructions on how to participate in the contest, inviting them to come up with a creative idea for a short film about HIV. The winning ideas in each contest are selected by local juries and, following adaptation, transformed into short fiction films by prominent directors. By 2014, the process had generated 39 short films (www.youtube.com/globaldialogues) and an archive of over 75,000 narratives (Winskell, Singleton, & Sabben, 2018) written by young people in almost every country in sub-Saharan Africa.

Distinctive characteristics of the data source

Our reflections in this manuscript are, then, grounded in a specific kind of data – creative narratives elicited in the context of a health communication program. These are not personal illness narratives, testimonials, or life histories, nor are they contemporary folkloric legends, though the young authors may draw on these cultural resources and models as they seek to make narrative sense of HIV. They have a distinctive focus on plots, characters, consequences, and messaging.

By virtue of their presumed (and often explicitly) fictional status, the narratives are symbolic representations in line with Hall’s (1997) definition above par excellence and hence arguably a particularly apt data source for a study under the aegis of SRT. The distinctive characteristics of the data and the unusual discursive context of their collection – through a scriptwriting competition – warrant further consideration as we reflect on the opportunities and challenges they represent from a social representations research perspective.

In explaining social representations, Moscovici (2001) encourages us to call to mind Magritte’s painting “The Treachery of Images” – a representation of a pipe that declares “This is not a pipe”. Per Hall (1997), representation enables us to communicate intersubjectively about both the “real” world of objects, people or events, and the imaginary worlds of fictional objects, people, and events. Although fictional stories are invented depictions of social fact, they are still culturally-determined social facts in and of themselves and are, as Rabinow (1986) describes, a source of insight into how people make sense of the world, and how they communicate those understandings to others in their cultural community.

Although Moscovici (1986) wrote about the value of literature, specifically Marcel Proust, as a source of insight for social psychology, he did not address such creative cultural representations as a source of social representations. Social representations studies have tended to use more traditional data sources (particularly in-depth interviews, focus group discussions, media sources, and word association tasks), sometimes in combination. This has led to calls for methodological diversification, particularly with a view to eliciting more implicit, symbolic data (Joffe, 1999).

Creative data sources like our own lay bare challenges that are often overlooked in more traditional studies. As Flick (1995) argues, the performative nature of qualitative data and questions around their referential status are too easily ignored in more traditional studies. Our study focuses on the post hoc analysis of fictional narratives, collected for another purpose (a scriptwriting contest). Like all narratives, these are oriented to an audience. They allow us to analyze young people’s representations of HIV-related phenomena in their narratives, rather than their responses to specific interview or focus group questions. Thus, the young authors are able to situate HIV within their own cultural and moral logic without contemporaneous co-construction with the researcher.

The narratives may be embedded within cultural norms of performance, discourse, and persuasion (Farmer & Good, 1991) specific to the contest, reflecting the young authors’ motivation, for example, to tell what they consider to be a good story and thereby win the contest, or to educate their communities about HIV. These conditions of production, which are implicit in other forms of data collection, are explicit in ours.

Methodological strengths and weaknesses from a social representations perspective

Unlike most social representations studies, Moscovici’s seminal study was mixed-methods, combining surveys, interviews and content analysis of newspaper articles. Our own study is distinctive in using an unusually large qualitative dataset that combines ethnographic richness with amenability to systematic cross-national and longitudinal comparison. This facilitates examination – in line with SRT priorities (Flick, 1992) – not only of the social production, but also the distribution and transformation of social knowledge.

Like virtually all social representations studies, our data are not representative: young people self-select to participate in the scriptwriting contests. Participants may be better educated, and more knowledgeable and motivated about HIV than the general youth population. In addition, the availability of the data is informed by programmatic realities, including sociopolitical circumstances (e.g. the first contest, in 1997, was only carried out in three countries, while social unrest in Kenya in 2008 impacted that year’s contest there). We have little demographic information about individual participants other than their sex, age, country of origin, and urban/rural place of residence. In light of the circumstances in which the texts were written, it is also impossible to know which depict lived versus imagined experience. We do not have information on the extent of young authors’ direct personal experience of HIV, for example as people living with HIV themselves or through family members. However, with the above caveats and to the extent possible in light of the programmatic realities of the scriptwriting competitions, the collection of these secondary data is consistent across sites and over time, allowing for meaningful systematic cross-cultural and temporal comparison of a large number of narratives.

