# Seger's Making a Good Script Great — Rewriting Frameworks

> Source: Linda Seger, *Making a Good Script Great* (1st ed. 1987; 3rd ed. 2010); *Advanced Screenwriting* (2003); *Writing Subtext* (2011)

---

## Core Thesis

Great scripts are not written — they are rewritten. The first draft captures the vision. Subsequent passes refine it through **focused diagnostic lenses** — each pass isolating a single dimension of craft. Trying to fix everything at once fixes nothing. The multi-pass approach is a scalpel, not a sledgehammer.

For AI-generated microdrama, this is critical: the generation pass creates the raw material. The script-doctor pass IS the Seger process — systematic, lens-by-lens revision.

---

## The Multi-Pass Diagnostic Approach

Each pass isolates ONE dimension. Do not attempt to fix multiple dimensions simultaneously — they interfere with each other.

### Pass Order (Recommended)

| Pass | Focus | Question | Priority |
|------|-------|----------|----------|
| **1. Structure** | Three-act turning points, sequence architecture, momentum | "Does the story move forward with escalating stakes?" | Highest — structural problems make all other fixes irrelevant |
| **2. Character** | Arc, motivation, consistency, transformation | "Does the protagonist change? Is the change earned?" | High — character problems undermine audience engagement |
| **3. Theme** | Thematic coherence, subtext, controlling idea | "Does the story mean something beyond its plot?" | Medium-high — theme gives the story weight |
| **4. Dialogue** | Voice, subtext, efficiency, character distinction | "Does each character sound unique? Does dialogue serve multiple functions?" | Medium — dialogue is the most visible symptom of deeper problems |
| **5. Scene** | Individual scene purpose, turning points, economy | "Does every scene turn? Does every scene earn its place?" | Lower — scene-level work after structure/character/theme |
| **6. Polish** | Prose quality, action description, pacing | "Is the writing itself working?" | Lowest — polish last, never first |

**Anti-pattern:** Do NOT start with dialogue polish. Brilliant dialogue in a structurally broken script is lipstick on a pig.

---

## The Structural Pass

### Three Key Structural Questions

1. **Is the catalyst strong enough?** Does it permanently disrupt the status quo? Can the protagonist go back to normal? If yes, the catalyst is too weak.

2. **Is the midpoint a genuine turning point?** Does it change the nature of the problem? After the midpoint, the protagonist should be solving a DIFFERENT problem than before — or the same problem should have revealed a new, more dangerous dimension.

3. **Is the climax the result of the protagonist's choices?** If the climax could happen without the protagonist (deus ex machina, coincidence, external rescue), it's structurally broken. The protagonist's transformation must be the CAUSE of the resolution.

### Momentum Diagnostic

For each sequence, ask:
- **What does the protagonist want in this sequence?**
- **What's preventing them from getting it?**
- **What do they do about it?**
- **How does the result change the situation?**

If any answer is "nothing" or "they wait," momentum has stalled.

### The "And Then" vs. "Therefore/But" Test

Events connected by "and then" are episodic — they happen in sequence but without causation. Events connected by "therefore" or "but" are dramatic — each causes or complicates the next.

| Connection | Type | Quality |
|-----------|------|---------|
| "X happens AND THEN Y happens" | Sequential | Weak — coincidence |
| "X happens, THEREFORE Y happens" | Causal | Strong — consequence |
| "X happens, BUT Y happens" | Complicating | Strong — complication |

**Every episode transition should pass the Therefore/But test.** The script-doctor's Pass 1 already checks this — Seger provides the theoretical foundation.

---

## The Character Pass

### Character Arc Diagnostic

| Question | Healthy Answer | Problem Indicator |
|----------|---------------|-------------------|
| What does the character want? | Specific, concrete external goal | Vague or absent goal |
| What does the character need? | Different from want, internal | Same as want (no inner journey) |
| What's preventing them? | Both external AND internal obstacles | Only external obstacles |
| When do they change? | At a specific, identifiable moment | Gradual fade (unearnable) |
| What catalyzes the change? | A crisis that forces a choice | Time passing or external rescue |
| Is the change visible? | Shown through changed BEHAVIOR | Stated through dialogue |

### Character Dimensionality

Seger identifies three dimensions of character:

| Dimension | What It Covers | How to Develop |
|-----------|---------------|----------------|
| **Backstory** | Who they were before the story | Implied through behavior, not stated in exposition |
| **Paradox** | Internal contradictions | The tough character who's gentle with children. The coward who acts brave. |
| **Transformation** | Who they become | Must be both surprising and inevitable — rooted in paradox |

**The paradox IS the transformation engine.** A character without internal contradictions has nowhere to go. The contradiction between what they are and what they could be IS the arc.

