magnus919_agent-skills/genius-life/references/creative-process.md

# Creative Process — The Five-Phase Spine

This reference describes the shared process spine used by both modes of
genius-life. Practice mode draws on it informally — an open conversation about
blocks or habits can pass through any of the phases without announcing them.
Development mode runs the five phases explicitly, in order, with worksheet
support.

The model is the evidence-compatible kernel of the unified picture that
Denise Shekerjian assembled from forty MacArthur Fellows in *Uncommon Genius:
How Great Ideas Are Born* (1991), cross-checked against modern creativity
research (see [Evidence basis](../references/evidence-basis.md)). The phases
overlap in real life and are not a rigid checklist; they are a rhythm of
**Prepare → Immerse → Incubate → Capture → Verify**. Cross-cutting practices
from the catalog — perspective-shifting, risk and courage, conditions,
mission, resilience, love of the work — apply at any phase.

Each phase below has four parts: what the research says, what the fellows'
practices are, concrete agent moves, and pitfalls.

## Phase 1 — Prepare: find and frame the real problem

### What the research says

The strongest evidence for spending deliberate time on the problem itself
comes from Jacob Getzels and Mihaly Csikszentmihalyi (1976), who watched art
students compose a still life. The students who explored the objects, handled
them, and kept re-arranging the problem instead of rushing to execute produced
drawings that critics rated more original, and five years later they were
still working as artists while many of the quick solvers had left art
entirely. Problem finding is treated in the literature as a learnable habit,
not a gift.

Arthur Koestler (1964) contributed the related idea of bisociation: a
creative act joins two frames of reference that usually stay separate.
Reframing techniques — re-describing the problem more broadly or narrowly,
posing hypotheticals, expressing it as a metaphor — are the practical moves
that generate those new framings. Classic product-design examples (a can
opener redefined as "the problem of opening," attributed to Arthur D. Little)
show how a broad first definition leads to non-obvious solutions.

### What the fellows' practices are

The book's most patient problem-finder is astronomer Alar Toomre, who narrowed
a boyhood fascination with space into one almost absurdly specific question —
what happens when two galaxies pass near each other — and kept working it for
more than twenty years. The photographer Alfred Stieglitz photographed the
same brick wall many times as one theme with endless variations. Stephen Jay
Gould, whose own talent was making connections, argued that the romantic image
of inspiration hides the plain work of choosing a real problem and staying
with it. Filmmaker Frederick Wiseman and poet Douglas Crase both protected
unstructured, not-yet-defined openings rather than rushing to a plan.

### Agent moves

- Ask what the real problem is before asking how to solve it. Offer reframing
  moves: broader or narrower phrasing, other verbs or nouns, hypotheticals
  ("what if the constraint were gone?"), or a metaphor for the problem.
- Protect the ambiguous start. When the person wants a plan immediately,
  slow down enough to name the question well — but follow their lead on how
  much structure they want.
- Use the [project worksheet](../templates/project-worksheet.md) to hold the
  problem statement and its reframings, if the person wants a worksheet.

### Pitfalls

- Jumping to solutions. Solving the first phrasing of a problem often misses
  the real one.
- Selling a "eureka from nowhere" story — treating the moment of insight as
  if it needed no preparation. Insight without a well-formed problem does not
  produce work.
- Treating the first framing as fixed. Reframing is part of the phase, not a
  failure of it.

## Phase 2 — Immerse: learn through doing

### What the research says

Graham Wallas (1926) called the opening of creative work Preparation —
conscious investigation of the problem. Teresa Amabile's componential theory
names domain-relevant skills (knowledge and technique) as one of the
components of creative output, alongside motivation and the environment.
The deliberate-practice literature (Ericsson and colleagues; Macnamara's 2014
meta-analysis) shows that feedback-rich, focused engagement is the modifiable
driver of skill — with the honest caveat that hours alone do not guarantee
mastery.

### What the fellows' practices are

The book's strongest statement of learning-by-doing is visual artist Robert
Irwin, who gave up an already successful practice to paint the same two-line
painting for two years until he had emptied his own habits. Art historian
Henry Kraus, entirely self-taught, formed his own impressions in the Louvre
before reading what experts said. Wiseman treats the shooting itself as the
research: he films scenes he does not yet understand and finds the film in
the editing. Kirk Varnedoe, reading Picasso's sketchbooks, noted that
repeating a drawing two or three times surfaces something present in none of
the versions. Political scientist Robert Axelrod deliberately read outside
his field, and economist Lester Brown scanned broadly and hired generalists
to keep his categories from hardening. The book adds Hemingway's iceberg rule
in paraphrase: you may leave out what you know, but not what you do not know.

### Agent moves

- Build an immersion plan with the person: what they will make, do, or study
  — weighted toward doing the work itself, not only reading about it.
- Encourage scanning outside the field: books, art, music, or disciplines with
  no obvious relation to the problem.
- Collect raw material without immediate interpretation. Let fragments
  accumulate before explaining them.
- Set up conditions: environment, tools, time. Use the
  [conditions audit](../templates/conditions-audit.md) to reverse-engineer
  what already works.

### Pitfalls

- Endless study without doing. Reading about craft is not the same as making
  the thing.
- Interpreting everything immediately. Premature meaning narrows the material.
- Ignoring conditions. The same person can be highly productive in one
  setting and blocked in another.
- Skipping the doing entirely and relying only on "research."

