Illusions of objectivity: The two functions of country of origin information in asylum assessment

Abstract

The article argues that the intertextuality of country of origin information (COI) plays a significant part in constructing asylum decisions and therefore the lives and futures of those affected by forced migration. COI in asylum decisions is based on COI reports, which in turn are based on other COI reports, interviews, or media materials. This makes them intertextual, that is, referring to and transforming other texts. That is, COI reports typically include dozens or even hundreds of pages of information, which are abbreviated in order to be used in a few paragraphs of asylum decisions. This is done through recontextualisation, that is, by selecting certain parts of the text and reformulating them to the new context. The article analyses COI in 67 negative first-instance asylum decisions and in 38 COI reports and similar sources through the concept of intertextuality. It argues that COI has two functions in the asylum decisions, personal and general. Personal COI is used to assess the truthfulness of the asylum applicant’s persecution narrative, whereas general COI is used to evaluate the overall security and human rights situation in the country of origin. The article shows that even though the decisions tend to present COI as objective information, it can easily be biased, irrelevant, and illogical, therefore creating an illusion of objectivity.

1. Introduction

A statement made by the Finnish Immigration Service in May 2016 aroused international attention. According to the Service (2016a), the safety situation in Iraq, Afghanistan, and Somalia had improved; therefore it would become more difficult for citizens of the said countries to receive international protection in Finland. More precisely the Service stated that the countries no longer had geographical areas to which a person could not be safely deported. At the time ISIS governed notable parts of Iraq, including the city of Mosul, yet the statement implied that even Mosul was a safe place to return. The announcement was made in the aftermath of Finland, along with the rest of Europe, receiving an increased number of asylum applicants. Unlike in the European Union, in Finland Syrians were only the fourth largest group of asylum applicants after citizens of Iraq, Afghanistan, and Somalia (Eurostat 2022). While the European states mainly welcomed Syrians, they at the same time became more restrictive to asylum seekers from other countries. The statement of the Finnish Immigration Service should perhaps therefore not be looked at as an objective fact, but rather as knowledge politics (Van der Kist, Dijstelbloem, and de Goede 2019) in the ‘race to the bottom’ to make Finland more unattractive for migrants, as was done by various European states at the time (Rosset and Achermann 2019). As such, the change in policy can be interpreted as an example of the symbolic use of country of origin information (COI), that is, such information is used not only to inform about the country of origin but also about the asylum politics of the country of destination (Rosset, Liodden, and Maia 2015).

Seeking asylum is a strain of human migration that has become widely mediatised and politicised in debates. COI plays an important part in asylum decision making and has seen an increased research interest in recent years, although the total number of studies remains fairly small. COI has mainly been approached through the perspective of legal studies (Dauvergne and Millbank 2003; Gibb and Good 2013; Bednar 2015) and social anthropology (Good 2004; Affolter 2021; Rosset 2021), but also through studies of politics (Van der Kist, Dijstelbloem, and de Goede 2019) and public administration (Johannesson 2012). It has also been approached from the perspective of knowledge production (Good 2004; Van der Kist, Dijstelbloem, and de Goede 2019; Van der Kist and Rosset 2020; Rosset 2021) and usage by the decision makers (Affolter 2021; Liodden 2022) in the decision documents (Rousseau 2002; Dauvergne and Millbank 2003; Pettitt, Townhead, and Huber 2008; Johannesson 2012; Liodden 2022). COI exists, in essence, in the form of texts, yet it has not really been studied as such. Still understanding how country information takes shape and transforms from document to document is important in understanding what kind of an effect it can have on asylum decision making.

The article analyses COI as illusions of objectivity, that is, it presupposes that COI is often presented and understood as objective, but is actually not objective by nature (Gibb and Good 2013). Rather, formed by choosing from and recontextualising intertwined texts, COI is inherently intertextual (see Fairclough 1992; Reisigl and Wodak 2001; Farrelly 2020). The question the current article aims to answer is: how is COI intertextually constructed in asylum decisions? The article shows how the intertextual choices related to COI can affect the asylum assessment and how the COI in the asylum decisions can be contradictory with the COI reports it is based on. The article contributes to the growing literature about the knowledge politics of COI and therefore, on a wider scale, to studies about asylum and migration. As Bennett (2018: 95–101) has shown, migration policy texts are discursively connected to other phenomena, such a concepts, legislation, and political discourses. Therefore, the article also contributes to understanding the links in asylum decision making to the migration political restrictions made after the increased influx of asylum applicants.

2. COI in asylum assessment

COI is an essential part of asylum decision making (Pettitt, Townhead and Huber 2008; Gibb and Good 2013), as the applicant’s country of origin forms the basis of asylum assessment (Schittenhelm and Schneider 2017). In general, COI describes and estimates the security situation, political rights, and rights of minority groups, access to services such as housing and healthcare, as well as risk profiles and possible agents of persecution. The COI reports are usually gathered and produced by a specialised branch within the national asylum institutions albeit one that is detached from the decision-making unit (Rosset and Achermann 2019). This is also the case with the national focus of this article, Finland, where COI is produced by a specialised unit within the first instance decision-making institution, namely the Country Information Service (Maatietopalvelu) of the Finnish Immigration Service. The asylum decisions are made by the Asylum Unit (Turvapaikkayksikkö), and the Country Information Service supports the decision making by compiling reports about the security and overall situation in selected countries as well as of the selected groups (e.g. women, children, and political and religious groups), possible reasons of asylum (forced recruitment and tribal practices) and armed groups. They are also responsible for updating the COI database and providing information for specific queries by the decision makers. The information provided by the Country Information Service is used by the decision makers of the Finnish Immigration Service as well as the appeal organs (Staffans 2012: 152) which emphasises the important role of COI, as the same information is used at all levels of the process and is therefore not independently tested at the appeal stage.

