The design and implementation of credibility assessments in Australian asylum decision-making frequently draw on problematic assumptions about language and can therefore result in inconsistent and unfair outcomes. This article builds on existing critical examinations by presenting a case study of a successful appeal in the Federal Court of Australia (FCA) which overturned a decision involving one such problematic credibility assessment. This case demonstrates the positive impact that sociolinguistic-aligned reasoning can have in such difficult appeals.

The article first provides some background on Australian asylum policy and procedures. It summarizes the findings of a critical discourse analysis of a set of credibility assessment guidelines and merits review decisions, where it was demonstrated that credibility assessments rely on flawed “language ideologies”, or understandings about how language and communication work. It then provides some history of the FCA decision, CRL18 v Minister for Immigration, Citizenship, Migrant Services and Multicultural Affairs [2020] FCA 917 (“CRL18”). It summarizes the judicial review process in Australia for migration decisions and contextualizes the CRL18 decision and earlier application process within this framework.

The article then examines how the judge in this case deals with three of the premises upon which the original negative decision was made, ultimately leading to him overturning the lower-level decision. For each premise, the analysis identifies relevant sociolinguistic scholarship which aligns with the judge’s reasoning. The article finishes with a concluding discussion, which includes an exploration of the structural limitations to appealing refugee visa applications that have been rejected on the basis of adverse credibility assessments. However, based on the preceding analysis, it concludes that when judges have an appreciation of sociolinguistic factors, they have the potential to denaturalize the problematic ideologies underlying credibility assessments. The examination of the CRL18 decision thus provides a powerful example of the potential benefits of communicating sociolinguistic research to law students, legal practitioners and decision-makers.

Credibility Assessments in Australian Refugee Visa Decision-Making

Asylum seekers who travel to Australia without a valid visa, for example by boat, face a number of barriers to gaining protection. Current migration law grants the government discretion to decide whether or not to allow these individuals to apply for a protection visa (the name given to refugee visas in Australia) and, even then, restricts them to only applying for a temporary visa, which requires periodic renewal through ongoing reassessment of the person’s asylum claims. Such individuals usually have some form of initial entry interview after being intercepted and detained. Later, if given the opportunity, they submit a written application and have an interview with an Immigration Department officer to have their refugee claims assessed.

If their claim is rejected, they face a very limited review process, conducted by the Immigration Assessment Authority (IAA), where they rarely have an opportunity to have another interview or present new information. Generally, the IAA decision-maker will make a decision “on the papers”, considering the application forms and initial decision, and reading an interview transcript and/or listening to an audio recording (or, sometimes, watching a video recording) of the Immigration Department interview and any other available documentation. On this basis, they will decide whether they agree with the Immigration Department officer’s decision to reject the application or whether they should overturn this decision and grant the visa. This “fast track” process has been widely criticized, with many voicing serious concerns that this type of review process in particular does not guarantee procedural fairness (Kirk, 2017; McDonald & O'Sullivan, 2018).

By contrast, people who arrive in Australia with a valid visa, for example as tourists or for study or business, have the right to apply for a permanent protection visa which, if granted, allows them an ongoing right to remain in Australia and eventually to apply for citizenship.Footnote 1 If their initial application to the Immigration Department is rejected, they have access to a more comprehensive review process, where they appear at a hearing and can present new evidence and have the merits of their claims reviewed by the Administrative Appeals Tribunal (AAT) (for a full explanation of the two processes see Kirk, 2017).

However, even with these better options, the design and implementation of the credibility assessments used in the AAT rely on problematic language ideologies. Introduced by linguistic anthropologists and now commonly explored in sociolinguistics, language ideologies are “taken-for-granted assumptions about how language works” (Eades, 2012, p. 474). Reliance on such ideologies can result in inconsistent and unfair outcomes, especially for those who do not have access to legal assistance. Critical discourse analysis of the AAT’s credibility assessment guidelines and a set of anonymized published decisions has found that credibility assessments rely on flawed language ideologies (Smith-Khan, 2017a, 2017b, 2018, 2019). In asylum decision-making, such beliefs about language are similar to and in some cases overlap with lay assumptions about psychology and human behaviour, which existing studies have identified as equally problematic for credibility assessments (see, e.g. Dowd et al., 2018; Herlihy et al., 2010).

