Amid signs of change: language policy, ideology and power in the linguistic landscape of urban Rwanda

In this paper we explore the nexus of language policy, ideology and power in the linguistic landscape of urban Rwanda. In post-genocide Rwanda, English has been promoted and gained status. This has led to an increased usage of English on shop signs in the streets of Kigali and other towns in Rwanda at the expense of both French and Kinyarwanda. Unique quantitative language data documented in streets before 2008 are in this study compared to data collected in 2018, in the same streets. This forms the background for analysis of official discourse, targeting language policy changes, especially after the 2008 decision to appoint English as the language of administration as well as the medium of instruction throughout the educational system from grade 1 on. This decision was made despite the fact that Rwanda has a national language, Kinyarwanda, known by 99.4 per cent of the population. The analysis shows that political aspects of language policy decisions are downplayed. Officially, both discursively and in practice, the Rwandan government, that is the English speaking elite in power, legitimize their decisions by pretending that imposing English is an inevitable, pragmatic and rational measure for economic development. This narrative reveals ideas about inherent qualities of specific languages while simultaneously discarding others. Additionally, in public discourse all four official languages are equal. Reality is different. In a society where the status language is only acquired through education and used as medium of instruction, power differences and socioeconomic inequalities are neglected and obscured.


Introduction
When returning to the Rwandan capital Kigali after a decade, a noticeably transformed urban environment with new skyscrapers, and a new linguistic landscape appeared. Even the street signs had changed. Gone were the former French-Kinyarwanda signs. Instead, new street signs using a new language, English, and a totally different system were found on all street corners. In a similar way, even language practices on private shop signs had changed, with an increased use of English. This does not just happen. It forms part of political decisions and strategies, and we should approach the changing linguistic landscape as an expression of post-genocide politics.
It is frequently assumed that an increased use of English is inevitable, due to globalisation and the myth of English as the international language par excellence, and that English is used extensively in public space to add an international air. However, in Rwanda, we see a different situation. English was added and introduced as co-official language after the genocide in 1994 and was appointed the sole medium of instruction in 2008, although English has no colonial or historical background in Rwanda. Furthermore, English is not needed as a lingua franca, as Rwanda's national language, Kinyarwanda, is spoken by almost everyone and used in all domains of life. Thus, the changed status and use of English must be investigated against historical developments and narratives voiced by official representatives. These ideologies are imprinted in practices in the socially constructed linguistic landscape of urban Rwanda. Building on Shohamy's (2006) three-dimensional understanding of language policy, unique empirical data comparing language use on private shop signs collected in 2008 are in our study compared with data from the same streets in 2018, that is after the 2008 change of language policy. These unique diachronic data illustrate clearly that public space is never neutral. Linguistic public space is always somebody's space, imbued with power exercised by somebody. This space is socially constructed by top down decisions and emplaced signs, even signs on private shops, which construct narratives about languages and language policy. Therefore, it is of vital interest to analyse how leading politicians and official representatives through public discourse reveal official ideologies behind developments over the past decades. To further identify factors influencing discursive ideas and norms reflected in language practices, and to investigate how changes are motivated and legitimatised in official public discourse, we use Thompsons's (1990) modes of operation of ideology used to exercise power.
In our analysis we see signage in public space as a part of the linguistic landscape. In the following we give a short overview of our theoretical points of departure. This is followed by a summary of Rwanda's history, specifically focusing on the post-genocide period and language policy changes which have taken place. This background is needed for the analysis, which understands the 1 3 Amid signs of change: language policy, ideology and power in… linguistic landscape as a nexus where situated norms, the embodied history and circulating beliefs and ideologies together construct a nexus of practice. The aim of the study is to unveil how language use in the linguistic landscape reverberate with the policy discourses circulating in media and thus intersects with official language policy, as a de facto language policy mechanism that translates ideology into practice.

Theoretical framing, epistemological perspectives and background
The interdisciplinary field of linguistic landscape (LL) studies adds to our understanding of language, society and people. As stated by Shohamy and Gorter (2009a, p. 2), texts displayed in the linguistic landscape " […] shape the ecology in local, global and transnational contexts". The linguistic landscape is traditionally defined, referring to Landry and Bourhis (1997, p. 25), as "the language of public road signs, advertising billboards, street names, place names, commercial shop signs, and public signs on government building […] of a given territory, region, or urban agglomeration". As earlier demonstrated by among others Lanza and Woldemariam (2009) in their study of urban Ethiopia, the display of certain languages and the lack of others provide a clear ideological message as to the value, relevance and priority of the languages" (Shohamy & Gorter, 2009a, p. 7). Their study stresses the link between official policy and practices.
Starting in the 1990s, linguistic landscape studies have provided insight into sociolinguistic topics such as language attitudes and language vitality (Landry & Bourhis, 1997;Spolsky & Cooper, 1991). Lately, the field has contributed to our understanding of contextualizing public space and issues such as identity, language policy, and political and social conflicts (Shohamy & Gorter, 2009a). Studies within the field of linguistic landscape have clearly demonstrated that language practices in this space do not replicate speech by individuals, but instead reveal the symbolic construction of public space (Ben-Rafael et al., 2006Gorter, 2013;Shohamy & Gorter, 2009a). Studying the visual linguistic order and patterns of language choice in the linguistic landscape enables us to infer such symbolic values (Hult, 2018).
