Introduction

Dual language (DL) education is one of the fastest growing enrichment educational models in the United States, providing grade-level content knowledge through English and another language to achieve high academic achievement, bilingualism, biliteracy, and intercultural awareness (Howard, Sugarman, Christian, Lindholm-Leary, and Rogers 2007). Research has shown that students who attend a well-implemented DL program tend to academically surpass those who do not, and this holds true regardless of students’ race, class or dominant language (Lindholm-Leary and Howard 2008; Marian, Shook, and Schroeder 2013; Thomas and Collier 2002, 2012). However, DL can be particularly effective at closing the educational achievement gap for EL students (Lindholm-Leary and Borsato 2006; Rolstad, Mahoney, and Glass 2005; Thomas and Collier 2002). It is important to note that there are three general types of DL programs implemented nationwide: one-way foreign language immersion models (serving primarily monolingual English speakers), one-way developmental bilingual education models (serving ELs and heritage speakers of the target language), and two-way immersion (TWI) models of DL (serving both populations of students) (Hamayan, Genesee, and Cloud 2013).

DL education in the U.S. emerged within a policy era when ELs were being offered an array of multilingual educational programs in several states, with DL being the most effective (Lindholm-Leary and Hernández 2011; Thomas and Collier 2012; Valentino and Reardon 2014). Nationally, this era came to a close at the end of the 1990’s when in many U.S. states, including California, Arizona, and Massachusetts, there began a dismantling of bilingual educational models that provided instruction exclusively for ELs in both English and other languages; meanwhile TWI programs in these states avoided this dismantling because they included non-ELs (Gándara, Losen, August, Uriarte, Gómez, and Hopkins 2010; Flores and Murillo 2001; Linton and Franklin 2010). This new era has been pro TWI but anti bilingual education and has spurred cautionary notes from leading scholars in the field who saw DL, which overwhelmingly appeared in its TWI variety, often being reserved for “gifted and talented” students and other already privileged groups while often “whitestreaming” the non-privileged students that it was serving (Cervantes-Soon 2014; Flores 2015; Morales and Rao 2015; Valdés 1997).

In the midst of this national pro-TWI but anti-bilingual-education era, there has also been an explosion in the number of DL programs, particularly in Utah. By 2008, Utah had passed state legislation that dramatically shifted language education policy toward a unique form of state-level support for DL programs. Utah’s 106 DL programs represent approximately 10 % of the 1049 self-reported one-way foreign language and TWI programs in the U.S. (Center for Applied Linguistics (CAL) 2011, 2015)—an influential portion considering that Utah has less than 1 % of the U.S. population. Recent reports cite TWI programs appearing to constitute the predominant form of DL in most of the country (Anderson 2015a; Maxwell 2015). Utah, however, represents a different trend—a surge in one-way foreign language immersion DL programs. Nationally, until Utah’s 2009 surge, the one-way foreign language immersion variety was rare and slow growing compared to the other DL types (Lenker and Rhodes 2007). As of 2014–2015, more than three fourths of Utah DL programs were of the one-way foreign language immersion variety, the remainder being TWI.

Keeping in mind the English hegemony (Macedo, Dendrinos, and Gounari 2003) that has periodically stifled bilingual education options in the U.S., these developments raised cautious optimism in us about a resurgence of equity-enhancing, educational opportunities. Yet as Utah is a largely White (80 %), conservative state, which in 2000 enacted English-only legislation, we were skeptical of the interests served by the new policy. Our previous research on how news media coverage evolved as the policy emerged revealed a shift away from discourses of achieving equity through using home and heritage languages for instruction and toward discourses that reflect more privileged constituencies (Valdez, Delavan, and Freire 2014). It also revealed state officials being quoted using rhetoric that we found problematic. For example,

“Our main goal is to mainstream immersion,” said … the world language specialist at the Utah State Office of Education. “In the past, it has been a boutique program for elite private schools.” (Stuart, 2010, emphasis ours)

In light of our understanding that DL had largely been used for equity purposes in language marginalized communities—and thus our disagreement with the claim that it had been largely used for elites—we wondered what was accomplished by the discourse of mainstreaming to point out the uniqueness of Utah’s policy (Leite and Cook 2015; Roberts and Wade 2012).

Grounded in a critical language policy framework focusing on hegemonic discourses of normative whiteness, globalized human capital, and English privilege and their material effects, we view educational language policies as a central gatekeeper to education (Tollefson and Tsui 2014). Thus, our critique keeps foremost in mind the academic gaps that exist between student groups and considers equitable educational policies to be those that seek to achieve equal academic outcomes for students often through unequal means (Carter and Welner 2013). Taking a mixed methods approach, we first undertook a critical discourse analysis of core state DL policy documents—those that best defined the shape and details of the 2007/2008 DL policies. We then conducted quantitative analysis of school demographic data comparing all Utah DL programs in place across two periods of time: (a) the period of time after the 2008 state-level DL policy was in place, which we refer to as the state-model period, and (b) the period of time prior to 2006 when DL programs were clearly not influenced by the DL policy, which we call the pre-state-model period.

We specifically posed the following research questions: (1) How do the core Utah DL policy documents reflect or counter, if at all, the three hegemonic discourses in operation on U.S. language education across race, class, and language status and with what effects on the targeting of beneficiaries for the policy? (2) How do the types of DL programs implemented in the pre-state-model period (before 2006) compare with those implemented in the state-model period (after 2008)? (3) What are the demographic distributions and their statistical significance in terms of race, class, and language status across the student populations of the Utah schools in which DL programs have been housed in the pre-state-model period versus the state-model period? Using critical discourse analysis, our discursive findings suggest that the core policy documents largely participate in the hegemonic discourses that help enforce these privileges. Our demographic findings suggest with statistical significance that Utah’s DL policy has led to the placement of programs in schools that tend to have more students privileged with speaking English, with having wealth, and with white racial markers.

