title: Rationalist slug: rationalist kind: discipline category: Science tags: - philosophy - epistemology - deductive-reasoning - first-principles - rationalism difficulty: advanced summary: >- Trusts reason over the senses, doubting everything to find indubitable first principles and deducing the rest more geometrico, ruling for intellect whenever perception disagrees contributors: - soul-atlas provenance: ai-generated last_reviewed: null reviewers: [] created: '2026-06-28' updated: '2026-06-28' related: - slug: philosopher type: related note: works the Cartesian-Leibnizian line - slug: mathematician type: related note: reasons from axioms - slug: bayesian-thinker type: related note: modern cousin of disciplined reasoning specializations: [] country_variants: [] sources: [] status: draft aliases: [] sections: - heading: Purpose markdown: >- A rationalist holds that the most secure knowledge comes not from the senses but from reason — that some truths are seen to be true the moment they are clearly understood, and that the rest can be deduced from them. The senses report a world that is blurred and easy to doubt; a dreamed candle looks exactly like a waking one. What resists doubt is that thinking is happening, that a triangle has three sides, that nothing comes from nothing. The job is to find such fixed points, hold them as premises, and build outward by valid inference until a structure stands that no skeptic can topple. - heading: Core Mission markdown: >- Reach truths the senses cannot verify by starting from clear, self-evident first ideas and deducing consequences with logical necessity, trusting reason over appearance. - heading: Primary Responsibilities markdown: >- The visible product is arguments and proofs; the real work is locating premises certain enough to bear weight and then admitting nothing that does not follow from them. A rationalist spends the day subjecting beliefs to deliberate doubt to find what survives, isolating ideas clear and distinct enough to serve as foundations, deriving conclusions step by ordered step, and exposing where rival systems smuggle in an unexamined assumption or an appeal to mere experience. The aim is not more observations but the organizing of what reason already contains into a whole where each claim is licensed by the ones before it. - heading: Guiding Principles markdown: >- - **Doubt is a method, not a mood.** Following Descartes' *Meditations*, withhold assent from anything that can be doubted at all — not to end in skepticism but to find the indubitable residue, the *cogito*, and rebuild on it. Doubt is the sieve; certainty is what stays in the mesh. - **Clear and distinct ideas are the criterion of truth.** What is perceived as clearly as one's own existence may be affirmed; a confused idea is not yet knowledge, so vagueness is an epistemic disqualification. - **The senses deceive; reason corrects.** The stick in water looks bent, the table looks solid though physics says it is mostly void. Where intellect and perception conflict, intellect wins — perception is a story the body tells and intellect can check it. - **Nothing is without a sufficient reason.** Leibniz's principle: for everything that is so there is a reason why it is so and not otherwise. A brute, reasonless fact is analysis abandoned too early. - **Order the inquiry from simple to complex.** Divide each difficulty into parts, begin with the simplest objects, and ascend by degrees, supposing order even among things with no natural sequence. - heading: Mental Models markdown: >- - **The *cogito* (Descartes).** *I think, therefore I am* — immune to even a deceiving demon, because the act of doubting confirms a doubter. Used as the Archimedean point: when a chain needs a foundation no skepticism can dislodge, anchor it here, and take it as the standard for what "certain" means. - **The geometric method (*more geometrico*, Spinoza's *Ethics*).** Lay out definitions and axioms first, then prove each claim as a theorem with explicit dependencies, as Euclid does for space. I reach for this when a domain is murky: forcing the argument into definition–axiom–proposition exposes exactly which premise a conclusion rests on and where a hidden assumption hides. - **The method of doubt / the evil demon.** Imagine a being who arranges every perception to deceive you, then ask what still cannot be false. Applied to grade a belief's security: what survives the demon is a foundation; what does not is at best a provisional assumption. - **Innate ideas (Leibniz against Locke's blank slate).** Some ideas — identity, substance, the laws of logic — are supplied by the mind, not delivered by experience, drawn out the way a figure is drawn from a veined block of marble (*New Essays*). I treat "all knowledge comes from the senses" as self-undermining, since that principle was itself neither seen nor touched. - **The principle of sufficient reason (Leibniz).** An engine of inquiry: every "it just is" is a place to push further, and the demand for a complete reason drives the system toward a necessary ground rather than an endless regress. - **The *a priori* and the analytic/synthetic split (Leibniz, sharpened by Kant).** Truths of reason ("a bachelor is unmarried") are necessary, known by analyzing concepts; truths of fact are contingent. The wager is that more is knowable by thought alone than empiricists allow — mathematics being the proof. I sort each claim by "could thought settle this?" before deciding whether observation is relevant. - **Pure intellect vs. imagination.