Tool/Mental Model/No. 0386
First Principles Thinking
First principles thinking is a method of reasoning that separates facts and constraints from inherited assumptions, then builds solutions from explicit premises. Rooted in philosophy and studied by Aristotle, it is used in mathematics, science and engineering to examine what a solution must satisfy.
- Evidence
- Useful, modest evidence
- Read
- 6 min
- Links
- 9 connections
01You've seen this when…
- in life
Your bedroom stays cold, so you shop for a bigger heater. Before buying, you check where heat escapes and discover a steady draft around the window.
- at work
A new employee needs eleven approvals before getting access to basic tools. You list what each approval protects against and find that several exist only because the old process included them.
- out in the world
A city considers adding another downtown parking garage. Planners start instead with how many people need to reach downtown, at what times, and from where.
02The idea
The first proposal often inherits the shape of the existing solution: a bigger heater, a faster approval process, another garage. First principles thinking pauses before accepting that shape.
It separates what the problem requires from how people currently solve it. You define the goal and identify the relevant facts, including the constraints that bind you. Then you construct a solution from those starting points, keeping the existing arrangement open to revision.
A first principle is a starting point for this particular argument, not necessarily an eternal truth. Conservation of energy is a different kind of premise from an estimate of customer demand. Both can inform a design, but the estimate needs testing and may change. A budget can be a binding constraint without being a law of nature.
The discipline is to expose those differences. Write down what you know, what you assume, and what someone has merely decided.
This differs from root cause analysis, which traces why an outcome occurred. First principles thinking asks what a solution must satisfy. It also goes beyond simply rejecting precedent. An unfamiliar proposal needs sound reasoning and evidence to justify a claim of improvement.
03How to use it
- Describe the required result before choosing a solution. Replace a request for a larger warehouse with a requirement to hold enough stock to meet a specified service level. Include who benefits and how you will recognize success.
- List the starting points and label their status. Separate well-supported facts, uncertain estimates, binding commitments, and inherited practices. Include units and ranges where possible: orders per day, minutes per task, available dollars. A guess remains uncertain even when it looks precise.
- Challenge the constraints one at a time. Identify why each exists. Physical limits, laws, contracts, and preferences constrain choices differently. Distinguish fixed limits from terms that can be renegotiated. Use Chesterton’s fence before discarding a rule whose purpose you do not understand.
- Build a small model of the problem. A sketch, an equation, or a short chain of reasoning is enough. Show how the inputs produce the required result. A complicated spreadsheet can hide important dependencies. Keep them visible.
- Construct more than one solution. Look for different ways to satisfy the same requirements. Reuse proven components when they fit. Analogical reasoning is welcome here: borrowing a solution is sensible once you understand why it should work.
- Test the premise most likely to break the design. Measure the draft, observe the approval process, or trial the proposed service. Choose a falsification test that could reveal you are wrong. A demonstration that only looks reassuring can leave that uncertainty unresolved. Revise both the solution and its starting assumptions when the results disagree.
The useful output is a proposal whose reasoning another person can inspect, including the starting points you treat as bedrock.
04A worked example
In the 1930s, MIT graduate student Claude Shannon worked with the differential analyzer, a mechanical calculating machine whose controls included relays. Complicated switching circuits posed a design problem: how could an engineer reason systematically about which combinations of switches would complete a circuit?
What it looks like An electrical engineering task requiring careful circuit drawings and familiarity with existing arrangements. A designer could modify a known circuit and trace its possible states.
What’s actually going on For the switching behavior Shannon wanted to analyze, a contact could be treated as open or closed. Contacts in series conduct only when all are closed; parallel paths conduct when at least one path is complete. These relationships could be represented with Boolean algebra, the mathematics of logical conditions. Circuit behavior became something engineers could calculate and simplify.
For a small illustration, imagine a lamp that should light only when a master switch is closed and at least one of two other switches is closed. Those requirements suggest a master switch in series with two parallel branches. The arrangement follows from the requirement, which can be specified independently of an existing circuit. The lamp is a hypothetical illustration of the logic, distinct from Shannon’s actual apparatus.
What made it work Shannon isolated the behavior relevant to the problem and connected it to an existing mathematical framework. His 1937 thesis, published as a paper in 1938, provided a systematic method for analyzing and synthesizing switching circuits. He drew on electrical knowledge and existing components. The abstraction focused on switching logic, leaving other physical details of relays outside its scope.
05When to reach for it
06When it misleads
- You promote an assumption into a fact. A forecast of demand or a claim about human motivation remains an assumption even when you write it at the bottom of a diagram. Mark uncertainty and test it.
- You remove the inconvenient parts of reality. A cheap design may depend on perfect coordination, free maintenance, or users behaving exactly as instructed. Those omissions can erase the apparent advantage.
- You confuse theoretical feasibility with practical feasibility. Raw materials account for only part of the cost of a finished product. Manufacturing, reliability, distribution, and support still count.
- You dismiss accumulated expertise. A strange-looking rule may preserve a lesson from an earlier failure. Ask experienced people what breaks when it is removed. First principles thinking should expose functional fixedness and guard against overconfidence that beginners know everything experts missed.
- You rebuild what already works. Deriving every choice anew consumes time and attention. Use the method where inherited assumptions are costly or doubtful. Applying it as a ritual for choosing office supplies wastes time and attention.
Human systems are especially difficult to reduce. A plan may have to sustain relationships while earning trust and legitimacy. A technically elegant plan that people will not accept may fail on its own stated goal.
07Roots
Aristotle, writing in fourth-century BCE Greece, faced a problem at the heart of proof: every demonstration depends on something already accepted. A proof about a triangle cannot rest on an endless chain of earlier proofs. In Posterior Analytics, he examined the starting principles from which demonstrations proceed and how people come to know them.
That account addressed knowledge and proof, a different task from running a modern brainstorming workshop. It established an enduring distinction between the premises of an argument and what follows from them. Later scientific and mathematical traditions repeatedly put that distinction to work: make the starting points explicit so you can derive consequences and examine whether those consequences fit the problem.
Shannon’s relay circuits show how that habit traveled into engineering. A tangle of contacts became tractable when he isolated their switching behavior and expressed it mathematically. The modern label covers a broader problem-solving practice with no single inventor. The practice works from assumptions, and its most defensible inheritance is the responsibility to show which ones your answer depends on.
08How solid is this?
Explicit premises and derivation are standard in mathematics, science, and engineering. First principles thinking is a broad framework with unstandardized methods. Claims that it reliably outperforms expertise or analogy in everyday decisions require evidence from those settings, beyond successful technical applications.
09Connections
- Often confused withRoot Cause Analysis
- Helps counterFunctional Fixedness, Status Quo Bias
- Includes Constraint Relaxation
- See also Fermi Estimation, Falsification Test, Abstraction, Chesterton’s Fence, Analogical Reasoning
10Origin and sources
An ancient philosophical tradition rather than a technique with one inventor. Aristotle examined first principles in Posterior Analytics in the fourth century BCE; reasoning from explicit foundations became widely used in mathematics, science, and engineering.
- [1]Aristotle (1993). Posterior Analytics. Translated with a commentary by Jonathan Barnes (2nd ed.). Clarendon Press.
- [2]Shannon, C. E. (1938). A symbolic analysis of relay and switching circuits. Transactions of the American Institute of Electrical Engineers, 57(12), 713–723.
- [3]Petroski, H. (1985). To Engineer Is Human: The Role of Failure in Successful Design. St. Martin's Press.
Suggest an edit· Updated 2026-10-02