The Rohingya Refugee Crisis: A Social Semiotic Study of Visuals in The New York Times and The Washington Post

Rohingya Crisis

The Rohingyas are an Indo-Aryan ethnic group mostly concentrated in Myanmar’s Rakhine State (Hunt, 2017). They are predominantly Muslim and trace their history in Myanmar as far back as the eighth century. After Myanmar gained independence in 1948, the Buddhist-majority state “worked to cast out the Rohingya people, excluding them from its constitution” (Blakemore, 2019). Clear evidence exists of systemic discrimination against the population; these include Myanmar’s refusal to recognize Rohingyas as one of the “135 national ethnic groups” thus rendering them stateless, as well as leaving Rohingyas out of the census count (Blakemore, 2019). Rohingyas were neither allowed to vote nor run for political office in Myanmar’s first democratic elections (Hunt, 2017). These actions are in addition to overt state-sponsored violent campaigns such as “Operation King Dragon” and “Operation Clean and Beautiful Country” squarely aimed at exterminating Rohingyas or driving them out of Myanmar (Ahmed, 2012; Blakemore, 2019). Religious violence against Rohingyas increased in 2012 (Fortify Rights, 2014; Human Rights Council, 2017; Rohingya emergency, n.d.). The sustained discrimination eventually led to Rohingya insurgents attacking and killing nine police officers patrolling the Rakhine State border (Hunt, 2017).

Myanmar military responded with “clearance operations” that included “burning down villages and triggering a mass exodus of Rohingya to Bangladesh” (Hunt, 2017). Human Rights Watch and the UN Refugee Agency have documented nearly 10,000 deaths and the rape and abuse of women and young girls. By the end of 2018, nearly 1 million Rohingya refugees were present in Bangladesh (UNHCR, 2018). The current official position of the Myanmar government is that Rohingyas are not a national “indigenous race,” but illegal immigrants from neighboring Bangladesh (Ferrie, 2013). Myanmar’s government has even stopped recognizing the term “Rohingya” preferring to use “Bengalis” (Ferrie, 2013).

Studying visuals of the Rohingya crisis in the U.S. press is important because it will provide insights into how journalistic depictions of distant suffering engage with the biases of broad discourses such as Western humanitarianism. The study will provide more comprehension of how international news coverage in the U.S. press remains implicated with power structures that perpetuate hierarchies between the global North and South. The issue of Western journalism’s intimate ties with global hierarchies have been revived in recent times with the accusations of racial bias in reporting that favored Ukrainian refugees, who were displaced because of the ongoing Russian invasion, over other groups of forced migrants (Allsop, 2022; Hellyer, 2022). The present study of Rohingya refugee visuals will provide a yardstick for comparison with depictions of other groups of forced migrants.

Tropes and Themes of Western Humanitarian Visuality

Media images of refugees circulate within the ecosystem of humanitarianism defined as “paying moral attention to others who are beyond one’s own immediate sphere of existence” (Tester, 2010, p. vii). Given Western cultural moorings of the NYT and WP, the production, circulation, reception, and validation of these visuals occur within dominant ideologies of “the Euro-American humanitarian project” (Kurasawa, 2013, p. 202). Such a project contains “a constitutive paradox”; it challenges as well as reinscribes hierarchies between the global North and South (Kurasawa, 2013, p. 201). To cover up this paradox, Western humanitarianism employs a mode of “sentimentalism” through narrative and visual devices with the aim of evoking “sympathy, repugnance, pity, and nobleness” among Western audiences toward represented subjects (Kurasawa, 2013, p. 202).

