The Map and the Territory: The Use of Country Information in Asylum Assessments

When determining who should be accepted as a refugee, decision-makers use information about asylum-seekers’ home countries to assess the credibility of the claim and the risk of future persecution. As such, country information plays a decisive role in the outcome of asylum claims. Based on asylum case files and interviews with decision-makers in Norway, I investigate the use of country information in the refugee status determination process and compare the specific pieces of country information that decision-makers used in their assessments to landmarks on maps. Landmarks here are understood as decision-makers’ interpretations about places, customs, and political and social conditions in asylum-seekers’ home countries. To come across as credible, applicants had to demonstrate knowledge of landmarks familiar to decision-makers, but they also needed to present a story that testified to their personal experience with the landscape in their home countries. Minor deviations from the landmarks could undermine a claim’s credibility. The metaphor of the map as a seemingly objective representation of reality illustrates the authority of country information in the refugee status determination process. As I demonstrate, however, decision-makers based their knowledge of such landmarks not only on formal sources of information, but also on the narratives of other applicants, assumptions about rational behavior, and their own everyday experience with places. In line with the legal mandate to produce a binary decision, decision-makers had to consolidate uncertain information into solid landmarks that enabled them to clearly distinguish between refugees and non-refugees. Because of their important role in enabling such distinctions, landmarks are key in refugee protection on the one hand and migration control on the other.


Introduction
How does one determine who is a refugee? When an asylum-seeker arrives in a European country and submits an application for asylum, she places her life in the hands of an employee in a public bureaucracy who must make a decision that ultimately can be a matter of life and death. A crucial element in this decision is information about the applicant's home country (hereafter "country information"). Country information can be about the human-rights situation in the applicant's home country, as well as about that country's social, political and economic conditions, religious and cultural practices, history, languages and geography (Accord 2006).
Country information has two central functions in the process of refugee status determination. First, decision-makers use such information to determine the risk of future persecution (i.e., whether the applicant has "a well-founded fear of being persecuted for reasons of race, religion, nationality, membership of a particular social group or political opinion"; United Nations 1951, 14). Second, country information often plays an important role in assessing the claim's credibility. This assessment of credibility has an internal and an external dimension (Good 2015, 122). Internal credibility is closely connected to decision-makers' perception of the claim's authenticity and the applicant's "capacity to give an account in a selfexperienced manner" (Johannesson 2012, 72). External credibility, on the other hand, refers to whether the applicant's story is consistent with the "generally known facts" about the asylum-seeker's home country (UNHCR 1998, para 11) -that is, the country information that decision-makers come to establish as "facts" in the decision-making process.
There is limited research on country information in the refugee status determination process. Most knowledge on the topic comes from the "gray literature"reports from non-governmental organizations and other stakeholders, particularly in the UK (e.g., Amnesty International UK 2004; Williams et al. 2009;Tsangarides 2010; Independent Chief Inspector of the UK Border Agency 2011; Stern 2013; Independent Chief Inspector of Borders and Immigration 2017). Among other things, these reports note that asylum decision-makers often use country information selectively and inconsistently; that country information is not sufficiently referenced or up to date; and that the interpretation of country information too often is guided by policy interests.
In recent years, some scholars have investigated the role of country information in French and British asylum procedures (Gibb and Good 2013) and its use in collaboration and policy-making in the European Union (EU) (Engelmann 2015). Other studies have investigated the production of country information (Rosset and Liodden 2015;Rosset 2019; van der Kist, Djistelbloem, and de Goede 2019), as well as the different ways in which the producers of country information bolster the public legitimacy of such knowledge (van der Kist and Rosset 2020). The most notable scholarly contribution on the topic is Good's study of anthropologists who act as country experts in asylum cases in UK courts (2007,2015). Good describes significant differences between anthropological and legal approaches to the notion of "facts." While the legal approach requires expert witnesses to provide clear-cut facts that enable yes/no answers, anthropologists tend to view facts as provisional, contested, and contingent on theory (Good 2007(Good , 2015. As I return to later in this article, the legal approach to facts, with its requirement that decision-makers produce a binary outcome of accepting or rejecting a claim, constitutes the dominant framework in the Norwegian asylum bureaucracy. None of the scholarly studies have, however, looked into the use of country information in asylum decisions or conducted interviews with asylum decisionmakers, both of which could produce a finer-grained understanding of how country information functions in legal context characterized by high stakes and limited evidence. This article addresses this knowledge gap by examining the use of country information in asylum decisions in the Norwegian Directorate of Immigration (hereafter "UDI") -more specifically, in decisions about asylum-seekers from Somalia and Iran. In the article, I explore the following questions: What type of country information do asylum decision-makers use to corroborate their decisions? What is the relationship between formal sources of country information, the interpretation of decision-makers, and applicant stories?
This article contributes to the field of migration studies in two ways. First, it presents empirical knowledge and theoretical insights about a topic that so far has received limited attention -namely, the role of country information in the refugee status determination process. In the analysis, I refer to asylum decision-makers' knowledge about applicants' home countries as a type of mental map and to the specific pieces of information they use to distinguish refugees from non-refugees as landmarks on this map. The term "landmark" sometimes references geographical information in a literal sense, such as features in the landscape, places, and distances. At other times, I use the term metaphorically to reference other kind of country information that plays an important role in justifying asylum decisions, such as information about the way the authorities in Iran act in specific situations. These landmarks carry the kind of authority often ascribed to maps, as objective depictions of reality (Duncan and Ley 1993). In reality, however, there is often a precarious link between such information and realities on the ground in asylum-seekers' home countries.
Second, the article illustrates both the important role of country information in the migration control system and the uncertainty involved in interpreting and using such information. This uncertainty is rarely visible in public debates where the distinction between refugees and other migrants is usually portrayed as clear cut (e.g., Erdal and Oeppen 2013). Thus, the insights of this article can be relevant for practitioners in the field of asylum, such as producers of country information, decision-makers, and asylum-seekers and their legal representatives, who in many instances deal with rejections based on specific interpretations of country information.
