Week 25 of 32
Dialogue pass
Rewrite Dialogue Pass
Run a dialogue pass for distinct character voice and line-level compression across your highest-density scenes.
Lecture
You've fixed the arc. You've trimmed the scenes. The shape is right and the pace is clean. Now you drop to the smallest unit of screenwriting craft that's visible to every reader on every page: the dialogue. Dialogue is where screenwriters are most exposed. Action description can be serviceable without being brilliant — if it's lean and specific, it does its job. Structure can be competent without being inspired — if the chain holds, the story moves. But dialogue is where the reader hears the writer's ear. A character who sounds like a person — who speaks with rhythm, specificity, avoidance, and contradiction that belong to no one else in the screenplay — creates the illusion of life. A character who sounds like every other character — who speaks in the writer's voice with the writer's vocabulary, delivering plot information in complete sentences — creates the feeling of reading a transcript, not watching a movie. The dialogue pass has two objectives: voice (each character sounds unmistakably like themselves) and compression (each line is shorter than the writer's first instinct, because real people under pressure don't explain themselves).
Voice. Every character in your screenplay should sound different. Not in accent or dialect — that's surface, and it's the casting director's territory, not the writer's. Different in how they think out loud. Different in what they reach for first: precision or vagueness, confrontation or deflection, questions or statements, abstraction or concrete detail. Voice is the character's cognitive fingerprint — the pattern of word choice, sentence structure, and conversational strategy that reveals how their mind works. A scientist and a politician can discuss the same contamination crisis, and the scientist will reach for data ("the levels exceeded state threshold by a factor of three") while the politician reaches for framing ("the community's concerns deserve a measured response"). Neither is wrong. Both are character.
Five dimensions of voice. When revising dialogue for voice distinctiveness, examine each character along these five dimensions. No two characters should score the same across all five.
Dimension 1: Sentence length. Some people speak in short, clipped phrases. Others in long, winding constructions. A character who speaks in fragments — "Not possible. Tried that. Ask Torres." — inhabits a different vocal space than one who speaks in measured paragraphs. Sentence length reflects cognitive style: fragments suggest a mind that processes quickly and delivers conclusions without showing its work. Extended sentences suggest a mind that thinks while speaking, building arguments in real time. Assign each character a default sentence length and let it vary under pressure — a character who normally speaks in paragraphs shortening to fragments when frightened is voice under stress, which is the most revealing form of voice.
Dimension 2: Directness. Does the character say what they mean, or do they circle? A direct character names the problem: "You lied to me." An indirect character approaches from the side: "That meeting felt different from what you described." Directness reflects a character's relationship to conflict — direct speakers are comfortable with friction, indirect speakers manage it through implication. In your screenplay, the protagonist and antagonist should occupy different positions on the direct/indirect spectrum, because their conversations will feel richer when one person is naming things and the other is deflecting.
Dimension 3: Vocabulary source. Where does the character's language come from — their profession, their class, their region, their education, their obsessions? A teacher draws metaphors from learning and process. A former military officer draws from operational language — tactical, sequential, objective-driven. A person who grew up in a religious household may use moral language instinctively even if they've left the faith. Vocabulary source is the most powerful voice differentiator because it makes each character's dialogue feel like it comes from a specific life, not from a writer who gave all characters the same access to the same dictionary.
Dimension 4: What they avoid. What a character refuses to say is as distinctive as what they say. A character who never uses the word "sorry" has a different relationship to accountability than one who apologizes constantly. A character who avoids personal pronouns — "the situation requires attention" instead of "I need to deal with this" — is distancing themselves from their own agency. Avoidance patterns are subtext made habitual: the thing the character always deflects from tells the audience what the character is always thinking about.
Dimension 5: Relationship to silence. Some characters fill every silence. Others weaponize it. A character who talks to prevent quiet — who narrates, explains, qualifies, adds disclaimers — is afraid of what silence might reveal. A character who lets silence sit — who answers a question with a look, who holds a pause until the other person flinches — uses silence as a power tool. The contrast between a silence-filler and a silence-wielder in the same scene creates natural friction even before the content of the dialogue generates conflict.
Compression. The second objective of the dialogue pass is making every line shorter. Not shorter for brevity's sake — shorter because real people under pressure communicate in compressed, incomplete, overlapping bursts rather than in the orderly, grammatically complete sentences that writers produce by default. Compression makes dialogue feel spoken rather than written. It creates pace. And it trusts the audience to fill in what's unsaid — which, as you learned in Week 14, is where subtext lives.
