Long division

In 1299, the city of Florence banned Arabic numerals from commercial ledgers and contracts. The official objection was fraud prevention: the digits 0 through 9 were easier to alter than Roman numerals. The practical effect was to preserve the city's professional class of abacus masters, whose expertise depended on merchants not being able to calculate for themselves. By 1491, Filippo Calandri was publishing long division algorithms in Florence itself in a printed arithmetic textbook called the Pitagora aritmetice introductor. Two centuries of institutional resistance had collapsed under commercial pressure.

Long division is a specific answer to a specific problem: how do you divide one large number by another using only pen and paper? The algorithm works by decomposing the problem. Take the dividend's leftmost digits — as many as needed to be larger than the divisor. Estimate the largest integer that fits. Multiply back, subtract, bring down the next digit. Repeat until nothing remains. Each step reduces a hard problem to a simpler one by handling one place value at a time.

The algorithm itself predates Calandri by a thousand years. The Sunzi Suanjing, a Chinese mathematical text written around 400 CE, describes a digit-by-digit division process on a counting board that is structurally identical to modern long division. Al-Khwarizmi performed equivalent calculations in Baghdad around 825 CE. Indian mathematicians had parallel methods. Renaissance Italian merchants needed it independently. The same procedure was discovered repeatedly, on different continents, by people who had no contact with each other, because any culture that adopts positional notation tends to arrive at the same algorithm. Positional notation makes it possible. The algorithm makes it practical.

Before long division standardized the process, European mathematicians used the galley method (also called the scratch method or the vessel method). The galley method produces the same result with less written work: partial products are written below the dividend and crossed out as they are superseded. The final layout resembles the hull and masts of a sailing ship, hence the name. It was the dominant technique across the Arab world and medieval Europe for seventeen centuries — far longer than long division has been the global standard.

The galley method lost not because it was mathematically inferior but because it was typographically impossible. The crossed-out digits that make galley division compact require cancellation marks that early printing presses with movable type could not easily reproduce. Long division, which requires no crossed-out digits — just clean columns of numbers written and subtracted in sequence — transferred perfectly to print. The Italian arithmetic textbooks printed in the 1470s through 1490s standardized long division because long division was what their medium could handle. A method used for seventeen centuries was displaced in a generation by a method that looked the same on the printed page as it did on paper.

Electric organs in fish evolved independently at least six times in separate lineages: electric eels in South America, electric rays in the Atlantic, weakly electric mormyrids in Africa, and several others. No two lineages share a common electrogenic ancestor. Each group modified different muscle or nerve tissue to produce electrical discharge by a different cellular mechanism. The convergence is explained by physics: there are only a limited number of ways to generate electric fields with biological tissue, and selection pressure from the same environmental problem tends to converge on the same structural solutions. The long division algorithm was invented the same way. China, India, the Arab world, and Renaissance Italy each needed a method to divide large numbers by hand. Each adopted positional notation. Each arrived at the same digit-by-digit algorithm independently, because positional notation makes it the obvious solution. The algorithm is not an insight. It is the natural consequence of the notation.