Multilingual Practices and Multiple Contestations in the Linguistic Landscape of Selected Towns in Oromia: A geosemiotic perspective

This study analyzed a linguistic landscape, which is a relatively recent development in sociolinguistic research committed to the investigation of the visual manifestations of the relationships observed between language or semiosis and society in the public space. There has been less attention on the semiotic interpretations in the public environments despite the ubiquitous nature of such signs in Ethiopia. This study intends to fill this gap through Semiotic Landscape research. By means of mixed methods, qualitative and quantitative, it aims to discover the contestations in the Linguistic Landscape (LL, henceforth) of selected urban environments in Oromia. The study employed a theoretical concept of geosemiotics from social semiotics as an analytical tool. Visual data consisting of 1275 photographs of signs were collected from the main streets of Adama, Jimma and Sebeta towns from different domains of sign using bodies. Besides, data were collected through in-depth interviews with major LL actors, selected based on purposive sampling technique. The analysis demonstrated that some top-down and most of bottom-up signs showed varieties of multilingual practices, contestations, identity constructions and other related social concerns. The absence of a clear policy on language use on signs at both federal and regional levels has resulted in contestations.


Background of the study
Urban environment provides present-day societies with multifaceted visual images that become components of their daily experience. These visual messages have different modalities which involve either language or image, or a blend of both (Kress & van Leeuwen, 2006;Scollon & Scollon, 2003;Leeuwen, 2005). According to Backhaus "language on signs is a form of impersonal communication via written messages […] in public space. These messages make up a vital part of information provision in urban everyday life. A city without signs would be more than inconvenient -it would be largely unintelligible " (2007:54). Generally, these signs visible to the public in different forms can be considered collectively as representing the LL. And "LL is a most important indicator capable of providing relevant information about societies, vitality and the inter-relationship of groups, especially in linguistic contested regions" (Shohamy & Gorter, 2009:2).
One of the focuses of LL research is multilingualism. Multilingualism manifests itself in either spoken or written form, or a combination of both. And, language used on signs is among the ways of practically implementing the linguistic diversity of our country, Ethiopia. In this study, multilingualism subsumes bilingualism. However, due to competitions among different groups, urban environments are sites of multiple contestations-covert or overt conflicts among LL-actors. This seems to be a reality in contemporary Ethiopia. For example, as Moriarity (2014:457) asserts "… the LL of a given community has become an influential site of contestation and negotiation, appropriation and resistance..." The current Ethiopian government structure is a federal administrative system that is based on relative ethnolinguistic and cultural similarity and difference. Accordingly, the country is divided into nine regional national states and two city administrations, namely Addis Ababa and Dire Dawa. It was after the collapse of the Dergue regime in 1991 and the coming of the Ethiopian Peoples' Revolutionary Democratic Front (EPRDF) to power, that Ethiopia was restructured into Federalism. This was first indorsed during the Transitional Government of Ethiopia (TGE) based on the proclamation of the Transitional Charter (TC) in 1991 that gave for the first time the right to self-determination of Nations, Nationalities and Peoples (FDRE, 1991:Article 2). Hence, the federal government was formed based on ethnic and linguistic composition. Besides, the 1995 constitution reconfirmed the earlier division of the federal government structure (FDRE, 1995:Article 47). Among these federal regions Oromia became one. And it was established as a regional government in line with National Regional Self-Government Proclamation No. 7/1992 and Article 47, No. 1 of the Federal Constitution. (FDRE, 1992) Among major towns in Oromia, this study focuses on Jimma, Adama and Sebeta. As a commercial town, Jimma has been attracting different ethnic groups from all over Ethiopia. As a result, according to Ethiopian Statistical Agency (ESA, 2007), among the total population of 120,960 residents of the town, speakers of different languages such as Afan Oromo (46.7%), Amharic (17.1%), Dawuro language (10%) Gurage language (6.4%) are living in Jimma town. Therefore, with a population of very diverse linguistic groups, despite the fact that the majority of its inhabitants are Afan Oromo speakers, and that the regional official working language is Afan Oromo, the town becomes the right context for LL research from the perspective of multilingual practice and inevitable contestation issues over the competition for visibility in the public space.
