(Socio)linguistic and Narrative Challenges in the Franco-Dutch Asylum Procedure - Implementing Language Analysis in France

(1)

(Socio)linguistic and Narrative Challenges in the Franco-Dutch

Asylum Procedure – Implementing Language Analysis in France

MA Thesis in Literary Studies: Literature, Culture and Society – French

Graduate School of Humanities

Universiteit van Amsterdam

Flora Elisabeth Rasmussen McQuibban

12340596

Under the supervision of Dr Rina de Vries Dr Caroll Clarkson

(2)

ABSTRACT

This thesis focuses on sociolinguistic, linguistic, and narrative challenges occurring within the Belgian, Dutch, French and Swiss asylum procedures, with a particular focus on the complex nature of Language Analysis in the Asylum Procedure (LAAP) and Language Analysis for the Determination of Origin (LADO). The Franco-Dutch framework was chosen as these countries have (1) a high influx or intake of immigrants and refugees (France and Belgium), and/or (2) an in-house language analysis unit tasked with verifying asylum seekers’ claimed country of origin (the Netherlands and Switzerland). Sociolinguistic and narrative pitfalls of language analysis, and conversely, the pitfalls of asylum procedures that have no standardised method of performing language analysis, are investigated by analysing and critiquing existing literature, which includes transcribed interview recordings, and examining responses to questionnaires sent to the Immigration and Naturalisation Service (INS/IND) in the Netherlands, the French Office for the Protection of Refugees and Stateless Persons (OFPRA) in France, and the Swiss State Secretariat for Migration (SEM) in Switzerland. Belgium was not sent a questionnaire given the limited scope of this paper, not least because language analysis is reportedly largely unused today in the Belgian asylum procedure (Patrick 2016). Ultimately, this project demonstrates that analysing an applicant’s (socio)linguistic and narrative features for the purpose of verifying their country of origin is beneficial to applicants as well as the asylum procedure, in cases where the practice is carried out in a highly informed manner and with great caution, and suggests the implementation of formal language analysis in the French asylum procedure.

(5)

1. INTRODUCTION

1.1. Asylum and language analysis

In the European Migration Network’s most recent Anniversary Report, “immigration” was listed as the most important issue according to Europeans in late 2017 – marginally above terrorism (EMN 2018, 4). In 2018, approximately 66 million people were registered by the United Nations (UN) as displaced, with 24 people fleeing their homes every minute in order to escape persecution or war (Edwards 2016, 411-22; Nick 2018, 180).

Although immigrants, migrants, and asylum seekers can all fall under the EMN Report’s “immigration” label, asylum procedures rely largely on the meticulous sifting of applications to determine who might have a demonstrably well-founded fear of persecution, and who might be attempting to enter the country in search of better (socio)economic opportunities. Part of the process usually consists of one or more interviews which serve the purpose of assessing the applicant’s narrative and reason(s) for seeking asylum, and verifying their claimed identity; i.e. their claimed country of origin. Academics have highlighted a number of problems in asylum application interviews (Blommaert 2001, Corcoran 2004, Maryns 2006, Eades 2010, Jacquemet 2011, Verrips 2011, Gómez Díez 2013, Channon et al. 2018, McNamara et al. 2018, Muysken 2019, Patrick 2019), including hiring untrained officials who lack a basic knowledge of regional linguistic and cultural training.

In order to tackle these challenges, and to add yet another means of assessing claims to asylum, some Member States’ governments have set up language analysis units tasked with performing Language Analysis in the Asylum Procedure (LAAP), or Language Analysis for the Determination of Origin (LADO), or both. By using linguists, native speakers, and forensic specialists, institutions hoped to improve the quality of their analyses and strengthen the

(6)

language analysts are likely to encounter (sub)conscious difficulties related to language or culture.

This thesis analyses sociolinguistic, linguistic, and narrative challenges – occurring within the Belgian, Dutch, French and Swiss asylum procedures – that, for example, occur due to lack of, or improper use of, language analysis (LAAP/LADO). Whereas the Netherlands and Switzerland are the only countries that currently have language analyses done by specialised in-house agencies (Cambier-Langeveld 2018, 4), and Belgium has a research and country of origin unit (once) including a language desk (Maryns 2004, 249-50), France is the one major Franco-European1_{country that can at best be described as using informal language analysis,}

with no language analysis desk and no LAAP or LADO used. This project hopes to demonstrate that language analysis can be beneficial to asylum procedures, and that the absence of accredited formal language analysis can be just as harmful as ill-informed language analysis.

The investigative nature of this project focuses on analysis of transcribed asylum interviews provided by existing literature, and critiques of the aforementioned asylum procedures, supplemented by responses to questionnaires sent to the Immigration and Naturalisation Service (INS/IND) in the Netherlands, French Office for the Protection of Refugees and Stateless Persons (OFPRA) in France, and the Swiss State Secretariat for Migration (SEM) in Switzerland. Specific questionnaire responses are examined in detail in Chapter 6.

1.2. Theoretical framework

This project operates within several theoretical frameworks, including sociolinguistics and Labov’s variationist sociolinguistics, Giles’ accommodation theory, Preston’s view on folk

1_{Luxembourg and Monaco’s respective asylum procedures are not treated in this thesis as the}

(7)

linguistics and perceptual dialectology, Blommaert’s ethnocoherence, and Jacquemet’s definition of entextualisation.

Sociolinguistics posits that social factors, such as class and gender for example, influence language and speech. Variationist sociolinguistics suggests that linguistic change follows a highly structured hierarchy that is based on social factors (Labov 1994); not only does language use differ from country to country, but people from the same country or speech community use language differently (intra-speaker variation), and the same person’s language or speech can vary (sub)consciously depending on the context (Maryns 2004, 256).

Accommodation theory posits that speech and language not only depend on context, but follow a strict structure in which, for example, subordinates modify their speech in the presence of superiors. This could take the form of upward convergence; an interviewee speaking a dialect might (sub)consciously take on the interviewer’s standard linguistic variation and accommodate to them (Bell 1984, 145; Giles et al. 2007, 40).

Perceptual dialectology refers to the ways in which non-linguists and untrained (lay)people or native speakers perceive dialects and linguistic variation (Preston 1999); folk linguistics is its pejorative umbrella term, often used in reference to the idea that (lay)people without training in (socio)linguistics use a basic understanding of how language works and evolves (Labov 1994, 186), which can lead to incorrect stereotypes and assumptions.

Finally, ethnocoherence is the idea that the coherence of a narrative is also variable and ultimately dependent on (socio)cultural background, diverging narrative traditions, and context (Blommaert 1994; Blommaert 2000, 5), and entextualisation is the removal of a narrative from its original context so as to place it within another framework.

