Multilingual Language Practices in Education in Pakistan: The Conflict Between Policy and Practice

This study examines the language practices in educational settings in Pakistan, taking the multilingual groupings in society into account. In Pakistan, each province is linked to the single or multiple identities of its people and the languages spoken by the majority. The national language Urdu is limited to educational settings and its function as a lingua franca. English serves people in authority, in offices, and in educational settings. Through discourse-ethnographic analysis, this study examines the individual and joint actions of policymakers and teachers to understand the role of language in educational policy and its practice in educational settings. The interview data identified issues regarding the relationship between language, identity, nation, region, religion, power, and personal attainment in regional, national, and international settings. Moreover, the power of national education policy to produce adequate results is limited by the regional discourses that policymakers ignore. This study concludes by arguing that policy practices for language-in-education in multilingual societies require thoughtful planning which should be informed by local conditions and requirements for its better implementation.

Creative Commons CC BY: This article is distributed under the terms of the Creative Commons Attribution 4.0 License (https://creativecommons.org/licenses/by/4.0/) which permits any use, reproduction and distribution of the work without further permission provided the original work is attributed as specified on the SAGE and Open Access pages (https://us.sagepub.com/en-us/nam/open-access-at-sage).

Original Research
Language studies that employ social and cultural methods have exposed a complex picture of language practices and language learning. Language learning is a process of learning the social and cultural practices of a community, of which language or linguistic performance is an integral part (Heath, 1983;Ochs & Schieffelin, 2008). In addition, language learning shapes learners' multiple identities and is an investment in creating a national identity (Gellner, 1983;Peirce, 1995). Related studies have demonstrated that access to language learning is a major problem for learners, particularly for learners of a second language (Duff et al., 2002;Miller, 2005;Pomerantz, 2007). Individuals with low proficiency in the dominant language of the area in which they live are likely to fall into the trap of not speaking any language well, which ultimately decreases their chances of securing good jobs and improving their status in society (M. A. Ashraf & Tsegay, 2016;Butler, 2015;Khalid, 2016;Phillipson, 1992). For example, the current dominance of English over other languages is a feature of verbal exchanges at international, national, and local levels in many countries, including Pakistan. Despite its international status, and ownership of major advantages, the English language has become a commodity for the speakers of other languages.
The debate over personal attachment to linguistic groups has been conducted in different ways and according to the demands of particular settings (Bourdieu, 1991;Gellner, 1983;Pennycook, 2017). In the Pakistani context, language status planning is based on the notions of nationalism, religion, and development (M. A. Ashraf, 2018b;Shamim, 2008). Whereas nationalism and religion appeared to support the national aspiration of constructing unity among Muslims of different ethnic identities, development supported by expertise in the English language became crucial in Pakistan's rise internationally. This notion of nationalism projected to define the nation, as "false" if nationals had no claim to national self-determination, and consequently could not claim sovereignty over a nation state (Gellner, 1983). However, the term nation is more difficult to define than nationalism, as nationalism is primarily a political principle that provides a baseline description to 1004140S GOXXX10.1177/21582440211004140SAGE OpenAshraf et al.

research-article20212021
1 Hunan University, Changsha, China 2 Beijing Normal University, China 3 Nanjing Normal University, China connect a people with their land (Hall, 1993). To make sense, it requires the political and national units to be consistent. Meanwhile, while language is associated with both the nation and nationalism, the structure of this intersection between language, nation, and nationalism is very complex, particularly in multilingual societies such as Pakistan (Rahman, 2001), where this structure is controlled via language ideologies in administrative and educational settings. For example, Article 251 of the Pakistani Constitution (National Assembly of Pakistan, 2018) requires the Pakistani government to replace English with Urdu in official settings, but in reality this has never been implemented (Channa, 2017). In this context, it is timely to explore the connections between power, ideology, and language education in Pakistan, which have important implications for policy development in multilingual societies.
This study aims to examine the language adjustments in policy, contentious responses from different stakeholders, and plans for realizing the language adjustment in multilingual societies of Pakistan. This study exposes complex processes whereby policy is negotiated and implemented, and discusses the effects that may influence the course of future policy implementation, particularly in teaching spaces. The study is significant as it discloses current progress in language-in-education policy in Pakistan and discusses how conflict in policy practices impact sociopolitical and educational practices. The study also examines how different language practices in educational institutions consecutively impact policy implications, and how language and policy practices (re)produce different language ideologies. This study contributes to extant literature by arguing that policy choices for language-in-education in multilingual societies require deliberate planning for better implementation and should be informed by local conditions and requirements that are necessary for successful policy implementation.
This study adopted discourse-ethnographic methods to examine how stakeholders including teachers and policymakers in Pakistan negotiate language choices by challenging local and national ideologies in educational settings where English is the dominant language, Urdu the national language, and local languages most accurately signify the speaker's identity. By focusing on four groups of provincial language speakers (i.e., Sindhi, Punjabi, Balochi, Pashto), this study explores how language ideologies have been perceived in policymaking and educational settings, and how these ideologies affect individuals' decisions regarding which language to use for which purpose. This study illustrates the complexities of language choice for teachers and policymakers and the conflicting ideologies in Pakistan. The study focuses on three questions: Research Question 1: How do policymakers and teachers view the issue of language in educational settings? Research Question 2: How do teachers from different linguistic backgrounds perceive language choice in educational settings?
Research Question 3: What issues do teachers consider when making language choices in their teaching practices?