We use systematic sampling strategies to mitigate limitations of our data source and optimize its strengths. We selected countries in which at least 500 narratives were contributed in 2005, and which have diverse epidemiological and sociocultural profiles. Our interest in examining the distribution of social representations demographically, in addition to longitudinally and cross-nationally, led us to determine that only narratives that were written by individual authors were eligible for inclusion in our study (a substantial proportion are authored by groups, many of them of mixed gender). As a result, it is not possible for our study as currently executed to assess group-level - or community-level - influences on social representations, which would have been desirable from an SRT perspective.

We stratified our data into 12 categories by sex, urban/rural location and age of young author (10–14, 15–19, 20–24) and randomly selected 10 narratives (if available) from each of the 12 strata (Winskell, Singleton, et al., 2018). From a pool of 32,759 narratives for the five countries across the eight time points, an overall sample for the longitudinal study of 1,937 texts resulted.

Our methodological approach triangulates between three primary analytical components:

1. Analysis of quantifiable characteristics of the narratives (“quantitative attributes”). We quantified discrete components of each narrative, double-entering them into Qualtrics research software (Qualtrics, Provo, UT). Examples of these quantitative attributes of the narratives include whether an HIV-related death occurred, whether the narrative ended with hope, and whether transmission resulted from sexual intercourse, blood route, and/or mother-to-child transmission. In some cases we complement descriptive statistics with more sophisticated statistical methods.

2. Thematic qualitative data analysis, focusing on thematically-related text segments and memoing for emergent themes. Narratives (overwhelmingly handwritten) were transcribed into English or French and entered verbatim into MAXQDA 12 qualitative data analysis software (VERBI Software, 1989–2016), where they were labelled with reference to a codebook of 54 thematic codes.

3. A narrative-based approach, focusing on plot summary and thematic keywords. Each narrative was summarized and independently double-coded with up to six of 44 keywords. Discrepancies in coding were resolved through dialogue. The narrative summaries served as an aide-memoire and data-management tool, allowing us to easily identify common story arcs or to rapidly situate segments of data within the context of the overall plot or argument.

The codebook was developed via an iterative team-based process, drew on the recommendations of African colleagues who read the narratives when selecting the winning contest entries (Winskell & Enger, 2009), and includes a detailed description of each code, inclusion and exclusion criteria, examples of the code in use, and specific guidance for coders (e.g. “things to look for specifically”). Codebook refinement was the product of an extensive process of discussion and consensus-building in which provisional codes were applied to the data and iteratively refined. The same codebook, in English, is applied to English and French-language narratives.

These methods, developed to enable systematic comparison across demographics, countries, and time, have the advantage of grounding the analysis in three distinct, though complementary, dimensions of the data and facilitating the generation and exploration of interpretive hypotheses. We conduct intersecting quantitative, narrative-based and thematic analysis. We place less emphasis here on the thematic dimension of our analysis, focusing instead on the categorical and narrative dimensions, which – it may be argued – are more distinctively associated with social representations research. While our study triangulates between the three analytical components, the balance between them shifts based on the research goals of the sub-study. For example, our comparison of symbolic stigma across countries with varying HIV prevalence levels (Winskell, Hill, et al., 2011) focused on descriptive statistics of quantifiable characteristics of almost 600 narratives contributed in 2005, while drawing on the thematic and narrative-based components to interpret those findings. We focused on moralisation and “othering” by assessing the prominence given in plotlines to certain categorical dimensions: the circumstances of infection; the association of HIV with stigmatised populations or behaviours; expressions of individual blame and shame; and the demonization of PLWHA. As the affective framing and outcomes of a narrative often serve as vehicles for communicating the moral of the story, we also examined the tone of the endings of the narratives and the prevalence of HIV-related death across the country samples. This allowed us both to examine the content of stigmatising representations in the distinct country settings and to compare them cross-nationally.