Similarly, Adama, as a town of commercial center due to its location on the main trade route of the country, has attracted many Ethiopians from different corners of the country, and therefore, it is home to many linguistic groups. In the 222, 212 total population of the town, speakers of Afan Oromo (38.6%), Amharic (34.2%) Gurage language (11.8%), Tigrigna (3.3%), etc., are living in the town (ESA, 2007). Moreover, as a seat of Oromia parliament, there is special attention given to the town by the regional government. Thus, Adama is also a multilingual town and a meeting point of two official languages; hence, it is suitable for LL study. By the same token, in the central part of the country, Sebeta was chosen for this study due to its proximity to the capital city of the country, Finfinnee (Addis Ababa), which is home to many big national and international organizations that pave the way for both local and global influences regarding language use in public spaces. In this respect, Adama, Sebeta and Jimma can be taken as multilingual environments where the study of public signage may yield good results.
Based on this context, such LL research is important in contemporary Ethiopia. During the last two decades, Ethiopia has gone through fundamental social and cultural redefinition (Moges & Blackwood, 2016) and political transformation (Keller, 1998) as a result of government change in 1991. The change has brought among other things crucial transformation on the semiotic and linguistic practice of the country, as power is relatively decentralized to regions.
1.2 Statement of the problem Languages on signs have been used in Ethiopia for centuries as a means of communication. Apart from communication, these signs have other roles to play as they carry ideological, attitudinal, political, historical and other issues that can influence society in different ways. Of course, those groups in power try to guide which languages to use in public spaces. Most of the sociolinguistic studies in Ethiopia or elsewhere have been focusing on spoken language, whereas written language and signs have not been given much attention. The towns, however, are not only "a place of talk" as has been emphasized by different scholars (Halliday, 1978:154). It is a place of "writing and reading, too" (Backhaus, 2007:1).
Needless to say, in Ethiopia, how much languages and signs reflect the multilingual issues of the country and how problems related to multilingualism can be revealed by LL analysis have not received attention. Thus far, the few studies conducted in the area of LL in our country have been centering on ideology, attitude and indexing modernity. This study therefore intends to fill such a gap. By analyzing varieties of written languages and other semiotic characteristics, the multilingual practices and contestations in the selected multilingual settings can be made explicit.
1.3 Objectives of the study The purpose of this paper includes: (a) Analyzing the multilingual practice and the preferred languages of the public and private signs. (b) Investigating the rationale of the LL actors for their language and sign practices and choices. (c) Analyzing the contestations in the LL.

Analytical framework
This study employs geosemiotics as a theoretical framework, adopted from social semiotics (Scollon & Scollon, 2003). Geosemiotics focuses on how language on signs makes meaning in relation to how, where (and when) they are physically placed. Particularly, the main focus of geosemiotics is on indexicality which gives priority to the fact that the meaning of signs depends on their context. In this case, language on signs in a certain environment indexes something, such as the linguistic community, the geographic boundary, the political, cultural and social situation of the area.
Geosemiotics views LL signs dialogically and considers all semiotic aggregates, which is "multiple semiotic systems in a dialogical interaction with each other" (Scollon & Scollon, 2003:12). Regarding the analysis of signs, this framework helps not only to interpret the content, order, organization, layout, and structure of signs, but also to see how signs are context-bound within a specific social context. Focusing on geosemiotics is important because, "language in public spaces is not limited to written words and multiple languages but rather encompasses a broader construct, that of multimodality" (Shohamy, 2015:153).
Similarly, Blommaert and Maly (2016) described signs from the point of view of geosemiotics and stated that signs can be interpreted by considering three "axes". These are: (i) signs point towards the past and tell about their origins, modes and processes of production. The elements of the compositions on signs and the linguistic make-up indicate who produced the signs, under which conditions they were manufactured, who owned them, what were the resources used, what was the rationale behind and so on. They further note that "the history of the sign … leads us towards the broader sociolinguistic conditions under which the sign has been designed and deployed". (ii) Signs also point towards the future, which means to the intended preferred readers, with the intention of making specific effects in mind among specific addressees. (iii) "Signs also point towards the present, through their "emplacement" … their location is not a random given, and neither is their "syntagmatic" position relative to other signs".