Using Norway as an empirical case begs the question of whether the insights presented here are relevant to other contexts. The institutional set-up of the refugee status determination process differs across European countries (Gill and Good 2019). For example, the mandate and scope of units that compile and/or produce country information vary (Rosset 2019, 13-17), as do appeals proceedings and legal traditions (Craig 2019). As I return to in the discussion, some of the institutional differences have implications for the interpretation and use of country information that matter for the possibility to generalize this article's findings. Nonetheless, studies from different national contexts illustrate that the dilemmas first-instance decision-makers in Western countries encounter in making asylum decisions are strikingly similar (e.g., Thomas 2007Thomas , 2011Dahlvik 2014;Bayrak 2015;Affolter 2017;Miaz 2017). As these studies show, decision-makers conduct their work in a highly contested and politicized field characterized by time pressure, limited resources, and ambiguous goals. They must make distinctions between refugees and non-refugees on the basis of uncertain information and in a context where they rarely receive reliable feedback about the accuracy of their decisions (Thomas 2011, 46-47). In particular, asylum decision-makers must negotiate two seemingly contradictory goals of providing protection for refugees and controlling and limiting migration, thereby navigating a constant "tension between being either unduly mistrustful or unduly lenient" (Thomas 2007, 56). The similarity of these basic decision-making conditions means that the empirical findings presented here are likely to be of relevance beyond Norway.
The rest of the article proceeds in the following manner. First, I present the data material and briefly outline Norway's refugee status determination process. Second, I recount a story from an interview with a decision-maker as an introduction to theoretical perspectives from the field of human geography, which I use to develop the framework on landmarks on maps. Third, I present empirical analysis based on asylum cases from Somalia and Iran, before concluding the article with a discussion of the findings and future research directions.

Research Design
In 2012 and 2013, I conducted 24 interviews with decision-makers in the UDI. All interviewees held master's degrees, most in the social sciences. They had worked in the UDI between three months and 13 years and had experience with asylum applications from 10 different countries. In addition, the data examined here include case files from Iranian (14) and Somalian (15) asylum-seekers. I selected the case files from the Immigration Appeals Board's public archive, where summaries of asylum decisions are published. After selecting the cases, I got access to all documents from the UDI. All chosen files involved cases with substantial doubt about applicants' credibility. The empirical data analyzed here also include some case files made available to researchers on another research project about credibility that the UDI had commissioned in the same time period (see Bollingmo, Skilbrei, and Wessel 2014).
The case files contained a report from the asylum interview and decision letters from both the UDI and the Immigration Appeals Board. Some also contained letters from lawyers, country reports, and health documentation. I took notes about the main arguments made in the claims, decision letters, country information reports, and any other material that corroborated the claim. The main goal here is not to map the country information in use at the time in its entirety but to identify some mechanisms that shaped the interpretation and use of such information more generally. Transcribed interviews and notes from the case files were coded thematically in Nvivo, a program for qualitative data analysis. All data presented here have been translated from Norwegian to English.

The Refugee Status Determination Process in Norway
In Norway, the vital task of assessing asylum claims lies in the hands of decisionmakers at the UDI. The UDI's asylum division consists of several units that specialize in a limited number of countries, enabling decision-makers to develop in-depth knowledge of a certain area. 1 The UDI regularly publishes guidelines for assessing asylum claims from major asylum-producing countries. 2 In the guidelines, they present interpretations of the situation in origin countries and the implications this information has for the legal assessment of the most frequent claims.
If asylum-seekers receive a negative decision from the UDI, they can lodge an appeal to the Immigration Appeals Board. If the Immigration Appeals Board also rejects the claim, applicants can appeal to the regular courts at their own expense. All levels of decision-making in the refugee status determination process in Norwaythe UDI, the Immigration Appeals Board, and the courts -rely on country information from "Landinfo," an independent unit with country experts who compile information, produce country reports, go on fact-finding missions to countries of interest, and serve as expert witnesses at the Immigration Appeals Board and in court. 3 Thus, Landinfo plays a significant role in assembling and presenting information that constitutes the basis of the "maps" that decision-makers use in their assessment of asylum claims.

The Map and the Territory
In 1931, the Polish-American mathematician and philosopher Alfred Korzybski famously stated that "the map is not the territory." 4 Korzybski was talking about language, using the map as a metaphor. Languages and maps are similar in the sense that they are representations of the real world. Ideally, they have a similar structure as reality, but they can never fully mimic that reality because perception and interpretation always intercede (Korzybski 1994, 58). Korzybski's words came to mind during an interview with a decision-maker whose story poignantly illustrated his point. When the interviewee was relatively new at the job, she assessed claims from a country where many applicants needed to cross a bridge to get to the neighboring country. She explained that when she talked to applicants, she imagined a bridge that resonated with her own experience of border crossingsnamely, the "Svinesund Bridge," which connects Norway and Sweden. The Svinesund Bridge is about 90 meters tall and 700 meters long and, therefore, a dominant landmark. "I had this giant image of [the bridge] in my mind," she said. "I imagined it to be at least as big as the Svinesund Bridge." After some time, she had the opportunity to travel to the country in question and, at one point, decided to go for a trip that took her across the bridge that she had encountered so many times in applicant narratives: I get on the bus, and the bus drives into a kind of a no man's land, and far away I see a bridge, and I am excited. I was really excited -and then the bridge was behind us.
[The "river"] was a tiny creek of about one meter, and the bridge was only an iron gate before you came to the other side of the border.
The border crossing was called "the bridge," but it was hardly a bridge at all. The "river" was merely a small creek. The landmark that she had encountered numerous times in applicants' narratives did not exist as she had imagined it. She commented: "It is possible that we have confused many applicants by asking a lot of questions about this bridge." The episode illustrates how easily a landmark, when imagined from a distance -in this case, from an office in Norway, thousands of miles away from its actual location -can grow out of proportion compared to its local significance. It also illustrates how decision-makers' personal experiences, such as knowledge of the border crossing between Norway and Sweden, can shape their view of phenomena in ways that, while appearing to be objective and concrete, diverge from applicants' lived realities.
