Detects the specific ways AI over-explains, narrates emotions that should be dramatized, and underestimates the reader's intelligence — then provides concrete rewrites.

## Key Points

- Prose feels "hand-holdy" — the narrator explains what should be obvious from context
- Readers report understanding the events but not feeling them
- The narrator tells the reader how to feel about scenes instead of letting scenes create feeling
- AI-generated sections have a "book report" quality — summarizing rather than dramatizing
- After AI Tell Detection and Prose Polish, as the deepest layer of prose quality work
- "She felt sad."
- "Anger surged through him."
- "A wave of relief washed over her."
- "He was overcome with grief."
- "She experienced a profound sense of loss."
- "She agreed to go because she didn't want him to feel rejected."
- "He decided to lie, knowing that the truth would hurt her more."

## Quick Example

```
She slammed the door so hard the picture frames rattled on the wall.
She was furious.
```

```
"Get out," he said, his voice barely above a whisper. His hands
were shaking.
He was barely keeping his composure.
```

Show-Don't-Tell Enforcer

Detects the specific ways AI over-explains, narrates emotions that should be dramatized, and underestimates the reader's intelligence — then provides concrete rewrites.

When to Use This Skill

• Prose feels "hand-holdy" — the narrator explains what should be obvious from context
• Readers report understanding the events but not feeling them
• The narrator tells the reader how to feel about scenes instead of letting scenes create feeling
• AI-generated sections have a "book report" quality — summarizing rather than dramatizing
• After AI Tell Detection and Prose Polish, as the deepest layer of prose quality work

---

The Telling Taxonomy

Type 1 — Stated Emotions

The narrator names the emotion instead of creating it.

The AI pattern:

• "She felt sad."
• "Anger surged through him."
• "A wave of relief washed over her."
• "He was overcome with grief."
• "She experienced a profound sense of loss."

The fix: Replace the named emotion with the specific, physical, behavioral evidence of that emotion.

Told	Shown
"She felt sad."	"She turned the wedding ring on her finger. It was loose now."
"Anger surged through him."	"He set the glass down too hard. Wine jumped across the tablecloth."
"A wave of relief washed over her."	"She exhaled. Her hands unclenched. The pen she'd been holding had left marks in her palm."
"He was overcome with grief."	"He opened the fridge for milk and saw the yogurt she'd bought. Still two weeks from expiring. He closed the fridge."

The rule: If you can feel the emotion without the narrator naming it, the naming is unnecessary. If you can't, the scene needs more dramatization, not more labeling.

Type 2 — Explained Motivations

The narrator tells the reader WHY a character does something instead of letting the action speak.

The AI pattern:

• "She agreed to go because she didn't want him to feel rejected."
• "He decided to lie, knowing that the truth would hurt her more."
• "She kept the secret to protect her sister."
• "He pushed her away because he was afraid of getting close."

The fix: Show the action. Let the motivation be inferred. If it can't be inferred, show earlier scenes that make the motivation clear in retrospect.

Told	Shown
"She agreed to go because she didn't want him to feel rejected."	"'Fine,' she said. She grabbed her coat." (The reader infers the reluctance from 'fine' and the abruptness.)
"He pushed her away because he was afraid of getting close."	Let 3-4 earlier scenes show him pulling back whenever intimacy deepens. The reader connects the dots. No narrator explanation needed.

When motivation-telling is acceptable:

• The character is deliberately hiding their true motivation (gap between stated and real reason is the point)
• The motivation is so counterintuitive that the reader would be confused without help
• The POV character is analyzing someone else's possible motivations (speculation, not certainty)

Type 3 — Narrated Subtext

The narrator explains what's happening beneath the surface instead of letting the surface speak.

The AI pattern:

• "The tension between them was palpable."
• "His words carried an unspoken accusation."
• "There was more to her smile than met the eye."
• "The silence said more than words ever could."
• "An undercurrent of hostility ran through the conversation."

The fix: Write the dialogue and action so well that the subtext is self-evident. The reader should feel the tension without being told "the tension was palpable."

Told	Shown
"The tension between them was palpable."	Write the scene: short sentences, characters avoiding eye contact, someone rearranging objects on the table. Trust the reader.
"His words carried an unspoken accusation."	Write the dialogue so the accusation is audible in the word choice and rhythm: "No, you were at the office. Obviously." The italicized obviously carries more accusation than any narrator explanation.
"The silence said more than words ever could."	Just write the silence. Show what each person does during it. The reader will feel it.

Type 4 — Theme Announcements

The narrator explicitly states the novel's themes instead of embodying them through story.

The AI pattern:

• "She realized that love wasn't about possession — it was about letting go."
• "In that moment, he understood that true strength came from vulnerability."
• "The journey had taught her that home wasn't a place."
• "And perhaps that was the greatest lesson of all: that change was the only constant."

The fix: Delete every explicit theme statement. If the theme is not clear from the story events alone, the story needs revision — not more narrator commentary.

The deletion test: Remove the theme statement. Read the surrounding scenes. Is the theme still communicated? If yes, the statement was unnecessary. If no, the problem isn't the missing statement — it's that the scenes aren't doing their job.

Type 5 — Show-Then-Tell (Redundancy)

The worst hybrid: the AI shows something effectively, then immediately tells the reader what it just showed, as if it doesn't trust its own prose.

The AI pattern:

She slammed the door so hard the picture frames rattled on the wall.
She was furious.