The geosemiotic analysis focuses on critical observation of the presence or absence of languages, language choices, the amount of information in each language, code preferences, text vectors, symmetrical composition, and emplacement issues of signs, relating them to different contextual factors. "One possibility for analyzing languages contextually is to focus on the written information that is available on language sign in a specific city" (Gorter, 2007:3). The language choice and the writing systems can convey the socio-historical information of the setting under investigation, not only by mere analysis of the languages presented on signs, but also by the way they are emphasized differently (color, font size, contrast, placement, etc.). Scollon and Scollon (2003:209) confirmed that language choice includes code preference, that "the relationship between two or more languages in bilingual (multilingual) signs and pictures […]". Accordingly, Lou (2010:207) witnessed that "geosemiotics provides a useful framework for understanding the importance of the physical, material, spatial context for the meaning of language and signs." Therefore, this study interprets both linguistic and non-linguistic signs collected from the selected urban environments supported by interviews with LL actors.

Methods
A pragmatic approach was adopted in this research. The main reason is the fact that pragmatism stands for a practical and applied research philosophy. Moreover, the pragmatic approach gives due attention to the historical, political, social and some other contexts (Creswell, 2009:23), and the issues of multilingual practices and contestations in LL are also under the direct influence of these contexts (Landry & Bourhis, 1997). Therefore, this study employed both quantitative and qualitative methods, with more emphasis on the qualitative one, because the LL data are usually more available in qualitative form.
Visual data consisting of photographs of signs were collected from the streets of Jimma, Adama and Sebeta towns from December 2015 to June 2016. Besides, data were collected from LL actors such as audience (sign readers), sign owners (public institutions and private), sign writers in general, and concerned government bodies through in-depth interviews to investigate their reactions to the written linguistic and non-linguistic signs of the towns. The motivation for selecting these towns was three reasons. These towns are relatively inhabited by heterogeneous ethnolinguistic groups that make the research on the subject possible. As a resident in Jimma, and as an occasional visitor of Adama and Sebeta, I have experienced some disputes and contestations between local government agencies and private sign owners on how to put their signs. Besides, at least two of Ethiopia's major languages, such as Afan Oromo, Amharic and English with no speech community, are expected to be found on signs in these towns. In this study, the whole semiotic aggregates of the sign establishment were considered as a unit of analysis based on the suggestion of Cenoz & Gorter (2006), because the communication role of a sign is not based only on the verbal part.
The major data collection tool was the photographing of signs using a digital camera. In total, 1275 signs were collected and analyzed. Interviews were conducted in two languages, Amharic and Afan Oromo, with 20 purposefully selected informants, based on the language skills of each interviewee. And then, the audio data were transcribed based on the method of orthographic transcription that enables the author to write down the spoken words using the customary spelling system (orthography) of the languages. Both quantitative and qualitative data were concurrently presented and analyzed in line with the specific objectives.

Data presentation and analysis 4.1 Multilingual practices in the LLs
Of the 1275 signs categorized as multilingual and bilingual as conceptualized in this study, further classification was made hereafter sorting the corpus according to Reh's (2004) taxonomy that is based on the content of the information on signs in each language. Multilingualism can be understood better by observing language-related practice in multimodal texts. Based on the classification of Reh (2004), the data of multilingual practice in the LL are divided into four categories. These are duplicating, fragmentary, overlapping and complementary multilingual signs. All of these categories were observed in the three towns, but their presence varied based on the domain of sign owners (regional or federal government, private, religious institution, NGO) and also from sign to sign. Duplicating multilingual sign, which presents all the required information in a balanced way in all the languages on a sign, was presented only on private banks. Such signs provide information to any reader proficient in one of the languages on the signs so that a reader can decode all the signs (symbolic, iconic and indexical). Ten different private banks strictly followed this multilingual practice. On the other hand, regional offices frequently use bilingual Afan Oromo-Amharic and sometimes multilingual Afan Oromo-Amharic-English signs in a balanced form; with a slight bias towards the regional language as the primary position was given to Afan Oromo on the signs.
