An Analysis of How Restrictive Language Policies Are Interpreted by Arizona ’ s Department of Education and Three Individual School Districts ’ Websites

Restrictive language policies for education have been passed in several states in the United States. In 1998, 2000, and 2002, California, Arizona, and Massachusetts passed the most restrictive of these policies, impacting 4.4 million students classified as English language learners (ELLs). This study examines how these policies are currently interpreted and presented to the public on Arizona’s Department of Education website, as well as how they are interpreted and presented on the websites of three of the state’s largest school How Restrictive Language Policies Are Interpreted 2 districts. We seek to understand how three key elements of the laws—one-year programmatic time limits, Structured English Immersion (SEI) programs, and waiver processes—are conveyed by each text. Using tools from critical discourse analysis (Fairclough, 2003, 2013, 2015), we trace the endurance or disappearance of these elements between texts and across time. Textual differences are discussed as reflecting and perpetuating important contextual differences among the districts.


Introduction
Restrictive language policies, which limit the ways that English is taught in K-12 classrooms to children who speak other languages, have been passed in several states in the United States.California, Arizona, and Massachusetts have the most restrictive of these policies, impacting an estimated 4.4 million students or 9.2% of the public school population (U.S.Department of Education, 2015) who are classified as English language learners (ELLs).California voters passed the first restrictive language policy, Proposition 227 (now California Education Code § § 300-340), on June 2, 1998, with a margin of 61% to 39%.Arizona followed, with voters passing Proposition 203 (now Arizona Revised Statute (A.R.S.) § 15-751-755) on November 7, 2000 by a margin of 63% to 37%.Massachusetts Ballot Question 2 (now Massachusetts General Laws chapter 71a (G.L. c.71a)) was passed by voters on November 5, 2002, 61% to 29%.While the passage of the policies themselves is significant, perhaps even more important is how these policies are interpreted at various levels-from state departments of education to teachers directly working with ELLs-as these interpretations drive decision-making and implementation on the ground.
In this article, we draw from Weatherly and Lipskey's (1977) idea of the role that "streetlevel bureaucrats" play in shaping the implementation of policy.We focus in this study on how the restrictive language policy in Arizona-the authors' state of residence-is interpreted and represented on its Department of Education website, and then on three districts' websites within the state.Specifically, we seek to understand how three key elements of the laws-one-year programmatic time limits, Structured English Immersion (SEI) programs, and waiver processes-are interpreted and presented to the public.We begin by providing a brief background on Arizona's English-only policies and a review of relevant work that has been done about their passage, their implementation, and their effects.This is followed by a brief overview of discourse analysis studies on ELLs related to these propositions.We then introduce Arizona's state and district level contexts and share our theoretical framework and methodology, before delving into our findings about how the three key elements of the policies are represented on the Arizona Department of Education and three school district websites in the state.In our discussion, we explore how these textual differences can be explained in relation to what we know about each context.We conclude with recommendations for major stakeholders in the education of ELLs.

Review of the Literature
California's Proposition 227 was the start of a successful voter initiative against bilingual education.On June 2, 1998, California voters passed Proposition 227 (now codified as California Education Code § § 300-340).Arizona citizens followed California's lead and passed Proposition 203 in November, 2000.On November 4, 2002, over 60% of Massachusetts residents voted in favor of Question 2, giving the restrictive language proposition the narrowest win of the three states.In Arizona, after Tom Horne became Superintendent of Public Instruction in 2003, he followed through on his campaign promises to make enforcing Proposition 203 a priority.Arizona passed legislation in 2007 that specified with greater precision how Proposition 2003 would be implemented.This policy mandated that all ELLs receive four hours daily of English Language Development in a Structured English Immersion (SEI) class.According to the Arizona Department of Education (ADE) (2008), the SEI model provides a clear direction for teachers, achievable targets for students, and student progression to proficiency in English.Four key principles of the model are that (a) English is fundamental to content area mastery, (b) language ability-based grouping facilitates rapid language learning, (c) time on task increases academic learning, and, (d) a discrete language skills approach facilitates English language learning, ideas that have been refuted by second language acquisition researchers for their lack of scientific evidence (Krashen, Rolstad, & McSwan, 2007).From the beginning, researchers in the field of second language acquisition such as Crawford (2003) and Krashen (1997) questioned the mis-presentation of information regarding English language learners and the attacks on bilingual education that fueled the debate just prior to each state's passage of the propositions.
Since the passage of the first Proposition in 1998, numerous studies have examined the implementation and impact of language policies emerging from the propositions.Gandara et al. (2000) reported on the initial impact of Proposition 227 in California and found great variation in how language policies, including resulting SEI programs, were being interpreted and implemented.Uriate, Tung, Lavan, and Dias (2010) focused on the widening academic achievement gap for Boston's ELLs as a result of the restrictive language policies.Krashen, Rolstad, and MacSwan (2007) were some of the earliest critics and published their concerns that Arizona's SEI model has shown little alignment with scientific evidence.More recently, several reports published by U.C.L.A's Civil Rights Project (www.civilrightsproject.ucla.edu)have documented concerns about and the effects of the implementation of the SEI model, such as the underidentification and underachievement of ELLs.In 2006, Wright and Choi (2006) surveyed Arizona teachers and reported that they believed that the state's language policies were failing to meet the academic needs of ELLs.The segregation of ELLs and their lack of access to rigorous core curriculum as a result of the restrictive language policies have been reported by numerous authors (e.g.Gándara & Orfield, 2012;Grijalva, 2009;Lillie, Markos, Arias, & Wiley, 2011;Moore, 2008).Furthermore, as Combs & Nicholas (2012) have pointed out, the policies have had a negative impact on Indigenous students, undermining indigenous native language revitalization efforts.Garcia (2016) argues that Arizona, as a result of these policies, is the worst state in the US for students learning English, and Wiley (2012) has referred to Arizona as "ground zero" for the most restrictive language policies in the country.
Longitudinal studies on the impact of these policies years after their initial implementation report that these policies have not been effective in closing the achievement gap nor in helping students learn English within the expected one-year time limit.Arias and Faltis (2012), in their edited volume, include a number of studies that analyze the long-term negative impact of the policies in Arizona, including the impact on teacher preparation.Jimenez-Silva, Gomez, and Cisneros (2014) examined the intersection of several policies affecting ELLs and argued that the convergence of these policies have significantly impacted how ELLs' academic needs have not been addressed in Arizona.
Despite the challenges presented by Arizona's restrictive language policies, there are examples of educators and school communities that have been able to negotiate these policies.For example, Newcomer & Puzio (2014) shared how a school community of practice mobilized for the facilitation of parent waivers to opt out of the state mandated SEI program for ELLs.As a consequence, the community created a space for a thriving bilingual community.In another example, the Teachers of Language Learners Learning Community (TL3C) at Mesa Community College created a thriving consortium that provides support for school communities seeking to include ELLs in their dual language programs (Jimenez-Silva, Garvey, & Gomez, 2015).
In recent years, scholars have explored language policy and planning (LLP) both through sociolinguistic and educational questions related to language rights, language endangerment, language evolution, language environments, and linguistic imperialism (Hornberger, 2002;Pennycook, 2004;Ricento, 2000;Tollefson, 2013;Wiley, 2013).This body of LLP work has often debated the role language policy, planning, and practices as they relate to one another, and to broader societal trends in language use.McCarty (2011) anchored her ethnography of language policy within this history, but emphasized that language policy, planning, and practice are conceptualized in her work as interrelated processes.Ruiz (1984) theoretically explored the ways that polices, programs, and discourse often explicitly and implicitly position minority language use as a problem, a resource, or a right.Whereas, Hornberger and Johnson (2007) described LLP as a metaphorical onion, with diverse contexts and interpretations of policy occurring and interacting at each layer of the onion.Specifically, they discussed the range of state level policies, and institutional and individual level practices as representing the diverse layers that interact in often unpredictable ways.The most recent developments within LLP research have therefore called for increasing our collective knowledge of the ways that language polices influence language behaviors, local practice, agency, identity, and social justice (McCarty, 2011;Ricento, 2000).At the same, there is also a continued focus on the ways that official polices can mediate local contexts through historically situated power structures and ideological spaces (Ricento, 2000).Therefore, we as researchers have made an attempt to address this important call by using critical discourse analysis as a tool that will help us understand online spaces which give voice to the Arizona Department of Education and local Arizona districts as language policy implementers in a restrictive language state.
From a discourse analytic perspective, researchers have looked closely at the language used by the campaigns and by the media, in order to understand how the campaigns succeeded in getting the propositions passed.Johnson (2005), in an analysis of newspaper coverage and voter pamphlets in Arizona, found that the debate over bilingual education was portrayed as a "war" between two sides, with ELLs positioned as "victims" of the "failures" of bilingual education and SEI as the pathway to the "American Dream" for ELLs.Gonzalez-Carriedo (2012) took a more general look at the ways linguistically diverse students were portrayed in Arizona print news media from 2006 to 2011 and found that specific ideologies emerged from different news media sources, with the newspaper that represents the greater East Valley of the Phoenix area emerging as the most supportive of SEI policies.