Language Pluralism and Language Ideologies in Pakistan
Pakistan is a multilingual and multicultural country with speakers of 77 languages (Eberhard et al., 2020). While each province is linked to a specific ethnic group that has its own language and culture, English is the language of bureaucracy together with Urdu, the national language. The Urdu language is generally used for communication and as a lingua franca between different ethnic groups. It serves to create national unity between the people of Pakistan, a function it also served among Muslims in the Indian subcontinent before partition in 1947 (Channa, 2017). However, English remains the language of the elites in Pakistan and is used for both official and informal interactions (Rahman, 2005). It has become embedded into Pakistani society and is considered a highly desirable language by a vast majority of the population as it has high status as the language of education, law, government, science, and technology (Mansoor, 2004;Rahman, 2002), whereas other languages are excluded from official and power structures of federal government, as well as from provincial power structures, such as the Punjabi language in Punjab.
Pakistan inherited its education system from the British colonial system and includes both English-medium and vernacularmedium education for specific administrative roles (Channa, 2017;Rahman, 1996). English-medium schools serve the country's elite class who can afford expensive fees, and as such these schools are considered symbolic of quality education (Rahman, 2002). Vernacular-medium schools include Urdu-medium, Sindhi-medium, and Pashto-medium. Urdu-medium schools comprise the majority of the schools that have been supported by the national government since Pakistani independence due to Urdu's status as a national language (Channa, 2017). However, questions continue to be raised regarding the quality of education in these schools. Due to the symbolic power of English as the language of the elites in Pakistan, a large number of nonelite private schools have arisen, promoting themselves as Englishmedium schools (Andrabi et al., 2008). Despite attracting middle and lower-income families to English-medium classes, the quality and language proficiency of teachers and students in these schools remains poor (Rahman, 2002). Several studies have been conducted on English teaching and learning and English as a medium of instruction (H. Ashraf, 2018a;Channa, 2017;Das et al., 2006;Halai, 2007;Khalid, 2016;Mansoor, 2004;Rahman, 2004;Shamim, 1993). These studies focus on educational discourses in classrooms and their effects on learning in Pakistani classrooms, which are diverse in terms of culture, language, and socioeconomic class.
In Pakistan, there are many linguistic and cultural ideologies that affect each ethnic group's view of language in society. One of these ideologies is the claim that national values are best approached and transmitted through Urdu, which, as mentioned above, was also the language of Muslims in subcontinental India before partition (Rahman, 2002). This ideology supports the labeling of English as a colonial language and implies that younger generations should not hold the norms of language and culture that English brings with it. A second ideology present in Pakistan is that an individual's identification with a group is mediated by their language and culture. Each ethnic group has its own language and culture, which is seen as an essential element of their ethnic values. In this ideology, both Urdu and English are considered a threat to local language, because government support for Urdu and English has restricted the opportunities for other languages. A third ideology is linked to economic growth and the international market for jobs and opportunities (Khalid, 2016;Shamim, 2008). This ideology is supported by many stakeholders from government and policymaking institutions, and it describes English as a language of opportunities (National Education Policy, 2009). Improving students' expertise in English is another important discourse in the cultivation of quality education. Because English and Urdu have been privileged by administrative powers, they have developed a functional stature and significance as well as exerting symbolic power to enforce a unified linguistic market, controlled by the official languages (Bourdieu, 1991). Therefore, a link has been created between power and inequality for those who aim to prioritize their own language over the national language (Rahman, 1996). These ideologies intersect in such a way that challenges and conflict are created among groups in society.
In policy contexts, the notion of one nation one language (Pakistan and Urdu) was adopted from the outset to create a sense of nationhood by those in power, and this caused major conflict among the majority language groups (see Rahman, 1996). Urdu was named a national language at the birth of Pakistan, and again in the drawing up of the 1973 constitution, and was later promoted as a unifying symbol of Pakistan (Haque, 1993) as well as a language of Muslim unity (Channa, 2017;Rahman, 1996). However, English remained the language of several domains of power, such as administration, judiciary, military, higher education, and commerce, as these domains had been under British rule before 1947. The scuffle between Urdu and English continued among governments from the 1990s, with support for both Urdu as a symbol of national solidarity and of English as the language of development (Shamim, 2008). The symbolic power of English was accepted by the government through its National Education Policy (2009), with the medium of instruction changed from Urdu to English for sciences and mathematics from Grade 4 onward. In addition, Pakistani judicial law, which aims to govern social behavior, exhibits another complex picture of ideologies. Article 251 of the 1973 Constitution of Pakistan claims that Urdu, as the national language, should be used for official purposes after a transition phase of 15 years from the date of the signing of the constitution. As such, on September 8, 2015, the Supreme Court of Pakistan ordered federal and provincial governments to fully implement Article 251, which was passed by the National Assembly in 1973 but never implemented. The Supreme Court decision has created a circumstance of urgency, where it is essential to understand the language ideologies that influence policymakers and educators in this complex and multifaceted setting.

