I wrote my article to critique the widespread assumption that translation between languages necessarily advances social justice, and to argue that it may instead discriminate against speakers of subordinated languages. I am deeply grateful to the four commentators for engaging with these points in insightful, inspiring, and personal ways, as I am to the editors for organizing this exchange. While the commentators write from different disciplinary perspectives and draw on different lived experiences with linguistic injustice, their observations are in many ways complementary, as they engage with the notion of punitive multilingualism and the underlying ideologies of language, race, and class, but also grapple with practical recommendations for improving linguistic justice. In responding to these comments, I would thus like to address them together, expanding on some of their helpful observations and building on synergies between them, with the aim to clarify and further develop my original analysis.
The relationship between linguistic diversity and social justice is central to my original article and to the four commentaries. Its examination requires a functional distinction between languages, where one language has a dominant role by virtue of its use in institutions, and other languages are subordinated to it. Yet, as Leung (p. 886) points out in her commentary, ‘some languages are more subordinated than others’, and such a gradation is also suggested by Haviland's postulation of a ‘cline … of public multilingual communications’ that range from celebratory to punitive’ (p. 872). Flores's commentary focuses on the origins of subordination, rightly emphasising its raciolinguistic dimension, as the communities that are targeted by discriminatory practices are often racialized communities, such as indigenous people and descendants of enslaved Africans in the United States, and as this linguistic subordination is an inherent aspect of their marginalization, and indeed ‘dehumanization’, as Flores writes. Di Martino's commentary in turn emphasizes social class as a factor in producing linguistic subordination, examining the relationship between the standard and subordinated sociolects in Italian society.
In exploring how translation practices relate to linguistic subordination, Leung draws on the distinction between source- and target-centered approaches, though noting that these do not ‘map neatly onto power relations’ (p. 883). But as the examples from her work on legal translation illustrate, this intersects with the direction of translation, and here it may be useful to consider Simon's (Reference Simon2006:40) distinction between translating ‘up’ (into a dominant language) and translating ‘down’ (from the dominant language into a subordinated one). In my article, I focused on discriminatory practices in translating ‘down’, both in court interpreting and in public order signage. Importantly, Di Martino shows us ways in which translating ‘up’ can also be discriminatory. Most notably this is the case in her middle-school experience of what she calls ‘punitive translation’, where a teacher compelled her to reformulate her class contribution in a more ‘refined’ sociolect. In contrast to the examples that I discuss in my article, it is the ‘translator’ who experiences discrimination here, as her speech is deemed inadequate by the institutional authority. With this example, Di Martino demonstrates that translation can be seen as a shibboleth, where linguistic difference is marked and at the same time evaluated, where one form is accepted and the other one is deemed unfit to ‘pass’ and its speakers subject to punishment.
‘Punitive’ ‘multilingualism’
In their commentaries, Di Martino, Haviland, and Leung each engage with the term punitive multilingualism and embrace it more or less, while examining its relationship to other frames of multilingual communication and other practices. Leung notes that the negative impact of punitive multilingualism is strongest on marginalized communities, whose languages are ‘more subordinated than others’ (see above). As I argue in my article, the degree of subordination of languages can be identified in part by examining the relative distribution of languages in the linguistic landscape, but importantly also by the extent to which signs project specific recognizable personae as addressees. At the same time, many languages are excluded from the linguistic landscape, particularly vernacular languages that are not habitually represented in writing. Their speakers may find themselves addressed in standardized, national languages that differ from their vernacular (e.g. Roma addressed in standard Hungarian, rather than Romani or Para-Romani (cf. Matras Reference Matras2002), Mayans addressed in Spanish, rather than in Tzotzil, etc.).
On the other end of the spectrum, examples of punitive multilingualism can also include other languages that do not index stereotypes in the same way. For example, in the German and Italian multilingual signs discussed in my article, the inclusion of English (For hygiene reasons forbidden to spit on the ground or Attention! Our goods are electronically protected against shoplifting) does not appear to have quite the same stereotyping indexical value as the inclusion of Turkish, Arabic, or Chinese does. This can be attributed in part to the fact that unlike some of the other languages, English is rarely limited to such punitive frames, but tends to be found also in other signage in the same environments, such as advertisement or way-finding signs. In fact, Leung argues that routine translation has a ‘positive norm-setting function’, and the more a language is present in the linguistic landscape, the less likely it is to be interpreted punitively (however, such a norm-setting routine can still be undermined with mocking parody, à la No smoking – el no a you smoko, as discussed in Hill Reference Hill2008:124).
