Skip to content
Published on

Thinking Tools for Reading the World — Mental Models

🇰🇷KO🇺🇸EN🇯🇵JA
Split View필사 모드
Share
Authors

Opening — Holding a Map in Your Hand

We make decisions every day.

We choose what to build, where to spend our time, and whose word to trust.

Yet the world is far too complex to hold in our heads all at once.

So we draw a shrunken picture of it in our minds.

We call this shrunken picture a mental model.

A good map does not draw every alley.

Instead it keeps only what you need to find your way.

A mental model works the same way.

It strips away the detail of reality and keeps only the frame you can use to judge.

This essay is about those tools.

First we define what a mental model is.

Then we look at what Charlie Munger called the latticework of models.

We compare first-principles thinking with reasoning by analogy.

We tour high-leverage models one by one, such as opportunity cost, inversion, and second- and third-order effects.

We examine the danger of overusing a single model, the man-with-a-hammer tendency.

We look at biases the models guard against, like confirmation bias and survivorship bias.

Finally we close with the limits every model shares, and how to build your own latticework.

This is not a grand theory book.

It is a practical guide to making better judgments in work and in life.

1. What Is a Mental Model

A mental model is a simplified representation of how the world works.

It is a thinking frame we reach for when we reason about something.

Let us take an example.

The sentence "when demand rises, prices go up" is a model.

That sentence does not contain every variable of reality.

It leaves out taxes, psychology, and government regulation.

Even so, this simple frame often helps us read a market.

The heart of a mental model lies in exactly this simplicity.

The world is complex, and our attention is finite.

If we tried to compute everything, we would never decide anything.

So we strip things down.

We keep the important relationships and set the rest aside for now.

The frame we keep lets us make good decisions quickly.

Models Are Already Inside Us

We already use countless models.

We simply do not notice that we do.

"A hot object burns you if you touch it" is a model.

"Someone who keeps their promises is trustworthy" is a model.

The trouble is that these models are usually unconscious.

An unconscious model rules us without ever being tested.

Studying mental models means bringing these frames out into the open.

It means naming them and examining when they hold and when they fail.

A conscious model becomes a tool.

An unconscious model often becomes a trap.

What Makes a Model Good

Good models share a few traits.

First, they are simple.

A frame with too many conditions is hard to use.

Second, they apply across many situations.

A frame you can use only once is barely worth learning.

Third, they have predictive power.

With the model, you can see a little more clearly what will happen next.

Fourth, they know their own limits.

A good model also tells you when it breaks down.

2. Charlie Munger's Latticework

Charlie Munger was Warren Buffett's longtime partner.

He was an investor, but he did not only talk about investing.

What he stressed was a latticework of mental models.

A lattice is a structure of many rods woven crosswise.

Munger's idea runs like this.

A few models from a single discipline cannot let you read the world properly.

Physics, biology, economics, psychology, statistics.

You must draw one core idea from each of these fields.

And you must weave them together like a lattice.

Why Many Disciplines

One field lights up only one face of the world.

Economics explains incentives well.

Psychology explains human irrationality well.

Biology explains competition and adaptation well.

Any one of them alone gives you only half the picture.

Overlay several lenses, and depth finally appears.

Munger put it this way.

You must hold the big ideas of the world in your head and use them together as a habit.

The Power of a Lattice

The power of the latticework lies in cross-checking.

One model reaches a conclusion.

If a model from another field points to the same conclusion, your confidence grows.

If the models disagree, that is a warning sign.

It means you are missing something.

Running several models at once reduces the risk of leaning too far one way.

This is why Munger stressed lifelong reading.

A lattice is not woven overnight.

You add one rod at a time by steadily reading across many fields.

3. First Principles vs. Reasoning by Analogy

The way we think divides roughly into two.

One is reasoning by analogy.

The other is first-principles thinking.

Reasoning by Analogy

Analogy means thinking by comparison to what already exists.

"A competitor does it this way, so let us do it this way too."

"This method worked last year, so let us do the same this year."

Analogy is fast and comfortable.

For most everyday judgments, analogy is enough.

But analogy has a limit.

If the thing you compare to is wrong, your conclusion is wrong with it.

Copying others makes it hard to get ahead of others.

Because you only move within an existing frame.

First-Principles Thinking

First-principles thinking is different.

It breaks a problem down to basic truths that cannot be divided further.

Then it builds back up from those basics.

It strips away the reason "because everyone does it that way."

Instead it asks from the ground up: "is that really so?"

A common example is the cost of batteries.

In the past, expensive batteries were taken as a given.

A first-principles approach changes the question.

What raw materials make up a battery, and what is the market price of those materials?

Go down to the floor like this, and a marginal cost far below conventional wisdom sometimes appears.

When to Use Which

The two are not a matter of better or worse.

They simply play different roles.

For everyday, low-risk decisions, analogy is efficient.

If you rethought everything from scratch each time, life would grind to a halt.