We are persuaded that Utah’s “mainstreaming” of DL is really a kind of gentrification, that is, an influx of more privileged inhabitants into a ghettoized neighborhood while less privileged residents are priced or pushed out. We coined the term “gentrification” to describe trends in DL that have pushed out ELs and other non-privileged students from multilingual education options (Valdez, Delavan, and Freire 2013). We see this process as broader than what Flores (2015) refers to as the “Columbusing” of TWI due to the fuller appropriation of DL that is occurring as one-way foreign language immersion models of DL grow at the expense of other types of DL that include non-privileged students. We build from the understanding that DL and similar forms of multilingual education originated in the U.S. primarily through grassroots struggle and as an equity measure for marginalized students and were of little interest to others; now, much like in a gentrifying neighborhood, privileged families are those primarily being invited to join DL programs, and they are poised to outnumber DL’s traditional clientele and thus dilute DL’s equity effects (Anderson 2015b). We have reason to believe these gentrification trends suggest a neoliberal pattern of diverting attention away from equity and heritage concerns for marginalized student populations (Cervantes-Soon 2014; Petrovic 2005) while allowing existing educational inequalities to remain uncontested (Macedo et al. 2003). Finally, we argue that DL education may soon constitute the next wave of inequitably distributed enrichment education in the U.S. by following a broader pattern of enrichment tracking practices, such as in gifted and talented programs (Baldwin 2005; Ford and Grantham 2003) and Science, Technology, Engineering, and Mathematics (STEM) programs that have tended to further privilege already privileged students (Museus, Palmer, Davis, and Maramba 2011). With Utah’s DL policy now influencing more than 30 states and with Delaware and Georgia explicitly emulating Utah’s model, we argue that anyone in favor of socially just education in the U.S. should be wary of losing the equity effects of DL education. We conclude by offering DL policy recommendations that can be advocated for in other states considering this educational option.

A Critical Language Policy Framework

Language policy and planning is a process of “language engineering” intended to influence the language practices within a society (Baldauf 2004). We situate ourselves within the field of critical language policy (Tollefson 2006); we view language education and its policies as political and direct our efforts toward equalizing power. Our understanding of power and its unequal distribution draws on the relationship between three concepts: James Gee’s concept of discourse, Antonio Gramsci’s concept of hegemony, and Pierre Bourdieu’s concept of multiple forms of capital. Through these frameworks, we view power as discursive as well as tied to tangible material effects. First we describe the three hegemonic discourses we see operating on DL education based on our review of the literature. We then theorize how student demographic data from schools where educational programs are located can reflect the privileging and marginalizing effects that correspond to these hegemonic discourses.

Hegemonic Discourses of Normative Whiteness, Globalized Human Capital, and English Hegemony

Gee (2012) describes discourse as “a socially accepted association among ways of using language, of thinking, feeling, believing, valuing, and of acting” or playing a particular social role (p. 143). Dominant discourses are (re)produced through the social action of everyday life in institutions, communities, families, and individuals to reflect the principles and values that comprise seemingly universal truths. They are inherited from the past and often passively accepted and reproduced in ways that maintain power imbalances, although with some degree of resistance. Gramsci argued that hegemony, similar to dominant discourses, resulted from passive consent to power imbalances rather than the brute force of the state or the dominant group. We adopt the term hegemonic discourse (and the potential for counter-hegemonic discourses) in order to merge these parallel conceptualizations of how power operates in society.

Three primary hegemonic discourses shape language politics in the U.S. The first is normative whiteness as described by Clayton Hurd, which Palmer (2010) used to theorize the subtle racism operating in DL classrooms. Hurd (2008) argues that hegemonic discourses of normative whiteness “privilege and sustain the dominance of white imperial Eurocentric worldviews” (p. 294) and their “particular ways of understanding history, citizenship, [and] notions of self and other” (p. 295). Normative whiteness is then a collection of discourses that positions those who look and act white as the normal, central, common-sense beneficiaries kept often unquestioningly in mind as policy is designed and implemented in the U.S. context.

The second hegemonic discourse affecting language education policies is neoliberalism, which promotes “an extension of market rules and principles to public and private sector organizational restructuring” (Olssen, Codd, and O’Neill 2004, p. 153), which often involves promises of greater equality of wealth through a discourse of “a rising tide floats all boats” or by channeling more wealth to the wealthy and trusting it will “trickle down” (Gould and Robert 2013). The post-9/11 upsurge in the valuation of multilingualism and language education (Wible 2009) has been theorized to be associated with neoliberal discourses of globalization and human capital (Fairclough 2006). This hegemonic discourse has been involved in signs of a major shift from an equity/heritage to a global or globalized human capital framework in DL policy discourse (Valdez et al. 2014). An equity/heritage framework is dominated by counterhegemonic value discourses related to educational equity for ELs and other marginalized groups as well as concern for heritage language and culture loss. A globalized human capital framework is couched within a larger neoliberal process of commodifying language diversity and learning (Cervantes-Soon 2014; Duchêne and Heller 2012; Heller 2003). It focuses on value discourses related to producing multilingual workers to compete in the global marketplace (Valdez et al. 2014). Henceforth we will use the concept of globalized human capital to describe the dominant neoliberal discourse specific to language education.

The third hegemonic discourse impacting language policies is English hegemony. Critical language policy scholars have used the concept English hegemony (Macedo et al. 2003; Shannon 1995) as a way of describing the power and privilege that the linguistic positionality of speaking English gives its speakers in the U.S. We differentiate the discursive formation by which English retains its commonly accepted prestige and perceived practicality, termed English hegemony, and the related material effect that we term monanglicization Footnote 1—language shift toward monolingual English competence that occurs for the vast majority of people in the U.S. despite home and community exposure to other languages (García 2003; Wong Fillmore 2000). English hegemony and monanglicization have meant that the U.S. education system has until recently treated multilingualism for the general public as either a threat or a low priority (Crawford 2007).

Material Effects of Hegemonic Discourses: Privilege via Capital

Though we understand power as discursive, such as the language used to set policy, we identify some of its material effects as demographic patterns and access to resources or opportunities. Pierre Bourdieu’s (1986) concept of capital helps us understand these material effects. The material effects of hegemonic discourses, although always contested, often translate into enduring unequal distributions of economic, cultural, and social capital in society where some remain privileged and others marginalized. We use the terms privilege and marginalization as shorthand for this understanding of power.