Contexts
Below, we outline the state, city, and district contexts in greater detail for readers who may not be familiar with them.
Arizona.In Arizona, approximately 28% of students have a primary home language other than English (ACF, 2008).Meanwhile, an estimated 7% (85,000) of the student population is classified as ELLs (AZDOE, 2014), a number that has decreased by nearly 40% since 2008 due to changes in how ELLs were being identified in Arizona as well as an exodus of students whose families have moved because of the political climate in the state (Kossan, 2010).The Arizona ELL student population is made up of elementary (79%), middle (13%), and high school (8%) students, with the primary language of most ELLs being Spanish (67%).The remaining home languages are reported to be English (17%), Other (6%), Arabic (2%), Vietnamese (1%), Navajo (1%), and unidentified (6%) (ADE, 2014).Once classified as an ELL, a student is assessed using solely the Arizona English Language Learner Assessment (AZELLA) annually to determine whether s/he has achieved proficiency in English.
Below we provide the basic demographics of the Arizona school districts documented in this study: Mesa Public Schools (MPS), Tucson Unified School District, (TUSD), and Flagstaff Unified School District (CHSD).
Mesa.MPS is the largest school-district in the state of Arizona, covering just over 200 square miles and serving approximately 63,213 students in 55 elementary schools, 9 junior high schools, 6 high schools, and a variety of choice and focus schools.Approximately 66.5% of MPS students receive free and reduced lunch (FRD) benefits, however the within-district variation in FRD ranges from schools with 97% of the student population on FRD to 10%.The ethnic breakdown of the district is 46% Anglo-White, 41.5% Hispanic/Latino, 4.4% American Indian, 4.3% African American, 1.5% Asian, 1.6% multi-ethnicity, and 0.6% Pacific Islander.Additionally, 7% of MPS students are ELLs and MPS is home to two dual language immersion (DLI) elementary programs.
Tucson.Tucson Unified School District (TUSD) is the second largest school district in the state of Arizona with 85 schools over 228 square miles serving over 48,000 students at 63 elementary schools, 26 middle/junior high schools, and 14 high schools.The student population is 64% Latino/a, 21% Anglo-White, 5.6% African American, 3.7% Native American, 4.1% mixedrace, and 2% Asian American.Approximately 69% percent of the district's student population receives FRL benefits, with schools that vary from 34% to 90% within the district.While ELL students represent 8% of the student population in TUSD, 29% of students have a primary home language other than English.TUSD has six DLI elementary programs, three middle school programs, and one high school program, making it the district with the most DLI programs in Arizona and the first district to offer DLI from kindergarten to twelfth grade.
Flagstaff.Flagstaff Unified School District (FUSD) is the largest school district in northern Arizona, serving approximately 10,000 students across 15 schools-three high schools, three middle schools, and 10 elementary schools-covering approximately 4,400 square miles of rural, small city urban, and suburban neighborhoods.44% percent of the district's current student population receive FRD benefits, however this percent varies from 12% to 95% at the school level.The ethnic breakdown in FUSD is 43% Anglo-White, 27% Hispanic/Latino students, 25% American Indian or Alaska Native, 1% African American/Black, 1% Asian American, and 3% multi-ethnicity, and 3% unidentified.Furthermore, 3.7% of the student population in FUSD is classified as ELL, which varies across school sites from 0.30% to 16.4% (FUSD, 2016).The district is home to a dual language school, Puente de Hozho Bilingual Magnet elementary school, which offers 50-50 DLI in Spanish-English or Navajo-English.The district also runs a school on the Navajo reservation with a well-established Navajo language and culture program.

