↓

Direction 1

Backward

Where the wound came from

→

Direction 2

Outward

What the behavior looks like now

↑

Direction 3

Forward

What the cost is doing

The excavation step sits between finding the pressure point and writing the actual song. Writers tend to skip it in both directions — either they go straight to writing without generating raw material first, or they stop generating before they've dug deeply enough. These objections cover both tendencies.

01"I have the pressure point. I'm ready to write. Why excavate first?"

What's true in this

Some writers work best by diving into the scene directly. The pressure point is generative. Writing inside it naturally produces images.

What it gets wrong

Writing inside the scene without first excavating in all three directions tends to produce lines about what is immediately visible in the moment — the outward direction — while leaving the backward and forward directions underdeveloped. The result is a song that shows the behavior but cannot explain how the wound shaped it or what the cost is doing. Excavation is not pre-writing. It is deliberate coverage of the full territory before the song decides which parts of it to enter.

Without excavation you get the surface of the moment. With it you get the depth behind it and the consequence ahead of it.

02"The lines I generate in excavation are rough. I don't want to use bad material."

What's true in this

Excavation produces raw lines, not finished lyrics. Most will not survive into the song unchanged.

What it gets wrong

Raw lines are the right output for excavation. The camera and narrator discipline is applied afterward, precisely to shape raw lines into ones that either show or claim without stating both at once. The excavation step is generative, not editorial. Trying to write polished lines during excavation collapses the two operations into one and produces neither good raw material nor good discipline.

Excavation produces material to discipline. Discipline is a separate step. Do them in that order.

03"I understand camera and narrator lines. I don't need to think about 'earning' claims."

What's true in this

Knowing the difference between camera and narrator lines is necessary and useful. It does not automatically produce good discipline.

What it gets wrong

Discipline is not just labelling lines. It is sequencing them so that camera lines arrive before the narrator claims they justify. A writer who knows camera from narrator can still write a song where the narrator claims things the camera never showed — where the song asserts rather than earns. Earning means the listener has been shown the evidence before the claim is made. That is a structural decision that requires deliberate sequencing, not just line-level identification.

Knowing which type a line is — and knowing whether it has earned its place in that sequence — are different skills.

04"Forward lines feel speculative — the character doesn't know what the cost is yet."

What's true in this

The character often does not see the cost yet. That blindness is part of the CI's argument. Forward lines can feel like editorial intrusion.

What it gets wrong

Forward lines do not require the character to know the cost. The narrator or the camera can show the cost while the character remains oblivious — which is the song's structural irony. The gap between what the character sees and what the song shows the listener is exactly where the forward direction lives. Forward lines that avoid naming what the character doesn't know still show the listener what is accumulating. That is not speculation. That is the song doing its job.

The character's blindness is not an obstacle to forward lines. It is the content of them.

The pressure point is a scene, not a song. It is the location the song enters — the specific recurring moment where the controlling idea's claim is visibly and simultaneously true. But a song cannot live entirely inside one moment. It needs history to explain how the wound formed. It needs observable present detail to show the behavior in action. It needs consequence to show what the behavior is accumulating, whether the character sees it or not.

Those three requirements correspond to three directions of excavation from the pressure point. None of them is the pressure point itself. They are what surrounds it — the territory the song needs to cover in order to show the listener the full argument rather than just its most compressed moment.

↓

Direction 1

Backward — Where the wound came from

Reach back from the pressure point into the character's history. The wound did not begin at the pressure point — it was shaped by specific prior events, patterns, and moments that explain why this particular behavior took root. Backward lines establish the wound's origin and give the listener the context to understand why the behavior in the present moment is not a choice but a compulsion. They answer the unspoken question: why is this person doing this?

Questions that generate backward lines:

What happened the first time this feeling appeared? What specific event, relationship, or pattern produced the wound?

What did the character learn to do, in earlier life, that became the habit that is running now?

What did they give up, lose, or leave behind that explains why the behavior became necessary?