The Euro-American humanitarian project draws heavily upon “humanitarian visuality” defined as “the interface of humanitarianism and visual representation” (Kurasawa, 2015). Western humanitarianism leans upon this regime of visuality as a tool to validate its interventions and shore up social and political capital. Visuality serves the cause of Western humanitarianism by naming and representing victims worthy of attention, foregrounding good deeds of Western actors (often ignoring local ones), representing Western humanitarianism as a universalist, apolitical discourse, and befogging how Western societies are implicated in the creation of humanitarian crises (Kurasawa, 2015). Humanitarian visuality strives for these goals by deploying a “semiotic structure of images of humanitarian crises . . . composed of a relatively stable system of formal relations between situational and compositional symbols” that serve “to establish the roles of various actors (victims, perpetrators, aid workers, etc.) who are part of the visual composition of a scene of emergency or mass suffering” (Kurasawa, 2015, p. 8). Humanitarian visuality also employs four familiar patterns or “iconographic conventions” to help translate the suffering of strangers and make them comprehensible to Western viewers: personification, massification, rescue, and care (Kurasawa, 2013, 2015).

In personification, the subject is represented as decontextualized and isolated, thus enhancing her vulnerability (Kurasawa, 2013). Massification depicts large groups “bunched together to form an undifferentiated corporeal mass” (Kurasawa, 2013, p. 207). It becomes impossible to delineate specific individual identities in massified depictions (Malkki, 1996). Rescue trope depicts humanitarian aid workers, primarily Westerners, helping crisis victims. Rescue trope reinforces prior hierarchies between the West and the East; aid workers are represented as agents able to transcend the suffering (Kurasawa, 2013). Care trope focuses on the physical and emotional recovery of victims, in addition to foregrounding their education through humanitarian institutions (Kurasawa, 2013).

Although the iconographic conventions are intended to “humanize” suffering strangers by normalizing and familiarizing differences for Western audiences, there are two dangers in the translation process: “reification” and “spectacularization” (Kurasawa, 2010). In reification, the image is abstracted from its specific socio-historical context while in spectacularization, the suffering of strangers is slotted into formulaic structures and recognizable types such that it becomes “fodder for a pornography of suffering” (Kurasawa, 2010, p. 141). To avoid the pitfalls described, Kurasawa (2010) proposes a “critical mode of visuality” that can initiate “phenomenological intensification”; that is, the visual representation must be able to temporarily displace viewers from their moorings enabling them to glimpse, even if fleetingly, “what it was like to be then and there” (p. 144). Kurasawa (2010) is empathic that viewers remain keenly aware of the distance between themselves and the sufferers; even at its most intense moments, “phenomenological intensification” can never enable viewers to fully experience the sufferings of others.

Even with these cautions, there are clear gains from using personified tropes (Bleiker et al., 2014; Šarić, 2019; Wilmott, 2017). Bleiker et al. (2014) found fewer subjects in images lead to viewers paying more attention to their suffering. Whether refugees are represented in large groups, medium-sized groups, small groups, or as isolated individuals play a substantial role in building relationships of emotional proximity or distance between viewers and subjects (Šarić, 2019; Wilmott, 2017). Social-psychological research has shown that technical decisions (camera distance, for instance) and number of subjects in visuals play a significant role in building relationships of proximity between viewers and subjects (Jenni & Loewenstein, 1997; Kogut & Ritov, 2005; Slovic, 2007; Small & Loewenstein, 2003). Absence of an identifiable victim in images has an important consequence: “We see no faces, no real people. We see just anonymous masses. We see an abstract and dehumanised political problem” (Bleiker et al., 2013, p. 411).

According to Wright (2002), many refugee visuals borrow from “iconographic themes” found in Christian religious painting and iconography (p. 60). Portrayal of forced migration predominantly draws from “The Expulsion from the Garden of Eden” and “The Flight into Egypt” (Wright, 2002). When refugee images follow prior themes, viewers are better able to relate to them (Wright, 2002). Refugee images in the media are classifiable into types including “‘Exodus,’ the mass movement of people, which may suggest the out-of-frame presence of a pursuer” or the well-known “Madonna and child” (Wright, 2002, p. 57). Beyond visual tropes and pre-established themes, humanitarian sentimentalism also uses visible markers for signaling refugee identity. For instance, wounds in refugee images often play an outsized role in stereotyped constructions of “the ideal refugee”; they serve as indicators for refugee-ness and are often accepted as objective evidence (Malkki, 1996). Depiction of refugees in the public domain occurs under strictly circumscribed ideological filters determining the themes of their representation. Refugees are often rendered silent and speechless (Bauman, 2002; Malkki, 1996; Rajaram, 2002). Wright (2004) points out refugee visuals often serve as “decorative wallpaper” while “expert knowledge is usually provided by . . . aid agencies. The voice of the refugee remains at the end of a chain of ‘framings’: contextualized by the anchorperson, reporter, non-governmental organization (NGO) representative and (perhaps) translator” (p. 108).

