We don’t perceive language neutrally. Long before we process the actual sounds, our brains are already busy predicting, filtering & filling in. Accent, accuracy, word choice, formality—none of these reach the mind untouched. Perception is never passive, shaped by experience, bias & the social meanings we attach to speakers. A paper from a few years ago by Kutlu, Tiv, Wulff & Titone’s in Applied Linguistics takes this everyday truth & pushes it into sharper relief by asking what happens when the face we see doesn’t match the voice we hear.
The Study
The authors tested 58 US undergraduates using normed recordings of American, British & Indian English, each paired with either a white or South Asian female face. Participants completed two tasks:
- Intelligibility: participants listened to short sentences played with background noise. The noise level was set so the speech was only slightly louder than the noise (a 4 dB signal‑to‑noise ratio), which makes accurate transcription genuinely challenging
- Accentedness: participants rated how strongly accented each speaker sounded using a 9‑point scale. Crucially, they weren’t judging the audio alone. They saw a face alongside each voice, then decided how “accented” the speaker seemed. This allowed the researchers to measure how visual cues—specifically racialised cues—influenced listeners’ sense of linguistic distance or “non‑nativeness”.
They also completed a detailed social network questionnaire to determine whether their close contacts were racially diverse or not.
The findings
Across the board, sentences paired with South Asian faces were transcribed less accurately. The audio never changed—only the face did.
Indian English was the least accurately transcribed overall, but crucially, Indian English became more intelligible when paired with a white face.
Accentedness judgments followed the same pattern: South Asian faces increased perceived accentedness for all varieties, including American English. But, unsurprisingly, listeners with more racially diverse social networks judged all varieties as less accented overall. Diversity softened bias; lack of diversity amplified it.
These results echo classic work by Rubin (1992) & Babel & Russell (2015), but extend it by showing how racialised expectations shape both subjective judgements & objective behaviour. The same sentence becomes harder to hear depending on who we think is speaking.
Context
This fits into a long tradition of research on native speakerism, racialised language ideologies & the politics of English varieties. Holliday’s work on native speakerism provides the broader frame: ideas about who “should” sound like what are socially produced, not linguistically grounded. Rubin’s early findings showed how expectations distort intelligibility, while Babel & Russell demonstrated how visual cues shift perceived accentedness. Together, they help explain why perception is never just about sound — it’s about the social meanings listeners bring with them.
Teacher Takeaways?
- Make ideology explicit: build activities that help learners notice how “accent” is socially constructed rather than a measure of ability.
- Diversify input: bring in a wider range of global Englishes (for example Indian, Nigerian, Singaporean or Filipino English) so learners regularly hear English as it’s used around the world, not just inner‑circle norms.
- Challenge expectations: use audio‑only tasks before revealing speakers to spark discussion about bias & perception.
Where do you see perception shaping communication in your own professional world?
Leave a Reply