Hence, there were four approaches to multilingual writing as encountered in the streets. The data in Table 1 focus on how the languages on a sign are visible, particularly on the contents of the messages on the signs. This multilingual writing approach could be taken as focusing on one component of geosemiotic analysis, which is an analytical tool in this study (Scollon & Scollon, 2003). According to the numerical data presented in Table 1, the four categories of multilingual practices -duplicating (15.21%), fragmentary (13.65%), overlapping (64.47%) and complementary (6.67%) -are used in the LL of the towns. The majority of the multilingual signs are the overlapping type and each multilingual type has its own meaning in relation to the language in the LL and also in relation to the readers of the signs.
Moreover, the specific data for multilingual Afan Oromo, Amharic and English in the three towns show that 58.45% of the signs are in the form of overlapping and 9.17% are in the form of fragmentary. The overlapping multilingual signs are the great majority because many of the signs contain additional information in most of the cases in Amharic and English. The signs under the overlapping category are those used by private business owners, and the duplicating signs were mainly used by the regional and federal governments (top-down signs).  (Table 1). Similarly, among the three types of bilingual signs identified in the towns, 68.59% of Afan Oromo and Amharic co-occurring signs and 48.28% of Afan Oromo-English signs are visible in the form of overlapping multilingual signs (Table 1).
Therefore, the general analysis shows that overlapping multilingualism is much recurrent. In overlapping multilingual writing, only a part of the message is repeated in another language. This means, in the other languages on a sign, different short additional information is usually given. Hence, some contents of the texts in the different languages on the sign overlap with or are repeated in some other language on the same sign. In this case, the language that contains additional information besides the part overlapping with the other language is considered as more valued and with better status than the other languages on the sign. This is the main reason behind Rhe's (2004) classification of bilingual and multilingual practices. Overlapping multilingualism is very frequent in the LL (64.47%) mainly on bottom-up (used by private sign owners) signs (53.57%) as shown in Table 1 and Table 2. On such multilingual signs, only a selected part of the message presented in one language completely overlaps with another language.
This multilingual sign ( Figure 2) is overlapping because two of the languages (Amharic and English) miss some of the information presented in Afan Oromo. This is an overlapping multilingual sign as such signs present similar information at least in two languages. On this sign (Figure 2), this role presenting similar information is assigned to texts of Amharic and English. On the upper right-hand side of the sign, there is an abbreviation that reads 'DhRTVO' (Dhaabbata Raadiyoofi TV Oromiyaa). The English and Amharic versions of the abbreviation are not included. In the same way, on the left-hand side of the sign, immediately below the logo of the organization, there is a phrase Mootummaa Naannoo Oromiyaatti, which means The Regional Government of Oromia.
This phrase intends to convey the message that the television organization is under the administration of the Oromia national regional state and this part of the information is not repeated in Amharic and English. Hence, this is one example of an overlapping multilingual sign that favors Afan Oromo.
The text in the middle written in Afan Oromo with bigger font size reads Dhaabbata Raadiyoo fi TV Oromiyaa, and is translated into Amharic and English as Oromia TV Organization. According to this sign, Afan Oromo's better status can be seen from the sign's multilingual practice. This is due to the fact that the sign is owned by the regional government that mainly uses Afan Oromo as a medium in the broadcast of the TV organization. In fact, both Amharic and English have minor coverage for a few minutes only to broadcast news.
Moreover, the classification of multilingual and bilingual arrangements can be further classified and analyzed by the sign type, i.e. top-down and bottom-up signs. Top-down signs are owned by governments (Federal & Regional), while bottom-up signs are owned by private business owners, religious institutions and NGOs. This differentiation helps to locate the sign using domains in the LL and their preference on the type of multilingual practice.  Table 2 shows that four of the multilingual writing strategies are involved in the construction of the LL of the towns under both the bottom-up and top-down sign categories. Top-down and bottom-up signs are similar in that overlapping signs are more frequent than other multilingual sign types. 10.90% of the top-down signs, i.e. signs of regional, federal and government signs and bottom-up signs such as private business and companies' signs and NGOs' signs accounted for 53.57% of the overlapping multilingual writing practice. The top-down signs were small compared to others. The other similarity in the above data between top-down and bottom-up signs lies in the fact that duplicating multilingual writing represents the second highest frequency, 48 signs (3.76%) and 146 (11.45%) respectively.