COI can refer to both COI reports and the COI presented as part of an asylum decision, the latter typically being based on the former. In practice, the COI reports are papers of dozens or even over a hundred pages in length detailing the situation in a given country. The themes they include vary from geographical details and human rights legislation to the situation of minorities and religious and ethnic groups, treatment of political dissidents, and conditions of living (Van der Kist, Dijstelbloem and de Goede 2019). The reports can discuss the situation overall or concentrate on specific issues: for example the situation of religious converts, certain political groups, or practices related to sexuality and gender. They can be based on fieldwork or desk research, and desk research can in turn be based on other reports by various immigration services, human rights organisations, experts, and the media (Dauvergne and Millbank 2003; Good 2004; Gibb and Good 2013).

The COI in the asylum decisions is based on the COI reports, with the decisions cutting dozens or hundreds of pages of information into just a few paragraphs. In the decisions, COI is used to assess the country of origin of the applicant (i.e. whether they are from the country they say they are), the applicant’s persecution narrative (i.e. how possible and likely it is that the events told by the applicant would happen in the said country), the behaviour of the applicant (Affolter 2021), and specifically the future risk of persecution. Therefore, the COI in the decisions is specifically related to the country of origin of the applicant and their persecution narrative, that is, the reason, events, and agents of persecution. Using shared COI increases the consistency and efficiency of decision making, as otherwise the decision maker should conduct fact finding independently for each asylum applicant (Bednar 2015). However, because of its repetitive nature, if done poorly, it can also cause harm to a number of asylum applicants (Bednar 2015).

In asylum decision making, COI tends to be taken as objective evidence (Dauvergne and Millbank 2003; Johannesson 2012; Gibb and Good 2013), but as Gibb and Good (2013) argue it cannot in actuality be ‘objective’ because the knowledge production of COI is interpretative by nature. The assumption of objectivity means that COI is taken as ‘an unquestionable truth’ by decision makers, a truth that easily overrides the narrative of the applicant if they are in contradiction (Sweeney 2007; Johannesson 2012; Liodden 2022). For example, if a man was to claim in his narrative that he is facing a threat because of having an extramarital affair, the decision can simply override the narrative based on COI stating that such threat is mainly aimed at women (Johannesson 2012; Bodström 2020). Here the COI is therefore treated as ‘the unquestionable truth’ and as a more reliable information source than the applicant, even though it is actually based on various decisions and interpretations about, for example, which topics are addressed in COI and which sources and parts of them are selected to be used in the COI reports or asylum decisions (see Van der Kist, Dijstelbloem and de Goede 2019). Therefore, the non-objectivity of COI is by default also a matter of intertextuality.

3. Intertextuality of COI

Intertextuality refers to how texts ‘can transform prior texts and restructure existing conventions (genres, discourses) to generate new ones’ (Fairclough 1992: 102) and how they link ‘to other texts (explicitly, implicitly, by referring to them or incorporating elements of them)’ (Farrelly 2020: 359). According to Reisigl and Wodak (2001: 37), the texts therefore ‘cross between fields, overlap, refer to each other or are in some other way socio-functionally linked with each other’. As COI is essentially formed by referring to other texts and is therefore ‘self-referential’ (Van der Kist, Dijstelbloem and de Goede 2019: 75), it is always intertextual.

With its various stages of reference, COI forms intertextual chains (Fairclough 2003: 130) or networks of intertexts (Farrelly 2020), with various sources and documents building on one another. If we take the asylum decision as a starting point, the decision usually includes references to COI. In Farrelly’s (2020) terms, the asylum decision is the text, and the COI sources are its preceding intertexts. This is the first link in the chain—the link between the asylum decision and its sources. However, usually, the COI reports are based on other sources, be it other reports, interviews, or media texts. These form the preceding intertexts for the COI report, the second link in the chain (see Figure 1). And in case they are also based on other sources, their sources form further links in the chain.

Figure 1.

The intertextual chain of COI in asylum decisions.

Every time a new link is created, the preceding text in the intertextual chain needs to be paraphrased to fit in a new environment and purpose, that is, it needs to be recontextualised, a term coined by Bernstein (1990). For example, media texts are usually produced for the purpose of public awareness, not for the purpose of asylum assessment, but media content can be changed and repurposed to be used as COI, that is, to give information particularly about the situation in the country. Similarly, the dozens of pages of the COI reports need to be repurposed to serve the asylum decision, which typically only includes a few paragraphs of COI. This is done by selecting certain parts of the text, giving them a new purpose, and contextualising and rearranging them (Linell 1998). Therefore, recontextualisation holds an important gatekeeping function in asylum assessment, as the way the COI is used can have a determining impact on the end result of the decision.

Intertextuality can be divided into manifest intertextuality and interdiscursivity, both playing a role in the COI. Manifest intertextuality refers to explicitly drawing from other texts—what is quoted, when, how, and why—through direct or indirect reporting or using references for example. Interdiscursivity refers to invisible intertextuality, that is, the kind that is drawn from but not marked nor manifested, such as genre (Fairclough 1992: 101–133). In asylum decisions, manifest intertextuality is visible in the COI sources referenced and the information based on them. As for interdiscursivity, the asylum decisions can be seen as part of the genre of ‘official notifications’ (Reisigl and Wodak 2001: 215–222). Interdiscursivity also manifests in other related genres, such as legislation, policies, and the media (Bennett 2018). Although the current article mainly focuses on manifest intertextuality, it is also informed by the concept of interdiscursivity.