(8)

1.3. Structure

Chapter 1 introduces the thesis and is followed by a deeper contexualisation of the refugee crisis, language analysis, the right of asylum, and the initial procedural issues that led to adopting Language Analysis in the Asylum Procedure (LAAP) and Language Analysis for the Determination of Origin (LADO) in Chapter 2. Chapter 3 makes use of the abovementioned theoretical framework and questionnaires, and presents an extensive analysis of the (socio)linguistic problems that can and frequently do occur in standard interviews, language analysis interviews and reports. Chapter 4, discusses contributing narrative and institutional challenges. Chapter 5 examines The Guidelines for LADO, and analyses the debate on who is qualified to perform language analysis for the purpose of determining origin. Chapter 6 is purely empirical in nature and presents a critical analysis of the information provided by the personalised questionnaires, and aims to hint at the cautious implementation of language analysis in France. Finally, arguments set forward by academics and professionals in the field on how to ‘move forward’ are summarised in the concluding remarks, and additionally, supplementary insight into how persisting hurdles might be tackled are proposed.

1.4. Research methods and limitations

Personalised questionnaires were sent to the Office for Country Information and Language Analysis (OCILA) in the Netherlands, the Specialised unit for analyses of origin (LINGUA) in Switzerland and the French Office for the Protection of Refugees and Stateless Persons, or

Office français de protections des réfugiés et apatrides (OFPRA) in France, in order to

investigate, update and add to existing literature, and in the case of France, answer pending questions concerning LAAP and LADO. A questionnaire was not sent to the Documentation Centre for the Different Instances of Belgian Asylum, or Centre de Documentation pour les

(9)

Différentes Instances d’Asile Belge (CEDOCA) in Belgium. Research conditions and

(10)

2. CONTEXTUALISING

ASYLUM

AND

LANGUAGE

ANALYSIS

2.1. The European migrant crisis

The European migrant crisis, or the refugee crisis, began in 2015 when the number of migrants entering the European Union (EU) rose dramatically; here, ‘migrants’ refers both to (socio)economic migrants and asylum seekers. Reasons for migrating to the EU include poor living conditions, poverty, famine, discrimination, violence, war, and even climate change (The

Guardian December 21 2017). European Parliament confirms that in 2015, over one million

asylum seekers and migrants arrived in Europe, and that in 2016, there were almost 1.3 million applications for asylum in Europe. This number decreased in 2017 to 728,470 applications, and of the 96,100 Syrian citizens granted protection in the EU, 70% were taken in by Germany (European Parliament 2017).

It is worth noting that most refugees are still predominantly hosted by countries outside of the EU (Figure 1); however, in 2017, Eurostat recorded approximately 618,775 “third country nationals” found to be illegally present in the European Union, with 18,285 in Belgium, 115,085 in France, 2,165 in the Netherlands, and 13,940 in Switzerland. These differences in numbers can perhaps be attributed to the lack of standardisation of EU asylum procedures, and the likelihood of asylum seekers applying to countries where the acceptance rate for their country of origin is high (de Volkskrant February 4 2016).

(11)

Figure 1. “Major host countries of refugees”, UNHCR Global Trends Report 2017.

Eurostat’s most recent records for 2018 (Figure 2) show 19,145 “third country nationals” found to be illegally present in Belgium, 105,880 in France, 2,790 in the Netherlands, and 14,420 in Switzerland (Eurostat 2018). As being present illegally in a country can also denote improper registration of a person, or departure from the country in which a claim for asylum is made, these numbers can refer both to illegal (im)migrants and to refugees. It thus becomes clear that EU Member States, and particularly countries that experience high influxes of (illegally present) migrants such as France, are under astronomical pressure in terms of assuring that their respective asylum procedures are carried out properly.

(12)

Figure 2. “Third country nationals found to be illegally present – annual data rounded”, Eurostat June 06 2019.

Despite the EU’s efforts, 439,505 people were denied entry at the EU’s external borders and as of June 26 2018, and more than 43,000 people have risked their lives in an attempt to enter Europe by sea, with thousands missing or feared dead (European Parliament 2017).

On the one hand, most would agree on the quasi-inviolability of providing international protection and recognise the basic humanity in granting asylum to those who are at risk of being killed or tortured, regardless of race, ethnicity, sex, gender, disability or political views. On the other hand, to borrow from Marco Jacquemet, who criticises institutional hostility in the asylum procedure, our society perpetuates an “ideology of mistrust” (Jacquemet 2011,

(13)

480), and one that casts a shadow over all types of migrants. Tackling immigration whilst honouring the right of asylum seems then to be a constant battle between a willingness to help and a deep-rooted mistrust of the Other.

2.2. A framework of mistrust

Immigrants, and sometimes refugees, are routinely mentioned by some media channels and people in positions of power within a framework of fearmongering that applies to several sectors of life; economic (job-stealing), social (profiting from unemployment benefits, difficulties integrating), political (radical agendas with links to terrorism), cultural (important differences in traditions or value systems), ethical or legal (behaviour or actions that might be out of line with the host country’s ethos or laws), and linguistic (unwillingness to learn the host language or unwillingness to communicate all together). This portrayal, led by respected and often professionally accredited people, contributes significantly in shaping public opinion. In November 2017, 54 percent of surveyed European citizens felt negatively about non-EU immigrants, with 20 percent of this same group feeling “strongly negative”. Amongst the 38 percent who felt “fairly positive” about non-EU immigrants, only 7 percent felt “strongly positive” (EMN 2018, 5). Along the same line of thinking, refugees are also at times viewed as a “homogenous” mass of people (Maryns 2006, 341) who lack a legitimate claim to asylum and who industrialised countries fear being “swamped” by (Eades 2009, 30; Eades 2010, 411); they are often branded with derogatory terms such as “queue jumpers,” “illegals,” or “boat people,” terms that have been used by high-ranking politicians, such as John Howard, who formerly served as Prime Minister of Australia (Clyne 2005).

The European Commission on Migration and Home Affairs notes that countries such as Austria, the Czech Republic, Slovakia and Liechtenstein took in less than 50 refugees last

(14)

Hungary, Poland, the United Kingdom, and Iceland took in none at all (EC 2018). In comparison, Belgium took in 1,171 refugees in 2018 as part of the EU’s refugee relocation system, France welcomed 5,030 refugees, the Netherlands received 2,775, and Switzerland took in 1,500 (EC 2018). Out of all the Member States, France came second after Germany in number of relocated refugees, despite already hosting a considerably high number of migrants. The findings provided by the European Parliament and the EMN reports only intensify the necessity of a smooth-running asylum procedure and standardisation of diverging interpretations of Human Rights.

