The Flood and Its Demands
We live in an environment saturated with claims. News, social media, advertising, political speech, scientific reporting, personal testimony, institutional statement, expert opinion, and amateur speculation — all compete for attention and credence in a volume that no previous generation has faced. The practical demand this places on the ordinary person is extraordinary: to navigate this flood without being drowned by it, to extract what is reliably true without being paralysed by universal scepticism, to maintain calibrated confidence in the face of constant noise.
This is not merely a technical skill. It is a moral one. The person who accepts false claims because they are emotionally convenient, who dismisses true claims because they are uncomfortable, who delegates judgement entirely to the sources they already trust — this person is not merely mistaken. They are failing a duty that the serious life imposes: the duty to proportion belief to evidence, to take reality seriously enough to engage with it honestly, and to resist the easy comfort of believing whatever flatters one's existing picture of the world.
The doctrine names this duty clearly: confidence must be proportioned to evidence and scrutiny. What follows is a practical guide to how that proportioning is done.
The First Question: What Kind of Claim Is This?
Not all claims are alike, and different types of claim require different standards of evaluation. Distinguishing them is the first step in any honest assessment. Empirical claims — claims about what is the case in the observable world — are in principle testable. They can, at least in theory, be confirmed or refuted by evidence. Normative claims — claims about what ought to be done, what is valuable, what is right — involve judgements that evidence alone cannot fully settle, though evidence remains highly relevant. Definitional claims — claims about what a word means or how a concept is to be understood — are settled by convention and argument, not by observation.
Much confusion in public discourse arises from the failure to distinguish these types. A scientific finding about the effects of a policy is an empirical claim; a judgement about whether that policy should be adopted involves normative considerations that go beyond the empirical finding. Conflating them — treating the science as if it settles the policy question, or treating the policy disagreement as if it casts doubt on the science — is a common error with significant consequences.
Within empirical claims, a further distinction matters: between claims based on well-established, extensively replicated bodies of evidence, and claims based on preliminary, single-study, or contested evidence. The scientific consensus on human-caused climate change is not epistemically equivalent to a preliminary finding from a single study that has not yet been replicated. Treating them as if they are — as if 'science' speaks with uniform authority on all questions it touches — is a misunderstanding of how scientific knowledge actually works.
The Second Question: What Is the Evidence, and How Strong Is It?
Once the type of claim is clear, the question of evidence can be engaged seriously. Good evidence is not simply evidence that supports a conclusion one prefers. It is evidence that was gathered rigorously, that is relevant to the specific claim being made, that has been subjected to scrutiny and critique, and that survives that scrutiny.
The hierarchy of evidence in medicine — a field that has developed particularly sophisticated standards for evaluating claims — offers a useful model. At the top sit systematic reviews and meta-analyses of multiple well-designed randomised controlled trials: studies that average over many independent investigations to produce more reliable estimates. Below them sit individual randomised trials, then observational studies, then case series and expert opinion. The further down the hierarchy a piece of evidence sits, the less weight it should carry in forming a confident conclusion.
Outside medicine, comparable hierarchies apply. In the social sciences, controlled experiments and large, pre-registered observational studies carry more weight than single case studies or uncontrolled observations. In history, contemporaneous primary sources carry more weight than later reconstructions. In everyday reasoning, direct observation and multiple independent corroborations carry more weight than single testimony or inference from general principles.
The question of source matters too. Is the source motivated to reach a particular conclusion? Does it have a track record of accuracy and correction? Has it been subjected to peer review or other forms of expert scrutiny? Is it transparent about its methods and data? These are not questions of cynicism; they are questions of epistemic hygiene. No source is perfectly reliable, and no source is perfectly unreliable. Calibrating one's trust to the track record and incentive structure of the source is a basic requirement of honest reasoning.
The Third Question: What Would Change My Mind?
Perhaps the most important and most neglected question in the evaluation of evidence is: what would it take for me to conclude that this claim is false? The person who cannot answer this question — who cannot specify any evidence that would change their view — is not reasoning. They are performing. Their claim to be evaluating evidence is a pretence: the conclusion has been decided in advance, and the evidence is being recruited rather than genuinely assessed.
This is the test the doctrine describes as distinguishing honest doubt from corrupt doubt: whether the uncertainty is answerable to reality or serves self-image. The same test applies to certainty. The honest conviction is one that could, in principle, be revised by evidence; the ideologically entrenched conviction is one that has immunised itself against revision.
Applying this test to one's own reasoning is uncomfortable but essential. In practice, it means actively seeking out the strongest version of opposing evidence and arguments, rather than the weakest. It means asking whether the same standards of scrutiny are being applied to evidence that supports one's view as to evidence that challenges it. It means noticing when one is experiencing a conclusion as obviously true in the absence of having actually examined the evidence carefully.
The goal is not perpetual uncertainty or the refusal to form any convictions. It is calibrated conviction — confidence proportioned to the quality and quantity of evidence. Some claims deserve high confidence. Some deserve provisional acceptance pending further evidence. Some deserve scepticism or outright rejection. The discipline of evidence evaluation is the art of placing claims in the right category — and remaining genuinely open to reclassification when the evidence demands it.
Correction refines rather than diminishes the serious person.