IIn the Beginning: Tokens and Writing

Accounting is an old trade. Older than money, older than the first city. As we will see, even older than writing.

The earliest accounting records are tokens from the agricultural villages of the ancient Near East, around 8000 BCE. These tokens are small clay objects shaped to represent specific commodities. A cone for a small measure of barley, a sphere for a larger one, a cylinder for livestock. To this day they turn up by the hundreds at sites in Syria, Iran, and Iraq.

According to archaeologist Denise Schmandt-Besserat, the pioneering scholar on this subject, tokens served the essential function of intermediating the redistributive economies of early farming communities. Levies set, deliveries verified, stores managed, distribution overseen. Each step required an early form of accounting, and the administrators who handled the tallies were an elite, their authority resting on a skill no one else in their communities possessed.

The token system served the economies of agrarian villages, essentially unchanged, for more than four thousand years. For this entire period, tokens had a one-to-one relationship with the commodities they represented. Each commodity was even counted in its own way, with its own number-words: one set for barley, another for oil, another for sheep, none of them running much past a dozen. There was not yet a general notion of “number” that applied to any possible object.

Then the system changed, spectacularly. The rise of the Uruk city-state in the fourth millennium BCE set off a burst of administrative innovation, and the token system expanded in several ways. New “complex” tokens appeared for the goods of urban workshops, marked with lines and shaped into forms for textiles, garments, and worked metal. Somewhat later, tokens representing debts or contracts began to be sealed inside hollow clay balls called bullae, carried with shipments or stored as records of obligation.

The bullae solved one problem but created another. They kept tokens together and under seal, but once closed, could not be verified without compromising the record’s integrity. The solution was to press each token into the wet clay surface of the container before sealing it. Three cones inside, three wedge-shaped impressions outside. The exterior carried an independent record of the interior, and any tampering became detectable. (This is the same instinct for structural, two-way verification that would eventually produce double-entry bookkeeping and the audit profession.)

With the exterior of the bullae now bearing the impressions of the tokens inside, a cognitive leap became possible: the impressions alone could carry the information. The tokens were eventually abstracted away; the mark itself represented the related good. The bullae, over time, were flattened into tablets, with the impression of the token now standing for its commodity.

This was a more portable record, but it was not yet writing. At first those marks merely mimicked the tokens, one impression per unit. A cone pressed into the clay left a wedge, standing for a small measure of barley; a sphere left a circle, standing for a large one. The count and the commodity were still bound together in a single mark.

Then, gradually, the count and the commodity came apart. Barley was the most traded commodity in the region, and its measures were so familiar that they came to count a wider and wider range of goods, until the measures broke free of barley altogether. The wedge came to mean simply one, the circle ten – they became numbers that could count anything at all. These were the first abstract numbers. It had taken four thousand years.

But a number alone is useless to an accountant: one of what? So beside it the scribe drew a small picture of the thing counted. A certain sign for oil, another for sheep, and so on. The impressed mark represented the quantity; the drawn picture represented the object. Together, these formed a pictograph capable of communicating increasingly complex information.

Those pictographs, inscribed onto clay accounting tablets, are the first writing.

Schmandt-Besserat attributes the breakthrough to the accountants of the Uruk administration – the unnamed clerks whose records became the basis of writing.

Writing emerged from accounting.

IIThe Historical Pattern

This is the pattern the rest of this essay traces. Each leap in economic complexity demands a leap in accounting, and each leap in accounting opens economic possibilities that have not existed before.

In Ancient Egypt, state power was built upon the scribes who tracked the royal granaries, measuring the Pharaoh’s wealth in stores of barley and wheat. The system was built for verification: grain was weighed as it arrived and the tally was checked against the granary’s standing totals.

Millennia later in Rome, accounting was a routine part of civic life. A propertied head of household maintained an “adversaria” – a wax daybook for real-time recording – along with a “codex accepti et expensi” – a bound ledger that recorded transactions from the adversaria each month. The codex was standardized enough to serve as admissible evidence in court, and provincial officials were required to account for the public funds they handled. Rome’s administrative capacity and reach were a function of its accounting practices.

Even later, the Islamic world conducted long-distance trade at a previously unimaginable scale through its “hawala” system, where brokers kept running accounts with one another without the need to exchange physical currency on each transaction. This only worked because brokers maintained reciprocal records, creating the infrastructure for trust.

