Unpeeling the Language Policy and Planning Onion in Rwanda

Language Policy and Planning is a complex endeavor that requires the intervention of different participants at different levels. In the onion metaphor which they use to represent language policy and planning, Ricento and Hornberger compare these participants and levels to onion layers. These levels together with the processes in which the policy is made permeate and interact with each other to varying degrees. They include legislation and political processes, states and supranational agencies, institutions and classroom practitioners. This desk-based research article discusses these â€˜layersâ€™ and the different roles which they play in language policy and planning in Rwanda, with a focus on the 2008 language-in-education policy. The article points to a unidirectional top-down approach, to lack of coordination in the way the layers work and interact and to a very limited role of classroom practitioners and practices in the language policy and planning process in Rwanda. These practitioners are conceptualized as mere implementers of policies decided on higher offices, which has negative effects on the effectiveness of the policies and their implementation.


Introduction
Language policy and planning is an important area for human societies because the language has always been a very useful tool and an identity resource for the humans (Ferris, Peck & Banda, 2013). Therefore, it is no wonder that several language scholars (for example, Baldauf, 2006;Haugen, 1966;Hornberger, 1994Hornberger, , 2006Lo Bianco, 2010;Ricento, 2006;Ricento & Hornberger, 1996) have devoted their studies to the field of language policy and planning (LPP), defining it in various ways. Baldauf (2006) defines language policy and planning as large scale and national planning, usually undertaken by governments with an intention to influence, if not change, ways of speaking or literacy practices within a society. Breaking down LPP into its two constituents, Lo Bianco suggests that a language problem leads to a language policy, which leads to language planning. In other words, LPP involves 'an identification of a language problem, the formulation of alternative ways of resolving the problem, deciding the norm to be promoted [language policy] and implementing it via the education system [language planning]' (Lo Bianco, 2010, p. 144). Lo Bianco's categorization concurs with that of Baldauf (2006) who indicates that policy is about selection (of which language(s) to teach, learn and/or use for what) and planning about implementation. It should be noted that while LPP concerns all areas of life, Lo Bianco (2010) notes that it has tended to see education as a field in which language policy is applied. This suggests that it is also the field which is most affected by language policy.
In the last twenty years, the field of LPP in Rwanda has been characterized by frequent and sudden language policy changes resulting from frustration with every policy option (Niyibizi, Makalela & Mwepu, 2015). These policy changes culminated in English becoming an official language alongside Kinyarwanda and French from 1995 and ultimately a solely medium of instruction at all levels of education from 2008 1 . While the top country management is the main language policymaking entity in Rwanda, different institutions and individuals played and still play a role in its implementation to varying extents and in different ways. This conceptual article investigates the role played by these with a focus on classroom professionals (or teachers) and discusses the implications of these roles. The focus on teachers is because, as argued by Ricento and Hornberger, whether they are consulted or not, these professionals are involved in shaping the process of LPP "whether consciously or unwittingly" (1996, p. 402).

Research Methods
This research is a desk research which involved reviewing the literature and research that has been conducted in the area of language policy and planning globally and in the Rwandan context. Language policy documents, academic papers, research reports, newspaper cuttings and views of policy makers in Rwanda constitute the main sources of data for this article.

Conceptual framework
This article is framed by Hornberger's (2006) proposed integrative framework for LPP and by Ricento and Hornberger's (1996) LPP onion metaphor.

