Q.Trace the route from the entrance to the inner court of the palace at Mari. What do you think would have been kept in the storerooms? How has the kitchen been identified?

Think of a crime scene. Nobody saw the thief, but the detective finds a muddy footprint, a broken window, and a half-eaten sandwich. From those leftovers, the detective builds a story: someone climbed in, got hungry, and left in a hurry. That’s the everyday version of archaeological inference.

Archaeology works the same way, except the “crime scene” is a buried village, a trash heap, or a temple ruin, and the “detective” has centuries of silence to work through. The people are gone, they left no written diary, and often no writing at all. What they did leave behind are artefacts — pots, bones, tools, coins, walls, ash layers. An artefact by itself is just a thing. A broken pot is a broken pot. But when you ask why it was made, how it was used, who used it, and what it tells us about their lives, you are making an inference.

So the precise meaning: archaeological inference is the reasoned process of drawing conclusions about past human behaviour, culture, and society from the physical remains they left behind. It is not guessing. It is a disciplined leap from the material (what survives) to the immaterial (what people believed, ate, traded, feared, or valued). The leap is guided by context — where the object was found, what lay next to it, what layer of soil it sat in — and by comparison with things we do know, either from written records of nearby cultures or from experiments in the present.

Why does this matter? Because most of human history left no written word. For the thousands of years before writing, and for the vast majority of people who never wrote even after writing existed, the only voice we have is the debris of daily life. Inference is how that debris speaks.

Here is what a good inference looks like in practice:

• You find a row of small clay pots, all identical, buried under the floor of a house. The fact is: pots under a floor. The inference might be: these were ritual offerings, because similar pots appear under thresholds in other sites, and because they were deliberately hidden, not thrown away.
• You find animal bones in a pit, but only the leg bones, and they are split open. The fact is: bones, split. The inference: people ate meat here, and marrow was valued enough to crack the bones for it — or, if the bones are from a species not local, there was trade.
• You find a layer of ash and charred wood between two layers of normal occupation. The fact is: a burnt layer. The inference: a fire, accidental or deliberate — but you cannot say which until you check whether the layer is thick and even (a planned destruction) or patchy and thin (a cooking accident that spread).

Notice the pattern. The inference is always a claim about people, never just a description of dirt. And it is always testable — another archaeologist could look at the same evidence and argue differently, which is exactly what makes the field alive.