We found that the association of HIV with outsiders (“othering”) and preoccupation with the circumstances of infection were more common in lower prevalence countries but varied substantially in tone depending on the sociocultural context. The highest proportion both of moralising narratives and of narratives with pessimistic outcomes came from South-East Nigeria and, to a lesser extent, from Kenya, countries in which evangelical Christian movements, including Pentecostalism, have sizeable followings. Social representations from the highest prevalence countries, Swaziland and Namibia, and from lower prevalence Burkina Faso offered potential models for the framing of HIV in ways that serve to increase social proximity and counteract symbolic stigma. Despite the palpable presence of AIDS in Swaziland, the characteristic tone of the stories was not one of tragedy, but of human resilience and agency, suggesting that the country’s extremely high prevalence may be favouring the emergence of cultural resources that promote coping.

In stark contrast to this bird’s eye view using the quantitative attributes as point of departure, another stigma-related study (Winskell et al., 2015) conducted a close reading of just three narratives, selected on account of the complexity of their representation of HIV stigma. The objective of this paper, which used stigma theory as a lens, was to examine stigma both in the referential dimension of the narratives (the claims they make about the ‘real’ world) and in their evaluative dimension (e.g. the often contradictory or ambivalent messages they communicate) in order to better understand the symbolic and social processes that perpetuate HIV stigma. The narratives thus provided access both to the explicit accounts of stigmatizing attitudes, behaviours and experiences of their characters and to the implicit, often contradictory, sense-making processes the young authors engage in, which may themselves be highly stigmatizing. We drew on conventional constructs of literary analysis (plot, narration, characterization, etc.) and a hermeneutic lens to articulate the relationship between parts and whole, plot and narrator’s commentary, and explicit and implicit meanings.

In their referential dimension, the narratives shed light on the contextual experience and mechanisms of stigma, revealing, for example, the role played by gendered power dynamics and sexual reputation in vulnerability to stigma. In their contradictions, tensions and ambiguities, they reveal the challenges young people face when trying to make sense of HIV and stigma with recourse to the plethora of divergent resources culturally available to them. The resulting tensions align with Moscovici’s (2008) concept of cognitive polyphasia, which allows for competing and potentially contradictory social representations to co-exist in the same group or individual at the same time. Cognitive polyphasia is a particularly important concept for the study of stigma in light of the sensitivity of the subject matter and its vulnerability to social desirability bias. It is also a useful tool in the conceptualization of sociocultural change processes as it makes it possible for traditional and modern representations to exist side-by-side, confronting rather than replacing one another (Wagner et al., 1999).

While our study uses triangulating methods, it would be ideal from a social representations perspective to complement these with triangulating data sources, for example, content analysis of newspapers or communication campaigns, and/or retrospective in-depth interviews with key informants. We seek to approximate to these sources by working in close partnership with colleagues in each of the five countries who have extensive programmatic experience working in HIV and can provide national and regional historic context on policies, communication campaigns, etc.

Delineating the intersection of narrative and social representations

Akin to Bruner’s (1986) distinction between paradigmatic and narrative ways of thinking, Moscovici drew a distinction between modern and “primitive” modes of thought, with the latter seemingly more narrative in focus and the former more diverse, fragmented and dynamic. We now turn to the question of how we understand the relationship between narrative and social representations in our study, addressing first the level of analysis at which the study operates and, second, the intersection between categorical and narrative elements.

Narrative and the sociogenetic or ideological level of analysis

Perhaps the most distinctive dimension of SRT is its dialectical dimension that links the cognitive with the social. Social representations have also been described as operating at multiple levels of analysis – at the ontogenetic (individual, developmental), microgenetic (interpersonal), and sociogenetic (or ideological) levels, with the latter being considered primary (Flick, Foster, & Caillaud, 2015).

Our study’s focus is squarely on sociogenetic or ideological dimensions of the narratives. Although individually authored, the narratives are oriented towards an audience and draw on existing sociocultural resources to make sense of HIV: as such, they provide insights into sociocultural norms within that community and this is the primary interest of our study. For our purposes, these norms include “story skeletons”, “a cluster of motivated events and states, sequential in time” (Schank & Abelson, 1995; 51), or the prototypical cultural narratives which – through their structure, plot, tone and thematic treatment – prescribe and constrain the types of stories that can be told (see László, 1997).