The innovations we see across Egypt, Rome, and the Islamic world are merely instances of the broader pattern: accounting technology can be a catalyst for economic possibility.

And by the late medieval period, European merchants were beginning to press against the ceiling of what their current accounting systems could handle. Commerce spanning the continent made it clear that the old recording system was insufficient; a new tool was needed to manage the explosion in economic volume.

It arrived in the merchant cities of northern Italy in the 14th century and would eventually bring about what we now know as capitalism.

IIIDouble-Entry and the Birth of Capitalism

“Capitalism and double-entry bookkeeping are absolutely indissociable; their relationship to each other is that of form to content.”Werner Sombart, Der moderne Kapitalismus (1916)

In 1494 the Franciscan friar Luca Pacioli published Summa de arithmetica, geometria, proportioni et proportionalità, a 600-page work on mathematics. Tucked within was a 27-page section containing the first complete printed description of double-entry bookkeeping.

The mechanism is elegant: every transaction recorded twice – as a debit and as a credit – with the two sides always in balance. The structure surfaces errors and fraud immediately. If the books don’t balance, something is wrong, and you can trace through the entries until you find the problem. Like the Uruk bullae 4,500 years earlier, double-entry makes tampering or error visible by design.

Pacioli’s system would be surprisingly familiar to a modern accountant: journals and ledgers, containing accounts for assets, liabilities, capital, income, and expenses. He proposed that trial balances be used to prove a balanced ledger at year end. Pacioli even urged his merchant readers not to go to sleep until the day’s debits matched its credits.

Pacioli did not invent this ledger-balancing system; merchants across the Mediterranean had been using versions of it for centuries. But what Pacioli did was codify the system, and publish it at the exact time when it would have its biggest impact, as printing was first becoming widely available. Within a generation, every merchant house in Europe was running some version of double-entry.

No one could have predicted the impact of this accounting breakthrough. For the first time, merchants could look at a ledger and see a coherent picture of a business as a unified economic entity, with its own claims on the future. Capital became a real, countable stock of wealth to be tracked, taxed, lent against, and invested.

Capitalism in any meaningful sense did not exist before double-entry, because the very concept of capital required the machinery double-entry provided. Max Weber agreed that the rational pursuit of measured profit (i.e., capitalism) rested on double-entry bookkeeping.

An immediate result of this new system was the joint-stock company. In 1602, the Dutch East India Company became the first enterprise to issue tradable shares to outside investors. Within a few years the Amsterdam Stock Exchange opened as the first formal market for them. Neither would have worked without double-entry. Shareholders who did not run the company, and did not know one another, held a claim on an enterprise they would never personally see.

In the centuries that followed, double-entry spread across every realm of organized economic activity. By the 18th century it was the standard of European commerce; by the 19th it was a precondition for capital formation at any scale.

Capitalism did not appear from nothing. It was built on a technology invented by accountants in northern Italy.

IVAccounting for the Industrial Revolution

Capital could now be pooled and invested at scale; the 18th century saw that capital flow into industry. The factories and railroads that followed were a new kind of industrial business, which demanded enormous up-front capital. Firms began investing in machinery that would yield revenue over decades while simultaneously wearing down with use.

Once again, the existing tools fell short.

So the accountants of the era answered with new methods, chief among them depreciation. As American railroad accountants worked out how to spread an asset’s cost across its useful life, railroads became financially intelligible.

Cost accounting followed the same logic. As factories grew, executives needed to understand not just whether the company was profitable overall, but which products or factory locations were profitable on their own. This technique had deep roots, tracing back to the 1770s when Josiah Wedgwood, struggling to understand which of his pottery lines made money, built a costing system that charged the use of his kilns to specific product runs rather than treating them as one-time expenses. Some lines that had appeared profitable were, it turned out, losing money.

That same insight would become central to Andrew Carnegie’s empire a century later. Carnegie Steel Company accountants carefully tracked unit costs, providing the operational visibility that competitors lacked. The next generation of corporations built on this system of modern industrial management with new accounting instruments like the return-on-investment formula (DuPont) and the flexible budget (General Motors).

As these new industrial businesses grew to staggering sizes, trust became a bottleneck. A railroad owned by thousands of shareholders who would never meet its managers needed an independent party to verify the books.

