Knowledge, reason and experience
Rationalism
Rationalism is a family of views about what reason can contribute to knowledge and where some of our concepts come from. A rationalist may hold that certain truths can be known independently of sense experience, or that the mind possesses some knowledge or concepts innately. This does not mean rejecting observation: the useful question is which claim, in which subject, reason is supposed to establish.
Updated
Three claims to keep separate
| Claim | Question it answers | What it does not establish |
|---|---|---|
| Intuition and deduction | Can some truths be grasped or inferred without sensory evidence? | That every person already knows those truths at birth. |
| Innate knowledge | Does the mind possess some knowledge independently of learning it from experience? | That the knowledge is consciously available to a newborn. |
| Innate concepts | Do some concepts originate in the mind rather than being acquired from experience? | That every proposition formed with those concepts is true. |
| A domain-specific position | Is the claim about mathematics, metaphysics or some other subject? | That its proponent takes the same position in every subject. |
| Reason in empirical inquiry | How do we infer conclusions from observations? | That using logic by itself makes an inquiry rationalist. |
What makes a position rationalist?
Begin with a claim about knowledge, not a personality type. A person can be patient, logical and attentive to evidence without endorsing philosophical rationalism. Likewise, a rationalist can collect observations and revise a scientific hypothesis. The disagreement concerns whether experience supplies all the relevant materials or grounds of knowledge in a particular domain.
Consider two questions: how many windows are open in this room, and whether the sum of two even integers is even. Looking is an appropriate way to answer the first. Inspecting a handful of calculations is not what establishes the second for all integers. Rationalist arguments often begin from this difference between finding particular facts and establishing a general or necessary truth.
The label is therefore relative to a subject and a thesis. A thinker might give reason a special role in arithmetic while relying on experiments in physics. Classifying a whole philosopher with one word can be a useful introduction, but it is an incomplete account of the arguments. Ask what the thinker says can be known and what makes knowing it possible.
Further reading: Rationalism vs. Empiricism
The basis of a belief differs from its origin
A belief can have an experiential occasion without having an experiential justification. Seeing pairs of counters may prompt a child to ask about even numbers. Later, a mathematical proof can establish a result independently of whether the counters were counted accurately. The event that started the inquiry and the ground that supports its conclusion perform different jobs.
The same distinction applies to concepts. Asking where a concept comes from is a psychological or philosophical question about possession and acquisition. Asking whether a proposition using that concept is justified concerns its warrant. Being able to think of a perfect triangle does not guarantee the truth of every geometric claim one makes about triangles.
This is why a slogan such as “knowledge comes from reason” needs unpacking. Does it mean reason supplies the concepts, establishes the propositions, or both? Does “independent of experience” concern the justification of a claim or the history of learning it? Without these distinctions, an argument for one thesis can be mistaken for an argument for a much stronger one.
Further reading: Rationalism vs. Empiricism
A worked example: why two even integers add to an even integer
Take any even integers a and b. By the definition of evenness, there are integers m and n such that a = 2m and b = 2n. Then a + b = 2m + 2n = 2(m + n). Since the sum m + n is also an integer, a + b is twice an integer and is therefore even. The letters cover every pair satisfying the assumptions.
Checking 4 + 8 or 12 + 16 illustrates the claim, but the proof does more than collect successful cases. It identifies a reason the conclusion follows for arbitrary even integers. If somebody objects that the examples omit very large negative integers, the proof already includes them: its assumptions do not restrict the integers to a small positive sample.
This example explains the appeal of knowledge through reasoning. It does not settle the origin of the concept of number or prove that the theorem was innately known. Someone can learn definitions and the practice of proof before coming to know the result. A rationalist still needs an argument for any additional thesis about innate knowledge or innate concepts.
Further reading: Rationalism vs. Empiricism
Intuition and deduction need more than a feeling of certainty
In these debates, intuition is an alleged direct intellectual grasp of a proposition, rather than a hunch about tomorrow’s weather. Deduction connects propositions so that a conclusion follows from its premises. A position that gives these methods a special role must explain both why the starting points deserve acceptance and why the inferential steps preserve the required support.