→

Direction 2

Outward — What the behavior looks like right now

Stay inside the present moment and describe what the camera can see. Outward lines are the most naturally available direction — they describe the observable surface of the pressure point scene, the specific physical details, actions, objects, and sounds that exist in the moment itself. These are the lines that make the song feel real and located. They are almost always camera lines. The risk with outward excavation is stopping too early — generating only what is immediately obvious and missing the telling details that are further in.

Questions that generate outward lines:

What does the camera see in this exact scene? What objects, settings, sounds, gestures, small physical details are present?

What does the character do with their hands? What do they look at? What do they avoid looking at?

What other person, if anyone, is present — and what do they do that the camera catches without explaining?

↑

Direction 3

Forward — What the cost is accumulating

Look at what the behavior is producing — not as a future event but as a present accumulation that is already visible if the listener knows where to look. Forward lines show the consequence: what is being lost, eroded, or foreclosed by the behavior the character keeps executing. The character may not see it. The listener does. This directional gap — the listener seeing what the character cannot — is the structural irony the song requires. Forward lines are where the song's argument becomes visible, and where narrator discipline is most needed to earn the claim rather than just state it.

Questions that generate forward lines:

What is the behavior costing the character that they haven't noticed yet? What is already gone that they think is still there?

What observable evidence exists right now — visible to the listener — that shows the cost accumulating?

What would the listener see, watching this scene, that the character is not watching?

Rule: Backward lines give the wound history. Outward lines give the behavior visibility. Forward lines give the cost evidence. A song that excavates all three directions has the material to earn its argument. A song that excavates only one direction has material for a scene, not a claim.

◈

Checkpoint

Quiz 1 — The Three Directions

1.Backward lines answer what unspoken question the listener brings to a song?

↳Backward lines establish the wound's origin — the history that explains why the behavior is not a choice but a compulsion. Without that history the listener watches the behavior but doesn't understand why it is inevitable for this particular person.

2.Outward lines are described as "the most naturally available direction." Why is this also a risk?

↳Outward lines are the first ones a writer finds because they describe what is immediately visible. The risk is stopping once the obvious details are written. The telling specific details — the ones only this song could have — are further in the outward direction, past the first-draft surface.

3.Forward lines show the cost accumulating "visibly to the listener but not to the character." Why is this gap structurally necessary?

↳The CI argues that the behavior costs more than it protects. For that argument to be visible in the song, the listener must be shown the cost — even as the character remains oblivious. That gap between character knowledge and listener knowledge is where the song's claim lives structurally. Without it the song can show the behavior but cannot make the argument.

4.A song that excavates only the outward direction will have material for what — and lack what?

↳Outward excavation produces excellent camera lines about the present scene. But those lines show what is happening — not why it is happening (backward) or what it is producing (forward). Without those two directions the song is visually specific but argumentatively incomplete.

Excavation is a generation operation, not an editorial one. The goal is to produce as many raw lines as possible in each direction before filtering or shaping any of them. The quality gate comes later. The excavation step is about coverage — making sure the writer has genuinely explored the backward, outward, and forward territory before deciding what the song will use.

The practical method is simple. Place the pressure point at the center. Set a timer or a quantity target for each direction — ten lines minimum, no cap. Write without filtering. Do not evaluate whether a line is good, useful, or well-phrased while writing it. Evaluation during generation collapses the two operations and produces neither good raw material nor good judgment about it.

Worked Example — Excavation from the Sunday Phone Call

Pressure Point

"She calls her mother every Sunday and describes the city the way a travel brochure would. Her mother says it sounds wonderful. She says yes, it really is."

↓ Backward — raw excavation lines

Left at twenty-two with one suitcase and a story that was already half-invented

Her mother cried at the airport and she didn't look back because she'd rehearsed not looking back

The town knew her as the girl who was going to leave — so she left before they were wrong about her

She told the story of why she left so many times the telling became the memory

What she left behind: a bedroom that still has her high school ribbons on the wall

Her father's face when she said she wasn't coming home for his birthday

→ Outward — raw excavation lines

Sunday morning, phone warm in her hand, window on the city behind her

The way her voice changes two syllables into the call — brighter, faster, not hers