The answer was the audit profession. Britain’s Joint Stock Companies Act of 1844 required new companies to publish accounts open to inspection. In 1854, the Society of Accountants in Edinburgh received a royal charter and became the world’s first professional accounting body. The American Association of Public Accountants, ancestor of today’s AICPA, followed in 1887. Their members audited what companies reported, legitimizing the numbers for outside shareholders.

U.S. accounting standards in the 1920s were largely voluntary – some firms were honest, but many were not. The result was financial statements that bore only a loose relation to reality. After the crash of 1929, the investigations that followed blamed, in large part, the absence of standardized accounting practices.

The Securities Acts of 1933 and 1934 created the SEC, which institutionalized what we now call GAAP. For the first time, one company’s financial statements had to mean approximately the same thing as another’s. Modern capital markets date from that requirement.

VThe Modernization of Accounting

Even amid these changes, an accountant in 1950 still used nearly the same tools as an accountant in Pacioli’s Venice: pen, ink, and a physical ledger. What had grown since then was the complexity of the work – and it was starting to become unmanageable.

A large American corporation could be expected to have tens of thousands of employees, dozens of subsidiaries, operations in many countries, and an obligation to roll everything into a single set of financial statements. Armies of clerks worked nights and weekends to close the books.

When computers first started being used for business, accounting work got faster but also remade. The first commercial computer installed for business in America was a UNIVAC, delivered to General Electric in 1954 to automate payroll. Within a decade, most large corporations were running payables, receivables, and inventory on mainframes. The clerks who had spent careers posting entries by hand now found themselves learning to type; work that had taken weeks now took days.

An arguably bigger shift came in 1979, when Dan Bricklin wrote a program called VisiCalc for the Apple II. It let an accountant lay out a grid of numbers with formulas linking them, and the machine recalculated everything the instant one of the inputs changed. By the mid-1980s, Lotus 1-2-3 and then Excel had made the spreadsheet the standard tool of business analysis.

The enterprise resource planning systems of the 1990s (SAP and Oracle) collapsed a global company’s many ledgers into a single system of record. Now, a shipment or an approved invoice posted its own entry the moment it happened.

Then came the internet, and cloud software, and with them kinds of companies that had never existed before. Accountants spent three decades adapting: new rules for recognizing software revenue, then stock compensation, then financial instruments no one had imagined a generation earlier.

For a sense of the scale of modern accounting, look at the supply chain for a single semiconductor. An advanced chip is the product of thousands of suppliers across dozens of countries, with components crossing the world several times before final assembly. Every transaction across that network is recorded; every contract is enforceable because the obligations beneath it can be audited. The chip in a modern laptop exists because that entire economic web is legible, and it is legible only because of accounting.

This is what we owe to the preceding era of accounting. And now even these systems are starting to strain.

Internal controls are failing. Public companies are restating earnings more often. The Pentagon has failed every annual audit since federal law began requiring them. The profession itself is shrinking, with experienced accountants retiring faster than new ones arrive, just as the demand for accounting continues to accelerate.

VIThe Horizon

The pattern that began with clay tokens in Uruk has not stopped, and it likely never will. Each great expansion of economic complexity has demanded a corresponding expansion of accounting, and each expansion of accounting has in turn created new economic realities. We are at the start of another such moment – it may be the largest yet.

AI agents are beginning to enter the economy as economic actors, able to negotiate contracts and execute transactions alongside humans. All this activity will need to be measured, verified, and acted upon.

The accountant of the next era will work differently. The repetitive parts of the work (data entry, transaction matching, reconciliations, and tie-outs) will go to agents that run on their own and deliver finished work for a person to review.

The accountant’s role shifts from doer to reviewer, from living in workpapers to managing teams of agents. The skills shift with it. Judgment and relationship building will become more valuable.

The Mesopotamians who first pressed tokens into bullae did not realize they were inventing writing, nor did the Italian merchants who refined double-entry know they were enabling capitalism. None of them could see, from where they stood, what great upheavals would come from their work.

Neither can we predict what will arise from the new era of accounting.

Yet the pattern is clear, and it is hopeful: each evolution of accounting has produced an economic reality the previous generation would have found impossible to imagine. We will not be the exception.