But when you want to change the game, you need first principles.

When conventional wisdom seems doubtful, when a big decision is at hand.

At those moments we must strip away inherited assumptions and go down to basics.

4. A Tour of High-Leverage Models

Now let us look at a few models worth actually using.

These are high-leverage models: useful out of proportion to the effort of learning them.

Opportunity Cost

When you choose one thing, you give up another.

Opportunity cost is the value of the alternative you gave up.

This is not only about money.

Time, attention, and energy count too.

If you spend two hours in this meeting, you gave up whatever else those two hours could have done.

Being aware of opportunity cost dissolves the illusion of "free."

Truly free things are rare in this world.

There is always an unseen alternative paying the price.

Inversion

Inversion means solving a problem backwards.

Instead of asking "how do I succeed," it asks this.

"How would I guarantee failure?"

Then it avoids each path to that failure one by one.

Munger prized this method in particular.

"I want to know where I am going to die, so I never go there."

Designing a happy life is hard.

But the list of actions that summon a miserable one is fairly clear.

Simply avoiding that list improves a great deal.

Inversion shines especially when the front door is blocked.

Second- and Third-Order Effects

Every action has an immediate result.

This is the first-order effect.

But that result produces another result.

These are the second- and third-order effects.

Shallow thinking stops at the first-order effect.

Deep thinking looks at what comes after.

Let us take an example.

A city introduces a cap on rents.

As a first-order effect, tenants carry a lighter burden.

But as a second-order effect, landlords may reduce supply.

As a third-order effect, over the long run housing may grow even scarcer.

Good intentions do not always produce good outcomes.

The habit of asking one more time "and then what happens" matters.

The Map Is Not the Territory

A map is a representation of the territory, not the territory itself.

This line comes from the scholar Alfred Korzybski.

Every model is a shrunken version of reality.

The shrunken version is convenient, but it differs from the original.

We often mistake the map for the territory.

We treat a plan as reality, and numbers as truth.

The moment we believe a metric captures reality perfectly, danger begins.

Good thinking always remembers the gap between map and territory.

It uses models without being trapped inside them.

Occam's Razor

Suppose several explanations account for the same phenomenon equally well.

In that case, choose the explanation with the fewest assumptions.

This is Occam's razor.

It does not mean the simplest explanation is necessarily correct.

It is a principle against piling on needless assumptions.

Complexity the evidence does not demand is usually clutter.

It is useful in daily life too.

When a colleague does not reply, consider this before imagining a conspiracy.

They may simply be busy.

Margin of Safety

When engineers design a bridge, they do not build it to fit the expected load exactly.

They leave room to bear loads far heavier than expected.

That room is the margin of safety.

The world does not unfold as predicted.

There is always error in the calculation, and shocks arrive out of nowhere.

So wise decisions need slack.

In investing it shows up as buying well below value.

In scheduling it shows up as not planning to the minute.

Slack looks like waste, but it is really insurance against uncertainty.

5. The Man-with-a-Hammer Tendency

Here is a famous saying.

"To the man with a hammer, everything looks like a nail."

It is the trap of leaning too hard on a single model.

Suppose a model works well a few times.

Then we want to apply that model everywhere.

We force it even onto situations where it does not fit.

Why This Happens

The familiar is comfortable.

Learning a new model takes effort.

So we try to solve every problem with the one tool in hand.

The more expert someone is, the easier they fall into this trap.

Someone who has dug deep into one field is at home with that field's lens.

They grow inclined to view everything through it.

An economist tries to explain everything with incentives.

An engineer sees everything as an optimization problem.

Each person's hammer manufactures its own nails.

The Cure Is Many Tools

The cure returns to the latticework.

We overuse a tool because it is the only one we have.

With several models in the toolbox, we can pick the one that fits the situation.

This is the practical reason Munger stressed many disciplines.

Ask yourself which hammer you are habitually holding right now.

Then look at the same problem again with a different tool.

6. Biases the Models Guard Against

The human brain loves shortcuts.

These shortcuts are usually useful, but they sometimes deceive us.

Good mental models act as shields against such biases.

Confirmation Bias

We tend to gather only the evidence that supports what we want to believe.

We quietly ignore the evidence that contradicts it.

This is confirmation bias.

It is as if we set the answer first and then go hunting for reasons.

Inversion is a fine antidote here.

Just ask first: "if I am wrong, what would the reason be?"

Deliberately keeping someone around who argues against you also helps.

Survivorship Bias

We tend to judge by looking only at what survived.

The habits of successful founders become books.

The countless people who had the same habits and failed go unrecorded.

Leaving out the unseen failures distorts the conclusion.

This is survivorship bias.

There is a famous story from the Second World War.

Looking at the bullet holes in returning aircraft, planners wanted to reinforce those areas.

A statistician pointed out something.

The planes that did not return had likely been hit elsewhere.

The places that truly needed reinforcing were the areas with no bullet holes.