Another factor in making the inclusion of English appear relatively innocuous is its status as a lingua franca, which causes these signs to be interpretable as addressing a general, rather than a specific audience (see also Haviland's example of No smoking). As a consequence, these signs do not address specific projected deviant personae, at least in places where there is no local community of L1-English speakers. The indexicality of English is likely quite different in those parts of Germany or Italy that are home to US military bases, where public order signs would have specific recognizable addressees and could potentially index racist stereotypes about African American soldiers and their families.
This shows that the distinction between punitive multilingualism and the other forms of public multilingual communications suggested by Haviland, such as ‘ambivalent’, ‘diplomatic’, ‘touristic’, or even ‘celebratory’ multilingualism, is somewhat fluid and depends on the indexicalities of a particular sign in a particular environment and on the recognizability of particular personae as addressees (e.g. ‘immodestly dressed tourists’ admonished by multilingual signs in Rome or Jerusalem; Spolsky & Cooper Reference Spolsky and Cooper1991:89). Whether a translation practice is punitive or not also depends on the perspective of the particular participant. Hungarian speakers who were not Roma did not feel offended (nor addressed) by public order signage in Toronto (Angermeyer Reference Angermeyer2017), and Di Martino notes that her teacher may have viewed translation as rehabilitative or reconstructive rather than punitive.
While not discussing the term directly, Flores's commentary reminds us that the second word in the term punitive multilingualism is also far from neutral, as the notion of multilingualism already reflects an institutional presumption of languages as discrete, bounded entities. From this perspective, institutional multingualism can be seen as inherently punitive, because it pathologizes the everyday translanguaging practices of communities in language contact situations. Flores and Leung both point out further that this essentializing of discrete languages is exacerbated by the language policies of corporations in the technology sector, be it through machine translation, through content moderation in social media, or more generally through Natural Language Processing applications that presume the use of the standard language. As Flores notes, ‘more improved technology might lead to more refined policing and pathologization of racialized communities whose language practices may be rendered further into unintelligibility by this increasing technology’ (p. 897).
Forensic translation and linguistic justice
The intelligibility of subordinated non-standard varieties is also at issue when their speakers appear in court. Di Martino discusses an example from an Italian court where a judge took the unusual step of appointing an attorney to act as interpreter for Neapolitan.Footnote 1 While this appears to have been in response to the defendants’ claim to not understand Italian, Di Martino notes that this ‘interpreter’ actually worked for the benefit of the judge. The attorney was ‘consulted for the translation and transcription of phone calls’, rather than translating from Italian into Neapolitan for the benefit of the defendants. Interestingly, even this limited use of Neapolitan provoked a significant backlash, as Di Martino shows, citing among others a blog post by Claudio Marazzini, chair of the linguistic academy Accademia della Crusca. In this commentary, Marazzini (Reference Marazzini2017) notably complains that the dialect interpreter is ‘not required to have a particular degree issued by any academy’, but is instead deemed appointable solely by virtue of having been born in Naples.Footnote 2 This illustrates the ideological relationship between institutional language use and language standardization: to be fit for use in court, a language ought to have a standardized form (ideally sanctioned by an academy), so that interpreters can be trained and examined in its correct usage. It also brings to mind Flores's point about abyssal thinking, namely the belief that the subordinated communities (in this particular case perhaps not racialized) have ‘no legitimate knowledges or practices that modern institutions need to recognize’.
Di Martino's example is reminiscent of other cases where the translation of subordinated varieties is limited to ‘forensic’ contexts, that is, where translation is used solely to facilitate the interpretation of language evidence by legal decision-makers. For example, in courts in the United States, speakers of African American Vernacular English (AAVE) are not provided with interpreters to facilitate their communication with judges or attorneys, but AAVE has been translated in cases that involved surveillance recordings of suspected criminals (Bucholtz Reference Bucholtz2009; Alim & Perry Reference Alim and Perry2010; Rickford & King Reference Rickford and King2016:955). This practice treats the subordinated sociolect like AAVE or Neapolitan as a ‘cant’ or ‘Gaunersprache’, a secret code that criminals use to avoid detection and which therefore needs to be deciphered by linguistic experts working for law enforcement.Footnote 3 Like punitive multilingualism, this association of subordinated sociolects with criminality is liable to stigmatize all of their speakers as socially deviant, rather than as members of a legitimate linguistic minority that is deserving of rights to court interpreting. Again, we have a case where translating ‘up’ becomes a discriminatory practice.