There are a number of reasons why such assumptions can be problematic. First, they overemphasize the responsibility of the asylum seeker in producing the texts that come out of the refugee application and decision-making process, essentially presenting them as communicating in isolation (Smith-Khan, 2017b). A sociolinguistic lens tells us that this is problematic because in reality, these texts are created in interaction with and with contributions from a range of actors, including, of course, the decision-maker who asks questions and writes the final decisions and other participants such as legal advisors and interpreters (Eades, 2012; Jacobs & Maryns, 2021; Maryns, 2013a; Reynolds, 2020). Furthermore, the whole process is guided and constrained by legal and procedural requirements that undoubtedly influence what is said, when and how (Smith-Khan, 2018, 2019).

Second and relatedly, there is an overemphasis on narrative consistency. Asylum seekers may not be believed if they do not recount an event consistently over a number of different interviews or written texts. This draws on the belief that fragments of text remain stable and retain the same meaning when removed from their original context (Rock et al., 2013; Smith-Khan, 2021a) and may thus be examined in isolation to provide evidence of a lack of credibility (Eades, 2012). Assigning ownership or responsibility for these fragments to one actor — the asylum seeker — is closely linked with this approach, as both approaches ignore the importance of textual, interactional and broader contexts in the production and meaning-making of these fragments. Furthermore, as pointed out by a broad range of scholars, expecting an asylum seeker to remember and recount traumatizing or even banal events from their past consistently over long periods of time is simply not realistic (e.g. Cohen, 2001). Guidance, both in Australia and beyond, warns against expecting asylum seekers to be able to remember and recount perfectly in this way and also advises that different individuals may remember different details (e.g. in Australia, for reviews seen by the Administrative Appeals Tribunal, see Administrative Appeals Tribunal (2015), paragraphs 29–31. For international guidance, see, e.g. UNHCR, 2013). However, decisions still routinely point to issues of vagueness and inconsistencies in the level or types of details given across different interactions, and/or by different individuals, as reasons for negative credibility assessments (Dowd et al., 2018; Smith-Khan, 2017a; Vogl, 2013).

Third, an all-or-nothing approach is often used in conceptualizing bilingual communication: monolingual ideologies inform how decision-makers conceptualize bilingual competence. This may explain the lack of attention paid to the role of interpreting and interpreters: many decisions do not even indicate whether an interpreter was present for a hearing, let alone describing in any detail their role or conduct. Likewise, in guidance, like the AAT Credibility Guidelines, they are mentioned only very briefly (Smith-Khan, 2017b, 2019). If an asylum seeker has access to interpreting, this often appears to be taken as sufficient to avoid any communication issues that could reflect adversely on credibility, for example, apparent issues with consistency or level of detail or “vagueness”. This means that if such issues occur, this is considered to be the asylum seeker’s fault. This approach ignores the myriad linguistic choices involved in interpreting (for an extensive discussion of these, see van der Kleij, 2015). Similarly, if an asylum seeker uses a second language within the application or appeal process, nuances regarding varying proficiency levels in their different languages may be ignored. This was the case for an asylum seeker deemed untrustworthy partly based on an apparent inconsistency for describing an injury to his “arm”, when using English, and “shoulder”, appearing in a medical report in his first language, and despite his later attempts to explain the proficiency-related reasons for this variation (Smith-Khan, 2019) (see similar discussion of overexpections of an asylum seeker’s second language proficiency in Maryns, 2005).

These problems with credibility assessment are particularly concerning given that there are limited legal avenues for the IAA’s or AAT’s credibility-related decisions to be reviewed and overturned in court. And even if they make it to court, there is the potential that the same problematic language ideologies could undermine a successful appeal.

While there is plenty to be concerned about, this article explores a good news story: a recent case of successful judicial review involving an asylum seeker whose temporary visa application had been rejected by the IAA based on adverse credibility findings. The decision in CRL18 v Minister for Immigration, Citizenship, Migrant Services and Multicultural Affairs [2020] FCA 917 (“CRL18”) implicitly incorporated understandings about communication aligned with sociolinguistics, rather than adopting the types of problematic assumptions summarized above. This ultimately contributed to a positive court decision for the asylum seeker involved.

