Abstraction Level

Input: $ARGUMENTS

Interpretations

Before executing, identify which interpretation matches the user’s input:

Interpretation 1 — Altitude adjustment: You’re thinking at the wrong level of abstraction — too zoomed in on details or too zoomed out in generalities — and need to find the level where things click (e.g., “I’m drowning in details” or “this is too vague to act on”). Interpretation 2 — Framing problem: You’re at a reasonable level of abstraction but looking at the situation through the wrong lens entirely — you need a different angle, not a different altitude (e.g., “I’ve tried zooming in and out and neither helps”). Interpretation 3 — Audience mismatch: You understand something fine but need to pitch it at a different abstraction level for someone else — matching altitude to your audience (e.g., “how do I explain this to my boss / a beginner / an expert?”).

If ambiguous, ask: “I can help with adjusting your thinking altitude, reframing from a different angle, or matching abstraction level to an audience — which fits?” If clear from context, proceed with the matching interpretation.

GLOSSARY

STEP 0: What type of situation are you in?

Answer this question: “What am I trying to do RIGHT NOW?”

SECTION A: Understanding Something

Goal: Find the altitude where things start making sense.

Step A1: State what you are trying to understand. Write one sentence: “I am trying to understand ________________.”

What you should see: A clear sentence. If you cannot write this sentence, you are not ready to choose an altitude — you need to first figure out what your question is. Write down three different versions of your question and pick the one that feels most honest.

Step A2: Notice your current pain. Read each description below. Pick the one that matches your experience right now:

• (a) “I have too many details and I can’t see how they connect.” —> Go to Step A3.
• (b) “I have a general sense but nothing concrete or testable.” —> Go to Step A4.
• (c) “I can describe it, but something feels off or incomplete.” —> Go to Step A5.
• (d) “I don’t even know where to start.” —> Go to Step A6.
• (e) “I feel like I understand it fine.” —> You may not need to change altitude. Test by explaining it to an imaginary 12-year-old. If you can, STOP — you are at a good altitude. If you cannot, go to Step A3.

Step A3: Zoom OUT. You are drowning in details. Do the following:

1. List the 3-5 most important details you have.
2. For each detail, ask: “What is this an example OF?” Write the answer. (Example: “Revenue dropped 12%” is an example of “declining financial performance.”)
3. Look at your answers from (2). Do they share a common theme? If yes, write one sentence that captures that theme. This is your new altitude.
4. Check: Does this new altitude help you see connections between your original details? If yes, stay here. If no, repeat Step A3 one more time (zoom out again). If zooming out twice does not help, go to Step A5 — you may have a framing problem, not an altitude problem.

What you should see: A sentence at a higher level of generality that makes your details feel organized rather than scattered.

Step A4: Zoom IN. You are floating in generalities. Do the following:

1. Write down your general understanding in one sentence.
2. Ask: “What would I expect to SEE if this were true?” Write 3 specific, observable predictions.
3. For each prediction, ask: “Can I check this right now?” If yes, check it. If no, note what you would need to check it.
4. Look at your predictions. Do they hold up? If yes, your abstract understanding is correct — you just needed grounding. If some fail, your abstract understanding needs revision — adjust it and repeat.

What you should see: Specific, observable predictions that connect your general idea to concrete reality.

Step A5: Reframe (stay at current level, change angle). You may not need to change altitude — you may need a different FRAME at the same altitude.

1. Write down your current explanation in one sentence.
2. Ask: “What if the opposite were true?” Write what that would look like.
3. Ask: “What would someone who disagrees with me say?” Write their version.
4. Ask: “What am I NOT looking at?” Identify what you have been ignoring.
5. If any of these questions produce an “aha” moment, adopt that new frame. If none do, try zooming out (Step A3) — you may need more context.

Step A6: Start from scratch with broad orientation. You are lost. Do the following:

1. Ask: “What KIND of thing is this?” (Is it a technical problem? A people problem? A resource problem? A timing problem? A knowledge problem?)
2. Ask: “Who else has dealt with this kind of thing?” Think of one example.
3. Ask: “What happened in that example?” Note the key events.
4. Ask: “Does my situation resemble that example?” If yes, use that as your starting framework and go to Step A4 to zoom in on your specifics. If no, try a different example.