Research Questions
Broadly we are curious about how language policies are interpreted from the time of their codification to their implementation in classrooms, by intermediaries such as a state department of education and school districts.In this paper, we ask specifically: 1. How are the key elements-Structured English Immersion (SEI) programs, one-year programmatic time limits, and waiver processes-of Arizona's restrictive language policies interpreted and represented on its Department of Education website? 2. Within Arizona, how are these elements interpreted by the largest school districts in each region of the state (north, south, central) and represented in their official websites?3. How do the differences across districts reflect and construct the local contexts in which they are written and read?
Theoretical Framework To answer these research questions, we draw on a variety of tools within critical discourse analysis (CDA) developed by Norman Fairclough (2003, 2013, 2015) using the principles of systemic functional linguistics (SFL) (Eggins, 2004;Halliday, 1994;Halliday & Matthiessen, 2013) to analyze language and text (e.g.Martin & Rose, 2003).A key idea in CDA is that while social conditions are realized in texts, texts also serve to (re)produce social conditions.For this reason, unlike other approaches to text analysis in which context only matters if explicitly referenced in the text (e.g.conversation analysis), researchers using CDA begin their work from both directions, seeking to understand the two-way relationship between text (concrete instances of written or spoken language) and context (broader social and cultural structures).
A second key idea in CDA is that of intertextuality-that texts relate not only to context, but to other texts.This notion comes from Bakhtin (1981), who wrote about the inherently dialogic nature of speech.The very nature of language means that each word that one uses has been used before and, thus contains a kind of internal history, bearing traces of its many contexts of use.Additionally, every utterance is both a response to an utterance that came before and an anticipation of what might come after.Through both words and utterances, then, speech is always linked to other speech.Julia Kristeva coined the term "intertextuality," connecting Bakhtin's ideas explicitly to writing.For Kristeva (1980), texts were both in dialogue with the reader (horizontal intertextuality) and with other texts (vertical intertextuality).Drawing on these concepts, CDA seeks not only to understand the relationship of text to context, but of texts to other texts.
A third key idea of CDA is that within a given cultural, historical, and social context, there are particular ways of understanding the world that make possible particular ways of speaking and writing.After Foucault (1971) and Bakhtin (1981), CDA refers to these understandings as discourses.In CDA, discourses are the links between text and context, in a relationship that Fairclough (2015) illustrated as nested rectangles (see Figure 1, next page), and for which he proposed three related phases of analysis: description, interpretation, and explanation.At the first level, the researcher uses systematic ways of looking closely at text; in the second level, the researcher distances herself somewhat from the text, looking for ways that the larger discourses or ideologies are realized in and produced by the text; in the third phase, the researcher moves further from the text, to explain how the text and discourses make sense within the larger social context of a certain time and place.
Many of the critiques of CDA rest in the third phase, as it depends on the researcher's own perspective of the relevant context.Yet, CDA researchers are upfront about having a perspective.This is where the "critical" part of CDA, as well as the fourth key idea, come in: CDA has both political and emancipatory aims (Titscher, Meyer, Wodak, & Vetter, 2000).CDA researchers are not simply concerned with texts, but in using texts as a starting place to question social and political life, ideologies, power, and justice.In Fairclough's words, "CDA is not, as one might assume, just a critique of discourse, it is a critique of the existing social reality (including its discourse) which begins with a critique of discourse" (2015, p. 7).
In our project here, we too come with a perspective.We, as researchers and educators, believe that languages are human rights and that families should have education options that allow for bilingualism and multilingualism to flourish.This belief is one reason for our interest in how restrictive language policies continue (or not) to be conveyed years after they were passed.

Focusing on Text: Systemic Functional Linguistics
At the center of Fairclough's nested analysis is the text.To understand texts, Fairclough draws on an understanding of language proposed by Michael Halliday (1994;Halliday & Matthiessen, 2013) in his systemic functional linguistics (SFL).A functional view of language proposes that pieces of language should be understood, not in terms of their grammatical roles, but their social functions.For Halliday, language simultaneously carries out three functions: interpersonal (representing social relationships between speakers), ideational (representing experience or states of affairs), and textual (organizing those relationships and experiences into text).Different kinds of interpersonal, ideational, and textual functions are carried out through a writer's or speaker's choice of words as well as her way of combining words into larger units of text.This is where the systemic part comes in.In SFL, the idea of choice is central: SFL views language as a system of options for meaning-making, so that any time we speak or write, we make a choice of one pronunciation, or word, or grammatical construction over others.In a discourse analysis based in SFL, the general research question might be phrased as, "What can I understand from the choices that this writer/speaker has made, in opposition to all of the other choices she could have made?" In this paper, we examine choices made in texts that were created in different times and places and for different audiences (voters, parents, districts, schools), but around the same topic (the education of English language learners) and for broadly the same purposes (to convey information regarding ELLs' education).We also analyze the way these choices reflect and construct discourses around language education.Finally, we relate these texts and discourses to the broader context in which each text was created.Because CDA is an in-depth technique-researchers often write articles about just one text-we needed parameters for selecting stretches of text to analyze.When the propositions were passed, the most significant elements of the new laws-in terms of their assumptions about language learning, their contrasts to existing ways of teaching ELLs, and/or their implications for classroom practice-were (1) the mandating of Structured English Immersion (SEI) as the default form of language education, (2) the establishment of parental waivers as the only way to access other programs for language education, and (3) a one-year time limit on students' placement in the SEI classroom.We therefore elected to trace these three elements first to the state's website and then across the district texts.

Methods
In this paper, CDA and SFL provide not only a theoretical orientation to language and text, but a methodological approach to analyzing them.For each text that we analyze, we follow Fairclough's model for a three-level analysis through description, interpretation, and explanation (Figure 1).Yet, unlike a typical CDA analysis, we also explicitly compare texts, across time (from the 2000 proposition to 2016 websites) and place (three districts within one state).In effect then, our analysis might be represented as in Figure 2, with comparisons occurring between texts and horizontal comparisons occurring between all levels of analysis-texts, discourses, and contexts.

Selection of State and Districts
Although California, Massachusetts, and Arizona all passed similar restrictive language policies between 1999 and 2002, Arizona was selected as the focus for this study because it is where we live (although Authors 1 and 2 have both lived and worked in California and Massachusetts as well).This is relevant because order to carry out the "explanation" phase in CDA, it is necessary to have solid contextual knowledge.Arizona is the state for which we have the best, most recent contextual knowledge, not only at the state level, but on a district-to-district basis.At the district level, we chose the largest school district in each of the state's three main regions: north, central, and south.Thus, while the paper contributes an understanding at the theoretical level of how multiple nested context(s) and text(s) interact, for us, it also sheds light on how teachers and parents in the districts around us might come to understand state policy.