A deduction can be valid while starting from a false premise. A person can also feel completely certain about a proposition that somebody else rejects. The fact that a claim appears obvious is not automatically a public demonstration of its truth. These are different challenges: one concerns the argument’s structure, while the other concerns the status of its starting points.
The Stanford discussion also cautions against using the intuition-and-deduction thesis alone as a simple boundary between all rationalists and empiricists. Empiricists may recognize mathematical demonstrations and relations between ideas. A stronger disagreement may concern what experience can supply, or whether there is innate knowledge or conceptual content beyond it.
Further reading: Rationalism vs. Empiricism
Innate knowledge is not the same as a built-in capacity
An innate-knowledge thesis says that some knowledge is possessed independently of acquiring it through experience. It need not mean that a baby can consciously state a theorem or defend an argument. Some versions describe knowledge that becomes available under appropriate conditions. But explaining what possession amounts to is part of the thesis, not a detail that can be assumed away.
Compare three claims: a child can learn a language; a child has an inborn structure that helps language learning; a child already knows a particular truth. The first describes an ability. The second gives a possible explanation of that ability. The third attributes knowledge. Moving from one to the next requires further argument, even if all three claims concern development.
A familiar rationalist question asks how repeated observations could establish a truth as necessary, rather than merely as something that has always happened in the observed cases. That challenge can support a special role for reason. It does not by itself establish that the truth was already possessed before learning. The necessity argument and the innateness argument must be evaluated separately.
Further reading: Rationalism vs. Empiricism
Innate concepts concern the materials of thought
An innate-concept thesis concerns the origin of at least some concepts. It is different from saying that a person has an innate belief. A person might possess a concept while holding mistaken beliefs involving it, just as possessing the concept of a circle does not prevent drawing a distorted figure or making a false claim about its circumference.
Empiricist accounts may explain complex ideas through combining or modifying materials derived from experience. Rationalist objections can ask whether this explains every concept, especially concepts whose content seems to exceed a particular sensory image. Neither side is well represented by the contrast between a mind that thinks and a mind that merely receives pictures.
Hume’s famous missing-shade case concerns a person imagining a shade of blue absent from an otherwise ordered range of familiar shades. It creates a difficulty for a strict copying account. It does not automatically establish an extensive stock of innate concepts, and it should not be used as though a single exception resolved every question about conceptual origins.
Further reading: Rationalism vs. Empiricism · David Hume
Descartes, Leibniz and Hume: compare the actual commitments
Descartes is a central example because he seeks secure knowledge through intellectual examination and argues for innate ideas. His project also includes optics, geometry, physiology and explanations of natural phenomena. Treating him as someone who simply refuses to look at the world misses the relationship he proposes between metaphysical foundations and particular scientific investigations.
Leibniz challenges the idea that experience can supply the necessity of necessary truths. Particular cases can suggest a proposition without explaining why its denial is impossible. This argument is more precise than the claim that the senses are useless. Experience can occasion an inquiry while reason supplies a kind of support that the observed instances do not.
Hume, usually introduced as an empiricist, still gives reasoning an important place in understanding relations between ideas and in inquiry about matters of fact. His position is not a declaration that people should stop inferring or thinking. The contrast becomes useful when we ask how ideas arise and what supports our expectations about events, rather than assigning reason to one side and thoughtlessness to the other.
Further reading: Rationalism vs. Empiricism · René Descartes · David Hume
A proof and its application to the world are different questions
Suppose a geometrical argument establishes a result within a specified system. Applying that result to a physical object requires more than repeating the proof. One must decide whether the object is represented adequately by the system and how accurate the measurements are. A rough triangular field is not automatically an exact mathematical triangle.