The first-and-last-line rule. The most reliable compression technique: for every speech of three or more lines, cut the first line and the last line. The first line of a speech is almost always a preamble — the character warming up, clearing their throat, approaching the point. "Look, I want to be honest with you about something" can be cut, leaving the honest thing itself to do the work. The last line of a speech is almost always a summary — the character restating what they just said, adding a conclusion the audience has already drawn. "So basically what I'm saying is we're out of time" can be cut when the preceding lines have already made the urgency clear. The middle of the speech is where the content lives. Cut the runway and the taxiing. Start on the takeoff.
This rule has a secondary benefit: it eliminates the "on-the-nose" quality that makes amateur dialogue painful. On-the-nose dialogue is dialogue where the character says exactly what they mean with no compression, no avoidance, no obliqueness. "I'm angry at you for betraying my trust" is on-the-nose. "Don't talk to me about trust" is compressed — it communicates the anger without naming it, refers to the betrayal without describing it, and the missing words (the character didn't say "I'm angry" or "you betrayed") become the subtext the audience processes. Cut the first and last lines, and on-the-nose often cures itself, because the explicit framing that made the dialogue too direct was in the preamble and the summation.
Page Craft
The first-and-last-line rule is the dialogue pass's primary mechanical tool. Apply it systematically across your twenty pages of dialogue-dense scenes. For every speech of three or more lines, remove the first line and the last line. Then read the remaining lines aloud. In most cases, the speech works — often it works better, because the middle is where the actual content lives. When cutting the first or last line breaks the speech's meaning, that line was doing real work and should be restored. But in practice, roughly seventy percent of first and last lines can be cut without any loss of meaning. That's where your page savings come from.
A related technique: cutting the response. When Character A asks a question and Character B answers, check whether the answer is necessary. Often the answer is implied by what happens next — a character's action or silence communicates their response more powerfully than a spoken line. "Will you help me?" followed by a cut to the next scene tells the audience nothing. "Will you help me?" followed by a cut to Nora working alone in her lab at midnight tells the audience everything — the answer was no, and the silence of the cut is the refusal.
10-minute drill: Choose the most dialogue-dense page in your draft — the page with the most lines of dialogue and the fewest lines of action. Apply the first-and-last-line rule to every speech of three or more lines. Then identify one response line (an answer to a question) that could be replaced by a physical action or a cut to the next scene. Rewrite the page. Count the dialogue lines before and after. The compressed version should have at least 30% fewer dialogue lines. Time yourself.
Core Reading
Assignment: Read 2 produced feature screenplays, chosen for distinctive character voices.
Script Pair Brief: Select one screenplay where every character sounds distinct — a script celebrated for its dialogue, where you could identify the speaker without character names from word choice, rhythm, and conversational strategy alone. Then select one screenplay that achieves dramatic power through dialogue economy — a film where characters speak rarely and briefly, and the restraint itself becomes a form of voice. The first teaches you how to differentiate voices within a dialogue-heavy screenplay — the craft of making five or six characters all sound like different people having the same conversation. The second teaches you that voice isn't only about what characters say — it's about how much they're willing to say, and what their silence communicates about who they are.
Where to find scripts legally: IMSDb, The Script Lab, SimplyScripts, studio FYC releases, and public library systems.
Why this pairing: Your screenplay needs both skills: differentiated voices for scenes with multiple speakers, and the discipline of economy for scenes where saying less communicates more. Reading both extremes calibrates your sense of when to let characters talk (because the voices are doing character work) and when to shut them up (because the silence is doing it better).
Reading Lens (track these while reading):
1. The name-removal test: choose a three-page scene with multiple speakers. Read the dialogue without looking at the character names. Can you identify who's speaking from voice alone — word choice, sentence length, directness, vocabulary source, avoidance patterns? If you can, what specific technique differentiates the voices? If you can't, are the characters sharing the writer's voice? 2. For each major character, identify their position on the five dimensions: sentence length, directness, vocabulary source, avoidance patterns, and relationship to silence. Which dimension is most distinctive? 3. Find the most compressed exchange in each screenplay — the shortest back-and-forth that communicates the most. How many words does it use? What's unsaid that the audience fills in? 4. Find the longest speech in each screenplay. Does it earn its length — is the speaker saying something that can't be communicated in shorter form? Or could the first-and-last-line rule improve it? 5. Identify a moment where a character's voice changes — where the way they speak shifts because of stress, revelation, or emotional rupture. What changes specifically (shorter sentences? loss of professional vocabulary? sudden directness where there was deflection)?