The present study with its mixed methods approach targets this symbolic construction of public space in a changing Rwanda. By quantitatively analysing languages visible on shop signs in urban Rwanda, we identify changed practices. However, it is not the numbers as such which are important, but what these quantitative data tell us about political and economic decisions and ideologies underlying what is quantitatively revealed by the data. Discourses by politicians and official representatives of the Rwandan state not only demonstrate major changes regarding official language policy that have taken place in Rwanda in recent years, including a change of the medium of instruction (MoI) from Kinyarwanda/French to English, but more importantly reveal ideologies and political agendas. In fact, language policy "should not be limited to the examination of declared and official statements. Rather, the real policy is executed through a variety of mechanisms that determine de facto practices" (Shohamy, 2006, p. 54). Discourses by official representatives function as powerful and covert means by which groups with status and power manipulate "people and their behaviors" (ibid., p. 22). The multilingual ecology reflected in the LL is in this way a result of social actions mediated by circulating discourses, formed by sociolinguistic conditions, history and political decisions, as pointed out by Hult (2009, p. 89). Furthermore, it stresses that ideology mediated by discourses is directional; it "always works to favour some and to disadvantage others" (Purvis and Hunt 1993, 478).
Specific language items, displayed in specific languages in a specific manner, "sends direct and indirect messages with regard to the centrality versus the marginality of certain languages in society" and transmits symbolic messages as to the legitimacy, relevance, priority and standards of language and the people and groups they represent" (Shohamy, 2006, p. 110). Thus, the value of specific languages is entextualised in the LL (Hult, 2015) and shows dominant discourses that reflect power relations, and both past and more recent political developments. Therefore, the linguistic landscape must be seen as a nexus of actions (Hult, 2015;Scollon & Wong Scollon, 2004), which offers an arena for analysing socio-political and ideological ideas about language (Hult, 2014;Kasanga, 2012;Lanza & Woldemariam, 2009) as well as "a window into ideologies and power" (Hult, 2018, p. 11).
Ideologies may be described as a system of ideas, which "construct meaning in the service of power" (Thompson, 1990, p. 23). Language in public space not only reflects language policy but also planning measures (Spolsky, 2004), ideological modes and associated strategies (Thompson, 1990). Thus, language policy and planning (LPP) is "one mechanism by which dominant groups establish hegemony in language use " (Tollefson, 1991, p. 16). In the era of late capitalism, neoliberalism and globalisation we need to investigate whose interests LPP serves and in the name of what (Heller, 2018, p. 3). This nexus of different factors which constitute LPP, links what goes on in the streets, on private signs, that is the quantitative representations of language in the linguistic landscape, to political decisions, ideas and ideological discourses behind language policy and how these have been and still are affecting practices in the LL. This interaction order, the embodied life experiences and discourses in place (Hult, 2015;Scollon & Wong Scollon, 2004) together form the actions we see in the linguistic landscape.
In Africa, exogenous languages earlier imposed by colonial powers have been maintained by post-independence regimes (Makalela, 2016;Rassool, 2007), through colonial habitus and cultural reproduction (Bourdieu, 1991), resulting in "linguistic choices that reinforce existing social, political and economic inequalities; and, in doing so, they collude in their own collective disempowerment and/or dispossession" (Rassool, 2007, p. 2). The imposition of French and English in Africa has not only been passive through cultural reproduction; it has also been steered through the creation of British and French councils as well as American initiatives in Africa, among other through teaching aid such as the Fulbright Teaching Assistant Program (United States, n.d.). In addition, influential organizations such as La Francophonie, promoting the spread of both the French language and French culture, and The Commonwealth, with their links to former colonial powers, have maintained the status of ex-colonial languages in Africa (Makalela, 2016;Phillipson, 1992).

3
Amid signs of change: language policy, ideology and power in… The preservation of ex-colonial languages as official languages is explained by Janks et al. (2014) as power in the hand of groups which exercise their power using modes through which ideologies can operate. Thompson (1990) identifies five main modes which may overlap: legitimation, dissimulation, unification, fragmentation, and reification. These modes can additionally be linked to different strategies of symbolic construction. Legitimation implies an appeal for support of ideas, through rationalization, universalization and narrativisation. Dissimulation, implies that the power relation is denied or obscured. Thompson exemplifies dissimulation as the use of figurative language to hide the unpleasant symbol of domination. Other dissimulation strategies are displacement and euphemization. Unification and fragmentation are contradictory modes: While unification is used to unite people, which creates a national identity, an 'us', fragmentation emphasizes differences in order to divide and rule. The fifth and last of Thompson's modes is reification-a social and historical creation is presented as a natural and inevitable outcome.
