Unreliable Narrator

What It Does

Tracks explicit claims made by a first-person narrator ("I never drink," "I'm not a violent person") and checks for subsequent actions or descriptions that contradict those claims ("poured a glass of wine," "slammed his fist into the wall"). Flags the contradiction as either an intentional unreliable narrator effect or an unintentional continuity error.

Why It Matters

The unreliable narrator is a powerful literary device — when done intentionally. The reader gradually realizes the narrator isn't telling the truth, which creates dramatic irony. But unintentional unreliability — where the author simply forgot what the character said three chapters ago — breaks the reader's trust in the writing itself. This analyzer catches both, letting you decide which contradictions are deliberate.

What Gets Flagged

Narrator Contradictions

Severity: Information

Example (flagged):

Unreliable narrator: the narrator claims "I never drink" (line 12) but "poured a glass of wine" appears at line 47 — intentional unreliability or continuity error?

Why: The narrator's explicit claim is contradicted by a subsequent action.

What Gets Tracked

Claims detected (first-person statements with absolute language):

• "I never [verb]"
• "I don't [verb]"
• "I'm not a [noun/adjective]"
• "I would never [verb]"
• "I always [verb]"

Contradictions checked: The analyzer maintains a map of claim topics and scans subsequent text for actions that conflict with those claims.

Scope Limitation

This analyzer only activates in first-person narration (text with high "I/me/my" pronoun density). Third-person narratives are skipped.

Configuration

No configuration options.

Technical Details

• Source: prose-craft
• Scope: Document-level (first-person narratives only)
• Method: Claim extraction via regex for absolute statements; contradiction detection via topic-based action matching