The story about the bridge has a parallel in debates in the social sciences about the relationship between representation and reality and scientists' ability to produce accurate accounts of the objects under investigation (e.g., Smith 1996). Such debates have been particularly important in the field of human geography (e.g., Duncan and Ley 1993). Throughout the last century, different geographical traditions have propagated different modes of representation. On the one side, there is the notion of mimesis -"the belief that we should strive to produce as accurate a reflection of the world as possible" (Duncan and Ley 1993, 2). In cartography, the idea of mimesis corresponds to the view that a well-designed map will more or less be an accurate depiction of reality. This mimetic model of scientific representation was called into question in the 1960s as part of the critique of positivism. This critique continued during the next decades from (broadly speaking) two constructivist approaches that problematized mimetic claims of representations (Duncan and Ley 1993, 5). One was the hermeneutic approach associated with "humanistic geography," which focused on the way representation is shaped by interpretation, local context, and meaning making (Peet 1998, 48). The second was the postmodern approach that emphasized the ways representation is entangled in power relations that shape representation and the authority of knowledge (e.g., Harley 1989;Crampton and Krygier 2006). In line with the latter tradition, maps became the object of textual analysis and deconstruction, with the goal of unpacking their seemingly unquestionable scientific quality by "reading between the lines of the map" (Harley 1989, 3). Scholars in this tradition of critical cartography argue that the act of making mapsand, thereby, presenting a specific image of the world -often serves political interest, particularly those of the state (Harley 1989). From this viewpoint, "maps make reality as much as they represent it" (Crampton and Krygier 2006, 16).
If we return to the topic of this article, the mimetic tradition in human geography, in which maps are seen as relatively accurate representations of reality, fits with the description of country information as a "collection of objective facts allowing rational and de-politicised decision-making" (Rosset 2019, 109). The units that produce and compile country information for the purpose of asylum assessments tend to emphasize that the presentation of objective facts is the goal of their knowledge production (van der Kist, Dijstelbloem, and Goede 2019). The mimetic approach to knowledge resembles, moreover, the positivist view of facts frequently found in legal discourse (Sweeney 2007, 30). Indeed, in legal training, the facts to which the law should be applied tend to be viewed as "philosophically unproblematic" (Good 2004, 31). The process of abstracting facts from the complexity of raw data receives little attention. Once facts are established, they are treated as "ultimate facts" subject to legal analysis, which, at its core, represents abstract, universal principles (Rigby and Sevareid 1992, 8-9).
The constructivist/critical cartography tradition, on the other hand, resonates with the work of scholars who critically investigate the production of country information and its role in the migration control system (Rosset and Liodden 2015;Rosset 2019; van der Kist, Djistelbloem, and de Goede 2019). From this perspective, the representation of country information as objective and neutral does not sufficiently account for the context of knowledge production and the ambiguity and uncertainty of the information itself, which requires a great deal of interpretative work. As such, producers and users of country information play a role in consolidating uncertain information into "facts" that serve to alleviate "the undecideability of asylum adjudication" (van der Kist, Djistelbloem, and de Goede 2019). The production of knowledge can also be shaped by state institutions' interests in maintaining authority and legitimacy when enforcing migration control (Rosset 2019;van der Kist and Rosset 2020). This perspective emphasizes that knowledge is inevitably influenced by the interests and interpretations of the institutions that produce or apply it.
In the analysis below, I use the metaphor of a map with landmarks to describe the information decision-makers used in their justification for asylum decisions to capture the authority of mimetic representation, keeping in mind the critical and constructivist arguments presented above. As already mentioned, I use the map metaphor in reference to not only geographical knowledge but also country information in general. As such, the use of the term "map" in this article bears similarities to frameworks related to "mental maps" and "environmental cognition" (e.g., Tuan 1975;Moore and Golledge 1976;Henrikson 1980;Weston and Handy 2004), which conceptualize the environment broadly, including economic and political aspects (Moore and Golledge 1976, 5; see also Weston and Handy 2004, 535). Another relevant parallel to the mental map tradition is the view that representations of the environment are often based on partial and incomplete information and that people's cultural filters, experiences, and knowledge shape such representations (Moore and Golledge 1976, 5;Axelsen and Jones 1987, 448). While people primarily acquire mental maps "when moving through" a landscape (Weston and Handy 2004, 536) the encounter with reality does not need to be direct (Henrikson 1980, 500). Asylum decision-makers have often not been in the landscapes depicted in their mental maps, which they acquire primarily through country information and their interactions with asylum-seekers. By reading country information and listening to applicant stories, asylum decision-makers move mentally through the landscapes of countries that are far removed from their personal experiences. In doing so, they engage in mental map-making from a great distance and use those mental maps to make consequential decisions about asylum applicants.

Somalia: Situating Oneself on a Map
In the case files for asylum applications from Somalia, some of the most important landmarks used to argue for the outcome of an asylum assessment were geographical facts and knowledge about local areas. Since Somalia's independence from Italian and British colonizers in the 1960s, the country has been riddled by conflicts and civil war (Janzen and Lewis 2020). In the wake of dictator Barre's fall in 1991, Somalia was divided into a northern region considered to be relatively stable and a southern region characterized by continued armed conflict (Wettergren and Wikström 2014, 6). For an extended period, the UDI typically rejected applicants from northern Somalia, while applicants from southern Somalia, particularly Mogadishu, often received humanitarian permits on the basis of the general security situation in their home area (UDI 2008b(UDI , 2013. Because of the centrality of geographical origin for the outcome for Somali asylum-seekers, the focus in asylum interviews was largely on determining where in Somalia they were from (UDI 2008b). To assess asylum claims from Mogadishu, decision-makers had access to maps and a database with information about the different city areas, mosques, markets, schools, and large buildings. Moreover, to assess the applicant's geographical origin, decision-makers used information such as the names of nearby cities and villages, the distance to rivers, and the area's physical features. The following excerpt from an interview with an asylum decision-maker illustrates her attempts to elicit both the applicant's authentic narrative and specific information that could be checked against landmarks: Today, I made a decision about an applicant from a small town. We have no information whatsoever about this town, other than the fact that I can look it up on the map. So verification of his home area mostly is about the applicant's narrative. I am after names of villages, or names of the different parts of the town. He can say that the town is by the river, which parts of the city that are on that side of the river, the order of the city parts, where his house is, he can describe the road going from A to B in town. And that's more to see if he talks about it in a natural way and has real-life references. I have no way of opposing the things he says, but of course, you can tell pretty fast according to the level of detail whether or not the applicant is telling the truth . . . . What we are able to check about this town are the power relations. I asked the applicant quite a lot about that, and then I asked Landinfo, and they answered back. Landinfo listed the power relations, the changes in power that have taken place . . . . The applicant's story corresponds with the things Landinfo says. The names of warlords. So I ask the applicant to talk about what it is like to live in an area where al Shabaab is in control. It's essential. Have you been there in the past few years? Have you lived in this town where al Shabaab is in control? And then he can talk about that, quite detailed. Half a page of free narrative about what he thought it was like to live under al Shabaab. And in addition to that, I take what the applicant says as a point of departure. Network is important. So the applicant says that "yes, I have an uncle who is a nomad outside of town, and another uncle who is a nomad in a different place outside of town." OK, what's the name of the villages in the area where your uncle is a nomad? Then you can list a bunch of villages around the town -and I can find those on the map.