CRL18 and Judicial Review of Migration Decisions

The appeal involved an asylum seeker from Lebanon who, along with one of his brothers (referred to in the decision as “M”Footnote 2), had travelled by boat via Indonesia to seek asylum in Australia. They each had an initial interview and processing on Australia’s offshore island territory, Christmas Island, before later each being allowed to make an application for a temporary visa. “CRL18” first had his refugee claim rejected by the Immigration Department, and when it was referred for merits review, the IAA also rejected it. The reason given for the IAA rejection was that the decision-maker had concerns about the asylum seeker’s credibility. In particular, the IAA decision-maker pointed to a number of apparent inconsistencies in his claim, both in terms of how he described events at different points in the procedures and also in that his accounts of various events were inconsistent with his brother, M’s. Based on these inconsistences, the decision-maker rejected some of the key claims supporting CRL18’s application and thus concluded that he did not need protection and should not be granted a visa.

With the help of lawyers, the asylum seeker (referred to in the decision as “the appellant”, i.e. the party bringing the appeal) sought judicial review in the Federal Circuit Court (FCC) in 2019 (referred to in this article’s citations as the “FCC decision”). When this review was unsuccessful, they made a further appeal to the Federal Court of Australia (FCA), which sits above the FCC in the Australian federal court hierarchy (referred to in this article’s citations as the “FCA decision”). This appeal was successful and is discussed in more detail below.

In Australian migration law, successful court appeals are limited to situations where there is found to be “jurisdictional error” in the original decision (for a detailed explanation, see Crock & Berg, 2011, Chapter 19). In practice, this means that the courts reviewing the original decision do not have the power to make findings about the facts of a case, or — in the case of a protection visa application — to consider the merits of the applicant’s refugee claims. Rather, they are limited to considering whether or not the original decision-maker followed the proper process for making the type of evaluation that they were legally authorized to make. This can pose a problem for rejections that are based on negative credibility assessments, as these are closely related to deciding whether or not to accept the events or facts an applicant claims are true. Some migration lawyers even opine that decision-makers may prefer to include credibility-based reasons in their negative decisions to best guard against their decisions being overturned in court (Smith-Khan, 2020, p. 120). Successfully challenging in court an unsuccessful asylum claim is very rare. For example, in 2014–2015, out of all unsuccessful merits reviews, only 0.7% went on to be appealed in court and to have a successful outcome for the appellant (14% of all reviews appealed in court were successful. Migration Review Tribunal & Refugee Review Tribunal, 2015, p. 25).

The limitations on their review powers mean that courts are cautious about appeals that involve disputing credibility assessments, and a very high bar will need to be reached for an appeal to be successful in such a case. In CRL18, the lawyers’ approach was to argue that there was illogicality in how the decision-maker used various apparent inconsistencies to reach certain negative conclusions about the applicant’s claims. However, when looking in closer detail at the inconsistencies relied upon by the IAA decision-maker and how these were dismissed as illogical, the court’s examinations — and the appellant’s lawyers’ arguments — align with sociolinguistic perspectives, even though no sociolinguistic or related research is explicitly discussed or cited in the decision. The following section explores these arguments and examinations and explains how they repudiate problematic assumptions about language and more closely reflect sociolinguistic perspectives.

The Apparent Credibility Issues and the Court’s Perspective

The judgment cites four premises, all related to apparent inconsistencies on which the IAA had relied to conclude that the asylum seeker lacked credibility and on that basis to reject his application. They were as follows:

  1. (1)

    The “Timing Premise,” concerning an internal inconsistency in the appellant’s evidence as to when the attack on his shop occurred

  2. (2)

    The “Subsequent Attacks Premise,” concerning an inconsistency between the evidence of M that agents clashed with A (their other brother) a further five times after the shop attack, with the evidence of the appellant who when questioned “gave a relatively vague response [and] said he had experienced many problems”

  3. (3)

    The “Tripoli Family Premise,” that the appellant had said in the interview that he did not have any family around Tripoli

  4. (4)

    The “Targeting of the Father Premise,” that M’s claim that the attackers were “mainly trying to attack the sons because they are younger compared to his father” was “vague and unconvincing and inconsistent” with the claim that the father was assaulted during the attack on the shop (FCA decision, paragraph 12, citing parts of the IAA’s written decision).

The court considered each premise in detail, looking at the IAA decision-maker’s reasoning and considering arguments made by each of the parties to the appeal. The fourth premise was found to be irrationally made as the decision-maker did not explain the basis for finding the accounts inconsistent, and on their face, they did not appear to be so (FCA decision, paragraph 67). Below, the first three premises and their treatment in the judgment are explained in greater detail. While neither the appellant’s lawyers’ nor the judge explicitly refer to sociolinguistic research, the below analysis demonstrates how the arguments and reasoning they adopt align with existing scholarship and avoid or challenge some of the problematic language ideologies that often arise in refugee credibility assessments.