In addition, refugee crises are usually represented through formulaic schemas such as the savages–victims–saviors metaphor; refugee victims are rescued from savage attackers by Western humanitarian actors (Mutua, 2002). Calhoun, (2004, 2008, 2010) identifies the lack of context as a major drawback in representations of humanitarian crises. The “emergency imaginary” de-contextualizes humanitarian crises representing them as sudden eruptions with no prior cause (Calhoun, 2008). This leads to lack of accountability for the human actions that initiate and aggravate crises.

The literature review identified some dominant visual tropes and themes used in humanitarian representations to trigger sentimental responses from audiences and represent “distant suffering” in culturally proximate terms. Such visual strategies present ethical challenges as sentimentalism works to deflect audience attention from material issues that cause and sustain global inequalities and the paradox of Western humanitarianism. Based on the literature review, we choose “Exodus” and “mother and child” themes for in-depth visual semiotic analysis. Exodus was chosen because, in its depictions of refugee groups, the theme is closely linked to massification trope. Mother and child theme was chosen because of its linkages to personification and sentimentalism tropes. In addition, both Exodus and Mother and child themes are derived from Biblical iconography. This will allow us to examine the use of culturally familiar tropes to help Western audiences grasp “distant suffering.” The analysis is driven by the following research questions:

Method

The field of semiotics is “concerned with everything that can be taken as a sign” (Eco, 1976, p. 7). To Berger (1994), substitutability is a central element of signs; hence signs are “things which stand for other things or . . . anything that can be made to stand for something else” (p. 1, italics in original). Construction of meaning through signs and the interpretation of meaning in alignment with the maker’s original intent are central elements of the semiotic field. Van Leeuwen and Jewitt’s (2011) explication of Barthian semiotics notes that meaning is composed of denotative and connotative layers. Initial denotative meaning constitutes recognition of the person depicted and her associated actions, while connotative layer consists of the broader ideas represented through depicted subjects (Van Leeuwen & Jewitt, 2011). Connotations can be “very broad and diffuse concepts” condensing “everything associated with the represented people, places or things into a single entity . . . ideological meanings, serving to legitimate the status quo and the interests of those whose power is invested in it” (Van Leeuwen & Jewitt, 2011, p. 6). Association of refugees with themes of threat or dependence on aid exists at the connotative level.

Semiotic tools are an appropriate fit for analyzing visuals of the Rohingya crisis. Awareness of sign systems can make us more critical media consumers; we become aware of mediation processes and “less likely to take reality for granted as something which is wholly independent of interpretive systems” (Chandler, 2017, p. 8). Critical insights gained through semiotic awareness make media consumers alert to broader processes and “behind-the-scenes” aspects that make them question what were previously assumed to be “natural” fits between signs and meanings. For instance, why does the image of a refugee mother with her child (sign) communicate vulnerability (meaning)? Or posing the question another way, why do media predominantly use images of refugee mothers and children to communicate vulnerability? This line of questioning inevitably involves contesting existing power structures. Kress and Leeuwen (2006) note: “Which metaphors . . . carry the day and pass into the semiotic system as conventional, and then as naturalized, and then as ‘natural,’ neutral classifications, is governed by social relations of power” (p. 8). Semiotic alertness makes us aware that meanings are not organic elements of signs; they are often ascribed and assigned by dominant cultural ideologies. For visuals, such meaning-assigning elements can include the larger discourse of humanitarianism, journalists, media outlets, and photo editors.

In the context of refugee studies, semiotics has been applied primarily at the linguistic/narrative level (Cole, 2017, 2018) From the semiotic perspective, the sign/word “refugee” is composed of the signifier/form of a person displaced from her country and the signified/meaning corresponds with such person’s inability to access protection from the state (Cole, 2018). However, the “refugee” sign is not frozen in time; it can acquire new meanings due to socio-cultural changes (Cole, 2018).

