1 Introduction

Despite the proliferation of humanitarian organizations and the expansion of the United Nations’ mandate, the numbers of refugees in the world continues to increase. In 2013, the United Nations High Commissioner for Refugees (UNHCR) recorded 11.7 million refugees worldwide. Over half of this total was composed of individuals in Protracted Refugee Situations (PRS). These exist when over 25,000 refugees of the same nationality have been in a given Country of Asylum (CoA) for over 5 years [1]. The fact remains, however, that whilst a large percentage of the population could provide some form of definition for what the term ‘refugee’ means, the term’s chronic, sometimes intentional and usually unnoticed misuse stands as testimony to how poorly understood the word remains.

Even those organisations that make up the refugee regime, such as states and UNHCR, demonstrate a lack of clarity when discussing who is defined as a refugee. This is despite the scope of their activities being determined by whether or not an individual fulfils the legal definition of a refugee. There are two main reasons for this. The first is that the financial and diplomatic resources available for supporting refugees are finite. In a perfectly functioning situation, the legal recognition of individuals as refugees serves to catalyse a range of responses by UNHCR, CoAs and Countries of Origin (CoOs). This system is already chronically over-stretched and challenging to coordinate. It would, however, be entirely unfeasible without a legal framework to delimit beneficiaries and designate which actor had responsibility for refugees at each stage of their displacement. Furthermore, if states felt that their responsibilities towards refugees could continue ad infinitum, the provision of asylum may become increasingly rare [2]. The second reason, therefore, is that without a legally-binding consensus over who becomes a refugee and when, establishing when individuals should stop being refugees would be impossible.

The moment at which refugee status should end is implicit, however, within the label’s original definition. This requires individuals to fulfil two reversible criteria. The first requires an objective form, in that an individual becomes physically displaced to outside their CoO; the second requires the fulfilment of a set of legal criteria which are laid out in the definition of a refugee contained in Article 1A(2) of the 1951 Convention Relating to the Status of Refugees (herein the 1951 Convention) or other related legal documents [3].Footnote 1 This Article reads that ‘the term “refugee” shall apply to any person who…owing to a well-founded fear of being persecuted for reasons of race, religion, nationality, membership of a particular social group or political opinion…is unable or, owing to such fear, is unwilling to avail himself of the protection of that country.’ When an individual stops fulfilling one of these criteria, they lose their right to access the protective mechanisms established by the 1951 Convention. Though this most commonly occurs through individual agency, if refugees return to their CoO or acquire a new nationality, the 1951 Convention enables a declaration by actors other than the affected refugees that ‘the circumstances in connexion with which he had been recognized as a refugee have ceased to exist’ and therefore ‘he can no longer…continue to refuse to avail himself of the protection of the country of his nationality.’ This constitutes Article 1C(5), commonly referred to as the ‘ceased circumstances’ Cessation Clause.

The legal definition of a ‘refugee’ is therefore critical throughout its institutional life cycle. There is nonetheless a very limited body of literature which specifically discusses the power and role of this label. Within Refugee Studies, articles by Shacknove [4] and Zetter [5] remain the main contributions. These approaches are lacking, however, in a coherent theoretical framework to structure their undoubtedly critical observations vis-à-vis the performative and malleable characteristics of labels. Their explanations of how the label is manipulated flit between very different types of semiotic transformation. They discuss the fragmentation of the category ‘refugee’ into many new descriptive labels (paradigmatic change) and the proliferation of meanings inherent to the word itself. They fail, however, to adequately acknowledge the conceptual and political differences between these processes, or the characteristics of language that make these transformations possible.

Refugee Studies thus lacks a grounded theoretical approach for explaining how and why the meanings of words change. This has been particularly apparent in analyses of periods when actors are contesting whether the refugee label is still relevant, such as during negotiations over the applicability of the Cessation Clause to a particular caseload of refugees. Academic approaches have consisted mainly of legal analyses of whether decision-makers have kept to guiding frameworks [6–9] rather than empirical assessments of how actors have to come understand and conceptualise the labels under scrutiny. These approaches lack established insights from fields such as legal semiotics. This discipline highlights how the flexibility of law is crucial in providing space for new legal interpretations to emerge. It also stresses that as law is made and interpreted in increasingly diverse national, socio-political and jurisdictional spaces, it must no longer be seen as immune to semiotic manipulation [10, 11].