Timing premise

The IAA decision-maker found an inconsistency regarding the timing of events the appellant described, involving the point in time at which his family opened a shop and the later date when an attack occurred there. They identified two dates in his written application, about six months apart. They also noted that he reported that there were several months between opening the shop and the attack occurring. The decision-maker decided this reference to “several” months, and the specific written dates were inconsistent with the appellant’s response of “a few months” in a later interview, reproduced in the transcript excerpt below:

Off: How long had you been running the shop before this happened, had you had it a few months, or was it open a couple of years? How long had it been open?

[App]: Just a few months, like--

Off: Just a few months.

[App]: Yeah (FCA decision, paragraph 23).

At the first level of judicial review, in the FCC, Justice Humphreys has accepted this premise. In his judgment, he concluded: ‘The word used by the applicant was “a few”. To my mind, that is less, much less than six months’ (FCC decision, paragraph 42).

Importantly, in the FCA appeal, the appellant’s lawyers drew ‘attention to the context in which the answer was provided…The question to the appellant posed two alternatives, “a few months” and “a couple of years”. It was not an open question…[his] answer was the most accurate of the two alternatives’ (FCA decision, paragraph 24).

Justice Stewart agreed with this argument and also noted that the officer’s interjection further limited the appellant’s scope to offer any other response. This approach emphasizes the importance of considering the immediate context of the interaction and how this contributed to the appellant’s response. This contrasts with the IAA’s and FCC judge’s focus on decontextualizing short phrases and examining them in insolation. Justice Stewart also acknowledged other elements of the interactional context, “the appellant was unrepresented and…required an interpreter” (paragraph 32), although he did not elaborate on the potential impact of these factors on the interview process or its transcription.

The judge further noted the broader context in which the communication took place, “many years after the events which are described” (paragraph 32), thus acknowledging another reason — aligned with the existing psychology scholarship on asylum claims and recall (eg Cohen, 2001) — that expecting this level of precision was inappropriate.

Finally, he once again reiterated the problems with focusing on isolated words or phrases, offering extra arguments for why this was especially inappropriate in this particular case.

It is not part of the visa application process to criticise an applicant’s choice of words with a lexicographer’s zeal for precision in the use of a phrase which even native English speakers struggle with, i.e. just how many months is “a few” months or “some” months? (paragraph 34).

In this way, Justice Stewart also emphasized the ambiguity of the particular terms themselves, rejecting the view that language and meaning are static and finite, and reinforced the inappropriateness of using this variation as a basis for a negative credibility finding, with implicit reference to the appellant’s (and potentially also the interpreter’s) non-native speaker status.

Subsequent attacks premise

The second premise also involved a number of points that the IAA decision-maker considered to be inconsistencies, but this time, between the appellant’s accounts and those of his brother, M. It contrasted M’s description of the shop being attacked “a further five times” and his descriptions of these incidents in his interview, with the appellant’s interview responses. The incident involved a third brother, “A”, who remained in Lebanon. The IAA decision-maker summarized the appellant’s responses:

However when the delegate asked the applicant if [A] had experienced any problems since the attack on their shop, he gave a relatively vague response and said he had experienced many problems and cannot go to Beirut and is very careful but is okay and did not refer to the specific incidents raised by [M] (cited in FCA decision, paragraph 39).

Once again, in addressing this reasoning, both the appellant’s lawyers and the FCA judge focused on the interactional context, looking at the specific questions that were asked of the appellant and of M (reproduced below in that order) and explaining how these questions influenced the different responses they offered.

Interview with the Appellant

Off[icer]: And has he had any problems since? Like, it’s been a few years now, is he having any problems now?

[Appellant]: [A]?

Off: Yeah.

[App]: He had – not all the time, but yes, he did.

Off: Okay, what problems? What’s – what problems is he facing?

[App]: Too many, like, in – he cannot go to Beirut.

Off: Okay.

[App]: Yeah, that’s the first thing. And he’s very careful at the place he’s living. Yeah, he’s very – going – very careful with everywhere, like, I mean, he’s living, and he’s okay in there, but he’s not really free, you know what I mean? Yeah (FCA decision, paragraph 40).