The treatment of subordinated languages as tools of deception also brings to mind John Haviland's (Reference Haviland2003:771) observation of ‘linguistic paranoia’, that is, ‘the presumption that when copresent persons use a language you cannot understand, it can only be because whatever is being said is “against” you’. In Haviland's account of this ideology in US courts, any use of another language is viewed with suspicion when its speakers are deemed capable of speaking English. Such speakers may then be perceived as threatening, secretive, and willfully disobedient (Reference Haviland2003:772). Of course, secret languages and codes do exist, but, as noted by Jones (Reference Jones2014:57), their secrecy does not need to be strategic, but may simply be ‘a surface effect of language ideology’.
While such forensic interpreting is clearly serving the interest of the institution, rather than those of speakers of subordinated varieties, Di Martino asks whether the opposite can ever be true, that is, whether the needs of these speakers can be prioritized over those of the institution. In my work on court interpreting (Angermeyer Reference Angermeyer2015), I have observed instances where this arguably did occur, as individual interpreters engaged in practices that ran counter to professional norms, but facilitated understanding by other language speakers. These practices include target-centered deictic shifts (2015:83–90), the summarizing of question-answer pairs into a single rendition (Reference Angermeyer2015:134), and in one exceptional case, the interruption of proceedings to actively verify understanding (Reference Angermeyer2015:97–99). While these practices represent exceptions, they are certainly echoed in Haviland's description of his own interpreting practice and his resistance to a professionalization that ‘greases the wheels of bureaucracy and routinizes otherwise difficult decisions’ (p. 877) or that prompts interpreters to engage in the suppression of ‘interruptions’. Flores similarly sees professionalization as tying interpreters to the institution.
Yet, even if interpreters resist some or even all of the practices that disadvantage speakers of subordinated languages, this cannot be relied upon as a sufficient remedy. Leung makes an important point when she argues for a ‘systems thinking’ that evaluates court interpreting practices in conjunction with other factors, such as the timeliness of judicial decisions. This point is also raised by Haviland, who notes that Tzotzil speakers with minimal proficiency in Spanish may prefer Spanish interpreting over Tzotzil ‘because it dramatically speeds up resolution of their cases’ (p. 878). ‘Systems thinking’ also helps scholars realize that some of the communication issues in court are not specific to interpreting. Leung notes that lay participants who speak the institutional language also lack the right to speak at will or to check their understanding. While these aspects derive from the hierarchical structure of courtroom discourse, reforms that center around verifying demonstrated understanding would arguably benefit all lay participants, not just those who rely on an interpreter, but also, and perhaps especially those who speak a stigmatized vernacular that is not afforded language status in court, such as AAVE or English-based creoles in US courts, or Neapolitan speakers in Italian courts.
Lay participants’ comprehension of legal language is already a central component of certain legal speech acts related to police cautions or guilty pleas, for example. Requiring legal professionals to sincerely verify comprehension in such cases would not be an undue burden and over time could perhaps challenge referentialist assumptions about communication. However, as illustrated by Haviland's example of a judge's admonition to a Tzotzil woman about the consequences of pleading guilty (n. 8), judges are typically content with presuming that their words are understood. Even when faced with clear evidence of misunderstanding, judges have been found to merely reiterate the same wording, rather than to paraphrase, elaborate, or explain (Dumas Reference Dumas2000).
These examples illustrate a tension faced by scholars working on language and social justice and which is evident in all four commentaries. On the one hand, many analyses locate the source of inequality in the underlying ideologies upon which institutions are constructed, and on the other, they also demonstrate a need to reduce and remedy the inequalities experienced by individuals who need to operate within these institutions as they exist today. The tensions between these perspectives are evident in Flores's account, when the embrace of translanguaging is viewed as a slippery slope towards exclusive use of the dominant language. As Di Martino notes, there are no ready-made solutions that work in all contexts, but I embrace her and Leung's call for linguists to engage in interdisciplinary work in order to link advocacy for linguistic justice into broader efforts and ultimately into the sphere of political institutions, as Haviland notes. This may be a path towards clearing local ‘hurdles’ without losing sight of the ‘horizon’ in Leung's terms, or as Flores puts it, to ‘simultaneously help to make the current world a better place while also working to imagine and create a new decolonial world’ (p. 900). The four thoughtful commentaries successfully strike a balance between these perspectives, in ways that I hope will inspire further debate, and I am grateful for the opportunity to be part of this discussion.