2.3. The right of asylum

Neither immigrants nor asylum seekers ought to be harassed or insulted, but conflating the reasons for their departure – reasons which, and it is worth pointing out, are not always worlds apart – has had a negative effect on how asylum seekers are received. By law, the right of asylum is a human right that is in theory offered to anyone considered a genuine refugee – i.e. any applicant who can demonstrate a “well-founded fear” of being persecuted and who finds themselves outside of their country of origin and is either incapable or unwilling to return due to this fear (UNHCR 2019, 18). The definition is subject to many interpretations and, as stipulated by the European Union’s Qualification Directive (European Union: Council of the European Union 2011), it is up to the asylum seeker themselves to prove their need for protection (Severijns 2019, 235) even though the right should essentially entitle all such persons to be legally protected in an EU country that has signed the 1951 Refugee Convention. Though it is ethically questionable to encourage a hierarchy of suffering (McNamara et al. 2014, 280), or right to protection, it is in this regard that asylum seekers and refugees differ, at least legally, from other influxes of people.

(15)

Although rejected applicants cannot be sent back to a country in which torture or inhumane treatment is occurring (McNamara et al. 2014, 277; Severijns 2019, 234) they can be repatriated, or sent (back) to a country that is considered ‘safe’. The European Commission provides a brochure in which several Member States have listed which countries they consider to be safe (EC 2019); however, these lists all differ from each other to varying degrees and countries are frequently added or removed. Coupled with the unreliability of lists plagued by variability are fears of persecution that are highly individual; preventing a person from being sent back to a war-torn country does not prevent the repatriation of applicants who base their claim on a personal fear that is not applicable to all citizens of their country. The consequences of poorly handling such cases can be disastrous:

In September 1998, a young Nigerian girl called Sémira Adamu died during her forced repatriation. Her asylum application, based on the argument that she would be forced to marry an old and violent man in Nigeria, had been rejected. She was put on a plane for Togo and seated in between two police men; when she started to shout, the police officers put a small pillow over her mouth. Sémira Adamu lost consciousness, went into a coma and died. (Blommaert 2001, 4)

Though not all Nigerian applicants will necessarily be forcefully repatriated, this shocking example nonetheless highlights that defining who qualifies for international protection is a delicate task; individual applicants might have well-founded fears that exceed the limits of those predefined by other nations.

2.4. Verifying claimed country of origin

Despite the lack of standardisation in the EU on how to verify claimed country of origin, most Member States share certain initial procedures that constitute the first steps in verifying claimed country of origin, such as checking the authenticity of “core” documents like passports, and “supporting” documents, like marriage certificates (Cambier-Langeveld 2018,

(16)

without any papers (Eades 2010, 412; McNamara et al. 2014, 263; Cambier-Langeveld 2018, 3); documents may have been stolen or lost at sea, and, undeniably, in some cases withheld by the applicant for a number of reasons. The European Migration Network (EMN) reports that asylum seekers have at times even been advised by lawyers to rid themselves of their documents in order to avoid forced repatriation (EMN 2013, 6). That said, motives for arriving undocumented need not be linked to the applicant’s personal history or journey. Documents may have been confiscated or destroyed by local authorities or by the government, and in some cases, official documents are simply not issued by the country of origin, or rarely made available: “For example, an immigrant entering the Netherlands and filing an asylum application claims to come from a village in south Somalia. Due to the absence of a functioning government, Somalia identity documents are seldom available.” (McNamara et al. 2014, 263). Other methods used to identify applicants, though rare, include fingerprints, photographs, DNA analysis, age assessments by way of radiological tests, social media, smartphones, or e-transactions (Cambier-Langeveld 2018, 3; Severijns 2019, 235). Most of these methods only prove useful, however, if an applicant’s data already exists in another (European) governmental registration database.

2.5. LAAP, LADO and LARP

In light of these persisting issues, EU Member States seem to have turned to language analysis (LA), a sub-field of Forensic Linguistics,2_{as a way of contributing to and improving the}

existing means of authenticating an applicant’s claimed identity. There are several areas of application in the field of Forensic Linguistics from which language can be used as legal evidence, but most relevant to the asylum procedure is a type of speaker identification that uses

(17)

forensic dialectology, or language analysis, in order to evaluate an asylum seeker’s linguistic profile.

Up until the late 1970s, interviews with undocumented asylum seekers relied heavily on the applicant’s narrative to determine the validity of a claim to asylum; however, in the 1980s, institutions were becoming wary of basing their conclusions on emotional testimonies alone (Jacquemet 2011, 480). In the 1990s, a new ‘method’ called Language for the Determination of Origin (LADO) was developed as a response to a number of issues such, as increased migration to Europe and the United States due to the collapse of the Soviet Union, civil wars in Somalia, and a rise in asylum applications (Wilson 2014, 219).

LADO is frequently referred to as Language Analysis in the Asylum Procedure (LAAP), or Language Analysis for Repatriation Purposes (LARP), and the terms LADO, LAAP and LARP are at times used interchangeably, but academics and LADO professionals alike have argued against the conflation of these terms (Reath 2004; McNamara et al. 2014, 277; Cambier-Langeveld 2016, 25-29). Whereas LAAP refers to language analysis used to compare an applicant’s speech to their claimed linguistic profile, it does not make any claims about their country of origin, their nationality, or citizenship. LADO and LARP, on the other hand, refer chiefly to language analysis for the purpose of repatriation, as an applicant’s country of origin must in effect be ‘determined’ for them to be sent (back). Both types of speaker identification examine an applicant’s dialect and its features, either in the initial interview concerned with the applicant’s reason(s) for seeking asylum (Gibb et al. 2014, 388), or in a second and separate interview that is specifically for the purpose of language analysis.

2.6. Interviewers, interpreters, and analysts

(18)

are involved in standard asylum interviews or asylum interviews for the purpose of LAAP, LADO or LARP, can hold several functions, which often intersect. Interviewers ask the questions and might be responsible for writing a summary or report of the interview.

Interpreters are responsible for bilateral translation and facilitating or enabling communication between interviewer(s) and interviewee. Other times, interpreters can be refugees (Jacquemet 2010, 140), and occasionally interpreters will not be used at all, even though they might be indispensable (Jacquemet 2009, 530).

Finally, in some cases analysts examine the applicant’s narrative and reason(s) for asylum, and in others, their linguistic features. Analysts usually write the subsequent interview report. At times, analysts are native speakers who do not make use of interpreters; language analysis performed in these instances is termed “direct”.

2.7. Belgium, France, The Netherlands and Switzerland

Countries that explicitly use LAAP or LADO today are Australia, Austria, Belgium, Canada, Finland, Germany, the Netherlands, New Zealand, Norway, Sweden, Switzerland, and the United Kingdom (Eades 2005, 505-06). Contextualising the Belgian, French, Dutch, and Swiss asylum procedures and their linguistic components provides a firm basis for understanding the ways in which asylum procedures across Europe differ from each other, including the extent to which they incorporate language analysis.