Interview with M

Off: … And when you say your brother [A] is still being harassed, what do you mean, what sort of harassment is he facing?

[M]: Yeah, they are still always trying to follow him up, but not too far, but just 21 kilometres from our area.

Off: Tell me everything you know about the problems your brother is having in Lebanon today.

Int[erpreter]: Can you say it again, please?

Off: Yeah, tell me everything you know about the problems your brother is having in Lebanon, continues to have in Lebanon.

[M]: You mean again to him, what happened again to him personally?

Off: Yes.

[M]: They start to clash with him for about five times. Once when he was shot in the shop, and the other time they shoot at him in the village, and also [inaudible] problems with him (FCA decision, paragraph 41).

Both the judge and lawyers for the appellant opined that there was no clear inconsistency or contradiction between these responses and that the appellant’s relative vagueness had more to do with the type of questions asked of him, pointing specifically to “the different levels of detail” involved in the specific questions.

Once again Justice Stewart went further in his examination and also pointed out that the questions asked of the appellant actually each involved “two different questions in quick succession” (paragraph 47). He noted that this resulted in the appellant answering only one part of each of these double-barrelled questions and also noted how these involved a mix of past and present tense which prompted particular time-specific responses. In particular, he contrasted the appellant’s last response in the present tense, which appeared to respond to the second part of the double-barrelled question (also in present tense), with M’s responses in past tense, dealing with past events. The judge therefore pointed to both “the different time periods” and “the different levels of detail” of the respective questions as possible reasons why the two brothers provided different answers.

Neither the lawyers nor the judge dealt explicitly here (at least from what is evident in the judgment) with the role that interpreting may have played in these questioning sequences. In fact, it is reasonably unusual to see the interpreter appear in a transcript of this kind, as they do in one line of M’s transcript. It reminds us that, in fact, the utterances assigned to the two brothers in these transcripts are actually the interpreters’ utterances, and, also, that more generally transcription is itself a process of entextualization, where choices are made about what and who to represent and how (Park & Bucholtz, 2009). It is possible, for example, that the interpreter in the appellant’s interview may have only interpreted one part of the double-barrelled questions. Here and elsewhere, choice of words and level of detail are at least partly attributable to the interpreter who produces them, rather than simply to the asylum seeker, as Justice Stewart briefly acknowledged in the Timing Premise, discussed above. However, the fact that the transcription, as is standard in this context, presents the appellant as the creator of these utterances reinforces the idea that he is responsible for them and backgrounds the contributions of the interpreter.

Nonetheless, the lawyers’ and judge’s focus on question detail and style echoes some of the concerns in sociolinguistic research on investigative interviewing. For example, in police interview research, scholars recommend using an open question that encourages expansion at an early point in the interview, often providing very similar wording to that of the question asked of M in the excerpt above, “tell me everything you know about …” (Heydon, 2012). They argue that this question type is best suited to gain a large amount of reliable detail and have recently extended this recommendation to Australian refugee review decision-makers, who have been found to be inconsistent in their questioning styles and strategies, as the examples in this case also demonstrate (Findling & Heydon, 2016; see also Luker, 2013). This expansive open question contrasts markedly with the more closed questions asked of the appellant in the form of “what” questions, which prompt a more specific and limited response, and especially the first double-barrelled question set, which appears to require a binary yes or no response. These latter types of questions may have a place in such interviews but are generally considered more appropriate for checking particular details, rather than eliciting large, expansive accounts (Findling & Heydon, 2016, pp. 27–28; see also Kjelsvik, 2014, for a detailed examination of interview structure, question types and agency).

Much like the scholarship on investigative interviewing, what the appellant’s lawyers and Justice Stewart concluded is that the type of questions asked will inevitably influence the form and content of the responses given. In this case, the judge agreed with the lawyers that difference in questioning style could account for the differences between the two brothers’ responses, and, therefore, that there was no clear contradiction between these two versions, meaning that this variation should not undermine the appellant’s credibility.

Tripoli family premise

The third premise upon which the IAA decision-maker found the appellant lacked credibility also relates to an apparent inconsistency between the two brothers’ accounts:

[M] claimed that, after the attack, his family moved to an area between Tripoli and Akkar to stay with family, however, the applicant claimed they did not have any family around Tripoli when asked by the delegate” (FCA decision, paragraph 52, citing paragraph 26 of IAA decision).

