First Principles Thinking: How to Reason From Scratch
First principles thinking is a way of reasoning that breaks a problem down to its most basic facts, the things you know are true without relying on convention or analogy, and then rebuilds a solution from those facts. It replaces “how is this usually done?” with “what is actually required?”
- Category
- Mental model (problem solving)
- Origin
- Aristotle, 4th century BC
- Best used for
- Problems where the standard answer feels too expensive or too slow
- Related
- Inversion, map is not the territory, five whys
Ask yourself
- Which parts of this do I know are true, and which am I assuming because everyone does it that way?
- If I had to explain this from zero to a smart outsider, what would I start with?
- What is the problem actually made of, in parts, costs and constraints?
- If nobody had ever solved this before, how would I solve it?
- Which constraint here is physical, and which is only a habit?
What is first principles thinking?
First principles thinking means breaking a problem down to the facts you are sure of and building your answer up from there, instead of copying how it is usually done. You stop reasoning from “what others do” and start reasoning from “what must be true.”
Most thinking runs on analogy. You price your service like competitors do, you plan a project like the last one, you study the way you were taught. That is efficient and often correct. The trouble is that analogy also copies the hidden assumptions of the original, including the ones that stopped being true years ago.
First principles thinking strips those assumptions out. It asks three plain questions:
- What is this problem actually made of?
- Which of those parts are facts, and which are conventions?
- If I only kept the facts, what solutions become possible?
The result is often the same answer you would have reached anyway. Sometimes it is a much cheaper, faster or simpler one that nobody tried because “that’s not how it’s done.”
Where does first principles thinking come from?
The idea comes from Aristotle, who described a first principle as the first basis from which a thing is known: a proposition that cannot be deduced from anything more basic. For him, real knowledge meant tracing a claim back to those foundations.
Two thousand years later, René Descartes turned the idea into a method. In his Meditations, he set out to doubt everything that could possibly be doubted and keep only what survived. “I think, therefore I am” was his first principle, the one thing he felt he could not doubt.
Physicists have always worked this way. A physics problem “from first principles” starts with established laws (conservation of energy, for example) rather than with an empirical rule of thumb. That is where the business use of the term comes from.
In the 2010s, Elon Musk made the phrase popular outside science. He has described in interviews how, when battery packs looked too expensive, he asked what a battery is made of and what those materials cost on the commodity market, and found the gap between material cost and price was large. Whatever you think of the person, the example shows the method clearly: price the parts, not the product.
How to use first principles thinking
You use first principles thinking by listing your assumptions, testing each one, and rebuilding from what survives. Here are the steps.
- Write the problem as one sentence. “Our onboarding takes three weeks” is better than “onboarding is broken.” A vague problem can’t be broken into parts.
- List every assumption behind the current solution. Include the boring ones: “it has to be in person,” “a manager has to sign off,” “it has to start on a Monday.”
- Sort each assumption into one of three buckets. Physical or legal fact, measured fact, or convention. Be strict. “Customers expect this” is a convention until you have asked them.
- Ask why on each convention. Keep asking until you hit a fact or a dead end. The five whys technique is a good structure for this.
- Rebuild from the facts only. Ask: given only what must be true, what is the simplest way to reach the outcome? Sketch at least two versions.
- Check the fences before tearing them down. Some conventions exist for reasons you can’t see yet. Before you drop one, apply Chesterton’s fence: find out why it was put there.
- Test the new version small. First principles produce hypotheses, not proof. Run a cheap trial before betting the budget.
A useful shortcut is to price things in their parts. Time, materials, steps, approvals. When the sum of the parts is far below the price of the whole, you have found where convention is costing you.
3 real-world examples
Business
Say a small software company believes it needs a sales team to sell to mid-sized firms, because every competitor has one. Breaking that down: what does a sales team actually do? It answers questions, builds trust, handles procurement paperwork and negotiates price. Which of those are needed for this product? If the product is cheap and easy to try, maybe only the paperwork is. The rebuilt answer might be a self-serve trial, a security document library and one person who handles contracts. Same outcome, a fraction of the cost. It might also fail, which is why step 7 exists.
Personal life
Say you have “always wanted” a bigger house. The analogy is your parents, your friends, the normal path. Break it down: what do you actually want? Maybe a quiet room to work in, space for guests twice a year and a garden. Those are the facts. A bigger house is one solution. A smaller house with a converted shed, plus a nearby rental for guests, is another. First principles thinking doesn’t tell you which to pick. It shows you that you had more than one option.