Yet, there is a difference among the top-down signs themselves in that the duplicated language signs owned by the regional government are Afan Oromo and Amharic, Afan Oromo and English bilingual signs or Afan Oromo, Amharic and English multilingual signs.
On the other hand, while the ownership of a sign is under the control of the federal government, the duplicating multilingual practice is in bilingual Amharic and English. And the religious institutions' signs are more of the fragmentary multilingual type.

Figure 3. Complementary Bilingual Information Arrangement on Private Sign
A sign is called complementary multilingual when different languages on a sign convey different messages. In such situations, a reader has to be familiar with all the languages used on the sign in order to understand the intended message. The difference between overlapping and complementary multilingualism is that there is some repetition in overlapping multilingualism and none in complementary. The bi-functional sign in Figure 3 is a typical example of complementary bilingual sign with Afan Oromo on the top telling Bar-ga-raii As-teer (Aster Berger) with unnecessary addition of the vowel "a" on the third syllable of the first word. At the bottom, it tells the reader about Aster Juice in Amharic orthography. The icons of burger and juices in glasses are more visible than the languages used. Hence, the sign is a bilingual complementary sign in that it shows two different pieces of information in two different codes, Afan Oromo and English. The total number of bilingual Afan Oromo and English signs is only 29 from the total corpus. From this, about half (48.28%) are overlapping where English dominates in the information content with some overlaps with Afan Oromo ( Figure 5).
Therefore, when all the data under this section are considered, the LL is dominated with categories of multilingual and bilingual arrangements with Amharic as a preferred language of communication on signs.

Figure 5. Bilingual Afan Oromo and English Signs
From the point of view of the expected readers, the four types of multilingual signs require different types of readers. The duplicating signs can be easily understood by monolingual readers because the readers have options to read only the text in one of the languages on the sign they know more to comprehend the full information, as all information is presented equivalently in all languages. Unfortunately, such signs are very limited in the LL. On the other hand, complementary signs require multilingual readers because the readers need to comprehend all the languages visible on the sign to know the intended complete message of the sign. In fact, such signs only account for 6.67% in the LL as a whole, considering the three major languages in the towns (Table 1). The tendency is similar when Amharic is excluded and Afan Oromo and English considered ( Figure 5). Furthermore, fragmentary and overlapping multilingual and bilingual signs show the status and hierarchy of the languages on the signs and the readers are also expected to read a language with a better status. As a result, 64.47% and 13.65% of the signs in the LL are categorized in the overlapping type and fragmentary. Therefore, the contestation between sign owners and culture and tourism office workers in the towns is mainly due to the fact that regional government wants the signs to be the duplicating type.

Rationales of the LL actors and the subsequent contestation
Although these towns are multilingual in terms of the spoken language, it is not sufficiently visible when languages on the signs are observed. It is not because different ethno-linguistic groups are forbidden to use their language on their signs, but because they don't think it is significant in business as their interview responses reveal. As writing in their language is one way of expressing their identity, they fear that it may entail that the service is only directed to a limited group of people in the community. According to Ben-Rafael (2009:48) language on sign is a manifestation of collective identity and it has "special ties binding given actors with specific segments of the public." Yet, this might result in negative outcomes, particularly in heterogeneous environments and to the minority. The following quote from the interview data opposes the use of language on signs as collective identity reflection. As it can be understood from the interview quote, readers associate the language used on the sign with the owner of the business. Therefore, the result may not lead to making a profit as it has implications with the potential customers. As a result, though everybody can use identity markers, including languages on the signs based on the local requirement, this sign owner considers it a threat to success in business. Hence, as towns are centers of heterogeneous dwellers, sign owners may prefer to use more widely used languages.
As the above quote (MAP) makes clear, there are business owners fussy about their particular identity, even regardless of the benefits. Of course, there are times when business owners benefit from identity marketing as Figure 6 shows. The figure reveals the front part of the multimedia shop which the above interviewee commented as inappropriate by showing her index finger. The sign of the business bears the name and service "Bilisummaa Maltiimiidiyaa Pirodaakshiin", which is a transliteration from English. The second language on the sign is Amharic, and it is also a transliteration of English. The third language on the sign is English, and it says "Bilisuma Multimedia Production". Bilisummaa is a term from Afan Oromo to refer to freedom and in this case, the business is named after the owner's wife. The sign owner's choice of this proper name for the business together with the languages and how they are written has also an identity marking visual semiotics.