Hornberger's integrative framework for LPP
Hornberger (2006) systematized LPP in an 'integrative framework' which identifies two language planning approaches and three language planning types. Holmberger (2006) represents the framework as follows: Sibomana, E. (2018). Unpeeling the language policy and planning onion in Rwanda  Hornberger identifies three language planning types which she presents on the vertical axis of the above figure. These are status planning, acquisition planning, and corpus planning. She explains status planning as "those efforts directed towards the allocation of functions of languages/literacies in a given speech community" (2006, p. 28). Myers-Scotton (2006) suggests that all language planning follows from status planning. Indeed, a language whose status is not known or unclear can hardly be planned for. Therefore, as Els (2005), in Baldauf (2006), suggests, status planning decisions should be based on community needs, irrespective of what their purposes are. However, this is not always the case in practice because, as Myers-Scotton indicates, language policies are not necessarily really planned in terms of being informed by contextual factors. Such a situation is likely to have a negative effect on the other LLP types and approaches and their success.
Hornberger defines acquisition planning as "efforts to influence the allocation of users of the distribution of language literacies, by means of creating or improving opportunity or incentive to learn them" (p. 28). Thus, as she explains, the goals of acquisition planning are identified in terms of the domains in which users are targeted to receive the opportunity or incentive. These domains are groups, education/school, literature, religion, mass media, and work. The management of these domains/groups usually avails the opportunities and incentives for their members to learn the target language. These opportunities, I suggest, may include education, jobs, and promotion (or demotion). Liddicoat and Baldauf (2008) (1977), it can entail raising the status level of a certain language or a change of its level of prestige by, for instance, changing the methods of teaching the language or requiring teachers to teach only in this language. In the words of Myers-Scotton (2006), acquisition planning describes policies regarding media of instruction and languages (to be) taught as school subjects. This may be the reason why Baldauf (2006) equates this type of planning with language-in-education planning. Cooper (1989) calls it a typical description of language teaching policies. As for corpus planning 2 , it deals with the language itself or what is done to the language in the process of its growth. Lo Bianco (2010) suggests that corpus planning involves processes such as proposing modifications to the internal resources of a language.

Language planning approaches
On a horizontal axis, Hornberger (2006) places two approaches to language planning: policy planning and cultivation planning. This scholar indicates that these approaches are form and function focused respectively. The policy-planning approach deals with matters of society and nation at the macroscopic level while the cultivation-planning approach deals with matters relating to language/literacy at the microscopic level and is mainly concerned with literary language. Elaborating further on these approaches, Hornberger (2006) explains that the policy approach, which is mainly concerned with standard language, is often interpreted to be the same as the status planning type, while the cultivation approach is often interpreted to be synonymous with corpus planning. However, Hornberger agrees that this match is not perfect. This match between language planning approaches and types may result from the fact that these approaches and types are cross-cutting and, according to Lo Bianco (2010), inseparable: an action of a status type can be taken and, depending on its goals, take a policy approach (when it is about the form) or a cultivation approach (when it is about the use).