4. Data and method

The data of this article consist of 67 negative first-instance asylum decisions made by the Asylum Unit of the Finnish Immigration Service and 38 related publicly available country reports or other similar sources. From an intertextual point-of-view, asylum decisions can be seen as social formations with shared conventions of producing texts (see Reisigl and Wodak 2001: 215–222). Therefore, in this article, the asylum decisions form the texts, that is, the starting point of the analysis and the country reports form the preceding intertexts (see Farrelly 2020). The asylum decisions in the data were made between March 2016 and March 2017, therefore during the time that Finland introduced new legislation and country reports affecting the decisions. The data include only negative decisions as the positive decisions from the time period are very short and may not include any COI at all, making them not very useful for the scope of this study.

The decisions in the data concern applicants from three countries: Iraq, Iran, and Afghanistan. This is due to the availability of the data and the consequent method of data gathering. As asylum decisions are classified as confidential in Finland, the decision data were gathered directly from asylum applicants as part of a grass-root project aiming to map their situation and support them. The applicants came from various parts of their countries and they represent various ethnicities and religions. Their reasons for seeking asylum vary from religion to ethnicity and political opinion. Most of the applicants are men, but some decisions concern families including women and children. Therefore, the decisions represent a fairly varied kind of applicants within the said geographical area. As asylum decisions include personal and sensitive information, they need to be handled with utmost care and respect. During data gathering, each applicant was asked to give informed consent about using their decisions in research and publishing. The decisions are discussed anonymously to protect the identities of the applicants.

The analysis follows the principles of Fairclough’s (1992) critical discourse analysis. In the analysis, I first identified the COI within the decision texts. I analysed the placement of COI within the decisions and the types of sources used. Then I examined the COI in relation to the reasons for asylum presented by the applicant. Finally, I selected certain parts of COI in the decisions for a comparative analysis with the COI reports they refer to. This selection was either done based on the similarity of the COI in the decisions or the clear differences or disconnections. The comparative analysis paid specific attention to the transformations (Fairclough 1992: 130–133), and the presence and absence (Farrelly 2020) in the texts and intertexts.

5. Analysis: the two functions of COI

In this analysis section, I present general findings about networks of preceding intertexts (Reisigl and Wodak 2001: 215–222; Farrelly 2020) as well as their incorporation through recontextualisation (topics, form, placement, and modality) and functions (Fairclough 1992: 119–120) present in the decisions. Preceding intertexts refer to the sources named and used in the texts and tell which preceding intertexts are included in the texts (Farrelly 2020). In the asylums decisions, this is typically done through the use of in-text references—similar to the current article—accompanied with a list of references placed at the end of the decision. The preceding intertexts in the data can be divided into four categories by their social practice and typicality (see Table 1). The first social practice represented in the data is the reports of national or supranational institutions and non-governmental organisations. They are characterised by being a result of lengthy research and are often produced with the specific aim to provide COI. A second social practice is mainstream media articles and blog texts by independent authors or think tanks. Even though they vary on how official they are, they can in the current context be seen as part of the same social practice because they are often reactionary and fast and are originally produced for other purposes—such as public information sharing—than providing COI. As for typicality, the two most often cited preceding intertexts are reports by national or supranational institutions, such as the national COI units or the United Nations, as well as mainstream media articles. In turn, the two more atypical sources cited are COI reports by NGOs, such as human rights organisations, as well as blogs and think tanks. This indicates that the asylum decisions especially form intertextual networks with intertexts from governmental organisations and mainstream media, rather than those of NGOs.

Table 1.

Social practice and typicality of COI in the asylum decisions

		SOCIAL PRACTICE
		Reports	(Social) media texts
TYPICALITY	Typical	Reports by national and supranational institutions	Mainstream media texts
	Atypical	Reports by NGOs	Blogs by independent authors and think tanks

Incorporation addresses the question of how the preceding intertexts are recontextualised in the decisions, affecting the kind of meanings constructed through intertextuality (Fairclough 1992: 119–120; Linell 1998; Farrelly 2020). The topics of COI discussed in the asylum decisions cover a wide array, including ethnicity, religion, work, armed groups, governmental officials, and honour violence as well as general security and societal situation in the area of origin. As for form, COI has a tendency to repeat identically or nearly identically in various decisions regardless of the decision maker. This indicates that the COI has been copy–pasted into the decisions (Dauvergne and Millbank 2003; Pettitt, Townhead and Huber 2008), which may serve the efficiency and uniformity of decision making (Tiililä 2000). The COI is typically placed in independent paragraphs alternating with the other information, such as the rephrasing the applicant’s narrative, and the judicial assessment. Within the decision, COI can be placed in three different sections: ‘facts’ (i.e. as part of assessing if the applicant is personally at risk of persecution); ‘judicial assessment’ (i.e. as part of evaluating the general security situation and living conditions in the country or area of origin); or ‘internal flight alternative’ (IFA, i.e. to evaluate, in case their home area has been deemed too unsafe, whether they can be obliged to seek safety in another area within their country rather than in the country of destination, see Schultz 2019). These are the names of the sections used in the decisions. Based on the content, ‘facts’ are about the credibility assessment based on both the narrative of the applicant and COI and, consequent to those, the future risk of persecution. ‘Judicial assessment’ is about reflecting the credibility and risk against the legislation in Finland (specifically the Aliens Act, 2004) and whether those make the applicant eligible to asylum, subsidiary protection, or—prior to the legal changes in May 2016—humanitarian protection.