The decision-maker explained that "you can tell pretty fast according to the level of detail whether or not the applicant is telling the truth." The applicant's personal narratives about his home area are important for the decision-maker's perception of the claim's authenticity, which matters for assessing the claim's internal credibility. Additionally, to assess the claim's external credibility, it is vital that the decisionmaker can match some of the applicant's answers to information from Landinfo, the Norwegian country information unit. In the assessment, the decision-maker goes back and forth between the applicant's narrative and landmarks that she can verify via Landinfo to check whether they correspond. The balance between the personal and the verifiable lends credence to the narrative. As the following excerpt from a rejection letter illustrates, however, that balance is not always easy to strike. Sometimes, applicants provided information that was too general or simply could not be identified on decision-makers' maps: We have noted that the applicant answers questions about his claimed area of origin both during the police registration and during the interview by referring to the districts in the region. This is considered to be commonly known country information and has not been given weight. We further comment that the applicant has provided information about his home area that is not verifiable. This has not been given weight either. (case 6, Somalia) As the above quote shows, when the applicant did not provide information about the right kinds of landmarks, s/he stood a limited chance of acceptance. The landmarks the applicant referenced may very well have existed, but if they were not drawn onto the maps used by decision-makers, such landmarks were not effective in securing the crucial link between the asylum-seeker's narrative and the map.
In some instances, the interviews contained geographical information that could be interpreted in different ways. In a report from an asylum interview, an applicant was asked to name the largest parts of the city, presumably to elicit information that could be checked against formal sources. He named the three parts of the city that were most well-known and added that there are other smaller areas, which were named after mosques or soccer fields. In the rejection, the names he provided of the city areas became a point of contention. The applicant had referred to his part of the city by its old name, which, according to an anthropologist, was changed after 1960. The UDI wrote: "We consider it unlikely that a young man does not know that his own city part today is called [new name]" (Case 1, Somalia). Using the old name could be, as UDI suggested in the decision, a sign that the applicant was not from the city. On the other hand, it is not unlikely that inhabitants informally continue to use old names even after they have been changed. The applicant's use of the former name could also be interpreted as a sign that he really was from the city, since presumably, someone who was not from there would learn the official and up-todate names merely by studying a map. Wettergren and Wikström similarly point to the many possible "truths" behind a given answer in their discussion of an excerpt from a Somali case file in Sweden. In the file, the caseworker noted the correct answers next to the applicant's responses to questions about districts, monuments, and bridges. The applicant answered correctly in some instances but gave different answers than expected at other times. These "failed answers" can be seen "either as an indication that the applicant is lying and thereby is not credible, or as an indication that the MB [Migration Board's] information is incorrect." Further, they note: There is also a third alternative, suggested by the applicant, that there are multiple truths regarding the names. This alternative undermines the very raison d'être of bureaucratic procedure as it strives to contain, not to embrace, ambivalence. (Wettergren and Wikström 2014, 573) In these kinds of situations, decision-makers held discretion that they could exercise in different ways. They could contain the ambivalence that accompanies the possibility of "multiple truths" about the name of geographical landmarks, or they could explore the ambivalence by questioning the landmarks themselves or seeking out alternative interpretations for the discrepancies between the landmarks and the applicant's explanation.
In many Somali cases, the information that decision-makers used had a mimetic quality because it referred to concrete objects and features in the landscape that appeared concrete, objective, or quantifiable. This kind of information was often difficult for applicants to challenge, as their only means to do so were personal narratives about their home area. In one interview, a decision-maker repeatedly asked the applicant about distances from her home village to other nearby places. The applicant gave short and vague answers that could be interpreted in at least two ways. The first interpretation is that she did not know the answers because she was from somewhere else and wanted to evade revealing questions. The second interpretation is that she did not know the answers because this kind of knowledge was not relevant or obtainable to her in everyday life in the village, which she claimed she never left before escaping Somalia. Throughout the asylum interview, the applicant attempted to forward the latter explanation several times, for example, by suggesting that she did not go to school and, therefore, could not explain distances. The following is an excerpt from the interview:

How far is it to the ocean?
It takes me about seven minutes to walk from my home to the ocean.
How far is it to the Shabelle River? I don't know. I have never been to the river . . . . Hard to say. Very far. If you go by car, how long would it take, closer to an hour or to 10 hours?

Is
It does not take 10 hours. My father would come back the same evening when he travelled to Mogadishu. It's a shorter distance to Marka.
The applicant consistently failed to provide answers that fit the decision-maker's landmarks. She tried to provide explanations based on her personal experiences instead, such as when she replied that her father would return the same evening when he traveled to Mogadishu and that Mogadishu was a longer journey away than Marka, another city in the area. In the rejection letter, her failure to place her village in accordance with the decision-maker's map and the fact that she did not know the names of surrounding villages constituted arguments for rejection. In the summary of the claim put forward by the applicant's lawyer to the Immigration Appeals Board, the lawyer stressed that there was a gap between the applicants' knowledge and the "map" used by the UDI: The interviewee and the interview object clearly have a different perception of reality. Many of the questions are based on matters such as the number of kilometres, time assessments, months and years. The sources of error are accordingly large. The same counts for the question about the direction of the road to Mogadishu, we do not know if this is the large main road or a small road. The Appeal's Board rather needs to base its decision on concrete descriptions, like the fact that the road is flooded several times a week. On this basis, the complainant has given a concrete and detailed explanation. She has no relationship to neighbouring towns, but she has accounted for those things that were important to her. (case 6, Somalia) The landmarks that the decision-makers used in this case were concrete and quantifiable, but their use depended on specific expectations about the way in which the applicant related to these landmarks in her everyday life. As the lawyer points out, there could be multiple reasons why someone might have a different relationship to a landmark than expected, such as the need to take a small road because the main road was flooded.
