Mediating Discourses of Displacement in the Literacy Practices of Refugees and Humanitarian Actors in Jordan, Kurdistan Region of Iraq, Lebanon and Turkey

Abstract Much has been written about the prevalence and politicization of public discourses on migration. Less prevalent are analyses of the role of literacy mediators in making texts, and the discourses they invoke, accessible to displaced people. With this in mind, this paper takes a discourse-ethnographic approach to the analysis of discourses of displacement. Literacy mediation in humanitarian settings is explored as a means by which refugees and humanitarian actors negotiate the unequal power dynamics of humanitarian interactions. Findings suggest that the institutional complexities of humanitarian efforts are reduced when mediators translate discourses in their literacy practices.


Introduction
This paper sets out the findings of a study which traces how texts are mediated and discourses appropriated by refugees in interactions with NGOs as part of their everyday literacy practices. Their access to the literacies of these institutions, their ability to mediate these literacies on behalf of others and the cultural brokerage which occur when refugees appropriate the discourses of these institutions are analyzed from a discourse-ethnographic perspective. The reason for this approach is that this journal has documented the power that discourse conveys in the socio-political field and the often privileged access to the production of discourse via media and legislation (Krzyzanowski et al., 2018;Rheindorf & Wodak, 2018, p. 15). This paper seeks to explore humanitarian actors' opportunities to produce discourse and appropriate existing institutional discourses in their interactions with each other, away from the socio-political field. To achieve this, the literacy practices of those whose lives have been uprooted by the conflict in Syria are analyzed. Literacy practices are the ways people use written language in their everyday lives. They involve values and attitudes as well as social relationships (Street, 1993). For example, the literacy practices of the refugees in this study include the reading and writing they do online to navigate journeys and to maintain friendships as well as the value that they give to involving others in these literacy practices. When refugees collaborate with others, the texts they produce in literacy events are imbued with the discourses that circulate within these social and cultural spaces. Literacy practices are the focus of this paper because, in his study of the interplay across supranational and state layers of governance in Lebanon and Jordan, Fakhoury (2019) found that Syria's displacement has intensified collaborative relationships but, due to institutional complexity, this interplay has yielded low effectiveness for the governance of large-scale displacement in Syria's neighborhood. Fakhoury's findings highlight several dilemmas that multi-level governance has entrenched. One of these dilemmas is the result of institutions' discourse on solidarity and rights-based approaches which have not reflected the realities of the need for assistance to and protection of refugees which, Fakhoury claims, amplifies inconsistencies between rhetoric and policy delivery (2019, p. 1311). It is the discourses of NGOs, invoked in the reading and writing that staff and clients do together, that this paper takes as its central concern. Street (1984) and Gee (2015), among many others working from a socio-cultural perspective, have rendered visible issues of power within everyday meaning-making by focusing on literacy practices. Literacy practices are the cultural ways people use written language in their lives. 'In the simplest sense literacy practices are what people do with literacy' (Barton et al., 2000, p. 7). Tracing the literacy practices of humanitarian actors and their clients illuminates the negotiation of institutional discourses because these discourses are invoked in written texts, where specific actions, behaviors and values are communicated through specific literacies (Papen, 2006, p. 10). The cultural contexts in which literacy is used, and the extent to which literacy uses are imbued with the values of displaced people and the humanitarian actors whose work is to support them, become sites for understanding the networks of relationships at work in the humanitarian sector and the collaboration which holds these networks together. In order to better understand these collaborative relationships, the current paper takes a discourse-ethnographic approach to observing literacy events in which refugees and humanitarian staff engage -and in doing so develops a portrait of literacy practices in refugee settings. This conceptualization of discourse is put to use to identify how literacy is used to challenge the very discourses that Wodak, Triandafyllidou and others have shown discriminate against refugees and migrants as part of the wider politicization and mediatization of immigration (Baker et al., 2008;Messer et al., 2012;Reisigl & Wodak, 2000;Wodak & Forchtner, 2014). Seeing texts as specific semiotic realizations of discourse means examining texts and their processes of production, which in turn means exploring how the discourses that texts invoke are mediated by individuals and groups. In other words, these groups' literacy practices illustrate how collaboration takes place across national borders and regional boundaries in the protection of those displaced from Syria.