Modality indicates the certainty of information (Fairclough 2003: 170–171), and the COI in the decisions presents a strong certainty. This is visible in three ways. First, the lack of modality markers in the text, words such as ‘can’ or ‘perhaps’, which would indicate uncertainty or hesitation. Secondly, the lack of contradictions or alternate information in the COI paragraphs, implies that the sources are by large in agreement. Thirdly, the strings of sources following the COI paragraphs, that is a paragraph can be followed by over 10 references, implying that the information presented is packed up by various sources, hence increasing its certainty. These three aspects are visible in the following example.

The situation in Baghdad has stabilised considerably since the civil war phase of years 2006–2008. The armed group calling itself the Islamic State (ISIS), which attacked Central Iraq, has not attempted to seize Baghdad. Its cells have, however, managed to conduct terror attacks, such as assassinations, kidnappings, and roadside bombings, around the Baghdad Governorate. (Landinfo 2015; Migri 2015; UNSC 2015a; Upper Tribunal 2015; Washington Institute 2015.)¹

The high modality indicates that the decisions present the COI as an indisputable and objective fact, even though this is actually not the case, as has been argued by previous research (Dauvergne and Millbank 2003; Good 2004; Johannesson 2012; Gibb and Good 2013). Therefore the COI in the asylum decisions is constructed as illusions of objectivity, as it is presented as objective, yet in fact formed by choices of inclusion, exclusion, and reformulations.

Based on the presented analysis, I suggest that COI has two different functions in the decisions referred to here as personal and general (see Table 2). While the types of preceding intertexts, form, and modality tend to be similar in both, the two functions are differentiated by the topics and the placement of the COI. The topics of personal COI are related to the applicant’s reasons for seeking asylum (e.g. ethnicity, religion, and work), the events narrated by them (e.g. threats, violence, and torture), and the reported persecutor (e.g. governmental actors and armed groups). The placement of personal COI is typically in the facts section, which rephrases and assesses the persecution narrative of the applicant, and the COI paragraph tends to immediately follow the retelling of the narrative and be followed by the assessment of the Service of the truthfulness of the narrative. Therefore, personal COI is strictly connected to the persecution narrative of the asylum applicant and its assessment. In turn, the topics of general COI are not related to the applicant’s narrative but rather to their geographical background—their country and area of origin. They may include known events of violence, legislation as well as other societal developments. General COI can be placed either in the facts, judicial assessment, or the IFA section of the decision. It tends to be placed independently, meaning without a clear link to the preceding paragraph, and in clusters with other paragraphs of general COI. For instance, the excerpt above about the situation in Baghdad, can variably be placed after either in the facts (after denying the credibility of the applicant’s persecution narrative and therefore risk of future persecution); in the judicial assessment (after denying that the applicant has well-founded grounds for asylum); or in relation to the internal flight alternative (after accepting that the applicant is in risk in their area of origin but assessing whether they can live somewhere else, in this case, Baghdad). The example is usually followed by one or more paragraphs about the general situation in Baghdad. Therefore the function of personal COI is to assess whether the applicant’s persecution narrative is credible and whether as a consequence they are likely to face a personal risk of persecution. In turn, the function of general COI is to assess—in the case, the personal risk of persecution has not been found likely (see Dauvergne and Millbank 2003: 305; Affolter 2021: 85)—whether the applicant faces a general risk because of the security situation in the country of origin and therefore whether the country is safe enough to return. The latter refers specifically to looking at the levels of generalised violence and the humanitarian situation to determine whether the deportation of the applicant, despite the lack of accepted risk of personal persecution, would break the principle of non-refoulement. These two functions of COI will be examined further in the following subsections through case examples and the concepts of transformations, and the presence and absence, respectively.

Table 2.

Personal and general COI in the asylum decisions

Function	Topic	Form	Placement	Function
Personal	Related to the applicant’s narrative, for example, reasons and events of persecution, persecutor	Repeated identically only in decisions of applicants with similar reasons of persecution or persecutor	In the facts section, immediately following the retelling of the narrative	To help assess the credibility of the persecution narrative and the personal risk of persecution
General	Related to place, that is, the applicant’s country or area of origin	Repeated identically in various decisions regardless their reasons of persecution or the persecutor	In the facts, judicial assessment or IFA sections, independently or clustered with other general COI	To help assess the general risk and whether the area of origin (or return, in case of IFA) is safe enough to return

5.1 General COI: transformations

Transformations refer to the ways in which the contents of the text change from the preceding intertext to the text (Fairclough 1992: 130–133; Farrelly 2020). To look at transformations in general COI, I use the case of Baghdad as an example, because Baghdad represents a clear change that happened during the period covered by the data. In the beginning of the period, the Finnish Immigration Service considered the city of Baghdad too dangerous for anyone to return. The statement made by the Service—discussed in the beginning of the article—marked a change in this view, as according to them the security situation had improved. The presupposition is that this change should also be visible between the COI reports of the Country Information Service and between the COI in the decisions. Therefore, I first look at transformations between the COI reports and then between the COI in the decisions.