During the meeting with the Immigration Appeals Board, the applicant explained many things in more detail than she had done during the arrival and asylum interview. Her narrative approximated the decision-makers' "map." When this kind of information was presented at a late point in the process, however, it did not necessarily strengthen the claim. Instead of being regarded as a sign of credibility, the match between the applicant's answers and landmarks on the map could be interpreted as a sign inauthenticity and fabrication. In this case, the Immigration Appeals Board remained unconvinced about the claim and maintained the rejection. Note how concrete and detailed the landmarks that decision-makers used were, such as the number of different kinds of fish near the applicant's village: The applicant's explanation about her home area was very limited shortly after her arrival to Norway, while she increasingly has explained herself in a more detailed manner about the place. Such knowledge can be acquired in different ways, and [the Immigration Appeals Board] does not put much weight on the fact that her knowledge appeared to be richer during the board meeting than during previous interviews. It is symptomatic that she during the arrival interview only was aware of one kind of fish outside of her village, while she during the last meeting mentioned three different kinds of fish. (Case 6, Somalia) An increased level of consistency between an applicant's narrative and decisionmakers' landmarks during an appeal constitutes a dilemma for applicants and decision-makers alike. For decision-makers, it is impossible to know whether an applicant's knowledge has improved for strategic reasons, creating a Catch 22 for applicants. If they try to respond to rejection arguments with more detailed accounts, it may be considered a strategic adjustment. If they do not, they are unable to redress the issues that led to a rejection in the first place.
A mismatch between the landmarks and the applicant's narrative by no means automatically entailed a rejection. In fact, decision-makers could give different weight to similar landmarks, depending on the case as a whole. For example, there are many similarities between the case above and another one concerning a young man who claimed to come from a small village. The applicant explained that he was unable to give accounts of the area surrounding his village because conflicts and roadblocks prevented him from leaving his home. In the rejection letter, the UDI noted, in a similar manner as the previous case, that the applicant had provided incorrect information about the names and geographical location of nearby villages, stating that the "applicant's explanations about where the towns are situated in relation to each other do not match the map" (case 3, Somalia). During the appeal at the Immigration Appeals Board, however, decision-makers came, "under doubt," to a different conclusion: The complainant's information about his home area is limited, but the Board cannot disregard the fact that this may be caused by the situation in the area and the fact that the freedom of movement of the population has been accordingly restricted throughout the complainant's entire upbringing. (case 3, Somalia) One plausible explanation for the different outcomes in these two cases is that the decision-makers perceived the internal credibility of these claims in a different manner. Thus, a claim's internal credibility can potentially make decision-makers interpret or at least use country information in different ways. The claim's internal and external credibility are not necessarily independent of each other. When an applicant's narrative appears highly authentic, decision-makers may be prone to accept deviations from landmarks and interpret them in a more benign manner. If the applicant's narrative appears thin and inauthentic, even minor deviations can be used in a decision as proof that the claim cannot be trusted. Thus, landmarks are malleable, and similar landmarks can be given different weight and significant in the context of a decision.
As the discussion above illustrates, many of the landmarks used in decisions about cases from Somalia were very concrete, even quantifiable. The information on which they were based had a descriptive and mimetic quality -they often appeared to reflect realities on the ground in a one-to-one manner. Even if landmarks were clear and concrete, however, decision-makers had to make assumptions about how applicants related to them in their everyday lives. For example, it may be beyond doubt that the name of the part of a city had changed officially, but how people in the city referred to their neighborhood remained uncertain. Geographical distances and placements on a map may likewise be unambiguous when looking at a map, but North, South, East and West may not have much meaning for someone on the ground who has never learned about geography. There may always be reasons for why an individual's circumstances and experiences translate into deviations from the landmarks.

Iran: The Reasonable Persecutor
In case files from Iran, the landmarks were not about geographical phenomena, but they retained some of the mimetic quality in the sense that decision-makers often appeared to use them as if they were accurate and direct representations of realities on the ground. Most case files from Iran concerned applicants who stated that they had been engaged in activities on behalf of forbidden Kurdish political parties. UDI's guidelines for assessing cases from Iran at the time stated that applicants who had been engaged in activities for such parties would most likely have a wellfounded fear of persecution (UDI 2008a(UDI , 2010. The core question was about the claim's credibility. One example of a landmark that appeared in several decisions was the size of secret cells of party members who worked together. According to a report from Landinfo, such cells consisted of three members. In one case, the applicant stated that he had participated in a cell that consisted of five members. In the rejection letter, the UDI stated that "according to the knowledge of the Directorate, such groups operate in cells of three persons" (case 5, Iran). In the complaint, the applicant argued against this assertion: "UDI's information that cells consist of maximum three people is incorrect. The cells should be minimum three persons and it is not unusual that they can consist of more people." The applicant could not substantiate this statement apart from referring to his personal experience, but a later report written by the Danish Refugee Council supported his claim, stating that secret cells often consisted of "three, five, seven, nine or 11 persons" (Danish Refugee Council 2013, 26). In a case from Ethiopia that constitutes an interesting parallel (Hauge 2016), the size of political underground cells also became an issue in the credibility assessment. While the applicant claimed that there were as many as 10 to 15 people at secret meetings, a witness from Landinfo explained that the political party in question normally operated with cells of three to five people, with a maximum of ten members. Hauge comments that it is striking how "concrete the country information is, and in addition it is striking that it is used in such an absolute manner" (Hauge 2016, 23). Although the asylum decisions did not depend on the argument about cell size alone, these asylum cases illustrate that information is sometimes used in a very literal manner, as if it constitutes a one-to-one reflection of realities on the ground.
