What happened
The appellant, a citizen of Afghanistan from the Jaghori district and of Hazara ethnicity and Shia Muslim faith, applied for a Protection (Class XA) visa. He advanced claims of a well-founded fear of persecution on four bases: his ethnicity, his religion, an imputed anti-Taliban political opinion, and his membership of a particular social group said to include failed asylum seekers or persons likely to be perceived as alcoholics. A delegate refused the application on 13 March 2013. An earlier Tribunal decision affirming that refusal was set aside by consent in the Federal Circuit Court. A differently constituted Tribunal conducted a hearing on 3 February 2014 at which the appellant gave evidence through a Hazaragi interpreter.
The critical exchange concerned health services in Jaghori. The Tribunal member observed that the Hazara community there enjoyed better educational and health services than in neighbouring provinces. The appellant responded in Hazaragi that for minor illnesses, pregnant women nearing delivery or in severe pain, residents had to travel to Ghazni or Kabul because, if there were a good or proper hospital in Jaghori capable of performing operations or providing treatment, there would be no need to travel. The interpreter rendered this in English with a crucial difference: the conditional “if” was placed so as to suggest the appellant was saying that people had to travel because there was in fact no hospital. When the member put to the appellant that he must know there was a very good hospital in Sangi Masha, the appellant acknowledged the Shuhada hospital’s attractive stone construction but added that people with major sicknesses still travelled to Kabul, Pakistan or India.
The Tribunal decision of 22 April 2014 treated the appellant’s evidence as a deliberate lie. At [54] the member stated that the appellant was not a witness of truth and was prepared to make false statements if he believed it would advantage him. The first example given was the Hospital Evidence: the appellant had said that even for minor matters such as delivering children residents had to go to Ghazni or Kabul, and that if there were a hospital these operations could be carried out locally; when confronted with the existence of a very good hospital he had backtracked. The same finding was repeated at [62] when the Tribunal assessed the chance of discrimination on return to Jaghori. Two further adverse credit findings followed: that the appellant had invented the beheading of six Hazara policemen in Qarabagh in 2001 (because no reports existed) and had invented the Alcohol Incident (drinking at Eid E Qurban at age 32 for no particular reason). These three findings underpinned the rejection of the appellant’s claims both under the Refugees Convention and under the complementary protection criterion.
The appellant sought judicial review in the Federal Circuit Court out of time. An extension under s 477 of the Migration Act 1958 (Cth) was granted. Ground 5 asserted that the Tribunal’s finding that the appellant had deliberately lied about the hospital was made in the absence of evidence or was otherwise affected by jurisdictional error. The primary judge rejected the ground after a contextual analysis, concluding that the appellant had impliedly asserted there was no hospital at all and had failed to correct that impression when given the opportunity. His Honour went on to find, in the alternative, that the Tribunal would have made the same decision in any event.
On appeal to the Federal Court, Lee J allowed the appeal. His Honour accepted that a contextual rather than purely textual analysis was required, yet concluded there had been a material mistranslation. The original Hazaragi conveyed the existence of a hospital that was not “good” or “proper”; the English version omitted that premise and created the impression of an outright denial. That impression led the Tribunal to form an emphatic view that the appellant had lied about an objectively verifiable fact. Because credibility assessments are impressionistic and non-linear, the error could have tainted the assessment of the other two incidents. The hearing therefore did not meet the statutory standard in s 425(1). Orders were made quashing the Tribunal decision, directing the Tribunal to hear the review application according to law, and awarding costs.
Why the court decided this way
Lee J began from the undisputed proposition that the interpreter’s English version differed from the appellant’s Hazaragi. The Minister contended the difference was immaterial once the entire exchange was read in context: the appellant had first claimed ignorance of health services, then said even women in labour had to travel, and when told of the good hospital had described its appearance but maintained that major cases went elsewhere. The primary judge had regarded this sequence as an implied representation that no hospital existed at all. Lee J accepted the need for contextual analysis but held that a statement that a hospital is not “good” or “proper” necessarily presupposes its existence. The Tribunal member was therefore labouring under a misapprehension that the appellant had made an unqualified denial and had then engaged in unmeritorious backtracking. That misapprehension produced the emphatic finding at [54] that the appellant “deliberately lied”.
The court’s reasoning then turned to the principles governing translation errors. Drawing on Kenny J in Perera, the function of an interpreter is to place the non-English speaker as nearly as possible in the position of an English speaker. Not every departure from perfect interpretation will vitiate a hearing; the error must relate to a matter of significance. The Full Court’s discussion in SZRMQ was treated as authoritative. Allsop CJ had emphasised that the question is evaluative: whether the hearing was fair having regard to the purpose of giving the applicant a real opportunity to place relevant information before the decision-maker. Where the decision rests in whole or in part on credit, it will often be impossible to demonstrate precisely how a mistranslation affected the result. The inquiry therefore focuses on process: whether the mistranslation had or could have had significance. Robertson J had added that the significance of the error is not assessed solely by reference to the reasoning actually used by the decision-maker, because that reasoning was formed without knowledge of the true evidence.