On the other hand, there are sign owners who assert that they have benefited from reflecting their identity using their own language and other identity marking semiotic systems on their businesses. The name of the business, the languages on the signs, how these languages are written, and other visible and audible semiotic characteristics that reflect the identity of the owner have positive socio-economic implications beyond identity reflections. In contrast, as shown above, for some of them this has a negative consequence because it has a tendency of targeting too specific a segment of the society.  Figure 6 shows, there were two traditional coffee sellers adjacent to each other, in front of the coffee sellers' main businesses, which were secretarial and internet service and multimedia production. The secretarial and internet service was owned by a Christian female and her coffee was served in a cup colored green, yellow and red from top to bottom (Ethiopian flag of the imperial regime). The multimedia business was owned by a Muslim male and his coffee was served with cups colored black, red and white from top to bottom. The contestation was between the colors of the cups (green, red, and yellow against black, red, and white). Besides, the main source of the contestation between the two business owners was the icon of the lion on the cup in the center of the three colors reflecting the Ethiopian flag of the imperial period. It symbolizes the flag of Emperor Menelik II, who was considered as invader according to the political ideology of the regional government, Oromia and Oromo people in general. This cup also had a lion holding a flag tied on a cross, a symbol of Christianity, indexing that the government and religion were different sides of the same coin during the time. As a result, the contestation was not only due to political ideology, but also due to religion. The owner of the cups with black-red-white was a follower of Islam. Hence, this is double contestation-political ideology and religion fused into market.
Informally, many people call such coffee cups (green-yellow-red) Menelik's cup. There are individuals who want to drink coffee with this cup. On the other hand, there are others who detest drinking with such cup. Of course, there are still others who are not concerned about it.
As a contestation to the Menelik's cup, coffee cups with black, red and white colors are placed at adjacent place. It represents a color choice of the Oromo Gadaa system, which is a democratic cultural, social and political administration, and it is also recognized by UNESCO as a world cultural heritage that also includes Irreechaa, Oromo thanksgiving ceremony (Legesse, 1973). Hence, it is highly valued by the Oromo regardless of political and religious orientation. The color order of the cup is slightly different from the flag currently used by the national regional state of Oromia. Such cups cannot be purchased from the market as they are not available. Therefore, the owner made a design for the contestation propose and decorated it in such a way using different inks that suite the intention. As a result, there is no positive social relation between these private business owners working side-by-side facing the main road. Often, they accuse each other, and the reasons are these ideologically antagonistic coffee cups and some other semiotics heard and seen around the business. As Scollon and Scollon (2003:16) showed, signs are usually in a dialogical interaction with each other, not only with the readers. In such a way "LLs are constructed by different discourse communities with multiple and often contradictory ideologies regarding the role of the shared public space" (Shohamy, 2012:547).
At the beginning there was only one cultural coffee business served in a normal white cup in that particular public space. According to the owner of the traditional coffee business, reflecting Oromo culture, who is also the owner of another business, they had been drinking coffee there because during that time, the coffee cups were 'normal', which means that they were not colored with Menelik's flag. And they had a good relationship. However, when the owner added the Menelik's cup to the 'normal' coffee cup the nearby business man thought about it. And, finally, he decided not to drink the coffee with a cup that had a symbol of Menelik on it. The interviewee puts the action as follows in his own words: As I don't have a right to stop this, I ordered the woodworkers to make me a cup seat that has an icon of odaa (oak) tree on it. This was to say come and drink coffee. I also thought how to modify the coffee cups. I decided not to use the white ones. I discussed with my wife. I gave her the design, and she painted it with appropriate ink. Then, we [the interviewee and his wife] started selling coffee … The government employees and other bodies who come to this town for a meeting, observe that these cups [Oromo's] and those cups [Menelik's] side by side, and they never pass without drinking from here. The primary aim was not to get money. We did that to reflect our identity. We are proud of those our customers who come here. The government bodies are also happy to be served here. (HAP, Translation) As a result, the cups not only helped to contest the nearby green, yellow and red cups but also served as an influential advertisement because of its unique semiotic features. As far as non-verbal signs in the LL of the three towns are concerned, the cups presented here are not the only cases. This is illustrated as one of the cases.