A theme that has been raised several times in the literature on asylum decisionmaking is decision-makers' tendency to make assumptions about what kinds of behavior that is logical and rational. These assumptions are applied both to individuals who apply for asylum and to their persecutors (Herlihy, Gleeson, and Turner 2010;UNHCR 2013, 176-184). In their report about credibility assessment, the UNHCR states, for example, that their review of case files "indicated that decision-makers speculate on how the applicant or a third party ought to have behaved, or on how events could have or should have unfolded" (UNHCR 2013, 181). A similar tendency could be seen in some of the case files about Iran. For example, in several cases, the UDI considered applicants' description of the way the secret police, Ettelaat, had proceeded to be unlikely. This landmark seemed to be based on a combination of informal knowledge and assumptions of the "reasonable persecutor," on which I elaborate below (Sweeney 2007). In one case, the applicant explained that Ettelaat had called his house and talked to his mother, who said that he was not at home. Two days later, they called again. When she maintained that he still was not home, they stormed and searched the house. In the rejection, the decisionmaker wrote: The UDI refers to the fact that it is unlikely that the Ettelaat would call home to the applicant if they suspected him of activity for [political party]. According to the knowledge of the Directorate, Ettelaat would arrest directly. They are not known to warn the persons they seek of their arrival beforehand. (case 1, Iran) Similar arguments about the Iranian authorities' expected behavior could be found in several cases. In another example, an applicant claimed that he had been summoned to a meeting with Herasat, a group responsible for monitoring morality and implementing Islamic law (Austrian Red Cross and Accord 2015). After the meeting, the applicant claimed that Herasat had interrogated him in the basement of his workplace for several days. In the rejection, the UDI stated that according to their knowledge, this kind of containment was not in line with Herasat's work methods (case 8, Iran). In the complaint, the applicant tried to counter this argument by stating that the information was not correct: "It is well known that these groups, which have a more or less loose relationship to the authorities of the country, are autonomous and will not be held accountable for their actions." In their petition for a renewed assessment of the case, the Norwegian Association for Asylum Seekers noted that it is "hard to understand what the UDI's assessment concerning Herasat's work methods are based upon, since there are no references to sources." Although the work methods of Ettelaat or Herasat constituted concrete pieces of information used in rejections, there were indeed no references to sources -merely a statement that the applicant's account was not in accordance with the "knowledge of the Directorate." In a report about asylum claims in court, the Norwegian Bar Association similarly observed that the Immigration Appeals Board frequently stated that an issue in a claim was "contradictory to Board's knowledge" without specifying the basis of this knowledge (Humlen and Myhre 2014, 29).
One possible explanation for the lack of references may be that some landmarks that decision-makers used were not based on formal country-information sources, but on decision-makers' experience with previous claims. In the words of one caseworker, "you have in-depth knowledge about what has been said in numerous previous cases, and you can see that it does not match well with the present case." Although she added that she was not able to use this kind of information in a decision, it is possible that when the UDI or Immigration Appeals Board referred to something as "not in line with their knowledge," without specifying it further, they were referencing this kind of experiential knowledge with other cases.
Another explanation for the lack of references to sources may be that decisionmakers sometimes, instead of basing their decisions on country reports, relied on certain "unstated assumptions about the motives, and 'efficiency', of persecutors" (Sweeney 2007, 27). According to Sweeney, there is a parallel between the way judges use the notion of "a reasonable man" as an analytical tool to determine liability and the idea of the "reasonable persecutor" in asylum assessments. In Sweeney's words, if the persecutor falls "short of the standard to be 'expected' of a 'reasonable persecutor', then the story is disbelieved" (2007,27). In one case that Sweeney refers to, the persecutor failed to search for the applicant at a family member's house. As Sweeney notes, "a 'reasonable' persecutor would have found the appellant and therefore the story must be false" (Sweeney 2007, 28). The model of the reasonable persecutor is primarily based on decision-makers' "impression of how the persecution could have been better executed" (Sweeney 2007, 27).
During asylum interviews, it was not unusual for applicants to be held accountable for what the Iranian authorities did or did not do. If the authorities did not seek out or arrest the applicant, the applicant was asked to explain this failure to act, as well as the opposite: if the applicant did not have a good explanation for why the secret police had searched his apartment and found forbidden material, the claim was difficult to believe. The gestalt of the reasonable and effective persecutor, thus, depends on the idea that an authoritarian regime's actions are predictable and that agents of persecution "act in a uniform, predicable fashion, treating like individuals in a like manner" (McDonald 2014, 453). As the applicant in the quote from an asylum interview below suggests, however, unpredictability can be an important tool in the hands of an authoritarian regime: Why do you think the authorities did not take action [to arrest your father] until after 1.5 years?
The Islamist regime's actions are unpredictable. You cannot know beforehand what they will do. They can leave someone in peace, and then suddenly arrest a whole lot of people. (case 2, Iran) The secret police's goal may not always be to arrest and interrogate, but instead to intimidate and deter. Unpredictability contributes to this end. Acknowledging such unpredictability on the part of Iranian authorities would, however, interfere with the very purpose of landmarks in the assessments -namely, to produce clear distinctions to be used as reliable tools to determine the claim.
As the previous section has illustrated, the landmarks used in the cases from Iran differed from those in the cases from Somalia. In the Iranian cases, the landmarks seemed to be made up of a combination of formal and informal knowledge. In addition to country reports that described the way underground political work was organized, decision-makers seemed to rely on knowledge that, in part, was obtained through asylum interviews and, in part, seemed to be underpinned by assumptions about rational behavior. The commonality between the landmarks deployed in the Somali and Iranian cases was decision-makers' tendency to use the landmarks in a very concrete, often absolute manner, as if the landmarks constituted an objective template against which applicant stories could be measured.

Asylum-seekers' Stories as an Impetus for Revising the Map
We're not familiar with the idea that threats and violence are common in recruitment to Al-Shabaab. Would you like to comment on this?
Well, it would be the people who live in Somalia those who can know something about that. (case 1, Somalia) The discussion in previous sections illustrates an asymmetry between the status of the information decision-makers used in the assessments and the information applicants brought to the table. The quote above illustrates this asymmetry. An applicant is confronted with the idea that his narrative is not in line with the country information about Somalia. The applicant's response is that it is the people "on the ground" who will have first-hand information about what is going on. His case was rejected, one of the main arguments being that "the use of force and violence is neither common nor necessary in connection with recruitment to various armed groups, among them Al-Shabaab" (case 1, Somalia). Contrary to this landmark, many applicants claimed that they had been subject to violent recruitment. Applicants did not have any means to corroborate their view, but the landmark was, in fact, later revised when new information indicated that such recruitment was more common than assumed (Landinfo 2012).