### Character Function Test

For each character: Remove them from the story. What breaks?

- **If nothing breaks:** The character is unnecessary. Cut or merge.
- **If the plot breaks but theme doesn't:** The character serves only mechanical function. Deepen.
- **If the theme breaks:** The character is essential. Protect.
- **If everything breaks:** This is a protagonist or primary antagonist. Ensure full dimensionality.

---

## The Theme Pass

### Subtext Architecture

Theme should be expressed through **subtext**, not text. Seger's hierarchy:

| Level | Description | Example |
|-------|-------------|---------|
| **Text** | What is literally said/shown | "I'm fine." |
| **Subtext** | What is meant beneath the text | (I'm not fine, but I won't show weakness.) |
| **Deep Subtext** | The thematic truth the scene embodies | (In this world, showing weakness gets you killed. Worth is performed, never felt.) |

**For microdrama:** With 450-500 words per episode, every line of dialogue should operate on at least two levels. Pure-text dialogue (meaning is on the surface) wastes the most precious resource: word count.

### Thematic Coherence Test

1. State the theme in one sentence.
2. For each sequence: How does this sequence explore, challenge, or complicate the theme?
3. For the climax: Does the climactic action PROVE the theme?
4. For the antagonist: Does their position represent a credible counter-thesis?

If any sequence doesn't connect to the theme, it's either off-topic or the theme statement is incomplete.

---

## The Dialogue Pass

### Seger's Dialogue Diagnostics

| Problem | Symptom | Fix |
|---------|---------|-----|
| **On-the-nose** | Characters say exactly what they mean | Add subtext — what they want vs. what they say |
| **Exposition dump** | Characters tell each other things they already know | Convert to ammunition — information used in conflict |
| **Voice convergence** | All characters sound the same | Isolate each character's vocabulary, rhythm, sentence length |
| **Speech-making** | Characters deliver themes directly | Break into fragments, make the other character push back |
| **Placeholder** | Dialogue that could be said by anyone | Rewrite so only THIS character would say it THIS way |

### The Cover Test

Cover the character names. Can you tell who's speaking from the dialogue alone? If not, the voices aren't distinct enough.

### Dialogue Economy in Microdrama

With max 40% dialogue (per CONSTANTS.md), every line must:
1. Reveal character (voice, values, intelligence)
2. Advance plot OR complicate the situation
3. Carry subtext

A line that only does ONE of these is underperforming. A line that does none is dead weight.

---

## The Scene Pass

### Scene Purpose Test

Every scene (episode in microdrama) must answer:

1. **What changes?** If nothing changes from beginning to end, the scene is inert.
2. **What's the conflict?** At least two levels of conflict should be active (McKee alignment).
3. **What's the turning point?** Where does the scene pivot?
4. **Could this be cut?** If removing the scene doesn't break the sequence, it should be cut or merged.

### Raising Stakes

Seger's stake-raising techniques:

| Technique | How It Works | When to Use |
|-----------|-------------|-------------|
| **Ticking clock** | Add a time constraint | When momentum stalls |
| **Personal jeopardy** | Threaten someone the protagonist cares about | When external stakes feel abstract |
| **Irrevocable choice** | Force a decision that can't be undone | At sequence transitions |
| **Raising the price** | Increase what the protagonist must sacrifice | When the protagonist seems comfortable |
| **Narrowing options** | Remove escape routes | During Bad Guys Close In |

---

## The Polish Pass

### Action Description Quality

| Quality | Description |
|---------|-------------|
| **Visual** | Describes what the CAMERA sees, not what characters think |
| **Active** | Uses strong verbs, minimal adverbs |
| **Economical** | One paragraph per shot (aligned with camera-test rule) |
| **Evocative** | Creates mood through specific detail, not adjective stacking |

### Pacing Control

- **Short sentences** = fast pace, tension, action
- **Long sentences** = slow pace, reflection, atmosphere
- **Sentence variety** = rhythm, prevents monotony

**In microdrama:** Pacing is controlled at the episode level. Action episodes should use shorter sentences. Emotional episodes can use longer ones. But within an episode, vary length to create rhythm.