Selection of Websites, Site Mapping, and Selection of Texts Within Sites
The websites we analyzed (see Appendix A for all URLs) were simply Arizona's Department of Education site and each district's official website.To select texts from each site, we began from the homepage and sought information relevant to English language learners using two approaches: First, we clicked through the site as a visitor would, following links that seemed promising (e.g., "Programs" >>> "English Language Learners") and mapping our paths through the sites, tracking site structure in terms of links and where links led.Second, in order to ensure that we had not missed any pages, we used the search function on each site to locate any other information related to English Learners, SEI, waivers, dual-language, bilingual programs, language immersion, or language acquisition.We then worked backwards from these pages to the homepage.Appendix B shows the paths to all of the documents we analyzed, as well as any other site information on language education and ELLs that was not selected for analysis.
Once we located all texts relevant to our key terms, if there was a single text (page or document) that addressed all three elements-SEI and program alternatives, waivers, and the oneyear timeline for learning-we chose that document for analysis.If there were multiple texts that addressed all three on a site, we chose the text that was most accessible (written for a general audience) and easily locatable (least clicks, most logical link path).If two texts addressed the elements in very different ways, or each addressed only some of the elements, we selected both.(See Appendix A for full text list.)

Approach to Description and Interpretation of Texts
As suggested by Janks (1997), we began our initial analysis by looking at multiple elements of language from an SFL perspective: Lexical choices, thematic structure, appraisal (evaluation), mood, modality, active and passive voice, transitivity, and cohesion, as well as what Fairclough (2003) calls assumptions, or the shared knowledge and beliefs that a text presumes.Only after initial exploration and the coding of extracts of text did we find that modality was the most fruitful for showing differences between texts.Modality is the amount of certainty and/or obligation that a text conveys.Modality can be broken into two types: epistemic modality and deontic modality1 .Epistemic modality indicates probability, or the degree of certainty that something will occur.For instance, saying "I might make dinner tonight" conveys a lower certainty on the part of the speaker than "I'll probably make dinner tonight," or, more certain still, "I'm absolutely going to make dinner tonight." The second type of modality, deontic modality, instead conveys obligation.To say, "I could cook for you," shows less obligation than saying, "I should cook for you," and much less than, "I have to cook for you." For the purposes of this paper, we analyzed both types of modality-epistemic (certainty) and deontic (obligation)-together, coding all of the original propositions in their entirety for these elements.We then coded all the parts of the website texts in which SEI, waivers, and the one-year timeline were addressed in the same manner.The umbrella category of modality is therefore the primary textual feature that we discuss in our findings, although we also discuss evaluative language and assumptions to a lesser extent.

Approach to Relating Text to Context
For the explanation phase of analysis, we drew, as most CDA researchers do, on our own understandings, created through our own collective many years of experience in the contexts in question, to explain the Why here?and Why now? of our analysis.In addition to our experiences, however, we also made calls to district offices and schools to gain a better understanding of the specifics of the student populations served in each district and to gather details surrounding the implementation of language education curriculum in schools with bilingual and/or SEI curriculum.

Findings
In Arizona's Proposition 203 SEI and programmatic alternatives.At the end of the opening rationale for Arizona's proposition, in which the reasons for children needing English are delineated, the text states, "It is resolved that: all children in Arizona public schools shall be taught English as rapidly and effectively as possible."The rest of the proposition is then spent outlining how this will be done: through Structured/Sheltered English Immersion [SEI].The proposition conveys that SEI will be the sole model for educating ELLs.One way that the text does this is through language of high certainty (epistemic modality) and high obligation (deontic modality) around SEI.In the following example, and throughout the rest of the document, HIGH CERTAINTY and HIGH OBLIGATION will be printed in bold capitals and underlined, Moderate Certainty and Moderate Obligation will be printed in bold, and Low Certainty and Low Obligation will be printed in italics.For instance, the proposition reads: ALL children in Arizona public schools SHALL BE TAUGHT English BY BEING TAUGHT in English.In particular, this SHALL REQUIRE that ALL children BE PLACED in English language classrooms.Children who are English learners SHALL BE EDUCATED through sheltered English immersion.
Words like "all" denote high certainty (contrast for, example, with the phrase "almost all" or "the majority of"), while modalized verbs like "shall be taught" and "shall be educated" denote high obligation, compounding the obligation already present in verbs like "require."Additionally, the propositions also convey high obligation/certainty that within these SEI classrooms, English is the only language that will be used.Arizona's proposition states that "'English language classroom' MEANS a classroom in which English IS the language of instruction used by the teaching personnel."The high obligation/certainty around English use continues (line breaks added for clarity; commentary in parentheses after each line): Books and instructional materials ARE in English (contrast with "should be" or "are usually") and ALL reading, writing, and subject matter ARE taught in English.(vs."most") Although teachers may use a minimal amount of the child's native language when necessary, NO subject matter SHALL BE TAUGHT in ANY LANGUAGE OTHER THAN ENGLISH, (contrast "any language other than English" with the weaker "another language" or "a language other than English") and children in this program LEARN to read and write SOLELY in English.
(contrast with "should learn" and "in English": "solely" augments, in terms of absoluteness, what "in English" would have conveyed on its own) All of the bold/underlined/capitalized words in this section serve to define SEI as an English-only program, through modalization as well as words that are in themselves high-obligation and highcertainty ("solely").It is also important to note that the only uses of low-obligation and low-certainty words-those is italics-are to discuss home languages in the classroom.
Waivers.Another way that Arizona's proposition presents SEI as the default program is by creating a legal procedure that parents must complete to have their children placed in a non-SEI program: the waiver system.All discussion of the steps in the waiver application process is written with high obligation language ("shall require," "to be provided"), while all discussion of the outcomes of the waiver process, after a waiver is granted, are written to convey low certainty and obligation.For example: The requirements of section 15-752 may be waived with the PRIOR WRITTEN INFORMED CONSENT, TO BE PROVIDED annually...Such informed consent SHALL REQUIRE that said parents or legal guardian PERSONALLY visit the school to apply for the waiver and that they there BE PROVIDED a FULL description of the educational materials to be used in the different educational program choices and ALL the educational opportunities available to the child.
The language around requirements for waivers contains high certainty/obligation.Furthermore, in Arizona, a significant number of school officials must be involved in the approval process and teachers, in addition to districts, may reject waivers, and may do this "without explanation or legal consequence."Thus, although the waiver process ostensibly created a path to bilingual classrooms, by codifying the procedures for obtaining waivers, it simultaneously constructs strong barriers to bilingual education, through numerous, obligatory, and repeated requirements to parents.This section of the proposition therefore also served to support SEI as the dominant path to educating ELLs in these states.
One-Year Timeline.The one-year timeline to learning English occupies just a small part of the proposition's text: "Children who are English learners SHALL BE EDUCATED through sheltered English immersion during a temporary transition period not normally intended to exceed one school year."The overall programmatic plan (SEI) is again written with a verb form conveying both high certainty and obligation ("shall be educated"), while the last part is somewhat less certain: while ELLs' placement in the program is "not normally intended" to last more than one year, it may last longer.However, this piece of the legislation is significant for its connection to the assumptions that young children can learn a new language quickly, given the right circumstances, and that any failure to learn quickly must therefore be attributable to bad (read: bilingual) programs.Additionally, its specificity, and lack of a research basis (August & Hakuta, 1997;Hakuta, Butler, & Witt, 2000), makes it a particularly good element to trace intertextually as we examine the state and district websites, as it is unlikely to have come from any source but the propositions.