Scholarly perspectives on the ideological workings of cultural narratives has tended to focus on totalizing “grand narratives”. However, the more literary, micro-level definition of narrative as a sequence of events also eloquently illustrates the workings of ideology. One story skeleton characteristic of the Nigerian sample contrasts “good girl” and “bad girl” archetypes to condemn active female sexuality, idealize female virginity, and connect HIV with immoral sexual behaviour. This specific story skeleton finds expression, for example, in a 2014 narrative by a 15-year-old rural male author in which Queenie is depicted as spoiled, sexually promiscuous and materialistic and ultimately dies a humiliating death from HIV, while poor and humble Munachi refuses to engage in transactional sex and rejects sexual advances as she works studiously through school, eventually marrying “one of the decent boys”.

In addition to revealing common story skeletons, our narrative data provide insights into young people’s appropriation and adaptation of dominant cultural scripts, a prescriptive form of micro-level narrative, which communicate “a set of expectations about what will happen next in a well-understood situation” (Schank & Abelson, 1995; 5). For example, those around gender and sexuality, “essentially instruct the narrative requirements of specific roles” (Simon & Gagnon, 1986; 98), serving to either constrain or foster agency within sexual encounters. For the purposes of our research and its applied aims, we find it helpful to view cultural – including sexual – scripts as a prescriptive form of social representation in narrative form that typically denies agency.

In our analysis of sexual abuse of power in our data, we found that many narratives reflected common cultural scripts embedded within two common story skeletons around relationships between teachers and female students and male employers and domestic workers (Singleton, Winskell, McLeod, et al., 2018). There was little in the way of critique of underlying assumptions of male sexual entitlement and female responsibility for controlling male sexuality in the context of unequal control of resources: the power dynamic of these encounters went unquestioned by the authors (both male and female) who conflated sexual abuse of power with consensual sex, particularly in Nigerian and Kenyan texts and those by younger authors in earlier study years, ultimately blaming the female characters for their sexual behaviour.

Julian Rappaport (2000; 4) defines dominant cultural narratives as “overlearned stories communicated through mass media or other large social and cultural institutions and social networks” and describes how coded words become shorthand categorical labels to index these narratives, often promoting stereotypes (e.g. “the welfare mother”). In the next section, we will address the categorical and narrative in social representations research and introduce its implications for applied practice.

Intersection between categorical and narrative dimensions

In the previous section, we addressed the question of how extensive social representations can be, arguing that – from the perspective of our research – it is helpful to consider them as taking narrative form and illustrating this with reference to story skeletons and cultural scripts. Ultimately, the question of the interrelation of social representations and narrative hinges on two additional issues: how social representations are interconnected and the intersection of methodology and research purpose.

Interconnection of social representations

On the question of interconnection, Moscovici’s references to social representations resembling “theories,” or “systems” of knowledge, has been interpreted (László, 1997) as meaning that they organize reality and perform an explicative and interpretive role similar to narratives. Moscovici compared social representations to theories

which order around a theme… a series of propositions which enable things or persons to be classified, their characters described, their feelings and actions to be explained, and so on. Further, the “theory” contains a series of examples which illustrate concretely the values which introduce a hierarchy and their corresponding models of action…. In fact, from the dynamic point of view social representations appear as a “network” of ideas, metaphors and images, more or less loosely tied together, and therefore more mobile and fluid than theories.

(Moscovici, 1998: 243)

Elsewhere, he describes the whole as being greater than the sum of the parts “and thus irreducible to a collection of individual representations” (Moscovici & Hewstone, 1983). This idea of an integrated system of social representations is conveyed through the concept of the “representational field” (Moscovici, 1973).

Narratives differ from other forms of text-based data in that they not only describe the unfolding of events in time leading to outcomes, but also provide an explanatory and evaluative framework for their interpretation, and are infused with characters’ (including the narrator’s) thoughts, motivations, and emotions (Fivush, 2010). In the case of the narratives created by the young participants in the scriptwriting competition, the story structure, characters, plot, and messaging intersect to create meaning. This integrative dimension of narratives has been compared by Jovchelovitch (2012) to the ways in which narrative structures a representational field: “In the same way that narratives contain a plot that organises events and semantically connect disparate elements of a story, the narrative form provides a core structure to a representational field, bringing together and investing with meaning the various notions, values and practices it contains” (p.446). This is not to be construed as meaning that the narrative organizes pre-existing social representations, but rather that narrative and social representations constitute one another in reciprocal ways.