In the COI reports, there is a clear transformation. During the period affecting the decisions in the data, the Country Information Service published two COI reports about Iraq, the first in October 2015 and the second in May 2016. The Finnish Immigration Service made a statement about the improvement of the security situation following the second report. Both reports are based on desk research and include overall information about the security and humanitarian situation in the country, including a section about the Baghdad area. In the 2015 report, the section about Baghdad is approximately one page long and discusses security forces, terrorist and other violent attacks as well as everyday life. In the 2016 report, the section about Baghdad is considerably longer, over two pages, and discusses violent deaths and terrorist attacks as well as restaurants, beauty pageants, sports, and culture. Indeed based on the contents, there was not a substantial change in the security situation in Baghdad between the two reports; the May 2016 report states that ‘the security situation in Baghdad has not essentially changed within the recent months’ (Finnish Immigration Service 2016b, emphasis added).

A transformation does, however, occur not in the information related to the security situation but in the topics covered in the reports. As can be seen in Table 3, whereas the 2015 report briefly addresses everyday life in one paragraph (paragraph 4), the 2016 report extensively discusses it in altogether four paragraphs (paragraphs 8–10). Indeed, in the 2016 report, everyday life (restaurants, beauty pageants, and sports) are presented side-by-side with the security situation (violent attacks and their victims) and given almost equal space. Dauvergne and Millbank (2003) suggest that this kind of approach to COI may be used to present the country of origin as safer than it actually is. They describe how, in the case of sexual minority asylum applicants in Australia, COI about existing gay clubs or cruising locations was included while COI about the risks and dangers related to such venues or behaviour was excluded. This may suggest that by emphasising the everyday life aspects and giving them almost equal space as the security situation, the 2016 report de-emphasises security threats in order to give a safer image of the area.

Table 3.

Topics in the 2015 and 2016 COI reports about Baghdad

Paragraph	2015 report	2016 report
1	General information	Demonstrations
2	Situation during previous civil war	Violent deaths
3	Situation after the civil war	Terrorist attacks
4	Security forces	Terrorist attack victims
5	Terrorist attacks	Terrorist attacks
6	Other violence	Violent attacks
7	Everyday life	Everyday life: restaurants
8	Situation on the Green Zone	Everyday life: beauty pageants
9		Everyday life: sports and culture
10		Everyday life: restaurants
11		Conservative power struggle

In contrast, in the COI in the decisions, there is almost no visible transformation. In the beginning of the period, the COI about Baghdad includes four paragraphs, repeating almost word to word in all the decisions. Any changes are mainly superficial, such as changing ‘in the neighbouring municipalities surrounding the city of Baghdad and their many towns and villages’ to ‘outside the city [of Baghdad] in the so called Baghdad Belt area’. The only significant change is adding a fifth paragraph discussing the growing number of bomb attacks, indicating rather deterioration than improvement of the security situation. Indeed none of the decisions referred to the May 2016 report, meaning that even several months after the report had been published, the decisions still referred to its predecessor, the October 2015 report.

The analysis of both the reports and the decisions indicates that there was actually no significant change in the security situation in Baghdad. However, after the Finnish Immigration Service made its statement about the improvement of security, there was a considerable change in the position of Iraqi asylum seekers in Finland: whereas in the latter half of 2015, approximately 60 per cent of them were granted international protection, in the summer of 2016, the percentage had dropped to 25 (Finnish Immigration Service 2022). If this was not explained by the improved security situation, what could explain it? A likely explanation is the changes made to the Finnish Aliens Act. That is, until May 2016, the Act included a category of humanitarian protection, which could be granted, among other things, if the person was unable to return home due to the weak security situation. Until May 2016, Baghdad was considered one of such areas, and people originating from Baghdad were therefore granted at least humanitarian protection. After the category was removed from the law, they were no longer entitled to protection (asylum or subsidiary protection, both included in the 25 per cent above) but were instead rejected. This was not considered a breach of the non-refoulement principle, since the coverage of humanitarian protection had been wider than strictly required by international agreements and since the applicants fulfilling the criteria of non-refoulement could technically be granted subsidiary protection instead (Proposal of the Government 2016).

Therefore I suggest that the statement of the Service about the improvement of security in Baghdad was not a matter of fact but rather an example of the symbolic use of COI. This means, following Rosset, Liodden and Maia (2015), using COI for public policy publicity in order to deter and discourage asylum-seekers from seeking protection rather than describing the actual situation in the country of origin. They describe how—very similar to the Finnish Immigration Service—the Ministry of Justice of Denmark in 2014 published a press release stating that it would become more difficult for Eritreans to get asylum, only to be later shown that the statement was based on incomplete and selective information. Even though there are differences between the two cases, they both can point to the possible drawbacks of government-produced COI: as governments are not independent parties in asylum matters, the information produced by them may not be unbiased (Dauvergne and Millbank 2003: 313).

5.2 General COI: presence and absence

This section looks at the presence and absence (see Farrelly 2020) of general COI. This refers to the way the information in the COI reports is recontextualised in the decisions and specifically what parts of them are selected to be presented and how (Linell 1998). Here I discuss the case of Afghanistan, as it provides a clear example of the presence and absence. It also shows a contradiction to the previous example of Baghdad, as COI about Afghanistan usually concentrates on the country as a whole rather than its cities or provinces. I start by looking at the presence and the absence of the decisions vis-à-vis their preceding intertexts, and then focus on one particular case of recontextualisation between the intertexts related to the internal flight alternative.

The decisions typically dedicate four paragraphs to the general COI about Afghanistan and thoroughly discuss the security situation, describing it as ‘fluctuating’.

The conflict in Afghanistan presents wide geographical and temporal fluctuations. In 2015, the conflict claimed over 11,000 civilian casualties (including those injured), which is 4 per cent more than in 2014 and the highest number since 2009. Still the number of civilian deaths decreased a bit from the previous year.