Arizona Department of Education Website
SEI and programmatic alternatives.On Arizona's website, SEI is ubiquitous.The first document analyzed, a brochure for parents titled, "Success for our English Language Learners," was subtitled "Information for Parents with a Student in a Structured English Immersion (SEI) Program."This suggests that SEI is synonymous with ELL education.Further evidence of this equivalency is the brochure's location, not within a thread dedicated to SEI, but under "English Language Learners">> "Program Information" >> "Parent Information," suggesting that it represents information for all ELLs' parents.Within the brochure, readers are informed in a section called "The Facts" that: Arizona has A LAW for students who are not proficient in English, also known as English Language Learners (ELL).An accelerated English language program, called Structured English Immersion (SEI), has been designed to help students learn English.
This statement does not utilize the highly obligation language that the propositions do, however, it presents the information as "facts" and "law."This sentence also uses the evaluative language, "help," to convey that the program is a good one.No other program models are presented in this document, or anywhere else on Arizona's website, although they are alluded to in the title of the waiver document available on the site, called the "Bilingual Program Waiver Request Form."Another document, "SEI Models," was the closest document on the site to a summary of program options, and it also only presents one model: SEI.Under the section of the document called "Structured English Immersion Model Components: Policy," the document states: ARIZONA LAW REQUIRES schools TO TEACH English.ARIZONA LAW REQUIRES materials and subject matter instruction TO BE in English.
ARIZONA LAW REQUIRES English language learners TO BE GROUPED together in a structured English immersion setting.
Like the "Success" pamphlet, this document refers explicitly to "Arizona law."Unlike the brochure, though, it adds the strong, high-obligation verb "requires," making clear that that all materials must be in English, that SEI is the required program, and that ELLs must be grouped in SEI classes.
There is a single document on the Arizona site, called the "Application for Alternate Proposed Program for English Language Learners (ELLs)" which lists a "Procedure for a school district or charter school to propose an alternate program for ELLs not already approved by the State Board of Education."This document, however, informs applicants that as part of the application, they must describe how all of the legal program requirements will be met, including: • Children SHALL BE PLACED in English language classrooms.
• ALL children taught in English using English materials.
• ELLs SHALL BE EDUCATED through Structured English Immersion (SEI).
• The period of SEI instruction is temporary -not normally intended to exceed one year.
This list, in effect, creates room for only one program model: SEI.
Waivers.On the Arizona site, there is no information about waivers in any of the documents, except for the waiver document itself.This absence was not surprising to us, given that the point of the waiver is to provide avenues to alternative programs and the Arizona site does not discuss alternative programs.The waiver document itself, however, can be found in a logical place on the site (see sitemap in Appendix B), and it does provide guidance as to the requirements to apply-visiting the school, applying every year, meeting one of the three acceptable reasons.Additionally, just before the signature line, the form reads in bold, "I was fully informed of my right to refuse to agree to this waiver" (Waiver, p. 1).This right to refuse a waiver, rather than to request one, is a right that is indeed stated in Prop 203 and provide another example of the Arizona site's close intertextual ties to the original proposition.
The one-year timeline.Arizona's site contains many references to the one-year timeline (parts that are verbatim from the proposition in each case are underlined).In the "Success" brochure, parents are informed that: "SEI is a program designed to help students learn English, in a time period not normally intended to exceed one year."The Application for Alternative Programs documents cites the law itself: "A.R.S. §15-752: 'Children who are English learners shall be educated through sheltered English immersion during a temporary transition period not normally intended to exceed one year'" (p.3) and also reminds applicants that they must "clearly describe how the proposed program design will meet the legal objective of English Language Learners reaching proficiency in a period "not normally intended to exceed one year" (p.6).In the SEI Models document, the one-year timeline is also mentioned in a way that does not quote verbatim: "The goal set forth in Arizona law is for ELLs to become fluent English proficient in a year.(A.R.S. §15-752.English language education)" (p.3).

Summing up.
As we have shown above, Arizona's Department of Education website has remained faithful to the original proposition, both textually and ideologically.Through language of strong obligation and strong certainty, SEI is presented as the sole option for ELL education and English as the sole language of the classroom.No other programs are discussed and the "Application for Alternate Proposed Program for English Language Learners (ELLs)" has the same effect on an institutional level that the waivers created on a family level: While they purport to provide avenues to bilingual education, they also set up significant bureaucratic hurdles.

Explanation: The State Context
Having described and interpreted the proposition and how it is presented on the state's Department of Education website, we now turn to Fairclough's third level of analysis, explanation, using our knowledge of the larger social contexts of Arizona.For the past several decades, Arizona's state politics have largely been controlled by Republicans, many from the most conservative wing of the political party.Politically and otherwise, Arizona prides itself on local control.Recently, Arizona chose not to adopt Common Core Standards on the grounds that local control needed to be maintained.In fact, Arizona's current Superintendent of Instruction, Diane Douglas (R), ran much of her campaign on the platform that Common Core is federal overreach (Rau, 2015).However, the privileging of local control does not extend to all populations, as in Arizona's one-size-fits-all model for educating ELLs.In Arizona, there is one test-the AZELLA -for assessing ELLs' English proficiency.Arizona has one model-SEI-for instruction.The state allows for one language-English-in classrooms.It also strives for one perspective for addressing issues of diversity, as seen in the controversy over the Mexican-American Studies program in Tucson.Given the antiimmigrant sentiment that continues to exist in Arizona (c.f.Sherriff Joe Arpaio or Arizona SB 1070), it is not surprising that policies that impact English learners are not afforded the benefit of local control so valued in our state.
CDA researchers, as stated previously, seek to understand how power relations influence and are influenced by text.Arizona's Department of Education's website presents Proposition 203 using very strong obligatory language and provides hardly any flexibility regarding SEI, waivers, and the one-year timeline.Language choices used in the website convey the message that the power to control the elements of the proposition is held by the ADE.When school districts have filed an "Application for Alternate Proposed Program for English Language Learners (ELLs)," the only applications that have been granted were to allow high schools to lower the four-hour ELD block of instruction to two hours.This demonstrates that even when school districts may appear to have choices and alternatives, in reality, the ADE significantly limits those choices.