What you should see: A named category for your problem and at least one analogous example.

SECTION A COMPLETE. You should now have an altitude where things make sense. Write down: “I understand this at the level of: [your sentence].” If you need to decide something, go to SECTION B. If you need to do something, go to SECTION C.

SECTION B: Deciding Something

Goal: Find the altitude where the decision becomes clear.

Step B1: Name the decision. Write: “I need to decide ________________.”

Step B2: Identify the decision’s natural level. Every decision lives at a natural altitude. Ask these three questions:

1. “Who is affected by this decision?” If the answer is just you —> the decision is naturally concrete. If the answer includes others, organizations, or systems —> the decision is naturally more abstract.
2. “How long will this decision matter?” If days or weeks —> concrete. If months or years —> more abstract.
3. “What breaks if I get this wrong?” If a specific task fails —> concrete. If a strategy or relationship fails —> abstract.

Write down where the decision naturally lives: CONCRETE / MIDDLE / ABSTRACT.

Step B3: Check one level UP from the decision. Ask: “What larger goal does this decision serve?” Write the answer. Then ask: “Does knowing this larger goal change which option I should pick?”

• If YES: You were about to decide at too low a level. Restate the decision in terms of the larger goal and proceed.
• If NO: Good. The decision is at its natural level.

Step B4: Check one level DOWN from the decision. Ask: “What specific, concrete thing will happen first if I choose Option A vs. Option B?” Write the immediate concrete consequence of each option. Then ask: “Does seeing these concrete consequences change my preference?”

• If YES: You were about to decide at too high a level. Restate the decision in terms of these concrete consequences and proceed.
• If NO: Good. The decision is at its natural level.

Step B5: Decide at the validated level. You now have a decision stated at its natural level, checked one level up and one level down. Make the decision at this level.

SPECIAL CASE — Irreversible decisions: If this decision is hard or impossible to undo, add one more step before deciding. Ask: “What would I see in 1 year / 5 years if this decision turns out to be wrong?” If you cannot answer this question, you are at too abstract a level to decide — zoom in until you can describe what failure looks like concretely.

SECTION B COMPLETE.

SECTION C: Doing / Executing Something

Goal: Get concrete enough to take the next physical action.

Step C1: State what you need to do. Write: “I need to ________________.”

Step C2: Can you do it right now, in the next 5 minutes?

• If YES: Do it. You are at the right level. STOP.
• If NO: Go to Step C3.

Step C3: Why can’t you do it right now? Pick the reason that fits:

• (a) “It’s too big / I don’t know where to start.” —> Go to Step C4.
• (b) “I don’t know HOW to do it.” —> Go to Step C5.
• (c) “I’m blocked by something outside my control.” —> Go to Step C6.
• (d) “I’m not sure this is the RIGHT thing to do.” —> Go back to SECTION B (you have a decision to make, not an action to take).

Step C4: Break it down (zoom IN).

1. Ask: “What is the FIRST thing that needs to happen?” Write it.
2. Ask: “Can I do THAT in the next 5 minutes?” If yes, do it. If no, repeat: “What is the first thing that needs to happen for THAT?” Keep breaking down until you have a 5-minute action.

What you should see: A single concrete action you can take in the next 5 minutes.

Step C5: Zoom OUT to learn, then zoom IN to do. You need knowledge before you can act.

1. Ask: “What do I need to know to do this?” Write the specific question.
2. Ask: “Where can I find that answer?” (person, document, experiment, search)
3. Go get that answer. This is your next concrete action.
4. Once you have the answer, return to Step C2.

Step C6: Identify the blockage and zoom to the right level.