These modes have frequently been employed in Africa, inter alia to make English a language of power (Janks et al., 2014;Phillipson, 2006). Batibo (2020) argues that post-independence nations have formed their language policies to accommodate inherited legacies from colonial powers, be it the results of the divide and rule strategy (British colonial rule), or assimilation by adopting the colonizers' culture, way of life, language and identity (French, Belgian and Portuguese colonial systems). As a consequence, most post-independence governments in Africa, in an attempt to restore and forge national unity and national identity, have maintained ex-colonial languages. In doing so, they have reproduced a colonial ideology of oneness; one nation-one language (Makalela, 2016), although most Africans are multilingual.
In Rwanda we see a rather different language situation compared to most African states that do not possess a single endogenous African language, known by the majority of the population and thus with possibilities to be used as medium of communication and instruction. As stated in the introduction, Rwanda has a national language Kinyarwanda, spoken by 99.4% of all Rwandans (Rwanda, 2005). In spite of this favourable situation, changes of power have historically led to changes in language policy by the introduction of European languages.
By using mixed methods, combining content analyses and quantitative data collection, we trace and analyse ideologies behind recent language policy decisions. We also investigate to what extent official discourse in place reverberate with the interaction order of the Rwandan linguistic landscape. The study was conducted in two steps. First, quantitative data were collected and compared to an earlier study to identify changes. Secondly, discourses by official representatives were identified and analysed to investigate motives behind these changes.
The data for the quantitative part were collected in 2018 in the same towns and streets as in an earlier study conducted between 2005 and 2008. These streets were all renamed in 2011-2012, as well as the towns where the streets are located. The data include the capital Kigali and Butare in the South, now renamed Huye, and Gisenyi, now named Rubavu, close to Lake Kivu in the North West by the Dem- St), Rue du Travail (now KN 2 St), Boulevard de Nyabugogo (now KN 1 Rd) and Avenue de la Justice (now KN 2 Ave). KN is the abbreviation of the district Kigali Nyarugenge. A quantitative full inventory of all shop signs in these streets in central Kigali was made. Additionally, the shop signs in the main streets in Huye and Rubavu were included. The main sign of each business or shop was categorised according to language, following the same methodology, with 16 categories, as used for the 2005-2008 data, given by Rosendal (2011, pp. 67-68,219-220). These categories grouped combinations of languages, based on amount and prominence of the language(s) used, which for multilingual signs was done on the basis of the amount of text in each language on the sign, and included the order and placement of text on the sign, and visual prominence, among other through spatial placement, text size and colour. In the Kigali streets, today 447 signs were found in the 5 streets versus earlier 792, due to recent physical urban restructuring and transformations. As the comparison between the data is based on the percentage of the total distribution of the different sign categories, these changes do not affect the diachronic comparison. The number of signs found in Huye /Butare and in Rubavu/Gisenyi slightly increased between 2006-2008 and 2018 (from 122 to136 and from 113 to 128 respectively).
As Rwandan newspapers align with official policy, both national and international media were selected to prevent a bias, and to ensure a broad representation of the ideological spectrum. Qualitative data from the period 2006 to 2020 were searched digitally in local and international media archives, mainly newspapers archives, and resulted in 4 150 news articles. Of these, duplicating citations, that is the same citation found in different newspapers, were taken out. The remaining citations were categorised into main themes, and reduced into 20 key newspaper articles (14 from international media and 6 from national/regional media), which were considered representative of official discourses preceding the major change in official education policy in 2008 and how official representatives explain and motivate the increased use of English in Rwanda after 2008. Although the framing and angle of the news articles were different in non-Rwandan media, the inclusion of international media did not result in differences in the official discourse.
The unique language policy development in Rwanda, with its specific historical and political background, is a necessary backdrop for an analysis of recent language policy changes, and is therefore presented briefly in the following.

Language policy and post-genocide developments
Until the first colonial administration in Rwanda, Kinyarwanda was the only recognised language and the language of instruction of the Rwandan Kingdom, for both young men attending Itorero, and young women attending Ibohero traditional schools (Muzungu, 2003;Niyibizi, 2014). Swahili and later French were introduced to serve the political interests of the colonial administrations when they entered the scene. When the Germans set up their colonial administration in 1898, they introduced Swahili, the administrative language of German East Africa (Deutsch-Ostafrika) in Rwanda (Ntakirutimana, 2012). Swahili was imposed despite the existence of a national language Kinyarwanda, to facilitate the colonial administration and to have one common language of administration in the colonised region East Africa, a "one language, one nation" ideology (Makalela, 2016;Ricento, 2000). In 1917, the Belgian army, assisted by the British, overthrew the German colonial administration. Rwanda was placed under Belgian colonial administration under a League of Nations mandate. French replaced Swahili which became the medium of instruction, administrative language and a symbol of the change of power (Rosendal, 2011).
