Introduction

Contemporary examples such as Darfur, Rwanda, Haiti, and Cambodia, and older ones, such as Chile, Argentina, or the Philippines, are transitional contexts, where a society is moving from war to peace or from authoritarianism to democracy, and addressing previous systematic abuses of human rights as a crucial part of the process. Associated with these socio-political transitions are calls for justice: the need to lay the past to rest and provide meaningful relief to victims, and—at least in some conceptions (Nagy 2008; cf. Roht-Arriaza 2006: 2)—to fashion a peaceful, just, and stable society in the future. Prosecution of perpetrators, truth commissions, reparations for victims, and reform of state and civil society institutions are the key transitional justice mechanisms (ICTJ 2009), but this article will argue that language policies can be empirically demonstrated to function as sites for transitional justice (TJ) in scenarios where oppression or conflict has sociolinguistic dimensions. Specifically, I present four brief vignettes as case studies that, while different, have similarities that can be located in a common theoretical framework for developing TJ-focused language policies (TJLPs). The cases analyzed are Taiwan, Sri Lanka, South Africa, and indigenous residential schools in Canada and (to a lesser extent) the US. These cases are empirical demonstrations of how TJLPs are or have been used in practice, which call for a theoretical framework to both explain them and devise optimum ways of applying TJLPs in the future.

A handful of sociolinguistic and language policy studies have engaged with fundamental TJ-related issues such as peacebuilding and reconciliation in countries including Malaysia (Nadzimah Abdullah and Sweehung 2012), Rwanda (Samuelson and Freedman 2010), and Sri Lanka (Herath 2015). However, these studies have avoided explicit use of the term TJ, and none have proposed a framework for understanding the issues comparatively. TJ as a field has not, in general, fared much better, though some chapters in Arthur (2011) deal with issues of language in the cases of Peru, Guatemala, and South Africa (Rubio-Marín et al. 2011; Fullard and Rousseau 2011), and indigenous residential schools in Canada (Jung 2011). Even here, however, language is treated somewhat peripherally to the core concept of cultural identity.

This article is structured as follows. It begins by outlining some definitions of TJ from the literature in legal studies and political theory, taking the position that TJ encompasses future dimensions of building or re-building peaceful and stable societies, and need not be concerned only with addressing past human rights violations. From there, it describes TJ measures and goals; the former being defined by practices in real-world contexts, the latter following de Greiff’s (2012) work on how TJ measures should be oriented towards certain broader and more abstract goals. The next section begins by delimiting case studies as only those contexts that are transitional and have sociolinguistic dimensions, then briefly outlines the relationships between language-based rights and human rights, emphasizing how language-based rights are often necessary for accessing other rights, before developing a normative TJLP theoretical framework. Four vignettes are then sketched as case studies—Sri Lanka, South Africa, and Taiwan and indigenous residential schools in Canada and the US—to understand how the framework might be applied; specifically, these outline the (often ethnolinguistic) causes of conflict and the language policies used as TJ mechanisms. The four case studies are then assessed together in terms of their relevance to the framework. I conclude by calling for scholars to develop aspects of this framework for use by practitioners, activists, and policymakers in real-world situations.

Defining transitional justice

TJ in practice has a storied genealogy (Teitel 2003; Elster 2004). Many societies in human history have addressed post-conflict situations beyond simply dividing up the spoils of war, and the Nuremberg trials, which took place between 1945 and 1946, had some (albeit limited) elements of TJ. TJ is conceptually wedded to human rights, which were codified with the UN Declaration of Human Rights in 1948. As Arthur relates, however, both in its application and as an academic field TJ is only around three decades old: spurred by work from a group of political scientists focusing on Latin America from the late 1970s (e.g. O’Donnell et al. 1986), Arthur points to the Aspen Conference of 1988 as being the crucial point for crystallizing an intellectual framework “as a response to … new practical dilemmas and as an attempt to systematize knowledge deemed useful to resolving them” (Arthur 2009: 324). It is often argued that a key reason for the emergence of TJ in its recognizable contemporary form only during the late 1980s and early 1990s was the end of the Cold War; as Forsythe (2011: 555) notes, the Cold War was “at least partially responsible for blocking a centralized international follow-on to Nuremberg”, owing to polarization between competing visions of geopolitical hegemony. For Arthur (2009: 342), however, a more compelling explanation is that former Western colonial powers “feared their own soldiers would be tried for violations they committed in the colonies”.Footnote 1

While the general principle of TJ is to address the legacy of systematic and massive human rights abuses, precise definitions of TJ are slippery and contested (Nagy 2008) and differ depending on the scope of what TJ mechanisms can be said to, or ought to, achieve. Teitel (2003: 69) argues that TJ can be defined as the “conception of justice associated with periods of political change, characterized by legal responses to confront the wrongdoings of repressive predecessor regimes”, but both Nagy (2008: 277) and Roht-Arriaza (2006: 1) problematize this definition. For Nagy, her difficulty is with Teitel’s almost exclusive “focus on (re)establishing the rule of law through legal mechanisms”, and Roht-Arriaza makes similar criticisms in terms of its “privileging the legal aspects of coming to terms with the past … [which] overvalues the role of law and legislation, and may give short shrift to the roles of education and culture and of distributional justice” (2006: 1, original emphasis). Roht-Arriaza (2006: 2) thus offers a broader definition; for her, TJ includes that.

set of practices, mechanisms and concerns that arise following a period of conflict, civil strife or repression, and that are aimed directly at confronting and dealing with past violations of human rights and humanitarian law.

