NiNiProblem statement
Credibility in asylum decisions is judged partly on how testimony is told: consistency, specificity, hesitation. In interpreted hearings, every one of those cues reaches the decision-maker only after passing through an interpreter who is professionally obliged to render meaning, not manner.
Existing critique of asylum credibility assessment focuses on decision-maker bias. This study relocates part of the problem upstream, to a routine and well-intentioned linguistic transformation.
Research questions
- RQ1. Which pragmatic features — hedges, self-repairs, discourse markers, narrative sequencing — are most frequently altered in interpretation?
- RQ2. How do decision-makers' written reasons take up those altered features as evidence of credibility?
- RQ3. Do alterations pattern by interpreter working conditions, such as consecutive versus simultaneous mode?
Materials and method
A corpus of 40 anonymised hearing recordings with matched written determinations, obtained through an established access agreement with a legal aid organisation and with applicant consent secured by the organisation, not by the researcher.
Recordings are transcribed in both languages by independent transcribers. Analysis proceeds in three passes: pragmatic feature coding of the source turn, comparison against the rendered turn, and tracing of surviving or lost features into the determination text.
Ethics
Participants are, by definition, people in a precarious legal position. Consent is sought after a decision is final, materials are pseudonymised before reaching the researcher, and no case is reported in enough detail to be identifiable. The project offers no benefit to individual applicants and says so during consent.
Limitations
One jurisdiction and a small number of language pairs. Access through a legal aid organisation over-represents represented applicants. Findings describe how testimony is reshaped, not how often that reshaping changes outcomes.