The information about the conflict and its casualties are indeed discussed in many COI reports, and the decisions make this present. The information is essential since an asylum-seeker cannot be deported if there is reason to believe they would be in serious danger upon return. However, the COI sources used in the decisions do not state that the security situation is fluctuating, but that it is deteriorating. This is an essential difference of recontextualisation between the texts and the intertexts since fluctuation indicates that although the security situation is sometimes deteriorating, sometimes it is improving. For example, the preceding intertext for the example above, a COI report by the United Nations Assistance Mission in Afghanistan (UNAMA 2016: 1), describes the situation as follows.

In 2015, the conflict in Afghanistan continued to cause extreme harm to the civilian population, with the highest number of total civilian casualties recorded by UNAMA since 2009. Following increases in 2013 and 2014, civilian deaths and injuries from conflict-related violence increased by 4 per cent compared with 2014. Between 1 January and 31 December 2015, UNAMA documented 11,002 civilian casualties (3,545 civilian deaths and 7,457 injured), marking a 4 per cent decrease in civilian deaths and a 9 per cent increase in civilians injured.

The decision starts by describing the situation in Afghanistan as fluctuating, whereas UNAMA describes it as continuing to cause ‘extreme harm’. In the following clause, the decision states the number of civilian casualties, their increase as well as this being at its highest for about five years. These same facts are found in the following clauses of the UNAMA report, although UNAMA also mentions that the number of civilian casualties has continued to rise the third year in a row—something left unmentioned by the decision. The decision concludes by emphasising that civilian deaths have decreased. This information is also found in the UNAMA report but, unlike the decision, UNAMA also states that the number of civilians injured has instead increased. By these small changes—talking about fluctuation rather than deterioration and emphasising the decrease in civilian deaths while overlooking the increase of those injured—the decisions recontextualise the situation in Afghanistan to seem less grave than what it is according to the intertexts.

Indeed, in many of the texts in the intertextual chain, the deterioration of the security situation is one of their key messages. This is indicated by placement, that is placing the information right in the beginning of the report or in the executive summary. This also constructs one of the main absences found in the data. That is, even though the security situation plays a key role in the COI reports, the decisions do not necessarily use them to discuss it, but instead use them to describe health care or the judicial system, for example. In practice, this means that the reports are named and listed as references in the decision, but their central message is left absent. This is what Farrelly (2020) has referred to as naming texts. According to him, naming can be an attempt to acknowledge the source of information, to show that it has been included and used, without actually including the main message of the source—in the current case, the deterioration of the security situation.

Similar play between the presence and absence can also be found between the preceding intertexts and their development over time. This is especially visible in the internal flight guidelines of the Finnish Immigration Service, which are separate from but based on the guidelines of the UNHCR. As the Service (2018) has stated: they closely follow the guideline of the UNHCR. There seems to, however, be a clear difference between the UNHCR guideline and the recontextualisation of the Service (see Table 4). In 2016, the Service considered the internal flight alternative to be generally available to Kabul, the capital of Afghanistan. However, at the same time, UNHCR considered IFA to be available to Kabul only for people with a support network or, without a support network, for single able-bodied men and married couples of working age. In practice, this means that the requirements for IFA were much more strict according to the UNHCR than to the Service. Furthermore, when UNHCR updated its guideline in 2018, the relationship between the guidelines remained similar: whereas UNHCR stated that IFA was generally not available in Kabul any more, the Service instead posited that IFA was available with a support network for married couples or without it for single able-bodied men. This indicates that the Service is selective in its recontextualisation of the UNHCR guideline, making certain aspects present and others absent and consequently ending up with a distinctively different perspective on the plausibility of IFA in Kabul.

Table 4.

Internal flight to Kabul by the Finnish Immigration Service and UNHCR

Year	UNHCR guideline	Finnish Immigration Service guideline
2016	IFA available: with a support network or without a support network in the case of single able-bodied men and married couples of working age	IFA available with or without a support network
2018	IFA not generally available (regardless of support network)	IFA available: with a support network in the case of married couples or without a support network in the case of single able-bodied men

5.3 Personal COI: presence and absence

Personal COI refers to the information that is directly related to the applicant’s narrative and reasons of persecution. The most prominent feature of personal COI in the data is its relational absence. That is, the decisions typically include far less personal than general COI, with altogether 12 decisions entailing no personal COI at all. The absence of personal COI manifests in two ways: either by excluding COI related to the personal situation of the applicant altogether or by presenting general COI as personal COI.

A poignant example of the complete absence of personal COI is a decision related to a reported blood feud in Afghanistan. Blood feuds can be defined as members of a family killing members of another family in retaliatory acts of vengeance, often triggered by murders of a member of one family committed by a member of the other (UNHCR 2016: 78–79). Personal COI is typically placed immediately after retelling of the persecution narrative to which it is linked, and before the assessment of credibility by the decision maker. In the example, the decision retells what the applicant has told about the blood feud, but instead of continuing with COI, the decision moves directly into the assessment, which reads as follows.

The Immigration Service views that it is possible that your brother was shot in [month year]. The Immigration Service does not however think it is credible that a family that has moved far away from your home village and that has not had anything to do with your family for four years had shot your brother. The Immigration Service views that possible revenge would have been done earlier than four years after your father and [the member of the other family] were killed.