She uses the same four words every week: it's so good here

She laughs at things that aren't quite funny so there's no gap for a question

Her mother's voice: it sounds wonderful — and the pause before she answers yes it really is

After the call: she sits with the phone still warm, the city making its city sounds outside

↑ Forward — raw excavation lines

Her mother stopped asking the hard questions two years ago — she answered too well

The city she describes on the phone is not the city she lives in

One day she will not remember what she thought the first week here — the version that knew it wasn't working

Every Sunday call is one more wall between the version who left and the version who is still leaving

She cannot go back without performing the person who didn't need to

The lease is up in April and she has already told her mother she is renewing

These are raw lines — unfiltered, some rough, some redundant. Several are too on-the-nose and will need to be shaped or replaced. But the coverage is there: backward establishes the wound's history and the original decision, outward gives the scene specific observable life, forward shows what the behavior is accumulating. The camera and narrator discipline step (Section 3) is what turns these raw materials into lines that earn their place.

Rule: Generate without filtering. The editorial step requires material to edit. A writer who filters while generating produces too little material and too early judgements. Write the obvious lines and write past them. The useful ones are in both halves of the list.

◈

Checkpoint

Quiz 2 — How to Excavate

1.Why should excavation and editorial filtering be kept as separate operations?

↳Generation requires suspension of judgment to produce enough material. Judgment during generation stops the flow before the full territory is covered. The useful lines — the specific, telling ones — are often in the second half of the list, past where a filtering writer would have stopped. Separate the operations to serve both.

2."She told the story of why she left so many times the telling became the memory" is a backward line because:

↳Backward lines establish how the wound was shaped in the past. This line shows the mechanism — the repeated retelling replacing the original event — which is the history that explains why the Sunday call is now a compulsion rather than a choice. The present behavior is an extension of this backward pattern.

3."The lease is up in April and she has already told her mother she is renewing" is a forward line because:

↳Forward lines show what the behavior is already producing. The April lease is not speculation — the character has already acted, already told her mother, already extended the compulsion into the next decision. The listener sees the cost (one more year, one more wall, one more Sunday call) before the character processes it. That is forward excavation.

4.What is the purpose of the ten-line minimum per direction?

↳The first lines in any direction are the obvious ones the writer already knew. Lines seven, eight, nine are where the specific and inevitable material hides. A minimum forces the writer to exhaust the obvious before finding the interesting. Without it the writer stops too early in every direction and has only the surface.

Raw excavation lines come in two types. Some describe observable facts — things a camera could point at without interpretation. Others make claims about what those facts mean. The first type is the camera line. The second is the narrator line. The discipline step identifies which type each raw line is, and then shapes it into the cleanest version of that type.

A camera line stripped of any interpretive layer is a line the listener can see. A narrator line that has been sharpened to its most precise claim is a line the listener can trust — because its camera line has already appeared. The discipline step is not about choosing between camera and narrator. A song needs both. It is about ensuring that each line is doing only one job and doing it cleanly — that camera lines point without claiming, and that narrator lines claim without pointing.

Camera Points. Does not interpret.

A camera line shows a specific observable detail without naming what it means. Its job is to produce the evidence — the image, the action, the object, the sound — that the narrator will later draw meaning from. A camera line that includes interpretation has contaminated the evidence with the verdict. The listener can no longer experience the image before being told what to think about it.

Contaminated camera line

"She laughs at things that aren't funny — desperately filling the silence so her mother can't ask the real question."

The image is there (the laugh). But the interpretation ("desperately filling the silence") is attached. The camera and the narrator are speaking at the same time.

Clean camera line

"She laughs at things that aren't quite funny."

The camera points. The listener sees the detail. The interpretation has been removed. The narrator can earn the claim later after the image has done its work.

Narrator Claims. Does not point.

A narrator line makes a claim about what something means — names the pattern, states the cost, identifies the irony. Its job is to interpret the evidence the camera has produced. A narrator line that is also pointing at something (a concrete image embedded in the claim) has taken on camera work. The two functions blur and the line does neither cleanly. The narrator should be given only the job it is supposed to do: make the claim the camera has earned the right to make.