According to Pandey, (2015:101) "public space in local contexts remains a site of contestation not just for competing languages per se (within specific speech communities), but more interestingly", it is also a reflection of competing ideologies in areas of religion as Woldemariam and Lanza (2012) showed. Such contestation is also observed in graffiti, a hand written sign or a drawing as a reflection of discontent, support or opposition. For example, in Figure 7 is an Amharic expression on signs in the form of graffiti which roughly means "Law shall be respected! Allah/God is Greater". It is used as political or social commentary. The graffiti sign in Figure 7 is communicating this message to the government as there was an assumption that there was a government involvement in an internal conflict within the Muslim religion followers favoring one of the groups. Since graffiti is usually an unauthorized sign, it is difficult to locate a writer. However, from the linguistic content of the graffiti and where it was written, it is possible to conclude that the writer targeted the federal government as intervening in the internal affair in an unjust way. "Graffiti is about conquering space" and accordingly unauthorized and anonymous writer has occupied the public space by masking the federal government sign. This is what Rubdy (2015:3) asserted, that signs reflect a form of religious activism as an index of "contestation and dissent in situations of social and ideological conflict". In this case, they indicate contestation and dissent with government.
The graffiti was written in a planned way. This can be understood from the fact that the sign chosen for the graffiti is federal government's legal sign. Members of the Muslim followers in the area also share this view as confirmed through interviews. It was overwritten on the legal sign as an act of deleting it. By doing so, the main sign was covered and the graffiti became more eye-catching. Therefore, other than the government, the message also targeted passersby, particularly the Muslims. Moreover, the ink used to overwrite was not the easily erasable one; the writer used red brown antirust paint. As a solution, the sign owner, Jimma University replaced the sign with a new one rather than displaying the contestation graffiti. "Graffito are transgressive because they are not authorized, and they may even be prohibited by some social or legal institution" (Scollon & Scollon, 2003:151).
Hence, such graffiti sign is linked to Scollon and Scollon's concept of "transgressive semiotics" which refers to the visible signs around semiotically prohibited zone in a broader public space, or against the expectation or in violation of a public expectation (Scollon & Scollon, 2003:147, 150). However, from the point of view of this graffiti writer, it seemed that this sign could be labeled as a transgressive sign used as a reflection of contestation regarding what happened during the time together with its solution, "The law shall be respected". Being there, the graffiti could get many readers, and most probably there were many who shared the view of the writer. Graffiti signs could be resistant to the existing hegemony of different types, and in this case contesting government interventions in religious practices. Hence, such markings could be made upon a wall, a fence or other surfaces like bigger billboards or like this one, in a way that is usually visually accessible to the public.
The main sign was put by Jimma University Law Faculty Legal Aid Center, but the sign was posted outside the university, around Jimma Zone Higher Court, where individuals with low income to pay for attorneys could get free legal service by the staffs of the faculty and their senior students. Like some other federal government signs, the sign is bilingual English-Amharic. The only difference is that English is on the top unlike other federal government signs that put Amharic on the top. This may be linked to the privilege English has in Ethiopian higher institutions. However, the presumed readers are the rural English illiterate customers in need of legal advisors, and also who are less fluent in Amharic. Other than the sign of the Jimma Zone Higher Court, there are many other regional signs around where this University's sign was put. Some of them are bilingual Afan Oromo-Amharic and some of them are multilingual Afan Oromo-Amharic-English. Yet, this contesting graffiti was written only on this federal sign working on the legal issue. Hence, demanding for the law to be respected by the government implies that there is a violation of law by the federal government itself.