This piece therefore outlines the explanatory potential of applying a semiotic framework to protracted negotiations over the future of a particular refugee caseload; Rwandan refugees in Uganda. Over the 10 years since the Cessation Clause was first suggested for this group, the relevance and value of the refugee label has been under constant review. Tripartite negotiations between the three actors engaged over this caseload—UNHCR, the Government of Uganda (GoU) and the Government of Rwanda (GoR)—have not only focused on the practicalities of how to coordinate the cancellation of these individuals’ statuses. They have also centred on rationalising disagreement as to whether the refugee label remains valid in current circumstances. Analysing the label’s use by different actors is thus critical if we are to understand the difficulties witnessed in obtaining a consensus around the future of this group. Similarly, exploring what the diplomatic contestation of this label has meant for refugees themselves is imperative. High-level dialogues have real implications, and can precipitate unanticipated outcomes and forms of contestation. In this example, the exclusion of Rwandan refugees from discussions over their future and (intentional) misinformation about the application of the Cessation Clause has resulted in this group imagining the law as extending beyond its actual influence. First, however, the semiotic framework which will be applied to explain these divergent understandings of the refugee label and the Cessation Clause will be identified.

2 A Barthean Approach

Multiple strands of semiotics exist to explain the dynamism inherent to signs. The complementarity between the work of Ferdinand de Saussure and Roland Barthes nonetheless provides a particularly intuitive and deployable heuristic framework. Saussure’s [12] work was fundamentally concerned with illustrating the centrality of language in how we constitute our world, and the arbitrary and yet conventional ways we use and interact with it. The latter point was critical. Though he maintained that signs are ontologically arbitrary, he acknowledged that they must nonetheless adhere to systems of convention in particular places and at particular moments in time to enable processes of communication. His semiotic framework presented the linguistic sign as being composed of two elements: the signifier and the signified. This model was designed to enable the analysis of how each constituent part evolves through the sign’s everyday usage in different contexts and over time, and thus how ‘a word can express quite different ideas without seriously compromising its own identity’ [12: 129].

Saussure’s observations on the changing nature of signs were not, however, accompanied by any signification explanation of when or why this happens, or with what possible effects. This same lacuna is evident in the work of other linguistic philosophers. Pierre Bourdieu, for example, argues that ‘one can only speak of the different meanings of a word so long as one bears in mind that their juxtaposition in the simultaneity of learned discourse (the page of the dictionary) is a scholarly artefact and never exists simultaneously in practice (except in puns)’ [13: 648]. Both Saussure and Bourdieu therefore failed to acknowledge that the same sign may be simultaneously associated with multiple signifieds and, as a result, interpreted in manifold different ways.

It is here that the work of Roland Barthes [14] has enriched this theoretical approach. He expanded the signifier to include sounds, objects and images and acknowledged that a sign could be simultaneously associated with multiple different meanings. Barthes thus advocated for the politically emancipatory task of denaturalising society’s signs to reveal the multiple ways that different socio-economic groups understood them. His semiotic approach, which illustrated the multiple levels of meaning possible within a sign, was structured using the Saussurean ‘sign’ and Hjelmslev’s [15] theory on the multiple orders of signification. Barthes proposed that the first-order of signification consisted of the Saussurean sign (later called the meaning), which was composed of the signifier and the signified. At the second order, the sign from the first order is transformed from meaning to form, thus becoming a signifier, and combines with a new signified. This becomes a second-order sign. Whilst the word ‘sign’ was maintained at the first-order, the term at the second-order was renamed ‘signification’, as shown in Fig. 1. This illustrates that the original sign is both meaning and form, ‘full on one side and empty on the other’ [14: 116].

Fig. 1
figure 1

The multiple orders of the sign. Adapted from Barthes [14]

Full size image

This tiered structure was intended to illustrate the characteristics of signs which enable an ‘epistemological slide’, whereby patterns of cognitive association change over time as a result of political processes and socio-historical conditions [16]. This ‘sliding’ is not, however, unidirectional and there is some confusion in Barthes’ writing about the role that it plays in perpetuating myths. Despite his assertion at times that ‘only the letters [from the sign] remain’ in the second-order schema, this is contradicted at other moments by his clarification that ‘myth hides nothing: its function is to distort, not to make disappear’. This process of distortion aims instead to intermittently deprive the first-order sign of its full meaning by concealing its original signifier and signified. Barthes [14: 117] states that ‘the meaning will be for the form like an instantaneous reserve of history, a tamed richness, which it is possible to call and dismiss in a sort of rapid alternation…It is this constant game of hide-and-seek between the meaning and the form which defines myth’. This constant switching between meanings, however, makes myths inherently unstable and fundamentally transient. He argued that they ‘come into being, alter, disintegrate, disappear completely’ because signs at both orders can never be restricted to a finite number of interpretations [14: 119].