This example illustrates more generally that discrepancies between applicant narratives and decision-makers' landmarks can be an indication that the map in use is not in sync with the territory. Several decision-makers described situations where applicants had forwarded claims that initially were dismissed because they did not resonate with the settled landmarks, but were later confirmed by more up-to-date sources. In the following quote, a decision-maker described this kind of situation, where information was used in a way that was "too absolute": There have been situations a few years ago where we used country information in a way that was too absolute, for too long. And then we began to realize that these things can happen anyways. This is an illustration of the fact that we're always lagging somewhat behind . . . . It takes quite a lot of time from a pattern develops on the ground until information about it reaches us. We've had situations before where we have used information in a manner that was too absolute. We always had additional arguments [for rejections]. But our point of departure was country information that turned out later not to be . . . we could not really use it, there were too many exceptions.
At an aggregate level, asylum stories can be considered a source of informal knowledge that figured in the background of assessments. As Sweeney (2007, 25) points out, decision-makers should "not underestimate the capacity of asylum seekers to provide new and interesting material that could contribute further to an understanding of the context." In some instances, decision-makers did heed applicants' narratives when they were not in line with the landmarks and even used those narratives as an impetus to revise their maps. Their inclination to do so appeared to depend substantially on whether they approached applicants with a presumption of truthfulness (Durst 2000). Such a presumption entailed a willingness to question not only applicants' credibility if there was a mismatch between their narratives and the landmarks, but also the landmarks themselves, as described by this decision-maker: We're in contact with applicants every day. We listen to their stories every day. I find it conspicuous if very credible applicants -over a period of several months -talk about political activity that Landinfo says does not take place. Then I believe we have to ask whether Landinfo is wrong. Because I cannot work under the assumption that everyone who comes to me is sitting there and lying about what they've experienced. I cannot work like that. I believe that if they substantiate what they are saying and they seem credible, then there's reason to question the country information.
The decision-maker here worked under the assumption that applicants were telling the truth, and if their claims did not match the country information, she considered that to be a reason to question the accuracy of the information, particularly if the claims were "very credible." As such, the presumption of truthfulness entailed that the asymmetrical relationship between the decisionmaker and applicants decreased somewhat, to the applicant's benefit. It was not, however, always the case that decision-makers approached applicants with a presumption of truthfulness. Some units appeared to be more characterized by a "refusal mindset" or a "culture of disbelief" (Jubany 2011;Baillot, Cowan, and Munro 2014) in which decision-makers from the outset were vigilant about finding contradictions and mistakes in applicants' statement that could support a rejection. In this situation, a mismatch between applicant narratives and landmarks did not seem to spur decision-makers' doubts about the map but merely led to further questioning of the applicant's credibility.
Some of the quotes above illustrate another important point -namely, that decision-makers often were conscious of the uncertain nature of country information. When they talked about country information, they did not appear to think of this information as a collection of unproblematic facts, as the following quote clearly illustrates: The country information on Somalia is not bullet proof enough for us to use it in a very categorical manner. It's a point of departure. If an applicant explains something that contradicts the country information, it takes more to substantiate [the claim], but suddenly you get a story that is really coherent and really detailed and then you have to accept it. You never use country information alone; you always need additional arguments. You cannot merely trust the country information.
Indeed, in interviews, decision-makers were often uncomfortably aware of the uncertainty connected to country information and of their own responsibility in interpreting that information. As one decision-maker put it, "country information reports are not written in a way that you can relate to them objectively. It's an interpretation." There is an interesting contrast here between decision-makers' awareness of the uncertainty of country information and the often-mimetic use of such information in the actual decisions, as illustrated throughout this article. I return to this discrepancy in the discussion below.

Discussion
In this article, I have used the metaphor of a map to describe the image decisionmakers construct of asylum applicants' origin countries. The map, I have argued, is a useful metaphor to understand how country information functions in the refugee status determination process because of its authority as a seemingly objective depiction of reality. Maps are, however, always drawn from a specific viewpoint and with a particular purpose in mind (Axelsen and Jones 1987, 447/461). The construction of maps among decision-makers in the asylum bureaucracy was shaped by the need to create distinctions between refugees and non-refugees, and their interpretation of country information was geared toward that goal. The specific pieces of information that decision-makers used to justify the outcome of a claim can be considered "landmarks" on a map that served as signposts in the direction of an acceptance or a rejection. The process of creating landmarks involved a mix of "hard" facts, based on country-information reports, and informal sources, such as exposure to applicant stories and assumptions about logical and rational behavior on part of persecutors.
For decision-makers, the credibility of claims depended on the relationship between the applicant's narrative and the map. To come across as credible, applicants had to strike a balance between conformity to the map and personalized detail in their narratives. They needed to follow familiar paths, pause at landmarks that were inscribed into decision-makers' maps, while providing details along the way that communicated authentic experience. If the narratives were too generic, deviated too much from the map, or complied with it too late in the process, the claim's credibility was difficult to recover. In some instances, however, highly credible applicants whose narratives deviated from the map could prompt decision-makers to question and revise their maps. Whether such a revision took place appeared to depend crucially on decision-makers' tendency to approach applicants with a presumption of truthfulness.
In several cases, decision-makers appeared to use landmarks to justify decisions as if there was a one-to-one relationship between the map and the territory. In this sense, they had a rather positivist approach to facts, in line with approaches in human geography that stress the primacy of mimetic representation (Duncan and Ley 1993;Hubbard et al. 2005) or the dominant legal view of facts (Rigby and Sevareid 1992). As the last empirical section demonstrates, however, decision-makers were often highly conscious of the many uncertainties surrounding country information and their own role in the interpretation of such information. How, then, can the mimetic use of landmarks be explained?