Language Ideologies as a Theoretical Framework
Language practice and language choice are commonly controlled and influenced by language ideologies. Language ideologies are defined as patterns of belief that are implemented in context by the connection between language and social power structures (Kroskrity, 2016;Mar-Molinero & Stevenson, 2016;Verschueren, 2012). Language ideologies build relations between languages and language practices by creating a particular opinion to shape the language structures as well as influence the manner of speaking (Gellner, 1983;Silverstein, 1979). This relationship also affects social concepts by generating the collective values, economic standards, and political control that reflect the social standards and principles of a society (Phillipson, 1992). This social control is governed by economic and political factors. Economic factors persuade marginal groups to acquire a particular national or international language to obtain a stable economic status, while political factors aim to strengthen policymakers' positions in controlling and spreading personal ideologies through a particular language (Phillipson, 1992;Rahman, 1996). Therefore, language ideologies impact language production through the systemic link between language, power, and society (Bourdieu, 1991;Piller, 2015). A dominant language is leveraged to preserve particular social and official structures by disrupting the social and official structures of other languages (Phillipson, 1992;Van Dijk, 2000). Language ideologies are connected with the material reality of language, which means that communication between people and the methods of communication may not be separated from their physical qualities that are connected with an individual's identity (Voloshinov, 1986). Language policy, education policy, political power, and cultural control are entirely interlinked and cannot be ignored in education studies that are directly impacted by such policy (Phillipson, 2008). Through this focus, language ideologies can be used to effectively explore the structures of belief that interact with language and language practices.

Research Setting
In this study, discourse-ethnographic methods were used to identify the language ideologies that informed learning practices in educational settings in Pakistan. This interdisciplinary technique uses perspectives from discourse studies (Barker & Galasiński, 2007;Fairclough, 1992;Halliday, 1994) and ethnography (Agar, 1986;Canagarajah, 1993;Fabian, 1983;Kirk & Miller, 1986). In discourse, this study accepts that language users establish social realities, identities, and relations in interacting with other social groups (Fairclough, 1992;Wodak & Savski, 2018). This connection helps us to portray discourses between language and social control through power (Wodak & Savski, 2018). In ethnography, researchers are required to reflect on and represent informed interpretations and representations (Fabian, 1983). This study paid attention to the things that give words meaning, including factors that affect surroundings and interact with different aspects of community practices (Kirk & Miller, 1986). In this way, it was necessary to go beyond simply direct reporting of the subject's situation to capture and signify the elements that give meaning to words (Agar, 1986;Canagarajah, 1993). This method allowed us to discover discourses in language practices. The study included participants from just six major languages in Pakistan.

Participants
With regard to the research questions, a combination of snowball and convenience sampling was used to locate and recruit participants based on the pre-established criteria of being (a) current school teachers working in four provinces (Punjab, Sindh, KPK, and Balochistan), (b) current university teachers, or (c) current members of staff who hold administrative positions in the policymaking process at major national policymaking institutions. Participation was voluntary and ethical formalities were fulfilled before conducting the formal interviews. Participants' anonymity was also assured.
All participating school teachers were currently working in public and private schools in different regions across four provinces of Pakistan. The majority of the teachers (16) were working in urban schools, whereas a few (8) were located in rural areas. There were 19 male school teachers and six females, aged from 25 to 50 years with working experience ranging from 4 to more than 20 years. All teachers could speak more than two languages (Urdu, English, and a native language). The university teachers had doctoral degrees in computer science and environmental sciences and were working at universities in Punjab and KPK. The policymakers held executive positions at their institutions in Islamabad and were responsible for conducting research and constructing guidelines for national education policy. These institutions are major sites for preparing national education policies in Pakistan. The policymakers in this study had experience of working 10 to 25 years in educational settings and were part of the policymaking teams responsible for making national education policies in 2009 and 2017.
To contact and recruit policymakers, a stepwise approach was used. In the first step, information on institutions that participate in preparing national education policy was collected through national education policy documents. Government websites were consulted to identify the stakeholders in national education policymaking. In the second step, emails and phone numbers were collected from the websites of each institution. In the third step, an email containing information about the researcher and research aims was sent to the email addresses provided on the websites. In addition, phone calls were made to the numbers made available on websites to ensure that the right person was approached in line with the research purpose.
In conclusion, two policymaking institutions (Policymaker 2 and Policymaker 3) responded to the initial email, while Policymaker 1 was approached via phone call. Appointments were made with each participant for interviews. Although few in number, the major role these institutions play in policy formation and the position of the policymakers in their institutions provided the comprehensive information necessary to complete this research. Descriptive information of participants can be seen in Table 1.

Data Collection
Data for this study were collected from January to August 2017 through semi-structured interviews with 24 school teachers (n = 6 from each of the four provinces), two university teachers, and three policymakers from three different policymaking institutions. Data were also collected through unstructured observations in the institutions of all the participants and the study of national education policy documents. All interviews were conducted in Urdu, but words from English and local languages were also present during discussions. All interviews were conducted face-to-face by the first author who visited each participant, and each interview lasted for 45 to 90 min.
During the data collection stage, participants were coded into different groups based on their regions and positions. For example, school teachers were coded as Sindh 1, Sindh 2, and so on; policymakers as Policymaker 1, 2, 3; and university teachers as University Teacher 1, 2. Throughout the data collection, the main aims were (a) to identify educational intentions and goals of education policy, (b) to understand the application of national policy and language practices, and (c) to identify the role of languages in their particular spaces in a multicultural and multilingual society. After data collection, all interviews were transcribed and translated into English for analysis.
Data analysis strategy. After transcribing the interview data, the research team divided it into different categories using the three-stage systematic analysis (Clarke, 2005;Thornberg & Charmaz, 2014;Wodak & Savski, 2018). We analyzed the data line by line, according to the importance of each word (Thornberg & Charmaz, 2014;Wodak & Savski, 2018). In the first stage, all interview transcripts were reviewed and examined for open coding to trace similarities and differences. Several coding themes emerged spontaneously at the beginning of data analysis (i.e., the different roles of different languages; differences in language ideologies among participants from different regions), but several themes were carefully devised only after reading the data many times (i.e., language practices in educational settings in each region). In the second stage, open codes were transformed into focused codes of meaningful categories by identifying the dimensions and relationships of each code (i.e., the role of each language in its specific region). In the final step, focused codes were developed into theoretical codes, previous research on language practice and language policy was reviewed, and data were compared with the literature (Figure 1).
These codes correspond to the answers to research questions: for first research question, issues relating to conflict in language policy, planning, and practice; the shifting concept of being educated, and for the second and third research questions, issues relating to language practices in teaching, language practices in curriculum, language practices in higher education, and language practices in religion. The analysis reveals the similarities and differences among teachers' and policymakers' beliefs about language use in educational settings in Pakistan.