This anti-foundationalism in Barthes’ work creates two major problems, however, when applying his theories to real-world scenarios. The first is methodological, and the second concerns agency. In discussing the former, Duncan et al. [16] argue that though ‘it is important to recognise the instability of meaning, it is equally important to realise that this plurality is finite’. It thus can be, and must be, exposed to empirical verification. They state that ‘interpretations are the product of social contexts of historically and culturally specific discourses; they are constructed by interpretative communities and they frequently, but not always, reflect hegemonic value systems’ [16: 120]. One way of rationalising the endless ways in which signs may be interpreted has been through organising these views into common understandings shared by ‘interpretative’ or ‘textual’ communities of actors [17]. This approach may homogenise the diversity of views held by those individuals aggregated into these communities [18, 19], but such a simplification is necessary from a methodological and analytical perspective.

The second weakness concerns an under-theorisation of the role of agency in the works of the authors above. Saussure argued that ‘language presumes nothing…pieces of language are shifted—or rather modified—spontaneously and fortuitously’. Barthes [14: 117] acknowledges some agency in processes of linguistic change by observing that in myth the signified ‘is determined, it is at once historical and intentional’. His observation is nonetheless limited by a singularly Marxist discussion of the bourgeois intentions promoting transformations in the meanings of words. This narrow focus fails to explain situations when multiple second-order schemes of signification exist and compete, or when texts and words are intentionally used to create ambiguity. Umberto Eco goes some way to addressing this conceptual lacuna in ‘The role of the reader’ through the notion of ‘significative intention’. This term refers to when authors use semiotic strategies both to constrain interpretative possibilities and to ‘create a halo of indefiniteness…to make the text pregnant with infinite suggestive possibilities’ [14, 20, 21]. Though this form of analysis is limited in Eco’s work to the level of an entire message rather than an individual sign, acknowledging the multiple intentions behind actors’ use of language provides a critical addition to any politically orientated application of semiotic theory.

This article adopts this framework to explain how the word refugee is susceptible to multiple, simultaneous understandings, and discusses the implications of this for how durable solutions are envisaged and negotiated in the refugee regime. It begins by contextualising the current status of the Cessation Clause in Uganda by providing a brief history of Rwandan refugees in the country. By doing so it shows how the Clause’s current status is made possible because of the changing significance of the word refugee to each of the actors involved. On the one hand, the term has come to mean different things to each actor over time; on the other hand, different actors have simultaneously held conflicting second-order interpretations based on what these ‘refugees’ have come to symbolise for them. Finally, the more that discussions over the future of this refugee caseload turn them into ‘myth’, the more their actual bodies, voices and futures are marginalized. This follows the work of Eco [20] who observes that though signs are defined by their substitutability in our minds for something else, ‘this something else does not necessarily have to exist or actually be somewhere at the moment in which a sign stands for it.’ Conversely, that it may ‘actually be somewhere’ is of little relevance to the value of the cognitive associations which emerge in its place.

3 The Cessation Clause for Rwandan Refugees

Though the movement of Rwandans out of their country has been ongoing for centuries, the current ‘old’ caseload of Rwandan refugees fled in the post-1959 period. The late 1950s saw violence in Rwanda following the abolition of the Tutsi monarchy and its replacement by a hard-line Hutu faction [22]. Politics became increasingly defined by ethnic affiliation, and anti-Tutsi propaganda was widespread [23]. As a result, between 1959 and 1967 over 250,000 Rwandan Tutsis were thought to have fled to Uganda, Tanganyika, Burundi and Zaire [24]. This was followed by a coup d’état in 1973, commanded by General Juvenal Habyarimana, which resulted in further large-scale displacement. Rwandans continued to flee the country in subsequent decades. The displacement during and following the militarized repatriation of the Rwandan Patriotic Front (RPF) from Uganda and the Genocide of 1994 was nonetheless unprecedented. 1.2 million Rwandan refugees entered the Democratic Republic of the Congo (DRC) alone, facilitated by the French controlled humanitarian corridors demarcating the way [24].

A large proportion of Rwandan refugees were nonetheless quick to repatriate. Alongside the outbreaks of disease in the camps in the DRC, Rwandan forces attacked these spaces in an attempt to locate ex-genocidaires suspected to be evading justice and re-mobilising for return [23, 25]. UNHCR simultaneously encouraged return movements. Within weeks of the end of the Genocide in July 1994, they had assured refugees that the new Government in Rwanda had the capacity to facilitate and ensure their safe return [26]: 132]. By November 1996, it was estimated that 700,000 refugees had returned to Rwanda from neighbouring countries for various reasons, both voluntary and coerced. The international community pronounced the refugee crisis as over. Throughout the next few years, the UN nonetheless continued to ‘stress the need for the Government of Rwanda to continue its efforts to support the voluntary repatriation’ of refugees from what was then Eastern Zaire [27]. This resulted in the suspected deaths of tens of thousands more Rwandan refugees, as the RPF had assigned them a collective culpability for perpetrating the Genocide [28].