In this study we measure the equity-related material effects of the three primary hegemonic discourses operating in language education across three higher-valued types of capital—white racial privilege, wealth, and English privilege—made visible in data that uses the demographic categories recorded by schools for race, free/reduced lunch status, and English proficiency. We see white racial privilege as the material effect of normative whiteness by which some individuals are racially classified as white and non-Latina/o, allowing them to wield this symbolic capital like cash. When we refer to wealth as a form of privilege, for purposes of this paper, we specifically focus on poverty as a descriptor of the lack of this form of capital. This lack of wealth has been identified by researchers as having the most significant implications for marginalizing students’ educational opportunities in U.S. schools (Reardon 2011; Rebell and Wolff 2012). Finally, we define English privilege in the U.S. as involving the capital advantage of entering school already speaking the nearly uncontested language of instruction. For purposes of this paper, we focus on English learner status as a reflection of low levels of the capital of English privilege. These three forms of capital often work in combination to create their material privileges. English learners, for example, often face marginalization across the three demographic categories we study: racial, linguistic, and socioeconomic (Gándara 2010). The concept of elite versus folk multilingualism (De Mejía 2002) offers us a key insight into how the intersectionality of these forms of privilege operates. Elite multilingualism (acquired by linguistically privileged groups via formal education and usually associated with a higher economic status) has long been treated as a more valuable form of cultural capital than folk multilingualism (acquired via family and community transmission in linguistically marginalized communities and usually associated with a lower economic status). Thus, built into the concept of elite versus folk multilingualism is a class distinction that often intersects with race. In the shadow of English hegemony, knowing a non-English language via folk multilingualism can racialize its users—de-whiten them to the same extent that the language in question is less associated with the western Europeans who have historically held power in the United States—whereas knowing a non-English language via elite multilingualism does not racialize or de-whiten its acquirers. Conversely, because of intersectionality with normative whiteness, a) someone classified as a person of color who acquires a prestigious language variety through formal schooling may not be awarded the full privileges of elite multilingualism, and b) someone classified as white who acquires a stigmatized language variety through contact with a local community might be able to display this knowledge without receiving the same stigma.

Background

To situate our study, it is important to establish a brief history of Utah’s ELs and the divergent influences of bilingual education and foreign language education on language education policy in the U.S. First, Utah has the eighth fastest growing population of students entering school speaking a language other than English (Pandya, Batalova, and McHugh 2011) going from an EL population of 41,000 in 1990 to 137,000 in 2010. The rapid growth of ELs led to growing pains in the state’s efforts to serve them in Utah schools. In the 1990s and into the 2000s, after the U.S. Office of Civil Rights (OCR)—the federal office responsible for enforcing ELs’ meaningful participation in educational programs—initiated and monitored compliance reviews, 10 of Utah’s 40 school districts were found to be out of compliance in services to ELs and were required to develop corrective plans. Utah efforts to resolve these OCR compliance issues around EL education stretched into 2009. Simultaneously, the Utah State Office of Education (USOE) assessment and course data has shown a persistent achievement gap exists in Utah for students with low levels of white racial privilege, wealth, and English privilege that range from two to three grade level disparities (Lambert 2009). These issues all point to the need for a continued focus on equity/heritage concerns in Utah EL education.

We believe the stakes of Utah’s policy choices on these equity/heritage issues have now risen as Utah takes a leading role in shaping U.S. DL policy. To understand this, one must recognize that DL is historically understood as developing along two divergent lineages within U.S. language education: foreign language and bilingual education (Baker 2011; Lenker and Rhodes 2007). At the most basic level, foreign language education has focused on serving those who already have English privilege, while bilingual education has focused on serving those for whom English is not a primary language and therefore lack English privilege. The most critical distinction between these two lineages is bilingual education’s recognition of the effects of English hegemony and monanglicization in the U.S. on ELs and the need for proactive, equity-based policy responses to balance out power differentials (Cummins 2000)—counterhegemonic equity/heritage discourses that have been largely marginalized within foreign language education.

In the wake of the nationwide negative public perception of bilingual education that swept the U.S. in the late 1990s (Crawford 2007), two divergent approaches to bilingual education emerged in high EL enrollment states and communities. The first approach centered in states such as Texas, Illinois, and New York that maintained bilingual education programs, including one-way developmental DL programs for ELs in addition to TWI. The second approach happened in states that explicitly passed anti-bilingual education measures such as California, Arizona, and Massachusetts who closed most bilingual education programs that were not TWI. For example, in California, after the bilingual education ban, bilingual education programs dropped sharply while TWI programs grew at an unabated pace driven by the enrollment of English privileged students with high levels of capital whose interests converged with those of ELs enrolling alongside them (Linton and Franklin 2010; Wentworth, Pellegrin, Thompson, and Hakuta 2010). We are now seeing a nationwide rise in foreign language immersion DL programs that serve primarily students with English privilege (CAL 2011). Although this trend predates the spread of Utah’s DL model (Dobnik 2007; Toomer-Cook 2007), there is mounting evidence that Utah is being looked to as the standard-bearer of this new DL landscape. More than 30 states have observed Utah programs and at least eight states are using its design and curriculum materials (USOE 2013). Furthermore, the son of a Utah state senator who championed the policy is leading a foundation backed by Chinese economic capital to spread Utah’s Chinese foreign language immersion DL model nationwide (Zhang 2014).

Methodology

Our research drew on data related to DL policy and school demographics to address our research questions. In any form of research, the questions posed and the interpretation of the findings are necessarily shaped by the researchers’ positionality (Kleinsasser 2000). Research bias does not come from having a position, but rather from not acknowledging one. Revealing one’s positionality helps to “unmask any bias that is implicit in those views” and “provide a way of responding critically and sensitively to the research” (Griffiths 1998, p. 133). As critical scholars we now offer our positionalities in relation to the forms of capital that are the focus of our study—language education, white racial privilege, wealth, and English privilege. As members of the research team, we bring a range of positionalities that include White and Latina/o identity markers. In regards to wealth, although one of us was raised in poverty, we are all middle class with current or prior work experience in Utah as teachers in a Spanish/English two-way dual language education (TWI) program, higher education teacher preparation program, or secondary education Spanish program. Our linguistic positionalities range from a simultaneous Spanish/English bilingual, to a native Spanish speaker who learned English through formal education, to a native English speaker who acquired Spanish through formal education. None of us are products of a DL program or anything similar; however, all of us have been participants in some form of foreign language education as children or adults and are trained in some form of multilingual education. Committed to promoting language education for all students with a specific interest in equitable access to education and the forms of capital it distributes, we next describe the process by which we collected and analyzed our data to address our research questions.

Critical Discourse Analysis of Utah DL Policy Documents

We gathered five core policy documents: Utah Senate Bill 80; Utah Senate Bill 41; Utah Administrative Code R277-488 (the Administrative Rule) (2012); the Utah Dual Language Immersion text (DL Overview) (USOE 2015c); and the 2013 Critical Languages: Dual Language Immersion Education Appropriations Report (the Legislative Report) (USOE 2013). In 2007, the state legislature passed Utah Senate Bill 80, followed in 2008 by Utah Senate Bill 41, which we gathered for study because these bills created, piloted, and began a process of funding the Critical Languages Program, a law that (in part) supports DL programs at public schools. We then gathered the most recent 2012 rules guiding implementation—the Administrative Rule. As a means of clarifying how the Administrative Rule and the USOE treat the required credentialing for DL teachers, we also gathered a USOE document used to provide an overview of Utah’s DL initiative and its model—a defacto policy document which we refer to as the “DL Overview” (USOE 2015c). Lastly, we gathered the Legislative Report, a report the USOE released to the Utah legislature that gave details on implementation of the DL policy framework and its results. After gathering these five pertinent policy documents, we then conducted a critical discourse analysis (CDA) of the legislation. Following Rogers (2004), we believe that there are “no formulas for conducting CDA” (p. 7), thus we sketch the particular iterative steps we took in this analysis of the power effects within the dataset texts.