Ricento and Hornberger's LPP onion
Ricento and Hornberger's (1996) LPP onion is made up of different layers. At the outer layers of the onion, they put "broad language policy objectives articulated in legislation or high court rulings at the national level" (p. 409). These scholars indicate that these objectives may be operationalized in regulations and guidelines which, in turn, are interpreted, implemented and eventually modified in institutional settings such as schools, businesses, and government offices. The interpretation, (un)implementation, modification of, and even resistance to, the guidelines depend on various factors including policymakers' interests, values and ideologies and whether the implementers share or not government's policies and vision.
Another onion layer or category which has a stake in LPP is States and Supranational Agencies. The State can make policies through Cabinet decisions or Ministerial Orders, as was the case in Rwanda before the creation of the Rwanda Academy of Language and Culture (RALC), or through a specialized Language Policy Body, such as the Pan African Language Board in South Africa (see http://www.pansalb.org.za/). The State has a lot of interest in education because, as Carnoy argues, education serves the State by fulfilling three functions … economic-reproductive (a process of qualification for work in the economy), ideological (the inculcation of attitudes and values), and repressive (the imposition of sanctions for not complying with the demands of school" (Carnoy, 1982, cited in Philipson, 1992: 68).
In other words, as Ricento and Hornberger (1996, p. 412) note, "education serves the sociopolitical and economic interests of the state so that the state can perpetuate and enhance its power." Such an approach to education and resultant reproduction of class structure is not only intranational but also international, which brings in supranational agencies. For instance, the United States and Great Britain, together have aggressively promoted the English language and Western culture in all parts of the world (Ricento & Hornberger, 1996) through education programmes in order to maintain and/or increase their power. Agencies such as the British Council and the United States Agency for International Development (USAID) are part of this drive. The politicoeconomic and linguistic power of English can be seen through the crucial role played by this language in international affairs (Bhatt, 2001;Altbach, 2004) and its displacement of hundreds of indigenous languages which had served as regional lingua francas (Ricento & Hornberger, 1996). Institutions, both public and private, constitute another layer of the LPP onion. Ricento and Hornberger (1996, p. 415) define institutions as "relatively permanent socially constituted systems by which and through which individuals and communities gain identity, transmit cultural values, and attend to primary social needs." Some examples of institutions mentioned by these scholars include schools, organized religion, the media, civic and other private and publicly subsidized organizations (e.g. libraries, musical organizations, magazine publishers), and the business community. In short, since language is virtually involved in all human activities (Ricento & Hornberger, 1996), all human institutions are involved in LPP in one way or another. These scholars indicate that institutions can work as policymakers, arbiters, watchdogs, opinion leaders, gatekeepers and reproducers or resisters of the existing social reality regarding LPP and, therefore, attempts to change language policies are hardly possible without the involvement of these institutions.
At the heart of language policy (the center of the onion) are the people who play a key role in the implementation of promulgated language policies: language classroom practitioners. A number of researchers (for example Calvo de Mora, 2012; Freeman, 1996; Throop, 2007; Li, 2010) view teachers as primary language policy makers or even "catalysts for policymaking" (Ricento & Hornberger, 1996, p. 418) in order for the policies to succeed. In spite of their assumed key role in policy making and implementation, Ricento and Hornberger (1996, p. 417) note, these practitioners are only seen as "an afterthought who implements what 'experts' in the government, the board of education, or central school administration have already decided." Li (2010) regards such teachers as passive or faithful adopters or conduits of pre-existing policies and plans which others have conceptualized and articulated. Interestingly, Ricento and Hornberger (1996) indicate that teachers are assigned the same role even in decentralized systems while it is widely recognized that education and social change in decentralized societies often begin at the grassroots. This failure to involve teachers in policy making may lead to difficulties in (and possibly failure of) policy implementation. Alternatively, it may turn the teachers into what Ricento and Hornberger (1996) call unwitting producers of social reality, especially when they 'implement' the policies uncritically. Arguing against such treatment of central onion layers by the outer ones, Ricento and Hornberger (1996) content that if social change is the main goal of language policy, teachers should be given a bigger say in the making of the policies such as those related to language because they affect them and their work considerably.

Results and Analysis
This section is a discussion of the roles which different layers play in different language policy types and approaches in Rwanda, and related implications for the acquisition and mastery of the languages used in Rwanda.

The role of onion layers in LPP in Rwanda
Before discussing the roles of the layers in LPP in Rwanda, it should be noted that the way language policy and planning in Rwanda is made has had a broad scope, referring to what Gafaranga, Niyomugabo and Uwizeyimana (2013) call as macro policy. For example, the Constitution of the Republic of Rwanda indicates that Kinyarwanda, French, and English are official languages and the language-in-education policy indicates that Kinyarwanda is to be used as a medium of instruction from pre-primary to Grade 3 and English from Grade 4. However, none of these documents provides further guidelines on how these languages should function in different areas in the Rwandan community. This situation leads to the state of unexpectedness in the use of these languages in different settings. One example is a high school learners' debate competition which I recently attended in one province. While the debates were in English and the audience understood this language, the officials who officially opened and closed the function spoke in Kinyarwanda in spite of them being proficient in English. Thus it can be argued that the policy does not provide guidelines or information on how the different LPP types and approaches are to be carried out. This may be one reason why acquisition planning endeavors in Rwanda were generally uncoordinated and ephemeral as will be elaborated on later. However, this does not mean that the constitution falls short of its role because, as Lo Bianco (2010) argues, the standing and public use of language is not necessarily determined by the constitution but by other procedures. The problem is that these 'other procedures' are not generally available in Rwanda.