First, the infrastructure of the maps used by asylum decision-makers was framed by law, which requires a "yes/no answers, not a 'yes but' or 'it depends'" (Good 2007, 31) and leaves no room for dialogue or hesitation (Johannesson 2012, 74;Johannesson 2017, 172). In this context, information was useful to the extent that it enabled clear-cut distinctions, while complex, nuanced information was of limited value. For landmarks to serve as signposts that pointed toward acceptance and rejection, the ambiguity of the information had to be contained.
Second, in a context of limited evidence, the evidence that is present, however tentative and uncertain, may increase in significance (see Lipsky 2010, 131). Country information often constituted the only concrete, external evidence in asylum cases -the type of evidence that ostensibly reduced the need for subjective assessments (Wettergren and Wikström 2014, 13). As Sweeney (2007) notes, an "'objective' point of reference seems required in order to defend against accusations of 'subjectivity', and where 'subjectivity' is linked to uncertainty" (p. 31). When landmarks were based on formal reports produced by experts, they appeared to be objectively true and could, thereby, gain "trump card" status in the decisionmaking process (Backe-Hansen 2001, 223-226), alleviating some of the burdens on decision-makers connected to uncertainty and discretion.
Third, when decision-makers used landmarks in a seemingly mimetic manner, with few reservations, they may have done so because the landmarks primarily served as justification for convictions formed on other grounds. As one decisionmaker put it, "we use country information to reject cases that we don't consider to be credible anyway." If a decision-maker was inclined to disbelieve a claim, she could use a slight deviation from the landmark as an argument for rejection. While the use of landmarks appeared to be very absolute in such cases, the landmark served as a justification for an outcome that rested primarily on an instinct or gut feeling that decision-makers developed over time through experience with other cases (see Liodden 2021). Although the focus here is on country information, this finding resonates with the view of legal realists, who similarly argue that legal rules function as "post hoc rationales for decisions reached on other grounds" (Leiter 2001, 281).
This article contributes to illuminating the relationship between law's requirement of clear and unambiguous categorization and the complex reality that lies beneath the "facts" upon which legal decisions rest. Its findings resonate with the work of Anthony Good (2004), who suggests that during the legal process, the anthropological view of facts as provisional and contextual is largely subsumed by the common-sense, empiricist legal approach, where facts are considered philosophically unproblematic. Country information is described in the refugee status determination process as "objective evidence" or "generally known facts" (UNHCR 1998, para 11). As this article has demonstrated, this terminology conceals the scope of uncertainty involved in establishing the facts as such in the first place.

An Agenda for Future Research on Mapmaking
The theoretical approach I have outlined here raises questions for further research. For example, how much do state interests shape the maps that decision-makers use to assess asylum applications? If, as critical cartographers suggest, the state has a stake in producing a specific representation of the world (Harley 1989), state interests may shape the interpretation and use of country information as well. The number of applicants accepted as refugees in a given country depends on the perceived situation in applicants' home countries. As such, country information is crucial to migration control, which is a highly salient political matter to most Western states. To shed light on the extent to which states leave an imprint on the maps used in asylum decision-making, we need comparative studies that investigate who is vested with the power to establish representations of the applicants' home country and under what conditions. The autonomy of the asylum bureaucracy varies in different national contexts (Hamlin 2014), and presumably, so does the independence of decision-makers with regard to interpreting and using country information. In Norway, "Landinfo" produces country information, but decision-makers themselves are largely responsible for making guidelines for the specific use of country information in asylum assessments. In the UK, by contrast, there is a special unit, the Country Policy and Information Team (CPIT), that conducts both tasks -the unit produces country information and issues policy guidance on the use of such information in asylum cases. Previously, these two tasks were kept separate, and the CPIT has been criticized for adapting country information to policy interests (Independent Chief Inspector of Borders and Immigration 2017). The blurred boundaries between the provision of country information and policymaking suggests that political interests may carry more weight in the interpretation of country information in the UK than in a country like Norway, but this question calls for comparative research.
Decision-makers' autonomy vis-à-vis the political sphere may differ not only between countries but also over time within the same country, depending on the political climate and specific events that put the question of asylum high on the political agenda. As I have described elsewhere (Liodden 2017), when there are many asylum-seekers, street-level decision-makers in Norway's asylum bureaucracy seemed to have less influence over the interpretation of country information. In these situations, the interpretation of country information appeared to be settled higher up in the asylum bureaucracy, closer to the political masters, and, presumably, more in line with the government's political interests. For example, in 2015, the Norwegian government instructed decision-makers in the UDI and the Immigration Appeals Board to consider Russia a safe third country for most asylum-seekers. The instruction was a response to the large number of asylum-seekers arriving at the Russian-Norwegian border, and it suggested that many applicants, at least in theory, could be returned to Russia without having their claim processed in Norway (Justisog beredskapsdepartement 2015). The controversial instruction is an example of a specific interpretation of country information that expressed state interests. The question about the influence of politics on the interpretation and use of country information requires further research, particularly in the aftermath of the arrival of a high numbers of asylum-seekers to a country.
A final question, which also calls for comparative research, is the extent to which applicants and others, such as their lawyers or non-governmental organizations, can challenge settled interpretations and present other sources of country information that may be important for the outcome of asylum cases. In the terminology of cartography, the question at stake is whether there are opportunities for producing "counter-maps" (Crampton and Krygier 2006) -in this case, representations of the situation in applicants' home country that emphasize other landmarks or question the accuracy of the landmarks that decision-makers use. There are structural and legal differences across countries that may shape the ability of external parties to produce such counter-maps. In Norway, decision-makers on all levels (the UDI, the Immigration Appeals Board, and the courts) rely on Landinfo's country reports and oral testimonies. Moreover, the Norwegian legal system's non-adversarial approach means that Landinfo does not serve one party but should present information in a neutral manner that allows the judge to make an impartial decision. Conversely, in a country like the UK that has an adversarial legal tradition, the two parties in the appeals process (the applicant's lawyer and the state representative) use country information actively as part of their artillery to challenge the other party's claims. In this situation, the production of "counter-maps" seems commonplace, but the multitude of competing maps also entails clear challenges, such as inconsistent, selective, and strategic use of information and the suspicion of country experts, who have often been accused of a partisan agenda (Good 2004). Comparative research can shed further light on the way various legal contexts affect the ability of different parties in the refugee status determination process to challenge, revise, or maintain landmarks and maps.