While SRT is eminently sensitive to the potentially non-referential status of social representations, the narratives addressed in social representations research are – for the most part – non-fictional. The young authors’ creative narratives in our study align closely with processes of anchoring and objectification. In the process of objectification, knowledge becomes saturated “with the core values and social norms of the culture” and assumes a moral dimension: “It is integrated into a moral system which regulates what is to be regarded as acceptable or unacceptable in a society” (Joffe, 1998; 24). As Joffe has demonstrated, AIDS was anchored in terms and objectified in imagery derived from past epidemics, the majority of which had been linked to foreigners, out-groups and perverse practices; much of this process was led by the mass media. It can be argued that in our narrative data, objectification happens not just at the level of the image, as in more conventional social representations studies, but at the level of the plot, characters, messaging, etc. Many of the young authors may be literally involved in a process of objectification – translating the abstract HIV knowledge they have acquired in didactic education into stories that draw on social representations and on their own personal experiences. Indeed, this was the original objective of the Global Dialogues scriptwriting competitions – to encourage young Africans to personalize the information they were receiving about HIV and thereby make it real and meaningful to them (Winskell & Enger, 2014). In this process, they are – at the same time – drawing on existing cultural narratives around gender, morality, etc.

Intersection of methodology and research purpose

On the subject of methodological considerations, Maxwell and Miller (2008) reflect on the distinction in qualitative data analysis between paradigmatic or categorizing approaches, on the one hand, and contextualizing, connecting or narrative-based approaches, on the other, and the advantages of combining the two. In a cross-sectional, cross-national study of condoms in the narratives (Winskell, Obyerodhyambo, et al., 2011) we found ourselves addressing both more categorical and more narrative representations in line with the pragmatics of achieving our applied purpose of informing condom-related HIV communication. We were interested both in categorical considerations – such as assessments of effectiveness, misinformation, identification of behavioural barriers and facilitators, isolated phrases or images, including vernacular expressions and euphemistic terms – and in narrative dimensions like summary of plot and overarching meaning, story skeletons and cultural scripts. In some cases, the two were explicitly linked: a categorical social representation (e.g. the phrase and concept “conduct not condom” in a Nigerian narrative) provided the coda for an entire story, or representations of condoms as ineffective drove plotlines in which condom use led to infection.

In qualitative research it is common to identify patterns and key themes based on their recurrence. While this strategy is helpful in determining the distribution of specific social representations, it is nonetheless important, in light of our applied purposes, to respect the diversity of the representations, including outliers. Our study, for example, often seeks to identify alternative narratives that could become important in processes of sociocultural change. In the condom study (Winskell, Obyerodhyambo, et al., 2011), such counternarratives or narratives of resistance were often humorous or offered models of relationship dynamics built on interpersonal communication. Narratives that reflected hegemonic social representations, in contrast, revealed patterns of harmful gender norms: for example, condom negotiation was initiated above all by female characters and frequently rejected by male characters who perceived it as an affront to their masculinity. We address the implications of such research findings for applied public health communication practice below.

Differences across time and place

One of the objectives of our study is to compare social representations of HIV across time and place. Systematic comparison of social representations is more easily envisaged with quantitative than qualitative methods, and with categorical than narrative approaches. The structural school of social representations, which focuses on quantitative methodologies of experimental psychology (particularly word association exercises), has been most concerned with developing a theory of change. The Central Core Theory (Moliner & Abric, 2015) differentiates between the core and peripheral components of a social representation, where the core is more stable over time, consists of fewer elements, and elicits greater consensus among group members, while the peripheral elements are more context-sensitive and more amenable to change. The core has been compared to Moscovici’s “figurative core”, the product of the process of objectification (Moliner & Abric, 2015).

According to Central Core Theory, it is the interaction between the central and peripheral systems of social representations that accounts for their stability and dynamics (Moliner & Abric, 2015). The abstract core relates to the group’s values, norms and history and determines meaning. The periphery is concrete and contextualized. It relates to group members’ individual experiences, absorbs contradictions and changes, and allows integration of newness, singularity or contradiction.

Central Core Theory has been criticized for its rigidity and over-reliance on statistical methods (Quenza, 2005). Differentiating between the core and periphery is certainly considerably easier in quantitative studies that focus on categorical phenomena. However, Jovchelovitch (2012; 452) has suggested that narrative principles plot “themes, characters, languages, times and events into a coherent core”. This core focuses on themata or antinomies, often framed in moral terms; and “fulfils normative functions and regulates social thinking, enforcing rules, prescriptions and moral codes… and consolidates plot and stories that stick and do not go away” (446).