There are clear gains of applying semiotic tools to refugee visuals. For one, researchers can illuminate how meanings (vulnerability, dependence, threat etc.) attach to form (mother and child, male refugees etc.). How do things come to be? Second, semiotic tools can help track changes in meaning over time. Researchers have noted the historical evolution of “the refugee” sign over time (Johnson, 2011; Malkki, 1995). With end of the Cold War, “the refugee” in the Western imagination changed from the heroic, anti-Communist, individual fleeing political persecution to large groups of people, especially from the global South (Johnson, 2011). Tracking changes in connotations of the refugee sign is important to counter predominantly negative meanings. Cole (2018) notes that “if certain actors can attach new concepts to a particular population or group of refugees, this can prove useful for advancing new ideas and behaviours in relation to them” (p. 7).

Social semiotic framework diverges in some ways from a strictly sign-centered, semiotic approach. In traditional semiology, “signs are usually thought to be ‘available for use’” while in social semiotics “the sign is not the pre-existing conjunction of a signifier and a signified . . . ready-made . . . to be recognized, chosen, and used as it is” (Kress & Leeuwen, 2006, p. 8). Catalani (2021) notes that “visual social semiotics moves beyond the traditional semiotics’ interest in structure and system” facilitating “a dynamic study of images as manifestations of social and cultural practices in-the-making” (p. 191). In other words, social semiotics focuses on process—the how—by which relatively independent concepts of signifier/form and signified/meaning are yoked together to connote/denote meaning. In part, the “social” refers to the presence of larger cultural forces in this process of sign-making. As Kress and Leeuwen (2006) write,

social semiotics is an attempt to describe and understand how people produce and communicate meaning in specific social settings, be they settings such as the family or settings in which sign-making is well institutionalized and hemmed in by habits, conventions, and rules (p. 266).

Social semiotic approach is distinguished by the attention it pays to context both at the level of production and reception of the sign. To a visual social semiotician, the image is not the result of a singular, creative burst; it is part of “a social process . . . its meaning is a negotiation between the producer and the viewer, reflecting their individual social/cultural/political beliefs, values, and attitudes” (Harrison, 2003, p. 47). In social semiotics, “texts,” defined as “the semiotic manifestation of material social processes,” take precedence over the analytical categories of “signs” (Iedema, 2011).

Kress and Leeuwen’s (2006) social semiotic framework lays out three concurrent functions of images during the meaning-making process: the ideational/representational, the interpersonal, and the textual/compositional. Ideational metafunction engages with “the people, places, and objects within an image—the represented participants (RPs)—and answers the question ‘What is the picture about?’” (Harrison, 2003, p. 50). Ideational can contain multiple processes including symbolization embedded in action process (Harrison, 2003). Interpersonal function projects the social relations between image producers and receivers (Kress & Leeuwen, 2006). This dimension answers how images engage viewers (Harrison, 2003). Compositional metafunction ensures internal coherence of all elements in a sign; coherence with external contexts is also maintained to make meaning (Kress & Leeuwen, 2006). Visual composition can be compared with linguistic syntax; a collection of rules enabling the arrangement of linguistic signs in grammatical order to make sense to the reader (Harrison, 2003). Any semiotic analysis of visuals will have to contend with these dimensions of meaning-making.

We examined visuals using the above three image functions. For representational function, we examine such issues as whether images contain vectors indicating action, whether RPs look at each other and “connect,” and whether images contain multiple embedded processes (Harrison, 2003). To Kress and Leeuwen (2006), images can belong to conceptual or narrative representational structures. Conceptual images show RPs in generalized terms foregrounding “timeless essence”; narrative images depict RPs in contexts of change processes, actions, and events (Kress & Leeuwen, 2006, p. 59). Vectors are present in narrative but not in conceptual images; they “are formed by depicted elements that form an oblique line, often a quite strong, diagonal line”; both inanimate objects and bodies/limbs can be used to form vectors (Kress & Leeuwen, 2006, p. 59). Vectors are essential for showing directionality, connections, and relationships. RPs connected by vectors indicate “they are represented as doing something to or for each other” (Kress & Leeuwen, 2006, p. 59). Vectors are also closely related to salience in images; RPs from whom vectors emerge are often the most salient in terms of size, placement, color, and focus (Kress & Leeuwen, 2006).

