In 1980, Luis and Walter Alvarez proposed that a massive asteroid impact wiped out the dinosaurs 66 million years ago. No one observed this event. No one can rerun it in a laboratory. Yet today, the impact hypothesis is among the best-supported explanations in all of science. How is that possible?

Sciences like geology, paleontology, and evolutionary biology face a fundamental challenge: they study events that happened once, long ago, and can never be repeated. This raises a deep philosophical question. If science depends on experimentation and repeatability, do historical sciences count as real science? The answer reshapes how we think about what scientific knowledge actually requires.

Experimental sciences work by manipulating variables and observing outcomes. You change one thing, hold everything else constant, and see what happens. Historical sciences can't do this. A geologist studying the formation of the Grand Canyon cannot rewind time and watch it happen. An evolutionary biologist cannot rerun the Cambrian explosion under controlled conditions.

Instead, historical scientists practice what philosophers call retrodiction—reasoning backward from present evidence to past causes. The world is full of traces: fossils embedded in rock layers, isotopic ratios in ice cores, genetic sequences shared across species. Each trace is like a clue at a crime scene. The event itself is gone, but it left marks on the world that persist. The scientist's job is to read those marks and reconstruct what must have happened to produce them.

This isn't guesswork. Historical inference follows rigorous logic. If a particular past event occurred, it would have left specific, predictable traces. Scientists identify those expected traces and then check whether they actually exist. When the asteroid impact hypothesis was proposed, it predicted a global layer of iridium-rich clay at the geological boundary marking the extinction. That layer was found on every continent. The reasoning is as disciplined as any experiment—it simply runs in a different direction.

Takeaway

Science doesn't require witnessing an event or recreating it in a lab. It requires that proposed explanations make specific, checkable predictions about what evidence we should find—even when that evidence points backward in time.

A single piece of evidence can usually be explained in more than one way. A layer of iridium might come from an asteroid—or from unusual volcanic activity. A fossil in an unexpected location might indicate migration—or it might have been transported by geological processes after death. Any one clue is ambiguous. This is why historical scientists rely on consilience: the convergence of multiple independent lines of evidence pointing toward the same conclusion.

Consider the asteroid impact hypothesis again. It's not supported by the iridium layer alone. Scientists also found shocked quartz—mineral grains deformed by extreme pressure—at the same geological boundary. They discovered the Chicxulub crater buried beneath Mexico's Yucatán Peninsula, dated to precisely the right time. They found tiny glass spherules, formed when molten rock was blasted into the atmosphere and cooled mid-flight. Each line of evidence is independent. Each could theoretically be explained away on its own. But the chance that all of them coincidentally point to the same event, if that event didn't actually happen, becomes vanishingly small.

This convergence strategy serves the same logical function as controlled experimentation. An experiment isolates a cause by eliminating alternatives. Consilience eliminates alternatives by showing that only one hypothesis accounts for all the diverse evidence simultaneously. It's a different method, but it achieves the same epistemic goal: ruling out competing explanations until one remains standing.

Takeaway

When you can't run an experiment, you let reality run it for you. Independent lines of evidence converging on the same conclusion can be just as powerful as a controlled test—because coincidence has limits.

Physics and chemistry typically explain by citing general laws. Water boils at 100°C at sea level because of laws governing phase transitions. The explanation works for any sample of water, anywhere, anytime. Historical sciences explain differently. They construct narratives—sequences of particular events, each one contingent on what came before, that together account for a specific outcome.

The explanation of the dinosaurs' extinction isn't a general law about asteroids and mass death. It's a story: a specific object of a specific size struck a specific location at a specific time, triggering a specific chain of environmental catastrophes—firestorms, darkness, cooling, acid rain—that cascaded through ecosystems in particular ways. Change any detail and the outcome might have been different. This is not a weakness. It's the nature of the subject matter. Historical events are shaped by contingency, by the particular arrangement of circumstances at a given moment.

Philosopher of science Carol Cleland has argued that narrative explanations are not inferior versions of law-based explanations. They are a distinct and legitimate form of scientific understanding, suited to a world where singular, unrepeatable events have real causal power. Understanding why something happened sometimes requires telling the story of how it happened, step by step. The narrative is the explanation.

Takeaway

Not all scientific explanation looks like a physics equation. Sometimes understanding the world means telling the right story—one where each event follows from the last and no single law captures the whole picture.

Historical sciences are not second-class citizens in the scientific community. They employ rigorous inference, demand convergent evidence, and offer genuine explanations—just not always in the form of universal laws and repeatable experiments.

Recognizing this matters beyond academia. It shows that science is broader and more flexible than the stereotype of lab coats and controlled variables suggests. The past is gone, but it is not beyond the reach of careful reasoning. Science can know what it cannot see.