Abstract general ideas from concrete facts #
In philosophy, abstraction means drawing general knowledge out of raw facts and cases. Without a clear method, agents fail at two extremes:
- Case-by-case recounting: Repeating what happened in each event. This traps the insight inside local names, dates, and one-off details.
- Hollow truisms: Stripping away so much substance that only a truism remains, such as “test well” or “plan ahead.” Statements that forbid nothing teach nothing.
abstract uses four classic thinking tools to turn raw facts and cases into portable, high-value rules:
🏛️ The 4 pillars of abstraction #
| Lens | Thinker | How it works in practice |
|---|---|---|
| 1. Essence vs. accident | Aristotle | Split essence, or what makes the pattern hold, from accidents like time, place, path, tool name, or actor that can change without breaking the rule. |
| 2. Inductive ascent | Aristotle & Francis Bacon | Climb from scattered facts to the causal law that explains why those facts happened before stating a general rule. |
| 3. Concept naming | C.S. Peirce | Turn a repeated finding, such as “this step failed because nobody checked the one-way action,” into a crisp noun handle: “the one-way action gate.” |
| 4. Flip test | Karl Popper | Judge a rule by what it forbids. If no sensible person would ever defend the reverse of your statement, it’s a truism, not knowledge. |
📋 Distillation workflow #
- Step 1: Anchor in concrete facts List the specific facts, examples, incidents, quotes, or findings in front of you. Keep them tied to real proof before you interpret them.
- Step 2: Strip accidents from essence For each group of facts, drop the accidents such as local names, IDs, dates, or tool quirks. Keep only the core structure: the actors, motives, limits, and state changes that repeat across cases.
- Step 3: State the causal law Explain the mechanism that links cause to effect. An abstraction without a causal why is just a summary.
- Step 4: Name the concept and write the rule Coin a short, clear handle for the pattern after Peirce. Write the rule in 1 to 2 sentences so it guides choices in any domain.
- Step 5: Run Popper’s flip test and set limits Check what the rule forbids, verify every gate below, and state the exact limits where the rule stops holding true.
✅ Quality gates #
Before sharing abstracted ideas, test every rule against four checks:
- Popper’s flip test: State the reverse of your rule. If nobody would ever defend the reverse, such as “write broken code,” your rule carries zero signal. Rewrite it to expose a real trade-off or non-obvious insight.
- Aristotle’s accident check: A reader who knows nothing about the source project, chat, or domain history must be able to grasp and apply the concept right away.
- Causal grounding: Every rule pairs what holds true with why it works that way, and cites the concrete facts that produced it.
- Clear scope limits: Every general claim names the exact boundary where it fails or yields to another rule.
📄 Output template #
markdown
## 1. Concrete fact base
- **Fact 1**: <Concrete event, example, or finding + source>
- **Fact 2**: <Concrete event, example, or finding + source>
- **Fact 3**: <Concrete event, example, or finding + source>
## 2. Essence vs. accident split
| Pattern | Accidental details (stripped) | Essential structure (kept) |
| :--- | :--- | :--- |
| <Pattern 1> | <Local names, tools, dates, one-off setup> | <Core relationship that repeats across cases> |
## 3. Abstracted concepts and rules
### 1. <Crisp concept name>
- **General rule**: <1–2 sentences stating the portable principle or law>
- **Causal law (why it holds)**: <The mechanism that drives this outcome>
- **What it forbids (Popper flip test)**: <The plausible habit or choice this rule rules out>
- **Grounded in**: <Facts 1..N>
- **Scope limit**: <Where this rule stops holding true>