Tech
A team inherits a nightly batch job that takes six hours and everyone works around it. The analogy says “batch jobs are slow, add servers.” The first principles question is: what does the job need to produce, and how much data actually changes each day? If only a small share of the records change, processing only the changes may cut the run from hours to minutes. Nothing about the old design was a law of nature. It was a choice made when the data was smaller.
Common mistakes
The biggest mistake is treating first principles thinking as permission to ignore experience. Here are the usual ways it goes wrong.
- Calling opinions facts. “People hate paperwork” feels like a first principle. It is an assumption. Real first principles are physical limits, laws, measured data and arithmetic.
- Using it on everything. Rebuilding every decision from zero is exhausting and slow. Most problems deserve a convention. Save the method for the ones where the standard answer clearly fails.
- Skipping the fence check. Conventions often encode lessons learned the hard way, like safety rules and legal requirements. Tearing them down without asking why is not bold, it is careless.
- Stopping at the teardown. Listing assumptions is the easy half. The value is in rebuilding and testing a new version.
- Forgetting that your model is a model. Your breakdown of the problem is itself a simplification. As the map is not the territory reminds you, a clean list of parts can still miss what matters.
First principles thinking vs reasoning by analogy
First principles thinking rebuilds from basic facts; reasoning by analogy copies a solution from a similar case. You need both, and the skill is knowing which one the moment calls for.
| First principles thinking | Reasoning by analogy | |
|---|---|---|
| Starting point | Basic facts and constraints | A similar case that worked |
| Speed | Slow, effortful | Fast, nearly automatic |
| Typical result | Sometimes a new, cheaper or simpler answer | Usually a reliable, familiar answer |
| Main risk | Overlooking hidden reasons behind conventions | Copying outdated assumptions |
| Best for | High-stakes problems where the usual answer fails | Routine problems with good precedents |
| Bias it guards against | Anchoring bias on the usual answer | Overconfidence in untested ideas |
A practical rule: start with analogy. If the analogy gives you an answer you can’t accept (too slow, too expensive, too risky), switch to first principles for that one problem.
The Mentarium take
First principles thinking is overrated as a lifestyle and underrated as a tool. The internet version says “great thinkers reason from first principles about everything.” Nobody does. You would never get out of bed. The real version is narrower and more useful: it is what you do when the normal answer has stopped working.
We place it in the Map step of the Mentarium Decision Loop (Frame → Map → Stress-test → Decide → Review). When you map your options, analogy gives you the options everyone has. A short first principles pass gives you the ones nobody wrote on the table. Then run the result through inversion in the Stress-test step, because a solution rebuilt from scratch has not been beaten up by reality yet.
Use the Decision Weight Score to decide whether it is worth the effort. A DWS of 1 to 3 does not deserve a first principles teardown; copy the convention and move on. At 7 or above, especially when the standard answer is the one you are unhappy with, spend the hour. Write the assumptions down. The list alone usually shows you which part of “how it’s done” you never chose.
Counters Anchoring Bias
FAQ
What is a simple example of first principles thinking?
Say a bike shop quote for a part seems too high. Instead of comparing it with other shops, you ask what the part is made of, what the raw materials cost and how long it takes to make. If the answer is far below the quote, you have found room to build, buy or negotiate differently.
Who came up with first principles thinking?
The idea goes back to Aristotle, who described a first principle as the first basis from which a thing is known. Descartes later built a method on doubting everything that could be doubted. In recent years Elon Musk has popularized the term in business by describing how he used it on batteries and rockets.
What is the difference between first principles and reasoning by analogy?
Reasoning by analogy copies what worked in a similar case: it is fast and usually good enough. First principles thinking ignores the similar case and rebuilds from basic facts. It is slower, so it is worth it only when the analogy is producing a result you can't accept.
Is first principles thinking always better?
No. It is expensive in time and attention, and most everyday problems are solved well by convention. Use it when the stakes are high, the standard answer clearly fails, or the field has changed enough that old analogies no longer hold.
How do you practice first principles thinking?
Take one decision a week and write down every assumption behind it. Mark each one as known, measured or assumed. Then ask why for each assumed one until you reach something you can check. The habit of separating facts from habits is most of the skill.