3.1 ‘First-order’ Discussions

Despite these assertive strategies in the late 1990s, over 55,000 Rwandans resisted repatriation and remained as refugees in the early 2000s [29]. This number included those that had experienced de facto local integration after decades in exile, and those that had ongoing concerns about whether their safety could be guaranteed in Rwanda due to possible retribution for their actual or alleged roles in the Genocide [28, 30–32]. Discussions nonetheless began in the early 2000s within UNHCR, and between its Offices and governments in Africa, over whether to apply the Cessation Clause to Rwandan refugees. In 2009, the Clause’s invocation was confirmed by the UN High Commissioner for Refugees, Antonio Guterres. A Joint Communique signed during his visit to Rwanda in October that year stated that it came ‘in the wake of the recent UNHCR…decision to positively respond to the insistent request of the Government of Rwanda to invoke the Cessation Clause for Rwandan refugees…[and] that given the level of peace, stability and development in the past 15 years and the political will to encourage repatriation, there is no longer any justifiable reason for Rwandans to remain in a refugee status abroad’.Footnote 2 CoAs were asked to follow a ‘roadmap of actions’ to provide all Rwandan refugees with durable solutions before the loss of their status by a generalised Cessation Clause. This was recommended to occur by the 31st December 2011 [33].

This announcement was much welcomed by the GoR due to its cancellation of the original signified and signifier of the sign refugee. These consisted of those individuals forcibly displaced out of their CoO [the signifiers] who fulfilled the definition of a refugee as laid out in the 1951 Convention [the signified]. The Cessation Clause confirmed that Rwandan refugees no longer fulfilled Article 1A(2) and thus should not be considered as remaining outside of the CoO due to either force or fear of persecution. This tacit support of Rwanda’s domestic politics was used by the GoR to delegitimise the claims of political detractors as unsubstantiated and unwarranted, and the refusal of refugees to return as being motivated purely by self-interest or criminality [34–36]. They had long feared refugees mobilising in exile and returning under arms, and there was clear evidence in the 2000s of dissenting voices establishing political momentum within the diaspora [37]. The Government’s previous attempts to coerce refugees across the border from the DRC and Uganda had been widely condemned as forced repatriation, but illustrated their strong desire for individuals to return [38, 39]. The Clause thus provided the GoR with an internationally sanctioned pathway through which to ‘encourage’ refugees to repatriate back to within their control.

The GoR’s attitude towards the Cessation Clause nonetheless concerned the other actors party to the negotiations at the time. Within UNHCR and across CoAs there were clearly conflicting views on the High Commissioner’s contention that Rwandan refugees no longer needed auxiliary international protection. Views expressed by UNHCR staff members ranged from total disagreement with the Clause’s application to this caseload, through to concerns about its scope and timing, to those that echoed Commissioner Guterres’ stance of applying the Clause to all Rwandan refugees before the commencement of 2012.Footnote 3

Though some CoAs have appeared more supportive of the Cessation Clause, for example Congo Brazzaville, others have given its announcement a much less encouraging reception. Despite the majority of Rwandan refugees residing there, the Clause has not been implemented in the DRC for a number of reasons. These include: the Congolese Government’s objections to the steps prescribed by UNHCR for applying and implementing the Clause [40]; the ongoing animosity between the GoR, the Government of the DRC, and the plethora of proxy armed groups in the region; and the sheer complexity and politicisation of refugee movements between the two countries.Footnote 4 In Uganda, the call to invoke Cessation was met by a less vocal, but equally hesitant, response. The GoU voiced concerns about the logistics of implementing such a vast project, and their continuing apprehension about the suitability of denying Rwandans access to refugee status considering their ongoing protection needs.Footnote 5

Numerous Tripartite Meetings between CoAs, the GoR and UNHCR were thus convened in the late 2000s without these parties achieving cooperation over when and how Cessation should be invoked. CoAs publicly disagreed over how the Cessation Clause should be implemented, and quietly contested that conditions in Rwanda were sufficiently improved to support the cancellation of refugees’ statuses. This opposition to the Clause’s announcement was extended by Rwandan refugees and their advocates, who objected to the suggestion that Rwandans were not in need of auxiliary state protection [41, 42]. They argued that this caseload still required the protective mechanisms catalysed by refugee status, and that their return to Rwanda could place some of their lives in danger.

The weight of this opposition, combined with the failure of multilateral discussions to achieve any consensus over the Clause’s practical implementation, resulted in UNHCR amending the Cessation Clause in 2011. The date of the Clause’s formal invocation was delayed to the 30th June 2013 to allow more time for all the actors to search for durable solutions, and its applicability was restricted to only those individuals who fled Rwanda between 1959 and 31st December 1998 [43]. These amendments ensured that those who fled after this date would continue to enjoy protection, and constituted the tacit recognition by UNHCR that the political structures in Rwanda were unable to ensure that all individuals could return safely.