When Rwanda gained independence from Belgium in 1962, this independence was in many ways symbolic as Belgium continued to exercise its power over Rwanda by providing aid, military specialists, technicians and other experts to train the new 'independent' Rwandan government (Mugesera, 2015;Semujanga & Galabert, 2013). French was maintained as the post-colonial official language and medium of instruction. In 1973, a coup by Major General Habyarimana, the Minister of Defence, ended president Kayibanda's rule. However, neo-colonialism continued, as France became a dominating power. This influenced both the political and the economic make-up of Rwanda (Durmaz, 2019;Mugisha, 2010). Since French was maintained as the medium of instruction, it may appear as if even the language policy was the same. However, Belgian French was discouraged and the French standard was promoted through textbooks written under French supervision and donations to Rwanda, combined with a corrective discourse in schools (Kajeguhakwa, 2001;Rurangirwa, 2010). Apart from a period in the late 1970s when the national and official language Kinyarwanda was designated as the medium of instruction from grade 1 up to grade 8 1 (Rosendal, 2011, p. 125), Kinyarwanda has only been used during the first 3-4 years of education, to later be replaced by French-and later English-as medium of instruction.
The introduction of English as the official language of Rwanda has its background in the long-time struggles between Hutus and Tutsis, and the displacement and exile of thousands of Tutsis starting in 1959. The family of the present president and the old aristocracy were forced to flee during the civil war in Rwanda, to the neighbouring countries Burundi, the Democratic Republic of Congo, Tanzania and Uganda, and returned to Rwanda after the genocide in 1994. The genocide, when more than 800,000 Tutsi Rwandans were killed, was stopped by the Rwanda Patriotic Front army that for decades had been living as refugees in neighbouring, mostly Anglophone, countries. The post-genocide government taking power in 1994 opened up for English as medium of instruction alongside French and Kinyarwanda, as most of the returning diaspora and earlier rebels who returned to Rwanda after the genocide did not know French, after decades abroad. Thus, English was introduced, first in the 1996-1997 revision and later in the 2003 law on the organisation of schools (Rwanda, 2003). This meant that Kinyarwanda was to become the medium of instruction in lower primary school (primary 1 to 3), and French or English were mediums of instruction in upper primary up to secondary school, depending on whether the school was following a Francophone or an Anglophone system (Rurangirwa, 2010). However, French remained the main medium of instruction within education.
The bilingual medium of instruction policy changed in October 2008, when English, at the expense of French, unexpectedly was declared to become the sole medium of instruction and to be used within administration (Rosendal, 2011;Rwanda, 2008).
The new language policy decision was implemented quickly, already in January 2009, with severe problems as practically no teachers knew English, and all academics in Rwanda were educated within a Francophone educational system. After pressures from among others UNESCO, Kinyarwanda was reintroduced to be used as MoI from Grade 1 to 3. However, private schools were allowed to use English. This later led to protests from parents and others who wanted English as the MoI even in state schools. As a result, in February 2011 the Department of Education reinstalled English as MoI from Grade 1 on (Ntakirutimana, 2012;Rwanda, 2019aRwanda, , 2019b. Simultaneously, the role of Kinyarwanda is said to be but of cultural importance. 2 Countries in the Global South are often labelled Anglophone, Francophone or Lusophone when they have ex-colonial languages as official languages. These terms are misleading. European languages such as French and English are often only learnt through the educational system, in crowded and often badly equipped classrooms. As a result, only an educated elite master the officially appointed languages of European origin. Data about language competence and practices show that the levels of competence in foreign languages such as French and English are very low (Rosendal, 2011, p. 78). 3

Ideologies in public discourse and the linguistic landscape
Official statements on the shift from French to English as the medium of instruction in 2008 reveal that the government of Rwanda mainly claims economic reasons for the decision to strengthen the status of English. This is an instrumentalisation of English. A common feature among the newspaper interviews is the idea that the language policy shift from French to English is a requirement for participating in the global economy. In this narrative, development is led by English speaking countries. This narrative additionally integrates Rwanda in the commercial regional and international world of English speaking communities such as the East African Community, composed of officially Anglophone member states (Kenya, Tanzania, and Uganda), as well as with the Commonwealth. An interview from 2007 published in the international newspaper The Christian Science Monitor, with the Rector of Kigali Institute for Science and Technology (KIST) who additionally served as the Chairperson of the Rwanda electoral commission, signals official attitudes and forebodes the coming change in policy. The rector says, when asked why the government favours English: The real driving force for Rwanda's preference for English is more economic than political, says Chrysologue Karangwa, who, as rector of KIST, oversees all departments. With a fast-growing number of foreign investors (most from Anglophone countries) coming to Rwanda, and with Rwanda joining the East African Community trade bloc, Rwanda can benefit from closer ties with its English-speaking neighbors while maintaining ties with French-speaking ones like Burundi and Congo. (Baldauf, 2007) This statement accords with several later official statements which claim that English is the language of business, capable of boosting the Rwandan economy and trade. English is portrayed as a superior language compared to other languages. An interview published by Reuters in April 2010 with the anchor of Rwanda's stateowned TV news, Arnaud Nkusi, illustrates this common discourse. Nkusi says that although he is partial to French, he sees the language change as an economic necessity: French is the language of the heart he says, but English is the language of work. And Rwandans are working hard to show they're competitive in an emerging African market. Every building in the country looks like it just got a fresh coat of paint, and the GDP is growing by an average of five per cent a year. The country's wealth is not in the soil, it's in the minds of its citizens, says Nkusi. (Hasselriis, 2010) The same idea is clearly voiced in an interview with a high official within the Ministry of Education, published in the international The Guardian