Similarly, a calculation can establish that a proposed inventory total follows from the numbers entered. It cannot establish that every item was counted correctly or that the list is complete. The arithmetic and the evidence about the warehouse each need their own check. Confusing them makes a mathematically correct calculation appear to guarantee an empirical conclusion it does not support.
This distinction helps explain how mathematical reasoning and observation can cooperate. It also identifies a useful objection to an overextended rationalist argument: even if reason establishes a formal relationship, what warrants the claim that reality instantiates it? An account of rational knowledge must state the scope of the conclusion rather than silently treating every application as equally certain.
Further reading: Rationalism vs. Empiricism · René Descartes
The strongest objections ask about grounds and limits
One objection asks how supposed intellectual starting points are justified. If two people report conflicting intuitions, the theory needs a way to assess them. Appealing to confidence alone does not supply that method. Another objection asks whether apparently innate content can instead be explained through learning, abstraction or the mind’s general capacities.
A further challenge concerns scope. Success in demonstrating an arithmetical result does not show that reason alone can determine every fact about politics, weather or the mind. The rationalist should identify the relevant domain and explain why the proposed method applies there. The critic should address that specific claim rather than attributing an unlimited claim the argument never makes.
These objections leave room for more than one position. One can accept some a priori knowledge without accepting innate knowledge; defend innate concepts without treating all beliefs formed with them as certain; or recognize intellectual methods while disputing their reach into metaphysics. The useful debate concerns these combinations and the reasons for them.
Further reading: Rationalism vs. Empiricism
How to read further without mixing the questions
For a first pass, follow the distinction between a priori and a posteriori support. Then read the comparison of empiricism and rationalism with three questions in view: what can be known, what supplies the grounds, and where the concepts originate. Keep an example beside each answer so that similar terminology does not hide a different commitment.
For Descartes, read the first two Meditations before assuming that the cogito already proves a complete theory of mind and body. The wax passage is a focused case about understanding an object through changes. It is a useful companion to this overview, but its conclusion should not be turned into a claim that observation is irrelevant to every scientific question.
For Hume, begin with the account of ideas and matters of fact in the Enquiry concerning Human Understanding, then ask what grounds inference beyond what has been observed. Comparing these projects is more productive than treating rationalism and empiricism as opposing temperaments. Write down the precise proposition at issue before deciding which objection or response bears on it.
Further reading: Rationalism vs. Empiricism · René Descartes · David Hume
Check your reading
Does using a logical argument make someone a rationalist?
No. Empirical inquiry uses reasoning too. Identify the claim about experience-independent knowledge or the innate origin of knowledge or concepts, and the subject to which it applies.
Does the even-number proof show that its conclusion was known at birth?
No. It establishes the result from definitions and integer arithmetic. The history or innateness of knowing those definitions and the theorem is a further question.
Can a sensory experience prompt an a priori inquiry?
Yes. What occasions a belief can differ from what justifies it. Seeing counters may prompt the question without supplying the proof for every pair of even integers.
Does an innate concept guarantee a true belief?
No. Having a concept and knowing a true proposition involving it are distinct. The belief still needs appropriate support.
Can a mathematical proof guarantee a warehouse count?
It can establish an arithmetic consequence of the stated inputs. It cannot establish that the inputs describe every item correctly; that requires evidence about the warehouse.
Sources and further reading
References checked on 2 November 2026. The bell, archive, arithmetic and lending examples are constructed for explanation.
- Rationalism vs. Empiricism
Stanford Encyclopedia of Philosophy
Used for domain-specific classifications, intuition and deduction, and selected discussions of innate knowledge and concepts. These are competing theses, rather than a single commitment shared by every rationalist.
- René Descartes
Stanford Encyclopedia of Philosophy
Used for the works listed here and the discussions of metaphysics, mechanistic science and passions. Its account distinguishes the cogito from later arguments about the mind and body.
- David Hume
Stanford Encyclopedia of Philosophy
Used for life and works, the understanding, motivation, ethics and religion. The entry presents competing interpretations of causation, moral sentiment and the Dialogues; the summaries here preserve those disagreements.