For Nagy (2008: 277), the problem with this definition is the absence of the sense that post-conflict peacebuilding or post-authoritarian democratization has a future dimension beyond simply dealing with the past. Roht-Arriaza (2006: 2) anticipates this limitation, but warns that “broadening the scope of what we mean by TJ to encompass the building of a just as well as peaceful society may make the effort so broad as to become meaningless”. However, for the purposes of this article—that is, in order to see how language policy and planning, with its inescapably forward-looking orientation, may be effective as a TJ measure—I follow Nagy’s broader definition; this also accords with de Greiff’s (2012) emphasis on the need for reconciliation and democratization not only to come to terms with the past, but prevent backsliding into conflict in the future.

In practice, TJ faces three key dilemmas. The first is the problem of “victor’s justice”, which occurs when one form of domination simply replaces another; the result is a new form of oppression, with the perpetrators now on the receiving end. Arguably, this is not “justice” per se but vengeance, which simply produces a new class of victims. Once again, this has temporal implications: in the glowing embers of a post-conflict society, victor’s justice may be a tempting prospect for dealing with the past, but it has little potential for normalizing social and political relations, and thus for rebuilding a just and peaceful society. The second, and somewhat related, predicament is the difficulty of time duration between rights abuses and the TJ measures designed to address them (Winter 2013), and this gives rise to a number of dynamics that may complicate TJ efforts. These include the problem of whether democracy and peace are outcomes of TJ or, as de Greiff (2012) points out, necessary conditions for TJ to take place. In other words, TJ in practice has to be careful about putting the cart in front of the horse: does TJ guarantee democracy and peace, or vice versa? A related complexity is the “post-” prefix of “post-conflict”, since the end of war or the replacement of an authoritarian government does not necessarily signify the end of social, ethnic, religious, or political conflict. Taken together, these dynamics mean that TJ may be operating in contexts where grievances that are generations or even centuries old are not merely unconducive to its implementation but might present seemingly insurmountable obstacles. A salutary example is the peace process in Northern Ireland, though many others can be pointed to. The third problem, one that de Greiff (2012) identifies, is the emphasis on establishing the “rule of law”. As Arendt ([1964] 2006) pointed out long ago, unconscionable things can be in the name of the law: authoritarianism and conflict are often undertaken “legally” in the loosest sense of the term, and what is needed is a specifically justice-centred rule of law oriented towards universal human rights norms. Furthermore, establishing justice, on its own, cannot establish the normalization of social relations, or foster senses of belonging and acceptance, that are key to effective TJ efforts.

The very real challenges of these dilemmas notwithstanding, as a set of core practical measures TJ is primarily concerned with (a) truth-seeking; (b) criminal prosecutions; (c) symbolic, material, or rights-based reparations for victims; and (d) institutional reform (ICTJ 2009). For de Greiff (2012), these are all contingent on one another: monetary reparations may be viewed simply as “blood money” without criminal prosecutions for wrongdoing; truth-seeking is an empty exercise without concrete institutional reform measures; and so on. There are, therefore, bidirectional dependencies; without reparations, truth-seeking gives rise to empty words, while reparations require truth-seeking in order to identify who are rightful recipients of reparations, and which specific abuses reparations are intended to atone for. As de Greiff (2012: 38) puts it, a holistic approach to TJ is necessary because “individual TJ measures are less likely to be understood as justice measures if they are implemented in isolation from one another”. Thus, TJ has a semantic dimension: it must have meaning and thus be meaningful to the victims of past rights violations.Footnote 2 To this end, de Greiff’s (2012) position is that these interrelated measures are or should be oriented towards achieving two pairs of proximate and longer-term goals. The first pair are the more immediate needs to recognize victims and the harms they have suffered and build or rebuild trust in civic and state institutions. The second pair are the more enduring projects to foster reconciliation between groups previously enmeshed in highly unequal and oppressive or violent power dynamics, and to achieve democratization, which for de Greiff entails the extension of full political participation to all citizens. Again, as with TJ measures there are dependencies between these goals: reconciliation is only possible when recognition of harms and victims takes place, and civic trust is essential for democratization.

Designing a framework for a transitional justice-focused language policy

In what follows, I outline how these understandings of TJ may be conceptually applied to a theoretical framework of a TJLP. I begin by ruling in and ruling out certain contexts in terms of their relevance to such a framework. Not all unequal sociolinguistic situations are specifically transitional contexts, and not all transitional contexts have sociolinguistic dimensions. I then briefly describe the concept of language-based rights, necessarily simplifying some of the contentious aspects of the debate for reasons of space but ultimately arguing that language-based rights are or normatively should be connected to universal human rights. In light of these discussions, a normative framework for a TJLP, oriented towards TJ measures and goals, is then outlined as the main contribution of this section.