Weekly. In this interview with the Director of policy planning at the Ministry of Education, the newspaper reports: Really it is not choosing English for its own sake, said Claver Yisa, the director of policy planning at the education ministry. This is a way to make Rwanda to be equal, to use English. English is now a world language, especially in trade and commerce. Rwanda is trying to attract foreign investors-most of these people are speaking English. It's choosing English as a medium of instruction so we Rwandans of today, and more importantly of tomorrow, will be able to benefit. If Spanish or any other language could get us to that, no problem. If Kinyarwanda could get us to that, that would be marvellous. It is not English for its own sake. (McGreal, 2009) The argument resonates what the President of Rwanda, Paul Kagame, voices in an interview with the French TV channel TV 5 Monde in 2018. The interview was conducted during the 17th summit of Francophonie, which took place in Armenia in October 2018. The president was asked by the interviewer, in French, if it is true as people say that Rwanda has abandoned French, referring to the fact that English has been made the compulsory medium of instruction within education. President Kagame, answering in English, categorically claims that there is no truth in the statement, and emphasises that to opt for English is a pragmatic choice, especially for business: We transact ....maybe 90% of trade, investments, all kinds of businesses with this English speaking…. group…of countries…in the region, really. So, realistically, what would be the reason of not embracing that as well, without having to leave or discard what you have been having. (TV5 Monde, 2018) This interview is also cited in other-both international and national-newspapers, among other the Rwanda-medium online newspaper igihe.com (Mukaneza, 2018). The idea that English per se brings development has also been claimed earlier by President Kagame, as in the following 2010 interview in the French medium newspaper La Croix: Nous donnons la priorité à la langue qui rendra nos enfants plus compétents et qui servira notre vision de développement du pays." (We prioritize the language that will help our children to be more competent and will serve our country's vision of development.) (d´Ersu, 2010) We find these ideas linking English to economic development in other talks by key state representatives in all official statements about the language policy change. The new status of English in 2008 also implied that English was to be used as language within administration. Even this decision is discursively linked to economic development. The Minister of Finance and Economic Planning, James Musoni, is for instance cited in the state-owned English-medium The New Times. While addressing financial institution representatives, including bankers, insurers and the Social Security Fund at the Ministry's head office in Kigali, he declared: "If Rwanda is to become a financial hub, then all financial institutions' staff should learn English" (The New Times, 2008).
The changed language policy in 2008, officially motivated as a necessity for Rwandan development, has resulted in major changes in public space in urban Rwanda. Language usage in the linguistic landscape after 2008 clearly shows this status change, which also implied a change in the way other languages are seen and used. Our quantitative data from 2005-2008 and from 2018 of signage on private shop signs in urban Rwanda demonstrate these remarkable changes, both in Kigali and in the two other towns that were investigated. Table 1 shows eight major categories which the shop signs were allocated to, in order to illustrate these diachronic changes. The symbol > in language combinations indicates the quantitative visual dominance, as described in the methods part.
As clearly seen, the increased use of English on shop signs, both on monolingual signs and in combination with other official languages, has transformed the linguistic landscape. The sign category 'English only' illustrates this trend very well. Even though English-only signs were found before the change of status in 2008, the use of English has more than doubled during the 10-year period, from 1 3 Amid signs of change: language policy, ideology and power in… 23 to almost 58 per cent in the capital Kigali (see Figure 1, which exemplifies this change). The increase of English on signs is less remarkable in the other towns, from just under 20 per cent to just over 32 per cent in Huye (Butare) and from just under 10 per cent to 35 per cent in Rubavu (Gisenyi) regarding the same category. Simultaneously, the decrease of the use of French on all kinds of shop signs is just as striking. In 2018, the use of English was more frequent than the earlier use of French: Almost 60 per cent of all signs were monolingual in English.
While English is attributed economic power and functions, all other languages are discarded or totally overlooked, as is the case of the national language Kinyarwanda. Accompanying this narrative about English as a tool for successful trade, we find statements especially degrading the functionality of French in the new era of modern development in Rwanda. Thus, in newspaper interviews, discourses promoting English at the expense of French are frequent. The following statement by Mrs. Mushikiwabo, Minister of Foreign Affairs, when asked why English was chosen as the new language of administration and education, echoes these ideas. Despite her current function as the Secretary General of Francophonie (OIF), in 2008 she said to the state-controlled Rwandan newspaper The New Times: It is known globally that French is losing value; there are about 3,000 publications from different scholars proving how the French language is losing value, why should we proceed with a language that would cost us a lot? (Musoni, 2008) By stressing the functional and economic assets of English compared to French, this quote resonates the statement of Mr. Nkusi, cited above, that even if French may be the language of the heart for many highly educated Rwandans, English is the language of economic development and growth. English forms part of the official success story and the official narrative of post-genocide Rwanda. The official narrative of English as representing the new and economically booming urban Rwanda is maintained despite the problems the very sudden and problematic transition within education from French to English as the medium of instruction caused. The pragmatic possibilities for a successful implementation were less important than the symbolic value of a new language, free from old ties and signalling a new start linked to economic growth. The international journal allAfrica (Plaut, 2012), when analysing language policy development in Rwanda, points out that in Rwanda beautification of the environment often is more important than infrastructural needs. Thus, the on-the-surface reshaping of the international image by introducing a new global language, whether it works or not in real life, is prioritized.