The decision concludes that they do not view it credible that the blood feud would have taken place as the applicant has told. As the main reason, the decision states that the act of vengeance took place four years after the original events. Since the decision does not include any COI about blood feuds, it is not possible to assess what the claim is based on. Indeed available COI about blood feuds in Afghanistan clearly contradicts the argument of the decision as, according to the UNHCR (2016: 78–79) for example, if retaliation is not possible immediately, the blood feud may reactivate after years or even decades after the original events. According to research literature, these kinds of arguments are not rare in asylum decisions. Indeed Johannesson (2012) has found that asylum decisions tend to reflect the personal assumptions of the decision maker. That is, when assessing the way certain kinds of people would behave, this is more often based on the assumptions of the decision maker than the available COI. This is something that Pettitt, Townhead and Huber (2008) have named speculative argumentation, meaning that the decision maker attempts to read the mind of a third party, in the current case the other party of the blood feud. More particularly, Sweeney (2007, see also Liodden 2022) has referred to the idea of a ‘reasonable persecutor’, in which the decision maker argues that a reasonable persecutor would not act in the way described by the applicant, and consequently, the applicant’s narrative must be false. However, as Sweeney points out, this argument tends to be based not on facts but on the decision maker’s assumptions on how the persecutor could have executed the acts of persecution better. These assumptions and even contradictions highlight the relevance of personal COI in asylum assessment and also the possible problems related to its absence.

Absence through using general COI as personal COI is visible in an example made for an applicant from Baghdad, reportedly being a target of a bomb attack approximately five years earlier. In the decision, the retelling of the applicant’s narrative about the attack is followed by a paragraph of COI, which is in turn followed by the assessment of credibility by the decision maker. The placement of the COI indicates that the COI is used as a personal COI as it follows the recounting of the narrative and is used to argue against the credibility of the applicant.

Considering also the country information above, the Immigration Service does not accept it as a fact that the bomb was targeted at you but views that the blast has been a consequence of the bad security situation of the area at the time.

However, the content of the COI paragraph ‘above’ is one of the paragraphs used as general COI in almost all the decisions made for Baghdadis, describing the security situation in the city at the time the decision was written. As the events narrated by the applicant took place five years earlier, and the COI describes the current situation, the COI is not actually related to the narrative of the applicant. Therefore even though the placement of the COI indicates that it is used to assess the applicant’s narrative of persecution, the information provided by the COI is not actually relevant to this. Therefore, personal COI—that is, COI relevant to the applicant’s persecution narrative—is in actuality absent in the decision. In general, the absence of personal COI may be due to time limitations, which may encourage or force the decision makers to use the minimum number of COI (Pettitt, Townhead and Huber 2008). Even though it may be difficult to find COI that would directly link to the individual claim, the analysis shows that the lack of personal COI is not only because of this, since there is a variety of personal COI between very similar decisions. For instance, even though the example above about a blood feud lacked any COI about blood feuds, other similar decisions included COI about the topic. The current example, in turn, presents a disconnection between the applicant’s narrative and the use of COI because the credibility of the applicant’s narrative is assessed based on COI irrelevant to their narrative. As the same COI is repeated in verbatim in other decisions to provide general information, it is likely that the COI paragraph has been copy–pasted to the decision and used despite the lack of connection between the narrative of the applicant and the content of the COI.

5.4 Personal COI: transformations

Large and significant transformations in COI, including personal COI, are scarce in the decisions. This already became apparent in the lack of transformations related to the general COI about Baghdad, as discussed earlier. The data do, however, include one exception where the transformation of personal COI is loud and clear: the personal COI related to apostasy, meaning giving up Islam, in Iran. The data include two decisions where such COI is included, and there is a very clear distinction between them. In the decision made chronologically first, the COI paragraph is as follows.

Giving up Islam aka apostasy has not been written in the Penal Code of Iran, but it is still possible to sentence for apostasy based on the Constitution. Paragraph 167 of the Constitution states that in the absence of written law the judge must rely on the principles of Islamic law. Paragraph 220 of the Penal Code confirms this (Constitution of Iran 1979, p. 167: Penal Code of Iran 2013, p. 220, IHRDC 2014). There has been no perceivable improvement in the situation of Christians during the term of President Hassan Rouhani. The most typical encroachments against Christians have been arrests, prison sentences, and criminal sentences. The arrests have targeted especially pastors who organise unofficial home church activities and conduct missionary work (ICHRI 2013; DIS 2014).

The example presents four statements. Firstly, it is possible to convict for apostasy in Iran. Secondly, there has been no improvement in the situation of Christians since the inauguration of the new president considered more progressive in 2013. Thirdly, Christians face arrests as well as imprisonment and criminal sentences. Fourthly, that especially pastors but also other Christians can face encroachments. This inclusion of other Christians is indicated by the use of the word ‘especially’, implying that particularly but not only priests face encroachment. This interpretation is confirmed by the COI reports used as sources in the paragraph, as they clearly indicate that also regular Christians may face arrest or harassment (e.g. DIS 2014).

The second decision, made a few months later, shares the first three sentences of the previous example, but then the paragraph continues:

The only person to have been executed for converting to Christianity is pastor Hosein Soodmand of the Jama’at Rabbani congregation, who was hanged in Mashhad in 1990. After Soodmand, the only apostasy sentences confirmed by the appellate organ have been given to Mehdi Dijab in 1994 and Yusef Nadarkhani in 2010. The sentence of Nadarkhani was however later changed to sentence for endangering public safety, more typical for Christian converts, and Nadarkhani was released after three years in imprisonment. (CSW 2016, DIS, Landinfo and DRC 2013, ICHRI 2013, IHRDC 2014)

The second example shares the first fact of the first example: conviction for apostasy is possible according to the law. However, the following statements are very different. According to them, only one person ever has been executed for apostasy, and two more have been given a sentence, but one of them has since been released. This would in essence indicate that apostasy does not cause problems in Iran, as only three people have been convicted for it.

