Origin of money
A Socratic walk-through of the origin of money — reasoned out one step at a time, not lectured.
The question we started with
THE QUESTION #How did people get from swapping goods to accepting tokens worth nothing in themselves?
Almost everyone has heard the story. Once there was barter; barter was clumsy; so people hit upon money, first as a useful commodity and later as a token. It is a satisfying story because it explains something strange — why we accept objects of no use to us — by showing the problem they solve.
But notice what kind of story it is. It reasons backwards from money's usefulness to money's history, and assumes the second follows from the first. Is that safe? A thing can be well suited to a job it was not invented for. So we should ask two questions and keep them apart: what problem does money solve, and where did money actually come from?
Reasoning it through
REASONING #Take the logical question first, because it is the cleaner one.
Suppose you have surplus grain and want a pot. You must find someone who has a spare pot and wants grain and wants it now, in roughly the quantity you have. Jevons named this the double coincidence of wants, and the trouble is combinatorial: every additional good you might want multiplies the number of pairings that must line up.
Is it only a matching problem? Push further and a second difficulty surfaces. How much grain is a pot worth? In a barter world every good must be priced against every other, so a hundred goods need thousands of separate exchange ratios, each negotiated fresh.
Now suppose one good is singled out — one everybody will accept, not because they want it but because they know others will. What happens? The matching problem collapses: you no longer need someone who wants grain, only someone who wants the common good, and everyone does. That is money as a medium of exchange. And the pricing problem collapses too: instead of every good against every other, each good needs one price, quoted in the common good. That is money as a unit of account, and it is arguably the deeper of the two — a shared ruler for value, which makes comparison, accounting, taxation and debt all expressible.
So the logic is sound. Money economises on coincidence and on comparison. Does it follow that this is how money arose?
Here we have to be careful, because this is where the received story is weakest. Anthropologists have looked for the barter economy the story requires — a community trading goods directly, awaiting its monetary innovation — and have not found one. Caroline Humphrey's much-quoted summary is that no example of a barter economy, pure and simple, has ever been described. What ethnography finds inside communities is not spot barter but obligation: gifts, tribute, reciprocal claims, running tallies of who owes whom, settled loosely and often left open on purpose. Direct barter shows up mainly between strangers who expect no future dealings, or in societies where a money system has collapsed — as an alternative to money, not a precursor.
And the earliest written records point the same way. The Mesopotamian tablets of the late fourth and third millennia BC are overwhelmingly administrative: quantities owed, allocated and reconciled, with obligations expressed in weights of silver that frequently never moved. So a unit of account is documented long before anything like a circulating coin — struck coinage appears in Lydia and Ionia around the seventh century BC, some two and a half millennia later.
David Graeber pressed this into a general argument: that credit and the recording of debt came first, and that the barter-to-money sequence is a just-so story economists inherited from Adam Smith and never checked. The strong form of that claim is contested — historians dispute parts of Graeber's reading of particular societies — but the negative finding underneath it is not seriously disputed. Nobody has produced the barter stage.
The analogy
THE ANALOGY #Think of a language and a dictionary. A dictionary is enormously useful for fixing what words mean, and you could tell a tidy story in which people, tired of misunderstanding one another, sat down and compiled one. But nobody did. The dictionary came thousands of years after the speaking, written down to record a practice already running — and it standardised the practice rather than starting it.
Words are not scarce and cannot be hoarded, whereas a unit of account is usually pinned to something that is — which is precisely why states, temples and creditors have always fought over who defines it, a struggle with no equivalent in lexicography.
Clarifying the model
THE MODEL #The correction here is not that the textbook account is wrong so much as that it answers a question it was not asked. The coincidence-of-wants problem is real and money genuinely solves it; that is a claim about function, and it stands. What does not stand is reading it as a chronology.
Two refinements follow. First, the functions did not arrive together. A shared measure of value can exist with no object circulating at all — Mesopotamian silver-weights, or cattle used as a reckoning unit in societies that did not hand cattle over at every transaction. Medium of exchange is a later and more demanding achievement, because it requires not just agreement on a ruler but general confidence that the next person will accept the thing.
Second, this reframes the "tokens worth nothing" puzzle. If money began as a commodity that was later abstracted, then a valueless token is a strange late departure needing explanation. If the unit of account came first, then the token is not a degraded commodity at all — it is the natural form for something that was always a record of a claim. What makes it work is not what it is made of but that others will take it, and behind that, generally, that some authority will accept it in settlement of what you owe.
Two honest caveats. Money's third textbook function — store of value — is left aside here, and so is the whole machinery of credit, which has its own account. And the evidence is thin in exactly the period that matters most: writing appears alongside accounting, so what came before accounting is inferred rather than read.
A picture of it
THE PICTURE #How to readRead down the columns as a sequence of what is attested, not of what is inferred. The striking feature is the gap: the unit of account — value expressed in a common measure — is documented from the earliest written records, while coinage, the thing most people picture when they hear "money", appears roughly two and a half millennia later. The barter stage the textbook story requires would belong at the top, and it is absent because no such stage has been found.
What became clearer
WHAT CLEARED #Money's logic and money's history are two different subjects, and the familiar story fuses them. As an explanation of what money does — collapsing the search for a double coincidence of wants, and giving every good a single price on a shared ruler — the coincidence-of-wants argument is sound. As an account of what happened, it is unsupported: the record shows recorded obligation and a shared unit of account long before anything circulated, and the barter stage it depends on has never been observed anywhere.
Where to go next
ONWARD #- Why an authority's willingness to accept a token in payment of taxes underwrites its value.
- How commodity money and token money coexisted for centuries rather than succeeding one another.
Key terms
TERMS #| Term | What it means |
|---|---|
| Double coincidence of wants | the requirement, under barter, that each party wants what the other offers. |
| Medium of exchange | a good accepted in trade not for its own sake but for onward use. |
| Unit of account | the common measure in which prices, debts and accounts are expressed. |
| Commodity money | money whose material has value in its own right, such as weighed silver. |
| Fiat money | money whose value rests on acceptance and authority rather than material. |
Every term the collection defines is gathered in the glossary.