In Belgium, assessing claims made by paperless applicants relied heavily on narrative before the introduction of the Documentation Centre for the Different Instances of Belgian Asylum (CEDOCA) in 2001, which disposes of a Country of Origin Information (COI) unit and a language analysis desk (Maryns 2004). The desk uses interpreters, though applicants have sometimes communicated directly with interviewers in a lingua franca (Maryns 2006, 70), and analysts who listen to a recording of the interview are not linguists but may be assisted

(19)

by linguists (Vanheule 2010, 180). Despite the existence of a COI unit and a language analysis desk, language analysis in Belgium seems to have been “largely discontinued” (Patrick 2016, 252), and the government’s website makes no explicit mention of it.

In the Netherlands, the Immigration and Naturalisation Service (IND) has a specialised unit named Office for Country Information and Language Analysis (OCILA) which analyses recordings of interviews where an interpreter and interviewer attempt to elicit speech samples from the applicant. OCILA uses a trained native speaker analyst and supervising linguist approach, where the analyst conducts the analysis and the linguist supervises, reviews, and signs off on the report together with the analyst (Cambier-Langeveld 2010a). The Netherlands is also the only country that is home to a contra-analysis bureau, De Taalstudio, which specialises in providing contra-expertise in asylum appeals.

In Switzerland, the State Secretariat for Migration (SEM) has a specialised unit called LINGUA, established in May 1997 (Baltisberger et al. 2007, 85; Baltisberger et al. 2010, 9). LINGUA also uses native speaker analysts and linguists, but in 1999 began training their native speakers in linguistics. LINGUA now hires analysts who have postgraduate training in linguistics (Singler 2004, 225), and uses direct analysis where the analyst plays the role of interviewer as well.

In France, the French Office for the Protection of Refugees and Stateless Persons (OFPRA) is in charge of asylum applications. There are no linguists and no formal language analysis ‘methods’ involved in France’s asylum procedure; instead, a recorded interview is done by a case officer, most often assisted by an interpreter, and the case officer forwards their proposed decision to their supervisor (Gibb et al. 2014, 288).

(20)

3. (SOCIO)LINGUISTIC CHALLENGES IN THE ASYLUM

PROCEDURE

Geolinguistic Challenges

3.1. Origin and socialisation

Certain countries such as France seem to remain unconvinced of the usefulness of LAAP and LADO in the asylum procedure. Although LADO was introduced as a means to improve informal language analysis in the asylum procedure, this technique has encountered

(socio)linguistic, narrative and institutional hurdles of its own. The first challenge in LAAP and LADO is metalinguistic: the description of the task given to interviewers, analysts and decision-makers of verifying or identifying origin. This description proposes a link between what a person sounds like and what language(s) they speak, and in turn, where they come from. Eades indeed mentions that the incontestable and unavoidable, yet simple, assumption that language analysis bases its practice on is that speech and language can, at times and to a certain extent, hint at an applicant’s country of origin (Eades 2009, 31).

Nonetheless, as Eades herself and many others have stressed, a person’s country of origin is not always their country of socialisation, i.e. the country they grew up in (Eades 2010, 414; Cambier-Langeveld 2010a, 29-30; Channon et al. 2018, 155). Most linguists agree that there is usually no inherent link between language and nationality (McNamara 2005, 363; Eades 2009, 32; McNamara et al. 2014, 263; McNamara et al. 2018, 159). This is especially true where deterroritorialised (im)migrants are concerned (Fraser 2009, 128; McNamara et al. 2010, 67), as they tend to originate from countries with ever-shifting national borders drawn by colonials and not by themselves. Migrants’ linguistic repertoires

(21)

are affected by a number of factors, including language attrition and constant contact with other linguistic varieties and cultures.

Language is then rather related to speech communities which, though at times rooted in certain regions or countries, can transcend national boundaries. In effect, to determine the country of origin of a deterritorialised or reteretorrialised person, language analysis would have to be “retrospective” (McNamara et al. 2010, 67), examining an applicant’s former speech patterns and habits. Yet asking an applicant to emulate speech that is no longer theirs can be awkward and unnatural, and even impossible in cases where the interviewee left their country at a young age.

Despite the aforementioned complications, Eades notices, for example, that Australian language analysis reports from the early to mid 2000s made no efforts to assert the lack of an intrinsic link between language and nationality (Eades 2009, 31).

3.2. (Socio)linguistic COI assumptions

Another obstacle in language analysis is lack of knowledge concerning the applicant’s claimed country of origin (COI), which can lead to a folklorisation of foreign culture, traditions, and socio-political and economic life. This is especially important in asylum procedures where COI knowledge takes precedence over (informal) language analysis, such as in the French asylum procedure (see OFPRA questionnaire response B8, Appendix 6).

Lack of COI knowledge can negatively impact the development of the interview, as applicants might feel confused or undermined, or the limited time allocated to each interview might be wasted on what Jacquemet calls “cross-talk” (Jacquemet 2011, 488-89); words with different ethnologic meaning(s) that can be interpreted differently, which in turn can lead to misunderstandings and talking ‘across’ each other instead of to each other. For example,

(22)

that all people, even deterritorialised asylum seekers, follow the Gregorian calendar (Jacquemet 2011, 488-89). Such an assumption could lead to inadvertent mistakes in regards to translating or reporting dates of important events, or misreporting a journey’s chronology.

Such an assumption can double in impact if these mistakes create inconsistencies between an initial interview and a subsequent appeal interview. In a Belgian case analysed by Maryns, a Cameroonian applicant named Koulagna is rejected and undergoes a second interview as part of the appeal procedure where the interviewer is an official who is familiar with at least some basic COI knowledge of Cameroon. Some of the applicant’s answers differ from those in his first interview, which is enough for his application to be rejected (Blommaert 2000, 2).

COI knowledge that is lacking can also intensify feelings of disbelief felt by officials towards applicants, as certain narratives might seem too incredible for a Western framework. Blommaert describes a case in which one Belgian judge displayed overtly discourteous incredulity at the suggestion of a place of origin in sub-Saharan Africa that had a hospital and a school, but no functioning phone lines (Blommaert 2001, 188). Whilst it might seem inconceivable in Belgium, this disbelief ignores the plausibility of such a set up in a non-Western environment, and is further aggravated by a non-Western understanding of terms such as ‘hospital’ and ‘school’ which can refer to establishments that differ entirely from the ones imagined by the judge.

