The Machine from the Sea
In the spring of 1901, sponge divers sheltering from a storm near the small Greek island of Antikythera came upon the wreck of an ancient ship, its remains scattered with bronze and marble statues that had lain on the seabed for some two thousand years. Among the treasures hauled to the surface was an unremarkable corroded lump of bronze, easily the least impressive object recovered, and, as it would slowly emerge over the following century, by far the most extraordinary. Today that object, the Antikythera mechanism, is widely described as the earliest known analogue computer: a geared machine that modelled the movements of the heavens with a precision that would not be matched again for a very long time.
In its complete form the mechanism was a hand-cranked box of interlocking bronze gears, no larger than a thick book. It now survives in 82 fragments, of which only about a third of the original remains, including some thirty corroded gearwheels. By turning a handle on its side, a user could set the device to a chosen date, whereupon its dials would display the positions of the Sun and the Moon together with a range of related astronomical information. What makes it so startling is not merely its ingenuity but its isolation: nothing of comparable mechanical complexity is known to have been built anywhere for more than a thousand years afterwards.
For the first half-century after its recovery, the device gave up few of its secrets. Too fragile and too thoroughly corroded to be prised open without destroying it, it sat in a museum as an intriguing curiosity. The first real progress came in the mid-twentieth century, when the British physicist-turned-historian Derek de Solla Price passed beams of radiation through the fragments and, from the shadowy outlines of the gears within, argued that the object had been an astronomical calculating machine. His study, published in 1974, established the broad purpose of the device, though much about how it actually worked remained obscure.
The decisive advance came in 2005, when a research team brought to Athens a purpose-built X-ray machine capable of imaging the interior of the fragments in three dimensions at extremely fine resolution. The machine had originally been designed not for archaeology but to inspect industrial components for hidden flaws. Its scans exposed thousands of minuscule Greek characters inscribed on surfaces buried within the corrosion, in effect an instruction manual etched into the bronze, describing the dials and the celestial motions they were built to represent.
The reconstructed picture is remarkable. On the front, a pointer tracked the Sun’s path around a ring marked with the zodiac, while a second showed the Moon advancing not at a steady pace but at its true, uneven rate, an effect achieved by a brilliant pairing of gears in which a pin on one disc ran along a slot in another. A small ball, one half dark and one half light, turned to indicate the lunar phase. On the back lay two great spiral dials: one followed a nineteen-year cycle that reconciled the lunar months with the solar year, while the other traced an eighteen-year cycle by which eclipses could be predicted, with engraved symbols marking the months in which an eclipse was due and even its likely time and colour.
The mechanism was not concerned only with the abstract heavens. A subsidiary dial tracked the four-year cycle of the great Panhellenic athletic festivals, among them the Olympic Games, tying the motions of the sky to the rhythms of civic and religious life. The level of skill embodied in such gearing, above all in the device that captured the Moon’s changing speed, implies a mature tradition of mechanical craftsmanship. Yet that is precisely the puzzle: of such a tradition almost nothing else survives, and the expertise the mechanism represents seems to appear in the record only to vanish from it again.
Work on the device continues to this day. A reconstruction published in 2021 proposed how the front display might also have shown the five planets known to antiquity, and scholars still debate finer points, such as how the calendar ring around the front was divided. Whatever remains unresolved, the mechanism has already overturned a comfortable assumption: that the Greeks excelled at abstract theory but had little taste or talent for practical engineering. It stands instead as proof that a civilisation two thousand years ago could capture the order of the cosmos in turning bronze, and then, somehow, let that knowledge slip away.