Journal Prompts:
1. For the voice-distinctive screenplay: write a one-sentence "voice profile" for each major character, describing their default speech pattern across the five dimensions. Then do the same for your own screenplay's characters. How many of your characters have genuinely distinct profiles? 2. For the economical screenplay: find a scene where a character communicates a complex emotional state in fewer than ten words. How does the context (what came before, what the audience knows, what's at stake) enable the compression? Could you write a comparable moment in your own screenplay? 3. Look at one dialogue exchange in your draft that you suspect is on-the-nose — where characters say exactly what they mean with no compression or obliqueness. Rewrite it using the first-and-last-line rule plus one avoidance pattern: have one character deflect from the direct subject while the other pursues it. Compare the versions. 4. Identify the character in your screenplay whose voice is flattest — the character who sounds most like every other character. What's missing from their profile? Which dimension (sentence length, directness, vocabulary source, avoidance, silence) could you adjust to make them distinct? 5. Write one line of dialogue for each major character in your screenplay — a line that only THAT character would say, that reveals their voice profile in six words or fewer. If you can't write the line, the voice isn't specific enough yet.
Writing Exercise
Deliverable: Dialogue pass on 20 pages (chosen for maximum dialogue density).
Constraints: Select the twenty pages of your draft with the highest dialogue density — the pages where the most conversation happens. These should include at least one protagonist/antagonist confrontation, at least one scene with three or more speakers, and at least one scene that you flagged during the cold read as having flat or on-the-nose dialogue. Execute the following revisions:
Voice pass: For every major character who speaks in the selected pages, write a voice profile — one sentence per dimension (sentence length, directness, vocabulary source, avoidance pattern, silence relationship). Then read through all of that character's dialogue in the twenty pages and revise any line that doesn't match the profile. If two characters' profiles are too similar, adjust one character's profile until the voices diverge, then revise their dialogue accordingly.
Compression pass: Apply the first-and-last-line rule to every speech of three or more lines. Cut the response where silence or action could replace a spoken answer. Eliminate on-the-nose preambles ("I need to tell you something"), summaries ("So what you're saying is..."), and emotional labels ("I'm angry" when the anger is visible in the behavior). Count the total dialogue lines in the twenty pages before and after the pass. Target a reduction of 25–35%.
Quality bar: After the pass, the name-removal test must be passable for every scene with two or more speakers — if you remove the character names, a reader should be able to identify the speaker from voice alone at least 80% of the time. No speech should exceed four lines without a compelling reason (a monologue at the climax, a speech that is structurally essential). The compression should make the dialogue feel more natural, not more cryptic — the audience should be able to follow every exchange even with the reduced word count. If compression makes a scene confusing, restore enough to maintain clarity.
Estimated time: 8–12 hours (voice profiles: 1–2 hours; voice revision pass: 3–4 hours; compression pass: 3–4 hours; reading: remaining time).
Human Draft Reminder: You rewrite every line. AI does not generate dialogue. The voices are yours — refined, compressed, and differentiated through your own ear.
AI Workshop
This week's narrowed reader roles: Reader A asks "Can I tell who's speaking without character names?" — evaluating voice distinctiveness through the name-removal test. Reader B asks "Where does dialogue feel like people performing a script vs. people talking?" — evaluating naturalism and compression. Submit a section of your revised dialogue pages (5–10 pages) to both readers.
Add this week's entry to your Disagreement Log. The dialogue-pass divergence: Reader A may identify voice consistency issues — lines where a character breaks profile, moments where two characters sound identical. Reader B may identify naturalism issues — exchanges that are structurally distinct (different vocabulary sources, different sentence lengths) but still feel written rather than spoken, because the distinctiveness is mechanical rather than organic. Both notes point to the same root: voice is not just a set of rules applied to dialogue — it's a way of thinking that inhabits the character so deeply the rules become invisible. If Reader A says the voices are distinct but Reader B says they feel performed, the revision isn't in the profile — it's in the writer's relationship to the character. Spend time thinking as the character before rewriting their lines. Ask: what would this person actually say — not "what would a person with this profile say," but what would THIS person, with THIS history, in THIS situation, actually let themselves say? The distinction is the difference between craft and ventriloquism.