IJSSH
Vol. 2 No. 2, August 2018, pages: 99~114 a) Status/policy planning Like in some other countries such as South Africa, Rwanda also has an institution which is in charge of language and planning: Rwanda Academy of Language and Culture (RALC). However, while part of its mission is to sustain the different languages used in Rwanda 3 , the role of this institution in the LPP has been rather unclear so far. So far, language policy and planning at the macroscopic level is regulated by the constitution and the Cabinet decision statements. This situation validates Lo Bianco's (2010) remark that language statuses are typically ascribed via public texts, such as constitutional provisions of sovereign nations.
Article 8 of the Constitution of the Republic of Rwanda of 2003 revised in 2015 stipulates that "the national language is Kinyarwanda. The official languages are Kinyarwanda, English and French" while the 08 October 2008 Cabinet's decision reads as follows: in order to increase the role of Rwanda in the East African Community (EAC) in particular, and in international affairs generally, the Cabinet asked the Minister of Education to put in place a quick programme of using English as a medium of instruction in primary, secondary and tertiary institutions which belong to the government and those which it subsidizes (GoR, 2008).
This Cabinet's decision is supplemented by guidelines from Ministry of Education. Thus the constitution, the Cabinet and the Ministry of Education constitute the outer layers of the onion: they make language status and policy planning decision at the national level. With reference to language policy at the microscopic level, it is regulated by micro policies of different institutions (schools, universities, churches, etc.) which, as has been pointed out, depending on how the managers of these institutions interpret and respond to the macro policy.