Literacy sponsorship and mediation
Of central concern in this paper is the manner in which the literacy practices of humanitarian actors shed light on what Fakhoury has called 'divergent sites of authority' (2019, p. 1311) in the Arab region. This is because, Stritzel claims, the Arab region's relational dynamics of power help contextualize these sites even though they are embedded in the complexities of migration politics in the region (2007, p. 345). Fakhoury has argued for an understanding of how these divergent sites of authority have 'shifted down' to the Arab refugee-hosting state, interacted with its power dynamics, and impacted refugee assistance and protection on the ground1 (2019, p. 1311). This paper seeks to understand this 'shifting down' in these power dynamics by asking what role literacy sponsors (with the power to promote or prevent literacy) and mediators (with the power to make it accessible) play in negotiating the relational dynamics of power in their literacy practices. The practices which are drawn on in these divergent sites of authority require an approach which enables the researcher to look across sites as well as identifying the blurring of boundaries between sites. The multi-sited ethnographic approach taken in this paper sees these sites as domains: spaces in which literacy occurs and in which literacy practices are given their shape by individuals and groups interacting with texts in institutional spaces (Barton et al., 2000). In this study these institutions are the national and international NGOs that were part of a broader study exploring language and resilience in the four countries neighboring Syria in 2015-2016 (Capstick & Delaney, 2016). They are the institutions that refugees come into contact with when displaced as it is within these institutional spaces that the relational dynamics of power are negotiated. The study looks at the ways in which institutions with the power to shape literacy both support access to literacy in, for example Arabic or English, or suppress access. Barton et al. suggest that literacy practices are patterned by social institutions and power relations, and some literacies are more dominant, visible and influential than others (2000, p. 12). In order to identify how this happens, we first need to see how institutional actors and their clients access the literacies that enable them to negotiate their displacement (part 1 of the analysis below). Access here is characterized by Judith Kalman's work in which she argues that using literacy in specific contexts means being able to respond to the specific requirements of participation because every literacy practice is shaped to fit the context in which it is employed. Kalman has suggested that contexts include physical spaces as well as the social conduct which is expected in them (2005). Literacy sponsors, Brandt has suggested, mediate this social conduct (2001). She suggests that literacy is part of broader social systems which confer value on reading and writing by the sponsors who withhold or regulate literacy in some way. Exploring literacy sponsorship, I have argued elsewhere (Capstick, 2016a), within the context of the access and availability of literacy, connects the relationship between people and the institutions which shape their literacy. This means taking account of 'any agents, local or distant, concrete or abstract, who enable, support, teach and model, as well as recruit, regulate or withhold literacyand gain advantage by it in some way' (Brandt, 2001, p. 27) in order to relate literacy sponsorship to the mediation of literacy on the ground. Baynham has suggested that a literacy mediator is 'a person who makes his or her literacy skills available to others, on a formal or informal basis, for them to accomplish specific literacy purposes ' (1995, p. 39), while later adding that the roles of translator of spoken language and literacy mediator in multilingual encounters can overlap (Baynham & Masing, 2000). For Baynham and Masing, in encounters such as these, literacy mediation means not only codeswitching between languages to assist those who are unfamiliar with those languages, but also switching between oral, written and visual modes. The literacy mediator therefore translates between codes (e.g. Kurdish and English) and modes (e.g. moving between online and offline) when reading, writing and speaking on behalf of others. Elsewhere (Capstick, 2016a(Capstick, , 2016b) I have discussed the link between encounters where migrants turn to literacy mediators for help with unfamiliar codes and modes and the sponsorship of literacy in everyday life. In the current study, the focus is on the forms of literacy mediation which take place in humanitarian settings with a view to establishing how the mismatch between institutional discourses and the lived experiences of displaced people are negotiated when discourses are invoked in written texts. Robins (1996) has used the term cultural brokerage to describe what happens when the power relations between dominant and non-dominant groups are asymmetrical and the discourses invoked by the former are unfamiliar to the latter. In Parts 2 and 3 of the analysis which follows, when the talk moves far away from an implicit text toward the discourses the texts invoke, the term cultural brokerage will be employed. This is because a cultural broker is able to translate dominant discourses relating to, for example, donor funding regimes, while a literacy mediator may be good at filling in forms with clients but not able to straddle the cultural contexts which grants them access to multiple discourses about funding, protection or camp management.