The District Websites
We now turn to the websites for three school districts in Arizona-Tucson (southern Arizona), Mesa (central Arizona), and Flagstaff (northern Arizona)-to examined the same elements which we examined the propositions and the state sites.However, we begin here with a brief discussion of the websites themselves, as some interesting contrasts were visible in their very structure.(It may be helpful to look at Appendix B's site maps while reading this section.) Flagstaff's site (see Appendix A for URLs) had the clearest path to information about ELLs.A click on "Departments" led to a directory with just 18 listings.The first listing, "Bilingual Education," led (and was the only path) to Flagstaff's main page on language acquisition, called "English Language Acquisition" but with the subtitle, "Bilingual Education." Tucson's website was trickier to navigate.One path from the homepage led to a page called "Language Acquisition," with the subtitle, "Two-Way Dual Language," which contained text and links about the Two-Way Dual Language program and schools, as well as the image at right.However, Tucson has a second page called "Language Acquisition (no subtitle)," that was significantly harder to find, accessed only by clicking "Directory" on the homepage, then "Departments/Services," then clicking "L" to go to the correct section of the directory, and finally selecting "Language Acquisition."This second "Language Acquisition" page contained documents about SEI in addition to TWDL.On Mesa's site, a site with more than 200 links on its homepage, it was surprisingly easy to find the main page for "English Language Acquisition," a page that strongly featured SEI.Mesa, however, dedicates a separate area of its website to "World Languages."On World Languages, only "Dual Language Immersion" is featured.SEI and programmatic alternatives.Flagstaff.The listing in Flagstaff's directory, "Bilingual Education," was a strong and immediate indicator that the information on Flagstaff's site would not feature SEI as the default program option.On the main English Language Acquisition ("Overview") page, SEI is not mentioned, although the page highlights the importance of learning English: "All students NEED TO ACQUIRE their highest level of English language proficiency and literacy."However, immediately before that, the page states, "People who can speak, read and write more than one language enjoy academic, social, financial and cultural advantages."And at the bottom of the page: "In addition, we support bilingual education through a dual language Spanish/English program and a nationally recognized Navajo language revitalization program."While all of these show moderate-to-high epistemic certainty, none of the high-obligation language of the Arizona DOE site or the propositions is found here about any program type, instructional method, or language of instruction.
Additionally, on the "What is an ELL?" page, arranged like an FAQ, the question, "What programs and services are available for English Language Learners?" is answered by listing: • In this response, the three programs are posed as equivalent options.SEI and ILLPs are presented factually (high certainty; epistemic modality), but with no obligation (deontic modality).The only high-obligation language is presented around waivers (see below for discussion).Similarly, the Flagstaff page, "Summary of Programs," presents a two-column chart comparing SEI/ILLPs (collapsed together on the chart) with bilingual education.Although SEI is in the left (primary) column and bilingual education on the right (secondary), the chart is constructed to highlight the comparable elements of the programs.For instance, for each program's "Goal," both say, "To become proficient speakers, readers, and writers of English." Tucson.On Tucson's site, the "Language Acquisition: Two-Way Dual Language" homepage, there is no mention of SEI.The page declares: "Dual language IS BASED ON linguistic, sociological, psychological and pedagogical theories that emphasize learning through the use of the student's first and second languages as mediums of instruction THAT WILL RESULT in STRONG bilingual and biliteracy development."Using high-certainty language ("WILL RESULT") and citing sources of this certainty ("BASED ON"), this page shows strong support for TWDL as the default programmatic choice for ELLs.In the most comprehensive document that links from this page, the "Two-Way Dual Language Program Handbook (PDF)," SEI is only mentioned once, on the first page, under Governing Board Policy, immediately followed by a statement about waivers: English Language Learners (ELLs) SHALL BE EDUCATED through Structured English Immersion (SEI).ALL students, however, whose parents HAVE REQUESTED and RECEIVED approval for waivers SHALL HAVE THEIR CHILDREN TAUGHT through bilingual education techniques or other generally approved methodologies.
Although this text has strong ties to the original proposition and to Arizona's state documents, underlining the same default program option (SEI) and procedure for participating in alternative programs (waivers), it takes the granting of waivers as a given, occurring in the past (HAVE REQUESTED and RECEIVED).Additionally, in the paragraph immediately following, the same middle-to high-certainty/obligation language is used to talk about Dual Language Instruction as the program of choice: In the majority of educational research studies, Dual Language Instruction (DLI) is considered THE MOST EFFECTIVE FORM of bilingual education and SHALL BE IMPLEMENTED, wherever possible, as part of the curriculum for students with an approved waiver.The goal of Dual Language Instruction is to promote individual student achievement, to provide students FULL access to the curriculum, to ENSURE students' rapid acquisition of basic English language skills, and TO SECURE for students the opportunity to demonstrate MASTERY of at least two languages, one of which will be English.
The rest of the 14-page document is then dedicated to TWDL.Recall, however, that there is another section of the Tucson website found via the Directory.Although it is less likely that someone seeking information would find this other "Language Acquisition" page, it is worth including because it differs substantially from the TWDL documents.On the second "Language Acquisition" page, there is an SEI Program outline that states, "The SEI program REQUIRES that ELLs be taught separately for four hours of ELD instruction... Schools that qualify for an Individual Language Learner Plan (ILLP) exemption may combine ELLs with mainstream students for all or part of the four hours."Here, SEI is strongly the default, with ILLPs as the only other option.The "Parent Program Option" document, also linked from the second "Language Acquisition" page, begins with the paragraph from the TWDI Handbook ("English Language Learners (ELLs) SHALL BE EDUCATED..."), but without the subsequent paragraph on dual language immersion.These Jekyll and Hyde "Language Acquisition" pages, and their relative locations on the site-one prominent, one buried-suggest that while Tucson does convey state SEI policy somewhere on the site, they choose dual language instruction as the face of language acquisition for the district.
Mesa.Unlike Flagstaff-where SEI, mainstream (ILLP), and bilingual programs are discussed and directly compared in one location on the site-or Tucson-where SEI programs are discussed in one place and bilingual programs in another with each program addressing the other, albeit briefly-in Mesa, the two programs are completely separate and neither addresses the other at all.In one section of the website, "English Language Acquisition," SEI is presented as the sole program (although it is worth noting that, other than in the "Success" brochure, the program is referred to as ELD, or English Language Development, rather than SEI).While we will not revisit the contents of the "Success" parent brochure (see Arizona website analysis above), the same highcertainty, high-obligation language that the brochure used to talk about SEI are used in the "ELD Program" document on the same page: ARIZONA STATE LAW REQUIRES that English language learners participate in 4 hours of English Language Development (ELD instruction on a daily basis).
ARIZONA STATE LAW REQUIRES that ALL English Development classes be focused on students gaining English language proficiency as quickly as possible.English Language Development classes MUST USE specific strategies, practices, methods and materials designed to help each child learn English.
This document, like the ADE's "Success" document, uses "Arizona state law" to add obligation and alludes to the assumption that "faster is better" for language learning.The document informs parents of their "right to decline these services," but does not offer any alternatives.Meanwhile, in the other section of the site, "World Languages >>>Dual Language Immersion," two-way dual immersion (TWDI) alone is discussed.On this part of the site, the only program presented is a 50-50 English-Spanish program with the following goals, listed at the top of the TWDI page: The Dual Language Immersion program is a program designed to empower students to achieve at a high academic standard.The program provides students with the opportunity to become college and career ready in a globally competitive world through academic excellence in content areas while acquiring scholastic achievement in two languages.Student success IS ENSURED through content knowledge and fluency in both languages […] A strong partnership between parents, teachers, and administration creates students who are bilingual, biliterate, and bicultural students who will cultivate an understanding and appreciation of diverse cultures.
This passage presents a very different stance on second language acquisition from that of the ELD page.It provides support for the TWDI program without any high-obligation language-just the high-certainty "is ensured"-and draws on discourses very different from the original propositions: neoliberal ideals of global competitiveness rather than national and nativist ideals of unity; goals of bilingualism, biliteracy, and biculturalism rather than quick transitions to English.However, this section of the website never mentions that the program serves, or is even intended to, serve ELLs.
Waivers.Flagstaff.Flagstaff's site mentions waivers once, on the "What is an ELL page?" under the description of bilingual programs (see previous section on SEI for exact text), and emphasizes the waiver requirement with high certainty/obligation ("Parents MUST SIGN"), describing possible acceptance with low certainty/obligation ("If the waiver is accepted, the student can participate").The site does not give details about the waiver process-what qualifies a student for a waiver or whether the default position is acceptance or rejection-but on this page, as well as the "Summary of Programs" page, bilingual programs are discussed as programs that include ELLs, implying that waivers are a viable path to entry into these programs.
Tucson.Tucson's site discusses waivers on both of the two different "Language Acquisition" pages.In both places, the text is the same.After describing SEI as the default option, both documents state: "ALL students, however, whose parents HAVE REQUESTED and RECEIVED APPROVAL for waivers SHALL HAVE THEIR CHILDREN TAUGHT through bilingual education techniques or other generally approved methodologies."Rather than the low-certainty language around waivers used in the proposition ("If a parental waiver has been granted"), this statement takes the waivers as a given, having already occurred in the past, at least for some group of students.This subtle difference reflects-and constructs-very different assumptions about the waiver process.Tucson's site also includes the waiver application itself, however it is found under the "Parents" link of the second, difficult-to-find Language Acquisition page.
Mesa.On Mesa's website, waivers are not discussed at all, nor can a waiver form be found on the site.Just as it made sense that Arizona's state site did not discuss waivers, given that it did not discuss any programs to waive to, it also makes sense that no waivers would be discussed on Mesa's site, given that waivers are associated with program choice, and on Mesa's site, SEI and TWDI programs are not presented as two choices, but as separate entities.The only (potential) allusion to waivers was buried in "Dual Language Immersion Program guidelines" for Hermosa Vista Elementary school: "English language learners (ELL) are eligible to participate in the program.Students who are identified as ELL by the English Language Acquisition Department and the (the district and state mandated assessment) will need to complete additional state required documentation."We imagine that the "additional state documentation" is a waiver.
The one-year timeline.Flagstaff.While Flagstaff does not use the propositions' language to discuss the one-year timeline to English learning, it does mention in the "Summary of Programs" chart that in both SEI/ILLP and bilingual education programs, "ELLs usually acquire basic English skills in one to two years."Yet, the goal that ELLs will learn "basic" English in one-two years is a very different goal from the propositions' goals to that students will be proficient enough to be exited from SEI (or English language development classes of any kind) after one year.
Tucson.Tucson does not mention the one-year timeline in any of its documents, nor is there any discussion of speed of acquisition.
Mesa.The only place on the Mesa site that the one-year timeline is mentioned is in the "Success for English Language Learners" brochure, created by the Arizona Department of Education and discussed above in the section on the state websites.The brochure also mentions that the goal is for children to learn English quickly.The ELD program document does not mention the one-year timeline, but it does say, "English Language Instruction Arizona State law REQUIRES that all English Development classes be focused on students gaining English language proficiency as quickly as possible," perpetuating the speed emphasis of both the propositions and the Arizona site.Summing up the districts: contrasts and links to the propositions.Using the critical discourse framework discussed previously, we can examine the intertextual relationship between Proposition 203 and each website, as well as how the district texts relate to each other.The three districts in this study represent Arizona's Proposition 203 quite distinctly on their sites.Flagstaff appears to be the district with the weakest intertextual ties to the original proposition.Flagstaff's website represents SEI (structured English immersion), ILLPs (individual language learning plans), and Bilingual Education as equivalent options with a common goal (of supporting students in becoming proficient speakers, readers, and writers of English).By also including statements that recognize the value of being bi-or multilingual and bi-or multicultural, the district demonstrates a commitment to promoting programs beyond just SEI by providing parents with choices about how to educate their children.Tucson's website shows close intertextual links to Prop 203 only on its hard-to-find Language Acquisition page.On its main homepage, it instead presents TWDI (two-way dual immersion) programs as the default program of instruction for ELLs, without mention of SEI.One document cites that the majority of research studies demonstrate that DLI (dual language immersion) is the most effective form of bilingual education, indicating that the choice of programs is based on research, not just an ideological stance.Only on Tucson's second, hard-to-find Language Acquisition page, is SEI presented as the default program for ELLs.Mesa seems to have stayed the closest to the original proposition both in terms of ideology and text.Mesa's website keeps separate its representation of SEI, referred to as ELD (English language development), from its presentation of DLI.When describing ELD, the website consistently refers to Arizona state law and emphasizes the need for ELLs to learn English as quickly as possible.The TWDI content aligns with ideals of bilingualism and biculturalism, but in contrast to Flagstaff's explicit inclusion of ELLs, it is unclear from the site whether ELLs are part of the TWDI program.