Sociolinguistic contexts with/without TJ dimensions

One of the first tasks in developing a theoretical framework of a TJ-focused language policy (TJLP) is to delimit contexts to which it may apply. Formulating such a typology involves, primarily, ruling out TJ contexts that do not have sociolinguistic dimensions, as well as those sociolinguistic contexts where languages or groups of speakers are in unequal power relationships and language-based rights may have been historically infringed, but which do not have TJ dimensions. We can see from Table 1 that Wales and France are examples where languages co-exist in unequal social status (Welsh/English and e.g. Breton/French, respectively); it should be noted that recent language policy efforts have attempted to ameliorate this, partly framed by the European Charter for Regional and Minority Languages. But despite the historical oppression of the Welsh language by successive English governments, Wales is not a society explicitly emerging from authoritarianism or conflict, and nor is France. In other words, these are situations of sociolinguistic inequality, but they are found in already-existing stable and peaceful liberal democracies, and do not constitute specifically transitional contexts (see Winter 2013). We can also rule out, for example, South Korea, though on different grounds. In the South Korean case, state/society relations have transitional dimensions, with calls for justice—accompanied and facilitated by democratization—emerging from around 1995 after a succession of authoritarian governments (Hanley 2014). However, as with Cambodia under the Khmer Rouge, political repression was not premised on identifying individuals or groups along ethnic or other lines that can be indexed by language.

Table 1 Sociolinguistic contexts with/without TJ dimensions
Full size table

This leaves us with two categories worthy of exploration, per the matrix in Table 1. The first is where transitional situations have sociolinguistic contexts with TJ dimensions. Here, we can see that South Africa, Taiwan, and Sri Lanka have experienced political repression or conflict that, at least in part, had linguistic roots or would benefit from language policies oriented towards TJ. We will examine these as case studies in more detail later. The final category is more complex, elusive, and somewhat rarer: this is where sociolinguistic contexts have transitional dimensions but are not located in obviously transitional situations more broadly. The example explored here is of residential schools to Anglicize indigenous groups in Canada or the US. As with Wales and France, the US and Canada are already-existing liberal democracies, and make the notion of “transitional” difficult to distinguish from “ordinary” justice measures (Winter 2013). But the brutality of the indigenous boarding school systems, as we will see, calls for certain transitional-like measures, not least because they were directly premised on the cultural and sociolinguistic genocide of indigenous groups and, furthermore, existed in “living memory”, making them historically proximate as transitional contexts.

Language-based rights

A divisive issue in sociolinguistics and political theory/philosophy, various theorizations of language-based rights exist, though they have only relatively recently (e.g. May 2001; Kymlicka and Patten 2003) begun to be unbundled from broader discussions of ethnicities and multiculturalism. Wee (2011: 49) critiques three related but distinct strands of work in this area, namely the linguistic ecology movement (Mühlhäusler 2000; Nettle and Romaine 2000), which need not be discussed here; the linguistic human rights movement (Phillipson and Skutnabb-Kangas 1995; Phillipson 2003); and the minority language rights movement (May 2001; 2005). For Wee (2011: 49), the minority language rights paradigm “warrants particular consideration because it provides the most thoughtful and developed articulation of language rights”. However, the linguistic human rights paradigm—that is, the articulation of language, in some form, as a basic human right—forms the bedrock on which the case for minority language rights rests.

Without engaging with the intricacies of Wee’s critiques, for our purposes we will take it as axiomatic that, in some form, language-based rights normatively ought to exist, and that instituting or strengthening language-based rights can contribute to an effective TJLP, which logically follows from the core mission of TJ to redress past human rights abuses and attempt to prevent them in the future. The term “language-based rights” used here refers to the types of rights that turn on two questions. First, and more obviously: to what extent should language-based rights guarantee the right to speak one’s language in public or private domains? Second, to what extent should language-based rights guarantee that access to other universal and inalienable human rights is not impeded on the basis of language? In slight contrast to some uses (e.g. Skutnabb-Kangas 2006), I refer to these respectively as language rights (LRs) and linguistic human rights (LHRs). This follows Rubio-Marín’s (2003: 56) formulation, in which language rights are framed as “expressive” rights, which “aim at ensuring a person’s capacity to enjoy a secure linguistic environment in her/his mother-tongues and a linguistic group’s fair chance of cultural self-reproduction”. Meanwhile, Rubio-Marín frames “instrumentalist” claims—that is, linguistic human rights—as those which:

aim at ensuring that language is not an obstacle to the effective enjoyment of rights with a linguistic dimension, to the meaningful participation in public institutions and democratic process, and to the enjoyment of social and economic opportunities that require linguistic skills (Rubio-Marín 2003: 56; my emphasis).