Methodology and data
This section discusses the approach to data collection and analysis taken in this study and the uses of fieldwork strategies in discourse-ethnographic research are explained. This ethnographically-informed approach combines linguistic and ethnographic perspectives with linguistic anthropology (Hymes, 1980) and focuses on the ways that texts (for example camp registration forms) are disseminated by, and implemented in, the work of non-governmental organizations. The paper takes the view this is now common in discourse studies that an analysis of linguistic elements of discourse presupposes a proper appraisal of their context of production and reception (Blommaert, 2008;Panagl & Wodak, 2004;Van Dijk, 1984). In order to carry out an analysis of these contexts of production and reception a multi-sited approach was taken given that displaced people and the humanitarian actors they communicate with come into contact in rapidly changing settings across multiple national borders and complex migration trajectories. The study emerged from a large-scale study exploring the language use and language learning needs of refugees and migrants in Iraq, Jordan, Lebanon and Turkey (Capstick & Delaney, 2016). From the original 47 governmental and non-governmental (NGO) agencies and schools that were part of the original research project, the data from the 16 NGOs from this wider study were selected for analysis for the current study. This meant re-coding and re-analyzing interview transcripts and fieldnotes made during visits to these NGOs including the notes taken during NGO staff meetings; meetings between beneficiaries and staff, volunteers and beneficiaries; and the participant observation notes taken during the four months of the original study in 2015 and 2016. 23 interview transcripts of interviews with NGO staff (paid and voluntary) and beneficiaries from across these 16 NGOs were re-coded based on the new research questions. The two overarching research questions which were formulated during FGDs with the NGOS were: what language and literacy practices do NGO staff and beneficiaries draw on in their day-to-day work and what literacy mediation takes places within these NGO settings? The following table describes the procedures which were followed and summarizes how ethnographic data were collected and analyzed in the present study: The objective of the current study was to identify how discursive processes play out in the literacy practices of NGO staff and beneficiaries. After the initial identification of the research questions in collaboration with the NGOs (Phase 1), in order to put the ethnography to work for studying the ways in which language is articulated and realized in these literacy practices, content analysis of interview data and fieldnotes was carried out to identify who sponsored literacy in these settings (Phase 2). At the center of the research were the interviews in which participants described their collaborative role with others in literacy events. However, these textual practices also extended to the 'informal' help that refugees offered those in their networks as there was never a clear delineation between where their formal role of offering advice overlapped with their informal literacy practices. Interviews with participants about these practices and the detailed content analyses of the transcripts (Phase 3) of these interviews were informed by ethnographic observation of meetings, interviews with beneficiaries, and NGO activities. This approach enabled the analysis of the literacy mediation that occurred in the NGOs as well as enabling the researcher to select the right documents for close linguistic analysis (Phase 4). The first step in Phase 4 was the selection of data for analysis of literacy mediation. This was carried out by identifying the discourse topics relating to displacement and the mediation of texts in the 32 interview transcripts. Following Krzy_ zanowski (2008), I define the basic analytic category 'discourse topic1 as 'expressed by several sentences in discourse … by larger segments of the discourse or by the discourse as a whole' (Van Dijk, 1984, p. 56). In this sense, a discourse topic is defined as the salient theme or idea that underlies the meaning of a series of sentences. Discourse topics therefore organize the interviews thematically. In the current study, the topics were addressed in different ways by the participants during semi-structured interviews. These interviews were primarily framed by questions about the participants' displacement. Thus, all the discourse topics relate to displacement, which I will call the macro-topic. In the coding of the data I identified parts of the interviews where the participants talked about the reading and writing that went on in their lives and in the NGO. This meant that I was able to identify sub-topics where the participants talked about displacement (the macro-topic) in relation to reading and writing (the subtopics). These sub-topics were navigating journeys; reciprocity; trust and ideology. These parts of the interviews were then analyzed for what the participants said about literacy mediation. In other words, sections of the interviews were selected where participants spoke in the greatest detail about reading and writing with others relating to displacement. In doing so I was working quantitatively by looking at the number of references to reading and writing within the context of displacement. This meant counting the statistical frequency with which participants referred to reading or writing. The macro-topic of displacement was framed by me as the interviewer, as were my questions about reading and writing whereafter, the sub-topics were 'put forth by the participants themselves' (Krzy_ zanowski, 2008, p. 174) in the sense that they offered up discussion of navigating journeys; reciprocity; trust and ideology in response to the semi-structured questions. In this way I was able to narrow down the data and focus on those sections analyzed below which relate to these sub-topics. For example, in interviews with refugee coaches, common positions about how to make a successful scholarship application were reflected in the observational data where I saw NGO staff sitting with beneficiaries filling in forms together. Such literacy mediation occurred often in the observational field notes of interactions between staff and beneficiaries and help explain the role of scholarships in improving refugees' lives. These ethnographic fieldnotes complemented the interview data rather than being used as a primary focus. In this way it was possible to examine different types of discourses produced and received within different NGO spaces and move toward an analysis of how literacy mediation was shaped by the production and appropriation of these discourses. Intertextual and interdiscursive relationships were explored in this final phase of the linguistic analysis with a specific focus on recontextualization. Recontextualisation is where the main arguments are transferred from one text to another and hence acquire new meaning in new contexts while interdiscursivity hinges on the notion that discourses are linked to each other in texts (Wodak, 2008).