As a response, language analysis units have developed COI documents to have at their disposal, which are used to train and inform staff members; however, Cambier-Langeveld, head of the IND’s language analysis unit OCILA in the Netherlands, or Bureau Land en Taal (BLT), stresses as recently as last year that COI is often lacking (Cambier-Langeveld 2018, 10). Extensive and up-to-date COI knowledge, documents and continuous training are thus essential to both LAAP and LADO, not only because this avoids gross folklorisation of the

(23)

complex features of a country, but because folklorisation of these features in a standard interview, language analysis interview, or report can (un)intentionally portray the applicant as insincere and incoherent, often by no fault of their own.

Conversely, making use of COI knowledge that is hyper-current can also be a disservice to the applicant, who might have left their country before certain political or social changes took place, or who could have been held in captivity without access to external information; in such cases, testing claimants on a national anthem, flag, or political party that has recently changed its name or colours would not be helpful. COI should therefore be used with caution, and interviewers and analysts would benefit from weighing standardised COI against the interindividuality of asylum narratives.

3.3. Perceptual dialectology and monolingual ideology

The sometimes poor quality of COI documents, or the absence of COI documents and knowledge in some asylum procedures altogether, could be due to the unfounded belief that all native speakers are inherently able determine if a person shares their country or region of origin or not. Perceptual dialectology is a subcategory of folk linguistics that is concerned with how non-linguists perceive dialects, and analyses (lay)people’s capacity to distinguish linguistic variations from one another, as well as distinguishing what their conclusions are based on (Preston 1999, Fraser 2009, 122; Maryns 2004, 256). Many academics have criticised certain aspects of perceptual dialectology; the belief that untrained native speakers can accurately distinguish closely related dialects from one another (Rose 2002, 63; 3_{Wilson 2009, 28); the}

dangerously weighted confidence with which untrained native speakers proclaim their results (Fraser 2009, 124; McNamara et al. 2018, 158) with no inherent link between confidence and accuracy (Foulkes et al. 2011); and the gross simplification of the complex and often

(24)

misleading relationship between nationality, country and language (Labov 1994, 186; Eades 2010, 413; Patrick 2010, 77-78; Bourassa 2013, 10; Patrick 2016, 238; McNamara et al. 2018, 158). Jacquemet has shown that during the Kosovo War in 1998 and 1999, untrained native speaker refugees who were in fact used as interpreters, and arguably implicit decision-makers, relied mostly on their own biases in terms of assessing an interview: “Instead of relying on hard evidence, in most cases interpreters’ first impressions were heavily influenced by such communicative factors as accent, looks or politeness.” (Jacquemet 2010, 140). It is worth noting that perceptual dialectology is not limited to untrained native speakers, but can extend itself to any person who holds biased or outdated beliefs about language and culture.

3.4. Multilingualism

A dominant ideology of monolingualism also fails to recognise the obvious and increasing multilingualism (Muysken 2010; Wilson 2014, 222; Patrick 2016, 238) of a globalised, and unfortunately colonised world in which even trained (non-)natives might assume a country is monolingual and that inhabitants perfectly master the standard national variant (Gardner-Chloros et al 2016, 96-97), failing to account for any multilingual communities. Nigeria, for example, is hosts the highest number of languages in Africa, and the third highest worldwide (Muysken 2010, 91). Furthermore, languages are mixed together frequently and at times indistinguishable from one another, yet a mixed linguistic repertoire in an asylum interview can still be viewed as 'inconsistent', and might cost the applicant their right to protection. For example, Nigeria is considered a ‘safe country’ by many Member States, and applicants displaying linguistic features typical of certain Nigerian regions (only) could be denied protection and sent (back); however, genuine refugees who have been socialised in Nigeria, or who have spent many years there, are likely to adopt certain of the country’s (socio)linguistic features or multilingual speech patterns, and are thus subject to the same consequences. It is

(25)

also worth noting that expecting multilinguals to be familiar with all of their region of origin’s linguistic variants, to the same degree and without subconsciously shifting to foreign linguistic features, would be unreasonable.

Many governments also fail to officially state their own multilingualism (Muysken 2010, 91), such as Turkey, which only recently declared having more than one official language, despite being host to 35 linguistic variants (Muysken 2019, 122). Basing COI knowledge solely on hearsay ‘official’ information acquired from the State instead of up-to-date empirical research can then lead to incorrect assumptions about a country’s languages. Rejecting an asylum application based on an applicant’s accent, multilingualism, proficiency level or vocabulary, which might seem, through an uninformed lens, ill-fitting with their claimed country of origin, could operate erroneously on the assumption of an ideology of monolingual dominance.

Monolingual ideology also underestimates the potency of loanwords (McNamara et al. 2014, 270), which are now quickly adopted and integrated into a linguistic variety, and relies heavily on homogenisation of language (Blommaert et al. 1998; Eades et al. 2004, 180; Maryns 2006, 322). This in turn misunderstands the often involuntary and subconscious power of variationist sociolinguists; i.e. how language necessarily evolves not only in itself but from region to region, from person to person – that is “interindividually” (Maryns 2004, 257; McNamara et al 2010, 67; McNamara et al. 2014, 272; McNamara et al 2018, 164-65, McNamara et al. 2019, 263-64) – and from context to context.

Sociolinguistic Challenges

3.5. Accommodation Theory

(26)

British-when speakers (sub)consciously adjust any aspect of their usual language features to those used by their conversational partner(s). Further linguistic shifts are likely to occur with each additional person present during an interview, as all participants individually present a new layer of sociolinguistic complexity. Speakers can also accommodate to non-participants who are simply witnesses to the conversation.

Communication Accommodation Theory, or CAT, which Giles and Ogay describe as a “barometer of the level of social distance” between two people (Giles et al. 2007, 294) can be classified into four principles: communication as influenced by the immediate context as well and by the socio-historical context which necessarily regulates latter context; communication as referential but also adaptable where participants adjust to each other by accommodating to their addressee; communication as governed by expectations often based on stereotypes about where the middle ground between sociolinguistic adjustment and maintenance lies where participants must calculate the amount of non-, under-, or overaccommodation that determines whether a conversation continues or ends; and communication as using several strategies of accommodation such as convergence and divergence in order to display a participant’s (un)willingness to communicate (Giles et al. 2007, 294-95).

The challenges that accommodation and sociolinguistic variation can cause in an interview have been extensively analysed by academics and linguists (Coupland 1995, 21-22; Maryns 2004, 257; Maryns 2006, 319; Giles et al. 2007; Wilson 2009, 222; Patrick 2010, 79; Detailler et al. 2011, 7; Patrick 2016, 238; McNamara et al. 2018, 157) and especially within the framework of the asylum procedure, where linguistic variation can have a negative impact on a language analysis report and on the decision to grant international protection.