The first level of our analysis was a qualitative content analysis to establish a concrete, surface-level, descriptive picture of the policy from the content of the policy texts to create a context for what we would look for more specifically. For example, we identified (a) parts of the policy texts pertinent to the elementary DL programs and (b) the genre of the various sections of the Legislative Report, noting that it included the DL history, growth patterns, program model, performance measures and student outcomes, a link to a DL promotional video, and a national magazine article that profiled the programs. The second level of analysis was to locate what was present in the texts that reflected, implied, or countered the three hegemonic discourses described in the theoretical framework, using them as a priori codes that paralleled the demographic categories of race, class, and language status found in our quantitative data. For example, with English hegemony, we looked for segments of the documents where language, languages or language groups were named or discussed; for globalized human capital, we looked for where its common keywords appeared or where economics, class, or socio-economic status were mentioned; and for normative whiteness, we focused on where racialization was mentioned or implied. The third level of analysis was to look for what was absent or discursively silenced (Thiesmeyer 2003) in the content. For example, as it became clear that the texts were countering the monanglicization that is most often the material effect of English hegemony, we were led to ask what parts of English hegemony were left silently un-countered in that process. The fourth level of analysis was to make sense of the patterns that emerged from the second and third levels in order to posit power effects across the local, institutional and societal domains in which discourses can operate (Fairclough 2014). This fourth level sought to describe whether and how the texts exerted power in the larger policy context in concert with or against the hegemonic discourses and what picture emerged of the beneficiaries within those patterns. For example, what at first appeared a surprisingly inclusive, counter-hegemonic definition of “dual language immersion” suddenly began to look like it did exclusionary, hegemonic work when successive re-readings of the segments identified in levels one, two, and three found the emphasizing and silencing of particular student beneficiaries.

Demographic Analysis to Identify Material Effects

We initiated our study in Spring 2010. While we hold an asset-based perspective of students, our critical lens makes it necessary to assess the presence or lack of certain markers of capital recorded by the elementary schools housing DL programs as a way to make equity concerns visible. For our demographic analysis, we used quantitative research methods to analyze our dataset on the schools housing the Utah DL programs and the equity trends between 2005 and 2014. We specifically drew on the Utah 2014–2015 school year state aggregated student enrollment spreadsheet of all Utah public schools and charters (USOE 2015a) and the 2014–2015 school year list of Utah DL programs (USOE 2015b) to create one comprehensive database of the schools housing the 106 DL elementary programs using SPSS20 statistical software. Data coded and included in the database from these two publicly available sources were: school and district name, target language, total student enrollment, white-non-Latina/o enrollment, free/reduced lunch (poverty) enrollment, and EL enrollment. White, poverty, and EL enrollment data was then recalculated into their respective percentages of overall student enrollment at each school to create our three demographic variable categories. We then supplemented the database with data related to the first year the program was implemented as well as the type of DL program (foreign language immersion vs. TWI). This data was obtained by reviewing district and school websites, local newspaper reporting, and then making phone calls to districts and elementary schools to resolve any inconsistencies related to the status and type of DL programs. We then used the year of implementation data to code the DL programs into three time periods: 8 pre-state-model (before 2006), 11 transition (2006–2008) when the DL programs had a mix of influences from the pre-state and state-model periods, and 87 state-model (after 2008) programs based on the state DL model funded by the Utah state legislature. DL programs that began during the pre-state-model period that converted to the state-model after 2008 were only counted as pre-state-model programs as our focus was on new DL programs that emerged from the state policy. Our analysis focused only on the 95 pre-state and state-model DL programs, avoiding the use of the 11 DL programs of the transition period in order to yield a clear comparison of the pre-state and state-model periods. The data provided an important snapshot of the schools in which these DL programs were housed. It is important to keep in mind that the school demographic data used represented the only level of data USOE reported they had available at the time. Statewide DL program’s student and classroom-level data were unavailable. We present further details of our statistical analysis of the demographic data alongside our findings.

Findings

A Critical Discourse Analysis of the Utah Model of State-Level DL Policy

Our critical discourse analysis looked for patterns vis-à-vis the hegemonic discourses around race, class, and language status as well as the resulting targeting of beneficiaries. Our policy findings are organized by the three major patterns that emerged from the five policy documents studied. First, there was a counterhegemonic but problematic treatment of a general equity discourse. Second, there was a strong reflection of globalized human capital discourses and their classist implications. Third, there was significant countering of English hegemony by elevating the value of multilingualism and acknowledging local multilingual communities. However, this pattern came with problematic limitations that lent greater weight overall to a longer list of ways in which the policy texts reflected English hegemony. In particular, the policy data suggested the plausibility that the capital of white racial privilege was intersecting subtly with the uncontested aspects of English hegemony through the policy texts’ discursive privileging of elite (school) over folk (home) multilingualism. Ultimately, all three forms of privileged capital we studied appeared to intersect under the discursive banner of “mainstreaming” discussed in the introduction. The data suggest that this oft-repeated idea—as when the DL Overview text invoked “a continuing goal to mainstream Dual Language Immersion Programs throughout the Utah Public school system”—operated as a coded descriptor of Utah DL policy’s most heavily targeted beneficiaries: students and communities that were whiter, wealthier, and identified more with English than other languages even when multilingual.

The Neutralization of the Counterhegemonic Inclusion of a General Equity Discourse

We found that only one policy text, the Legislative Report, included clear instances of a diversity or equity discourse that seemed to counter the three hegemonic discourses. The two instances were (1) the description of the Utah model as one in which “children of all abilities and socio-economic conditions, from urban to rural areas of the country and from homes in which many languages are spoken” (USOE 2013, p. 7); and (2) the inclusion of a summary of academic results of the programs highlighting the fact that “students with different demographic characteristics are represented in the DLI programs” and the fact that “students from ‘traditionally marginalized groups’ achieve at the same level as mainstream students” (USOE 2013, p. 19). Despite these remarks in the Legislative Report, there was a conspicuous absence in the other four policy texts of explicit discursive statements directed to local heritage language communities particularly regarding research evidence of DL effects in boosting the school achievement of ELs, bilinguals, and heritage learners from historically underserved and racialized groups (Pease-Pretty On Top, n.d.). This absence amounts to a clear instance of discursive silencing—when language is used “to limit, remove or undermine the legitimacy of another use of language” often by displacing “the silenced material by means of another” more “acceptable” discourse (Thiesmeyer 2003, p. 2).