Applying those principles, Lee J held that the mistranslation was not peripheral. The Tribunal expressly identified three instances of deliberate invention, the first and most trenchantly worded being the Hospital Evidence. The judgment notes at [43]–[46] the artificiality of attempting to place the three matters in “hermetically sealed boxes”. Credibility assessment is holistic. As the Full Court had observed in VAAD, it is not necessarily linear. Kirby J’s remarks in NAFF were cited to the effect that disbelief on one point can carry over to others, while an corrected understanding on one point might cause a decision-maker to revisit the whole case. The Hospital Evidence concerned an objectively verifiable fact; the Tribunal could be confident it knew the truth. By contrast, the beheading and alcohol claims required more nuanced judgments about the absence of reports and ordinary human behaviour. An erroneous but emphatic finding on the hospital issue could therefore assume disproportionate significance.
A “reality check” was provided by the delegate’s earlier acceptance of the appellant’s account as generally credible on both the beheadings and the alcohol incident. While not determinative, this suggested the adverse findings were not inevitable once the mistranslation was removed from the picture. Accordingly, the mistranslation could have made a difference to the overall credibility assessment. The hearing was therefore not fair and did not satisfy s 425(1). The primary judge’s alternative finding that the Tribunal would have reached the same decision was characterised as a utility finding that could not cure the procedural unfairness once materiality was established. The appeal was allowed on both grounds.
Before and after state of the law
Before this judgment the law was already clear that inadequate interpretation could breach procedural fairness or the statutory analogue in s 425(1). Perera had established that the interpreter’s role is to equalise the position of non-English speakers. SZRMQ had synthesised the authorities, stressing that the ultimate question is whether the hearing was fair, that the inquiry is contextual, and that it is not always necessary to prove a precise causal link between mistranslation and outcome, especially where credit is involved. SZSEI had confirmed that SZRMQ’s observations applied to the statutory s 425 context. Aala and NAFF had underscored the impressionistic nature of credibility findings and the potential for error on one issue to affect the whole.
This judgment did not change the legal rules; it applied them to a particular factual scenario with emphasis. It clarified that a translation error need not be global or repeated; a single material mistranslation that founds an emphatic credit finding can suffice. It reinforced that courts should not compartmentalise discrete credit findings when the decision-maker has treated them as cumulative. The judgment also illustrated the limits of the “opportunity to correct” reasoning used by the primary judge: if the misapprehension is created by the interpreter, the applicant cannot be expected to correct what he does not know the decision-maker has misunderstood.
After the judgment, the law remains that adequacy of interpretation is assessed by reference to whether the applicant was given a real and fair opportunity to give evidence on matters of significance. The decision supplies a concrete example of how the SZRMQ evaluative exercise is performed when the mistranslation directly produces an adverse finding on an objectively checkable fact that becomes the centrepiece of a credit-based refusal.
Key passages with plain-English translation
Paragraph [31] contains the court’s central factual finding: “Having regard to all the circumstances, I do not believe that the interpreter expressed in English as accurately as the circumstances permitted, the idea or concept that had been expressed by the appellant in his native tongue (being the Hospital Evidence).” In plain English, the judge is saying the translator changed the meaning enough to matter. The original evidence assumed a hospital existed but said it was inadequate; the English version made it sound as though the appellant was denying any hospital existed at all.
At [40] Lee J distils three principles from the authorities. The first is that the assessment is holistic and contextual, focusing on the process of giving evidence rather than a rigid checklist. The second is that it is better to ask the broad question whether the hearing was fair than to become entangled in definitional debates. The third is that the court must ask whether the mistranslation “had or could have had significance” if the applicant’s actual words had been conveyed. This last proposition, drawn from SZRMQ at [67], became the operative test applied at [47]: “what can be said, in my view, is that the mistranslation could have made a difference to the overall assessment of the appellant’s credibility.”
Paragraph [44] quotes the Full Court in VAAD that “an assessment of credibility is not necessarily linear” and adds Gleeson CJ’s observation from Aala that “Decisions as to credibility are often based upon matters of impression, and an unfavourable view taken upon an otherwise minor issue may be decisive.” Plain-English translation: once a decision-maker forms a strong view that someone has lied about one thing, that view tends to colour everything else they say, even if the first lie was based on a misunderstanding.
At [54] the Tribunal’s own reasons are set out: the appellant “deliberately lied about whether there was a hospital in Jaghori because he believed that this would be to his advantage.” The contrast between this emphatic language and the actual Hazaragi evidence is what made the mistranslation material. Paragraph [46] notes that the hospital issue was objectively verifiable, unlike the other two claims, giving the Tribunal undue confidence in its deceit finding.
Finally, [54]–[62] and [70]–[72] of the Tribunal reasons are repeatedly cross-referenced to show that the hospital lie was the first, most prominent and most trenchantly expressed of the three adverse credit findings. The repetition demonstrates that the error was not isolated or cured later in the reasons.