In questionnaire response A10 (Appendix 4), OCILA notes that sometimes applicants are unwilling to communicate in their L1, so an L2 or L3 must be used instead, and further

(27)

clarifies “[i]f there is no interpreter for the L1 the interview might have to be translated with an interpreter in L2 or L3 but we would still ask the applicant to provide a speech sample in the claimed L1.”. and LINGUA offers a similar response (A11, Appendix 5); however, elicitation of an L1 speech sample is complicated if the context presents the applicant with an interpreter for a different variant of their language, or an interpreter communicating in their L2 or L3. It is likely that the applicant will accommodate to their interlocutors, compromising the L1 speech sample. Interestingly enough, none of the questionnaire respondents mention accommodation as a sociolinguistic challenge when discussing language analysis.

3.6. Convergence

Two important types of accommodation which can impact an applicant’s chances of being granted protection, especially when the interviewer, interpreter or analyst holds an ideology of monolingualism to be true, or when they are ill-acquainted with variationist sociolinguistics, are convergence and divergence. Convergence can occur in an interview when the interviewee, or interviewer, adjusts their speech pattern to fit their interlocutor. An important reason for deciding to converge to a speaker’s dialect, accent, style, speech pattern, word choice, intonation, or even pitch, is the desire to gain approval (Coupland 1995, 22; Giles et al. 2007, 297); a desire most often expressed by the interviewee. It is then much more likely that the asylum seeker will accommodate and converge to the interviewer’s speech rather than the opposite, as the latter is in a position of power and impacts, at the very least on a surface level, the decision to grant asylum.

A basic understanding of convergence is that it can be upward or downward. Though less likely to occur, downward convergence could take place when an interviewer lowers their linguistic register to match that of their interviewee’s, for example in interviews with children,

(28)

an upper class interviewer (Giles et al. 2007, 295), a (sub)conscious strategy that is expected of subordinates (Giles et al. 2007, 297), and typical of asylum interviewees (Muysken 2010, 94; Patrick 2010, 79; Muysken 2019, 129); it is widely accepted amongst academics and some LADO professionals that the institutional and personal pressure of asylum interviews can have a considerable impact on an applicant’s speech (Patrick 2010, 79; McNamara et al. 2014, 269; Cambier-Langeveld 2016, 31; Patrick 2016, 238; McNamara et al. 2018, 157). Furthermore, the power structure affecting language is not just institutional or professional in nature, as it can extend itself to socio-political factors, gender, and age (Blommaert 2000, 29; Rose 2002, 63; Corcoran 2004, 212; Giles et al 2007, 301-03; Patrick 2010, 79; McNamara et al. 2014, 271; Channon et al. 2018, 156). The socio-historical context of interviews will often mean subconsciously conforming to a stereotyped standard of power relations that will expect not only lower classes to conform to higher classes, but women to men, and younger generations to older ones.

Applicants who interpret the asylum interview as a social gatekeeping interview that is concerned with who might be ‘fit’ to enter the country can even engage in ‘hypercorrecting’ in an attempt to appear educated or cultured. Such seems to be the case in Corcoran’s analysis of an asylum interview carried out by OCILA. The applicant pronounces the word “town” as a prosodic diaresis – two separate vowels instead of a diphthong where the vowel sounds glide together. He also categorises his native language, Krio, as “bad English” which points towards his interpretation of the interview as a gatekeeping interrogation (Corcoran 2004, 211). As the decision to not grant asylum can in effect be life-threatening, “hyperaccommodating” in order to appear as likeable as possible would not be an improbable scenario, and total convergence is often viewed as negative (Giles et al. 2007, 297) and insincere.

Convergence is also related to loss of (socio)linguistic identity (Giles et al. 2007, 296-97), whether this loss occurs during the interview, or beforehand by way of prolonged exposure

(29)

to different speech communities (Labov 1994, 5; Coupland 1995, 22). Therefore, although convergence might be difficult to consciously control in certain social contexts, or can be used (sub)consciously as a tactic to gain approval or encourage communication and entente, an asylum applicant risks detaching themselves from their usual speech community which goes against the purpose of a language analysis interview that seeks to elicit the applicant’s native language(s). The applicant is then in danger once again of seeming insincere.

3.7. Divergence and speech maintenance

At times, a person may also engage (sub)consciously in divergence, which occurs when a person (in)voluntarily reinforces their own linguistic features rather than adapting to that of their conversational partner’s. Speech maintenance, a closely related phenomenon, is described by Coupland as “the absence of detectable speech modification” (Coupland 1995, 22). Though both divergence and speech maintenance might be more helpful than convergence in terms of eliciting a speech sample of a native (regional) variety, these strategies run the risk of appearing uncooperative or hostile, as accommodation and especially upward convergence are seen as “pro-social interpersonal interaction[s]” (22). Furthermore, divergence need not be an immediate reaction to other conversational participants in an interview, but can be a result of sociohistorical detachment. Thus, a person might diverge from their original speech community, whether this is intentional or not, because they or others, view their speech community negatively (22). This is the case in the aforementioned asylum application analysed by Corcoran where the applicant equates his native Krio to “bad English.”

3.8. Linguistic shifting

(30)

only depending on other participants in the conversation, but depending on topos and locus. Discursive, or topical, shifts (Coupland 1995, 24; Maryns 2001, 78; Patrick 2010, 79; Patrick 2016, 238) can provoke the use of a standard linguistic variant or lingua franca where the topic of conversation is related to the host country, and the use of a local or regional variant where the topic of conversation is related to the home country. Such linguistic shifts may also be dependent upon the speaker’s geographical situation; the impossibility of eliciting an asylum applicant’s home variety abroad and outside of the context of their home country or country of socialisation is well-documented by several academics, some of whom have been directly involved in LAAP and LADO (Coupland 1995, 22; Maryns 2001, 77-78; Maryns 2005, 309-10; Muysken 2010, 97; McNamara et al. 2014, 268-69; McNamara et al. 2018, 163; Muysken 2019, 129).

An asylum seeker may do their best to converge culturally to encourage communication, and maintain or diverge linguistically to facilitate the elicitation of a speech sample, but might be unable to do so consistently depending on the sociolinguistic context of the interview. This could result in shifting between all three (sub)conscious strategies, and could be further exacerbated if there is more than one interlocutor present. Determining origin is difficult if the applicant adopts some of the interviewer’s standard vocabulary, involuntarily mimics the interpreter’s intonation, or in a diglossic context, shifts between varieties – all of which, especially within an ideology of monolingual dominance, might be interpreted as language ‘inconsistencies’ in an interview report.

If an interview report or a language analysis is based largely on folk linguistics, all the influencing aspects of communication accommodation, convergence, divergence and speech maintenance might not be recognised or reported. In fact, Detailleur and Spotti say of the IND’s language analysis unit: “in its eleven years of existence, all public documents brought out by BLT – and the IND for that matter – that clarify the procedure and use of the LADO-tests do

(31)

not mention the danger of language accommodation and/or precautionary measures taken to avoid its negative impact.” (Detailleur et al. 2011, 8). In such a scenario, all forms of linguistic variation in an asylum interview could be seen as illogical and unwarrantable sociolinguistic inconsistencies, at the expense of the applicant’s credibility.