---

## Mapping to Recoil Engine

| Seger Concept | Recoil Equivalent | Gap / Enhancement |
|---------------|-------------------|-------------------|
| Multi-pass diagnostics | Script Doctor agent | **Direct match.** Script Doctor already runs structural + dimensional passes. Add: Seger's pass ORDER (structure → character → theme → dialogue → scene → polish). |
| Therefore/But test | Script Doctor Pass 1 | **Direct match.** Already implemented. Seger provides the theoretical backing. |
| Subtext hierarchy | Dialogue quality evaluation | **NEW: Score dialogue on text/subtext/deep-subtext levels** |
| Character dimensionality | Character DNA / behavioral DNA | Add: verify paradox (internal contradiction) in each major character |
| Cover test | Voice distinction check | Existing (voice contamination checkpoints). Add: formal cover test gate. |
| Scene purpose test | Episode evaluation | Add: verify each episode has identifiable change, conflict, turning point |
| Stake-raising techniques | Intensity metrics | Add: verify at least one stake-raising technique per sequence |
| Momentum diagnostic | Treatment/episode arc | Add: verify protagonist has active want + obstacle + action in every sequence |

---

## Alignment with Script Doctor

Seger's multi-pass approach maps directly to the existing Script Doctor agent's architecture:

| Seger Pass | Script Doctor Equivalent | Enhancement |
|------------|-------------------------|-------------|
| Structural | Pass 1 (inter/intra-episode transitions) | Already implements Therefore/But. Add: catalyst strength + midpoint turning-point checks. |
| Character | Broad Diagnostic → Character dimension | Add: paradox check, transformation visibility check |
| Theme | Pass 2 (series-level dimensions) | Add: subtext hierarchy scoring |
| Dialogue | Focused Pass (voice, economy) | Add: cover test, on-the-nose detection |
| Scene | Close Read (per-episode) | Add: purpose test (change + conflict + turning point) |
| Polish | Deep Fix | Add: action description quality rubric |

---

## Gate Integration

### Binary Gate: Therefore/But Transitions
At each episode boundary: Is the transition causal (therefore/but) or sequential (and then)? Sequential transitions = gate failure. (Already implemented in Script Doctor Pass 1.)

### Rubric: Subtext Depth
Score 1-10: What percentage of dialogue operates on multiple levels (text + subtext + deep subtext)? Pure-text dialogue in microdrama wastes precious word count.

### Rubric: Character Paradox
Score 1-10: Does each major character contain a visible internal contradiction that drives their arc? Characters without paradox lack transformation potential.

### Rubric: Momentum Continuity
Score 1-10 per sequence: Does the protagonist have a clear want, a clear obstacle, and take clear action in every sequence? Passive protagonists = momentum stall.

### Pairwise: Structural Soundness
Two adversarial reviewers debate the weakest structural element. One argues the catalyst/midpoint/climax chain is sound. The other identifies the weakest link. Judge evaluates whether the chain holds.

### Script Doctor Enhancement: Gap-Maximizing Reversal Pass

Extends the existing Script Doctor pairwise comparison with a generative adversarial step (see `mckee_principles.md` — Gap-Maximizing Reversal Generation for full specification):

1. Script Doctor identifies key reversals during its structural pass
2. For each reversal flagged as "predictable" or "low gap": generate an alternative version that maximizes the gap between expectation and result
3. Pairwise score the original vs. the gap-maximized version
4. If the alternative wins, it becomes the revision recommendation

This transforms the Script Doctor from purely diagnostic ("here's what's wrong") to diagnostic + generative ("here's what's wrong, and here's what better looks like"). Aligned with Seger's principle that rewriting is not about identifying problems but about producing superior alternatives.

**Backlog:** Plan B — requires generative capability beyond current script-doctor scope.