The oldest mathematical object ever found is a small fragment of a baboon fibula, roughly 8 cm long, discovered in the 1970s by South African archaeologist Peter Beaumont during excavations at Border Cave in the Lebombo Mountains on the border between South Africa and what is now Eswatini. Carved into its surface are 29 distinct notches, deliberate sequential incisions made with multiple different stone-cutting edges over an extended period. The bone is roughly 42,000 years old, and 29 is almost exactly the length of a lunar cycle, 29. 53 days.
Someone, 42,000 years ago, was tracking the moon with a dead baboon and a sharp rock. That bone is not art. It is information storage. Somebody looked at the sky, noticed a pattern, and recorded it.
In 1960, Belgian geologist Jean de Heinzelin pulled a small curved bone from a layer of volcanic ash near Lake Edward in the Congo. It was a baboon fibula about 10 cm long with a chunk of quartz jammed into one end like a crude pencil. Carved into its surface were 168 notches organized into three distinct columns. The bone is roughly 20,000 to 25,000 years old, and the notch groupings are suspicious.
One column contains groups of 11, 13, 17, and 19, the four prime numbers between 10 and 20, whose sum is 60. Another column also sums to 60. The third column shows clear doubling patterns. Jean de Heinzelin argued this was an early arithmetic system.
Alexander Marshack, in his 1972 book The Roots of Civilization, argued the notches tracked lunar phases over roughly 6 months. Claudia Zaslavsky, in her 1973 book Africa Counts, proposed a menstrual calendar interpretation. The debate is unresolved, but nobody disputes that the marks are intentional, grouped, and follow a pattern. Somebody was counting something that happened in regular cycles 20,000 years ago.
There is also the wolf bone from Dolní Věstonice, discovered in 1937 by Czech archaeologist Karel Absolon at a mammoth hunter site in what is now Moravia in the Czech Republic. It is roughly 30,000 years old. Carved into its surface are 55 notches arranged in two sets, 25 and 30, grouped in clusters of five. The 25-notch series ends with a notch that is twice as long as the others, a structural marker, a divider.
That is base five counting, using the five fingers on a hand. The ability to notice quantity is not even uniquely human. Stanislas Dehaene, a cognitive neuroscientist at the Collège de France, published a book in 1997 called The Number Sense that changed how scientists think about the relationship between brains and numbers. He and his colleagues demonstrated that human infants, before they can speak or walk, already possess what he calls an approximate number system.
A six-month-old baby can tell the difference between eight dots and 16 dots on a screen, not by counting, but by a gut-level sense of magnitude. Andreas Nieder at the University of Tübingen discovered individual neurons in the brains of rhesus monkeys that are tuned into specific quantities. One neuron fires when the monkey sees three objects. A different neuron fires for four.
He found the same thing in crows, birds whose evolutionary lineage separated from primates over 300 million years ago. Nature built the number sense independently at least twice because tracking quantity is that useful for survival. But animals can estimate and compare. They cannot count past roughly four with any precision, and they cannot do anything abstract with the numbers they sense.
The linguist Daniel Everett spent years studying the Pirahã people of the Amazon basin and found that their language has no exact number words at all, just relative terms for few and somewhat more. Cognitive scientist Peter Gordon tested Pirahã adults in 2004 and found they struggled to reproduce exact quantities larger than three. Everett’s conclusion was stark. Exact counting is not an innate cognitive capacity.
It is a cultural technology, something humans invented. Archaeologist Denise Schmandt-Besserat at the University of Texas at Austin spent decades studying thousands of small clay objects excavated from sites across the ancient Near East. They look like toys, little spheres, cones, cylinders, disks. They are the oldest mathematical technology on Earth.
Starting around 8,000 BC, farming communities in Mesopotamia used these clay tokens to represent specific commodities. One sphere equaled a large measure of grain. One cylinder equaled a jar of oil. Around 3,500 BC, as cities like Uruk grew larger and trade became more complex, scribes started placing these tokens inside hollow clay balls called bullae and sealing them shut with a cylinder seal, creating tamper-proof contracts.
But there was a problem. You could not tell what was inside without breaking the seal. So scribes started pressing the tokens into the wet clay surface of the bulla before sealing it. Then someone realized that if the impressions on the outside told you everything the tokens on the inside did, you did not need the tokens anymore.
By roughly 3,200 BC, scribes at Uruk were pressing reed styluses into flat clay tablets, making marks that represented quantities. And by 3,100 to 3,000 BC, they had separated the number from the thing being counted. The abstract number was born, not in a university, but in a warehouse, and its purpose was accounting. From tokens in a bag to the concept of abstract number took roughly 5,000 years.
The Sumerians developed a base-60 number system, not base-10. 60 is divisible by 1, 2, 3, 4, 5, 6, 10, 12, 15, 20, 30, and 60, giving 12 divisors. Base-60 makes fractions clean and division easy. Your hands might be the reason your clock has 60 minutes.
Count the 12 phalanges of your four fingers using your thumb as a pointer, and track how many sets of 12 on the other hand. 12 multiplied by 5 is 60. The maths they did with that system was not simple. There is a clay tablet in the Yale Babylonian Collection, catalog number YBC 7289, dating to roughly 1800 to 1600 BC.