b) Cultivation planning
Cultivation planning which, according to Baldauf (2006), lays an emphasis on the functional extension of language development and use, was affected by the official status of English and its recently acquired status of the only medium of instruction from Grade 4. Different institutions have made several and varying decisions and actions at the microscopic level regarding cultivation planning, especially since English was made the only medium of instruction from pre-primary in 2008 and subsequently from Grade 4 in 2011. The diverse nature of these decisions suggests that the new language-in-education policy was interpreted severally by different layers, which validates Ricento and Holnberger's (1996) indication that the legislation, judicial decree, or policy guideline is interpreted and modified as it moves from one layer to the next. This situation also gives credit to Spolsk's (2007) argument that all institutions have their own (macro) language policies which may sometimes contradict or differ from macro policies. For instance, while the Cabinet decision did not say anything about French, which had been a medium of instruction and one of the main school subjects at primary and secondary levels for a long time, this language was made an elective school subject. Moreover, schools which had been using French as the main medium of daily communication switched to English. Also, while the decision recommended the use of English as a medium of instruction in public and subsidized institutions (only), even private ones followed suit.
This situation described in the above paragraph led to an increased use of English in various areas including schools, media, and religious organizations, among others. For instance, English has become the second most used language in different areas after Kinyarwanda (National Institute of Statistics of Rwanda 2014), some churches have introduced English services, others use English interpreters when the preaching is in Kinyarwanda, many religious choirs which sing in Kinyarwanda add English captions in their songs video clips and, according to Kwibuka (2013), most advertising billboards and posters in the most frequented places of Kigali are now in English. Institutions such as banks, migration offices, hospitals, travel agencies and other institutions which ask their customers to fill in various forms in English also take part in cultivation/acquisition planning by teaching English implicitly. Furthermore, these institutions send messages to Rwandan citizens and residents that if they need to have (better) access to their services, they need to be able to use English. Another message could be that English is a key of many social doors c) Acquisition planning Moving towards the center of the onion, the Cabinet decision was welcomed by different supranational agencies. The main of these is the British Council whose mission is, among others, to create cultural relations between the UK and the wider world 4 . As it appears on the Kigali British Council website (https://www.britishcouncil.rw/), this agency is involved in acquisition planning by spreading English in Rwanda. The Agency's office in Kigali helps those who want to sit for IELTS and TOEFL exams through its Kigali English Language Centre.
It also helps in improving the language level of teachers, creates self-study materials that allow teachers to incorporate their study into their work and family schedule and supports the Rwanda Education Board's (REB)'s school-based mentor programme 5 to help teachers to use English as a language of instruction. The British Council also offers self-access learning opportunities: (i) learn English SMS service, (ii) learn English by radio and (iii) learn English through newspapers. It has an English for Education System (EES) programme aimed at carrying out research in English teaching and learning. All these initiatives contribute to the aforementioned power of English in the current local, regional and global affairs. As has been pointed earlier, the overt aim of the British Council may be to promote English, but the covert one is undoubtedly to increase the influence of United Kingdom and spread its culture (British Council, 2012) Like in other developing countries where English is gaining more ground, there is a concern that the drive for the acquisition of proficiency in this language may overshadow the need for, and the value of, the national language, Kinyarwanda. Indeed, some parents indicate that they are ready to sacrifice Kinyarwanda, provided that their children get access to proficiency in English .
The decision to make English the only medium of instruction in Rwanda gave rise to the need for proficiency in this language as never before. Thus many institutions got involved in intensive acquisition planning activities by becoming hubs for the teaching of English to different categories of people, both young and adult. Many informal centres were also created to teach English, in both rural and urban areas (Kwibuka, 2003), local English language schools got filled up with students who included taxi drivers (McCrummen, 2008) and some schools hired private English coaches for their teachers in order to speed up their acquisition of proficiency in English. English teaching endeavors were joined by Tigo, a telecommunication company in Rwanda, which introduced a cell phone based English teaching programme 6 . These facts show that, as Baldauf (2006) suggests, the language-in-education policy does not just occur in schools; it also implicates less systematic teaching situations in the community or the workplace.
The classroom practitioners in Rwanda, who are supposedly at the heart of language policy, are not different from those in other parts of the globe regarding their actual role in LPP. While a number of linguistic scholars (for example, Calvo de Mora, 2012; Freeman, 1996; Li, 2010; Lo Bianco, 2010; Throop, 2007) consider teachers as primary language and curriculum policymakers, Pearson (2014) notes that the teachers in Rwanda were conceptualized as mere implementers of the language in education policy: they were instructed to implement a policy on which, apparently, they had little (if any) to say because the decision had been made from a very high level. Thus, the language-in-education policy making adopted a top-down approach (from the outer layers) with little evidence (if any) that it was informed by classroom contexts and practices (Pearson, 2013; Rwanda Ministry of Education, 2015). The slight change in the 2008 language-ineducation policy that occurred in 2011 is an example of the importance of classroom contexts in policymaking. After realizing that teaching in the medium of English was not productive in pre-and lower primary education, the Ministry of Education decided to use Kinyarwanda at these levels (Nyiraneza, 2011).
This failure to involve teachers in educational policymaking is common in Rwanda because, as VSO Rwanda (2004) indicates, Rwandan teachers are seen but not heard. This situation may be the main source of challenges in education policy implementation and one reason why, as REB (2014)  French to English as a medium of instruction has been one of the most challenging experiences Rwanda's education system has faced. In fact, research shows that most teachers and learners at all levels of education in Rwanda were not ready to use English to teach and learn other subjects (Pearson, 2013, Samuelson & Freedman, 2010; USAID, 2012; Uwambayinema, 2013) and the training which was offered to teachers was perceived by the very teachers as insufficient and inadequate (Pearson, 2013). Thus, it can be argued that if the teachers had been consulted before making English the main medium of instruction, the policy and/or its course of implementation would have been different. Shohamy (2006) argues that educational policies which do not involve classroom practitioners usually fail to examine whether they are feasible and are "imposed … for political and social reasons, without attention being paid to the needs and wishes of those affected by the policy and those expected to carry it out" (2006, p. 143). This may be one of the reasons why instead of working as reproducers of the policy, some teachers work as resisters of such policies. For example, some teachers who participated in Li's study of the role of teachers in policymaking in China explained the reasons for their resistance as follows: When we find that the new curriculum is not practical in classroom teaching, but there is no way to let policymakers know about our opinion, we just ignore the instructions in it. We just follow those we think suitable. The methodologies suggested in the curriculum sound good, but we use our own way according to the students' needs and their individual differences" (Li, 2010, p. 444).
It is unfortunate that no provisions have been made for these teachers' classroom practices to reach central policymakers and thereby influence policymaking. With reference to Rwanda, some of the teachers who participated in Pearson's (2013, p. 6) study on the implementation of the 2008 language-in-education policy indicated that the policy was a mandatory, "a kind of a law" to be followed and not questioned. This approach contributes to teachers' understanding of policy as a synonym of decisions made by leaders at higher levels of which teachers are mere implementers (Li, 2010). Indeed, some people with authority think that language policy making is a matter of getting financial resources and training and instructing teachers to implement policies. For example, Lo Bianco quotes a senior US Defense official who responded to the educators' suggestion to do research before rushing for policy making as follows: "[We] can do it NOW! We just need to say what language we need, get the money, train the teachers, and they do it. It isn't rocket science!" (2008, p. 172, capitals in the original). Thus it appears that challenges were inevitable in the implementation of the English medium of instruction policy in Rwanda mainly because teachers and learners had not been consulted and/or prepared in advance. In addition, they were expected to use a language in which they had limited proficiency and no appropriate teaching/learning resources were available (McGreal, 2009;REB, 2014). In order to address these challenges, some Rwandan teachers resorted to translating the content and resources such as textbooks which they had from French into English with the help of more knowledgeable colleagues and other bilinguals (Pearson, 2013), others registered for evening courses for English, while some schools hired private English coaches as has been noted. These strategies worked to various with limited levels of success.