For interpersonal role, we examine if visuals belong to “demand” or “offer” categories, whether viewers feel proximity to RPs, or if vertical or horizontal angles have been employed (Harrison, 2003). Demand images explicitly address the viewer; RP’s gaze and gesture demand viewers to “enter into some kind of imaginary relation with him or her” (Kress & Leeuwen, 2006, p. 118). In offer images, RPs exist as objects to be gazed at; no contact is established between RPs and viewers (Kress & Leeuwen, 2006). RPs are “offered” to the viewers “as items of information, objects of contemplation, impersonally, as though they were specimens in a display case” (Kress & Leeuwen, 2006, p. 199). For compositional metafunction, we examine placement of RPs, whether some RPs are rendered more salient than others, and distance between RPs (Harrison, 2003).

The social semiotic analysis of refugee visuals follows prior research in that area (Catalani, 2021; Dilmaç & Kocadal, 2018; Dogra, 2011). We examined images of the Rohingya crisis in two national U.S. newspapers—NYT and WP. We focus on visuals from the two outlets because of their reputation as “newspapers of record” in the United States (Benoit et al., 2005; Hansen & Martin, 1998; Meeks, 2013; Stecula & Merkley, 2019). Both NYT and WP are considered among the “most influential media sources in the U.S.” (Stecula & Merkley, 2019, p. 1). We focused on the American Press for two reasons. First, the U.S. remains “the largest single provider of humanitarian assistance worldwide” contributing more than $8 billion in humanitarian assistance in 2018 (U.S. Department of State, n.d.). Second, citizens gain awareness of distant sufferers such as Rohingyas mostly through refugee representations in the press (Dogra, 2012).

We collected all articles from the two newspapers from August 25, 2017 to August 25, 2019 that mentioned “Rohingya.” Two years was considered adequate for the analysis as prior studies have used similar timelines (Brooten et al., 2015). Articles along with the visuals were collected from Nexis Uni. In cases where visuals did not get saved along with the text, we saved images separately. Unit of analysis was each image. We did not use the newspaper article as a unit of analysis as many articles consisted of multiple images. A total of 55 articles were collected from Nexis Uni. Forty-one articles were collected from NYT and 14 from the WP. Many articles consisted of multiple images, adding up to a total of 137 images in both newspapers. Majority of these images 78.8% (N = 108) included refugees. Among those images, 50% (N = 54) contained both male and female refugees followed by 27.8% (N = 30) images showing only male refugees and 15.7% (N = 17) containing female refugees. In addition, children were present in 68.5% (N = 74) images.

For the analysis, we examined images exemplifying the themes of “mother and child” and “Exodus,” which were chosen because of their close relationship to the dominant humanitarian tropes of sentimentalism: personification, massification. Six (three for Exodus and three for mother and child) were chosen for the analysis. Care was taken to choose images from the earlier as well as the latter section of the timeline to track any changes in representation over time.

Themes in Selected Visuals

Discussion

Based on the extensive literature on refugee representations and Western humanitarianism, this visual study of the ongoing Rohingya refugee crisis in two major U.S. newspapers identified dominant tropes (personalization and massification) and two themes exemplifying those tropes: “Exodus” and “mother and child.” The analysis used semiotic tools including camera shots, camera placement, eye contact with viewers, horizontal angles, and relationships between RPs to examine how “signs” and “referents” cohere to communicate meaning. We also attended to deviations from dominant meanings in images. By and large, visuals subscribed to pervasive connotations of refugee vulnerability and gendered stereotypes. The sole deviation noted was in an Exodus-themed image that depicted a woman in the foreground leading the group.