3.9. (Socio)linguistic self-awareness

Finally, language in the asylum procedure is not only affected externally by cultural and linguistic assumptions and by uncontrollable linguistic shifts, but internally by lack of (socio)linguistic self-awareness. How aware is a person of their own proficiency in a language or dialect, and is this awareness less accurate in multilingual deterritorialised peoples who lose touch with their ‘L1’?

Language attrition refers to the process of losing a first or native language (L1); although it can happen for a number of reasons, it is usually due to being isolated from this language, whether it be from its speakers or from the regions where it is spoken, and aggravated by the acquisition of a second language (L2). As refugees are both a deterritorialised and reterritorialised group, often isolated from their native language and country, and frequently travelling with other migrants from various countries, one could assume that they would be particularly prone to language attrition. The process can be a socially disconnecting and perhaps a (sub)consciously painful one to admit to having undergone for applicants who are particularly proud of their home country and native or first language, and who are confronted with a fluent native speaker interpreter or analyst. On the other hand, applicants who view their language(s) and home setting negatively might welcome language attrition entirely and proclaim that they have no knowledge of their former language abilities, despite traces of an L1 usually being retrievable (Muysken 2010, 97).

(32)

Another concern is that proficiency or ability to communicate adequately in a language may well be highly indexical and culturally dependent, meaning that the term ‘proficiency’ not only hinges on the context in which it is used, but on the different meanings it can hold to someone who has been socialised outside of a Western framework. As previously mentioned, governments do not always recognise their country as being multilingual, and allowing for a the existence of legally ‘unofficial’ linguistic variants. If the government fails to recognise certain languages or dialects, it is a real possibility that the applicant will not recognise them either. Seeing as the very nature of the asylum interview is in fact official, it might make perfect sense for applicants to declare only the languages they are proficient in if they are recognised legally, administratively, or politically. Furthermore, some applicants risk accidentally occulting some of their languages as they may only count those in which they can speak and write (Muysken 2019, 124), thus involuntarily underestimating their own linguistic abilities.

Muysken also notes that “[s]peakers may overrate their knowledge of more prestigious languages, and underrate their knowledge of less prestigious languages” (97), and given that the asylum seeker is not only likely to want to make a good impression, but could view the interview as an act of social gatekeeping, boasting a knowledge of a prestigious lingua franca and underplaying any knowledge of a regional dialect might seem strategically prudent.

Overestimating proficiency could also result in interviewing without the presence of an interpreter, for example in cases where the applicant is very competent in the host language, but not fully acquainted with semantic and pragmatic nuances of colloquial speech (Blommaert 2000, 29), when such an intermediary might be necessary and can lead to a more detailed and accurate report (Maryns 2005, 311-12). OFPRA states in questionnaire responses A2 and C10 (see Chapter 6.4 and Appendix 6) that French-speaking asylum seekers will be heard in French if this is their wish, and that there is no interpreter present in interviews conducted in French, even if French is not the applicant’s L1. In fact, OFPRA further clarifies that interpreters were

(33)

not required to be legally present before 2015 (D16, Appendix 6), and that current profiles of interpreters used at OFPRA “include native speakers”, meaning that not all interpreters working at OFPRA are natives. This allows for interpreters and asylum seekers alike to overestimate their proficiency in an L2 or L3.

(34)

4. NARRATIVE CHALLENGES IN AN INSTITUTIONAL

CONTEXT

Narrative Challenges

4.1. Linguistic and narrative incoherence

As mentioned above, the way in which questions are phrased can affect the response patterns of an interview, and thus its positive or negative development. The first challenge is understanding the nature and purpose of the conversation. Misunderstanding the general interview or language analysis specific interview as a social gatekeeping interview is not uncommon, and can lead to the production of an incoherent narrative, such as in the aforementioned case investigated by Corcoran (2004):

Mr B is told, ‘this language analyst will listen to the recording of your language and they will come to a decision about you’ (lines 28–9). Like other gatekeeping interviews such as job interviews, academic admissions interviews and oral examinations, Mr B hypothesizes that he is meant to demonstrate his verbal and social fitness for the task at hand – in this case successful integration into the Dutch socioeconomic system. (206).

This misunderstanding is not only based on Mr B’s own hypothesis, but also on the very brief explanation of the interview that the IND provides, which by no means clarifies that the goal is to pit a linguistic repertoire against a country of origin and country of socialisation profile, rather than to demonstrate a proficiency in a prestigious standard language, or in a lingua franca. Furthermore, questions on immigration or the desire to one day seek work in the host country (Jacquemet 2009, 533-34) contribute to the image of the interview as a gatekeeping interrogation as the interview’s purpose is to verify the claim to asylum, but questions are related to socioeconomic intentions. Deciding on the right answer can provoke confusion and discomfort, as a willingness to seek work might make a genuine refugee seem like an economic

(35)

migrant, and an unwillingness to participate in the socioeconomic life of the host country might be synonymous with an unwillingness to integrate without being a financial burden. Such questions appear in the following case in an audio-recording analysed by Gómez-Díez (2013):

1 O4_: _{est-ce que vous avez l’intention d’émigrer=}

do you intend to emigrate =

2 = dans un autre pays par la suite?

=to another country afterwards?

3 A: pardon?

sorry?

4 O: est-ce que vous avez l’intention d’émigrer=

do you intend to emigrate=

5 =dans un autre pays par la suite ?

to another country afterwards?

6 A: non

no

7 O: si c’est légal eh ? si c’est possible/=

in the event it is legal eh? in the event it is possible /=

8 =vous pouvez dire oui/hein↑

=you can say yes

9 A: (1’’) pour l’instant je ne pense pas

(1’’) for the moment I am not thinking about it. (31)

Asking questions related to immigration and employment might contribute to an incoherent narrative as the interviewee’s interpretation of the interview risks differing considerably from that of their interviewer. Setting up an environment that leads to a report of a coherent narrative between interviewer and interviewee then necessarily depends on having understood the purpose of interview, as well as having understood each other.

A number of language-related challenges that can affect the construction of a coherent narrative are also picked up on through the use of Conversation Analysis by Channon, Foulkes and Walker (2018), such as not allowing the applicant enough time to answer, asking the applicant to speak in their native language without providing them with a topic of conversation, or using echo questions as a way of both confirming what the applicant has just said – as in, “IE5_{: I was two years, IR: two [?], IE: yes” – and at other times, to highlight that the answer}

(36)

might be problematic – as in, “IR: can you try to say the same thing in Yoruba now, IE: Yoruba no, IR: no [?], IE: Yoruba is just small just small I speak” – without specifying when which meaning is being used (Channon et al. 2018, 161-63). All of these factors can lead to significant misunderstandings and surface level incoherence, but most impactful on the direction of the conversation is perhaps the fraught relationship between repair initiators and hostile reformulations. Though the interviewer needs to remain in a position of authority, which can be difficult during a highly emotive narration, non-institutional conversations tend to prefer repair initiators that see problems of incoherence as accidental miscommunications (Pomerantz, 1984b; Svennevig, 2008, Channon et al. 2018).