The participants who appear in the analysis sections which follow were refugees, IDPs and host community though the boundaries around all three categories were not clear-cut, nor were boundaries around their roles in the NGOsome refugees were paid as 'coaches', some were unpaid volunteers but assisted on projects and some were beneficiaries. Pseudonyms have been created for the participants who were interviewed and observed. Interpreters and translators have helped with the data collection and transcription of data from across these different fields.

Part 1: Sponsorship of literacy in humanitarian settings
This section examines the sponsorship of literacy in humanitarian settings in the four countries and the opportunities that displaced people have for participating in reading and writing activities which I will characterize as access to literacy (Kalman, 2005). Using literacy in specific contexts means learning to respond to the specific requirements of participation. Each practice is shaped to fit the social context in which it is employed. Institutions are part of the larger social systems which confer value on reading and writing. In the following examples from an NGO in Turkey we see how literacy in English is sponsored by supra-national agencies in humanitarian settings.

Example 1
We have a field officer in each area. So we rely a lot on the information that we get from them. We subscribe to several NGO security information websites that will provide and feed us with constant conflict management data. It's all in English. I don1t know of any NGO security fora in Arabic. It might exist … All the fora that I know of are English-speaking -because they involve INGOs as well; they don't only involve Syrian bodies. Like, I tell people, you1re going to have to get used to the acronyms that they use for military activity.
This example illustrates the sponsorship of literacy in standard varieties of Arabic and English which were used in official and unofficial domains in humanitarian settings. Thus access to literacy in English is required for security information, conflict management data as well as in the military jargon used in these settings. In Example 2 from Jordan, the participant has just explained that very little English was used inside Syria prior to the conflict but that in the UN system, literacy in English was sponsored: Example 2 But still they developed their English skills in Syria. Two of them were working for UN agriculture organization that was based in Aleppo and therefore they had to use English because that was the language of communication within the UN structure.
There were many examples such as this in the data, where participants described how sponsors of literacy in English provided access to written material in English for those humanitarian actors who can read and write in English. The varieties of English included Standard British English and Standard American English though other varieties of English were also present in the data such as post-colonial Englishes from the Asian sub-continent. The analysis revealed how sponsorship of literacy in English is related to those opportunities to use and practise English in its written form in the reporting, evaluation and communications of many small, medium and large-scale agencies. Kalman argues that it is this availability of printed matter which influences how opportunities to access reading and writing practices are constituted and how, in turn, these opportunities facilitate the availability of printed matter. In the humanitarian context, written culture is not automatically accessed by the mere physical presence of written materials, since texts in English were available, but not everyone was able to read them in the same way, or in some cases read them at all.
The analysis here provided a link between the sponsorship of literacy at the macro, meaning institutional sponsor, and micro, meaning individual sponsor, levels. For example, humanitarian staff explained how the sponsorship of literacy depended on their colleagues' ability to use the language. Literacy in English practised in the offices of NGOs is related to the sponsorship of literacy by the wider humanitarian sector. When international humanitarian agencies promoted the availability of literacy in English through reporting, evaluation and other written material in English, but simultaneously restricted access to literacy in English by assuming all staff are already able to read and write in English, staff found alternative ways to help each other. This help often came in the form of literacy mediators.

Part 2: Analysis of literacy mediation in NGO settings
This section provides the findings from the analysis of the literacy mediation which relates to the four discourse sub-topics described in the previous methodology section.
Discourse Sub-topic 1: Literacy mediation for navigating journeys When orienting to this discourse, the participants spoke about how they used reading and writing collaboratively to navigate their journeys. Negotiating the written material of digital infrastructure of their migration trajectories was often as important as negotiating the physical infrastructure of their displacement. Similarly, mobility in the language and literacy practices of the literacy mediator was as important as the portability of the technology. That is, part and parcel of navigating websites, news sources, social media and mobile apps that participants described were the literacy practices that participants drew on to make sense of these technologies. Several participants in the study commented on how their literacy practices changed as the affordances of a smartphone allowed them to access the internet outside the home and take photographs which they introduced easily into their communications by asking and showing others. In the following example, Fateh makes his online literacies available to others when he explains how he showed them.
Fateh: It's all done together even if you're alone later, you show it, write it, someone else tries it and you know it's passed on like this. … like my pics, I used to show everyone. It's the best sites then it's the best way from this and that. If we didn't share it they might not know.
Fateh explained how, over time, this feature of Facebook was also used for navigating journeys. Many participants described how the smartphone was an essential tool as well as a potential danger when navigating journeys together. As a tool for literacy mediation, the smartphones enabled access to translation tools as well as access to medical and legal services and support networksincluding the services of NGOs. They also provided essential communications with friends and family, which was a core aspect of literacy mediation. However, smartphones were also seen as a potential threat by some literacy mediators because they felt that the digital traces that were left in their wake increased the vulnerability of displaced people. Surveillance by threatening groups was noted by one participant: Abdallah: you don't know who can get hold of your number, which violence groups and all, really, you don't know what these guys find on your phone This participant was more covert in his literacy mediation as a result. These threats heighten the significance of literacy mediators particularly when the digital literacies of those they mediate on behalf of are low or partial.