Do not look only at the visible data; ask about the data that is missing.

Other Biases

There are many more biases.

Anchoring is when your judgment is dragged toward the first number you saw.

The availability bias is overrating the examples that come to mind easily.

The sunk cost fallacy is continuing down a wrong path because the money already spent feels wasted.

Simply knowing the names of these biases is half the battle.

Because once it has a name, you can catch it in the moment it operates.

7. The Limits of Every Model

The statistician George Box said this.

"All models are wrong, but some are useful."

That sentence holds humility and practicality at once.

What "Wrong" Means

A model is by definition a simplification.

A simplification necessarily leaves something out.

Because something is left out, no model can be entirely correct.

This is not a flaw but the nature of the thing.

A model that captured the world exactly would be as complex as the world.

Such a model would not be a map but merely another territory.

What "Useful" Means

Being wrong does not make a thing useless.

Newtonian physics is, strictly speaking, wrong.

Yet it remains useful enough to build bridges and launch rockets.

What matters is not right or wrong, but whether it serves the purpose at hand.

Do not worship a model as truth.

Treat it as a tool fit for a purpose.

You can swap tools as the situation changes.

The Attitude of Knowing Limits

The most dangerous person is one who does not know the limits of their model.

Knowing when a model breaks down is the real skill.

On what assumptions does this model stand?

What happens when those assumptions break?

Keep these questions always at hand.

8. Building Your Own Latticework

A mental model is not stacked up overnight.

You weave it a little at a time over a lifetime.

Here are a few practices.

Read Across Many Fields

Digging deep into one field is fine too.

But if you want a lattice, you also need breadth.

Deliberately pick and read books far from your specialty.

If you are a physicist, try psychology; if a developer, try history.

One core idea from an unfamiliar field becomes a new rod.

Give Models Names

When you meet a good idea, name it and write it down.

With a name, it is easy to reach for again.

Opportunity cost, inversion, margin of safety.

Such labels tidy the toolbox in your mind.

Review Your Failures

Wrong decisions are the finest textbook.

Trace back what went wrong and which model you missed.

Next time you will be able to call that model up first.

Experience without review does not accumulate; it just drains away.

Slowly, but Steadily

Do not rush to finish the lattice.

Properly learning just a few good models a year is enough.

In ten years a substantial lattice takes shape.

What matters is not speed but direction and consistency.

In One Picture

Below is a sketch of several models converging on a single decision.

     [ Opportunity Cost ]      [ Inversion ]
              \                    /
               \                  /
   [ 2nd/3rd-Order ]---+---[ Margin of Safety ]
                       |
                ( One Decision )
                       |
   [ Occam's Razor ]---+---[ Map != Territory ]
               /                  \
              /                    \
   [ Check Confirmation ]   [ Check Survivorship ]

Do not look through only one lens.

When you overlay several lenses, the decision grows sturdier.

Closing — For the Sake of Seeing More Clearly

Mental models are not magic.

They are not machines that judge the world for you.

They are simply lenses that help you see the world a little more clearly.

The more lenses, the better.

Because a single lens always has a blind spot.

Opportunity cost reminds you of the unseen price.

Inversion opens a back door on a blocked problem.

Second- and third-order effects hold back a hasty conclusion.

Margin of safety prepares you for shocks you did not predict.

And above all these lenses, you need the attitude of humility.

The humility never to forget that all models are wrong.

Hold the tools, but do not blindly trust them.

Hold the map in your hand, but remember the map is not the territory.

In that way, widen your own lattice little by little.

Better decisions come from better thinking.

And better thinking, in the end, comes from steady practice.

Questions to Ponder

  1. Think of one big decision you made recently. Which mental model were you using unconsciously at the time? Was that model appropriate?

  2. What is the "hammer" you habitually hold? Reflect on whether you are swinging that one tool at every problem.

  3. Confirmation bias or survivorship bias may have been at work in some recent judgment. What evidence did you ignore, and which failure cases did you leave out?

  4. Is there a model you would like to add to your lattice over the next year? From which field's books could you gain it?

References

  • Peter Bevelin, "Seeking Wisdom: From Darwin to Munger" — a broad synthesis of Munger's thinking and the psychology of human misjudgment.
  • "Poor Charlie's Almanack: The Wit and Wisdom of Charles T. Munger" (Charlie Munger) — the primary source for Munger's talks and the latticework-of-models idea.
  • Daniel Kahneman, "Thinking, Fast and Slow" — the landmark work on cognitive biases such as confirmation bias and anchoring.
  • Farnam Street, "Mental Models: The Best Way to Make Intelligent Decisions" — https://fs.blog/mental-models/
  • George E. P. Box, "Science and Statistics", Journal of the American Statistical Association (1976) — the source of "all models are wrong, but some are useful."
  • Wikipedia, "Survivorship bias" — https://en.wikipedia.org/wiki/Survivorship_bias
Discuss on Twitter • View on GitHub