3.2 ‘Second order’ Discussions

The GoR welcomed neither the temporal limitations to the Cessation Clause nor the delay in its invocation. The amendments made the physical return of all Rwandan refugees impossible and confirmed that the political situation within the country was still producing new asylum claims. The RPF nonetheless still sought confirmation from UNHCR and CoAs that the Clause would ultimately be invoked and Rwandans declared to no longer be refugees, regardless of the form that this would take. Several reasons explain this insistence that a consensus on the Clause’s invocation be arrived at in light of the 2011 amendments. Most prominently, the GoR sought to end the politically damaging discussions around why the cancellation of Rwandan refugees’ statuses have been, and should be, delayed. As deliberations over the future of this caseload have extended, the refugees have increasingly come to be seen as an indictment of domestic politics rather than as a symptom of the ongoing difficulties in securing durable solutions [44, 45]. They have corroborated negative rumours about the RPF’s domestic human rights record and its declining influence in international arenas. One GoU Minister emphasised that Rwandan refugees had come to signify the GoR’s waning authority in multi-lateral negotiations, especially with its closest neighbours and the UN institutions. The GoR has thus responded to these dates by de-emphasising the original signified and signifier of the sign ‘refugee’ in two ways intended to bolster other actors’ support for the invocation of the Clause.

The GoR’s first strategy to encourage CoAs to support the Cessation Clause has been to partially obscure the legal grounding of refugee status and instead establish new definitions to distinguish between different groups of Rwandan refugees. This has constituted an attempt to change the signifieds associated with the refugee label. Refugees affected by the Cessation Clause are presented as having fled purely due to state-orchestrated violence by the Habyarimana regime and the rebel groups that fled across the border as the Genocide ended. This is despite the country having been run by the RPF between 1994 and 1998. Representatives of the GoR and heavily state-managed media have then created a clear ambiguity surrounding who those individuals who left after 1998 are, and why they were granted refugee status.Footnote 6 Members of the GoR made clear that they wanted ‘refugees’ in the post-1998 period to be understood as individuals who had fled due to economic insecurity or personal concerns [46], or because of ‘past criminal records’ and ‘fear of persecution for genocide-related crimes’ [47]. Whilst refugees who fled pre-1998 were thus argued to have fulfilled the definition of a refugee contained within the 1951 Convention, those individuals who fled Rwanda post-1998 were presented as having fled a qualitatively different set of phenomena. These push-factors involved neither persecution nor state-orchestrated violence.

The GoR’s second strategy has been their increasing flexibility towards which durable solutions CoAs can offer Rwandan refugees. The RPF has shifted away from framing return as the only acceptable durable solution for Rwandans. It has instead begun to support any proposal, “in any way possible”, which would expedite agreement between CoAs and UNHCR over the Clause’s applicability.Footnote 7 The GoR has therefore registered a ‘major victory’ [46] in the pledges of support from CoAs and UNHCR that refugee status could be cancelled because of changed circumstances in Rwanda. This declaration in itself has been celebrated and instrumentalised as a proxy validation of the country’s peace and stability. The fact that it is unlikely to result in the repatriation of individuals en masse, or any immediate reduction in the number of Rwandan refugees, has not detracted from this celebration.Footnote 8

Organisations that have worked with UNHCR and the various governments throughout the Comprehensive Strategy for Rwandan refugees attest to this shift away from applied solutions being the central emphasis of the Cessation Clause. They have observed a distinct lack of pressure from the GoR or UNHCR to make governments and organisations translate their rhetorical commitments to implement the Cessation Clause into tangible plans for action.Footnote 9 As a result, many interviewees felt that attaining agreement over the Cessation Clause had become a “game” for the GoR. Success for them was evident in the pledges of support from CoAs for the eventual cancellation of refugee status, not in the actual implementation of durable solutions.Footnote 10 This has proven an advantageous strategy as whilst governments and UNHCR continue to understand this caseload in many different ways, arriving at a consensus over the exact physical and legal solutions for these individuals was proving impossible. Outcomes for actual refugees were thus suggested to have been blatantly deprioritised in the GoR’s strategy for ensuring support for the Cessation Clause, and in the CoAs and UNHCR’s responses to this.Footnote 11

3.3 The Epistemological Slide

The behaviour of the GoR suggests that countering the second-order significance of the sign ‘refugee’ and the ‘Cessation Clause’, which has come to connote its own politically contentious history in this context, is of greater importance than brokering solutions for those experiencing protracted displacement. The label itself has thus acquired a value and significance in negotiations which exists almost irrespective of any association with those whose forced displacement first bought it into being [48: 262–263]. This has created an increasing divergence between the signification ‘refugee’ and its original signifier and signified. Baudrillard [49] argues that this phenomenon is increasingly common in our post-modern era. Processes of hyper-circulation cause signs to become fully disinvested of their original signifier and signified. This results in their total disconnect from reality and the emergence of a value wholly divorced from their original material or conceptual form. Eco’s theory of semiotics alludes to this as he argues that ‘signs can be used to lie, for they send back to objects or states of the world only vicariously’ [20: 179].