Explanation: The District Contexts
Referring back to the critical discourse analysis framework, it is crucial that we examine the district website texts within each's district's cultural, historical, and social context.We take each district in turn.
Flagstaff is a mountain town that borders the Navajo Nation and is in close proximity to the Yavapai, Havasupai, and Hopi reservation lands.It is also home to Northern Arizona University.Its geographic distance (140 miles north) from the state's capital of Phoenix, its proximity to significant American Indian populations, and college-town feel are not, we argue, insignificant factors in Flagstaff's long history and commitment to the cultural and historical diversity of its local communities.Flagstaff's American Indian population is nearly four times the percentage represented in Mesa, Tucson, or Phoenix.It is home to one of the most-studied dual language schools in the state, Puente de Hozho Bilingual Magnet Elementary School, which is open to all students in the area and reflects the linguistic diversity of its community, with all students at Puente receiving 100% of their day in either Spanish-English or Navajo-English.This school is another example, like the community discussed by Newcomer & Puzio (2014), of how restrictive language policies in Arizona can be negotiated.As a whole, the community in Flagstaff has demonstrated an unwavering commitment to financially supporting FUSD's programs, through overrides and bond elections.
Tucson, 60 miles north of the U.S./Mexico border and 110 miles south of Phoenix, has been at various points in its history, a part of both Mexico and the United States.When the first public school in Tucson opened in October of 1867, enrollment was limited to males whose primary language was Spanish, and one of the first teachers hired for the school, John Spring, taught by first giving instruction in Spanish, and then in English (Cooper, 1867).Given its history, it is not surprising that TUSD currently boasts 10 DLI programs.It was also the first district in the state to offer a DLI program from kindergarten to twelfth grade, and has a history of promoting bilingual and culturally relevant pedagogy as a key component of its K-12 curriculum.Five years ago, TUSD made national news when a super majority of the Republican-led state legislature passed House Bill 2281 outlawing TUSD's Mexican American Studies (MAS) program, as well as prohibiting any school or school district in Arizona from offering programs that are 1 designed primarily for pupils of a particular ethnic group or advocate ethnic solidarity rather than treating pupils as individuals (HB2281, 2010).This instance is just one example of TUSD's history of promoting bilingual and culturally relevant pedagogy, even when these are unpopular at the state level.Perhaps naturally then, TUSD has emerged as site of creative resistance to propositions and house bills that disproportionately affect their diverse, and majority Latino/a student population.
Mesa is located in Maricopa County, the same county as Phoenix, and is a suburb 20 miles east of Phoenix.It is the main city of the East Valley section of the Phoenix metropolitan area.Mesa was named the most conservative big city in the country in a recent study analyzing public policy in cities across the country (Tausanovitch & Warshaw, 2014).The most recent census data shows that Mesa continues to be majority white (non-Hispanic) (64%), middle class ($43,000 median income), and suburban, and it is unanimously represented at the state and federal levels by Republicans.MPS is the largest school-district in the state of Arizona and is one of the most diverse (just 46% Anglo/White students), yet the demographics of the city (64% Anglo/White) do not match the schools.This may help to explain why, Mesa was found to be one of the locations in the state where media coverage showed the strongest support of SEI policies (Gonzalez-Carriedo, 2012).Mesa's Human Relations recently hired Arizona State University's Morrison Institute for Public Policy to do a survey and found that about 75% of Mesa's residents believe that their community is an inclusive community, and the same number said they had not experienced or witnessed discrimination (Hunting, 2014).Yet, the same study found that discrimination continued to plague certain minority groups.Furthermore, about half of respondents-particularly those who reflect Mesa's diverse ethnic, religious, sexual and other identities-rated Mesa residents' overall cultural awareness as average or below average.This points to two different realities for those living in Mesa-one experience for those from dominant ethnic and linguistic backgrounds and one for those from nondominant backgrounds-a duality reflected on MPS's website in the completely separate programs portrayals: English-only education through SEI for English language learners, but dual language immersions for English speakers who want to learn a second language.