Discussion
Most of LPP activities in Rwanda are centered on acquisition planning, which validates Baldauf's (2006) argument that acquisition planning often constitutes the sole language planning activity in many polities. The proliferation of language acquisition activities may have resulted from the discourse which was built around English by government officials and other people with various kinds of authority (such as journalists and academics) since the Government decided to make it the medium of instruction. This discourse portrayed English as a passport to social, economic and academic power and implied that one is severely disadvantaged if they cannot use English . Therefore, the people who wanted to "maximize their opportunities for upward mobility" (Myers-Scotton, 2006, p. 372) decided to study this language. Such a discourse refers to what Baldauf (2006) calls informal or covert realization of language policy.
In spite of the above-mentioned discourse, the policymakers did not advise any a specific and research-based plan of how the different types and approaches of LPP should be carried out in order to facilitate the implementation of the new language-in-education policy. This may be one of the reasons behind the lack of coordination in, and elaborate approaches to, English teaching activities. Everyone did what they thought was effective, which led to some decisions which appear to have a negative impact on education. For instance, some educational officials and teachers treated multilingualism as a problem to achieving proficiency in English and, as a result, banned the use of other languages (including Kinyarwanda) in classrooms and on school premises (Gakwaya, 2014; Uwishyaka, 2015) in the mistaken belief that this will speed up the access to proficiency in English. This belief and resultant practices made some people take language policy to mean the promotion of English over other languages (see Lo Bianco, 2010), which refers to Calvet's (1998, p. 203) argument that "language policy is a civil war of languages." The fact that most of the educational policies in Rwanda are not research-based (Rwanda Ministry of Education, 2015) is likely to lead to the futility of efforts and resources mobilized to implement these. Indeed, people may be taking a route which will not necessarily take them where they want to go as they may have developed policies "without any recourse to empirical findings or advice" (Van Els, 2005 in Baldauf, 2006, p. 3). For example, English was made a medium of instruction "in order to offer learners access to proficiency in English" (Sibomana, 2015, p. 130) without any research-evidence showing that the same approach has yielded very limited levels of success in other contexts where English is a foreign language. Indeed, some years after the policy implementation started, teachers' and learners' proficiency in this language is still limited (Osae, 2015) and the policy implementation still a challenge (Rwanda Education Board, 2014).
Another example of a route which may not lead to the intended destination is banning the use of Kinyarwanda and French in some schools in order to speed up the access to proficiency in English. Such a decision has serious negative effects on education, on learning and teaching and on learners' and teachers' personal identity Sibomana & Uwambayinema, 2016). This is because, in the process of making sense of the world, learners and teachers use all the linguistic resources at their disposal to make sense of their learning (Boakye & Mbirimi, 2015;Canagarajah, 2011;Canagarajah & Wurr, 2011;Garcia, 2009;Makalela, 2015). Thus, preventing them from using some of these resources is limiting the exploitation of their learning and teaching potential, paving a way for difficulties in, and/or resistance to, using some teaching approaches which have been found to be appropriate for multilingual contexts. These approaches, one of which is translanguaging, recognize, value and capitalize on all learners' linguistic resources and languages and treat bi and multilingualism as an advantage and not as a problem (Canagarajah, 2011;Canagarajah & Wurr, 2011). Therefore, the decision may have serious implications for the quality of education as well.
It has been noted previously that the success of LLP in any context rests on the involvement of language teaching professionals and other teachers who, Ricento and Holnberger (1996) argue, are at the heart of LPP. As has been indicated previously, the LPP process in Rwanda took a top-down approach and, apparently, very limited consultation (if any) was undertaken by outer layers with other layers before the policy was made. This means that the permeation of the layers was also 'uni-directional' (Lo Bianco, 2010) and there was limited (if any) possibility for the central or inner layers to permeate and influence outer layers' decisions. Thus, policymakers might have overlooked or, at least, downplayed the realities at the onion center (the classroom), which may be one reason why, as Niyibizi, Makalela, and Mwepu (2015) note, language policy changes in Rwanda have had negative cognitive and pedagogical effects on both learners and teachers. It should be noted that in spite of the conceptualization of teachers as passive recipients of language policies, teachers may not be passive in practice: they play an instrumental role in classroom language policy (re)creation" and, therefore, they "are inevitably engaged in acts of language planning and policy each day" working within the constraints of, and possibly questioning, the existing policy (Throop, 2007, p. 45). For instance, after the 2008 language-ineducation policy, some teachers in Rwanda continued to use Kinyarwanda, because "the children don't understand enough" while others used both English and French because they themselves were not proficient in English (Pearson, 2013, p. 12). As Freeman notes, teachers have considerable autonomy in their implementation of high-level decisions, which leaves room for significant variation in the way they put the plan into practice on the classroom level ... considering teachers and administrators as planners allow an understanding of how practitioners potentially shape the language plan from the bottom up (1996, p. 560).
Thus whether (central) policymakers like or not, teachers' practices will impact on policy implementation at least locally. While they may not have considerable knowledge of LPP matters, their experiences, reflection, and  Baldauf (2003), in Baldauf (2006, p. 154), argue, "the impact of language planning and policy depends heavily on meso and micro level involvement and support." Throop also argues that effective LPP often depends on "an intimate knowledge of the context" (2007, p. 49) and several educational researchers (for example, Arif, 2002;Li, 2010;Throop, 2007) suggest that teachers are best positioned to access such knowledge. With such knowledge, teachers can construct, (re)construct resist policies to meet their students' needs effectively in their local contexts (Li, 2010; Myers-Scotton, 2006; Throop, 2007) and teachers in Li's (2010) and Pearson's (2013) studies mentioned previously are good examples. Thus local practices and their impact need to be given due consideration in language policy-making processes or "the inconsistency between theory and practice" (Li, 2010, p. 448) will keep on increasing.