Again, the judgment sets out the respective passages from the two brothers’ interview transcripts and looks at the specific questions asked and specific answers given.

Interview with the Appellant

Off: Do you have family near Tripoli?

App: Do I have family in?

Off: Or around Tripoli?

App: In Tripoli? No, no. No, no (FCA decision, paragraph 53).

Interview with M

Off: Yeah. So, after this incident happened, what happened? Did your family leave the area? Did they go back to Akkar?

[M]: No, we left an area before Akkar to – with relatives, stay with relatives.

Off: Okay, in Beirut.

[M]: No, in close – close Tripoli.

Off: After Tripoli?

[M]: After Tripoli.

Off: Okay, all right, and so the Sunni area--

[M]: Before Akkar.

Off: Okay, okay. So you have family around Tripoli?

[M]: Yeah, yeah.

With the agreement of the parties, the judge introduced information about the geography of Lebanon, which explained the respective locations of Beirut, Tripoli and the Akkar region, Tripoli being located between the other two. He discussed the difficulties of determining how one would describe their relative distance from and relationship with each other.

Furthermore, once again, rather than reformulating and summarizing the appellant’s speech in the way the IAA decision did (see paragraph 26 cited above), Justice Stewart emphasized the actual content of these conversation sequences. The appellant’s reformulations demonstrate that he understood the question as being whether he had family in Tripoli and answered this definitively. The judge concluded that this response “having family somewhere between Tripoli and Akkar, as identified by M, is different to having family in Tripoli”, making it reasonable and consistent for the appellant to have answered no to the question he understood. This meant that once again, there was no contradiction on which to base a negative credibility finding here (paragraph 61).

Research across a range of bureaucratic and legal settings has demonstrated how police or immigration interviews, examination or cross-examination in court are strategically cited and summarized in ways that support a particular line of reasoning or argument (Eades, 2012; Jacquemet, 2009; Maryns, 2013b; Rock et al., 2013). The judge’s approach here resembles the emphasis that scholars place on uncovering the transformational work that occurs in legal processes, with an insistence on going back and looking at the original text, rather than relying on later summaries alone.

Combatting Problematic Language Ideologies in Court: Challenges and Opportunities

The decision in CRL18 demonstrates how focusing on the sociolinguistic realities of refugee status determination (RSD) procedures can result in fairer assessments of credibility. In this case, the first three levels of decision-making — the Immigration Department officer, the IAA decision-maker and the FCC judge — adopted an approach that effectively drew on or at least accepted the problematic language ideologies on which negative credibility assessments seem often to rely. The FCA appeal was only successful because Justice Stewart was open to scrutinizing the apparent inconsistencies from a perspective where the interactional nature of RSD processes is foregrounded and where the asylum seeker should not be held responsible for decontextualized fragments of transcript, nor for the reports or summaries of their speech that effectively transform its meaning. Prompted by the appellant’s lawyers’ equal emphasis on these issues, he focused on interactional context and questioning style in such a way that he was capable of finding the IAA decision-maker’s reasoning illogical. This resulted in meeting the very high standard necessary to overturn what may otherwise be considered questions of fact, which would be outside the purview of judicial review.

With sociolinguistic awareness, migration lawyers and appeal judges can make arguments and decisions that accommodate asylum seekers more fairly by becoming aware of and addressing “common sense”, but incorrect, assumptions about how language works. This demonstrates the potential benefits of researchers who work on communication in RSD and related administrative and legal decision-making reaching out to the legal profession and future law professionals, legislators and decision-makers through law schools (as argued in Grey & Smith-Khan, 2021a, b). Providing and/or increasing the availability of sociolinguistic training for first instance and initial review decision-makers is also essential. The potential for this has been demonstrated through the research-training activity piloted by Heydon and Findling with Australian refugee review decision-makers (see Findling & Heydon, 2016; Heydon, 2019, ch 7). Ideally, rather than leading decision-makers to believe that they have the expertise to assess communication-related issues, such training would help sensitize them about the potential scope and complexity of linguistic considerations and issues, encourage them to critically reflect on their assumptions about communication and to seek or take on board expert input where needed.