There are clear conceptual and semantic difficulties in even talking about “abuses” of language-based rights, let alone quibbling over terminology, not least because they pale in comparison with more egregious violations such as genocide or torture that are obviously more urgent for TJ measures. But, to the extent that they can be reasonably called abuses at all, and regardless of whether abuse is premised on action, inaction, or the simple absence of language-based rights, such abuses can index broader political repression insofar as discrimination or repression may be based on ethnicity, social position, political ideology or other attributes that can be marked through language. Indeed, language may be itself one of the roots of—or at least a pretext for—conflict or oppressions (Kontra et al. 1999: 1; Patten and Kymlicka 2003: 4; Brown and Ganguly 2003). Framed another way, discrimination can take place against individuals or groups based on the language that they speak, and linguistic discrimination may be a cipher for other forms of discrimination; thus, linguistic discrimination can, under certain circumstances, rise to the level of systematic disregard for at least some human rights. If this is the case, then it is possible to pose this problem differently. It follows that sociolinguistic reform—specifically through language policies—can be used to address abuses of linguistic and other human rights and prevent further abuses in the future. Where language is at the root of or pretext for conflict or oppression, such reforms may be fundamentally necessary, even if they are not on their own sufficient. In other cases, they can be important adjuncts to other TJ initiatives.

A normative framework for a TJLP should aim to achieve a language policy that is both compensatory and equitable. First, it must be compensatory for abuses either of language-based rights themselves, or abuses of rights with a linguistic dimension, or for political, ethnic, or other repression that is based on language. Second, it must be equitable, meaning that guarantees against further abuses are extended to all linguistic communities in society, including groups who were not necessarily victimized directly but also—perhaps especially, given the need to avoid victor’s justice—speakers of the language(s) of the former oppressors.

TJ measures and goals

As well as being both compensatory and equitable, a TJLP should be oriented towards some (or all) of the TJ measures of criminal prosecutions, truth telling, reparations, and institutional reforms, as well as—and perhaps more fundamentally—to de Greiff’s (2012) proximate and mediate goals of, respectively, recognition and civic trust, and reconciliation and democratization. It must also avoid any form of victor’s justice and contribute to the establishment of a justice-minded rule of law. Let us consider these in turn, framed in terms of language-based rights, beginning with TJ measures.

In terms of criminal prosecutions, there is little relevance for a TJLP, though one of the few linguistic human rights guaranteed in international legal instruments (Article 14, 3f of the International Covenant on Civil and Political Rights) is the right to an interpreter if a defendant cannot understand the language of the court. This right must be upheld as a basic human right and to avoid victor’s justice. There is somewhat more relevance to truth-telling. The more obvious aspect is to uncover the realities of linguistic repression, which may provide insight into the realities of other forms of political repression, but there is a further aspect, namely deciding the languages in which the truth is actually told and in which languages the truth, once designated as such, is disseminated and thus understood as meaningful. Transgressors must have the right to tell the truth in their own languages, but victims must have the right to hear it in their own languages. In terms of reparations, material reparations may be paid for past infringement of language-based rights, and more indirectly economic resources may be given to revitalizing dead or dying languages, situated in a broader project of securing language-based rights in the future. Symbolic reparations can include officializing languages or making them more visible (or audible) in public spaces.Footnote 3 Institutional reform allows previously-repressed languages to be used to access state services and thus civil, social, and political rights, and human rights more broadly. A key site for reform is the education system, to which specific mediums-of-instruction may have restricted access, and the justice system may also need to be reconstructed multilingually. At the core of all of this, of course, are rights-based reparations that specify what language-based rights are or should be, and how they should be apportioned, particularly through designating which languages are legally (and not merely symbolically) “official”.

These are of course only sketches of potential interfaces between TJLPs and TJ measures, but it can be seen that in various ways they may contribute to de Greiff’s (2012) goals of TJ, which I summarize below. Recognition and rebuilding civic trust are proximate goals on the road to achieving more lasting peace and stability through reconciliation and democratization.

Recognition and civic trust

A key dimension of TJ is the notion of recognition: recognition that harm has been done, and recognition of those who have been harmed; and, as the philosopher Thomas Nagel pointed out, recognition is not merely about knowledge of wrongdoing but also, and more fundamentally, its acknowledgement (de Greiff 2012: 42). A TJLP can recognize that harm has been done, and that harm has been done in part on the basis of language repression, either to speakers via the denial of language rights or to citizens via the denial of linguistic human rights. This aspect of recognition goes beyond establishing the simple facts of harm; instead, it acknowledges that language discrimination was sufficiently grievous to warrant intervention and recompense. Second, it recognizes the victims of such harm; there are linguistic communities and/or individual speakers in society who have suffered from unjust treatment through their misrecognition—a symbolic or actual lack of respect or concern—as non-bearers of language-based rights.