Our methodologies allow us to track social representations in the narratives temporally and cross-nationally. For example, in a quantitative study (Winskell, Sabben, et al., 2020) using adjusted odds ratios in addition to descriptive statistics, we found that representations of death, treatment access and hopefulness improved significantly from 2005 onwards, when access to life-saving antiretroviral therapy (ART) was increasing dramatically in sub-Saharan Africa. However, a subsequent slight downturn – accompanied by an increase in the proportion of narratives that included infection – suggested that the hope-promoting and stigma-reducing dividend of ART may have partially run its course by 2011/2013. In more qualitative temporal studies, we found that while ART access may have alleviated the association of HIV with death, a thematic core of despair, blame, misogyny, and social stigma persisted, often finding expression in deeply moralistic cautionary tales (Winskell, Singleton, et al., 2020), aptly illustrating the stickiness of the narrative core described by Jovchelovitch.

While there are commonalities across country samples, one of the most salient findings from our study is the cross-national differences in the narratives, particularly as they relate to sexual morality and blame (Winskell, Hill, et al., 2011), with representations in the Nigeria sample being persistently most stigmatizing, with the highest prevalence of death, sexual transmission, intentional transmission, and blame, and the lowest prevalence of hope (Winskell, Sabben, et al., 2020; Winskell, Singleton, et al., 2020). Narratives from Swaziland, the country with the highest HIV prevalence, are – in contrast – those most likely to be hopeful. While this reflects HIV prevalence and resulting social proximity, it also suggests that the epidemic was framed in different ways at the national level and that this anchoring and objectification, based both on the local cultural and religious context and, by extension, on communication in the media, including via HIV-related campaigns, was tenacious. These distinctive cross-national differences in the narratives point to opportunities to reframe cultural narratives around HIV in less stigmatizing ways, particularly in countries like Nigeria (Winskell, Sabben, et al., 2020; Winskell, Singleton, et al., 2020). I address applied dimensions of the research in the final section.

Applied Considerations: Social representations and cultural narratives

The power of symbolic representations to create reality by constructing meaning and opening up or circumscribing possibilities points to their applied importance. For many years social representations research was encouraged (e.g. by Bauer & Gaskell, 1999) to adopt a disinterested stance, particularly in relation to communication, with a view to avoiding anything that resembled social engineering. However, an increasingly critical, engaged and activist strand within social representations research has gained ground (Howarth, 2006).

A dominant understanding of HIV – fostered in part by the media and by ill-advised prevention campaigns – is that people have acquired HIV through their own moral failings or deviancy. This understanding finds virulent expression in our narratives from Nigeria. The idea of “changing the cultural narrative” has been propelled into the mainstream in recent years (Poo & Shafir, 2018). The concept remains undertheorized, in part because it reflects different definitions of narrative depending on the context in which it is used and in part because the challenges of theory development may be compounded by methodological challenges inherent in the systematic study of narratives, which – particularly when understood in more literary terms, e.g. as story skeletons – requires a combination of quantitative, categorical and qualitative, narrative-based perspectives. As intimated by Murray (2004), the intersection of SRT and narrative has the potential to provide a theoretical frame and, through mixed methods approaches exemplified by Moscovici’s seminal study, the methodological tools to support such inquiry.

Narratives distil and organize complex social knowledge, amplifying its emotional charge. Jovchelovitch (2012) has proposed that an important question for social representations researchers is how certain stories can perpetuate some representations while others disappear. She argues that characteristics of specific stories inform the stickiness of the social representations they organize and their ability to mobilize communities to action. Liu and Sibley (2015; 275) relate this quality to “the proper use of narrative features – time, plot, characters, perspective, narrative intentions and evaluation” and the role they play in the narrative’s credibility, authenticity, relevance, and coherence, echoing Walter Fisher’s (1987) influential insights into narrative communication. If we seek to change dominant cultural narratives and related social representations, we need to be attentive to and leverage these communicative dimensions, in particular their power to cultivate empathy.