The reason for this transformation within a span of a few months is unclear. No new significant COI reports were published nor are such referenced in the later decision. The analysis indicates that there are three intertextual factors to the change: a change of an intertext, shifting the focus of the texts, and the absence and presence of information between the intertext and text. Both examples refer mainly to the same sources of information, as can be seen in Table 5. Yet, there is one significant change of intertext: the first decision refers to a 2014 report by the Danish Immigration Service (DIS), but the second decision instead refers to an older 2013 report published jointly by DIS, The Norwegian Country of Origin Information Centre (Landinfo) and Danish Refugee Council (DRC). In comparison to the other quoted reports, the added 2013 report is quite different in its content. That is, all the other reports clearly indicate that even though ethnic or born Christians may not be persecuted in Iran, Christian converts are. However, the 2013 report does not make this distinction, instead is collectively referred to as ‘Iranian Christians’ or ‘Christians in Iran’. For this reason also the conclusion of the report is more ambiguous than the conclusion of the other reports: according to the 2013 report, some sources say Christians are persecuted, while others say they are not. Essentially adding this source in the more recent decision allows the COI to argue that not all Christians are persecuted in Iran, but without explaining the difference between the situation of ethnic Christians and Christian converts.

Table 5.

COI intertexts about apostasy in Iran

	Example 1	Example 2
Coherent intertexts	Constitution of Iran 1979	Constitution of Iran 1979
	Penal Code of Iran 2013	Penal Code of Iran 2013
	ICHRI 2013	ICHRI 2013
	IHRDC 2014	IHRDC 2014
Transformed intertexts	DIS 2014	DIS, Landinfo and DRC 2013
		CSW 2013

Shifting the focus is visible between the COI paragraphs in the decisions. Whereas the first example talks about a wide array of persecution—arrests as well as prison and criminal sentences—faced by Christians, the second discusses only apostasy sentences. This implies a presupposition that only those who are officially convicted of apostasy are in fact persecuted for their religion. This ties in with the absence and presence of information, as in the second example the decision uses the lack of sentences for apostasy to indicate that those who have given up Islam do not face persecution in Iran. This is not untrue in itself, as is revealed by the COI reports. However, what the reports tell is that the lack of convictions for apostasy sentences is not because there would be no persecution, but because Iran is using political sentences such as spying or endangering public safety rather than the religious sentence of apostasy to convict converts (e.g. DIS 2014: 7–8). This explanatory information is, however, absent from the decision, therefore, creating an illusion that Christian converts would not be persecuted even though the COI reports state the contrary.

6. Conclusion

The article started from the presupposition that COI is by its nature intertextual. Through analysing the intertextuality of COI it argues that COI has two functions in asylum decisions: general and personal. General COI refers to information about the security and human rights situation in the country or specific area, whereas personal COI is related to the applicant’s reasons and narrative of persecution. Furthermore, the article has shown how transformations as well as presence and absence can be used to recontextualise COI so that certain areas can seem less dangerous and certain reasons for persecution less likely. Although COI is by its very nature not objective (Gibb and Good 2013), the decisions nevertheless portray it as if it was objective; as fact-like and without a hint of uncertainty or contradictions, even though it can easily be biased, irrelevant and even illogical, as shown in the analysis. Therefore, the COI in the decisions is not objective but forms an illusion of objectivity. The question of illusions is important because legality does not automatically mean legitimacy (Reisigl and Wodak 2001: 205–206). That is, the data are composed of legal decisions meeting the requirements of the law, but the analysis implies that their legitimacy may be put to question, as the justifications of the legality are less than robust. The idea of objectivity, therefore, aims at legitimising the legality of the decisions, but the idea of illusions contests this. This means that even though the asylum decisions may be legal—and may have even been confirmed as such at the appeal organs—they may nevertheless lack legitimacy.

This article has mainly addressed manifest intertextuality, but COI is also a matter of interdiscursivity. According to European Asylum Support Office EASO (2019), COI should be fact based and unbiased, aim for objectivity, be relevant to the matter, as well as logically and factually sound. The current analysis shows that this is not always the case with COI in the asylum decisions. Quite the contrary, the COI in the decisions may be selectively chosen to create a bias, lack relevancy to the matter in question, and also create logical fallacies in the asylum assessment. Therefore, the COI in the decisions does not seem to draw from the aim of objectivity and impartiality, but rather from a different genre, that of knowledge politics of COI (van der Kist, Dijstelbloem, and de Goede 2019), wherein its function is not to inform asylum decision making but to send a public message about the local asylum politics to current and prospective asylum applicants. As such the themes of the article are not only related to forced migration but also to the wider discussions about human migrations and the attitudes and public policies related to them and, discursively, the importance of recontextualisation in official documents and regimes of border control.

Acknowledgements

The author wishes to thank all the people who shared their asylum documents and helped to collect them, as well as the Deportation and Resistance workshop participants of the Nordic Migration Research (NMR) Conference and ETMU Conference at the University of Helsinki in 2021.

Funding

This work was supported by the Finnish Cultural Foundation [00210242].

Conflict of interest statement. The author has worked and volunteered in organisations assisting asylum applicants in their asylum procedures.

Footnotes

1

The translations from the original Finnish language material have been made by the researcher.