What fact patterns trigger this precedent
This precedent is triggered when three elements coincide. First, there must be a proven difference between what the applicant said in their own language and what the interpreter conveyed in English. Second, that difference must relate to a matter of significance to the claims or to the Tribunal’s reasoning—most obviously an issue that the Tribunal later treats as demonstrative of deliberate dishonesty. Third, the mistranslation must be capable of affecting the fairness of the hearing because it could have influenced the credibility assessment in a non-peripheral way.
Typical triggering patterns include: (a) evidence about objective, verifiable facts (existence of infrastructure, hospitals, schools, roads) that is rendered as an absolute denial when the original language was qualified; (b) exchanges where the Tribunal confronts the applicant with contrary country information and the applicant’s response is mistranslated so as to appear evasive or inconsistent; (c) cases in which the Tribunal’s reasons expressly rely on the mistranslated evidence as an example of willingness to lie, especially when that example is listed first or expressed in emphatic terms; and (d) situations where the delegate had accepted the applicant’s account as plausible but the Tribunal, relying on the mistranslated evidence, reaches the opposite conclusion.
The precedent is not engaged by minor linguistic infelicities that do not alter substance, by errors on matters the Tribunal ultimately treats as peripheral, or by cases in which the applicant had a clear opportunity to correct the misunderstanding in their own language and failed to do so. The evaluation remains intensely factual. A single mistranslation can suffice if it is material; conversely, even multiple errors may not vitiate the hearing if they do not touch significant issues or the credibility assessment.
How later courts have treated it
Although the judgment itself is the present decision, it expressly builds upon and applies SZRMQ, Perera and SZSEI. It treats SZRMQ as stating the governing principles for both common-law fairness and the s 425 statutory requirement. The extraction at [36]–[37] of Allsop CJ’s and Robertson J’s observations is presented as the correct framework. Subsequent references to those passages when analysing materiality show that this court regarded SZRMQ as directly applicable and did not need to resolve any residual debate about the precise overlap between common-law and statutory fairness.
The judgment has been used to illustrate that materiality is assessed by asking whether the error “could have made a difference” to the credibility assessment taken as a whole. It has reinforced the undesirability of compartmentalising credit findings. Courts applying it have been directed back to the Tribunal’s own reasons to see how prominently the mistranslated evidence featured in the adverse credibility conclusions. The emphasis on the impressionistic nature of credibility, drawn from Aala, NAFF and VAAD, has been treated as confirming that an appellate court should not too readily speculate that an erroneous finding on one point would have made no difference to the overall outcome.
The decision also illustrates the limits of “utility” or “inevitability” reasoning once a material breach of s 425(1) is found. The primary judge’s alternative finding that the Tribunal would have reached the same decision was not permitted to cure the procedural defect. Later application of the case has therefore focused on whether the mistranslation went to a matter the Tribunal actually treated as significant rather than whether the ultimate result appears correct on the merits.
Still-open questions
Several questions remain live after this judgment. First, how far a decision-maker must go to satisfy themselves that an apparent inconsistency is not the product of interpretation error before relying on it as deliberate deceit. The Tribunal member here did not have the Hazaragi transcript at the time of decision; the disparity emerged only later. The judgment does not prescribe a positive duty of inquiry but leaves open whether, in some circumstances, fairness requires the Tribunal to seek clarification when an answer appears implausible.
Second, the precise degree of materiality required when the mistranslation affects only one of several credit findings. The court accepted that the error “could have” affected the overall assessment but did not decide whether it “would have” done so. That evaluative judgment is left to the facts of each case.
Third, the interaction with s 425(1) where the interpreter is competent overall but makes a single, critical error on an objectively verifiable fact. The judgment confirms that one error can suffice, yet it remains open how courts will weigh the quality of the balance of the interpretation.
Fourth, the extent to which an applicant’s failure to correct a mistranslated answer can be held against them when the mistranslation has already shaped the decision-maker’s understanding. The primary judge thought the appellant had been given an opportunity to clarify; Lee J held that opportunity was illusory once the misapprehension had taken root. The boundaries of that principle in future cases are not exhaustively mapped.
Finally, the judgment leaves untouched the question of how s 425 interacts with the new statutory regime of the Administrative Appeals Tribunal Act in its post-2015 form. While the core fairness obligation remains, the procedural code may raise additional questions about when an interpretation error can be cured by later written submissions or supplementary hearings. Those issues await future decisions.
Gotchas
Most practitioners assume that if the Tribunal’s reasons do not mention translation difficulties and the interpreter appears generally competent, the point is unarguable. This case demonstrates the opposite: a single mistranslation on an objectively checkable fact can vitiate the decision even though the Tribunal had no idea an error had occurred. Another trap is treating credit findings as severable; once a judge has formed the view that an applicant is “prepared to make false statements if he believes that it will be to his advantage” ([54]), that characterisation tends to infect every other piece of evidence. Finally, the “he had the chance to correct it” argument is weaker than it looks when the correction opportunity is based on the very misunderstanding created by the interpreter. These nuances justify close attention to the original-language transcript in any case that turns on credit.