You have
Access
Open access
I wrote my article to critique the widespread assumption that translation between languages necessarily advances social justice, and to argue that it may instead discriminate against speakers of subordinated languages. I am deeply grateful to the four commentators for engaging with these points in insightful, inspiring, and personal ways, as I am to the editors for organizing this exchange. While the commentators write from different disciplinary perspectives and draw on different lived experiences with linguistic injustice, their observations are in many ways complementary, as they engage with the notion of punitive multilingualism and the underlying ideologies of language, race, and class, but also grapple with practical recommendations for improving linguistic justice. In responding to these comments, I would thus like to address them together, expanding on some of their helpful observations and building on synergies between them, with the aim to clarify and further develop my original analysis.
The relationship between linguistic diversity and social justice is central to my original article and to the four commentaries. Its examination requires a functional distinction between languages, where one language has a dominant role by virtue of its use in institutions, and other languages are subordinated to it. Yet, as Leung (p. 886) points out in her commentary, ‘some languages are more subordinated than others’, and such a gradation is also suggested by Haviland's postulation of a ‘cline … of public multilingual communications’ that range from celebratory to punitive’ (p. 872). Flores's commentary focuses on the origins of subordination, rightly emphasising its raciolinguistic dimension, as the communities that are targeted by discriminatory practices are often racialized communities, such as indigenous people and descendants of enslaved Africans in the United States, and as this linguistic subordination is an inherent aspect of their marginalization, and indeed ‘dehumanization’, as Flores writes. Di Martino's commentary in turn emphasizes social class as a factor in producing linguistic subordination, examining the relationship between the standard and subordinated sociolects in Italian society.
In exploring how translation practices relate to linguistic subordination, Leung draws on the distinction between source- and target-centered approaches, though noting that these do not ‘map neatly onto power relations’ (p. 883). But as the examples from her work on legal translation illustrate, this intersects with the direction of translation, and here it may be useful to consider Simon's (Reference Simon2006:40) distinction between translating ‘up’ (into a dominant language) and translating ‘down’ (from the dominant language into a subordinated one). In my article, I focused on discriminatory practices in translating ‘down’, both in court interpreting and in public order signage. Importantly, Di Martino shows us ways in which translating ‘up’ can also be discriminatory. Most notably this is the case in her middle-school experience of what she calls ‘punitive translation’, where a teacher compelled her to reformulate her class contribution in a more ‘refined’ sociolect. In contrast to the examples that I discuss in my article, it is the ‘translator’ who experiences discrimination here, as her speech is deemed inadequate by the institutional authority. With this example, Di Martino demonstrates that translation can be seen as a shibboleth, where linguistic difference is marked and at the same time evaluated, where one form is accepted and the other one is deemed unfit to ‘pass’ and its speakers subject to punishment.
‘Punitive’ ‘multilingualism’
In their commentaries, Di Martino, Haviland, and Leung each engage with the term punitive multilingualism and embrace it more or less, while examining its relationship to other frames of multilingual communication and other practices. Leung notes that the negative impact of punitive multilingualism is strongest on marginalized communities, whose languages are ‘more subordinated than others’ (see above). As I argue in my article, the degree of subordination of languages can be identified in part by examining the relative distribution of languages in the linguistic landscape, but importantly also by the extent to which signs project specific recognizable personae as addressees. At the same time, many languages are excluded from the linguistic landscape, particularly vernacular languages that are not habitually represented in writing. Their speakers may find themselves addressed in standardized, national languages that differ from their vernacular (e.g. Roma addressed in standard Hungarian, rather than Romani or Para-Romani (cf. Matras Reference Matras2002), Mayans addressed in Spanish, rather than in Tzotzil, etc.).
On the other end of the spectrum, examples of punitive multilingualism can also include other languages that do not index stereotypes in the same way. For example, in the German and Italian multilingual signs discussed in my article, the inclusion of English (For hygiene reasons forbidden to spit on the ground or Attention! Our goods are electronically protected against shoplifting) does not appear to have quite the same stereotyping indexical value as the inclusion of Turkish, Arabic, or Chinese does. This can be attributed in part to the fact that unlike some of the other languages, English is rarely limited to such punitive frames, but tends to be found also in other signage in the same environments, such as advertisement or way-finding signs. In fact, Leung argues that routine translation has a ‘positive norm-setting function’, and the more a language is present in the linguistic landscape, the less likely it is to be interpreted punitively (however, such a norm-setting routine can still be undermined with mocking parody, à la No smoking – el no a you smoko, as discussed in Hill Reference Hill2008:124).
