Philosophy of Science
Consensus Isn't a Vote: What 'The Science Is Settled' Actually Means
The most dangerous misunderstanding in public life isn't that people distrust scientific consensus — it's that almost nobody, including scientists, can explain what consensus actually is.
The Idea
Scientific consensus is not a poll. It doesn't mean every expert agrees, and it doesn't mean the question is closed forever. What it actually means is something more precise and more interesting: that a claim has survived sustained, adversarial scrutiny from people who were genuinely trying to disprove it, and that no credible competing explanation has emerged to take its place. The philosopher of science Imre Lakatos had a useful way of thinking about this. Every scientific field, he argued, operates around a 'hard core' of assumptions that are treated as settled — not because they're beyond question in principle, but because the entire research programme depends on them. You don't get to restart from scratch every time you enter the lab. Consensus is what makes normal science possible. It's the shared ground everyone builds from. The trouble is that hard core and truly settled are not the same thing. Consensus shifts — sometimes slowly, sometimes dramatically — when the weight of anomalies becomes too heavy to ignore. The shift from continental drift being fringe speculation to accepted geology took decades and required new instrumentation, not just new arguments. This is why 'the science is settled' as a rhetorical move is both defensible and slightly misleading: settled for practical purposes is not the same as settled in principle. Knowing the difference is what separates scientific literacy from scientific deference.
In the World
In the early 1980s, Barry Marshall, an Australian physician, became convinced that stomach ulcers were caused by a bacterium — Helicobacter pylori — not by stress or excess acid, as the settled consensus held. The medical establishment didn't just disagree with him; they found the idea nearly absurd. The stomach's acidic environment was thought to be sterile by definition. No serious pathogen could survive there. Marshall struggled to get his work published or taken seriously at conferences. The consensus was so entrenched that it shaped what counted as worth testing. Eventually, in an act that has become one of medicine's more dramatic footnotes, he drank a petri dish of H. pylori, developed gastritis, and then treated himself with antibiotics. He wasn't just making a scientific argument; he was staging a demonstration forceful enough to pierce the consensus. By 2005, Marshall and his colleague Robin Warren received the Nobel Prize in Physiology or Medicine. The ulcer consensus collapsed not because a vote was taken, but because the evidence accumulated to the point where the old framework cost more than it explained. This story is sometimes told as proof that consensus is always suspect. It's actually proof of the opposite: the system worked. It was slow, it was resistant, it was frustrating — but it eventually gave way to better evidence. That's what distinguishes scientific consensus from ideology: it has a breaking point, and it's designed to.
Why It Matters
How you hold the idea of consensus changes how you engage with it. If you treat it as a headcount, you'll be manipulated by anyone who can manufacture the appearance of a split — and there are entire industries built on doing exactly that. If you treat it as infallible decree, you've turned science into an authority rather than a method. The more useful frame is to think of consensus as a live record of the best-tested explanation available right now, held provisionally but not timidly. Strong consensus on, say, vaccine safety or human-caused climate change isn't strong because lots of scientists believe it — it's strong because the mechanisms are understood, the predictions have held, and decades of attempts to find a better explanation have failed. That framing gives you something the simpler versions don't: the ability to take consensus seriously without surrendering your judgment entirely. You can ask 'how robust is this?' rather than just 'do most experts agree?' And when someone tells you the consensus is wrong, you can ask the right question back: 'What would it take for the evidence to look like this if the consensus were actually correct?' Often, the answer is revealing.
A Question to Ponder
Is there a belief you hold because of consensus that you've never actually examined — and if that consensus shifted tomorrow, would you even hear about it?
Get a new one of these every morning.
Start learning with Thinkable