Contaminated narrator line

"She answered too well — her mother stopped asking the hard questions — like whether she was really happy in the city she chose."

The claim ("answered too well") is there, but it drags the camera with it. The specific detail of the mother stopping is buried in the claim rather than arriving on its own as evidence first.

Clean narrator line

"She answered too well."

The claim stands alone. Four words. Earned by the camera line that showed the behavior. The camera pointed; the narrator names what the pointing revealed.

The discipline step applied to raw excavation lines is mechanical in the best sense. For each raw line: identify which type it is trying to be. Then strip it down to that type. Remove the interpretation from the camera lines. Remove the images from the narrator lines. What remains is a line that does one job cleanly — and is therefore available to be sequenced with other lines that do the other job, so that the song earns its claims rather than just making them.

Rule: A camera line that interprets has already spent the narrator's credit. A narrator line that points has already spent the camera's credit. Each line should do one job. The song earns its claims by sequencing them so that camera precedes narrator — evidence before verdict.

◈

Checkpoint

Quiz 3 — Camera and Narrator Discipline

1.A "contaminated camera line" is one that:

↳A contaminated camera line has both image and interpretation collapsed into the same line. "She laughs at things that aren't funny — desperately filling the silence" has the image (the laugh) and the verdict (desperately filling silence) in the same breath. Strip the verdict to clean it.

2.The discipline step for narrator lines removes what from them?

↳A narrator line that contains an image is doing two jobs at once — pointing and claiming. Removing the embedded image lets the claim stand cleanly. The image then either belongs in its own camera line before this narrator line, or it can be cut. The clean narrator line can only be earned if the camera has already done its pointing separately.

3."She laughs at things that aren't quite funny" is a clean camera line. "She answered too well" is a clean narrator line. Why does the narrator line require the camera line to have appeared first?

↳A claim without evidence is an assertion — the listener is asked to accept it on the narrator's authority alone. A claim after evidence is a conclusion — the listener has been shown what the narrator is drawing from, and the claim lands because the evidence has done its work. This is the structural difference between a song that asserts and a song that earns.

4.The discipline step is described as "mechanical in the best sense." Why is the mechanical quality a virtue here?

↳The discipline step needs to be consistent to be useful — every line treated to the same question and the same stripping operation. The mechanicalness is the consistency. It removes the temptation to make case-by-case exceptions ("this line can do both") that produce contaminated lines and undermine the earning structure.

Applying camera and narrator discipline to individual lines produces clean lines. Sequencing those clean lines so that camera precedes narrator — evidence before verdict — is what makes the song earn its claims. The discipline step is line-level. The earning step is structural.

A claim is earned when the listener has been shown the evidence before being asked to accept the interpretation. A claim is unearned when the narrator makes it without the camera having first pointed at anything. The structural test is simple: for every narrator line in the song, ask — what camera line did this arrive from? If the camera line exists and preceded the narrator line, the claim is earned. If no camera line preceded it, the narrator is asserting rather than concluding. That is the difference between a song that earns its argument and one that merely makes it.

Unearned Narrator claims without camera evidence

The narrator states the claim — names the pattern, identifies the cost, draws the conclusion — without the camera having pointed at any evidence first. The listener is asked to accept the assertion on the narrator's authority. A listener who doesn't already feel what the song is claiming will not be moved by this; they will feel told rather than shown.

Example sequence (unearned)

"She's becoming a stranger to herself." [narrator claim]

"She calls home every Sunday and says she's fine." [camera evidence — arrives too late]

The claim arrives before the evidence. The listener is told the verdict before seeing the behavior that warrants it. The song asserts.

Resequenced (earned)

"She calls home every Sunday and says she's fine." [camera evidence — arrives first]

"She's becoming a stranger to herself." [narrator claim — now earned]

The camera points first. The listener has the evidence. The narrator's claim lands as a conclusion rather than an assertion.

Earned Camera points, then narrator concludes

The camera line appears first — points at the specific observable detail, the behavior, the object, the gesture. The narrator line arrives after — draws the meaning, names the pattern, states the cost. The listener has been shown the evidence before being asked to accept the verdict. The narrator's claim is available to land because the camera created the conditions for it. This is the debt-and-payment structure: the narrator takes out a debt (a claim that needs proving) by borrowing from the camera (the evidence that proves it).