Conflict in Language Policy, Planning, and Practice
In this study, first observation was the nameplates on the doors of each room inside the institutions. In Policymaker 1's organization, the nameplates were written only in Urdu, whereas in Policymaker 2 and 3's organizations, they were written only in English. Policymaker 1 stated that his organization had recently changed the language of nameplates from English to Urdu, to demonstrate its commitment to meeting the requirements of the law following the Supreme Court's 2015 decision. However, Policymakers 2 and 3 referred to the fact that people do not understand the meaning of titles written in Urdu on nameplates. Policymaker 1 also revealed this to be the case, because these titles had not been used in official settings before and are not part of any curriculum. Therefore, a common person who has learned Urdu at school may be able to read the titles, but also unable to make sense of them. However, the major factor in this discussion is that each institution uses a single language (English or Urdu) on the nameplates. This represents the historical power structures of English and Urdu, both in administrative and Note. Language in parentheses denotes participants' view of shifting their mother language (i.e., they used the regional language to communicate with their parents, plus Urdu with their children).
educational settings (Rahman, 1996). To trace the issue in terms of identity, the nameplates should be written in Urdu to reflect the national character of nationalism (Gellner, 1983;Hall, 1993). However, nameplates written only in English or only in Urdu display the power of bureaucratic settings (Phillipson, 2008), which may create social inequality in national and regional situations as other languages are left out from national bureaucratic settings.
In the context of policy, the government's language policy is highlighted in the National Education Policy (2009), which states that the Ministry of Education aims to develop a comprehensive plan of action for implementing English language policy in the shortest possible time, in consultation with the provincial and regional education departments. The aim is to provide better opportunities for disadvantaged groups in Pakistan who do not have access to quality education and English learning. However, our study of government documents did not find any relevant English language policy, nor a plan issued since the publication of this education policy. Participants, including policymakers, reported their perspectives on the English language. For Policymaker 1, English had been used in the region (subcontinent) for a very long time, and felt it was not possible to replace it, particularly in science and technology. Policymaker 2 had a different perspective; he felt that there was a need to consider both education practice and language practice. According to him, policymakers had never measured the costs and benefits of using any particular language in the country. As members of a policymaking team, both policymakers perceived themselves as being in positions where their voices must be heard and felt that their opinions should be considered authentic and valuable; however, they did not share common perspectives on language use in administration. Furthermore, during the conversation at his office, Policymaker 1 asked his secretary to type a letter for official use. He gave instructions in Urdu, but asked that the letter be written in English. This was not a unique situation; in most offices, staff communicate in Urdu, or in local languages if both speakers belong to the same language group, but official documents are released in English. When asked about this, Policymaker 1 laughed before giving his opinion. He pointed out two major roles of language: for communicative purposes and for work purposes. For communicative purposes, we learn most of our language from society, but for work purposes, we learn the language at schools, colleges, and universities. Therefore, he learned the local language from society to communicate with people, but learned English and Urdu at educational institutes for work purposes. Although he had learned Urdu at  school, he had never learned official practices in Urdu. Therefore, both he and his secretary are incapable of writing official documents in Urdu.
The Pakistani government has been training its workforce in English for years . . . since independence or even before that . . . and at this stage (now), when we have more trained workforce in English, policymakers cannot consider going back and beginning to train the workforce in Urdu . . . after the (Supreme court) decision, Urdu would be a new thing in official documents. (Policymaker 1) Policymaker 2 was of the view that the majority of the policymakers were not being judged on their ability in educational settings as well as their expertise in policy studies; more social and political agendas were involved. He pointed toward the very serious problem that the majority of the stakeholders in educational policymaking had no experience in education. Thus, they failed to see the problems present in society, specifically in education. As a result, policy documents were not based on local realities. He also outlined the major role of higher education institutions (HEIs) in the policymaking process. Such institutions are supposed to conduct research and provide solid feedback to the policymaking process, but instead they put pressure on the basic education system by demanding that English dominate the curriculum, because the curriculum in higher education is completely in English. They continuously tell national policymakers that English is the only available option for competing in a global world, particularly in science and technology. Therefore, from the perspective of HEIs, English should also be taught in basic education.
In my view, policymakers are still confused about the use of language, since we have many voices in the battle over language and everyone wants to win the fight. As a result, the issue of language remains unchanged, and is becoming more multifaceted, but everyone is satisfied since they are not losing the battle. (Policymaker 2) This statement is a strong indication of the conflict that exists in policymaking institutions, alongside official coercion that was rarely successful in promoting linguistic standardization. A similar threat of coercion and subtle practices was often essential to shape agreement among different stakeholders (Ives, 2004). This issue reflects the inability of government and citizens to implement the law. In other words, English retains its dominance within the government and society.
The participants from four provinces discussed the selection of language since independence. Participants from the Punjab agreed with the government's policy of recognizing one language as the national language. For them, it would be difficult for the government to support all languages as national languages. They identified with Urdu, which they felt was integrated into their national and religious identities as Urdu emerged as the language for Muslims in the Indian subcontinent prior to independence, and Pakistan was subsequently created for the Muslim population. Therefore, recognizing Urdu as the national language of Pakistan was considered logical in view of the language's connection with religion (M. A. Ashraf, 2018b). For participants from the Punjab, the country is one unit and one language is necessary to represent the unity of the country. These opinions reflected the operation of political power in Punjab politics and policymaking (Nawab, 2012;Rahman, 2002). Similarly, participants from the Punjab in this study displayed a relaxed attitude toward decisions made at the national level.
However, participants from other provinces, especially Sindh, criticized the language politics in Pakistan. For them, the Sindhi language has existed for centuries, before the creation of Urdu and was firmly attached to their identity as Sindhi people. Each minority language has its own ethnic recognition. Therefore, the politics of language in national government since independence have been generally rejected by the Sindhi people (Malik, 1997). These participants reported two major factors that had been handled badly in the past; first, loyalty to the country, and second, loyalty to the Sindh. To understand the first, Sindh 3 and Sindh 4 gave examples of times when Sindh people were criticized for being disloyal to Pakistan. The example below relates to the behavior of rulers, the army in particular, toward Sindhi language.
Let me give you one example. The Pakistani army stormed into Sindh University to abolish Sindhi language at the university in the 1960s. Do you know the reason? Because the nephew of one army general (General Tikka Khan) could not pass his Sindhi language exam at high school, which was necessary to get admission to the university. We all (Sindhi people) have witnessed such behavior from Pakistani rulers. (Sindh 4) Participants from KPK and Balochistan also disagreed with the government's actions on language in the past, but now perceive English as the only language that can guarantee success in the current political and national scenario. They held this view because they compared their students with students from the Punjab, who have better opportunities for gaining admission to university and acquiring good jobs because of their English skills.

