mental-models（心理模型）

Mental Models (Munger-Style Latticework)

Mental models are simplified representations of how the world works, drawn from multiple disciplines. Charlie Munger's key insight: you need a latticework of mental models from many fields — not just one hammer looking for nails. By viewing a problem through multiple disciplinary lenses simultaneously, you get triangulated wisdom that no single perspective can provide. The goal is to have ~100 models and know when each applies.

Analyze the current topic or problem under discussion through the latticework of mental models. Apply models from at least 6 different disciplines. Cross-reference and triangulate. Apply this framework to whatever the user is currently working on or asking about.

Discipline 1: Physics & Engineering

Apply relevant models:

• Leverage — Where is the point of maximum leverage? Small input, large output?
• Friction — What friction exists in the system? What would reducing friction enable?
• Critical mass — Is there a threshold that, once crossed, triggers self-sustaining change?
• Entropy — Is this system tending toward disorder? What energy is needed to maintain order?
• Feedback loops — Positive (amplifying) or negative (stabilizing) feedback?
• Redundancy & margin of safety — Where are the single points of failure?
• Resonance — Is there a frequency/timing that amplifies the effect?

Which physics/engineering model is most illuminating here, and what does it reveal?

Discipline 2: Biology & Evolution

• Natural selection — What is being selected for? What traits survive in this environment?
• Adaptation vs. extinction — Is the subject adapting fast enough to environmental change?
• Ecosystem dynamics — What is the ecosystem? Who are the predators, prey, symbiotes, parasites?
• Red Queen effect — Do you have to keep running just to stay in place?
• Niche construction — Is the subject changing its environment, not just adapting to it?
• Immune system — What are the defense mechanisms? What gets past them?
• Punctuated equilibrium — Long stability interrupted by sudden change?

Which biological model fits best, and what does it predict?

Discipline 3: Psychology & Behavioral Science

• Incentives — What behaviors are being rewarded? (Munger: "Show me the incentive and I'll show you the outcome.")
• Loss aversion — Are people weighing losses ~2x more than equivalent gains?
• Social proof — Is behavior driven by what others are doing rather than independent analysis?
• Commitment & consistency bias — Are people doubling down because they've already committed?
• Availability heuristic — Are recent/vivid events distorting probability estimates?
• Narrative fallacy — Are we constructing a story that feels right but doesn't match the data?
• Pavlovian association — What conditioned responses are at play?
• Dunning-Kruger — Who is overconfident? Who is underconfident?

Which psychological model explains the most about the human behavior in this situation?

Discipline 4: Economics & Business

• Supply and demand — What shifts the curves? Where is equilibrium moving?
• Comparative advantage — What can be done here that can't be done better elsewhere?
• Opportunity cost — What is being given up? Is it worth it?
• Marginal thinking — What does one more unit cost/produce? (Not average — marginal.)
• Moats — What creates durable competitive advantage? How wide and deep is the moat?
• Principal-agent problem — Whose interests are misaligned? Who's playing with other people's money?
• Tragedy of the commons — Are shared resources being depleted because individual incentives diverge from collective interest?
• Creative destruction — Is something new destroying something old? Is that good or bad?

Which economic model is most relevant, and what does it prescribe?

Discipline 5: Mathematics & Statistics

• Power laws vs. normal distributions — Is this a domain of averages or extremes?
• Compounding — What is growing exponentially that looks flat now?
• Regression to the mean — Is current performance sustainable, or will it revert?
• Bayes' theorem — What should the prior be? How much should this evidence update it?
• Order of magnitude — Are we arguing about 10% differences when there's a 10x factor we're ignoring?
• Combinatorics — How many possible configurations exist? Are we exploring enough of the space?
• Asymmetry of outcomes — Is the upside/downside symmetric, or heavily skewed?

Which mathematical model provides the clearest insight?

Discipline 6: History & Philosophy

• Chesterton's Fence — Before removing something, understand why it was put there.
• Lindy effect — Things that have survived a long time are likely to survive longer.
• Ozymandias effect — All empires fall. What's the half-life of this?
• Hegelian dialectic — What is the thesis, antithesis, and potential synthesis?
• Via negativa — What should be *removed* rather than added?
• Skin in the game — Who bears the consequences of the decision? (Taleb)
• Historical analogy — What historical parallel illuminates this situation?

Which historical/philosophical model offers the deepest wisdom?

Cross-Disciplinary Triangulation

Now synthesize across all six lenses:

• Convergence: Where do multiple models from different disciplines agree? (High confidence)
• Divergence: Where do they disagree? (Needs deeper investigation)
• Blind spots: Which models are missing from the analysis? What discipline hasn't been consulted?
• Dominant model: Which single model provides the most explanatory power for this specific problem?
• Ensemble insight: What understanding emerges from combining the models that no single model alone could produce?

Actionable Output

• What is the recommended course of action based on the multi-model analysis?
• What are the key risks identified by the models?
• What mental model should the decision-maker keep at the forefront?
• What would Charlie Munger probably say about this situation?

"To the man with only a hammer, every problem looks like a nail." The cure is a toolbox full of mental models from every discipline. Use them all, trust no single one, and pay attention when they disagree.