Camera debt

"She uses the same four words every week: it's so good here."

"The city she describes on the phone is not the city she lives in."

The camera produces the evidence (the four repeated words). The narrator draws the conclusion (the described city is not the real one). Evidence, then verdict.

Why it works

The listener hears the four words first and forms a suspicion. The narrator then confirms what the listener was already beginning to feel. The song and the listener arrive at the claim together.

This is the structural experience of a song that earns: the listener is not told what to feel — they are shown what is happening, and they feel it before the narrator names it.

The three directions of excavation map directly onto this earning structure. Backward lines tend to produce camera evidence for narrator claims about why the wound formed. Outward lines produce camera evidence for narrator claims about what the behavior looks like right now. Forward lines produce camera evidence for narrator claims about what the cost is doing. Each direction generates raw material that can be sequenced into earned claim structures — but only if the discipline step has separated the camera lines from the narrator lines first.

Rule: For every narrator claim in the song, there must be a camera line that preceded it. If the camera line doesn't exist yet, the claim is unearned. Either write the camera line or cut the narrator line. There is no third option.

◈

Checkpoint

Quiz 4 — Earning Claims

1.The structural test for whether a narrator claim is earned is:

↳The earning test is structural: every narrator claim must have a prior camera line as its evidence. The sequence matters — camera before narrator. A claim that precedes its camera evidence is an assertion. A claim that follows it is a conclusion. The test makes this distinction operational.

2.Resequencing "She's becoming a stranger to herself" to appear AFTER "She calls home every Sunday and says she's fine" transforms the narrator line from an assertion into:

↳The resequencing changes the structural relationship between the lines. The same narrator claim, arriving after the camera evidence, is now a conclusion the listener can confirm rather than an assertion they must accept. The words are identical. The earning status changes entirely based on sequence.

3.How do the three excavation directions map onto the debt-and-payment earning structure?

↳Each direction generates raw material in both types — camera lines and narrator lines. The backward direction produces images of the wound's history AND claims about why the wound formed. The forward direction produces images of the cost accumulating AND claims about what the cost means. Each direction's camera lines earn that direction's narrator claims. The discipline step separates them; the earning step sequences them correctly.

4.If a narrator claim has no preceding camera line, the only two options are:

↳An unearned claim has two resolutions: find or write the camera line that precedes it, or remove the narrator line. There is no middle position. "Strong enough to stand on its own" is the narrator asserting — which is precisely what the earning structure is designed to prevent. Every narrator claim needs its camera evidence.

The full process — excavation in three directions, discipline applied to raw lines, claims earned through sequencing — fails in three specific ways. Each failure corresponds to a different stage of the process.

FM1Single-Direction Excavation

The writer excavates in only one direction — almost always the outward direction, because it is the most naturally available. The song has strong present-tense camera lines but cannot explain why the behavior is a compulsion (no backward material) and cannot show what it is accumulating (no forward material). The song is visually specific but argumentatively thin: it describes the behavior without demonstrating the claim about what the behavior costs.

A song with vivid present-scene images — the phone, the window, the voice change, the four words — but no lines about how the choice was formed, and no lines about what the Sunday calls are foreclosing. Specific but one-dimensional.

Problem: The listener can see the behavior clearly but cannot feel why it matters. The CI's argument — that defending the choice is how it comes to own you — is not visible because the forward material (the ownership accumulating) was never excavated.

Fix: After excavating outward, deliberately set a timer for the backward and forward directions separately. The three directions are not naturally equal in availability. Outward lines come easily. Backward and forward lines require the writer to push past what the present scene shows.

FM2Skipping Discipline

The writer excavates in all three directions but skips the camera-narrator discipline step and moves directly to sequencing and drafting. The raw lines — which mix camera and narrator within individual lines — go into the song without separation. The result is lines that do two jobs at once: pointing and claiming, showing and telling, simultaneously. The song loses the structural gap between evidence and verdict. The listener is told and shown at the same time rather than shown first and then helped to name what they saw.