The Shifting Concept of Being "Educated"
The notion of being literate was another theme that cropped up during interviews, particularly with Policymaker 1, who gave his opinions regarding society's behavior toward language. He used the Urdu word "Babu" to summarize his answer: In our society, parents wish to see their child like "Babu." Parents want to see their child speak English and dress up neatly, and comb their hair. (Policymaker 1) Here, the concept of "Babu" used by Policymaker 1 which is widespread in Pakistani society has historical roots. Babu was a term of respect used for literate or educated men. Before partition in 1947, English was primarily used by India's elitist administrative services, that is, the Indian Civil Service (ICS) and Indian Administrative Service (IAS), the officers of the armed forces, and in higher education (Rahman, 1996). The native Indians who worked in offices in British India were given the title "Babu" by their local societies. In addition, learning English and adopting the Western dress code were necessary for those working at ICS, IAS, and other government offices (Kachru, 1983;Rahman, 2002). Based on this concept, the view holds that everyone must learn English and wear Western clothes if they wish to work in official positions, or, in other words, to be called Babu.
Participants from the Punjab province emphasized the importance of English for official occasions and meetings and that English and Urdu were the only possible languages for official speeches, related to the ideology that English and Urdu speakers are superior to others. For them, speaking in English suggests superiority and is necessary to become a leader.
However, this notion was rejected by many participants from Sindh and Balochistan who mentioned that speaking the local language during official meetings would give the opportunity to other officers, mainly from other regions (particularly the Punjab), to laugh at them and call them illiterate.
They shared examples from friends and family who had witnessed such occurrences: My brother is an officer in the mineral department. He went to Islamabad for a meeting. He gave a speech in Urdu (which is a national language of Pakistan), but other officers (from other provinces) asked him why he did not speak in English. (Balochi 2) My friend went to attend an official meeting on education, and officers from Punjab clearly emphasized the importance of English, and demanded that all provinces use English in official settings. (Sindh 4) These discussions indicate the perception of language in official settings and beliefs about language inside elites. They also raise the issue of regional differences. Hence, Toolan's (1997) labeling of English as the language of high flyers seems true in the Punjab, and it is growing in other regions of Pakistan. To achieve this supremacy, the Punjabi language disappeared, first from higher education and later from all types of education (Rahman, 2002). Therefore, this approach has affected the politics of language in the country, where different social groups are struggling to establish their own positions.