Discourse Sub-topic 2: Literacy mediation as a form of reciprocity Many of the participants in this study that described some form of literacy mediation oriented to a discourse of reciprocity. Kamal in Turkey explained how he consumed increasing amounts of news posted in Facebook groups and would look at news stories on his phone and on his laptop in the company of others, commenting on these stories in his words 1to help the others on their ways1.
Kamal: When we're looking at it [the news story] we're doing it to help the others on their ways 'cause you know they're gonna do it for you later Tony: In what way do it for you? Kamal: They share stuff, they make guidance or they do something totally different … like helping out on accommodation This can be seen as one of the many forms of reciprocity that literacy mediators reflected on. Kamal's Tweets, which he presented on his smartphone, used Arabic as well as English though he explained that for international Twitter accounts literacy in English was required. In this way Kamal is both individual sponsor of literacy in English when he tweets in English but also literacy mediator when making these tweets accessible to those in his network that have low literacy in English. People with low levels of formal education have been seen to draw from resources in their immediate surroundings in order to overcome difficulties with specific texts. Fingeret (1983) demonstrated how people with low literacy tap into their existing social networks for literacy skills which they themselves do not possess but which others are able to provide. The reciprocal arrangements in these networks mean that in return for help with a literacy task, other services may be provided in return. Reciprocity in these relations means that traditional boundaries between 'online' and 'offline' are no longer helpful as literacy mediators help family, friends and clients across domains, contexts and countries. Literacy here is a shared resource which exists for members of the group who rely on others in order to cope with certain literacy demands of being displaced. For example, Khalid in Iraq explained that he shared access to international news sources from his smartphone by sitting with friends and translating the news in English into Arabic and Kurdish.
Khalid: translation on Google is so easy but they didn't know it, or they know it but they didn't practice, so we sit and we show them with news and Facebook of course In these instances, Khalid acted as a literacy mediator in that he code-switched between Arabic and English and mode-switched between oral and written. He explained that the English of those he mediated for 'improved, little by little, till they did it alone'. Khalid also explained that he would often read the stories before meeting up with other refugees. It seemed that he was preparing in private for what would then be shared publicly.

Discourse Sub-topic 3: Literacy mediation and trust
The discourse of reciprocity was closely related in the analysis to a discourse of trust. Producing and consuming social media news was part of the reciprocity across the social network but several literacy mediators also explained that they preferred Facebook to Twitter when mediating information about routes because they felt that they had more control over personal information. Mohammed (Lebanon) presented the news feed apps which he most frequently used with his friends. Here Mohammed explains how he avoids mainstream national news sources and relies on sources that had come to him recommended by others.
Mohammed He added that he used Twitter in a similar way though was concerned to show how he only followed trusted people on Twitter. He explained that he considered those who were 1close to the ground1 in Syria as the most reliable. He explained how he would often start to follow other followers of popular Twitter feeds, sponsoring literacy in Arabic and English in turn as he mediated for others using both of these languages.
Similarly, trust was an important aspect of literacy mediation where mediators feared reprisals from those that they felt may be monitoring social media. Kamal in Turkey feared for his family when he mentioned not wanting to bring harm to those he mediated for inside Syria with whom he used similar social media sites for communication.
Kamal: I always take care of my won and for my family inside Syria. Assad regime can know everything so its my job to take care of my posts and not make mistakes in different places.
The analysis of the interview data here revealed that of those research participants who reported regularly acting as literacy mediators, many were of the belief that they had a responsibility to protect their own and others when mediating in online spaces. A small number of participants working in humanitarian agencies reported that when they did not feel that they had personal security to act as literacy mediators. Using pseudonyms was not enough protection for those that reported this aspect of literacy mediation. Literacy mediation, it seems, becomes imbued with power imbalances if those involved are unsure whether to trust those in their networks.

Discourse Sub-topic 4: Literacy mediators and ideology
The analysis revealed that literacy mediation extends to negotiating ideological positions. Many physical and social media spaces are politicized which has ramifications for the work of the literacy mediator. Participants in all four countries explained that literacy mediation involved mediating different ideological positions and stances. An example from Khan in Jordan suggested that when he explained the details of NGO programmes to those in his network, it was difficult for him to find the words which relayed the original stance of those posting online if that stance was different to his own. The reason for the difficulties, he explained, was that those that he acted as literacy mediator for often held ideological positions which were different to his own. He explained that he preferred to mediate for those with whom he was able to connect with ideologically as well as socially. Again, this was related to trust and a feeling that literacy mediation was simpler when the mediator shared values or an ideological position with those he was mediating on behalf of.