## Phase 3 — Incubate: deliberate break with a capture plan

### What the research says

Incubation — setting a problem aside — reliably improves later performance,
most strongly on open-ended problems (Sio & Ormerod's 2009 meta-analysis).
The mechanism is debated: undemanding distractor tasks and mind-wandering
have both been linked to gains, though a preregistered 2025 study found that
the specific break task mattered less than how much the person actually
wandered, and a 2019 survey of physicists and writers found most ideas still
arrive during work sessions rather than during wandering. The honest summary:
breaks help, but only after focused effort, and they are not a substitute
for the work.

### What the fellows' practices are

Howard Gardner's method included putting intractable problems aside and
letting the mind work in the dark, then returning with fresh eyes — and
"strobing" a project by forcing repeated early drafts with staggered
deadlines so each return sees the work anew. Robert Shapley kept a garden of
questions: many interesting open problems, so no single failure mattered.
David Stuart let ideas surface from boredom and play.

### Agent moves

- Schedule a genuine break after focused effort, never instead of it. Name
  the return time.
- Set a capture plan before the break: where and how the person will jot
  fragments (phone note, paper, the
  [incubation log](../templates/incubation-log.md)), so a thought is not lost.
- Keep the person's open questions alive. Review the garden of questions when
  they return.

### Pitfalls

- Presenting the break as passive waiting or as a magic pill. Breaks work
  after effort, and no one can promise a reproducible "aha."
- Forgetting the capture plan. A break that loses the thought it surfaces
  gains nothing.
- Letting incubation replace immersion or verification.

## Phase 4 — Capture: quick, low-judgment capture

### What the research says

Volume precedes quality. Dean Keith Simonton's historiometric work (the
"equal-odds rule") shows that within a career, the chance any given work is a
"hit" is roughly constant, so total output predicts total impact. The insight
literature (Jung-Beeman and colleagues; Kounios and Beeman) finds that "aha"
moments carry a measurable neural signature but appear only after extended,
prepared processing — the flash is the cheap part. A 2019 diary study found
most creative ideas occur during work, so the capture habit matters in
work sessions, not only on walks.

### What the fellows' practices are

Wiseman edits fifty to ninety hours of footage into a two-hour film — he
collects vastly more material than he can use, then finds the film. Bill
Irwin gathered street gestures, hallway movements, and mirror practice as raw
parts for pieces he had not yet conceived. David Stuart absorbed hundreds of
glyph images and let matching and variation do the work; even his mistakes
showed something. Ian Graham credited trained attention rather than luck for
his finds.

### Agent moves

- Prompt quick, low-judgment capture: fragments as they arise, without
  editing or explaining.
- Teach the volume logic: generate many possibilities before judging any.
  Quantity is input to judgment, not the product.
- Defer evaluation. Judge late, once fragile thoughts have matured.
- Frame everything captured as work-in-progress, not polished output.

### Pitfalls

- Curating at capture time. Editing while collecting kills the material
  before it can develop.
- Judging too soon. A fragile new thought rarely survives immediate scrutiny.
- Confusing capture with completion. Fragments are raw material.

## Phase 5 — Verify: develop, judge, ship

### What the research says

Wallas made verification an explicit stage: testing the idea and reducing it
to exact form, and he insisted the flash of illumination is worthless without
it. Modern assessments agree that most real creative work happens here, not
at the moment of insight. Amabile's research adds that feedback and
informative evaluation help, while controlling pressure hurts. The systems
view (Csikszentmihalyi, 1996) adds that the work only becomes "creative" when
a field receives and judges it — shipping is how the work enters the world.

### What the fellows' practices are

Derek Walcott is the book's voice for ruthless selection: a professional
knows what is good and discards what is not, while an amateur defends it.
Howard Gardner wrote a large book in six days and practiced cool-eyed
examination of his own failures. John Sayles moved fast to give fragile ideas
form and kept producing even when a project sat shelved for years. Robert
Coles modeled harmonizing instinct with judgment: instinct proposes, judgment
disposes. Bill Irwin's very titles — signaling incompleteness — treat each
piece as a temporary resting place in a continuous process.

### Agent moves

- Help the person give the idea form: outline, draft, prototype, first
  version.
- Guide selection. Encourage discarding weak material without defending it.
- Verify against the original vision: what was the question, and does this
  answer it?
- Reframe shipping as a temporary resting place, not a final monument.
  Review what changed and what to try next.

### Pitfalls

- Skipping verification. The real work of creative practice happens in
  developing, judging, and shipping — not in the flash of insight.
- Polishing too early or defending weak material out of attachment.
- Treating the finished piece as final. The work-in-progress frame is what
  makes closing and moving on liberating.

## Cross-cutting pitfalls (both modes)

- **No eureka-from-nowhere.** Never present insight as a bolt from the blue.
  Every useful idea in the research and in the book's interviews follows
  preparation, effort, and often a period of incubation. If the person talks
  as if ideas should arrive free, gently re-frame: preparation is what makes
  the flash possible.
- **No skipping verification.** Capture is not completion. Developing,
  judging, and shipping is where the work actually gets made.
- **No guaranteed outcomes.** Raise the odds with conditions, volume, and
  process; never promise a result.
- **The person stays the author.** Offer process support, not ghostwriting.