It shows a square with both diagonals drawn. Along one side is the number 30. Along the diagonal is inscribed in sexagesimal notation 1 24 51 10. Convert that to decimal and you get 1.
414211296. The modern value of the square root of 2 is 1. 41421356. They were accurate to six decimal places, roughly 3,800 years ago.
Plimpton 322, a clay tablet purchased around 1922 for $10 by publisher George Arthur Plimpton from antiquities dealer Edgar Banks, the real-life inspiration for Indiana Jones, dates to roughly 1800 BC. It contains 15 rows of numbers arranged in columns, and those numbers are Pythagorean triples. Row one, 119 squared plus 120 squared equals 169 squared. The Babylonians were generating Pythagorean triples over a thousand years before Pythagoras was born.
Eleanor Robson at the University of Cambridge argued in 2001 that the tablet is a scribal school exercise in reciprocal pairs. Daniel Mansfield and Norman Wildberger argued in 2017 that it is the world’s oldest trigonometric table. Either way, it is sophisticated mathematics performed in Mesopotamia in the second millennium BC. Meanwhile, in Egypt, a scribe named Ahmes was copying the most famous homework assignment in history.
The Rhind Mathematical Papyrus, purchased in 1858 in a Luxor street market by Scottish antiquarian Alexander Henry Rhind, is a scroll roughly 5 m long containing 84 mathematical problems. It dates to around 1550 BC, but Ahmes notes at the top that he is copying a much older text from the reign of Pharaoh Amenemhat the third, around 1850 BC. He titled it “Accurate reckoning for inquiring into things and the knowledge of all things, mysteries, all secrets. ”
Problem 50 gives a rule for calculating the area of a circle.
Subtract 1/9 of the diameter, then square the result. Run the numbers, and you get an implicit value for pi of approximately 3. 1605. The actual value is 3.
14159. They were off by about 0. 6%. Problems 56 through 60 calculate the slope of pyramids using a unit called the seked.
Problem 56 gives a seked that corresponds to a slope angle of roughly 51 degrees and 50 minutes, the incline of the Great Pyramid of Giza. The Moscow Mathematical Papyrus, acquired by Russian Egyptologist Vladimir Golenishchev in the 1890s and now housed in the Pushkin State Museum of Fine Arts, dates to roughly 1850 BC. Problem 14 calculates the volume of a truncated pyramid using the correct formula: 1/3 of the height times the sum of the base squared, the top squared, and the product of the base and top sides. The Greeks did something different, something nobody had done before.
They asked why. Starting around 600 BC, they began treating mathematics as something that needed to be proven, not just shown to work in a specific case. Thales of Miletus, born around 624 BC, is traditionally credited as the first person to prove a mathematical theorem. His theorem, that an angle inscribed in a semicircle is always a right angle, is not especially dramatic, but the idea behind it is revolutionary.
Pythagoras of Samos, born around 570 BC, founded a secretive brotherhood in Croton, in southern Italy, that was part math school, part religious cult. They believed with absolute conviction that the universe was built entirely from rational numbers. When Pythagoras discovered that vibrating strings produce harmonious sounds at ratios of 2 to 1, 3 to 2, and 4 to 3, it confirmed everything. Then Hippasus of Metapontum, around the 5th century BC, proved that the diagonal of a unit square, which equals the square root of two, cannot be expressed as a ratio of two whole numbers.
It is irrational. The number never ends. It never repeats. According to the ancient writers Iamblichus and Pappus, the Pythagoreans were so horrified by this discovery that they took Hippasus out to sea and drowned him.
A man was killed because a number did not cooperate. Euclid of Alexandria, working around 300 BC, compiled 13 books called The Elements. 23 definitions, five common notions, five postulates. From that foundation, he built the entire known structure of geometry using nothing but logical deduction.
Book 9, Proposition 20, proves that there are infinitely many prime numbers. The Elements was used as a mathematics textbook for over 2,000 years. Abraham Lincoln taught himself logic from it. Archimedes of Syracuse, born in 287 BC, trapped the value of pi between two bounds by inscribing and circumscribing regular 96-sided polygons around a circle.
His result was that pi is greater than 3 and 10/71 and less than 3 and 1/7. He proved that a sphere has exactly 2/3 the volume of its circumscribing cylinder and asked for the diagram to be engraved on his tombstone. Cicero found the grave a century later and confirmed the engraving was still there. The rest of the world was inventing things the Greeks never imagined.
In India, the mathematician Baudhayana, writing around 800 BC, stated the Pythagorean theorem explicitly, roughly 300 years before Pythagoras was born. The Sulba Sutras contain an approximation of the square root of two accurate to five decimal places. India’s greatest mathematical gift to the world is zero. In 628 AD, a mathematician named Brahmagupta, working in Bhillamala in what is now Rajasthan, India, composed the Brahmasphutasiddhanta.
