Most people assume historical facts are straightforward. Caesar crossed the Rubicon. The Berlin Wall fell in 1989. Napoleon died on St. Helena. These statements seem as solid as mathematical truths—either they happened or they didn't.

But philosophers of history have long recognized something unsettling: the concept of a historical fact is far more slippery than it appears. What exactly are we referring to when we call something a historical fact? The event itself, which no longer exists? The evidence we have for it? The statement we construct about it? Each answer leads to different—and troubling—implications for what historians can actually claim to know.

This isn't merely an academic puzzle. The epistemological status of historical facts determines what kinds of arguments historians can legitimately make, what standards of evidence apply, and whether historical disputes can ever be genuinely resolved. Understanding why historical facts are constructed rather than simply found transforms how we read history—and how we evaluate competing claims about the past.

Here is the fundamental confusion that trips up most discussions of historical knowledge: we use the word 'fact' to mean two entirely different things. Sometimes we mean the actual occurrence—the event itself, happening in real time in the past. Sometimes we mean the true statement we make about that occurrence.

The philosopher Michael Oakeshott made this distinction with characteristic precision. The event—say, the assassination of Julius Caesar—was a singular occurrence in 44 BCE. It happened once, in its full chaotic complexity, with all its causes and consequences radiating outward. That event no longer exists. It cannot be observed, measured, or directly accessed by any means available to us.

What we call a 'fact' about Caesar's assassination is something quite different: it is a statement we make in the present, using present concepts and language, about that vanished occurrence. The fact 'Caesar was assassinated on the Ides of March' is a linguistic construction. It selects certain features of the event as relevant, ignores others, and frames what happened in terms meaningful to us.

This means facts are not lying around in the past waiting to be discovered, like fossils in rock. They are produced through an active process of selection and formulation. The past supplies the raw material—the events that actually occurred—but the facts are our creation. We make them by asking questions, examining evidence, and constructing statements we believe to be true.

The implications are significant. If facts were simply 'out there,' historical disagreements would be merely technical—better research would resolve them. But if facts are constructed, then the frameworks we bring to history shape what facts we can even formulate. Different questions generate different facts from the same past.

Takeaway

Historical facts are statements we construct in the present about a past that no longer exists—they are products of inquiry, not objects waiting to be found.

If we cannot directly observe past events, how do we arrive at facts about them? Through inference from evidence. This might seem obvious, but the inferential nature of historical knowledge has profound consequences that are often overlooked.

Evidence consists of traces—physical remnants, documents, artifacts, and testimonies that survive from the past into the present. A medieval charter, a Roman coin, an eyewitness account recorded decades later. These traces are not facts; they are the raw material from which facts must be constructed through a chain of reasoning.

The historian examines a document and infers what it tells us about past events. But this inference is never automatic or unmediated. We must consider the document's provenance, authenticity, and reliability. We must interpret its language in historical context. We must assess what the author knew, intended, and might have misrepresented. Each step involves judgment.

R.G. Collingwood described this process as the historian acting like a detective, constructing an account of what happened by reasoning from clues. But unlike a detective, the historian cannot interrogate witnesses, revisit the crime scene, or conduct experiments. The evidence is fixed and incomplete. Large portions of the past left no traces at all.

This inferential gap means that historical facts always carry a degree of uncertainty that empirical sciences avoid. A physicist can repeat an experiment; a historian cannot repeat the past. Our facts are conclusions drawn from fragmentary evidence through fallible reasoning. They may be well-supported, even beyond reasonable doubt, but they remain inferences—not observations.

Takeaway

Historical facts emerge through inference from surviving traces, not direct observation—every fact carries the epistemic weight of the reasoning that produced it.

Not all historical statements occupy the same epistemic ground. Recognizing different levels of factuality provides essential tools for evaluating historical claims and understanding where legitimate disagreement is possible.

At the base level are what we might call hard facts—statements so well-attested by multiple independent sources that denying them requires rejecting the entire evidential framework. That World War I occurred, that Lincoln was assassinated, that the Black Death devastated medieval Europe. These facts are not beyond all possible doubt, but doubting them means doubting historical knowledge as such.

Above these sit soft facts—statements well-supported by evidence but more dependent on interpretation. The specific number of casualties at a battle, the exact date of a document's composition, the route of a medieval journey. These facts can be contested by new evidence or alternative readings of existing evidence without challenging the broader historical framework.

Higher still are interpretive claims—statements about causes, meanings, significance, and patterns. That the Reformation was primarily about religious conviction rather than political economy, that the French Revolution marked the birth of modern politics, that industrialization improved overall living standards. These are not straightforwardly 'factual' in the same sense; they involve judgment about how to weight and connect lower-level facts.

Confusion arises when we treat all three levels as equivalent. Critics of historical objectivity sometimes point to interpretive disagreements as evidence that history is 'just opinion.' Defenders of objectivity sometimes claim more certainty for interpretive claims than the evidence warrants. Keeping these levels distinct clarifies where genuine knowledge is possible and where reasonable people may legitimately differ.

Takeaway

Historical statements range from virtually undeniable hard facts to contested interpretive claims—epistemic clarity requires recognizing which level any given statement occupies.

The question 'What is a historical fact?' turns out to be a gateway into the deepest problems of historical epistemology. Facts are not transparent windows onto the past but constructed statements, built through inference from incomplete evidence, carrying varying degrees of certainty depending on their level of generality and interpretation.

This should not lead to despair about historical knowledge. Recognition that facts are constructed does not mean they are arbitrary or that all constructions are equally valid. Some inferences are better supported than others. Some interpretations fit the evidence more coherently. The constructed nature of facts simply means we must be epistemologically humble—aware of the processes that produce our knowledge.

What we gain is precision. We can distinguish between what we know with near-certainty and what remains genuinely contested. We can recognize where new evidence might change our conclusions and where disagreements reflect different interpretive frameworks rather than factual errors. Understanding what historical facts are is the first step toward understanding what we can legitimately claim to know.