Concept

Rationality

philosophycognitive-scienceepistemologyenlightenmentconsciousnessmulti-source
Teach me this →

Rationality

The capacity to use knowledge to achieve goals. Not a synonym for intelligence, and not the same as logic — rationality is the set of cognitive tools and habits that allow a mind to track truth, form accurate beliefs, and make decisions that serve long-run rather than momentary interests. It is contested at every level: what it is, whether humans have it, what grounds it, and whether Western culture’s dominant conception of it is too narrow to be useful.

Steven Pinker, Iain McGilchrist, and John Lennox each address rationality from very different starting points. Their convergence: all three regard the dominant cultural conception of rationality as deficient, though for opposite reasons. Their divergence: they locate the deficiency in different places and propose different remedies.

Pinker: rationality as cultivated skill

Pinker’s Rationality: What It Is, Why It Seems Scarce, Why It Matters (2021) treats rationality not as a natural endowment but as a culturally transmitted and individually cultivated skill. The tools of rationality are external systems for correcting the default tendencies of the human brain:

  • Logic: formal inference that preserves truth from premises to conclusions. The brain does not naturally follow modus ponens; it follows familiarity and plausibility. Logic forces consistency.
  • Probability and Bayesian inference: the correct way to update beliefs in response to evidence. Bayes’s theorem tells you how much to revise a prior probability when new evidence arrives. Most people are systematically miscalibrated: they over-update on vivid anecdotes (availability heuristic) and under-update on statistical base rates.
  • Decision theory: how to choose under uncertainty given goals and probabilities. Distinguishes expected value from worst-case thinking; explains why sunk costs are irrelevant to forward decisions; shows when risk aversion is rational and when it is not.
  • Game theory: how to reason when others are also reasoning strategically. Explains cooperation failures, arms races, and commitment problems.
  • Falsification and hypothesis testing: not collecting evidence for your view but actively seeking evidence against it.

Pinker’s central claim: humans have evolved intelligence (adaptive response to the local environment) but not rationality (truth-tracking capacities that work in the abstract, the statistical, and the long-run). Rationality must be learned. The Enlightenment created the cultural infrastructure — journals, universities, scientific communities, public discourse norms — that amplifies individual rationality beyond what any single mind can sustain alone. See Enlightenment.

The political dimension: Pinker argues that rationality is unintuitive. Liberal democratic institutions (checks and balances, freedom of speech, universal rights) do not flow naturally from tribal human psychology. They are constructed rational solutions to coordination problems. Without active maintenance, human political behaviour reverts to tribalism, deference to authority, and in-group/out-group moralisation. The current apparent decline in rationality is partly a failure of cultural transmission.

Cognitive biases as the antagonists of rationality:

  • Availability heuristic: judging probability by how easily examples come to mind (making terrorism feel more dangerous than car accidents).
  • Confirmation bias: seeking evidence that confirms existing beliefs.
  • Sunk cost fallacy: letting past investment affect present decisions.
  • Base rate neglect: ignoring statistical rates in favour of specific, vivid cases.
  • Motivated reasoning: deploying intelligence to defend conclusions already reached, not to track truth.

Pinker’s optimistic note: these biases are systematic, not random — which means they can be described, predicted, and corrected. Awareness of biases (debiasing), structured forecasting (Superforecasting), and institutional design (adversarial collaboration, peer review) are all techniques for ratcheting rationality upward.

McGilchrist: narrow rationality as the problem

McGilchrist’s critique is not that people are irrational but that the kind of rationality that dominates Western intellectual culture is structurally limited — a left-hemisphere mode of attending that mistakes itself for the whole of knowing.

The left hemisphere, on McGilchrist’s account, is the hemisphere of analysis, sequential reasoning, abstraction, and explicit formalisation. It works by decontextualisation: stripping away the particular to expose the general; reducing the complex to the simple; translating the embodied into the propositional. This capacity is genuinely useful — it is how we do mathematics, law, and logistics. But it is not the whole of cognition.

The right hemisphere attends to the world in its particularity: to context, to implicit meaning, to the body, to the living, to what cannot be fully formalised. It is the hemisphere that holds the whole picture — including the parts of the picture that logic and calculation cannot reach. McGilchrist argues that the right hemisphere is epistemically prior: it provides the ground from which the left hemisphere’s analysis proceeds, and to which conclusions must return for evaluation in full context.

The cultural problem is that the Enlightenment — while achieving genuine goods — systematised the left-hemisphere’s mode of attending and elevated it to the status of the only legitimate form of knowing. The result:

  • Scientism: the view that only what can be measured and formalised has epistemic standing
  • Reduction: treating the explained-in-terms-of-parts as equivalent to the whole
  • Loss of wisdom: the phronesis of Aristotle — practical wisdom, ethical judgment, contextual sensitivity — is not formalizable and therefore disappears from the category of ‘knowledge’

McGilchrist does not oppose rationality. He opposes the monopoly of a narrow form of rationality at the expense of a fuller form of knowing that includes perception, intuition, metaphor, and the body. Goethe’s ‘tender empiricism’ — attending to the phenomenon with patience and sensitivity rather than forcing it through a pre-existing conceptual grid — is his model of how to think well.

