Writing the Unreliable Narrator Without Cheating the Reader

The unreliable narrator has become one of contemporary fiction's most abused devices. Used well — in 'We Need to Talk About Kevin', in 'Atonement', in 'Gone Girl' — it creates a retrospective revelation that recontextualises everything you have read. Used poorly, it is simply a plot twist that requires the author to have withheld information that the narrator would obviously have shared.

The distinction matters. An unreliable narrator deceives herself. She does not deceive the reader on behalf of the author. The reader should always be able to see the unreliability operating, even if they cannot yet name what it conceals.

The contract with the reader

The implied contract in first-person narration is that the narrator is reporting her experience as she understands it. Unreliability enters through the gap between what she reports and what is true. The reader's job — and pleasure — is to inhabit that gap.

• Unreliability through self-deception: the narrator believes what she says but is wrong (Stevens in 'The Remains of the Day')
• Unreliability through selective memory: she knows more than she tells, but for psychologically credible reasons
• Unreliability through limited perspective: she can only report what she can see

The reader must be able to reconstruct the truth from the evidence the narrator provides. If the truth can only be reached through information the narrator would never have given us, the narrative has cheated.

The test: can a careful reader, on re-reading, find the clues seeded throughout? If yes, you have written an unreliable narrator. If no, you have written a plot twist.