Rebuilding civic trust is necessary to refashion a functioning society (de Greiff 2012: 44–46) because a fundamental dynamic is the ability to trust (reformed) state and civil society institutions. As mentioned, educational institutions are perhaps the most direct site for the TJ dimensions of compensatory multilingual policies: those that have marginalized languages, or punished children for speaking proscribed languages, are clearly not trustworthy in the eyes of those language speakers. Similarly, reforming the institutions of the police and courts has TJLP dimensions. If citizens cannot understand, in their own language, the charges put to them then—even with the rule of law nominally operating—mistrust of the judicial system is not unreasonable. A key issue here is that rebuilding civic trust will not happen when institutions are simply relexified with another hegemonic language: compensatory and equitable multilingualism must be the guiding principle.

Reconciliation and democratization

Reconciliation can be defined as peaceful co-existence based on mutual trust, respect, tolerance and understanding. Communities riven with animosity along ethnic, racial, religious or linguistic lines can only co-exist in the future if the past is addressed; this means more than simply “wiping the slate clean”, since that leaves justice undone and resentments to fester (de Greiff 2012: 51). Instead, reconciliation requires the normalization of social relations, such that the potential for future injustices is limited. Normalizing and equalizing sociolinguistic relations may encourage previously conflicting groups to think about other aspects of identity—ethnicity, religion, or culture—along similar lines. As such, compensatory and equitable multilingual policies are thus a highly visible and wide-reaching demonstration of the commitment to broader reconciliation, both symbolically and materially.

Democratization is a more mercurial category than the foregoing. One conception is of the establishment of the rule of law, but there is an obvious problem in the fact that, as mentioned, unconscionable things can be done legally (Arendt [1964] 2006). As such, it is necessary to think of “a rule of law that ultimately involves a commitment to a more substantive conception of justice, one that calls for political participation” (de Greiff 2012: 55). Without full political participation, rights—linguistic or otherwise—are not rights at all, but simply privileges granted at the whim of the powerful. The relationship between compensatory and equitable multilingual policies is somewhat oblique, but it hinges on the notion that citizens should be able to understand the political process and be free to deliberate democratically in their own languages if we are to achieve, in Fraser’s words, “participatory parity” (1992: 118). At some level, this requires guarantees of access to multilingual and multiple public spheres (Fraser 2014: 25) and not merely, for example, the provision of multilingual ballot papers.

Applying the framework: four vignettes

In this section, I intend to briefly explore how several empirical contexts that have introduced sociolinguistic or language policy reforms as part of TJ measures can be understood through this theoretical framework. These are the countries identified in the matrix in Table 1, three of which (Sri Lanka, South Africa, and Taiwan, in no particular order) are both transitional in themselves and have sociolinguistic contexts with TJ dimensions. I also explore indigenous residential schools in the US and Canada, which are sociolinguistic contexts with TJ dimensions, but are found in (arguably) non-transitional state–society relations. For each vignette, I give an overview of the scenarios warranting TJ interventions; an empirical description of current language policies; and how these policies can be analyzed through the theoretical framework outlined above. The case studies here are deliberately explained in terms of their specific differences in historical, social, and political contexts, though in the following section they are addressed as having similarities that can be woven into a common theoretical framework of a TJLP.

Sri Lanka

From 1983, Sri Lanka was embroiled in civil war for 26 years, with government forces engaged in combat with the Liberation Tigers of Tamil Eelam (LTTE, or Tamil Tigers). The latter were claiming an independent state for the minority Tamil-speaking population, who had long been discriminated against by the Sinhalese-speaking majority after independence from Britain in 1948. The roots of the conflict itself in fact had sociolinguistic dimensions, and in briefly explaining them here I draw on the account in DeVotta (2003, 2004: Chapter 3; see also Herath 2015). British colonialists elevated Tamils to positions of administrative and political power due to their superior English-speaking abilities, leading the Tamils to be viewed as collaborators in the oppression of the Sinhalese. Post-colonial independence in 1948 was an opportunity for the Sinhalese to take their resentment out on the Tamils, beginning more symbolically with what is often referred to as the Sinhala Only Act of 1956 which replaced English as the official language, but, as the name suggests, gave no official status to Tamil. Tamils, as Herath (2015: 252) points out, were never prevented from speaking their own language in their communities, and had access to primary, secondary, and tertiary education in their own language; in 1978, Tamil was recognized as an official language in an attempt to placate Tamil nationalists. Despite these basic language-based rights being at least nominally respected, however, language was a cipher for ethnic division. Social, political, and ethnolinguistic tensions would eventually boil over to spark a conflict that would claim, according to UN estimates, upwards of 100,000 lives.