It's easier when you know them well and when you agree with their point. It gets tough, you know, when you don't … we mostly agree but sometimes it's tough to know what to write when there's one or two who fight and you're in the middle of this to try explain it … clearly … but these guys just don't have it, so I say leave it However, it must also be said that for many of those acting as literacy mediators, the social networks of which they were part provided them with opportunities to 1try out1 a range of ideological positions. For example, one participant reported a literacy event where he had been asked by friends to explain funding opportunities for an NGO which provide scholarships for refugees. The institution offering these scholarships was known to have previously been involved in work which was seen to be biased toward refugees from a particular region of Syria. The literacy mediator was conflicted about his own position on this. He described the value in trying out different perspectives when reading and writing on behalf of others but he also recognized that he was not acting as a faithful scribe when doing this. When sharing their literacy skills, mediators were not only accessing information about Syria or mediating information about migration and refugee issues, they were also negotiating the informal power dynamics and hierarchies of the ideological stance of those in positions of power. Analyzing this discourse topic illustrates how literacy mediators who are not refugees themselves reinforce and maintain the ideological positions of the original message and are less likely than literacy mediators who are humanitarian actors to challenge or dispute the ideological stance in the literacies that they mediate. One participant explained that he was willing to 'let opportunities pass by' if the literacy event he was involved in required him to mediate an ideological position which was different to his own. However, it should be noted that this was not a common feature of literacy mediation and in some sense counters the discourse of reciprocity discussed earlier. It seems from the interview data in these cases that a lack of trust was also an important aspect of this reluctance to mediate in these situations. Where shared ideological stance was low, but trust was high, participants appeared to be willing to act as faithful literacy mediators. However, when trust was low and shared ideological stance was low, the literacy mediators did not feel compelled to be faithful transcribers or scribes. It is on this point that ideological stance is related to cultural and social practices as ideological positioning did not always emerge as separate to literacy mediators' cultural and social practices. This is a common observation in the social practice paradigm, where reading and writing appear highly socially and culturally contextual, interwoven into local ways of life, and sensitive to ideological complexities of time and place (Barton & Hamilton, 1998;Besnier, 1995;Heath, 1983;Scribner & Cole, 1981).
Analysis part 3: Cultural brokerage and straddling different cultural contexts As we saw in the analysis above, literacy mediation has been seen as 'a process that can challenge the power of dominant literacies and discourses by allowing those not commonly familiar with these practices -via a mediator -to access and deploy them for their own needs' (Baynham & Masing, 2000, p. 79). The aim of this section of the analysis is to establish how far the practices of reading and writing and the practices of challenging dominant discourses are accessed and deployed when navigating the mismatch between institutional discourses and the realities of displaced people's lived experiences.
Vignette 1. This vignette is based on interviews and ethnographic observations at one NGO in Gazientep Turkey. Their workshops which run inside Syria support initiatives and organizations which were operating in five provinces of Syria. Since the border between Turkey and Syria had closed they had identified trainers who were able to deliver training on the ground. The main respondent is the Office Manager, Ayman. The organization sponsors literacy in Arabic and English.
Ayman showed me conversations which he had conducted on his Facebook pages where he and his colleagues discussed the materials design for organizational management trainings delivered inside Syria but designed in the office in Turkey. He explained that he had first learned about organizational management while working for a US business in Lebanon. He also described his computer-based literacies, learning informally by searching for materials on the internet. He spoke most frequently of teaching himself digital literacies related to project presentations and reports. I concluded that this access to literacy in English for organizational management had emerged from a combination of his existing English language writing skills which he put to use interacting with others online. Literacy practices from across humanitarian settings and the wider NGO networks converged in the literacy mediation of materials design. NGO staff and beneficiaries brought with them different literacy practices which mingled in the NGO space as each domain generated and spread other literacy practices across the network. Whereas Kalman argues that members of different generations take up new opportunities to participate in reading and writing events and to learn new literacy practices, the staff in this NGO scaffolded each other's literacy when they produced training materials in Turkey for delivery inside Syria, thereby 'creating literacy-generating situations where different practices are displayed and appropriated by participants' (2005, p. 40). Ayman described how he sat together with colleagues in the office in Turkey using English and colleagues online in the office in Syria using Arabic:because the concepts … Especially when we talk about specific topics like project cycle management and like financial management across the border, the concepts need a certain educational background, and I don't know how much they can do. They can do the basics. So we can introduce concepts like cashbooks, for example … We just tell them this is the tool and respond … this is how you use it by talking