"She laughs at things that aren't funny, desperately filling the silence so her mother can't ask the real questions — she's been performing so long she's forgotten what she actually thinks."

Problem: One line contains a camera image (the laugh), an interpretation (desperately filling), a narrator claim (forgotten what she actually thinks), and a second camera gesture (so her mother can't ask). All of it in one breath. The listener is overwhelmed rather than shown, then told. The earning structure is gone.

Fix: Apply the discipline step to each raw line before sequencing. Strip each line to its type. "She laughs at things that aren't quite funny" — camera. "She's forgotten what she actually thinks" — narrator. Separate. Sequence. The earning is in the gap between the two.

FM3Narrator-First Sequencing

The writer has clean camera and narrator lines but sequences them with the narrator before the camera — verdict before evidence. The narrator makes a claim that the subsequent camera line would have earned, but arrives before the listener has seen the evidence. The listener is told what to think before they have been shown what to think it about. The song asserts. It may be emotionally correct but it does not earn. The listener feels pushed rather than led.

Sequence: "She answered too well." [narrator] — "She used the same four words every week: it's so good here." [camera]

Problem: The claim arrives before the evidence. The listener must accept "she answered too well" on the narrator's authority — the four words that would have proven it arrive afterward, too late. The earning has been reversed. The song tells, then shows, which is the structural formula for assertion rather than conclusion.

Fix: Resequence. Camera first: "She used the same four words every week: it's so good here." Narrator second: "She answered too well." The same lines. The same words. The earning is in the sequence, not in the lines themselves.

Diagnostic Questions

FM1

Test: Did you excavate all three directions, or only the outward one?

Fix: Set a separate timer for backward and forward. They are not naturally available. The discipline of covering all three directions is what gives the song its argument.

FM2

Test: Did you apply camera-narrator discipline to each raw line before sequencing?

Fix: Apply the discipline step line by line. Strip each to its type. Do not draft until the lines are separated. The earning structure requires clean lines.

FM3

Test: For every narrator line, does a camera line precede it?

Fix: Resequence. Camera before narrator, always. If the camera line doesn't exist, write it or cut the narrator line. The sequence is where the earning lives.

Rule: FM1 gives the song a scene without an argument. FM2 gives the song an argument without clarity. FM3 gives the song an argument in the wrong order. All three produce a song that tells the listener what to feel rather than giving them what they need to feel it themselves.

◈

Checkpoint

Quiz 5 — Three Failure Modes

1.FM1 (single-direction excavation) produces a song that is "visually specific but argumentatively thin." What is missing?

↳Outward-only excavation produces a song that shows what is happening but not why it is a compulsion (backward) or what it is costing (forward). Those two directions are where the CI's argument lives. Without them the song is a scene, not a claim.

2.FM2 (skipping discipline) produces lines that "do two jobs at once." Why is this structurally harmful?

↳The earning structure requires a temporal gap: camera arrives first, narrator arrives second. A line that contains both collapses the gap — the evidence and the verdict are simultaneous. The listener cannot feel the evidence before being told what it means. The song asserts rather than earns.

3.FM3 (narrator-first sequencing) can occur even when all lines are clean and correctly typed. Why?

↳Clean lines placed in the wrong order produce the same result as unclean lines: assertion instead of earning. The discipline step makes lines clean. The sequencing step makes claims earned. Both are required. A song with clean lines and wrong sequence is FM3 — it has done the first step and skipped the second.

4.All three failure modes produce a song that "tells the listener what to feel rather than giving them what they need to feel it themselves." What is the structural mechanism that produces this shared outcome?

↳The evidence-before-verdict structure is what allows the listener to arrive at the feeling before the song names it. FM1 lacks the evidence (no backward or forward material). FM2 collapses evidence into verdict (contaminated lines). FM3 reverses the sequence (verdict before evidence). The structural outcome is the same: the listener is pushed toward the feeling rather than led to it.

Full assessment across five categories: the three directions, how to excavate, camera and narrator discipline, earning claims, and failure modes.