The Limits and Frontiers of Science

Science is extraordinarily powerful at answering one kind of question: what is the case? It measures, predicts, and explains the physical world with a reach no other method matches. But that very success tempts us to ask it questions it was never built to answer — about what we ought to do, what is just, or what makes a life good.

Knowing where science's authority ends is not a criticism of science; it is part of using it well. Someone who understands the limits can lean on evidence fully where it belongs and stop pretending it decides matters it cannot. Mistaking the boundary in either direction — dismissing solid findings, or dressing a value judgment as a fact — leads people badly astray.

Questions of fact versus questions of value

A question of fact asks what is, was, or will be: how fast a virus spreads, whether a bridge will hold, what the climate is doing. Evidence can, in principle, settle it. A question of value asks what ought to be: whether a policy is just, how to weigh one person's freedom against another's safety, what we owe future generations. Science can inform such questions richly — telling us the likely consequences of each option — but the final weighing of goods rests on values that measurement alone cannot supply.

The permanent frontier

Every answered question opens new ones. Science does not shrink toward a finished map; its edge of open problems moves outward as it grows. The nature of dark matter, the physical basis of conscious experience, the origin of life — these are not failures but frontiers, live questions being worked on with real methods. A field with no open questions would be a dead one.

"We don't know yet" is an answer

In everyday talk, admitting ignorance feels like losing. In science it is a precise, respectable position: it says the evidence so far does not justify a confident claim. That honesty is exactly what lets science self-correct. The alternative — inventing an answer to fill the silence — is the move that separates inquiry from superstition.

Suppose a town must decide whether to fluoridate its water. Science can answer the factual parts: at what concentration fluoride reduces tooth decay, at what higher level it risks harm, and how the costs compare. Those are measurable, and reasonable people should accept the best evidence on them.

But whether to fluoridate also turns on questions science cannot settle: how to weigh collective dental health against individual consent, how much risk is acceptable, and who decides for everyone. Notice the two layers. The data narrows the choice and rules out claims that contradict evidence; it does not, by itself, tell the town what it ought to value. A good decision uses science for the first layer and honest deliberation for the second.

Consider someone who argues, "Studies show that policy X makes people happier, therefore we ought to require X by law." The leap hides a value premise — that we ought to maximize this particular measure of happiness, through compulsion, at whatever cost to freedom. The study may be flawless and the conclusion still unearned, because a factual finding was quietly promoted into a moral command. The error is not bad data; it is asking evidence to do a job evidence cannot do.

In A Treatise of Human Nature (1739–40), the Scottish philosopher David Hume noticed a recurring slip in the moral arguments of his day. A writer would reason for a while using ordinary statements of fact — propositions joined by is and is not — and then, without warning, every claim would suddenly be connected instead by ought or ought not. Hume observed that this change is "imperceptible" yet of "the last consequence," and pressed that a reason should be given for how the new relation could be deduced from the old. He did not declare the leap always impossible, but insisted it needs a justification that is almost never supplied.

This observation, later nicknamed "Hume's guillotine," is one of the most durable ideas in philosophy: you cannot derive an ought purely from an is. It is closely related to, though not identical with, what G. E. Moore in Principia Ethica (1903) called the "naturalistic fallacy." The distinction endures because it captures exactly what the fluoridation example shows: no pile of measurements, however tall, contains a conclusion about what we should value. Science supplies the is; the ought requires something more.

• "If science can't answer it, the question is meaningless." Value questions are real and consequential; they simply need tools beyond measurement.
• "An open question means science has failed." Live frontiers like dark matter mark a healthy field, not a broken one.
• "'We don't know yet' is a cop-out." It is an honest report of the evidence and the precondition for finding out.
• "You can read your values straight off the facts." Facts constrain and inform values, but the is–ought gap means they never fully determine them.

Take a debate you care about and split it into two columns. In one, write every part that a good study could in principle settle; in the other, every part that turns on what we ought to value. Which column holds most of the actual disagreement? Often the loudest fights are value clashes wearing the disguise of a factual dispute.

Think Like a Maester: When an argument jumps from "this is how the world is" to "this is what we must do," stop and ask which value was smuggled in to bridge the gap.

1. In your own words, what is the difference between a question of fact and a question of value?
2. Give an example of how science can inform a value question without settling it.
3. Why is an open question such as the nature of dark matter a sign of a healthy science rather than a failure?
4. What did David Hume observe about the move from "is" to "ought," and why does it matter?
5. When is "we don't know yet" the most scientific answer available, and why?

Science answers questions of fact with unmatched power, but it cannot, by itself, settle questions of value — it can only inform them. Its frontier of open problems is permanent, and "we don't know yet" is a precise, honest answer rather than a defeat. Hume's is–ought observation names the boundary cleanly: no set of facts alone yields a conclusion about what we ought to do. Using science well means leaning on it fully where it belongs and refusing to let it impersonate a moral authority it does not have.

• Measurement and Uncertainty
• How Science Changes Its Mind