1. Write down exactly what is blocking you.
2. Ask: “Is this blockage at a higher level than my task?” (Example: you need permission from management.) If yes, zoom out: restate your need at the level where the blocker lives and address it there.
3. Ask: “Is this blockage at a lower level than my task?” (Example: a specific tool isn’t working.) If yes, zoom in: fix the specific blocker, then resume.

SECTION C COMPLETE.

SECTION D: Explaining / Communicating Something

Goal: Match the abstraction level to your audience.

Step D1: Identify your audience. Write: “I am explaining this to ________________.”

Step D2: Assess your audience’s current altitude. Ask: “How much does my audience already know about this topic?”

• (a) They know nothing. —> Start ABSTRACT. Give them the big picture first, then offer concrete examples as illustrations. Use the structure: “[General principle]. For example, [concrete case].”
• (b) They know the basics. —> Start at MIDDLE altitude. Connect your point to what they already know, then zoom in to the new specific information.
• (c) They are experts. —> Start CONCRETE. Give them the specific new information. They already have the abstract framework.

Step D3: Verify altitude match. After drafting your explanation (even mentally), ask: “If I were my audience, would I know what to DO with this information?”

• If YES: Your altitude is right.
• If NO, and the problem is “too vague”: Zoom in. Add specifics.
• If NO, and the problem is “too detailed / lost”: Zoom out. Add framing.

Step D4: Add altitude bridges. Good communication includes BRIDGES between abstraction levels. For each key point, include:

• The abstract version (why it matters)
• The concrete version (what it looks like in practice)
• A bridge sentence connecting them (“This matters because…” or “In practice, this means…”)

SECTION D COMPLETE.

SECTION E: Creating Something New

Goal: Oscillate between abstract (structure) and concrete (content) deliberately.

Step E1: Start abstract — define structure. Before creating content, create the SKELETON:

1. What is the purpose of this thing? (One sentence.)
2. What are its major parts? (3-7 sections, categories, or components.)
3. How do the parts relate to each other? (Sequence? Hierarchy? Network?)

What you should see: An outline, blueprint, or structural diagram — NOT content yet.

Step E2: Go concrete — fill one section. Pick the section you understand best. Fill it with concrete content. Do not worry about other sections yet.

What you should see: One section with real, specific content.

Step E3: Zoom out — check coherence. After filling one section, zoom out:

1. Does this section still fit the structure you defined in Step E1?
2. Does it suggest changes to the overall structure? If yes, revise the structure. If no, proceed.

Step E4: Repeat the oscillation. Continue: fill a section (concrete), check against structure (abstract), adjust structure if needed, fill next section. Do this until all sections are filled.

Step E5: Final zoom-out pass. Read the entire thing at the ABSTRACT level: Does the overall structure accomplish the purpose from Step E1? Then read at the CONCRETE level: Is every section specific and complete?

SECTION E COMPLETE.

QUICK REFERENCE CARDS

Card 1: The One-Minute Version

Feeling overwhelmed by detail?     --> ZOOM OUT. Ask "what is this an example of?"
Stuck in vague generalities?       --> ZOOM IN. Ask "what would I expect to see?"
Tried both and neither helps?      --> REFRAME. Ask "what am I not looking at?"
In a crisis?                       --> SKIP THIS. Act on instinct. Review later.

Card 2: Altitude by Goal Type

Understanding  --> Start abstract, zoom in to test
Deciding       --> Find the natural level, check one up and one down
Executing      --> Go concrete immediately
Communicating  --> Match your audience
Creating       --> Oscillate deliberately between structure and content

Card 3: The Four Zoom-Out Axes

When someone says "zoom out," there are FOUR ways to do it:
1. TIME:        Think about a longer timeframe
2. SCOPE:       Think about more people/things affected
3. SYSTEM:      Think about higher-order effects
4. PERSPECTIVE: Think from a different stakeholder's view
Try all four. Usually only one of them helps for your specific situation.

Card 4: Hard Stop Rules

STOP zooming out after 3 times.  If it hasn't helped, the problem is framing.
STOP zooming in after 5 times.   If it hasn't helped, you're on the wrong problem.
STOP oscillating if you lose the thread.  Pick ONE altitude, stabilize, then resume.