In chapter 18, he defined zero as an independent number and laid out its complete arithmetic rules. The oldest undisputed physical carving of a circular zero in a decimal place value system appears in an inscription at the Chaturbhuj Temple at Gwalior Fort in Madhya Pradesh, India, dating to 876 AD. Aryabhata, working in 499 AD, calculated pi as 3. 1416.
He also developed the first comprehensive sine table using a Sanskrit word jya, which was transliterated into Arabic as jiba, then misread by later scribes as jibe, meaning fold or bay, which the 12th century Latin translator Gerard of Cremona rendered as sinus. That is why trigonometry uses the word sine, because of a translation error of a translation error of an Indian mathematician’s work. In the 14th century, Madhava of Sangamagrama, working at a school of astronomy in Kerala in southern India, discovered infinite power series expansions for pi, sine, and cosine. The formula pi over four equals 1 minus 1/3 plus 1/5 minus 1/7 and so on forever, discovered by Madhava, is still called the Leibniz series in most Western textbooks.
Leibniz published it in 1676 AD. Madhava had it roughly 300 years earlier. In China, The Nine Chapters on the Mathematical Art, compiled between roughly 200 BC and 100 AD, and given its canonical commentary by Liu Hui in 263 AD, solves systems of simultaneous linear equations with up to five unknowns by arranging coefficients on a counting board and reducing them through row operations. That procedure is identical to what Western mathematics calls Gaussian elimination, named after Carl Friedrich Gauss, born in 1777 AD.
The Nine Chapters also introduced negative numbers formally. Red counting rods for positive numbers, black counting rods for negative numbers. European mathematicians would not reluctantly accept that negative numbers were even real for another thousand years. Liu Hui, in his 263 AD commentary, calculated pi by inscribing polygons inside a circle until he reached a 3,072-sided polygon.
His result was pi approximately 3. 14159. Zu Chongzhi, working around 480 AD, pushed the method to a 24,576-sided polygon and calculated pi as lying between 3. 1415926 and 3.
1415927. His fractional approximation, 355 divided by 113, remained the most precise value of pi anywhere on Earth for over 1,100 years. In the Islamic Golden Age, around the 9th century AD, Muhammad ibn Musa al-Khwarizmi, working at the House of Wisdom in Baghdad around 820 AD, wrote a book called Kitab al-Mukhtasar fi Hisab al-Jabr wa al-Muqabala. The word algebra, meaning restoration, became the English word algebra.
Al-Khwarizmi also wrote a separate book on Hindu numeral arithmetic that was translated into Latin in the 12th century as Algoritmi de numero Indorum. The Latinized version of his name, algoritmi, became the English word algorithm. Omar Khayyam, the 11th century Persian poet who wrote the Rubaiyat, was also a mathematician who systematically classified and geometrically solved cubic equations by finding intersection points of conic sections. He investigated Euclid’s parallel postulate using what is now called the Khayyam-Saccheri quadrilateral, directly anticipating non-Euclidean geometry 700 years before anyone in Europe explored it.
With zero contact with any of these traditions, the Maya invented a complete positional number system with a base of 20, using a dot for one, a bar for five, and a stylized shell glyph for zero. The oldest Mesoamerican inscriptions using a zero date to the 1st century BC at sites like Tres Zapotes and Chiapa de Corzo. The Dresden Codex contains astronomical tables tracking the 584-day synodic cycle of Venus with an error of less than 2 hours over 500 years. At least three separate civilizations, the Babylonians, the Indians, and the Maya, independently invented the concept of zero.
At least four, the Babylonians, the Indians, the Chinese, and the Greeks, independently arrived at what we call the Pythagorean theorem. The House of Wisdom translated Greek and Indian works into Arabic. European scholars translated those Arabic texts into Latin in 12th century Toledo. Leonardo of Pisa, known as Fibonacci, learned the Hindu-Arabic numeral system from Arab merchants in Bougie, in what is now Algeria, and published it in his 1202 AD book Liber Abaci.
Descartes unified geometry and algebra with the coordinate system in 1637. Newton and Leibniz independently invented calculus in the late 17th century. Newton even hid his discovery in an encrypted Latin anagram in a 1676 letter, afraid that Leibniz would steal the credit. Decoded, it says, “Given an equation involving any number of fluent quantities to find the fluxions and vice versa.
”
So how did humans invent maths? They did not invent it once. They did not invent it in one place. And they almost certainly did not set out to invent it at all.
The oldest evidence of mathematical activity is a baboon bone with 29 notches from a cave in Southern Africa, roughly 42,000 years old. The evidence overwhelmingly points to multiple independent origins. The basic tools of mathematics, counting, geometry, the Pythagorean theorem, the concept of zero, were developed independently by cultures with no contact with each other on different continents. Given the cognitive hardware wired into every human brain, and the practical demands of trade, agriculture, construction, and astronomy, mathematics is a convergent invention, the same way eyes or wings are convergent biological adaptations.
Nobody set out to discover eternal truths about the structure of reality. They were trying to get through the day. And yet the maths they built, piece by piece, accident by accident, century by century, turned out to describe the universe itself. The orbits of planets follow equations.
The structure of atoms obeys symmetries. The expansion of the cosmos is governed by calculus.