Explicit Globalized Human Capital Discourses as Implicit Classism

In an email interview (Speaking in Tongues 2010), the USOE official leading the DL initiative responded as follows to a question about the factors that prompted Utah’s DL policy decision: “Economic, Economic, Economic! Utah is a small state, so for our economic survival and the national security of our country we MUST educate students who are multilingual.” This official’s emphatic attesting to the globalized human capital discourse foregrounded in Utah’s DL policy planning is echoed strongly in the five policy texts themselves. All policy texts we analyzed showed some degree of a pattern of privileging economic themes and interests by consistently using phrases such as “compete in a global society” and “economic development benefits” (Utah Senate Bill 41 2008; Utah Senate Bill 80 2007). The documents also allowed such economic themed phrases to dominate descriptions of the program goals such as the excerpt below.

The USOE engaged business and education leaders regarding the economic and social benefits of a multi-lingual citizenry. The group followed through by creating the Utah World Language Roadmap, a comprehensive plan to prepare the populace—linguistically and culturally—for full participation in the global marketplace. [Legislative Report] (USOE, 2013, emphasis ours)

An interesting finding is that the Time Magazine article included in the Legislative Report claimed to eschew the competing nations aspect of globalized human capital discourse yet actually still participated in the discourse of human capital by making the claim, for example, that multilingual people “work faster and expend less energy in doing so, and as they age, retain their cognitive faculties longer” (p. 20). These are human capital concerns in that they betray a discourse of seeking a productive, efficient workforce and view the brain as part of the new “knowledge economy” (Peters 2001).

A last example of a more subtle reflection of globalized human capital in the policy texts was the silencing of an initial opportunity for the funding of DL programs teaching Navajo (or Diné as many in this Native American tribe prefer to be called). This early gesture at supporting a local language in addition to the languages of international trade simply fades from sight with no Navajo programs ever appearing. The Administrative Rule quietly replaced Navajo with Portuguese in 2012 on the list of languages from the original legislation. Moreover, the Legislative Report and the DL Overview’s narration of the history of Utah DL provided a list of the languages legislated for DL but discursively silenced Navajo. Thus a local language was erased from both the law and from the history told about the law. The effect of this silencing was to further narrow the focus of the policy on languages best suited for the hegemonic discourse of globalized human capital.

Globalized human capital discourse exerts power subtly by hiding the issue of class and the inequality of current wealth and opportunity conditions via neoliberal promises to “float all boats” by increasing the wealth of the wealthiest. These discourses assume that the opportunities to apply language skills in global markets are automatically and equally distributed across socioeconomic groups. Under these assumptions, what was discursively silenced was any consideration of whether using language learning for equity/heritage purposes might create more effective routes to economic empowerment for marginalized language groups.

The Countering of English Hegemony, or Is It?

There was clearly a pattern in the policy texts of a strong opposition to English monolingualism (a key indicator of English hegemony) through a strong elevation of the value of multilingualism and the recognition of local multilingual communities as potential program participants. For example, the Time Magazine article within the Legislative Report quotes the USOE official in charge of DL lambasting monolingualism as “the illiteracy of the 21st Century” (USOE 2013, p. 20). In a subtler way, the stated goals of the Administrative Rule—to “increase the number of students who reach proficiency in a critical language”, “build overall foreign language capacity in the state of Utah”, and “increase the number of biliterate and bilingual students”—also work to counter English hegemony. The last goal is an example of the implicit or understated ways in which the DL policy texts acknowledged the existence of local communities that were already multilingual and the value of serving them. Examples of more explicit ways these communities were acknowledged were when the Legislative Report mentioned “homes in which many languages are spoken” (USOE 2013, p. 20) and when the Administrative Rule defined “dual language immersion” as “a distinctive dual language education program in which native English speakers and active speakers of another language are integrated.” The policy’s definition of the Utah model promised the inclusion of students normally marginalized by English hegemony as a defining feature of the program.

However, though this group of discursive moves countered English hegemony by critiquing monolingualism and promising inclusion of students with diverse capitals within Utah’s array of languages and language abilities, its overall discursive impact was ultimately overwhelmed by four pieces of evidence that suggested a reassertion of English hegemony across the policy texts that was subtly intersecting with the normative whiteness implicit in elite multilingualism. First, there was discursive and legal silencing of 90–10 DL models shown through research to better combat monanglicization. Second, folk multilinguals were portrayed as second-class participants in comparison to English-privileged elite multilinguals. Third, two types of Utah DL were defined in the policy texts which served to exclude the third type of DL used in other states to serve folk multilinguals—developmental bilingual education. Finally, there was inequitable regulation of teacher credentialing that suggested those with the least English privilege were being marginalized instructionally within DL programs.

The rejection of 90–10 DL models Research has shown that well implemented 90–10 DL models are more likely than 50–50 programs to solidify long-term, balanced bilingual and biliterate abilities, especially for ELs (Thomas and Collier 2002). Our content analysis revealed frequent reiterations across the five policy texts of the prohibition on funding 90–10 DL models, such as, “All state-sponsored schools with Dual Language Immersion programs are required to implement the fifty–fifty model and use two teachers.” In a Speaking in Tongues (2010) email interview, the USOE official in charge of DL wrote, “I personally abhor anything but a 50–50 model for instructional and political reasons!” We were thus persuaded that one of the reasons the policy mandated its strict 50–50 model was to not alienate or make wary those parents and powerbrokers who were seeking elite multilingualism for Utah children. Thus, ultimately the Utah model reflected English hegemony, asserting that English would not be marginalized in the division of DL instructional time.

The second-classing of folk multilingualism Though acknowledged, the goal of serving Utah’s already multilingual local communities did not receive as much attention as the goal of increasing elite multilingualism for the monolingual majority of English speakers. For example, there was no implicit acknowledgment of local folk multilingualism in the stated goals of the original DL legislation excerpted below,

(1)(b)(i) the importance of students acquiring skills in foreign languages in order for them to successfully compete in a global society; and (ii) the academic, social, and economic development benefits of the acquisition of critical languages. (Utah Senate Bill 80; Utah Senate Bill 41, emphasis ours).