The implication for moral and philosophical reasoning: questions about meaning, value, and consciousness resist formalisation not because they are subjective but because they concern aspects of reality that the left hemisphere’s tools cannot reach. Demanding that all valid knowledge be explicit, propositional, and measurable does not make the excluded knowledge disappear — it makes the culture blind to it. See Hemisphere Lateralisation and Consciousness.

Lennox: rationality requires a rational ground

Lennox’s engagement with rationality is focused on a single question: what grounds the reliability of human reason? If the materialist account of mind is correct — if human cognitive faculties are the product of natural selection operating on random mutations — there is no reason to trust them as truth-tracking. Natural selection optimises for survival, not for truth. A faculty that generates systematically false beliefs may be reproductively advantageous; a faculty that generates true beliefs about mathematics, logic, and physics would be epistemically valuable but is not guaranteed by the evolutionary process.

This is the evolutionary argument against naturalism (EAAN), developed by Alvin Plantinga and drawn on by Lennox:

If our cognitive faculties were produced by unguided evolution, the probability that they are reliable (i.e., tend to produce true beliefs) is either low or inscrutable. But if we cannot trust our cognitive faculties, we cannot trust any of our beliefs, including the belief that evolution produced our cognitive faculties. The position is self-defeating.

Lennox extends the argument: the very practice of science and philosophy requires that we trust human reason to track truth, not merely to track survival advantage. The atheist or materialist who engages in rational argument is already implicitly treating reason as reliable — but on materialism, there is no explanation for why it should be. For Lennox, the rationality of the universe — its mathematical intelligibility, Wigner’s ‘unreasonable effectiveness of mathematics’ — and the reliability of human reason to grasp it both require a rational ground. The Logos of John’s Gospel: ‘In the beginning was the Word [Logos]’ — the rational principle underlying all things. See Language and Mind and Levels of Explanation.

Lennox’s corollary: faith is not a retreat from reason but an exercise of it. Pistis (the Greek word translated as ‘faith’) means trust based on evidence, not belief in the absence of or against evidence. The distinction between faith and reason is misdrawn in the standard Enlightenment account; the real distinction is between evidence-grounded belief and credulity. A scientist who trusts a colleague’s data is exercising faith of the same kind — trust based on track record and coherence.

The formal frontier: self-locating belief

Pinker’s toolkit treats Bayesian conditioning as the correct way to update on evidence, but the philosopher Adam Elga works at a frontier where that ‘correct way’ is genuinely contested. His signature case is the Sleeping Beauty problem: a coin is flipped; on heads Beauty is woken once, on tails twice with her memory wiped between wakings; on waking she learns nothing new about the coin itself, yet has she acquired self-locating evidence — evidence about which waking she now occupies — that should move her credence in heads from one-half to one-third? Fully-informed, competent agents disagree, and the disagreement is not a bias to be trained away; it exposes that standard conditioning is silent on how to update beliefs about one’s own place in the world. The same structure recurs in cosmology, where anthropic reasoning and Boltzmann-brain scenarios ask an observer to reason from the bare fact of their own existence in a very large universe — and where, Elga notes, taking the reasoning at face value can be ‘the equivalent of finding out that the encyclopedia you have been basing all of your life upon was typed by monkeys’. Elga also grounds the coherence the whole edifice presupposes not in intuition but in vulnerability: an agent whose degrees of belief violate the probability axioms can be turned into a money pump, accepting a sequence of individually acceptable bets that together guarantee a loss (a Dutch book). Rationality’s formal core, on this view, is at once more precise and less settled than the toolkit framing implies.

Where mainstream views differ

The mainstream debate in cognitive science and psychology follows Pinker’s framework: rationality is real, biased, and correctable. The question is which biases are universal, which are cultural, and which techniques most effectively reduce them. Kahneman and Tversky’s ‘heuristics and biases’ programme, Tetlock’s superforecasting research, and Sunstein and Thaler’s nudge theory all work within this frame.

The harder disagreement is between Pinker and his critics — from both directions:

The ‘rationality is insufficient’ critics (converging from McGilchrist and Lennox, though for different reasons): Both McGilchrist and Lennox argue that Pinker’s vision of rationality, while internally consistent, is too narrow for the tasks it is assigned. McGilchrist: it cannot ground ethical judgment or metaphysical insight because both require a mode of knowing the left hemisphere cannot supply. Lennox: it cannot ground its own reliability because a naturalistic account of reason is self-undermining.

The ‘rationality is culturally relative’ critics: postmodern and critical theory traditions argue that ‘rationality’ as defined by the Enlightenment is not universal but culturally specific — a particular tradition disguising itself as the universal. This is a weaker argument than either McGilchrist’s or Lennox’s because it does not engage with the specific content of Enlightenment rationality (probability theory, logic, evidence evaluation) but merely contextualises it historically. Pinker, McGilchrist, and Lennox all reject this move, though for different reasons.

The question of grounding: Pinker believes rationality needs no transcendent ground — it is a tool humans developed, and its justification is its track record. Lennox believes it requires a transcendent rational ground, because otherwise the track record argument is circular (we trust reason to evaluate the track record of reason). McGilchrist is closer to Lennox on this point than to Pinker: the right hemisphere’s integrative mode of knowing, which he treats as prior to and richer than the left hemisphere’s formal rationality, opens toward what he calls the sacred as one of the four epistemological pathways. See Science and Religion.

In the wiki