Examination of the visuals underlined media as a deeply cultural institution leaning on familiar Biblical themes to translate the “distant suffering” of Rohingyas. The bigger question remains: What is lost in that process of familiarization? Irom (2019) notes the process of translating the suffering of strangers through culturally proximate lenses poses ethical challenges. Even without the added problems of cultural distance, it is impossible for any representation to attain complete commensurability with the lived refugee experience. To Kurasawa (2010), familiarization—reduction of the complexity of non-Western lives to narratives recognizable and relatable to Western audiences—is necessary to generate humanitarian aid and audience empathy. However, from an ethical standpoint, such reduction should not attempt to completely erase the “Otherness” of Rohingyas. In short, the study argues for critically aware representations driven by a flexible optics. Such optics should aspire toward demonstrating “common humanity” as well as maintaining distance from impulses to reduce life stories of the Rohingya strangers to familiar terms.

The semiotic analysis underlines that, despite the rich scholarly literature on pitfalls of refugee representations, very little has changed in media discourse. This demonstrates the urgent need to bridge the gap between academic/critical work on journalism and journalistic practice in the field. Journalistic practice can initiate steps to delink refugees from primarily negative connotations. This would be in the best traditions of semiotic approaches that reject “the notion that ‘things’ coincide with inherent and immutable meanings” (Cole, 2017, p. 305). As a central element of Western humanitarian discourse, media images are intimately tied in with the ideological assumptions and inherent hierarchies embedded within humanitarianism of the global North. Visual themes such as “mother and child” reflect humanitarianism’s pervasive sentimentalism. When powerful images come to reductively define the complexities of entire humanitarian crises for audiences in the global North, refugee subjects become associated with terms such as dependence and emotionality. This is not a far cry from colonial attitudes that depicted the West in terms of efficiency and strength while the East was seen as feminine and pathologically weak.

The current study is not without limitations. We examine two newspapers for a period of two years. Future research can extend this by examining more media sources, including coverage from other countries. The Rohingya crisis has been going on for many years with the latest round of violence beginning in 2017. Longitudinal study of coverage from previous years might provide comparisons helpful to understanding media representations. The current study examined only visuals. Although visuals by themselves are useful, examining narrative coverage will provide a more holistic view of the crisis. Visual-verbal congruence is an essential factor in the communication of persuasive meaning (D’Angelo et al., 2019). The study also focused on media products without looking at institutional constraints shaping how news photographers work. What routines did photojournalists follow when covering the Rohingya crisis? To Lowrey (1999), the “routines approach sheds doubt on how much control over signage the visual journalist has, as evidenced by the routine of previsualizing photo shoots” (p. 12). Future research, guided by work in organizational communication, can shed light on this gap. Among the five factors that influence media content delineated by Reese and Shoemaker (2016) in their “hierarchy of influences” model, two might be of particular interest in this regard. Upcoming research can include studies of journalistic routines; “those patterns of behavior that form the immediate structures of mediawork . . . unstated rules and ritualized enactments that are not always made explicit” (Reese and Shoemaker, 2016, p. 399). Researchers can also focus on organizational factors by studying ethnographic makeup of newsrooms and linking it to media content (Reese & Shoemaker, 2016).

We collected images from newspaper websites. Examining images posted on social media accounts of the two papers could shed light on coverage of the crisis on social media. It is beyond the scope of the present study to analyze social media content, but future research could explore that aspect of representation. Despite some of these limitations, our study is possibly one of the first social semiotic visual analysis of the ongoing Rohingya refugee crisis. The authors hope the study will contribute to “a critical mode of visual engagement” that “compels us to recognize relations of power, as well as socio-economic and cultural asymmetries between represented and viewing parties, in a manner that links the normative validity of empathetic responses to the foregrounding of structural hierarchies” (Kurasawa, 2010, p. 146). Following Kurasawa (2010), images of distant suffering can best be put to use only when they undergo a de-reifying process thus enabling viewers go beyond these images and begin recognizing and acknowledging the structural inequalities that lead to humanitarian crises in the first place. The present study is a modest step toward the de-reification of images of humanitarian disasters by helping us recognize the ways in which signs and referents cohere to construct persuasive meanings.