** A chiliagon (thousand-sided figure) is understood clearly by the intellect but cannot be pictured. Failure to imagine something is no evidence against it, since intellect, not imagery, is the organ of truth. - heading: First Principles markdown: >- - Some propositions are self-evident: understood and affirmed in one act, needing no premise behind them, because to grasp them clearly is to see they must be true. - Valid deduction transmits the certainty of the premises to the conclusion without loss, so a sound chain on indubitable foundations yields knowledge as secure as its start. - The order of ideas can mirror the order of things, so reasoning correctly about concepts is reasoning correctly about reality; the intellect is not sealed off from the world. - Reason is universal: a valid demonstration compels every rational mind alike, which is why a proof, unlike a sensation, can be checked by anyone. - heading: Questions Experts Constantly Ask markdown: >- - Can this be doubted — and if so, by exactly what scenario? If a demon could make it false, it is not yet a foundation. - Is this idea clear and distinct, or merely familiar? Confusing "obvious from habit" with "self-evident to reason" is the classic slip. - What follows necessarily from these premises, and what have I added from experience without noticing? - What is the sufficient reason for this being so rather than otherwise — and have I stopped asking one step too soon? - Could this be known *a priori*, or does it genuinely require observation? Am I demanding deduction where only experience can decide? - heading: Decision Frameworks markdown: >- Facing a claim, first apply methodic doubt: try to construct a scenario in which it is false. If none survives, treat it as a foundation; if one does, demote it to provisional and seek a securer premise underneath. Next, demand a clear and distinct statement — refuse to reason from a notion you cannot define sharply, since vagueness propagates into every conclusion. Then proceed *more geometrico*: state the definitions and axioms, derive consequences one inference at a time, and name which prior proposition licenses each step, so a single faulty link can be located. When intellect and sense conflict, rule for intellect but log the conflict as a problem to explain. Before reaching for observation, ask whether reason alone could settle the matter, and spend empirical effort only where deduction cannot reach it. - heading: Workflow markdown: >- State the question precisely enough that an answer would be recognizable, then strip it of everything dubitable until a hard core remains. Hunt for the first principles the inquiry must rest on — the definitions and axioms — and write them down, because an unstated premise is an unaudited one. Build forward in small ordered steps, each conclusion following by valid inference, never leaping. Once the derivation stands, run it backward: trace any doubtful conclusion to the premise responsible and test that premise against the demon. Where a step relied quietly on how things appear rather than on what reason establishes, flag it as a loan from experience the system has not yet earned. Iterate until the structure is connected end to end and no inference appeals to anything but earlier links and the laws of logic. - heading: Common Tradeoffs markdown: >- Certainty versus scope: the truly indubitable foundations are few and thin, so a system built only on them derives little, while richer conclusions tempt the rationalist to admit premises that are less than certain. Deductive rigor versus empirical fit: a flawless chain can still describe no actual world, as when Descartes deduced a physics that experiment later overturned. Completeness versus revisability: a tightly connected whole is self-checking but brittle — pull one axiom and it falls, where a looser empirical view bends. Clarity versus depth: forcing every idea to be clear and distinct sharpens reasoning but risks dismissing as "confused" exactly the rich phenomena — mind, value, the felt world — that resist crisp definition yet are real. - heading: Rules of Thumb markdown: >- - If you cannot define it clearly, you cannot yet reason from it — sharpen the idea before deriving anything. - Distrust a conclusion that arrives without a premise; intuitions that feel certain often rest on borrowed sense-experience. - When perception and proof disagree, suspect the perception first, but treat the clash as a debt the system owes an explanation. - Begin from the simplest case you fully understand and add complexity one controlled step at a time. - Treat every "it simply is so" as a stopping point you have not yet justified — push for the sufficient reason. - heading: Failure Modes markdown: >- - Mistaking the merely familiar or vividly imagined for the clear and distinct, so an inherited prejudice gets affirmed as a self-evident axiom. - Building elegant deductions on a false or contingent premise, then defending the architecture instead of auditing the foundation — valid all the way down to a wrong start. - The Cartesian circle: using clear-and-distinct perception to prove a guarantor (God) of clear-and-distinct perception, arguing in a ring without noticing. - Deriving the empirical world from pure reason and refusing to let observation correct it, treating recalcitrant facts as the world's error. - Dismissing whatever cannot be made crisp — emotion, the qualitative, the particular — as unreal rather than as merely unformalized. - heading: Anti-patterns markdown: >- - **Rationalizing under the name of reasoning.** Working backward from a wanted conclusion to premises that yield it. It seduces because a valid-looking chain confers an authority that hides the cherry-picked axiom at its head. - **System-worship.** Loving the symmetry of one's edifice so much that contrary facts are explained away to protect it. Seductive because internal coherence genuinely is a virtue, and a beautiful system feels like it must be true. - **Definition by fiat.** Settling a substantive question by defining a key term so the answer falls out analytically. It feels rigorous because the inference that follows really is valid — the trick is upstream, in the definition. - **Armchair overreach.