about a concept in Arabic to a concept in English like cashbooks … In this example, Ayman appears to act as literacy mediator as he describes code-switching and mode-switching but the extent to which he brokers concepts that draw on the specific cultural knowledge of those he mediates for is not clear. This is largely due to the diversity of different participants in the materials design here (NGO refugee staff, NGO non-refugee staff and Syrian trainers on the ground in Syria). Using a mixture of standard and nonstandard varieties of Arabic, NGO talk about the materials design using chat on Skype meant that the different varieties of Arabic and English were used by staff, each with their own conventions, and each opening up access to new literacy practices and new opportunities for collaboration. Ayman explained that he drew on his knowledge of the concepts of project management itself as well as his knowledge of funding requirements as he explained how important it was to 'get the right kind of training delivered to please the funder as well as the beneficiary.' Ayman described the long hours of information-seeking about funding as part of that mediation. He explained that this included using search engines and becoming familiar with technical language, as he often found himself on websites in different languages which were intended for finance or human resource management practitioners as well as websites related to humanitarian funding which he said the trainings needed to align with: English is the prime … Especially most of the programmes … I mean, Syria being under a French mandate means that the Syrian law, Syrian civil structures and Syrian administrative structures are very close to the French system. But most of the support programmes that have been implemented in areas that are outside the control of the regime are funded by Anglo-Saxon actorsthe United States, United Kingdom mainly. So the administrative structures that are coming along, the concepts, etc., are all English. Therefore, you cannot … I mean, you can1t do French. And then we1ve told them that a lot of times. Now, once peace comes in, how are you going to bring those two together again will be a challenge … Ayman's mediation is built on his understanding of the different cultural practicesdominant and non-dominant -that have emerged in the region since the colonial period when Syria was under French mandate. His understanding of the institutional discourse of humanitarian is shaped by this knowledge of different cultural norms and practices. Wodak (1996) found that in institutional discourse, those entering the institution from outside are unable to act on their own initiative and must react to the information received from insiders. This form of discursive control, of who has access to institutional discourses, can be seen in Ayman's organizational management literacy practices. Ayman learned to overcome this discursive control as he learnt to develop his understanding of different 'administrative structures' and concepts. An example of this at the micro-level of literacy mediation is an example he gives of moving between concepts in English, French and Arabic to explain the concept of 'evaluation' in English.
This (evaluation) is a difficult one for Syrians that know some French but not English. Strictly you could say 1un entretien annuel d1 evaluation but people just say 1 evaluation1 when they want to check the performance of someone against something. (taqdir) will be suitable to provide this meaning but it isn't the sort of meaning funders want to see in any application.
Ayman gave the impression that he was able to help those he mediated for in Syrian by drawing on his cultural knowledge of 1evaluation1 in French and taqdir in Arabic by giving examples that form a bridge between the different cultural meanings of the two concepts.
Ayman, as a literacy mediator, worked with colleagues to explain the English information that he found on the internet but also drew on his knowledge of the discourses of humanitarian funding when he explained 1but this isn't the sort of meaning that funders want to see in applications1. He developed the ability to translate discourses and make them available to others in the NGO, thereby extending their ability to cope with humanitarian funding regimes. Literacy mediators like Ayman not only need to understand the concepts of organizational management but also need to be able to negotiate what Wodak calls 1power registers1. Wodak describes these power registers as the linguistic behavior, or symbolic capital in Bourdieu's terms, of the powerful elite which is invested in knowledge expressed in specific institutional genres (1996, p. 40). Ayman negotiated the power registers of institutional discourses which he was able to do through developing his literacy practices. In the case of humanitarian funding, the asymmetrical power relations demand that Ayman had first to find, collect and collate the various documents which make up the entire training materials before crossreferencing these materials with the ever-changing funding regimes. Ayman needed to engage with non-dominant institutions, such as refugees in his social media network, about issues that they felt they most needed help with, before setting to work on reading through the source material from other NGOs which he found online. Reading official documents was central to this process, but what also emerged from the analysis was Ayman's ability to understand and help others understand dominant and non-dominant discourses. This social power is what Wodak, drawing from Foucault (1984) and Van Dijk (1984), understands as 1discursive control: who has access to the various types of discourse, who can and cannot talk to whom, in which situations, and about what1 (1996, pp. 65-66). In Ayman's case this social power comes from his ability to ask questions of official documents and access discourses about funding and training as part of his developing organizational management literacy practices.
Vignette 2. This vignette is drawn from interviews and ethnographic observation in an NGO in the Kurdistan region of Iraq. The interviewee, Kahlid, is a refugee member of staff responsible for helping newly arrived IDPs complete their camp registration forms.