COMMON MISTAKES

1. Treating abstraction as inherently superior to concreteness. Many people believe that “big-picture thinking” is smarter than “getting into the weeds.” This is wrong. The smartest thinkers match their altitude to their goal. A surgeon who “thinks big picture” during an operation will kill the patient.

2. Changing altitude when you should change frame. If you have zoomed out AND zoomed in and neither helped, you almost certainly do not have an altitude problem. You have a framing problem — you are looking at the situation through the wrong lens. Try Step A5.

3. Zooming out as procrastination. Asking “but what’s the REAL question here?” can be a genuine reframe or it can be avoidance of doing the concrete work. Test: after zooming out, do you know what to do next? If yes, it was genuine. If no, and you want to zoom out again, you are probably procrastinating. Apply the hard stop rule.

4. Zooming in as anxiety management. Focusing on tiny details can feel productive when the big picture is scary. Test: is the detail you are focused on the BOTTLENECK? If no, you are probably hiding in details to avoid confronting a harder question.

5. Assuming your audience shares your altitude. The most common communication failure. You are explaining at the altitude where YOU understand it, but your audience lives at a different altitude. Always check: “Does my audience have the context to understand this at the level I’m presenting it?”

6. Only zooming on the “generality” axis. Forgetting that you can also zoom out in time, scope, system level, or perspective. If zooming in/out on the generality axis is not helping, try one of the other axes.

7. Thinking you need to pick ONE level and stay there. For complex problems, the answer is often to oscillate deliberately. The mistake is oscillating UNCONSCIOUSLY (which produces incoherent thinking). The fix is to oscillate DELIBERATELY (which produces rich, multi-level understanding).

8. Using jargon as a substitute for genuine abstraction. Replacing specific words with general words is not the same as zooming out. “We need to leverage our core competencies to drive synergies” is not abstract thinking — it is vague thinking. Genuine abstraction identifies PATTERNS. Test: can you give 3 different concrete examples of your abstract claim? If no, it is jargon, not abstraction.

WHEN TO OVERRIDE THIS PROCEDURE

Override completely when:

• You are in a genuine emergency requiring immediate action (seconds matter)
• You are an acknowledged expert in the specific domain AND you have strong intuition about the right altitude AND the stakes of getting the altitude wrong are low
• You are in a creative flow state and the procedure would break your momentum

Override partially (skip to the relevant section) when:

• You already know your goal type (skip Step 0)
• You have already identified your pain signal (skip the diagnostic steps)
• You are iterating on something you have done before (start at the section that matches your current stuck point)

Seek expert help when:

• You have followed this procedure and STILL cannot find an altitude that works — you may be dealing with a “wicked problem” that resists all framing
• The decision is irreversible AND high-stakes AND you are not an expert — this procedure helps but is not a substitute for domain expertise
• You find yourself overriding the procedure repeatedly for the same type of decision — you may need a domain-specific procedure instead of this general one

WORKED EXAMPLES

Example 1: Software Developer Choosing How to Fix a Bug

Situation: A developer gets a bug report: “The checkout page sometimes shows the wrong total.”

Step 0: Goal = Execute (fix the bug). Go to SECTION C.

Step C1: “I need to fix the checkout total calculation bug.”

Step C2: Can I do this in 5 minutes? No — I don’t know what’s causing it yet.

Step C3: Why not? (b) “I don’t know HOW to do it” — I don’t know the cause. Go to Step C5.

Step C5:

1. “What do I need to know?” —> What specific conditions cause the wrong total.
2. “Where can I find that answer?” —> Reproduce the bug, check the logs, read the calculation code.
3. Concrete next action: Open the bug report and identify the reproduction steps.

The developer reads the reproduction steps. The bug only happens when a discount code is applied AND the cart has more than 3 items.

Now back to Step C2: Can I fix it in 5 minutes? Still no — I know the conditions but not the cause. The developer looks at the discount calculation code and finds: the discount is applied per-item but the rounding happens after the total, creating a floating-point discrepancy.