TJ was declared as a political goal by the Sri Lankan government in 2015, 6 years after the end of the civil war. Given the sociolinguistic roots of the conflict, it is unsurprising that post-civil war reconstruction efforts relied on substantial reform of language policy as one means to defuse ethnolinguistically-based historical enmities (Herath 2015). Indeed, the remit of a single government entity—the Ministry of National Co-existence, Dialogue, and Official Languages—indicates how deeply entangled the language issues are with TJ concerns. The National Language Project, among other initiatives, “is aimed at strengthening relationships between Sinhala and Tamil speaking citizens, increasing respect for language rights and linguistic diversity, and thereby fostering social cohesion and peace” (Herath 2015: 254). Sinhala speakers, particularly in the education system, are being encouraged to learn Tamil and vice versa, while the enhancement of the English skills of all Sri Lankans is being prioritized, ostensibly for national economic development in the context of globalization, but also (presumably) to avoid the re-establishment of an English-speaking elite. Symbolically, a report by the Lessons Learnt and Reconciliation Commission released in 2011 recommended that the national anthem be sung simultaneously in Sinhalese and Tamil. These measures are both compensatory and equitable. Herath (2015: 258) is sceptical that simply restoring language-based rights, whether symbolically or in actuality, is sufficient to address the past; specifically, he points out that merely officializing Tamil in 1978 by decree was not sufficient in itself to prevent civil war erupting half a decade later. Nonetheless, such efforts go some way to preventing abuses in the future, fostering civic trust, and reconciliation and democratization. Here, we can see that the narrow definition of atoning for the past advanced by Roht-Arriaza (2006: 2) is fundamentally limiting, and Sri Lanka’s adoption of TJLP measures is a practical and empirical demonstration that TJ must look ahead to the reconstruction of societies.

South Africa

The taalstryd, Afrikaans for “language struggle”, has been a fulcrum of South African politics since its colonial beginnings (Reagan 2002: 422). English became dominant in the nineteenth century as the British exercised control over the region, much to the chagrin of the majority Afrikaans-speaking descendants of the seventeenth century Dutch colonists; conflict between the two groups led to the Boer War between 1899 and 1901. Both languages were officially on an equal footing from 1925, but the rise and rule of the National Party (NP) between 1948 and 1994 increasingly prioritized Afrikaans, though English retained a role in higher education and business. The association of Afrikaans with the NP’s apartheid system—the separation and unequal treatment of black and white South Africans—meant English was also adopted by the opposition African National Congress (ANC) party, partly as a neutral lingua franca between indigenous African languages that were marginalized from official political life. Although complex, overlapping, and subject to abrupt shifts, fundamental to apartheid-era language policies was, as Orman writes, framing racial distinctions within language politics: “ensuring the survival of the Afrikaners as a distinct volk [people] was always the overriding concern … language policy was used to promote separate identities based upon a biological/racial hierarchy” (2008: 86–87). The “question of language in South Africa has been at the core of some of the most brutal stagings of apartheid’s institutional force” (Clarkson 2008: 270), and battles over language during apartheid were particularly evident in the education system, in part leading to the infamous Soweto uprisings of 1976.

The injustices and gross human rights violations—too complex and widespread to list here—committed under apartheid constitute the rationale for TJ efforts in South Africa. In 1990, the NP under F. W. de Klerk began to dismantle apartheid as a system of social and political governance; in 1994, a truth and reconciliation commission was set up, and is generally acclaimed as one of the more successful examples of the model. In terms of language policy, South Africa has the most multilingual constitution in the world, guaranteeing official status at the national level to nine African languages—Siswati, Northern Sotho, Southern Sotho, Tsonga, Tswana, Ndebele, Xhosa, Venda, and Zulu—as well as English and Afrikaans as of 1996. The constitution also established a Pan South African Language Board to promote and develop non-official minority languages, including sign language(s), the indigenous Khoi, Nama, and San languages, and immigrant and religious languages from Arabic and Greek to Tamil and Urdu. As Heugh points out, this has been accompanied by a shift away from the separatist policies under apartheid, with instead “a move towards principles that espouse the equal status and functions of eleven of the country’s languages in addition to the promotion of respect for, and use of, other languages” (2002: 449). These policies are compensatory and equitable, recognize past injustices based on language discrimination, and are oriented towards reconciliation and democratization. However, the exclusion of sign language(s) and several indigenous languages from official status raises questions about whether reconciliation can be truly achieved if some groups are left out. There are also wider structural forces at work, including limited infrastructure and poverty (especially in rural areas and among black populations) as well as the global and de facto national status of English, that might prevent equality in actuality, particularly if democratization is defined as full political participation.

Taiwan

Taiwan’s entire political history can be recounted through its languages (Price 2019), exemplifying the adage that a history of a language is a history of its people. From early contact between isolated indigenous Austronesian-speaking groups with China in perhaps the twelfth century, to Dutch colonization between 1624 and 1662, and settlement by Qing-era farmers and entrepreneurs over the next two centuries, Taiwan has long been a fluidly multilingual island. The transitional dimension, however, comes only in the last century or so of its history: colonization by the Japanese between 1895 and 1945, and—perhaps especially—the authoritarian and often brutal rule by Chiang Kai-shek’s Kuomintang (KMT) or Nationalist Party who fled from China in 1949 after defeat by Chairman Mao’s Communist Party forces after the civil war. Under martial law for nearly 40 years, Hsiau (2000) relates, KMT language policy was to spread Mandarin Chinese at the expense of the Sinitic languages Hoklo and Hakka, brought by Chinese colonists and migrants from the seventeenth century onwards, and a dozen or so indigenous Austronesian languages. Education was the main medium for Mandarin-spread, punishing pupils for speaking non-Mandarin languages and incentivizing adults to learn the language outside of formal schooling. Although Mandarin did not truly replace the majority language Hoklo, it led to substantial language shift among Hakka and indigenous languages. All these languages were marginalized from public and political life.