Conclusion
All the onion layers play irreplaceable roles each in making the onion what it is. Therefore, all these layers should be involved in all LPP types and approaches as much as possible, for its relevance and successful implementation. In order for this to be effectively done, the LPP process should not just be a one-way (topdown) approach but also and especially a bottom-up approach by giving a central place to teachers. This, Cochran-Smith and Fries (2002) argue, will help in better addressing learners' needs and, as Li (2010) notes, make the policies more implementable. Therefore, policy-making initiatives should be based on grassroots level experience or, in other words, on the center of the onion where the key policymakers (teachers) are placed, conceptualizing them as actors rather than passive recipients of instructions from above. Furthermore, the different layers need to be educated about their role in LPP processes and equipped with relevant knowledge and skills which they need in order to play these roles effectively. Particularly, language teacher education programmes should include courses about LPP in order to help teachers to understand how their own practices create and recreate language policies. There is also a need for more systematic and regular coordination and monitoring of LPP activities by policymakers to make sure that what is happening in classes is research-based, context-relevant and in line with articulated policies, which will help in 'proposing realistic remedies' (Ricento, 2006) to language problems and preventing a waste of energy in language policy implementation.

Conflict of interest statement and funding sources
The author declared that he has no competing interest. The study was financed by the author himself.

Statement of authorship
The author has a responsibility for the conception and design of the study. The author has approved the final article.

Biography of Author
Dr. Emmanuel Sibomana holds a Ph.D. in Applied English Language Studies from the University of the Witwatersrand, South Africa. He has worked as a high school teacher in Rwanda and a lecturer of applied linguistics at the University of Rwanda and the University of the Witwatersrand in South Africa. Dr. Sibomana has presented papers at local and international conferences and published extensively in the areas of academic literacy, language education, language policy and language teacher education. He is now working as the Director of Policy and Programs at the Wellspring Foundation for Education of which mission is to be a catalyst for the transformation of education in Africa and to foster vibrant communities which address poverty in all its forms. Email: esibomana1@gmail.com