Discussion
The sources of the conflicts between the regulating bodies in the municipalities and other bodies who use language on signs such as private business owners, religious institutions and federal government bodies are basically a language exclusion on signs in the linguistic landscape; in other words, working for and against multilingualism has become a cause of contestation. Exclusion in this case is marginalized visibility, or less visibility of languages, particularly Afan Oromo around some domains of language use on signs. The exclusion is the consequence of a top-down unofficial policy and bottom-up practice, which assigns prestige or stigma to particular languages considering themselves as a separate island within the regional government, Oromia.
Negotiating these facts requires careful planning in consultation with all concerned bodies. As officially recognized multilingual country with multiple interests, unless properly managed, "… the public space remains an area of contestation and subversiveness between civil society, private businesses and the state" (Rubdy, 2015:3). This happens due to the regional state's interest in promoting the regional language, and that may be resisted by individuals at grass-roots level.
From the point of view of contestation, Woldemariam and Lanza's (2012) study has shown how language use in different forms manifests conflicts between different denominations of Christianity. And the example presented here manifests conflict more in the form of abstract cultural and political symbols displayed in the form of non-linguistic symbols and iconic signs in the public space. This is how LL serves as a place for displaying the ideas of relationships between people and socio-political issues (Jaworski & Thurlow, 2010:11).
According to Ben-Rafael et al.'s (2006) study in Israel, linguistic landscape items are not necessarily a reflection of Israel's multilingual nature that exists in speech form. The result of the current study is also similar. Languages such as Dawro language, Kafinoono, Gurage language, and Tigrigna which are spoken to some degree in the towns are absent on the signs. The use of English on the signs in the LL of the towns does not reflect the everyday spoken language use of the community in the studied area. The residents are more of bilingual in Amharic and one of other Ethiopian languages, or monolingual Amharic speakers. Hence, the multilingual (Amharic, Afan Oromo and English) signs observed are not an actual representative of a linguistic repertoire of the setting of the study. Therefore, the role of this foreign language's visibility with the local languages on multilingual sign is associated more with people's attitude rather than communication which is "impersonal multilingualism" (Edelman, 2009:142).
It is a fact, however, that the stigmatization of regional language is not carried out explicitly via legislation, but through implicit institutional practices of sign owners. According to Spolsky's (2008:4) terminology, even in the absence of consistent top-down language management, as in the towns studied, language beliefs and language practices have been strongly influenced by what Fishman (1991:383) called the systematic delegitimization of minority languages as anti-modern. That is why local languages, including Amharic have negligible presence in a prestigious business area.
Generally, the LL actors involved in the contest are the CTO workers in the municipalities and the municipal administrators in the towns, the local community or those who own the signs, some invisible groups or individuals and organizations.

Conclusion
Though in this study bilingual sign is treated under multilingualism, when separately seen, the most common signs in the LL are bilingual and not multilingual. Where there is multilingual practice, it is chaotic, that there is no harmony among sign users. And, the languages visible on signs cannot represent the towns' multilingual nature vis-à-vis the various ethno-linguistic groups living in the towns. Therefore, the multilingual nature of the country, though recognized by the federal government, as stated in the constitution, some government institutions have less commitment to tell their presence among the local people in their own language. They communicate and advertise their services and all that they demand from the local people, in Amharic and English.
Multilingual practice is somehow observed among private business owners. Many of them use Afan Oromo and Amharic from the local languages, and additionally they use English. Therefore, duplicating multilingual signs are more frequent, and detail required information is described in Amharic. The religious institutions' practice of multilingualism on signs is also limited. Only few protestant churches practice fragmentary multilingual sign which is incomplete to be called multilingual.
Based on the domains of use, the LLs of the selected towns are bilingual Amharic-English from above-the federal government, and bilingual Afan Oromo-Amharic from middle-the regional government and multilingual Afan Oromo-Amharic-English from below-private business owners.
Besides, there is a clear contestation in the LL that is based on language, history, politics, culture and religion as reflected from the semiotic aggregates in different domains of languages on signs. And the contestation has created dialogical interaction between anonymous individuals and government in the form of graffiti, between CTO regulators and private sign owners, between sign owners themselves and between sign readers and sign owners through advertisement of addiction inducing signs. By way of and due to such contestations, there are multiple identity constructions in the LL, i.e. political, cultural, religious, and linguistic and signs are rarely viewed as a tool of communication.