Although the official discourse positions English as merely a vehicle for development, and not as a decision against French and France, it is clear that the longterm political conflict the Rwandan government has had with the government of France is present in public discourse. Critics have claimed that Rwanda changed their language policy due to this political conflict and accusations by Rwanda of French participation in the genocide in 1994. In the effort to normalise the relationship between Rwanda and France, the president of France, Nicholas Sarkozy, has accepted the role of France but not apologised, as wanted by the Government of Rwanda and survivors of the genocide. This was commented on in the following Rwanda will need to hear a more apologetic tone from France before it restores relations: What we want is for France to accept that at a certain period and at a certain extent, the French government played a role in the genocide, he says. (Baldauf, 2007) Similar accusations against France have been common, as for instance voiced by Abdoul Khalim Harelimana, a member of the Rwanda Elders Advisory Forum and former Minister of internal security, who in a 2019 interview said: "Instead of protecting the people who were being killed who are the Tutsis, they came and protected the government.
[…] and […] the Interahamwe militias […]" (Durmaz, 2019). The position of France during the genocide has been described as a way of maintaining political and economic power in the region, "rooted in the centuriesold great game in Africa, particularly the competition between two colonial powers and two cultures: Anglophone and Francophone. Rwanda was where Francophone hegemony ends and where Anglophone colonial culture starts" (ibid.). Therefore, as claimed by President Mitterand's adviser Hubert Vedrine in a 1998 National Assembly inquiry, France took a global engagement of security (Wallis, 2006). It is thus hard to neglect the impact of the genocide when discussing language policy and the change to English, and the devaluation of French as main language of instruction, 14 years after the genocide. This was a major change, since French had been used since colonial times, throughout the post-colonial period, and thereby produced Rwandans who identified themselves as Francophone.
Another predominant narrative in official discourse in Rwanda is that Rwanda is accommodating a number of languages, and not choosing one to the detriment of another. In the interview with TV5 Monde (2018) with President Kagame, referred to above, the president also stressed that introducing English is not excluding French. He exemplifies this claim by saying that some schools teach Mandarin and German. English is added to French, he underscores, without touching on the higher status of a language as medium of instruction versus as a subject among other subjects within education. The fact that the use of English is problematic within education, that knowledge of English is linked to the elite in power, and that English is not spoken in Rwanda, not even in urban environments, is neglected. In this official narrative English is just added as a practical medium of global communication. Today, this discourse is also echoed among ordinary citizens. A La Croix journalist in 2019 cites that the common answer was, when asking about the switch to English: "Non, we didn't switch from French to English, but we have added a new language which is English" (Cochez, 2019). This is in part true. As a matter of fact, Rwanda has never left Francophonie, and today the Secretary General of Francophonie, Louise Mushikiwabo, is Rwandan. In 2017, TV5 Monde Info commented on the maintenance of French as well as the 2017 addition of Swahili as official language as an "economic pragmatism" (Eyguesier, 2018). The maintenance of French in Rwanda is, according to Secretary General Mushikiwabo, due to the identity of Rwandans as belonging to the Francophone world. However, the General Secretary stresses that Rwandans' identity today is linked to all four official languages: "So we're East Africans -Swahili speaking, English speaking, members of the Commonwealth, and at the same time we're Central Africans -Kinyarwanda speaking and Frenchspeaking" (Sittoni, 2018).
Still, multilingualism in the four official languages is something of a utopian idea, and the difficulties and implications of the transition to English as MoI has to some extent been targeted in local media in Rwanda (see, for example, Mbaraga, 2017). Especially university staff, but also international newspapers and scholars have questioned the language policy change, due to the severe problems introducing a language which is not known among the population causes. This is true both for teachers and for learning outcomes (Mbonyinshuti, 2020;World Politics Review, 2017).
Although the official ideology is that all official languages in Rwanda are equal and that there is no rivalry between the languages, historically there has for long existed a competing situation between French and English. In an interview with Jeune Afrique in 2005, President Kagame, while acknowledging the competition, clearly took a marked economic stance, declaring:

Nous avons ici trois langues officielles: le kinyarwanda, l´anglais et le français. Chacun évolue, progresse ou régresse dans le cadre d´un marché linguistique ouvert et libre. L´état n´intervient pas.