Conclusion
While educational policy certainly impacts students, key stakeholders in the education of K-12 students may not always access policy directly, instead relying on intermediaries such as departments of education and local districts to relay educational policy.For this reason, it is important to understand how policy is conveyed in texts found in those venues and how the local context in which the texts are produced comes to bear on the "translation" of those policies.Using a critical discourse analysis framework, we have examined the text of Arizona's restrictive language policy, Prop 203, in relation to how it is conveyed to the public a decade and a half after its passing.By contrastively describing, analyzing, and explaining the state proposition, as well as state and district websites, we have highlighted differences in both the texts as well as the social and ideological contexts in which these texts were produced.

Implications
The contributions of the paper are significant for both theory and practice.First, theoretically and methodologically, this study illustrates in a way that most CDA studies cannot, just how tight the relationship between context and text can be.This study looks at four website texts which, based on the law, should be very nearly the same, and shows that instead they are quite different.Thus, the relationship between the textual contrasts and the contextual variation is clearer in this study than it might be in a study where only a single document is examined.
At the level of educational practice, one of the practical applications-and future directions-that we see for this research comes from our work in teacher preparation.As we work with teacher candidates to help them understand research about language acquisition, we encourage them to use this research to turn a critical lens on what they have experienced in their student teaching placements.Sometimes, they conclude that they would like to teach in a district that does not approach English language teaching the way that their current district does, yet, they wonder how they will know-before taking a job-whether another district does things differently.We therefore plan to build on the findings from this paper to create a tool to support our teacher candidates is using school and district websites as indicators of a school or district's stance toward educating ELLs.After introducing the idea that language reflects and constructs context, the tool will lead them through a mini-CDA, asking them to look for evidence of assumptions about language learning and ELLs, stances toward multilingualism, and alignment or distance from the three key elements of Arizona's English-only law.The tool will promote students' informed decision-making about their future employment, foster their development as critical consumers of text, and support their continued engagement with questions of who and what are valued in various approaches to teaching ELLs.

Authors' Note:
The authors would like to thank Wayne Fan, designer extraordinaire, who transformed our site maps from spaghetti on paper into something useful.Thank you also to the guest editors and the three anonymous reviewers for their very helpful comments and suggestions.

Figure 2 .
Figure 2. Multilevel CDA carried out in this paper These programs are offered at Puente de Hozho Tri-lingual Elementary Magnet School and Puente de Hozho MS Academy.Parents MUST SIGN a parental waiver for their child to participate in a bilingual program.If the waiver is accepted, the student can participate in either the Spanish/English Dual Language Program or the Navajo Immersion Program.The goal of these programs IS for students to become proficient speakers, readers, and writers of English and another language (Spanish or Navajo).
Education Policy.He teaches graduate and certification courses to students in the Educational Leadership Program.His research examines education policy issues broadly related to human resource management, instructional supervision, and school finance.Gabriel University of Connecticut rachael.gabriel@uconn.eduRachael Gabriel is an Assistant Professor of Literacy Education at the University of Connecticut, and is an associate of the Center for Education Policy Analysis (CEPA), and the Center on Postsecondary Education and Disability (CPED).Her research inte rests include: teacher preparation, development and evaluation, as well as literacy instruction, interventions, and related policies.Rachael's current projects investigate supports for adolescent literacy, disciplinary literacy, state policies related to reading instruction and tools for teacher evaluation.