Another factor in making the inclusion of English appear relatively innocuous is its status as a lingua franca, which causes these signs to be interpretable as addressing a general, rather than a specific audience (see also Haviland's example of No smoking). As a consequence, these signs do not address specific projected deviant personae, at least in places where there is no local community of L1-English speakers. The indexicality of English is likely quite different in those parts of Germany or Italy that are home to US military bases, where public order signs would have specific recognizable addressees and could potentially index racist stereotypes about African American soldiers and their families.
This shows that the distinction between punitive multilingualism and the other forms of public multilingual communications suggested by Haviland, such as ‘ambivalent’, ‘diplomatic’, ‘touristic’, or even ‘celebratory’ multilingualism, is somewhat fluid and depends on the indexicalities of a particular sign in a particular environment and on the recognizability of particular personae as addressees (e.g. ‘immodestly dressed tourists’ admonished by multilingual signs in Rome or Jerusalem; Spolsky & Cooper Reference Spolsky and Cooper1991:89). Whether a translation practice is punitive or not also depends on the perspective of the particular participant. Hungarian speakers who were not Roma did not feel offended (nor addressed) by public order signage in Toronto (Angermeyer Reference Angermeyer2017), and Di Martino notes that her teacher may have viewed translation as rehabilitative or reconstructive rather than punitive.
While not discussing the term directly, Flores's commentary reminds us that the second word in the term punitive multilingualism is also far from neutral, as the notion of multilingualism already reflects an institutional presumption of languages as discrete, bounded entities. From this perspective, institutional multingualism can be seen as inherently punitive, because it pathologizes the everyday translanguaging practices of communities in language contact situations. Flores and Leung both point out further that this essentializing of discrete languages is exacerbated by the language policies of corporations in the technology sector, be it through machine translation, through content moderation in social media, or more generally through Natural Language Processing applications that presume the use of the standard language. As Flores notes, ‘more improved technology might lead to more refined policing and pathologization of racialized communities whose language practices may be rendered further into unintelligibility by this increasing technology’ (p. 897).
Forensic translation and linguistic justice
The intelligibility of subordinated non-standard varieties is also at issue when their speakers appear in court. Di Martino discusses an example from an Italian court where a judge took the unusual step of appointing an attorney to act as interpreter for Neapolitan.Footnote 1 While this appears to have been in response to the defendants’ claim to not understand Italian, Di Martino notes that this ‘interpreter’ actually worked for the benefit of the judge. The attorney was ‘consulted for the translation and transcription of phone calls’, rather than translating from Italian into Neapolitan for the benefit of the defendants. Interestingly, even this limited use of Neapolitan provoked a significant backlash, as Di Martino shows, citing among others a blog post by Claudio Marazzini, chair of the linguistic academy Accademia della Crusca. In this commentary, Marazzini (Reference Marazzini2017) notably complains that the dialect interpreter is ‘not required to have a particular degree issued by any academy’, but is instead deemed appointable solely by virtue of having been born in Naples.Footnote 2 This illustrates the ideological relationship between institutional language use and language standardization: to be fit for use in court, a language ought to have a standardized form (ideally sanctioned by an academy), so that interpreters can be trained and examined in its correct usage. It also brings to mind Flores's point about abyssal thinking, namely the belief that the subordinated communities (in this particular case perhaps not racialized) have ‘no legitimate knowledges or practices that modern institutions need to recognize’.
Di Martino's example is reminiscent of other cases where the translation of subordinated varieties is limited to ‘forensic’ contexts, that is, where translation is used solely to facilitate the interpretation of language evidence by legal decision-makers. For example, in courts in the United States, speakers of African American Vernacular English (AAVE) are not provided with interpreters to facilitate their communication with judges or attorneys, but AAVE has been translated in cases that involved surveillance recordings of suspected criminals (Bucholtz Reference Bucholtz2009; Alim & Perry Reference Alim and Perry2010; Rickford & King Reference Rickford and King2016:955). This practice treats the subordinated sociolect like AAVE or Neapolitan as a ‘cant’ or ‘Gaunersprache’, a secret code that criminals use to avoid detection and which therefore needs to be deciphered by linguistic experts working for law enforcement.Footnote 3 Like punitive multilingualism, this association of subordinated sociolects with criminality is liable to stigmatize all of their speakers as socially deviant, rather than as members of a legitimate linguistic minority that is deserving of rights to court interpreting. Again, we have a case where translating ‘up’ becomes a discriminatory practice.