You will find conflict between the people, especially they're from different areas and conflict on resources. They tell us 'we have no water' or 'we have no enough food for staying alive'. There is lack of facilitieswash facilities. I mean, they call it lack, but we call it, this is the minimum standards that we should be providing for the IDPs. But for them, you know, most of them … many of them they have their own houses before, but now their situation has changed. Especially in Kirkuk they have no freedom outside the camp so we write on the forms 'what is required is food distribution, wash facilities, a network of kerosene distribution … .' In this extract, Khalid orients to both the non-dominant discourse of 'staying alive' and the dominant discourse of 'minimum standards'. In order to position himself within these competing discourses, Khalid begins by explaining what the beneficiaries tell him about the privations they face. His stance is conflicted when he moves between the in-group of beneficiaries (he is a refugee too) and NGO staff. Firstly, his use of the pronoun 'we' signals membership of the in-group of refugees. This is later evidenced by his understanding of the beneficiaries' housing situation prior to displacement and his use of the out-group 'they' to describe officials. However, shifts in the talk occur when he includes himself in this out-group of official staff 'so we write on the forms'. His stance on beneficiaries' needs is recontextualized here at the end of the extract when he reformulates the wording on the form from the words of the beneficiaries, 'we have no water' 'we have no enough food' to 'what is required is food distribution, wash facilities, a network of kerosene distribution' which invokes the dominant discourse and the power register of 'minimum standards.' This is also an example of interdiscursivity as Khalid is able to draw on discourses about minimum standards when filling in registration forms. Khalid's understanding of the institutional literacy practices of the NGO make him both literacy mediator and cultural broker in this instance can be seen to straddle both dominant and non-dominant cultural contexts.

Conclusion
With increasing securitization of migration (Bigo, 2002;Scheibelhofer, 2012), mediatization and politicization of the refugee crisis in Europe (Krzyanowski et al., 2018), and an increasing shift to new types of discourses on immigration which draw on discriminatory rhetoric (Krzyanowski & Wodak, 2011), this paper sought to employ a discourse-analytic perspective to refugees' own production and translation of discourses. The methodological apparatus that was assembled to do this drew from a literacy practices approach as well as work in the discourse-ethnographic tradition. By establishing the sponsorship of literacy in these settings, the first part of this paper identified how many humanitarian actors sponsor literacy in English and Arabic, making access to these literacies difficult for those who have low or partial literacy skills in Arabic or English. In the second part, the analysis focused on how these refugees turned to literacy mediators for help and set out the discourses that these literacy mediators oriented to when describing their mediation of various literacy practices in relation to displacement. These were discourse sub-topics about navigating journeys, reciprocity, trust and ideological positioning in literacy mediation which all related to the macro-topic of displacement. Through this analysis, relationships on social media and in office spaces in humanitarian settings were seen to converge in the literacy practices of mediators who often overcome the challenge of conflicting ideological positions by developing trusting and reciprocal relationships with others.
In the third part of the analysis, we saw how one humanitarian actor, Ayman, sponsored literacy in English while at the same time mediating this literacy for others while giving his colleagues opportunities to use their spoken Arabic when writing training materials across national borders. Similarly, Khalid in vignette 2 was able to translate discourses from non-dominant to dominant cultural contexts when helping IDPs complete their registration forms. Barton and Hamilton (1998) argue that individuals move in and out of different domains and occupy the borderlands between them. For displaced people and the humanitarian staff who support them, NGOs becomes places where different aspects of life are negotiated and fitted in with each other to make institutional discourses accessible to clients. In this process new, hybrid practices are produced, such as those Ayman and Khalid develop when drawing on their knowledge of the discourse of humanitarian funding when mediating the design of training materials and the registration of IDPs. The demands for literacy are resolved through collective practices (Kalman, 2005) with each participant providing support for other members across their transnational networks.
Given that NGO staff work across national boundaries, displacement gives space for literacy mediators and cultural brokers to provide links between different social and cultural contexts. In order for cultural brokers to appropriate discourses it was important for them to employ concepts which focus on the translation of discourses as well as the translation of language varieties such as English, Arabic and Turkish. Robins' term 'cultural brokers' was useful here as it refers to mediators who provide access to registers and discourses for those individuals who struggle with complex legal and bureaucratic literacy practices (1996). When the emphasis was not on reading, writing or the translation of spoken and written language but rather on the translation of discourses, the term cultural brokerage was useful when marking out the translation of discourses about humanitarian organizational management and funding. In other words, in a climate where Syria's displacement has intensified collaborative relationships but, due to institutional complexity, has yielded low effectiveness for the governance of large-scale displacement in Syria's neighborhood, the findings of the current paper suggest that cultural brokers are able to make this institutional complexity accessible to others when translating discourses in their literacy practices, thus narrowing the inconsistencies between rhetoric and policy delivery.