Maintaining a partial relationship to the original signified and signifier of the sign ‘refugee’ is nonetheless advantageous for the GoR. Had they dismissed this caseload of individuals as ‘fake asylum seekers’ or ‘migrants’, the political significance of the Cessation Clause would have been lost. The GoR could not have then capitalised on the international recognition of changed circumstances within the country which accompanies a declaration of Cessation. The agreement between all actors that refugees exist in the first place also justifies the allocation of responsibilities for this caseload between the CoO, CoA and UNHCR, and produces a first-order sign which acts as a point of departure from which actors’ different interpretations of refugees can be pursued. The initial label thus enables the Barthean ‘epistemological slide’, which allows actors to draw on different meanings of the word ‘refugee’ as it suits them [16]. The risk is that the status becomes associated with such an array of different meanings that its original legal definition, and the responsibilities that UNHCR and governments owe to individuals recognised as requiring additional protection, are partially obscured and thus harder to defend.

Figure 1 shows, however, that the orders of meaning in a sign are inextricably connected. In order to have and sustain the multilateral discussions outlined above, physically displaced persons fulfilling the relevant legal definition must exist. The legal-normative vocabulary may then have remained constant throughout negotiations over Cessation, but this importantly provides no guarantee that different actors are interpreting the meaning or form of these terms in a consistent and shared way. Understanding the refugee label as a multiple order sign therefore explains two interconnected processes. The first is how it is possible for discussions over refugees to become overburdened with second-order signifieds and thus descend into negotiations dominated by empty conjecture. The second, however, is that this shift towards the second order of signification represents a slide, and not a rupture, and thus has not meant the absence of consequences for Rwandan refugees. The act of interpreting words ‘is a political practice which has material consequences’ [16], and this caseload of refugees has interpreted discussions over their future at the literal, rather than the symbolic, level. The final section thus outlines these ‘material’ ramifications for Rwandan refugees. This highlights the danger of presuming that prolonged, high-level and often hypothetical legal discussions about refugees can proceed without impacting upon these individuals’ lives.

4 Unanticipated Implications

As Ricoeur [50] notes, a ‘text’s career escapes the finite horizon lived by its author’. Once it is produced and circulated it inevitably possesses its own agency. It may cause changes in the attitudes and behaviours of its readers, or experience considerable distortion [20], regardless of the author’s initial intent [51, 52]. The lack of formal procedures to support the implementation of the Cessation Clause has not therefore resulted in these discussions causing zero change for Rwandan refugees. For example, since the dates of the Cessation Clause were amended in 2011, CoAs have in theory been free to continue granting asylum to individuals who left Rwanda after 1998. In practice, however, maintaining opportunities for some refugees to acquire status whilst simultaneously cancelling that of others has proven a complex position for CoAs to maintain diplomatically.Footnote 12 Positive outcomes for Rwandans undergoing RSD have thus reduced, as previously predicted by UNHCR staff.Footnote 13 This has been worsened by the GoR’s widespread propaganda, which presents the Cessation Clause as proof of the absence of any continuing forms of persecution or human rights abuses within the country.Footnote 14 This has undermined asylum seekers’ ability to fulfil the nexus required in Article 1A(2) of the 1951 Convention.

This politicisation has impacted upon the availability of all durable solutions for this caseload. Opportunities for third-country resettlement have reduced as the Cessation Clause’s stop-start implementation and the GoR’s narrative of changed conditions within Rwanda has served to delegitimise refugees’ claims for protection. At the Eighth Tripartite Commission Meeting on the Repatriation of Rwandan Refugees in Uganda, held in May 2010, it was clearly stated in a summary document signed by all three parties that,

In response to concerns raised by the Government of the Republic of Rwanda on possible pull factors due to resettlement processing, UNHCR Uganda informed the Meeting that Rwandan refugees are currently not being considered for resettlement, save for exceptional cases…as voluntary repatriation is being promoted.Footnote 15

Though this statement was made before the amendments to the Cessation Clause were announced in 2011, organisations providing legal counselling to Rwandans in Uganda have reported that they have been denied opportunities to work on resettlement cases. They claim that these cases have been taken away from them and handled as ‘administrative issues’ by the GoU. Furthermore, a Protection Officer at UNHCR’s main Implementing Partner in Uganda stated that a bottle neck has been created in the resettlement pipeline because of the significant pressure being applied by the GoR on traditional resettlement countries. The RPF stresses that conditions in Rwanda have changed, as evidenced by the Cessation Clause, and therefore lambasts these countries for harbouring Rwandans in exile who they claim have no genuine right to protection. Even individuals with extreme protection concerns have thus struggled to be accepted for resettlement by any country.Footnote 16