Back to Step C2: Can I fix it in 5 minutes? YES — change the rounding to happen per-item before totaling. Done.

Altitude trace: The developer stayed CONCRETE throughout, which is correct for execution. They briefly zoomed out (from “what’s wrong” to “what conditions trigger it”) but only enough to identify the next concrete action. They did NOT zoom out to “what’s our overall code quality strategy?” because that would have been wrong-altitude procrastination for this goal.

Example 2: Manager Deciding Whether to Reorganize Their Team

Situation: A manager notices their team is slow to deliver. They are thinking about reorganizing into sub-teams.

Step 0: Goal = Decide. Go to SECTION B.

Step B1: “I need to decide whether to reorganize my team into sub-teams.”

Step B2: Natural level of the decision:

1. Who is affected? The whole team (8 people) plus stakeholders. —> More abstract.
2. How long will it matter? Months to years. —> Abstract.
3. What breaks if wrong? Team productivity and morale. —> Abstract. Result: This decision naturally lives at an ABSTRACT level.

Step B3: Check one level UP. “What larger goal does this serve?” —> “I want the team to deliver faster and with less friction.” “Does knowing this change which option I should pick?” —> YES. Reorganizing is one solution to “deliver faster.” But there might be others (better processes, fewer projects, different tooling). The manager was about to decide on a SOLUTION without confirming the PROBLEM.

Restatement: “I need to decide what is causing slow delivery and what to do about it.” This is now at the right level. Reorganization is one option, not the framing.

Step B4: Check one level DOWN. “What happens concretely if I reorganize vs. not?” —> Reorganizing: 2-3 weeks of disruption, new communication patterns, possible morale issues. Not reorganizing: continued slow delivery, growing frustration. “Does this change my preference?” —> Somewhat — the reorganization cost is real. The manager now wants to check if there’s a lower-cost intervention first.

Result: The manager zooms in to identify the specific bottleneck (it turns out to be too many concurrent projects, not team structure) and addresses that instead.

Altitude trace: The initial framing was at the SOLUTION level (“should I reorganize?”). Zooming out one level revealed it was really about CAUSES of slow delivery. Zooming in one level revealed the costs of the proposed solution. The right altitude turned out to be the PROBLEM level, not the SOLUTION level.

Example 3: Student Writing a Research Paper

Situation: A student needs to write a paper about climate change policy.

Step 0: Goal = Create. Go to SECTION E.

Step E1: Start abstract.

1. Purpose: “Argue that carbon pricing is more effective than regulations for reducing emissions.”
2. Major parts: (1) The problem, (2) Regulatory approach, (3) Carbon pricing approach, (4) Comparison, (5) Conclusion.
3. Relationship: Sequential argument building.

Step E2: Fill one section concretely. The student picks Section 3 (carbon pricing) because they know it best. They write specific details: types of carbon pricing (tax vs. cap-and-trade), real examples (EU ETS, British Columbia carbon tax), specific results (emissions reductions of X%).

Step E3: Zoom out — check coherence. After writing Section 3 in detail, the student realizes their evidence is stronger for carbon taxes than cap-and-trade. This changes the structure: the thesis should be narrower (“carbon TAXES are more effective” not “carbon PRICING is more effective”).

Revised structure from Step E1: Adjust thesis and update Section 3 to focus on carbon taxes specifically.

Step E4: Continue oscillating. Fill Section 2 (regulatory approach) with concrete examples. Zoom out: does the comparison still work? Yes. Fill Section 4 (comparison). Zoom out: is the argument coherent? Yes.

Step E5: Final passes.

• Abstract pass: The argument flows logically and supports the thesis. PASS.
• Concrete pass: Section 1 (the problem) is still too vague — needs specific emissions data. FIX: add data.

Altitude trace: The student oscillated between structure (abstract) and content (concrete) throughout. The key moment was Step E3, where concrete work at one level (writing about carbon pricing) forced a revision at the abstract level (narrowing the thesis). This is the oscillation pattern working correctly.