Narratives have multiple advantages as resources to counteract symbolic violence (Bourdieu & Wacquant, 1992) whereby symbolic systems promote the interests of dominant groups. These advantages include the ability of narratives to contextualise human experience holistically, to elicit empathy and identification, and to foster the social renegotiation of meaning (Winskell & Enger, 2014). These advantages offer important opportunities to challenge representations of gender and sexuality that sustain HIV stigma. Findings from our research have strong translational implications both within and beyond the communities in question and can inform the framing not only of categorical (e.g. slogans and image-based) but also of narrative-based messaging, which is particularly important in the context of increasingly prominent narrative-based health communication strategies.

Mass-mediated entertainment–education (e.g. pro-social soap operas and radio dramas often used in social and behaviour change communication in low and middle-income countries) often employ emotionally-compelling narratives to propose alternative cultural models, while participatory approaches drawing on Paolo Freire (2003 (1973)), including Boal’s Theatre of the Oppressed (1979), may empower affected communities to deconstruct and challenge existing ones (Campbell, Foulis, Maimane, & Sibiya, 2005; Singhal, Cody, Roger, & Sabido, 2004; Storey & Sood, 2013). A key component of a social change agenda is empowering alternative narratives, ideally recounted by alternative voices, namely the voices of those most affected (Winskell & Sabben, 2016).

CONCLUSION

The categorical, symbolic, image-rich dimension of SRT is the most distinctive and easily understood. However, as illustrated by our study, a narrative-based perspective allows us to better discern the interconnections between such categorical social representations and the integrative, systems-oriented characteristics of representational fields. Rather than narratives organizing social representations, it is more accurate to speak of narratives both drawing on and shaping social representations in a reciprocal, dialectical process (Murray, 2008).

Our creative narratives fit well within a social representations framework as they presume co-construction by teller and audience and draw on shared cultural models, revealing ideological patterns. Moreover, SRT accommodates and even legitimizes the fictional, symbolic, and unresolved referential status of these narratives. Our triangulating methodologies privilege more categorical to more narrative dimensions of our data depending on the purpose and applied objective of the sub-study, and accommodate epidemiological studies, close readings, and mixed methods approaches depending on our applied needs.

Our own distinctive data reinfuse categorically indexed story skeletons (e.g. the good girl/bad girl archetype) and cultural and sexual scripts (e.g. the boss-maid script) with their narrative dimension. In this way they illustrate the ways in which ideology operates through dominant cultural narratives. Given both the distinctiveness of our data and the diversity of understandings of narrative and social representations, we do not seek to extrapolate from our study to make generalizing statements about the narrative basis of social representations: depending on how social representations are conceptualized, the research purpose, and the mode of elicitation applied, they span the spectrum from more categorial to more narrative-based. However, we do view the dominant cultural narratives revealed in our data as hegemonic social representations in narrative form and believe that SRT offers a theoretical, mixed-methods methodological and critical framework to support the study of their distribution and evolution.

Central Core Theory provides a framework to articulate stability and change within narrative representations over time, across countries, and by age of author, particularly when considered in combination with Jovchelovitch’s narrative model of the central core. Moralistic dimensions of the narratives, particularly in relation to gender and sexuality, reflect power-laden positions and are least susceptible to change, even in the context of dramatic increase in access to ART.

In creative narratives, the key SRT process of objectification, whereby an image comes to stand for the new phenomenon, happens not just at the level of the image, but at the level of the plot and characters. This process is both informed by existing cultural narratives, story skeletons and scripts, and actively framed by those engaged in HIV communication, including governments, NGOs, and the media. By allowing us to unite norms, values, and emotions and eliciting empathy, narratives are particularly powerful tools for sociocultural change. Findings from our study can help inform more effective narrative-based HIV communication strategies.

As Phoenix et al. (2017) point out, SRT is most helpful when applied in conjunction with other social theories, particularly regarding issues of power and exclusion. We complement SRT with Bruner’s highly compatible perspectives on narrative sense-making, like other researchers at the intersection of social representations and narrative, and with theories of narrative communication. While a social representations frame is not essential to the sociocultural study of narratives (see Plummer, 1995, for example), it can add important analytical, methodological and applied dimensions, not least in relation to the distribution of narrative-based representations (drawing on mixed-methods approaches) and a focus on processes of change and on applied communication practice.