Hostile reformulations and interpretations of the interviewee’s answers include the use of the turn-initial ‘but’ – as in, “IR: so you don’t speak French, IE: no I don’t [speak French], IR: [but you said in] school they were teaching you in English and French, IE: no you have to choose, IR: ah ok” (Channon et al. 2018, 166). Here, the interviewer chooses to question the statement previously made by the applicant using the arguably confrontational and disbelieving turn-initial, instead of reformulating the question and prefacing it with “I think I have misunderstood, let me rephrase my question.”

Strategies such as these, though supporting an image of authority, might create a fissure in the conversation which can be both unpleasant and unproductive. Failing to explain the purpose of the interview, cutting off the interviewee, lacking in linguistic clarity, and relying on hostile reformulations such as “but you said” and “I don’t think you understand” (Channon et al. 2018, 166-69) that hold the interviewee accountable for any miscommunications, can all lead to further splintering of the asylum interview(s).

That being said, it is worth repeating that conscious manipulation of language is very difficult, especially in emotive settings, and the difficult task the interviewer has of covering certain essential questions all whilst minding their language, at times eliciting a speech sample

(37)

for language analysis, and attempting to be authoritative, respectful, inquisitive and trusting in equal measure, should not be underestimated.

4.2. Ethnocoherence

Coherence is not just contingent on structural and linguistic misunderstandings that occur within the interview, and metalinguistic misunderstandings in regards to the description of the interview itself, but are culturally-dependent. Blommaert defines ethnocoherence as the “codes and norms of coherence that are anchored in different narrative traditions.” (Blommaert 1994; Blommaert 2000, 5). Narrating in a second or third language is difficult as it is, and the difficulty is bifold if the interviewer, interpreter, analyst, or courtroom participants expect the asylum seeker to naturally comply with Western narrative traditions. That said, limited linguistic resources in a foreign language do not necessarily render a narrative simple, unstructured, or incoherent. Blommaert (2000) illustrates this point with the example of Habiba, a Somali asylum seeker speaking in a “broken” and “simple” English who has been accused by the Belgian government of displaying narrative incoherence. Breaking down the narrative, Blommaert shows that Habiba introduces new information in every single one of her statements, highlighting the salient signs of a structured narrative:

1. So Somalia is starting war *nineteen ninety one 2. So until ninety one to ninety five I was in Somalia

3. And [baby starts crying] wa [laughs] and I have *four children at that time 4. And euhm… My husband comes from euh *north Somalia

5. And I *south Somalia is fighting north at=at south is fighting\ 6. So my=my husband and my children have no. *safety for their lives

The main statements are marked by ‘so’; the subordinate ones by ‘and’. The structure is clear, and Habiba accomplishes it by means of only two cohesive markers: ‘and’ and ‘so’. In conclusion: stories told in ‘simple’ language are not necessarily ‘simple’ stories. (12-13).

(38)

are to tell their “home narratives” in considerable detail (Blommaert 2001, 4; Maryns et al. 2001, 63), is that coherence is indexical of truth-telling, at least from a Western perspective. An applicant that is partially or fully incoherent is thus associated with an unwillingness or inability to tell the truth; this phenomenon could be termed truth expression. Truth expression might refer purely to the use of culturally-dependent narrative patterns, which, in a Western framework, might be indexical of and even synonymous with ‘truth-telling,’ though not inherently related to truth itself: “The point, however, is that the criteria used in such acts of evaluation are derived from a strongly literacy-based European understanding of textuality and coherence […], and that these criteria may clash with the criteria of – text-based – ‘truthfulness’ and ‘reliability’ used in societies different from ours.” (Blommaert 2000, 5-6).

Truth expression is then, in a poststructuralist sense, open to interpretation and culturally-dependent; for example, Westerners might interpret silence as confrontational, or in the face of an accusation, as a confirmation of the allegation – i.e. as gratuitous concurrence. Nevertheless, it has been show, for example, that speakers of Aboriginal English and traditional Aboriginal languages frequently perceive silence to be positive and helpful in encouraging productive communication (Eades et al. 2003, 125), and thus even the meaning of the absence of narrative is rooted in ethnic traditions.

OCILA, LINGUA, and OFPRA do not mention the possibility of gratuitous concurrence in their questionnaire responses. Participants were asked how it is checked that applicants understand the purpose of the interview, to which OCILA replied “It is checked verbally by asking. The purpose is also announced on paper in advance.” (A16, Appendix 4), LINGUA ticked “Asking the applicant in their native language whether they’ve understood”, and further clarified that they avoid administrative jargon (A16, Appendix 5). OFPRA stated that this was tested with a “formal question” and must be

(39)

“kept in mind by the interviewing officer for the duration of the assessment” (A1, Appendix 6). How to be sure, then, that saying yes to “having understood” is not purely an instance of gratuitous concurrence?

4.3. Entextualisation

The credibility of a narrative is also affected by entextualisation, a process that combines both decontextualization – removing a narrative from its original context – and recontextualisation – placing the narrative in a new context. Entextualisation can also simply refer to the rendering of speech into text (Jacquemet 2011, 481), which necessarily demands some recontextualising changes, or to a written public record that symbolises the reproduction of institutional authority (Park 2009, 487). This happens frequently in the asylum procedure, as narratives told in the context of an interview often do not follow the required format for an interview summary or subsequent report (Maryns 2004, 241).

Entextualised interview reports not only summarise the asylum seeker’s narrative however the author pleases, but necessarily weigh its credibility by way of expressions of belief or doubt. This will mean that the decision-makers, whether they remain the relevant (im)migration authorities, or whether they are judges and legal experts in a court, often base their verdict on the report alone and do not refer back to the original narrative, with its original wording and context. Blommaert (2001) shows in italics how expressions of doubt can undermine the credibility of a supplanted narrative, as if a judgement has already been pronounced:

It has to be noted that the concerned remains very vague at certain points. Thus he is unable to provide details about the precise content of his job as political informant. Furthermore the account of his escape lacks credibility. Thus it is

unlikely that the concerned could steal military clothes and weapons without

being noticed and that he could consequently climb over the prison wall. It is

(Socio)linguistic and Narrative Challenges in the Franco-Dutch Asylum Procedure - Implementing Language Analysis in France