What a Dropped Coin Proves
People have been losing small change for as long as there has been small change. It falls through a gap in the floorboards, out of a hole in a bag, into the mud of a road, down the back of everything.
Nobody notices. Nobody writes it down. And that is why coins are one of the most useful things anyone digging can find.
A coin is an object with a name on it
Almost nothing else from the ancient world introduces itself.
A cooking pot does not say who was in charge when it was made. A wall does not carry a date. A coin carries a ruler's name, usually a portrait, often a string of titles, and it was struck in a workshop that was keeping track. Coins were made in enormous numbers, so they are common. They are small, so they travel. They are hard, so they survive.
A named, datable, portable, tough, common object. There is nothing else quite like it in the ground.
The floor, not the ceiling
Here is the thing everyone gets backwards the first time.
You are digging, and in one layer of earth you find a coin struck in the year 250. What have you just learned about the layer?
The tempting answer is the layer is from about 250. It is wrong, and the reason it is wrong matters more than the answer.
Walk it through one step at a time. Watch which claims survive and which do not.
- A coin sits in a layer of earth. It was struck in the year 250 — that part is not in doubt, because the coin says so.
- The layer formed around the coin. So the layer cannot have been there before the coin existed. It cannot be from 200, or 100, or 20.
- That is a real and solid finding. You have put a floor under the date. The layer is from 250 or later. Call it the not-before date.
- Now try the other direction. Is the layer from 250 or shortly after? Careful. Coins do not vanish the year after they are made.
- A coin can stay in use for decades. This one may have been struck in 250, carried around for forty years, and dropped in 290 by somebody who had had it for ages.
- Worse: it could have been kept as a curiosity, or found by someone else and lost again. There is no upper limit at all. The coin sets a floor and no ceiling.
- So the honest answer is: this layer is from 250 or later. Not "about 250". The floor is knowledge. The ceiling was never on offer.
Check it in your own house
If there is loose change anywhere in the house, look at the dates on it.
Some coins will be from this year or last. Others will be surprisingly old — ten years, twenty, sometimes more. They have been in circulation the whole time, going in and out of tills and pockets and jars, and they are still perfectly good money.
Now imagine dropping one down a drain today. In a thousand years somebody lifts the drain cover and finds a coin dated twenty years before you were born.
They would be right to say the drain is at least that old. They would be quite wrong to say that is when it was blocked.
One coin, or a pattern of them
A single coin a long way from home is a good story and a weak argument.
It proves contact of some kind, at some point, by some route. It does not tell you whether that route was busy or whether one object drifted along it over a century, changing hands a dozen times.
What changes the argument is numbers. Roman gold and silver coins have turned up in southern India in large quantities, in hoards, again and again. One coin there might be a souvenir. Thousands of them, in the same region, over the same stretch of years, is a trade route — and it fits neatly with Roman writers grumbling about how much silver was leaving the empire to pay for pepper and other goods from the east.
The same reasoning maps power. Where a particular ruler's coins turn up thickly, that ruler was probably paying people: soldiers, officials, suppliers. Where they turn up as one-offs, that is trade or travel, not control. The pattern is the evidence, never the single find.
A layer of earth contains a coin struck in the year 250. What does that establish?
- The layer formed in about the year 250
- The layer formed no earlier than 250, and could be very much later
- The layer formed no later than 250
- The coin was dropped within a year or two of being made
The trap underneath all of it
There is one more way to get this wrong, and it is the reason nobody dates a layer from a single coin if they can help it.
The floor only holds if the coin arrived when the layer did. Dig a pit through old ground and the earth you throw back in is full of older material, so a very old coin can end up in a much younger layer. It still gives a floor — just a uselessly low one, hundreds of years below the truth.
So the useful rule is not find a coin, read the date. It is: find several, take the latest one, and check whether the ground has been disturbed. The latest coin in an undisturbed layer gives the tightest floor anyone can offer.
That shape of answer — solid in one direction, silent in the other — turns up all over this work. A beam with its outer rings trimmed away gives exactly the same half-answer. Knowing which half you are holding is not a technicality. It is the whole job.