Translational Justice

Language is failing at what it should do. Communicate.

I follow what happens when interpretation becomes legal fact, when AI works unevenly across languages, and when software treats dialect as an error.

The project asks who carries the cost when fluent language is trusted too quickly.

CF-01 / Asylum
Utterance vs. record "I cannot say this here" rendered as  I do not know . Tribunal finding: claim not credible.
Active investigation
CF-02 / Language Technology
Google Translate, autocorrect and Singlish The words can be right while the human connection disappears.
Fieldwork active
CF-03 / AI systems
Low-resource language bias Automation does not flatten the hierarchy. It encodes it.
Research ongoing
  1. 01 / Surface

    A sentence can sound fluent.

    Its grammar can be clean, its vocabulary plausible and its tone completely assured.

  2. 02 / Knowledge

    Fluency shows knowledge.

    It can show that someone knows the words, the rules or the expected equivalent. That matters, but it is only a beginning.

  3. 03 / Limit

    Knowledge does not prove understanding.

    A person, institution or machine can produce convincing language without grasping the person, context or consequence behind it.

  4. 04 / The gap

    A distance can open between the two.

    On one side sits what the language appears to say. On the other sits what has actually been understood.

  5. 05 / The principle

    False Fluency names the confident mistake.

    It describes fluent output being accepted as proof of understanding and used as if nothing important could be missing.

  6. 06 / Mechanisms

    The gap is widened in recognisable ways.

    Three mechanisms organise the theory. Each gives confidence a different route past the available understanding.

  7. 07 / Mechanism 01

    Bias decides what gets priority.

    Assumptions, stereotypes and institutional aims influence what a translation keeps, edits, domesticates or leaves out.

    Read the Bias mechanism
  8. 08 / Mechanism 02

    Metacognitive failure hides the limit.

    If we do not examine how well we understand, confidence can survive long after the evidence for it has run out.

    See where this mechanism sits
  9. 09 / Mechanism 03

    Missing knowledge can still sound complete.

    Gaps in specialist, cultural, contextual or source-language knowledge can produce confident, incomplete readings.

    See where this mechanism sits
  10. 10 / Asylum

    Fluent error can become legal fact.

    An interpretation enters the record, and the record can shape credibility even when the original words were never preserved.

    Follow the asylum case file
  11. 11 / Technology and AI

    A machine can return the words and miss the person.

    Language technology may produce accurate-looking text while losing tone, dialect, hesitation, idiom and rapport.

    Read the language technology case
  12. 12 / Classics

    A translation can become the whole text.

    Students may learn one English rendering fluently while the source and the translator's choices remain out of view.

    Open the Classics case file
  13. 13 / Complete framework

    Start with the gap. Then inspect how it was made.

    The complete framework connects False Fluency to its mechanisms and to the case files where the consequences become visible.

    Enter the False Fluency framework

Theory map

One central gap. Three mechanisms that widen it.

False Fluency is the parent concept. Bias, lack of metacognitive awareness and knowledge deficits explain how confidence can outrun understanding.

Manifesto / June 2026

The Translational Justice manifesto.

A book-length argument about language, power and the official record, developed through four cases and one recurring claim: something is always lost.

TEDx talk / Where the project began

The Words I Had to Leave Behind

This talk is where I first set out the question that became Translational Justice: what do people have to leave behind to be understood?

Watch on YouTube
Policy briefing / July 2026

Interpretation Failure in the UK Asylum System

Three reforms: independently accredited interpreters, original-language testimony retained beside the English record, and AI operating on what the applicant actually said.

Download DOCX