What was re-emphasized here were globalized human capital discourses and a pattern of “foreign language” being used (instead of a more neutral term) to mean “non-English language,” pointing to a reliance on the foreign language lineage by the crafters of the DL policy. This discourse of foreign worked to undermine any countering of English hegemony in the policy texts because it clearly positioned English as (a) the point of reference for all other languages mentioned, and (b) the proper, normal, common-sense, or native language of Utah and the United States, thus marginalizing folk multilinguals as supposed foreigners in their own land.

DL definitions that silence folk bilinguals and exclude developmental bilingual education DL options The Utah DL policy describes two program models:

1

(1) “One-way” immersion is a program in which a student population consists of

2

English language speakers with limited to no proficiency in the foreign immersion

3

language. In such a model, less than 30 percent of the students have a native

4

language other than English.

5

(2) “Two-way” immersion is a program in which a student population consists of

6

a majority of English language speakers and a minority of language speakers other

7

than English with dominance in their first language and home language support for

8

this language. A 1:1 ratio is ideally maintained for these two language groups, but a

9

minimum of one-third of each language group (such as 2:1 ratio) is required.

10

(Critical Languages Program 2012) (emphasis ours)

The DL Overview document adds headings to each of these program models with one labeled “One-way immersion programs serve one group” and the other labeled “Two-way immersion programs serve two groups.” What emerged from our analysis of the two program model definitions were contradictory statements concerning the composition of their participants that result in the silencing and de-prioritization of EL and bilingual participants. In the one-way model, ELs and bilingual students are omitted from the heading “serves one group” and from lines 1–2 of the policy text excerpt. Yet lines 3–4 contradictorily assert that up to “30 percent” of participants can be ELs and bilinguals. The effect of the contradiction is to position folk multilinguals as an afterthought to the more important group—English privileged elite multilinguals. In the two-way or TWI model, the heading’s phrase “two groups” and excerpt line 8’s reference to a “1:1 ratio” suggest a parity among participant groups that is then contradicted by line 6’s relegating of folk multilinguals to a “minority” status. The effect of this analogous contradiction is to position elite multilinguals as the primary beneficiaries of the TWI model as well.

Our content analysis revealed the absence of a third type of DL referenced in the research—“one-way” developmental bilingual education—that mainly benefits folk multilinguals. Thus, the inclusion of folk multilinguals in the definitions within the Administrative Rule ironically functioned primarily to ensure the exclusion of any programs not designed to include English-privileged participants. This along with the discursive contradictions and silencings in the definitions suggested that the main purpose of listing the two DL models was not to make distinctions between the types of credentialing required of teachers, as the Administrative Rule suggested, but rather to ensure that the DL model often seen as threatening to English hegemony and its beneficiaries was silenced and did not exist alongside them. The inclusiveness of diversity in the definitions—even the lofty and highly disputable claim that this inclusiveness was what was most distinctive about Utah DL—thus functioned to ensure that diversity remained marginalized while making English-privileged participants the most discursively important in the policy texts.

Teacher qualification requirements prioritize the needs of English-privileged students over ELs Our content analysis revealed that the endorsement requirements for target language DL teachers were significantly higher than for the English DL teachers. Our discourse analysis showed that higher value was placed on the importance of the training of the target language teachers. Both of these patterns reflected English hegemony by deprioritizing the educational needs of the ELs and bilinguals when they were the recipients of the promised DL “integration” in the definitions of Utah DL. Only the Administrative Rule and the DL Overview text spoke directly to teacher credentialing. We point to two key pieces of evidence from our content analysis to underscore the absence of planning and consideration of the instructional needs of folk bilinguals present within Utah DL programs while privileging the instructional needs of English beneficiaries. First, despite the possibility that up to 30 % of folk bilinguals could be enrolled in one-way foreign language immersion DL programs, the English DL teacher was not expected to have any training in ESL or dual language methods. Second, in TWI programs, all required to enroll folk bilinguals, English DL teachers were only “strongly recommended” to hold an ESL endorsement. Although this was an equity improvement over a complete silencing of the needs of folk bilinguals, a “recommendation” still constituted a discursive second-classing in comparison to a “requirement.” The English DL teacher credentialing requirements were in contrast to the more rigorous requirements for the target language DL teachers who also had to demonstrate “advanced mid-level or higher” proficiency in the DL target language in addition to supplemental endorsements. Two endorsements were required for those working in one-way foreign language immersion programs—“a world language endorsement” and a state “dual language immersion endorsement in the language of instruction”; while those working in TWI programs were only required to obtain one state “Dual Language Immersion endorsement.” This disparity signaled that having highly qualified teachers that could meet the target language needs of English-privileged students in DL was a priority—especially in models that were intended to serve them exclusively. Our discourse analysis further found that the Administrative Rule’s language—“schools shall hire qualified language teachers” who have the appropriate “endorsement in the language of instruction”—discursively silenced the fact that English was also a language of instruction and English teachers were also “language teachers,” particularly and crucially for ELs. Thus the perspective and educational needs of folk bilinguals were silenced and ignored while those of English privileged beneficiaries were centered, showcasing another instance of English hegemony and the privilege it bestows on its speakers.

The discursive moves we have described that counter English hegemony in the Utah DL policy texts are thus overwhelmed or marginalized by the hegemonic currents of discourse that suggest that it is Utah’s English privileged or, better stated, the potential multilinguals who are most personally and culturally identified with English—that is, those acquiring elite (school) rather than potentially racializing folk (home) multilingualism—who are most targeted as beneficiaries. When added to the evidence in the previous section of the ignoring of folk bilinguals’ educational needs, this begins to point to the role of normative whiteness in the privileging of elite multilingualism in Utah’s DL policy—Utah DL’s multilingual version of English hegemony. It also reinforces the implicit classism coded within the globalized human capital discourses embedded in the elite multilingualism privileged in Utah’s DL policy. Since the racism and classism at work in U.S. policymaking operates in coded forms and is not overtly championed (Coates 2011) as our findings confirm, it is important to pay attention to what is discursively being silenced in any educational policy. In the case of Utah DL policy, the effect of these subtle or coded forms of racism and classism suggests that the discursively targeted beneficiaries of the policy are not just those primarily identified with English but also the whiter and richer of those primarily identified with English. We thus argue that one can see these five policy texts—one of which refers explicitly to “mainstream” as an antonym for “historically marginalized groups” (Legislative Report, p. 19)—furthering a perhaps unconscious, broader cultural agenda where to “mainstream” DL is code for gentrification. This gentrification process thus drastically reduces the prevalence of DL’s equity/heritage discourses and material effects by prioritizing access for more privileged students, as the following quantitative findings confirmed.

