Field notes / primary interviews
Six anonymised case files from the fieldwork archive. The notes are evidentiary: they record how solicitors, practitioners, interpreters, standards bodies, translation scholars, and refugees describe the systems that turn language failure into institutional fact.
Not language alone. Interpreters step over the line. Interview records become consistency tests. Bad interpretation produces different portrayals of evidence. Recordings reveal repeated errors. Judges dismiss them. Early-stage proceedings have no control. Client and interpreter fail to understand each other. The same interpreter is booked again.
In the moment, the interpreter can still be changed. After the fact, much less likely. If answers return oddly long, short, or off-question, something has shifted in translation. Interpretation requires domain knowledge, including religion. The Home Office is trying to catch people out. The system begins from a standing suspicion: could they be lying?
The DPSI is a "Driving Test" for cognitive load and high-end linguistic capacity. Current systems rely on agency interpreters who may not have passed this test, creating a systemic risk. Enforcement is currently missed due to a deregulated legal framework.
Interpreting is not merely knowing two languages. Systemic pressure from officers and courts creates a zero-clarification environment, turning minor misunderstandings into recorded legal errors. The duty to point out errors is often suppressed by the demand for speed.
Too many interpreter qualifications lowers the value of qualification itself. The framework is fragmented and unregulated: qualifications may exist, but they are not mandatory, and anyone can enter some routes without proving contextual competence. AI training compounds the same hierarchy. Languages with little online presence become low-resource languages, making them harder to model and easier for automated systems to ignore.
Refugees often do not know where to ask for help, or how to begin asking, because the system assumes English before access can even start. The language barrier is not only inside interviews, forms, and hearings. It appears earlier, at the point where a person must identify the right service, understand what it does, and explain a problem in terms the institution can recognise.