We have three official languages: Kinyarwanda, English and French. Each develops, prospers or regresses within a linguistic market which is open and free. The state does not interfere (Rosendal, 2010, p. 84) In fact, the state does interfere. English has clear ties to the post-genocide political power. President Kagame and most of his administration knew English, and not French, after returning from long-time exile in Anglophone countries. In addition to politico-economic and bilateral reasons given in official declarations, the new language policy of Rwanda has accommodated the post-genocide leaders', including the president's, language skills. 4 Ignoring the unequal access to English and that status is accorded to a language appointed MoI and used by the elite, English is portrayed as a neutral tool.
Our quantitative data on language use on shop signs clearly show that one exogenous status language has been replaced by another exogenous status language. In fact, the use of Kinyarwanda, the national language, also has decreased, both in Kigali and in the two other towns. However, we find more signs in Kinyarwanda or with combinations of Kinyarwanda and other languages outside of Kigali. Also worth noting is that, in contrast to the situation in Kigali, multilingual signs were more frequent in smaller towns than in Kigali. A closer look at the English-only signs also shows that there is a considerable variation between different streets in Kigali. Table 2 illustrates this.
Two of the streets, KN 2 St (earlier Rue du Travail) and KN 59 St (earlier Avenue du Commerce), show lower percentages of English-only signs. This was the case even a decade earlier. Even if all streets have a mix of different kinds of small shops and commerce, these two streets are popular shopping streets for ordinary people, often with low education, which even the earlier names (streets where commercial and manual work is found) in French and Kinyarwanda indicated. Nevertheless, even in these streets, English has become more common than before, used on almost 55% and 45% of the signs respectively. A similar trend is found in Huye (Butare) and Rubavu (Gisenyi). As seen from Table 3, more signs using English only are found in the main street of Huye (Butare) than in the smaller street off the main street, SH 1 Rd. 5 This change in the linguistic landscape of urban Rwanda has also been noticed by journalists, who have noted with surprise a reduced use of French, for example in restaurants and other functions in public space (see, for example, Cochez, 2019). In the following part, we discuss further these changed practices and their implications.

Discussion and conclusions: language policy, ideologies and practices
By officially and uniformly describing the 2008 status change of English as a pragmatic measure and a tool enabling economic development, ideologies about the unique and positive qualities of English are spread and legitimised. English is presented as the language of development, science and technology that serves the Rwandan state. Thus, all Rwandan citizens should learn it in order not to be left behind in worldwide trade and commerce. This universalization and rationalisation discourse (Thompson, 1990) is paralleled with an official devaluation of the functionality of French, a language which is losing ground. Both through dissimulation and figurative language, French is reduced to merely a language of the heart, a language to express emotions not suited for trade and business. Neither the role nor the potential of the national language Kinyarwanda is discussed in public discourse. This concurs with Janks et al.'s (2014) claim that ex-colonial languages are kept in African countries to keep power in the hands of dominant groups and their patrons in Europe or the United States of America.
The ideology revealed in this discourse appears to be a strategy for constructing a new national identity, where English represents the modern and independent future of Rwanda, free of past history. Through official statements spreading the ideas that English is the only language which enables trade and economic development, successful and long-time past trade with Arab countries and China as well as earlier trade with officially Anglophone neighbouring countries in the region is neglected. Furthermore, the fact that not all successful, rich and so called developed countries in the world have downgraded and replaced their national languages with English, for example within education, is also ignored.
The unexpected introduction of English as the MoI on all levels of education has been interpreted as the result of crashed diplomatic relations between Rwanda and France (Rosendal, 2011;Semujanga & Galabert, 2013), which the fact that Rwanda sought new partnerships and joined English speaking organisations such as the East African Community (EAC) and later the Commonwealth (Rosendal, 2011) indicates. The language policy change in 2008 has also been seen as an intentional political and economic long-term Anglo-American policy to overthrow the influence of the French language in Africa (Samuelson & Freedman, 2010), since Rwanda has a strategic geographical position on the linguistic Maginot line between Francophone Africa to the West (DRC) and Anglophone Africa to the East. In the same vein, Leclerc (2019) argues that the Anglo-American support to Rwandans based in Uganda in their struggle to regain their country was aimed at stabilising the region. The government of Rwanda has consistently denied that the shift from French to English had anything to do with the diplomatic fallout with France. By downplaying the political conflict between Rwanda and France, the political factors behind language policy decisions as well as the impact of the genocide and other historical and political developments are concealed. Even the elite in power's favourable mastery of English, through years in exile in Anglophone environments, is obscured. Furthermore, ideologies behind the official narrative given in the data of this study reveal underlying ideas about languages. French is officially not a language suitable for trade, while English is. English is even the sole language which is deemed suitable as medium of education, although Rwanda has a well-developed national language known and spoken by at least 99.4% of the population, and Kinyarwanda was successfully used as MoI up to grade 8, as well as language of printed school books in the late 1970 (Rosendal, 2011, p. 125). Despite this fact, the Ministry of Education reduces Kinyarwanda to having merely a cultural function. Additionally, official discourse reveals seemingly naïve ideas about multilingualism-that being taught in a language implies that you learn the language fluently and this makes you bilingual-in a society where people speak only their national language in all kinds of situations and in all domains.