The availability of the local integration for Rwandan refugees in Uganda is equally uncertain. The GoU and UNHCR have appeared content to agree over the future of this caseload in abstract terms, rather than make politically contentious decisions which might upset the status quo. The continuing legislative ambiguity surrounding whether refugees have a legal right to acquire Ugandan citizenship provides an example of this. Despite the law being clear in allowing refugees to naturalise [53], representatives of the GoR, GoU and UNHCR maintain that a central impediment to the implementation of the Cessation Clause has been a definitive ruling on whether or not refugees can naturalise in Uganda. Each actor therefore claimed during interviews that they were constantly engaged with lobbying to resolve this legal uncertainty. In reality, however, there was no evidence to suggest that any of these actors had pushed this agenda forwards in the Constitutional Court where it is under deliberation. This was in part due to the issues’ immense politicisation, compounded by the potential implications of its resolution for thousands of refugees living in Uganda. It was also, however, the result of the value of the Cessation Clause no longer lying in the actors’ attaining concrete solutions for the affected caseload of Rwandan refugees. This has appeared to disincentivise the lobby for a definitive ruling on the issue of naturalisation.Footnote 17

Finally, the number of refugees voluntarily repatriating has been much lower than the main actors to these negotiations had predicted. The lack of preparation for the reintegration programmes designed to accommodate these individuals has nonetheless suggested no real conviction within UNHCR, the GoR or the GoU that refugees were ever likely to have returned en masse. Footnote 18 , Footnote 19 Even amongst those that have tried to repatriate, high rates of recycling have been observed as individuals have returned to CoAs following unsuccessful attempts to reintegrate in Rwanda.Footnote 20 As a commentary in All Africa stated as far back as 2011, voluntary repatriation thus ‘appears only on paper and reinforces the argument of those who oppose cessation’.

This ambiguity vis-à-vis durable solutions has resulted in confusion within the Rwandan refugee community about what impacts the Cessation Clause might have upon their status. Accessing reliable information to counter this has proven challenging for refugees. The formal systems of communication responsible for reporting updates on the Clause’s implementation have proven inert. This has been worsened by the speed with which rumours have spread through refugee communities,Footnote 21 and the false and irresponsible coverage being produced by domestic and international newspapers [57, 58]. UNHCR has arguably fallen short in its responsibility to redress this. One example concerned an article in The East African entitled ‘No More Refugee Status for Rwandans Abroad’. This stated that none of the 100,000 Rwandans outside the country ‘can enjoy international protection and assistance as refugees’ following the invocation of Article 1C(5) [44]. UNHCR quickly protested to the paper about the article’s misreporting of events, and was purportedly given the opportunity to buy a full page spread to respond with an accurate statement of events. It declined to do so, however.Footnote 22

Furthermore, the information and sensitisation campaigns for refugees have been compromised by the Clause’s application to only a certain caseload of Rwandans. UNHCR has facilitated delegations from the GoR to enter camps in Uganda to answer refugees’ questions about processes of repatriation and conditions upon return [59]. This has occurred without an acknowledgement that these camps are providing protection to a new caseload of refugees who will not be affected by the Cessation Clause. The clear presence of representatives of the Rwandan state in the spaces where pre- and post-1998 refugees are residing has resulted in these individuals showing a reluctance to attend meetings which they consider may be used for reconnaissance by the GoR. This has heightened their suspicions of ulterior motives within organisations such as UNHCR, which they have suspected to be sympathetic to the GoR.Footnote 23

This has ironically appeared to have been worsened by certain NGOs and academics who have conflated UNHCR’s more recent position on the Cessation Clause with the GoR’s misleading announcements that no Rwandan should be considered a refugee [47, 60]. Encouraged by a group of Rwandan refugees who adamantly reject the claim that Rwandans are no longer in need of international protection, reports by these organisations have entered refugee camps and fuelled speculation about how this community will be treated in the future. Rwandan refugees have responded through worrying patterns of secondary migration [47]. The fear that continuing residency in the refugee camps in South-Western Uganda would make individuals vulnerable to forced repatriation—regardless of whether they belong to the caseload affected by Cessation or not—has resulted in many Rwandan refugees leaving the camps. They have appeared to favour either de facto local integration in rural Uganda or a move to urban centres such as Kampala.Footnote 24 The danger of this, however, is that urban refugees are not covered by a clear protection framework. This has appeared to increase their vulnerability to operatives acting extra-judicially for the Rwandan state, who have been reported to violently target opponents of the RPF living in exile. [61–63]. It is thus clear that whilst the GoR, GoU and UNHCR have increasingly understood refugees in an abstracted sense, as entities signifying a host of wider political preoccupations, those marginalised on the ground have failed to distinguish between these different ‘orders’ of meaning. Rwandan refugees and those organisations working alongside them have responded by translating these symbolic discussions into very real consequences, with visible implications for the physical and psycho-social security of this caseload.