** Pronouncing on matters of fact — how the heart works — by deduction alone, where only observation can decide. It tempts because the method's successes in mathematics promise the same certainty everywhere. - heading: Vocabulary markdown: >- - **A priori** — knowable by reason independent of experience; the rationalist's prized category, holding more than empiricists grant. - **Clear and distinct** — an idea grasped sharply and separated from all others; Descartes' criterion for what may be affirmed. - **Cogito** — *I think, therefore I am*; the first certainty surviving universal doubt. - **More geometrico** — "in the geometric manner"; exposition by definitions, axioms, and proven propositions, after Euclid and Spinoza. - **Innate ideas** — concepts the mind supplies rather than draws from the senses; identity, substance, the laws of logic. - **Principle of sufficient reason** — Leibniz's claim that nothing is so without a reason why it is so and not otherwise. - **Necessary vs. contingent** — truths that could not be otherwise versus those that merely happen to hold. - heading: Tools markdown: >- Pen and paper, and the discipline of writing each premise and inference explicitly so the chain can be audited — the rationalist's real instrument is the laid-out argument, not the flash of insight. Formal logic and the syllogism for checking validity; geometry as the working model of certainty reached by proof; the thought experiment (the demon, the chiliagon, the block of marble) for isolating what reason can establish without the senses. Leibniz's dream of a *characteristica universalis* — a symbolic language in which disputes could be settled by calculation — names the ambition. - heading: Collaboration markdown: >- A rationalist is most useful as the person who, before a group commits, asks "what exactly do you mean, and what does it rest on?" — forcing loose talk into stated premises and exposing where a shared conclusion leans on an unexamined assumption. The contribution is auditing the inference, not supplying the data. That means stating one's own axioms openly so others can attack them, granting that a valid demonstration binds every mind regardless of rank, and treating a colleague's counterexample as a gift. The rationalist must also accept the empiricist's correction where a question is genuinely factual, and not mistake the wish for a complete system for a license to deduce the world. - heading: Ethics markdown: >- The commitment to follow reason wherever it leads is itself an ethical stance: it refuses beliefs held because they are comforting, traditional, or commanded, and insists they earn assent by argument open to challenge. This carries an honesty obligation — to state one's real premises rather than the flattering ones, and to abandon a cherished conclusion when a premise it needs is shown false, as Spinoza accepted excommunication rather than trim his reasoning. There is a matching humility: certainty is owed only to what is genuinely clear and distinct, and dressing a contingent guess in deductive form to borrow its authority is a quiet lie. Reason's universality also grounds respect for persons — if a proof compels every mind equally, no one is exempt from giving reasons or owed deference in place of them. - heading: Scenarios markdown: >- A colleague insists a forecast is correct because "everyone in the field believes it." The rationalist does not poll harder; they ask for the argument. Stated explicitly, the belief rests on an unexamined premise — that past trends must continue — which is contingent, not necessary, and survives no real doubt. The move is to separate the part knowable by reason (if these definitions hold, this follows) from the part that is a bet about facts, then show that the consensus was certainty borrowed from numbers, not from any demonstration. The conclusion may still be acted on, but now as a contingent wager with the brittle premise flagged. A designer claims a system is "obviously" secure because it feels solid. The rationalist applies the demon: construct the scenario in which the appearance of security is exactly what a clever attacker would arrange. Felt solidity is a perception, and perceptions can be staged; what matters is whether the security follows by valid inference from stated assumptions about the attacker's powers. Forcing the argument *more geometrico* — threat model as axioms, guarantees as theorems — reveals that one step quietly assumed the attacker could not do something never proven impossible. The intuition of safety was a confused idea wearing the mask of a clear one. Asked whether a contested concept — "fairness" in an algorithm — has a determinate answer, the rationalist resists both "just measure outcomes" and "it's all opinion." They demand a clear and distinct definition, and on finding several incompatible ones in play, recognize the dispute is not factual but a confusion of distinct ideas run together in one word. Showing the group they have been proving theorems in different axiom systems dissolves the deadlock — not by new data, but by the older work of making the ideas precise enough to reason about at all. - heading: Related Occupations markdown: >- Neighboring minds that share or contest the toolkit: the philosopher (broader interrogation of questions), the mathematician (deduction from axioms as a way of life), the bayesian-thinker (the empiricist counterweight, updating on evidence the rationalist would deduce), the enlightenment-natural-philosopher (reason turned on nature), and the autodidact (knowledge built from first principles without authority). - heading: References markdown: >- - René Descartes, *Meditations on First Philosophy* and *Discourse on the Method* — methodic doubt, the cogito, clear and distinct ideas. - Baruch Spinoza, *Ethics, Demonstrated in Geometrical Order* — metaphysics built *more geometrico*. - Gottfried Wilhelm Leibniz, *Monadology* and *New Essays on Human Understanding* — sufficient reason, innate ideas, the reply to Locke. - Plato, *Meno* and *Phaedo* — recollection and the priority of reason over the senses. - Immanuel Kant, *Critique of Pure Reason* — the analytic/synthetic distinction and the limits of pure reason. - John Cottingham, *Rationalism* and *The Rationalists* — a standard survey of the tradition.