Language Practices in the Curriculum
Participants demonstrated diverse perspectives toward both the languages in the curriculum as well as the language of the curriculum. For policymakers, the language of the curriculum has been a very complex issue both in the past and at present. In the past, the idea was that Pakistan is one country, Urdu is the national language; therefore, Urdu should be the language of the curriculum in state schools. Policymakers strongly approved the implementation of Urdu, which they considered beneficial for all regions in helping to establish their Pakistani identity. They preferred adherence to the Pakistani identity rather than the differently constructed regional identity. As education has been a national matter since 1947, policymakers considered that education should espouse Pakistani identity, and thus the Urdu language was privileged over regional languages. Participating teachers moderately acceded to these decisions, with some differences between regions. Economic impact was the major factor that affected participants' views about language practices as well as the shift from regional and national languages to English. They stressed the importance of the economic impact on language choice, not only for the language of the curriculum, but also as a state language; a lingua franca, or as a means of communication (Pomerantz, 2007;Rahman, 2005). Most participants in this study and the policy documents investigated perceive English as the only route to success for students and the job market as English is required for all good jobs, for high-status leaders, as well as admission to higher education.
Nevertheless, participants from Sindh seemed to value knowledge of Sindhi and appreciate the significant meaning and role it has in Sindhi society. Several participants stated that they felt comfortable using Sindhi and saw themselves as nationalists of the Sindhi language. However, they believed that it had not been sufficiently developed in the past and thus could not be used as the language of science and technology. They identified different political and administrative shortcomings that had limited the development of their language. Moreover, the Sindhi language had not been promoted as the language of science and technology, because those in power supported the dominance of English and Urdu.
Sindhi is not pure to be the language of science and technology. Sindhi has been the medium of instruction in Sindh for many centuries. We always feel comfortable with Sindhi. But policies after Pakistani independence restricted the role of Sindhi, by employing Urdu as the medium of instruction. So, we could not develop it as we could have if we had had a chance to develop it. (Sindh 3) These interviews revealed the varying positions and views regarding the role of Sindhi as the language of the curriculum. Opinions were divided on whether Sindhi could serve as an adequate medium to deliver scientific knowledge. There was a strong belief, particularly among those who oversaw higher education, that Sindhi was not suited for the teaching of modern complex notions of science. However, some believed that it could have been possible if efforts were made in the past, or even now, making it easier for students and teachers. They gave examples of Sindhi-medium schools, where the language of instruction was Sindhi and the whole curriculum had been developed in Sindhi language. However, they noted that the national government always treated these schools badly.
We have many Sindhi medium schools. If you visit different cities in Sindh, you will be surprised to see the curriculum and development of Sindhi language . . . but they did not get support from national government because they (schools) were using the Sindhi curriculum, which was in contradiction to the national curriculum (Urdu/English). (Sindh 4) The Pashto, Balochi, and Punjabi languages have not progressed in Pakistan. Participants from these language groups reflected on how these languages could not be used in the curriculum. One major reason given was that the languages could not adequately handle application to the sciences. A second reason was that these languages did not have any history of practice in official and economic settings. Therefore, constructing the curriculum in these languages was perceived as a huge task with unknown benefits.
The development of Pashto is too poor (in Pakistan). It cannot teach science . . . and math . . . and other subjects at higher (education) levels. (KPK 3) It would be difficult or impossible (I think) to apply the Balochi curriculum because it lacks language development in literature, science, technology and in each aspect (compared with English language). (Balochistan 3) However, this statement does not imply that the Pashto language inherently lacks the potential to teach sciences. Rather, these comments contained a critique of the forces that had impeded development of the Pashto language, along with other languages during the establishment of national control over languages. KPK 2 gave the example of Afghanistan, where Pashto is the national language, and the language of the curriculum. However, in Pakistan, the national government took control of language policy and impeded the development of Pashto inside Pakistan (Rahman, 2002). As it had never been used at an official level or for schooling, Pashto had remained a language of communication only. The majority of the participants from all regions felt that, at present, when everyone is moving toward English, selecting a regional language for the curriculum would be dangerous.
In addition to the regional languages, different perspectives were also presented on the role of Urdu, and Urdu as the language of the curriculum. After independence, Urdu was adopted as the single national language of Pakistan. Since then, the government has promoted it by stating that it is the language of the curriculum. Furthermore, it has become fashionable for people to adopt words from English into Urdu, not only in daily communication, but also in the curriculum. However, most participants felt that Urdu is unsuited to the curriculum, because while it was promoted at the official level, in practice its use was immature.
Research development, publications, and training of researchers in Pakistan marked another major influence on language development.

Language Practices in Teaching
Teaching practice in all provinces, including language learning, focused solely on reading and writing. Most participants see reading as a part of speaking, because reading means being able to read and speak. There were no separate classes for speaking, nor any particular speaking practices inside schools. Most teachers from public schools and some private schools believed that it is very difficult for students to change their language behavior inside the classroom, because classrooms and schools are the only places where most students practice speaking Urdu and English. Outside the classroom, students mostly use local languages with their classmates, friends, and families. English is used in the classroom only during learning and memorizing words. It was difficult for teachers to teach primary school students science and math in English because students could not even recognize the English alphabet. Moreover, although high school students have better English proficiency, the medium of communication and lectures was still Urdu.
The differences between private and public schools are the major factors that produce various perspectives and approaches to language practices in teaching. For participants in private schools (i.e., Punjab 1), language is strictly controlled in both the school and classroom, not only in terms of teaching and learning, but also in terms of practice. They stressed the role of languages in education, which are mainly Urdu and English. All teachers from private schools indicated that their schools require students and teachers to speak only Urdu or English. These schools also request parents to speak Urdu or English with their children at home because doing so affects their speech at school.
Our school has a very strict policy about speaking inside the school as well as at home. We ask our students to speak in Urdu or in English only. We do not allow students to speak other languages at school. Besides, we also ask parents to speak Urdu or English at home. Because it will affect their speaking at school. (Punjab 1) However, public schools do not have such rules regarding language use at school, nor ask parents to practice any particular language at home. Nor do teachers have any rules regarding speaking a specific language at school. Most teachers revealed that they often communicate with their colleagues and students in local languages or in Urdu.
The divide between urban and rural areas is the second factor affecting language practices. Participants explained that local languages remain the dominant form of communication among people in rural areas, and students from rural areas face extra problems during language interactions at school as well as in the curriculum. Participants described multiple influences, including parents' education, language practices for communication, and ethnic and cultural values. However, such practices are not limited to rural areas, as they also occur in urban societies. Students from low socioeconomic families practiced local languages with their families and communities. For a few participants, teaching in Urdu appeared easy, but it was difficult to make pupils speak Urdu at school.
Another perception of local languages shared by most participants was that language is associated with behavior. In the Punjab, students from a Punjabi language background mostly spoke Punjabi during violent confrontations. Punjab 3, 4, and 5 described the use of Punjabi language for emotional behavior. Similarly, other languages were also linked to students' behavior, with local languages also preferred for expressing emotions. KPK 3 shared an example of his individual behavior, saying that he had an impatient personality, a habit he had inherited from his father. He always relaxes by speaking or shouting in Pashto.