However, there are limits to what increasing awareness can achieve. For example, existing research has found that even where migration advisors demonstrate sociolinguistic awareness, they are constrained to act within existing procedural structures. Also, when advising their clients or preparing submissions on their behalf, they draw on their understandings and predictions about how decision-makers will assess credibility (Smith-Khan, 2020, 2021b). Sometimes, increased knowledge about communication can even be used against minority participants in legal interactions (see, e.g. in Eades, 2013, discussed in Grey & Smith-Khan, 2021b).

At the judicial review level, lawyers must also make sure their arguments align with the rules about which parts of the decision-making process the courts have the power to scrutinize. In deciding whether to pursue an appeal in court, lawyers are particularly constrained by strict rules requiring them to certify that they believe on reasonable grounds that the appeal has reasonable prospects of success, and if they are found to have encouraged their client to pursue litigation where there were no reasonable prospects of success, they face a range of adverse personal costs orders (see Migration Act 1958 (Cth), Sects. 486E and F). Ultimately, this means that lawyers may hesitate to pursue a line of reasoning that seems to require an expansion or new interpretation of legal principles or where the argument risks being seen as reopening scrutiny of the merits of the claim (which can often be the case in appeals related to credibility).

Furthermore, decision-makers work in institutional settings, and as such are influenced in how they do their work by the culture of such settings, their employment conditions, institutional guidelines and most probably also by broader political discourses. This means that the politics of the governing party, especially of those responsible for the Immigration Department itself, can result in what has been termed a culture of disbelief or suspicion (Bohmer & Shuman, 2018; Jubany, 2011; Piller et al., 2021) or one in which scepticism is viewed as impartiality (Johannesson, 2018). Such an environment is likely to affect civil servants working within this part of government, including Immigration Department and IAA decision-makers. Indeed, both major political parties in Australia have negatively politicized asylum seekers for at least the past two decades, and questioning refugees’ credibility is among the most popular themes in mainstream political discourse (Clyne, 2003; Macken-Horarik, 2003a, 2003b).

Moreover, despite the fact that the AAT is supposed to be an independent body, restructuring in 2015 led to a purge of existing decision-makers who had made politically unfavourable decisions. Investigative journalists also revealed that of those newly appointed, 64 had close political allegiances with the governing party, and some had no legal training, thus drawing into question their independence and competence (Hobbs & Williams, 2019; Landis-Hanley, 2019). Where there is room for discretion and variation in how credibility assessments are applied, these institutional cultural issues have the potential to undermine the incorporation of sociolinguistic understandings in this area of decision-making, where such an incorporation may be viewed as tilting the scales in favour of asylum seekers.

Issues at the merits review stage are particularly concerning, as the ability to successfully seek judicial review of IAA and AAT decisions depends heavily on having access to legal assistance. Unrepresented asylum seekers have faced significant challenges appealing in the FCC (Smith-Khan, 2015). Of course, such court challenges require not only identifying sociolinguistic issues where these are relevant but most importantly presenting sophisticated legal arguments. Furthermore, recent research has identified substantial variations in decision-making between different FCC judges, with lower rates of success for unrepresented appellants (Dorostkar, 2020). Successive decreases in government funding for legal assistance means a greater likelihood of asylum seekers being unrepresented at any level of the process but even more so for judicial review (Kenny et al., 2016; McDonald & O'Sullivan, 2018). Furthermore, given the limits on what courts can scrutinize and thus the low chances of success at the judicial review stage, lawyers must be selective when it comes to deciding who to assist. Finally, even for those who have a successful judicial review, the court does not have the power to grant them a visa. Generally, success at this level results in the court ordering lower-level review decision-makers in the AAT or IAA to remake the decision in the proper manner. This may ultimately lead to applications being rejected again but with revised written reasons. This is yet another reason why addressing issues at the initial application and administrative review levels is crucial.

Yet, there is a glimmer of hope: judicial review decisions have influential power. Nobody likes for their decisions to be negatively assessed and set aside in court. If lawyers and judges like those in the CRL18 decision adopt sociolinguistic perspectives to contest the application of problematic language ideologies in lower-level credibility assessments, this creates pressure on IAA and AAT decision-makers, and also on those responsible for creating policy instructions and guidelines for them, to rethink their reliance on these ideologies. While cultures of disbelief and lack of legal assistance still pose significant challenges, this shift would raise the bar in terms of how decision-makers must account for adverse credibility findings. Hopefully, this would mean that credibility assessments could play a less prominent role in RSD, and when used, asylum seekers would be more likely to be given the benefit of the doubt.