Despite martial law ending in 1987, for a variety of political reasons formal TJ measures were not announced until 2016 (Rowen and Rowen 2017). In the interim period, reform of language policies arguably stood in as symbolic means to achieve certain TJ goals, though they were contextualized in a broader project of de-Sinicization and Taiwanization on the part of the KMT’s successors, the Democratic Progressive Party (DPP). In terms of official status, the DPP had three options. The first was to de-officialize Mandarin, a tempting prospect for those wishing to underscore the break with KMT rule and the island’s independence from China. This, however, would have failed the victor’s justice test, as would (perhaps to a lesser extent) the second option of co-officializing the majority Hoklo language without officializing Hakka and the indigenous languages. Neither would they contribute much towards genuine institutional reform—in the case of officializing Hoklo only, this would be simply the relexification of institutions with a different hegemonic language—nor, for perhaps obvious reasons, towards reconciliation or democratization. The third option—to officialize all languages—avoids these problems; while this option was not ultimately achieved due to political and practical reasons, clearly only truly compensatory and equitable language policies meet TJ criteria. In the absence of opportunities to officialize non-Mandarin languages, DPP language policy entailed emphasizing bureaucratic and non-legislative channels (Dupré 2016: 417). As Roht-Arriaza urges for TJ in general, this required a shift in focus from law and legislation to education and culture. One of the salient achievements of the DPP was to implement mother-tongue language education in elementary schools from 2001 (Scott and Tiun 2007), although strictly speaking this had begun from the early 1990s as the KMT undertook certain reforms (or at least tolerated local governments undertaking them) as part of the process of democratization between 1987 and 1996. Similarly, the revaluation of non-Mandarin languages in public and political life was a symbolic form of reparations and recognition, oriented towards reconciliation and democratization (Price 2019).

US and Canadian indigenous residential schools

Stated quite plainly, the US and Canada were both founded on the genocide—whether literal, sociolinguistic, or cultural—of their indigenous populations by European colonists. Both contexts thus involve systematic and massive human rights abuses, but fundamentally demonstrate the difficulty in distinguishing “transitional” from “ordinary” justice measures in established liberal democracies; that is, whether TJ is required or can operate in non-transitional contexts (Jung 2011; Winter 2013). In sociolinguistic and cultural terms, a notable issue is the existence of residential boarding schools for indigenous children, whose “explicit purpose … was to destroy aboriginal language and culture, to ‘take the Indian out of the child’” (Jung 2011: 224) or, as U.S. Army Lt. Richard Pratt infamously put it, “kill the Indian to save the man”. Rife with sexual and physical violence and other abuses of basic human rights, including language-based rights, these schools existed from the late nineteenth century to at least the mid-1970s in the US and 1996 in Canada, though the effects still linger: some 10% of children in care in Canada are indigenous, disproportionate to their population of about 2%; many of these children are taken from parents with substance abuse problems, but many parents were former residents of indigenous boarding schools (Jung 2011: 239). In sociolinguistic terms, the residential schools can only have hastened the decline (or death) of North American indigenous languages and cultures, a situation which may be impossible to reverse even with the most well-intentioned policies.

The US has an ambivalent relationship with addressing past injustices, at least in part due to its suspicion—under the aegis of “American exceptionalism”—of universal human rights (as opposed to US Constitutional rights) as pertaining to itself or its citizens.Footnote 4 Canada, however, has been more forthright, taking steps to right past wrongs in more-or-less formal TJ terms, particularly through the federal Indian Residential Schools Truth and Reconciliation Commission (IRSTRC), which reported in 2015 and was established as part of the 2006 Indian Residential Schools Settlement Agreement (Nagy 2013) that also earmarked CAN$2 billion for material reparations to victims of sexual and physical abuse, as well as—perhaps unusually from the perspective of jurisprudence—compensation for lost languages and cultures that courts would usually fail to recognize. It is not yet clear what the outcomes will be in terms of a TJLP, though measures and proposals so far seem to be compensatory and equitable. Significantly, a federal Canadian Indigenous Languages Act was proposed in 2016, giving official status to eleven indigenous languages; up to now, Haque (2009) argues, Canada has been multilingual or multicultural only within a bilingual framework of French and English. This represents a symbolic move that atones for past sociolinguistic injustices—residential schools being one, if perhaps the most emblematic, aspect—and, assuming it has any teeth, guarantees language-based rights to indigenous speakers in the future. The IRSTRC called for the creation of “university and college degree and diploma programs in Aboriginal languages” (see Sterzuk and Fayant 2016), an effort at reparations with profound symbolism and one with a subsidiary outcome of building civic trust in the education system. For Jung, these efforts are fundamental to reconciliation, which “will entail revaluing indigenous languages and cultures, lost and degraded in part through the residential school system” (Jung 2011: 245). She goes further, extending the recognition of victims and wrongdoings—and emphasizing the necessity of a future dimension of a TJLP—to “include not only individual harms suffered by former students themselves, but also collective and cultural harms suffered by aboriginal communities, languages, and cultures” (Jung 2011: 227).

