Things I Have Withheld
Kei Miller
WINNER OF THE OCM BOCAS PRIZE FOR NON-FICTION SHORTLISTED FOR THE BAILLIE GIFFORD PRIZE In this astonishing collection of essays, the award-winning poet and novelist Kei Miller explores the silence in which so many important things are kept. He examines the experience of discrimination through this silence and what it means to breach it: to risk words, to risk truths. And he considers the histories our bodies inherit – the crimes that haunt them, and how meaning can shift as we move throughout the world, variously assuming privilege or victimhood. Through letters to James Baldwin, encounters with Liam Neeson, Soca, Carnival, family secrets, love affairs, white women’s tears, questions of aesthetics and more, Miller powerfully and imaginatively recounts everyday acts of racism and prejudice. With both the epigrammatic concision and conversational cadence of his poetry and novels, Things I Have Withheld is a great artistic achievement: a work of beauty which challenges us to interrogate what seems unsayable and why – our actions, defence mechanisms, imaginations and interactions – and those of the world around us.
Award History
0 wins · 2 total
| Award | Year | Result | Category / Notes |
|---|---|---|---|
| Baillie Gifford Prize for Non-FictionMajor | 2021 | Finalist | Secondary source |
| Orwell Prize: Political Writing | 2022 | Shortlist | Secondary source |
Experimental book profileGenerated by GPT-5.4 nano · may contain inaccuracies
This is an unverified interpretation of the catalog description, offered as an opt-in discovery experiment—not as bibliographic fact.
01 Central figures
None extracted with sufficient confidence.
02 Central places
None extracted with sufficient confidence.
03 Suggested argument
Silence surrounding discrimination is breached by risking words and truths, reshaping inherited histories as meanings shift with movement and shifting privilege or victimhood.
Model confidence 63%
04 Reading orientation
45 / 100 academicTrade / academic crossover
An estimate of intended readership and scholarly apparatus—not quality or importance.
Confidence percentages are the model's own estimates. Profile confidence: 78%.