Shedding Light on Utah’s State-Model Demographic Inequities

To shed light on the state-model demographic inequities, it is important to understand the historical growth trends of elementary DL programs in Utah. These growth trends were noted to occur across three key phases in Utah’s DL history—the pre-state-model period (1978–2005), the transition period (2006–2008), and the state-model period (2009–2014). Figure 1 illustrates these growth trends by the type of DL model implemented during each of these time periods. Overall, we identified marked increases in the growth of DL programs after the implementation of Utah Senate Bills 41 and 80’s dual language education initiative. Growth in new DL programs during the pre-state-model period remained relatively stable with an average of one program added every two years, seven out of the eight DL programs being TWI and one being a one-way foreign language immersion DL program. Pre-state-model DL programs as a whole suggested an alignment with an equity/heritage framework because of (a) their nearly unanimous choice of a locally prominent non-English language and (b) the fact that six out of the eight pre-state-model DL programs were housed within high poverty schools with higher enrollments of non-whites and ELs. During the transition period, the ratio of one-way foreign language immersion programs to TWI flipped from pre-state model levels with only two out of the 11 DL programs started during the transition period being TWI programs and the remaining nine DL programs being one-way foreign language immersion models. Thus, the transition period marked a time of emerging growth in one-way foreign language immersion DL programs, including the start of the first non-Spanish DL program in a school with privileged demographics, while TWI programs grew at a slower rate than in the pre-state-model period. It was not until after the state DL funding mechanism was legislated and put firmly in place that Utah saw dramatic increases in one-way foreign language immersion and TWI models of DL. Eight Chinese and four French programs burst onto the scene in 2009 within schools that were heavily White, affluent, and English privileged. There were also two Spanish one-way foreign language immersion DL programs started that year in high poverty schools with over 20 % Latina/o enrollment and 11–23 % EL enrollment, enrollments that would suggest the viability of a TWI model. As of 2014, 87 new DL programs had been implemented in the state-model period with only 16 of them being TWI programs. One-way foreign language immersion DL programs represented 82 % of the new DL programs started during the state-model period representing a significant and sustained shift toward “mainstreaming” of one-way foreign language immersion models of DL.

Fig. 1
figure 1

Number of DL programs started by type (one-way foreign language immersion and two-way immersion DL types), year started (1978–2014), and DL implementation period (pre-state-model, transition, and state-model periods)

Full size image

The specific quantitative analysis methods used to compare the white racial privilege, wealth, and English privilege demographics of the schools that housed these DL programs in the pre-state-model period versus the state-model period were as follows: We first performed a two-sample independent t test assuming equal variances using a pooled estimate of the variance to test the hypothesis that the resulting mean percentages of students with white racial privilege in schools housing DL programs pre-state-model and state-model were equal. The same analysis was conducted to test the hypothesis that the resulting mean percentages of the variables of ELs (students lacking English privilege) and students in poverty (lacking wealth) enrolled in schools housing DL programs pre-state-model and state-model periods respectively were equal. Table 1 summarizes the pre-state-model and state-model DL programs’ descriptive and statistical information related to the mean percentages of indicators of white racial privilege, wealth, and English privilege. Our analysis of the mean percentages of students with white racial identification, poverty, and ELs enrolled in Utah elementary schools housing pre-state-model DL programs and state-model DL programs revealed that state-model DL programs were more often located in schools with a higher mean percentage of students identified with white racial privilege (M = .7796, SD = .1576, N = 87) than pre-state-model DL programs (M = .4498, SD = .2479, N = 8), a statistically significant difference, MD = − .3298, 95 % CI [− .5378, −.1217], t (7.53) = − 3.70, p < .01 (two-tailed based on unequal variances). This statistically significant result on white racial privilege lets us know that the result was not likely due to sampling error, the effect size d of −1.99 indicates that the strength or practical significance of the difference in white racial privilege between schools housing pre-state-model and state-model DL programs was significantly larger than typical.

Table 1 Comparison of schools housing pre-state (n = 8) and state-model (n = 87) dual language education programs on the student enrollment variables of white racial privilege, poverty, and English learner status
Full size table

We also found that state-model DL programs were more often located in schools with a lower mean percentage of students lacking the privilege of wealth (poverty) (M = .3479, SD = .2134, N = 87) than pre-state-model DL programs (M = .6409, SD = .2631, N = 8), a statistically significant difference, MD = .2930, 95 % CI [.1334, .4526], t (93) = 3.65, p < .01 (two-tailed). The effect size d of 1.35 indicates that the strength or practical significance of the difference in enrollment of students lacking wealth between pre-state-model and state-model DL programs was also larger than typical. Finally, we found that state-model DL programs were more often located in schools with a lower mean percentage of students lacking English privilege (ELs) (M = .0724, SD = .0897, N = 87) than pre-state-model DL programs (M = .2513, SD = .1599, N = 8), a statistically significant difference, MD = .1789, 95 % CI [.0448, .3130], t (7.41) = 3.12, p < .05 (two-tailed based on unequal variances). The effect size d of 1.83 indicates a larger than typical strength or practical significance in this difference in means. Overall, state-model programs were found to be often located in schools with students that were more often whiter, wealthier, and more English privileged compared to pre-state-model schools.

Discussion: DL as the Next Chapter in a Larger History of Inequitable Access to Enrichment Education?

Our discursive findings suggest the Utah DL policy is shaped by hegemonic discourses that center on and target mainly white, wealthy, English-privileged students as the beneficiaries of these programs while silencing other beneficiaries. In this way, the targeted beneficiaries are given an opportunity to acquire even more privileged capital via the elite multilingualism they will learn at school, while the educational value of using DL to further develop folk multilingualism is largely silenced. Our demographic findings show the material effects of Utah’s state model of DL education largely being used to privilege the already privileged by often housing these programs in schools with high-value-capital demographics. What is silenced in a state DL educational policy with this sort of discursive targeting and with these sorts of inequitable material effects are the possibilities for maintaining the equity effects of the pre-state-model period when home language support in educational settings was employed as a means of closing the achievement and attainment gap for marginalized students.

