I knew I had relevant frameworks in my vault (Question Generation SOP, Research Response SOP, Q-I-ST Framework) but couldn’t piece them together for studying.
The gap: my systems were built for research → decisions, not learning → exams. I needed to see how they connect.
Secondary problem: I was treating all questions equally. LO sub-questions, confusion flags, and Feynman’s 12 Favourite Problems all felt like they needed the same atomic note treatment. This created hoarding anxiety and retrieval overwhelm.
The Core Insights
1. Learning objectives ARE Feynman questions. The course already did the question-generation work. LO says “Explain X” → that’s the question you’re trying to answer.
2. Most questions are process, not product. They exist to guide attention. Once they do that job, they dissolve. The answer might become a note. The question itself usually doesn’t.
3. My existing SOPs already solve studying. Question Generation SOP → Research Response SOP is the same shape as LO decomposition → textbook reading → understanding. I just didn’t see the mapping.
4. Ideas should be note titles, not buried inside question-titled notes. “Opportunity cost is the next-best alternative foregone” is searchable and linkable. “What is opportunity cost?” with the answer inside is not.
How This Connects to My Existing Systems
| My Existing Framework | Studying Equivalent |
|---|---|
| Question Generation SOP Step 1: State what you’re accomplishing | ”I need to understand [LO] so I can explain it on the exam” |
| Question Gen Steps 2-3: What could go wrong → Questions | Already done—LOs ARE this |
| Research Response SOP Step 1: Question understood | Decompose the LO using structural parsing |
| Research Response Step 3: Supporting evidence | This is what highlighting on phone does |
| Research Response Step 5: Connection made explicit | Claude dialogue—“I highlighted X, help me connect it” |
| Research Response Step 7: Formatted output | Atomic note if worth graduating |
The textbook is just another evidence source. Highlighting is just evidence collection. Claude dialogue is where connection happens.
The Question Hierarchy
Not all questions deserve the same treatment:
| Level | Example | Lifespan | Storage |
|---|---|---|---|
| Feynman 12 | ”How do systems self-organize?” | Years | Permanent note (hub) |
| Research question | ”What size hot water tank?” | Until answered | Becomes alias in the Idea note |
| LO sub-question | ”What is opportunity cost?” | Until exam | Scaffolding; becomes alias if Idea graduates |
| Confusion flag | ”Wait, why does scarcity force choice?” | Minutes | Not a note—just a thought |
Questions don’t dissolve—they become wayfinding. The Idea note gets the claim as title. The questions that led to it become aliases in YAML, making them searchable entry points.
See Question Hierarchy - Four Levels for details.
The Two-Tier Note System
| Tier | Purpose | Lifespan | Format |
|---|---|---|---|
| Scaffolding notes | Guide reading, capture confusion, process understanding | Until end of course | Bundled by chapter, disposable |
| Atomic notes | Permanent knowledge assets | Forever | One idea per note, fully linked |
Most study artifacts are scaffolding. Only ideas with cross-domain value graduate to permanent.
See Scaffolding Notes vs Atomic Notes for the full breakdown.
The Complete Workflow
PHASE 1: Pre-Reading (5 min)
├── Create scaffolding note for chapter
├── Decompose LOs into sub-questions
└── These questions guide your reading
PHASE 2: Reading on Phone
├── Highlight = "This answers one of my questions"
├── Quick note = "I'm confused about X"
└── Don't try to understand everything—just flag
PHASE 3: Processing (Claude session)
├── Open highlights/flags as your agenda
├── Dialogue until it clicks
└── Update scaffolding note with understanding
PHASE 4: Feynman Test
├── Close everything
├── Explain from memory
└── Where you stumble = gap repair needed
PHASE 5: Graduation Decision
├── "Did I think something not in the source?"
├── Yes → candidate for atomic note
└── No → stays in scaffolding, dies after exam
See Five-Phase Study Workflow for phase details. See Highlighting as Flagging for the phone reading method. See Worked Example - LO2 Market Economy for a complete trace through all five phases.
The Ownership Filter (What to Keep)
Instead of: “What might I need later?” (hoarding frame)
Ask: “Is this MY thinking, or someone else’s?” (ownership frame)
| Type | Can You Find It Again? | Keep? |
|---|---|---|
| Textbook definition | Yes—textbook exists | No |
| Study example | Yes—Google it | No |
| Research citation | Yes—Google Scholar | No |
| YOUR insight connecting concepts | No—it was in your head | Yes |
| YOUR framework that clicked | No—you built it | Yes |
See The Ownership Filter for the full reasoning.
Handling Trivia vs Concepts
Different professors test different things. You can’t know which game until you have data (past exams, first midterm, upper-year intel).
| Game | What’s Tested | Strategy |
|---|---|---|
| Conceptual | LOs, mechanisms, application | Feynman method, understanding |
| Trivia | Random details, specific wording | Extraction, cramming |
Don’t mix them. Do concepts during studying. Do trivia extraction separately before exams.
See Trivia vs Concepts - Two Games for the hedging strategy.
Key Linked Notes
Workflow:
- Five-Phase Study Workflow
- Scaffolding Notes vs Atomic Notes
- Highlighting as Flagging
- Worked Example - LO2 Market Economy — complete trace through all phases
Question Management:
- Question Hierarchy - Four Levels
- Decomposing LOs with Structural Parsing
- Seven Lenses for Decomposing Claims — mechanism, actors, conditions, scale, sequence, trade-offs, scope
What to Keep:
Handling Exam Uncertainty:
SOP Integration:
The One-Sentence Summary
Decompose LOs into questions before reading, flag confusion while reading on phone, process via Claude dialogue, Feynman test yourself, and only keep ideas that are YOURS—everything else is recoverable.