Somewhere around 2700 BCE, in the cities of Mesopotamia, someone arranged a set of pebbles on a marked clay tablet and realized that the position of a stone could carry meaning beyond its presence. This was the counting board—the first abacus. It is still in use today.
No tool in the history of calculation has survived as long, adapted as thoroughly, or resisted obsolescence as stubbornly as the abacus. It outlasted the slide rule, the mechanical calculator, and the first generation of digital computers. In Japan, it is still taught in schools.
The Ancient World: Counting Boards
The earliest counting devices were not frames with beads but flat surfaces with loose counters. The Mesopotamian counting board used clay tablets with marked columns; the operator moved pebbles between positions to represent different powers of ten. The Latin word calculus—from which we get "calculate"—means pebble.
Rome had a more sophisticated version. The Roman hand abacus, examples of which survive in bronze, used slotted grooves with sliding counters. It could represent numbers in the millions and handle fractions for the monetary system. Roman merchants, tax collectors, and engineers used it throughout the empire's history.
The Greek abakos (from which the English word derives) referred to a counting surface, typically a wooden or stone board strewn with sand on which figures could be drawn and erased. The Salamis Tablet, discovered on the Greek island of Salamis in 1846, is a marble counting board dating to roughly 300 BCE—the oldest surviving abacus of any kind.
The Chinese Suanpan and Japanese Soroban
The framed bead abacus—the version most people picture—was developed in China, where it is called the suanpan (算盤, "calculating tray"). Its earliest clear documentary reference appears in texts from the 2nd century CE, though the device itself is likely older.
The suanpan has a distinctive two-deck structure: two beads above the dividing bar, five below. This arrangement can represent values beyond nine per column, accommodating hexadecimal arithmetic used in Chinese units of weight and currency. In practice, the extra beads function as carries, allowing operations without resetting columns mid-calculation.
Japan adopted the suanpan sometime during the 14th or 15th century, calling it the soroban (算盤). Japanese mathematicians then simplified it. The modern soroban uses one bead above the bar and four below—exactly enough to represent the digits zero through nine, no more. This reduction made the soroban faster for decimal arithmetic. The unnecessary beads were eliminated, and with them the cognitive overhead of managing extra positions.
The visual difference between a suanpan and a soroban is immediately obvious once you know what to look for. The suanpan is heavier, more complex; the soroban is leaner, optimized. The two-five structure versus the one-four structure reflects different theories of what a calculating instrument is for.
Medieval Europe: The Counting Board
Western Europe had its own parallel tradition. Medieval merchants used cloth or leather counting cloths marked with lines, moving coins or tokens (called jetons in French, rechenmarken in German) between columns. The cloth could be folded and stored. The tokens doubled as monetary tokens in some contexts.
This tradition was so embedded in commercial life that the word "counter" in English derives from it—the counter in a shop was originally the flat surface on which accounts were reckoned, not merely a place to stand and transact. The word "exchequer" comes from the French eschequier, a chessboard-patterned counting cloth used by the English royal treasury.
European counting boards coexisted with written arithmetic for centuries, serving different populations with different needs. Merchants who couldn't write or calculate on paper could use the board. The introduction of Hindu-Arabic numerals gradually displaced them from educated use, though they persisted in popular commerce well into the 17th century.
1946: The Contest That Decided Nothing
On November 12, 1946, in Tokyo, a Japanese sergeant named Kiyoshi Matsuzaki sat across a table from Tom Wood, a U.S. Army private who had been selected as the best electric calculator operator in the region. Wood used a Remington Rand electric machine. Matsuzaki used a soroban. They were given identical arithmetic problems: large additions, subtractions, multiplications, and divisions.
Matsuzaki won four of the five rounds. The soroban beat the electric calculator at addition and subtraction decisively; the machine was faster at complex multiplication. Stars and Stripes, the U.S. military newspaper, covered the story under the headline "Abacus Beats Electric Calculator."
The contest was later described as proof of the abacus's superiority. In context, it was something more specific: proof that a skilled operator using a simple mechanical device could outperform a skilled operator using the best available electric calculating machine at the tasks that constituted the bulk of daily commercial arithmetic. The electric calculator was faster at tasks people rarely did. The soroban was faster at tasks people did constantly.
This was 1946. The first general-purpose electronic computers were just becoming operational. Within twenty years, electronic calculators would make the contest moot. But the contest itself revealed something important about what calculation actually requires in daily life: mostly addition and subtraction, performed quickly on numbers of modest size.
The Soviet Schoty
Russia and the Soviet Union maintained their own abacus tradition through the schoty (счёты), a distinctive design with ten beads per wire (usually colored in groups of two with the middle two beads a different color for visual orientation) on horizontal wires in a tilted frame. Unlike the suanpan and soroban, the schoty places beads on wires that tilt rather than slide vertically.
The schoty was standard equipment in Soviet shops, offices, and classrooms well into the 1980s. A visitor to a Soviet grocery store in 1985 might see a checkout clerk using a schoty to total a customer's order before announcing the price. This was not backwardness—it was reliability. The schoty required no electricity, no batteries, no maintenance, and never crashed. In a system where electronic equipment was expensive, imported, and frequently unavailable, the schoty made practical sense.
Flash Anzan and Competitive Soroban
In Japan, competitive soroban has evolved into something that barely resembles ordinary calculation. Flash anzan (フラッシュ暗算) is an event in which competitors perform mental arithmetic at speeds that seem impossible by imagining a soroban in their minds and manipulating imaginary beads.
In competition, strings of three-digit numbers are flashed on a screen, each for fractions of a second. Competitors must add fifteen or more numbers shown at 0.2-second intervals—well below the threshold at which most people can consciously read the numbers. The winners are performing what amounts to visual computation on a mentally simulated device.
Brain imaging studies have confirmed that expert flash anzan practitioners use visual-spatial processing areas of the brain during calculation, not the language areas used in conventional arithmetic. They are literally computing on an imaginary abacus, and they are faster than calculators at addition and subtraction for the same reason Matsuzaki was in 1946: the soroban encodes addition visually, and humans process visual information extremely quickly.
4,700 Years of Continuous Use
The suanpan/soroban family has been in more or less continuous use for at least 1,700 years, with precursor counting boards in use for another 2,700 years before that. If we date the abacus to Mesopotamian counting boards around 2700 BCE, we arrive at something like 4,700 years of uninterrupted use—making it almost certainly the oldest continuously used computing device in human history.
By comparison, the slide rule was invented in the 1620s and obsolete by 1975: a lifespan of about 350 years. The mechanical calculator appeared in useful commercial form in the 1880s and was largely gone by the 1970s: less than 100 years. The personal computer has existed since the late 1970s: under 50 years.
The abacus's durability is not an accident. It requires no power, no maintenance, no software updates, no network connection. It produces no errors from battery failure or corrupted memory. It is as fast as its operator, and skilled operators can be very fast. It is tactile, visual, and embodied in a way that electronic calculation is not—many practitioners describe the physical motion of beads as part of how they think.
Whether the abacus is still the right tool for any particular task is a separate question. But that a tool invented in Mesopotamia before the pyramids were built is still being actively used, taught, and competed with in 21st-century Japan says something about the relationship between tools and the tasks they were built for. Some problems are old enough that old solutions remain competitive.
More from Anethoth: builds.anethoth.com — public build dossiers for software projects in progress.