The treatment of subordinated languages as tools of deception also brings to mind John Haviland's (Reference Haviland2003:771) observation of ‘linguistic paranoia’, that is, ‘the presumption that when copresent persons use a language you cannot understand, it can only be because whatever is being said is “against” you’. In Haviland's account of this ideology in US courts, any use of another language is viewed with suspicion when its speakers are deemed capable of speaking English. Such speakers may then be perceived as threatening, secretive, and willfully disobedient (Reference Haviland2003:772). Of course, secret languages and codes do exist, but, as noted by Jones (Reference Jones2014:57), their secrecy does not need to be strategic, but may simply be ‘a surface effect of language ideology’.
While such forensic interpreting is clearly serving the interest of the institution, rather than those of speakers of subordinated varieties, Di Martino asks whether the opposite can ever be true, that is, whether the needs of these speakers can be prioritized over those of the institution. In my work on court interpreting (Angermeyer Reference Angermeyer2015), I have observed instances where this arguably did occur, as individual interpreters engaged in practices that ran counter to professional norms, but facilitated understanding by other language speakers. These practices include target-centered deictic shifts (2015:83–90), the summarizing of question-answer pairs into a single rendition (Reference Angermeyer2015:134), and in one exceptional case, the interruption of proceedings to actively verify understanding (Reference Angermeyer2015:97–99). While these practices represent exceptions, they are certainly echoed in Haviland's description of his own interpreting practice and his resistance to a professionalization that ‘greases the wheels of bureaucracy and routinizes otherwise difficult decisions’ (p. 877) or that prompts interpreters to engage in the suppression of ‘interruptions’. Flores similarly sees professionalization as tying interpreters to the institution.
Yet, even if interpreters resist some or even all of the practices that disadvantage speakers of subordinated languages, this cannot be relied upon as a sufficient remedy. Leung makes an important point when she argues for a ‘systems thinking’ that evaluates court interpreting practices in conjunction with other factors, such as the timeliness of judicial decisions. This point is also raised by Haviland, who notes that Tzotzil speakers with minimal proficiency in Spanish may prefer Spanish interpreting over Tzotzil ‘because it dramatically speeds up resolution of their cases’ (p. 878). ‘Systems thinking’ also helps scholars realize that some of the communication issues in court are not specific to interpreting. Leung notes that lay participants who speak the institutional language also lack the right to speak at will or to check their understanding. While these aspects derive from the hierarchical structure of courtroom discourse, reforms that center around verifying demonstrated understanding would arguably benefit all lay participants, not just those who rely on an interpreter, but also, and perhaps especially those who speak a stigmatized vernacular that is not afforded language status in court, such as AAVE or English-based creoles in US courts, or Neapolitan speakers in Italian courts.
Lay participants’ comprehension of legal language is already a central component of certain legal speech acts related to police cautions or guilty pleas, for example. Requiring legal professionals to sincerely verify comprehension in such cases would not be an undue burden and over time could perhaps challenge referentialist assumptions about communication. However, as illustrated by Haviland's example of a judge's admonition to a Tzotzil woman about the consequences of pleading guilty (n. 8), judges are typically content with presuming that their words are understood. Even when faced with clear evidence of misunderstanding, judges have been found to merely reiterate the same wording, rather than to paraphrase, elaborate, or explain (Dumas Reference Dumas2000).
These examples illustrate a tension faced by scholars working on language and social justice and which is evident in all four commentaries. On the one hand, many analyses locate the source of inequality in the underlying ideologies upon which institutions are constructed, and on the other, they also demonstrate a need to reduce and remedy the inequalities experienced by individuals who need to operate within these institutions as they exist today. The tensions between these perspectives are evident in Flores's account, when the embrace of translanguaging is viewed as a slippery slope towards exclusive use of the dominant language. As Di Martino notes, there are no ready-made solutions that work in all contexts, but I embrace her and Leung's call for linguists to engage in interdisciplinary work in order to link advocacy for linguistic justice into broader efforts and ultimately into the sphere of political institutions, as Haviland notes. This may be a path towards clearing local ‘hurdles’ without losing sight of the ‘horizon’ in Leung's terms, or as Flores puts it, to ‘simultaneously help to make the current world a better place while also working to imagine and create a new decolonial world’ (p. 900). The four thoughtful commentaries successfully strike a balance between these perspectives, in ways that I hope will inspire further debate, and I am grateful for the opportunity to be part of this discussion.