Reason

The capacity to work out what follows, what to conclude and what to do - and the question of what it is for.

Theme: Knowledge, truth and reason
Tradition: Cross-tradition
Term: λόγος · logos
Term fixed: c. 500 BC

In plain terms

Reason is the thing that is supposed to distinguish us, and it is not obvious what it is. Following a valid inference is part of it. So is weighing evidence, revising in the face of objection, and giving grounds that another person could assess.

That last part matters more than it looks. Reasoning is not obviously an inner process at all - much of it is public, done out loud, between people who disagree. And humans are demonstrably better at it in company than alone, and better at finding flaws in an opponent's argument than in their own.

Which raises a question about what the capacity is for. If it evolved to find truth, its performance is patchy in a specific way. If it evolved to produce and evaluate arguments in a social setting, the same performance is exactly what you would expect.

As a philosopher would put it

The rationalist tradition takes reason to deliver substantive knowledge independently of experience: Descartes, Spinoza and Leibniz all hold that some truths about reality are established by reason alone. Empiricism restricts it to relations between ideas, and Hume's conclusion is that reason is inert with respect to action - it can determine means but not ends, and is the slave of the passions.

Kant's critique is the pivot. Reason has legitimate use within the bounds of possible experience and generates antinomies beyond them, so the critique of reason is conducted by reason itself. Practical reason is then given a role Hume denies it: the categorical imperative is a demand of reason on action.

The twentieth century has largely naturalised the question. Empirical work on human reasoning found systematic departures from logical norms, which produced a long argument about whether people are irrational or the norms are the wrong standard. Mercier and Sperber's argumentative theory takes the second line: reasoning evolved for producing and evaluating arguments, which explains why confirmation bias and the group advantage are features rather than defects.

Where you meet it

In 1966 Peter Wason gave people four cards showing A, D, 4 and 7, and a rule: if a card has a vowel on one side, it has an even number on the other. Which cards must be turned over to test it?

The answer is A and 7. Fewer than ten per cent get it. Most people choose A and 4, looking for confirming instances rather than the case that could falsify the rule.

The same task in a different dress produces a different result. Told that they are a bouncer enforcing a rule that anyone drinking alcohol must be over eighteen, and shown cards reading beer, cola, twenty-five and sixteen, most people immediately turn over beer and sixteen - the identical logical structure, solved by around three quarters.

The logic did not change. What changed was whether the problem was about detecting a cheat. Fifty years of work has not produced agreement on why, and the result is one of the most robust findings about human reasoning there is.

What is actually disputed

What reason is for, and whether it has a single standard.

If reasoning aims at truth, the systematic biases are malfunctions and the interesting question is how to correct them. If it aims at argument and persuasion, the biases are adaptations, individual reasoning is expected to be poor, and the fix is not training but discussion with people who disagree. The evidence that groups outperform individuals on exactly these tasks supports the second reading and does not settle it.

Underneath sits the older question of scope. Hume's claim that reason cannot set ends and Kant's claim that it can is the central disagreement in moral philosophy, unresolved after two and a half centuries. And the standard against which human performance is judged is itself contested: departures from formal logic may be departures from a norm that was never the right one for reasoning under uncertainty with limited time.

Commonly confused with

being unemotional Nothing in the concept requires the absence of feeling, and there is good evidence that people with impaired emotional response reason worse rather than better about what to do.

Defined against

Emotion The contest between them organises most classical accounts of the person, and almost every tradition rules on which governs which

Philosophers

Heraclitus Logos as the order of things before it is a faculty
Aristotle Reason as what is distinctive of the human, and the study of its forms
René Descartes Reason as delivering substantive truths without experience
David Hume Reason as inert, the slave of the passions
Immanuel Kant Reason critiquing its own limits, and legislating for action

Works

Fragments Logos in its earliest philosophical use
Nicomachean Ethics The rational part, and practical wisdom
Meditations on First Philosophy Reason as a source of knowledge on its own
A Treatise of Human Nature The demotion of reason in motivation
Critique of Pure Reason The limits of reason, established by reason
Groundwork of the Metaphysics of Morals Practical reason as legislating

Schools

Heraclitean School Logos as cosmic order
Peripatetic school (Aristotelianism) Reason as distinctive and as trainable
Early Stoicism Reason as both faculty and world-order
Cartesianism Rationalism at its most confident
British Empiricism Where its scope is cut back
German Idealism Where it becomes self-critical

Related concepts

Inference c. 330 BC · Knowledge c. 369 BC · Wisdom c. 340 BC · Emotion c. 340 BC · Deduction c. 330 BC

Heraclitus's logos is an order in things as much as a faculty in people; the narrowing to an individual capacity for inference happens gradually and is complete by the early modern period.