You’ve now got the two big warnings: a model is a simplification (lesson 1), and it’s never the territory (lesson 2). Which raises an uncomfortable question — if every single model is incomplete and wrong somewhere, how do you ever reason well?
The answer isn’t a better model. It’s more models, working together. The investor Charlie Munger called this a latticework of mental models: a web of ideas from many different fields that you hang facts onto and check against each other. Where one model has a blind spot, another sees clearly. This lesson is about why the collection is the real prize.
Before you read — take a guess
Guess: why would a web of many models beat having one really good model?
The man with a hammer
Munger borrowed a proverb to name the danger of owning just one model:
The law of the instrument
“To the man with only a hammer, every problem looks like a nail.” When you know one model well, you’ll cram every situation into it — even the situations it fits badly. This is also called the law of the instrument.
Picture an economist who explains everything with supply and demand — including why their marriage is struggling. Or a therapist who reads every business failure as a childhood wound. Each has a genuinely powerful model. The problem is they have only the one, so they swing it at problems that need a different tool entirely. The model isn’t wrong; it’s being used where it doesn’t belong.
Notice this is the map–territory trap from last lesson, scaled up. One model maps part of reality. Believe it maps all of reality and you’ve mistaken your one favourite map for the whole territory.
Name the failure mode.
Pick the right option for each blank, then check.
The problem — also called the — is forcing every situation into your .
A nutritionist insists every health problem — including a friend's broken arm — comes down to diet. Which trap is this?
Why the latticework works: blind spots don’t overlap
Here’s the mechanism, stated plainly. Every model has a blind spot — a region of reality it can’t see (that’s lesson 2). But different models are blind in different places. So if you check a situation against several models from different fields, the spots where one is blind tend to be exactly where another sees fine.
Take a single decision — “should our company cut prices to win customers?” — and look at it through several models at once:
| Model | What it reveals | Its blind spot |
|---|---|---|
| Supply & demand (economics) | Lower price should lift quantity sold | Ignores how rivals react |
| Game theory (strategy) | Rivals may cut prices too → a price war | Ignores customer psychology |
| Incentives (psychology) | A “cheap” brand may signal low quality | Ignores the raw economics |
| Feedback loops (systems) | A price war can spiral out of control | Ignores the one-off numbers |
No single row gives the full answer. Read together, they catch each other’s misses — the economics blind to rivalry, the strategy blind to perception, and so on. That is the latticework doing its job.
Why 'from different disciplines' matters
Three models from economics share economics’ blind spots — they’ll all miss the same things. The power comes from reaching across fields: physics, biology, psychology, math. Munger’s whole pitch is that the big ideas from many disciplines, held together, are what let you see a problem whole.
Munger's latticework is strongest when its models come from MANY fields. Sort each set: does it give you wide coverage or shared blind spots?
Place each item in the right group.
- Incentives (psychology) + feedback loops (systems) + supply & demand (economics)
- Natural selection (biology) + game theory (strategy) + compounding (math)
- Three slightly different versions of the same supply-and-demand model
- Supply & demand + opportunity cost + comparative advantage (all economics)
A worked example: diagnosing a struggling team
Watch a latticework beat a hammer on a real problem. A team keeps missing deadlines. The single-model thinker reaches for their one tool and stops. The latticework thinker runs the problem past several:
- Incentives (psychology): Is anyone actually rewarded for hitting deadlines — or does heroically fixing late work get more praise than quietly shipping on time?
- Bottlenecks (systems): Is one overloaded person a chokepoint everything waits on, so the team’s speed is really just their speed?
- Feedback loops (systems): Does missing one deadline create chaos that causes the next miss — a doom loop?
- Opportunity cost (economics): Is the team simply saying yes to too much, so every deadline competes with three others?
Any one of these might be the real cause — and the single-model thinker, swinging their favourite hammer, would have confidently “solved” it with the wrong fix. The latticework doesn’t just give more answers; it stops you from being confidently wrong.
The payoff
Notice you didn’t need to be a psychologist, a systems engineer, and an economist. You just needed one solid model from each, hung on the latticework and ready to reach for. Breadth of models beats depth in any single one — for the generalist decisions most of life is made of.
Which of these are real benefits of running a decision past several models from different fields? (Select all that apply.)
The pitfall: a latticework can be abused
Owning many models has its own failure mode, and it’s worth naming. With a big toolkit, it’s tempting to rummage until you find the model that gives you the answer you’d already decided on — then call it “rigorous thinking.” That’s motivated reasoning in a lab coat.
The fix is the same discipline as before: a latticework is for finding where models disagree, not for cherry-picking the one that agrees with you. If three models point one way and your favourite points the other, the interesting thing is the disagreement — not an excuse to trust your favourite.
Breadth isn't a buffet
A latticework is not a menu you order your preferred conclusion from. Its value is that the models check each other. The moment you’re using extra models to justify a foregone conclusion, you’ve turned a thinking tool into a rationalising tool.
When to reach for the latticework
- On any decision that crosses domains — which is most important ones. Hiring, pricing, strategy, big life choices all involve psychology and economics and systems at once. One model can’t cover them.
- Whenever you notice you’re using your favourite model again. That noticing is the cue to deliberately ask, “what would a different field say about this?”
- Not for trivial, single-domain calls. Choosing lunch doesn’t need game theory. Reserve the full latticework for decisions where being one-sidedly wrong actually costs you.
Match each idea to its meaning.
Pick a term, then click its definition.
Recap
Big picture
The latticework, in one picture
- Latticework of models
- The disease
- Man with a hammer
- Law of the instrument
- Why many beat one
- Blind spots differ by field
- Models check each other
- Reach across disciplines
- Economics + psychology + systems
- Breadth over depth
- The abuse
- Cherry-picking to confirm
- The disease
Quick check — the latticework
What is a "latticework of mental models"?
Check your answer to continue.
You can simplify, you can stay humble about the gap, and you can stack models to cover each other’s blind spots. The last foundational idea ties them together by asking the most honest question of all: which territory do you actually know? That’s the circle of competence.