Utah’s “mainstreaming” of DL provides a fast-track to advanced placement and concurrent college enrollment courses in high school (USOE 2013, 2015c) resulting in a localized instance of the larger nationwide pattern of DL gentrification (Morales and Rao 2015; Flores 2015). As English hegemony and normative whiteness conspire on the one hand to diminish multilingual schooling options designed specifically for those with less English privilege (bilingual education lineage), on the other hand a new globalized human capital discursive climate is leading policymakers and more privileged parents to envision DL as a necessary educational enrichment (foreign language education lineage). Thus, the gentrification of DL and other similar forms of U.S. language education reflects the receding influence of the bilingual education field as a megaphone of counterhegemonic equity/heritage discourses against English hegemony and monanglicization. It is because of this national shift toward multilingualism but away from equity/heritage discourses and policy that we argue that a larger educational struggle over access to enrichment education—and the capital it provides—appears to be incorporating language education as its next battlefront. If we define enrichment tracking as any form of educational opportunity—whether documented to be more engaging and effective or merely perceived to bestow exclusive prestige—that is offered as a scarce benefit via the separation of students into curricular tracks within schools or across schools (Montgomery County Education Forum 2002), then DL definitely stands poised to fit this mold.

Enrichment tracking constitutes a mechanism for the unequal distribution of capital and exists in opposition to compensatory tracking, that is, remedial tracks, such as English as a second language courses. Examples of enrichment tracking are honors classes, advanced placement courses, gifted and talented programs, STEM programs, and enrichment based charter schools. Past attempts to include more students of color, bilingual and EL students, low-income and other underserved students in these enrichment programs have been minimally successful leaving these students largely underrepresented in enrichment programs (Esquierdo and Arreguín-Anderson 2012; Oakland and Rossen 2005) and overrepresented in compensatory programs (Losen and Orfield 2002; Brisk 2005). Pérez and Flores (2002) argue that TWI “programs have been called enrichment education” in order to be “‘marketed’ to the majority population and policymakers” (p. 358). In his blog, Nelson Flores (2015) warns that the discovery of dual language programs by privileged parents is the columbising of bilingual education, leading to what we have been calling the gentrification of DL (Valdez et al. 2013). This gentrification process reflects our concern that DL may soon constitute the next chapter in a larger history of enrichment tracking whose benefits have gone largely to the already privileged.

Concluding Remarks: Advocating for an Equity Effect

We want to leave readers with the understanding that we do not oppose allocating public resources for DL programs that enrich privileged students with multilingualism. It is commendable that foreign language education has moved from the margins of U.S. education to take a more prominent position. However, in our critique of policy decisions that drove the Utah DL policy, we show readers how these educational outcomes can come about at the expense of the equity effects of the bilingual education lineage in language education policy, evidenced in a shift from an equity/heritage to a globalized human capital framework in DL programs and surrounding public discourses (Valdez et al. 2014) that draws primarily from a foreign language education perspective. What is required is equitable access to DL education for its traditional residents (marginalized language students) so that their educational needs are not pushed out of sight as they have been thus far in the Utah model. A more inclusive DL version could forge a collaborative dual framework where equity/heritage and globalized human capital concerns get equal treatment—backed institutionally with truly equivalent financial resources and representation for their respective student beneficiaries. Such a compromise would require institutional bridges between the foreign/world language and the equity/alternative language services/bilingual education departments. To guide more states and districts in such a collaborative vision for inclusive enrichment education, we offer five recommendations.

In response to concerns that DL programs cannot escape asymmetrical power dimensions (Palmer 2010; Scanlan and Palmer 2009; Valdés 1997), the first recommendation calls for all DL programs and related policies to explicitly plan for how they will address power differentials within their recruitment and admission process as well as in the day-to-day aspects of DL program implementation. Second, programs should be geographically located to restore equity effects by establishing a process of outreach and targeted recruitment with equitable demographic goals rather than just accepting applications without selection criteria. Third, local marginalized language groups should be given a special opportunity to use DL programs for the preservation or recovery of their languages. Dorner (2015) describes parents seeing multilingualism through DL as “a means to integrate within their local communities and to identify with extended family members” (p. 126). In Utah, later-arriving language groups such as Pacific Islanders (i.e., native Hawaiians, Samoans, Tongans) whose numbers are over 29,000 in the state (U.S. Census Bureau 2015) are an example of local language groups amenable to DL education. In addition, future research should investigate the possibilities of coordinating DL policy with policies guiding indigenous language education. This becomes particularly important in a state such as Utah that has five historically indigenous groups—Goshute, Navajo, Paiute, Shoshone, and Ute—whose languages deserve some of the state resources and support allocated to DL education. This localization would bring Utah in line with other states that are integrating equity/heritage concerns. For example, Portland Public Schools (2015) has one Vietnamese DL program and is considering adding Somali and Cantonese programs; and Minnesota has six DL programs in Dakota/Ojibwe and three programs in Hmong (Marty 2014). Fourth, one-way DL programs should not be the sole prerogative of the English-privileged through foreign language immersion programs, a situation that is not merely inequitable but unequal. Local marginalized language groups are entitled to the option of choosing to design one-way developmental bilingual education programs for speakers dominant in a language other than English. The policy should not continue to specify this option as unfundable or offer discouragement or disincentives for it as is happening today in Utah. Again, these programs are not rare in the U.S. as a whole and are especially common in the state of Texas.

The final and perhaps most important recommendation would be to re-conceptualize the way the programs are talked about both in policy and in its promotional marketing. It should be possible to give equity and heritage concerns as much importance as globalized human capital concerns when we talk about DL rather than continuing to see one framework as an impediment to the other, or the concerns of marginalized groups as a threat to the appeal of programs to the privileged. We offer the Portland Public Schools DL department as an example of how some of our recommendations can be implemented. It is ironic that Utah state DL officials credit visits to a Portland Chinese program that began in 1998 for inspiration in how to run their programs (Conley 2009). Yet, had they followed the evolution of this program, they would have learned that though Portland began its DL history with a focus on privileged students, it has since transformed itself to include equity/heritage goals. Portland Public Schools (2015) initiated a DL department in 2012 with a mandate that combines the traditionally separate purviews of an equity department and a foreign/world language department.

Closing the opportunity gap for historically underserved students is a priority. To this end, the department provides multiple pathways and entry points for students to become bilingual and biliterate. A variety of programs are offered to support the diversity of levels of proficiency in the partner languages. Programs for native speakers include native language literacy and dual language immersion. Programs for non-native speakers include world language and dual language immersion.

More states and districts should emulate such discourses of a commitment to multilingualism for students at all levels of privilege that works against normative whiteness, wealth inequality, English hegemony and monanglicization rather than working subtly in cahoots with them. Those in favor of social justice in education should continue the fight to avoid DL becoming yet another form of enrichment education accessible only to those who need the enrichment least.