Student Self-Check
Budget Dial
Budget tier this week: Your chosen tier, applied at the dialogue level.
Top 3 production implications of the dialogue pass: 1. Compressed dialogue reduces page count — and page count directly correlates to shooting days. A screenplay that drops from 110 pages to 102 through dialogue compression has eliminated roughly eight minutes of screen time, which translates to roughly one fewer shooting day depending on the schedule. That's a meaningful cost savings at every tier. 2. Voice-distinct characters are easier to cast — when a character has a specific voice profile, the casting director can hear what they're looking for. A character described as "talks in fragments, avoids emotion, draws metaphors from mechanical processes" gives a casting director a clear audition target. A character described as "smart, caring, determined" gives them nothing — every actor can play those adjectives. Distinct voices produce more specific casting, which produces better performances, which is the cheapest production value in cinema. 3. Compression reduces coverage requirements — a five-line speech requires the camera to hold on the speaker for an extended take. A two-line speech allows faster cutting between speakers, which means more flexibility in the editing room and fewer takes needed per setup to cover the dialogue. Shorter speeches are literally easier to shoot.
Cheaper equivalent: A scene that runs four pages because two characters deliver extended speeches explaining their positions to each other can be compressed to two pages by replacing the speeches with the exchange itself — short volleys, interrupted lines, action lines between dialogue blocks. The same content, delivered through interaction rather than presentation. The scene shoots faster, cuts faster, and plays faster.
Worth-it spend: The scene where your best dialogue lives — the exchange with the sharpest voices, the most compressed subtext, the most precise rhythm. That scene deserves extra rehearsal time (for the actors to find the rhythm), extra takes (to capture the precise timing of the pauses), and extra attention in the sound mix (so that every syllable and every silence registers). Great dialogue is performance-dependent. Give the performance room to breathe.
Editorial Tip
Professional script readers read dialogue faster than action. Their eye is trained to scan down the narrow dialogue column rapidly, processing exchanges in seconds. This means two things for you: first, dialogue problems are more visible than action problems because the reader spends more conscious attention on dialogue quality. A clunky line sticks out. A generic voice registers. An on-the-nose speech makes the reader wince. Second, great dialogue creates disproportionate goodwill. A reader who encounters three or four genuinely sharp exchanges — lines that are surprising, compressed, and specific to the character — will carry that goodwill forward through rougher patches of the script. The dialogue pass is where you earn the reader's trust at the line level, the same way the opening ten pages earned it at the structural level. Make the dialogue worth hearing.
Journal Prompt
During the dialogue pass, which character's voice came most naturally to you — the character whose lines you could hear without effort, whose word choices felt inevitable rather than constructed? And which character's voice resisted you — the one whose dialogue you had to force, whose lines felt assembled rather than heard? Write about the difference. The character whose voice comes naturally is probably the character closest to your own cognitive patterns. The character whose voice resists is the one you need to work hardest to inhabit — and paradoxically, that character's voice is often the most interesting to the audience, because it sounds different from everything else in the screenplay. The voice that challenges you is the voice that rewards the reader.
Week Summary
By the end of this week you should have:
• Read 2 produced screenplays with a focus on distinctive character voices (dialogue-rich vs. dialogue-sparse)
• Written voice profiles for each major character across five dimensions
• Executed a voice pass on 20 dialogue-dense pages, ensuring each character's lines match their profile
• Executed a compression pass using the first-and-last-line rule, reducing dialogue by 25–35%
• Eliminated on-the-nose preambles, summaries, and emotional labels
• Completed the dialogue compression drill (densest page revised + one response replaced with action, 10 minutes)
• Run revised pages through both Revision Specialist prompts
• Updated your Disagreement Log with this week's entry
Looking Ahead
Week 26: Clarity Pass (Readability + Production Clarity). Voice is set. Compression is applied. Now the final technical pass before packaging: can a reader follow the geography, blocking, and logic on every page? The clarity pass examines the screenplay for moments where a reader — or a DP, or a director — would be confused about who is where, what they're doing, or how the action connects spatially. You'll tighten action lines for visual precision, verify that every scene's geography is communicable through what's on the page, and ensure the screenplay is "camera-neutral" — written in a way that any director could visualize without bumping into the writer's assumptions about staging or coverage. Clarity is the final craft layer. After clarity, you package.