5 Conclusion

The negotiations over the Cessation Clause for Rwandan refugees have played out at a number of interconnected levels. Discussions were initiated by UNHCR and the GoR in the early 2000s to establish concrete durable solutions for these individuals prior to what was envisaged would be the eventual termination of their refugee status en masse through the Cessation Clause. Each major actor—the GoR, the GoU and UNHCR—approached these negotiations, however, with different concerns as the Rwandan refugees had come to be synecdochical to each of them for a more expansive set of issues that needed resolving. These competing priorities initially served to undermine a consensus being reached over the future of all Rwandan refugees. The amendments to the Cessation Clause in 2011 alleviated some of the actors’ concerns, such as ongoing protection-based and practical worries about the advisability of repatriating the entire Rwandan caseload. These changes, however, caused the GoR to see this caseload as predominantly signifying a different and more symbolic range of problems, beyond the initial security concerns about this group mobilising in exile. The regime in Kigali therefore began promoting any pathways that would encourage CoAs to agree to the Clause’s applicability to this caseload, regardless of whether these resulted in the return of Rwandan refugees or the establishment of any durable solutions. This provided those actors that were hesitant about invoking the Cessation Clause with increasing space to implement it, or not, in line with their interpretation of its applicability. The result was the emergence of a disjuncture between the refugee at the level of the sign and the refugee at the level of signification. Negotiations over Rwandan refugees became increasingly divorced from established legal processes and the lived experiences of the refugees themselves. This was seen through extensive commitments to the Cessation Clause, but very few attempts to institutionalise this through the provision of durable solutions.

This disjuncture has not played out so innocuously in the lives of refugees. As the application of the Cessation Clause is generally determined through Tripartite Agreements between UNHCR, the CoO and the CoA, their voices have been excluded from any formative role in the Clause’s application and execution. Multi-lateral and predominantly symbolic discussions over their future may not have corresponded with practical action by the GoU, GoR and UNHCR but this has not prevented refugees on the ground in Uganda from experiencing unexpected outcomes of the as-of-yet unimplemented Cessation Clause. The spread of inaccurate information through the social networks of these refugee communities has resulted in some organised resistance, such as a petition co-organised by a Rwandan refugee against the Clause’s application.Footnote 25 More prevalently, however, it has resulted in refugees over-imagining the reach of the law and responding in ways which have undermined their physical and psycho-social security. Furthermore, as CoAs have become increasingly unsure about how to respond to Rwandan refugees as a result of the GoR’s attempts to distort who these ‘refugees’ are and what the Cessation Clause should mean for them, these governments have hesitated to institutionalise durable solutions.

Semiotic theory highlights how the multiple orders of meaning implicit within the word ‘refugee’ enable this interplay of action and inaction. Whilst the first-order sign remains critical for catalysing an international response to refugees and for uniting CoOs, CoAs and UNHCR through a clear legal framework, the sign’s malleability enables the word ‘refugee’ to take on auxiliary symbolic meanings. These can come to obscure the term’s original legal definition and the responsibilities that states agreed to uphold upon ratifying the 1951 Convention. PRS accentuate this phenomenon as, with the passage of time, the label ‘refugee’ can become invested with evermore layers of meaning and politics. Arriving at solutions for refugees then also implicitly entails resolving the broader issues that refugees have come to signify to each of the actors involved. This has worrying implications, however, if actors stop seeking durable solutions for refugees when their political concerns have been addressed.

This theoretical approach, which synthesises insights from refugee studies and legal semiotics, challenges a pivotal conclusion often made vis-à-vis labelling in the refugee regime. Whilst certain authors have argued for a more expansive conceptualisation of the refugee label, including through accommodating refugees’ understandings of their own needs and situation, this heuristic framework implicitly cautions against this approach [4, 64]. Once individuals are recognised as refugees, the fundamental characteristics and responsibilities constitutive of the sign refugee—including its original signifier and signified—should instead be reinforced. This in turn should be complemented by the recognition of what characteristics constitute second-order signifieds, and what intentions are motivating the investment and divestment of the term with these new meanings. Understanding the refugee label through a semiotic framework highlights the characteristics and mechanisms that make language, including legally defined terms, susceptible to manipulation and distortion. This aims to ensure that the marginalised bodies and rights of refugees are held constant at centre stage, and assured of the forms of assistance that states and UNHCR have a responsibility to provide.