Evaluating the framework: some considerations

The four vignettes above are very different in many respects, including the root causes and outcomes of conflict or oppression, the historical proximity of TJ measures to the events in question, practical and political barriers to both TJ and TJLP measures, and the specific ways in which language policies have been used to address TJ concerns. This indicates the need to fashion specific TJ measures according to particular national or subnational contexts, a fact that has long bedevilled attempts at an overarching theory of TJ. There are, however, a number of similarities, which we can use to elaborate a theoretical framework of a TJLP. These are several-fold: their relationships to de Greiff’s (2012) theorization of TJ goals; the location of these contexts within the language-based rights paradigm; the difficulties posed by resource allocation and enforcement, including the problem of “symbolic” reform without actual reform; and the unintended danger that TJLP may lead to the politicization (or re-politicization) of ethnic identities. This penultimate section discusses these in turn.

Although there are differences in the precise ways that TJLPs are geared towards specific TJ measures, in the cases presented, TJLPs have been oriented towards achieving de Greiff’s (2012) two pairs of immediate and mediate TJ goals, namely civic trust and recognition, and reconciliation and democratization, respectively. All four cases seem to make contributions to these goals, with the functions of recognition and reconciliation perhaps the strongest; this is particularly the case for indigenous residential schools, which are specifically transitional dimensions of broadly non-transitional, already-existing liberal democracies. Arguably, the most salient bidirectional dependencies are not between the two goals in each pair, but between recognition and reconciliation and rebuilding civic trust and democratization; recognition of victims of harm via linguistic oppression is essential for sociolinguistic reconciliation, while rebuilding civic trust in (multilingual) institutions is essential for the project of democratization as full political participation (including in multilingual public spheres).

All four cases are located within the paradigm of language-based rights in Rubio-Marín’s (2003) theorization, whether in terms of ensuring a linguistic/cultural group’s reasonable chance of self-reproduction (“expressive” or language rights), or accessing rights with a linguistic dimension (“instrumental” or linguistic human rights). Rights are granted or restored (or, at least, claims to rights are recognized), and they are done so in ways that constitute policies that are compensatory and equitable. In the cases of Sri Lanka, South Africa, and Taiwan, these have been articulated in deliberations over the official status of various languages, the implication being that only officialization grants full rights to speakers of a given language; that is, that officialization is the organizing legal and political logic for apportioning rights. There are good reasons, however, to be suspicious of this. It is quite possible that officialization is only symbolic, especially without political commitment to overcoming practical barriers, such as poverty or unequal access to other resources, that hamper the ability of speakers and citizens to access and exercise their rights. A TJLP can enhance various other measures, but it is also dependent on their robustness for success. In other words, a TJLP is not a panacea; it may be a necessary component of TJ, but is not sufficient on its own.

In this regard, a genuine risk for a TJLP is the political mobilization (or re-mobilization) of sociolinguistically-defined ethnic identities. In other words, a TJLP is in danger of working precisely oppositely to the goal of reconciliation, by inadvertently sharpening ethnic or ethnolinguistic distinctions that may have been at the root of conflict or oppression in the first place. However, this should not be taken to mean that a TJLP should be abandoned altogether as a TJ project. Where this could be a problem is in scenarios where the TJLP process stalls, and where compensatory and equitable language policies are not achieved; that is, where only certain groups are accorded language-based rights, or where recognition of victims is limited to certain groups, or where victor’s justice takes precedence. In fact, this suggests that TJLPs may play a role in guaranteeing that other TJ measures are worked through. It is partial solutions to TJ and TJLP dimensions, rather than the solutions themselves, which risk backsliding into conflict or oppression. There is, of course, a distinction between partial and imperfect solutions: perfection is to be strived for, though political and practical reasons will almost always stymie its achievement, but this is not the same as partial solutions that do not, in good faith, attempt at full and fair dispensation of sociolinguistic justice.

Conclusion

It can be seen that, empirically, TJLPs have been put into practice in a number of contexts, including in the case studies outlined above. In the cases of Canada, South Africa and Sri Lanka, they have been directly related to TJ measures, while in the case of Taiwan TJLPs have stood in for other TJ measures in the absence of political will or practical means to achieve them. In all four cases they have all been oriented to de Greiff’s (2012) goals of recognition and rebuilding civic trust and, especially, reconciliation and democratization. This suggests that practitioners and policymakers see value in using TJLPs in real-world contexts, but what has so far been lacking is a theoretical framework to explain how they work or offer normative guidance on what is optimum. I do not suggest that this article offers a fully worked-through framework, but that one is worth exploring collaboratively by scholars, practitioners, and policymakers. While insufficient on their own, they may be necessary in transitional contexts with a sociolinguistic dimension. TJLPs can augment other TJ measures and be effectively oriented to TJ goals; they can help atone for the abuse of language-based rights in the past, and contribute towards building stable, just, and peaceful societies in the future.