There is a hole at the heart of Roman civilisation. The empire that built aqueducts, codified law and taxed half the known world had no way of writing nothing. Roman numerals can express four thousand or four million, clumsily, but ask a Roman clerk to record zero and he had to fall back on a word — nulla, nothing — because no symbol existed. The gap was not a quirk; it was a handicap. Without zero there is no true place-value arithmetic, which is why Roman calculation leaned so heavily on the counting board and the abacus, and why long division in Roman numerals remains a punishment no schoolchild has ever deserved.
Zero had to be invented, and it was invented more than once. The earliest zero-like symbol belongs to Babylon. Mesopotamian scribes had used a base-60 place-value system since the early second millennium BC, but for centuries they simply left an empty gap where a modern writer would put a zero — with predictable ambiguity, since 61 and 3,601 could look identical. By around 300 BC, in the Seleucid era, scribes were marking the empty column with two small slanted wedges. It was punctuation, not a number: the sign never stood alone, never ended a number, and nobody ever added or multiplied with it.
Half a world away, the Maya did the same thing entirely independently. Their Long Count calendar, in use in the early centuries AD, ran on a base-20 place-value system with a genuine zero, often drawn as a stylised shell. It is one of history's great independent inventions — and one of its great dead ends, because Mayan mathematics never crossed the ocean and left no trace on the numbers the rest of the world came to use.
Brahmagupta makes nothing a number
The decisive step was taken in India. Indian mathematicians had developed the decimal place-value system, and the Sanskrit word sunya — emptiness — named the vacant place. Then, in AD 628, the astronomer Brahmagupta wrote the Brahmasphutasiddhanta and did something no Babylonian bookkeeper had attempted: he treated zero as a number in its own right, with rules of arithmetic. A number subtracted from itself, he wrote, gives zero; zero added to a debt leaves a debt; zero added to a fortune leaves a fortune. In the same work he laid out rules for negative numbers, framed in that same merchant's language of debts and fortunes. He stumbled only at the last hurdle, declaring that zero divided by zero is zero — division by zero would not be properly tamed until the age of calculus, a thousand years later.
The physical evidence survives. A wall inscription in a temple at Gwalior in central India, dated to AD 876, records a garden measurement using a small round zero that any modern reader would recognise. Older still may be the Bakhshali manuscript, a birch-bark mathematical text unearthed in 1881 in what is now Pakistan, which marks zero with a dot; radiocarbon tests published in 2017 suggested parts of it might date from the third or fourth century AD, though scholars dispute this, because the manuscript's folios returned wildly different ages.
The long road to the ledger
From India, zero travelled west with scholarship and trade. In ninth-century Baghdad the mathematician al-Khwarizmi described the Hindu system of reckoning; his own name, mangled by Latin translators, gave Europe the word algorithm. Sanskrit sunya became Arabic sifr — empty — which Latin writers rendered as zephirum, which Italian wore down to zero. The same Arabic root, borrowed a second way, produced cipher, a word that meant zero long before it meant secret code: a fair measure of how occult the little circle seemed to medieval Europe.
The great populariser was Leonardo of Pisa, better known as Fibonacci, whose Liber Abaci of 1202 introduced the nine Indian figures and the sign 0 to European merchants. The reception was wary. In 1299 Florence banned the new numerals from guild account books, partly on the grounds that they invited fraud — a 0 could too easily be inflated into a 6 or a 9, or have extra digits quietly appended. Respectable accounts were to be kept in Roman numerals or written out in words. For two centuries the abacists, loyal to the counting board, fought a rearguard action against the algorists with their pen-and-paper arithmetic. Commerce settled the argument. Double-entry bookkeeping, codified in print by Luca Pacioli in 1494, made the ledger a machine for finding zero: the whole method rests on debits and credits cancelling to nothing. The banker's balance is Brahmagupta's arithmetic of debts and fortunes, institutionalised.
One absence still trips us up. The medieval monks who devised our year-numbering had no zero to work with, so the calendar jumps straight from 1 BC to AD 1 — which is why purists insisted the third millennium began in 2001, and why astronomers, needing sane arithmetic, quietly use a year zero of their own. From the empty column of a Babylonian scribe to the empty balance of a bank ledger, the story of zero is the story of civilisation learning to count what is not there — and finding that nothing, properly written down, changes everything.
Quiz nuggets
- Babylonian scribes of the Seleucid era, around 300 BC, marked empty places with two slanted wedges — a placeholder, never a true number.
- Brahmagupta's Brahmasphutasiddhanta of AD 628 gave the first rules for arithmetic with zero, though it wrongly declared zero divided by zero to be zero.
- A temple inscription at Gwalior, India, dated AD 876, contains one of the oldest surviving recognisable written zeros.
- Arabic sifr, meaning empty, gave English both zero (via the Latin zephirum) and cipher.
- Florence banned Hindu-Arabic numerals from guild account books in 1299, partly for fear that a 0 could be doctored into a 6 or a 9.