Language Practices in Higher Education
In this study, participants from schools mentioned the influence of higher education on the language practiced in educational settings and in society. If a student wanted to become a doctor, engineer, or other science-related profession, then he or she must learn English, because English is the only language of instruction for the medical sciences, engineering, and other scientific fields. Medicine and engineering were the two most preferred professions among students and were considered to be high-status professions. According to participants, there is intense competition among students to become doctors or engineers and this is the main reason why students need to learn English.
However, University 1 and University 2 had very similar views to those found in schools. Both participants were Assistant Professors: University 1 in Computer Science and University 2 in Environmental Science. During discussions, participants categorized linguistic practices at university into two types: verbal and written. It was easy to recognize the written language, which was English at both universities. However, the verbal language varied by level. Senior teaching and management staff preferred to use English for official meetings and speeches, whereas lower-ranked staff used Urdu or local languages. Teachers mostly used Urdu to communicate with each other, but teachers from the same language background communicated in the local language. This demonstrates that even in higher education, the application of a single language is unrealistic. In summarizing their reflections, University 1 said, Both university teachers suggested that English should be learned through school education. For them, learning English is the only possible language for education, particularly higher education, because the majority of research is in English. They supported the government's idea of implementing the English language curriculum from the first grade. For them, students would have better opportunities in higher education if they had learned through English at school. They linked international development to the country's development, whereby speaking the English language would be very important.

Language Practices in Religion
Participants in this study share parallel patterns of language practice in society. All participants recognized the association of language with religion, for which they were required to learn the Arabic language. They use Arabic every day but only for religious purposes, for example, offering prayers, reciting the Quran, during the call to prayer, and in reading Islamic texts. However, they mostly use Urdu to read Islamic literature, translated from Arabic or written by scholars in the Urdu language. In addition to using Arabic for prayer, Urdu is the language most used inside the Mosque or in religious ceremonies. In some regions, local languages are also used in religious ceremonies. Religious scholars mainly use Urdu to communicate during religious gatherings. However, there are still examples of local languages being used during religious ceremonies, where local religious leaders teach religious texts to local people. However, in all scenarios, all participants were only able to read Arabic. They had been taught the Arabic language when they were young, mainly through reading the Quran and a basic introduction to the Arabic alphabet. This explained the current pattern of learning Arabic; even if Arabic is a part of the curriculum, included in the subject Islamyat (Islamic studies), many parents still prefer to send their children to the mosque or madrassa to learn Arabic. Some parents hire private tutors to teach the Quran to their children at home. Nevertheless, Arabic language learning is mostly directed toward reading the Quran and performing religious practices.