Official statements indicate that all the four official languages in Rwanda have their unquestioned place; in official discourse it is underscored that Kinyarwanda, French, English and Swahili are accorded the same status and are equal. This discourse stresses unification. English is said to be just an added language. However, this is not the case, neither regarding official status allocation nor practice. Through the appointment as the MoI and language of administration in 2008, English is allocated a superior status. Statements by top politicians dissimulate this fact and legitimise the 2008 language policy decision by utilising the mode of narrativisation (Thompson, 1990) by presenting the status of the official languages as being on the same footing. This disguises socio-economic differences and educational inequality and demonstrates the unidirectional character of ideologies; these work in favour of certain groups at the expense of others, as stated by Purvis and Hunt. The English speaking elites and educated Rwandans are in a superior position, even though the earlier French speaking elite has had to learn English, which they have, but with problems. Those who are disadvantaged today, when English is used as medium of instruction and earlier when French was employed, is the general population who speak Kinyarwanda and only obtain a few years of education. This illustrates that the choices made are far from neutral and that the Rwandan LPP ignores the function that the national and co-official language Kinyarwanda could play, at all levels of society. In this regard, Rwanda is in an exemplary position in the African context, where multilingualism is the norm and it is difficult to assign functions to one indigenous language. As most communication in Rwanda is conducted in Kinyarwanda, neither of the official exogenous languages are heard or used on TV, the radio or in the streets and markets of towns. Knowledge of French (earlier) and today English is thus an issue of socio-economic conditions, such as parents' possibilities to send their children to private schools, if they can afford sending their children abroad during holidays in order to learn English, and naturally also for how long parents can afford paying for their children's education. By not promoting and using the potential of the national and official language Kinyarwanda, language policy and planning in Rwanda upholds socio-economic differences.
As pointed out by Shohamy, the modes and mechanisms which are used to explain political decisions, reveal power differences. English is an integral part of the official success story and the official narrative of post-genocide Rwanda and the building of a new Rwandan identity. Simultaneously, the way the recent language policy change is presented to ordinary Rwandans obscures relations of domination and power. The official discourse, the discourse in place, has trickled down to the streets, now visible in the linguistic landscape, and thus turned into practices. These social actions mediated by circulating discourses, ideologies about values, status and use of languages, are as pointed out by Hult (2015), interpretations of institution policies in daily interactions. Despite their priority to sell products, shop owners have commissioned signs that to a high degree use English-neglecting the fact that the vast majority of their customers do not read or speak English. This has shaped a changing linguistic landscape where the symbolic value of English, as part of the official narrative of a modern state which identifies with the Anglophone world, is showcased.
Although the Rwandan state has changed public street signs 6 to align with recent changes in language policy, from using Kinyarwanda-French to English, official language policy does not stipulate or delimit how languages are to be used on signage in the LL. However, as single policy actions are discursively connected to the larger policy nexus, they become de facto policy (Hult, 2018;Shohamy, 2006;Spolsky, 2009). This indirect connection between visual language use and the circulating discourse used as powerful and covert means to manipulate people, reproduces language ideology and in this way serves as a policy mechanism. Officially, Rwanda has four official languages, formally claimed to be equal. The recent promotion of English and discourse by the elite in power in combination with the visual representations in the LL clearly showcases this de facto language policy, where only English is promoted. Shohamy (2006) stresses that decisions about language use in public space are found on different layers of decision making about language use, from national to individual levels, and are always linked to ideologies By investigating the nexus of large scale discourses resonating history and political developments and actions mediated through semiotic tools, we have in our study seen that these social actions reverberate with the interaction order in the LL, even outside of the capital Kigali, and have become sedimented into practices.Thus, we have investigated historical perspectives and ideologies put forward by people in power as well as social actions in the LL. The diachronic empirical data of the presented study clearly have demonstrated the significantly increased use of English on shop signs, both on monolingual signs and in combination with other official languages and illustrated the transformations in the linguistic landscape. This transformation underscores the impact of LPP mediated by discourses in practice and by nationally circulating ideologies.
By status attribution through the designation of English as the MoI, the Rwandan government, in power since returning from exile in neighbouring Anglophone countries after the genocide, has strengthened its position. Discursively, by using narrativisation, the political elite legitimise their power by giving prominence to English as the only cultural and economic capital, and by pretending that to impose English is an inevitable development and the only way forward for Rwanda towards the economically prosperous future. By simultaneously downgrading French, a language associated with both the colonial past and the genocide, socio-political problems originating from the introduction of English are dissimulated. Future research may in more depth analyse recent developments in the light of the ongoing healing of the past and post-genocide restoration, reconciliation and unification. Still, the narrative of a new collective Rwandan identity where one language is the key for success while other languages and speakers are left out constitutes the new LPP. The majority of Rwandans who only master Kinyarwanda are excluded in this narrative, even if shop owners, as the quantitative data accounted for in this study shows, attempt to be part of the national development narrative by identifying with official public discourse.
Funding Open access funding provided by University of Gothenburg.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.