The first sentence SHOWS fury. The second sentence TELLS it. The second sentence is not only unnecessary — it actively undermines the first by implying the reader can't infer fury from a slammed door.

"Get out," he said, his voice barely above a whisper. His hands
were shaking.
He was barely keeping his composure.

"Get out" + whisper + shaking hands = composure breaking. The final sentence is redundant.

The fix: When you find show-then-tell, delete the tell. Always. The show is doing the work. Trust it.

Type 6 — Metaphor Explanation

AI deploys a metaphor or image, then explains what it means.

The AI pattern:

The house stood empty, windows dark like closed eyes. It was as
though the building itself was sleeping — or perhaps dead.

"Windows dark like closed eyes" is a good image. "It was as though the building itself was sleeping — or perhaps dead" explains the image, killing it.

He watched the last leaf detach from the branch and spiral downward.
Like his hopes, falling away one by one.

The leaf image works. "Like his hopes, falling away one by one" bludgeons the reader with the symbolism.

The fix: Deploy the image. Stop. Let the reader make the connection. If the image is good, they will. If they won't, the image isn't right — replace it rather than explaining it.

---

The Show/Tell Ratio

Counting Method

Label each passage (paragraph or sentence cluster) as:

• SHOW: Action, dialogue, sensory detail, or dramatized scene
• TELL: Named emotions, stated motivations, explained subtext, theme statements, summary
• SHOW+TELL: Show-then-tell redundancy (counts against)

Calculate: Show / (Show + Tell) × 100 = Show Percentage

Interpretation

Show %	Assessment
80-100%	Strong. The prose trusts the reader.
65-79%	Moderate. Some over-telling, mostly in transitions.
50-64%	Heavy telling. Significant revision needed.
Below 50%	The narrator is writing a book report about events instead of dramatizing them.

Genre Calibration

• Literary fiction: aim for 80%+ showing
• Commercial fiction: 70%+ is appropriate (slightly more narrator guidance is conventional)
• YA fiction: 65%+ (some telling is age-appropriate)
• Middle grade: 60%+ (more telling is expected)
• Thriller/action: 85%+ showing (momentum depends on dramatization)

---

Scanning Process

Step 1 — Emotion Word Search

Search for all instances of named emotions in narration (not dialogue):

• felt, feeling, emotion, emotional
• angry, sad, happy, afraid, anxious, relieved, overwhelmed
• All variants: fury, sorrow, joy, terror, anxiety, relief
• Emotional metaphors: "a wave of," "a surge of," "a flood of," "overcome with"

For each instance: is the emotion also shown through action/detail? If yes, flag as redundant. If no, flag as tell-only.

Step 2 — Motivation Search

Search for explanatory constructions in narration:

• "because [pronoun] + [verb]"
• "in order to"
• "so that"
• "knowing that"
• "wanting to"
• "[pronoun] did X to [achieve Y]"

For each: is the motivation inferable from context? If yes, flag as over-explanation.

Step 3 — Subtext Search

Search for meta-narration about communication:

• "unspoken," "unsaid," "beneath the surface," "between the lines"
• "the tension," "the silence," "the atmosphere"
• "more than met the eye," "what [they] really meant"
• "palpable," "thick enough to cut," "hung in the air"

Every instance is a flag. If you have to tell the reader the subtext exists, the text isn't creating it.

Step 4 — Theme Statement Search

Search for narrator epiphanies and universal statements:

• "realized that," "understood that," "saw that," "knew that"
• "perhaps," "in the end," "the truth was," "what mattered was"
• Any sentence that could be cross-stitched on a pillow

---

Output Format

# Show-Don't-Tell Audit
**Title**: [Novel title]

## Show/Tell Ratio: [N]% show
**Assessment**: [rating]

## Telling Instances by Type

| Type | Count | Worst Chapter |
|------|-------|--------------|
| Stated emotions | [N] | Ch. [N] |
| Explained motivations | [N] | Ch. [N] |
| Narrated subtext | [N] | Ch. [N] |
| Theme announcements | [N] | Ch. [N] |
| Show-then-tell redundancy | [N] | Ch. [N] |
| Metaphor explanation | [N] | Ch. [N] |

## Flagged Passages with Rewrites

### Chapter [N]

**[Location]**
- Type: [Stated emotion / Explained motivation / etc.]
- Original: "[text]"
- Issue: [why this is telling]
- Rewrite: "[revised text]"

[Repeat for each flagged passage]

## Summary
**Most common telling type**: [type]
**Chapters with heaviest telling**: [list]
**Recommended revision approach**: [guidance]

---

Anti-Patterns

• Eliminating all telling. Some telling is efficient, necessary, and correct. "Three weeks later" is telling. It's also the right choice — nobody needs three weeks dramatized. Telling is appropriate for transitions, backstory summaries, and low-stakes information.
• Turning every emotion into a physical symptom. If every feeling becomes "her stomach clenched" or "his throat tightened," you've replaced one problem with another. Show through ACTION AND CHOICE, not just physiology.
• Demanding constant inference. If a character's motivation is genuinely unclear and the reader needs to understand it for the plot to work, a line of interiority is fine. The goal is to eliminate unnecessary telling, not all telling.
• Ignoring POV. In first person, the narrator IS the character. Some "telling" is actually voice — the character naming their own feelings is what people do. The line between character voice and over-telling is finer in first person.
• Applying the same standard to every scene. A crucial emotional scene needs maximum showing. A transitional scene connecting two important scenes can afford more telling. Calibrate effort to significance.