Discussion
In this study, a discourse-ethnographic analysis of participants' language ideologies and language practices in Pakistan established a multifaceted interaction between language, power, and ideology. The language ideologies demonstrated the domination of English in policymaking institutions that have the power to align specific languages with rewards, with a variety of opinions on the issue from different linguistic participants. These ideologies are supported by a system of national and global forces that empower national and global languages over local languages. However, global forces appeared to dominate, primarily due to the elites that linked their supremacy to English language use (Butler, 2015;Phillipson, 1992;Rahman, 2002). This points to a system that cannot act without English to achieve global competency as well as maintaining elite positions inside the country. In this system, participants also accepted the importance of local languages in their particular societies, as these languages gave them a feeling of personal belonging. This suggests that linguistic groups are not simply passive receivers of language from global and national forces, but also resist the language ideologies imposed by such forces through learning, institutions, employment, or economic pressure.
Explicit language policy was the first arena in which different conflicts and practices of language ideologies played out, and which provides the answer to the first research question. Language policy is an activity that organizes language rules and vocabulary for the direction of writers and speakers in a nonstandardized language community (Bourdieu, 1991). It implies an attempt to guide the development of language in a direction desired by the organizers of that language. Language planning and policy have been described as the efforts of governments or administrative agencies to deal with language problems in the process of nation building (Fairclough, 1992;Phillipson, 2008). However, the participants from policymaking institutions in this study viewed themselves as in a complex situation where government sought English dominance in administration, and even ignored the law to convert the system from English into Urdu. Failure to implement the law reveals that laws alone do not work if they are not consistently enforced (Tyler, 2006). The government and administrative institutions have not applied the law, even after 2015 when the Supreme Court ordered government and other administrative institutions to enforce it. The participants were still in conflict with the current requirement to use Urdu in administration, which had never been used in this setting before. This was revealed, for example, when Policymaker 1's institution changed the nameplates from English to Urdu, whereas other institutions did not, because of their lack of familiarity with this language in administrative domains.
This study demonstrates that in Pakistan, language choice is in practice informed by various considerations (i.e., answering the second research question). Several participants closely linked language to their personal identities, whereas others saw varying options for future benefits. As an individual act, sociocultural, political, and economic conditions were prominent among participants' reasons for selecting a specific language (M. A. Ashraf & Tsegay, 2016). The double-edged nature of beliefs was apparent in contemporary ethnic discourses, which were embedded in specific historical and sociopolitical contexts (Fairclough, 1992). Various components influenced participants' language choice, but the strength of this influence varied from one participant to another, depending on their origin, social perceptions, local sociocultural setting, education, and social, political, and economic power. The concept of a divided society was fully illustrated in the context of Pakistan, where different ethnic groups make claims for self-determination, including a unique claim to a specific piece of territory under the criteria for nationalism (Hall, 1993). In this case, ethnic politics might help to address the group identities based on a mixture of shared family, region, religion, tradition, and language. This was an important point in identifying the ethno-political problems that have deep roots in divided societies. A divided society such as Pakistan is ethnically diverse and performs through politically significant cleavage, wherein benefits are controlled for political purposes. Therefore, participants from Sindh, KPK, and Balochistan view Punjab as the dominant force that controls the politics of power through language.
Participants described the different positions of language in the curriculum and teaching (i.e., answering the third research question). It was claimed that language had a specific role in society, as well as among political structures (Pennycook, 2017). The domination of English was linked with the colonial educational ideology wherein English was prioritized to secure better opportunities. Higher education was considered an important component in triggering the language ideologies that position English as the only language of the curriculum (Shamim, 2008). Participants see this as a way to produce a skilled workforce that can create economic development by attracting businesses to those areas where educated workers live (Mansoor, 2004). However, the teaching practices in all sectors of education were similar, transferring knowledge of English through national or local languages. Even higher education, which has used English as the medium of instruction from the outset, still experiences teaching practices in the national language. Few private schools, such as in the case of Punjab 1, implemented English-only or English/ Urdu-only language in school, indicating that only the elite have free access to privileged languages. However, this raises the argument as to whether local languages have been neglected; it seems that such languages are on their way to becoming languages of verbal communication only, as is already the case of the Punjabi language in the Punjab. Furthermore, religious ambitions require individuals to learn and practice the Arabic language, which is linked to religious identity. Language learning for religion is mainly aimed at being able to read the Arabic texts and to memorize scripts for performing religious duties. However, while this study excluded religious education, including language, it nevertheless found some evidence of religious language within collective educational settings.

Conclusion
This study demonstrated a contemporary understanding of the increasingly multifaceted configuration of language and power in education. Significant distinction and complexity were reflected in the language practices, ideologies, and power reported by participants. Hegemonic spaces in which dominant language ideologies were mechanically reproduced subverted the ideologies of local languages. The situation of languages in educational settings in each province was far more multifaceted than either description would suggest. The situation of languages in society as well as in educational settings can best be understood as a hybrid space in which hegemonic language practices and ideologies were both reproduced and challenged. In addition, supplementary hegemonic practices and ideologies were established without considering their impact on individual societies. It is crucial for people to learn their native language and the ethnic beliefs and wisdom embodied in that language (Fairclough, 1992;Rahman, 2001). This was one of many statements by the participants that views language and identity in hegemonic spaces, which participants considered essential and connected to the dynamic nature of identity. However, policymakers and teachers, especially from the Punjab, viewed national identity and success in hegemonic spaces when considering language selection, and as such neglected the Punjabi language, which was reserved for communication only. This supported the notion that both identity and language are flexible and hybrid, and not tightly connected to one another. However, this notion was rejected by other participants, particularly those from Sindh and at a moderate level by participants from KPK and Balochistan. There are complex connections between humans as social agents and the broader social forces that inhibit their subjectivity (Gellner, 1983). Language should not be thought of as a solid and impervious entity, and individuals did not consider the language they use as the message they want to get across at societal and institutional levels, such as the private school's control over language choice in Punjab.
The results of this study have some implications for policy development and educational settings. This study reveals the key role of language ideologies in the process of policy construction, policy implementation, and language policy inside schools and classrooms. These language ideologies are complex and diverse in nature, and vary among each linguistic group, that could certainly generate language conflict among different linguistic groups in multilingual society. Specifically, policymakers seeking the aggressive promotion of national and global languages should also consider the complexities of such actions on different ethnic and linguistic groups, and at least evaluate the reaction of different linguistic groups to such policies. In addition, teaching practices in school and classrooms should be measured carefully, before and after policymaking process, as teachers are important mediators of classroom-level language policy. However, there is need for more studies on language practices inside the different communities of Pakistan to understand how individual language practices struggle against national and international forces. For example, how does learning a language by reading alone, as is the case for Arabic, enable individuals to construct meaning without learning the meanings inherent in the language? National identity is linked to Urdu, whereas individual identity is linked to a local language, global identity to English, and religious identity to Arabic. This gives some indication of the complex learning environment for students that requires further detailed investigation to understand students' learning outcomes under the umbrella of four languages.