The Antikythera Mechanism: Cosmion of Scientific History

Ancient Greek Science Final Paper

The Antikythera Mechanism: Cosmion of Scientific History

“The animated figures stand
Adorning every public street
And they seem to breathe in stone, or
Move their marble feet.”

-Pindar, speaking of the island of Rhodes in his 7th Olympic Ode

So much of our own history is unknown to us; recording events as they pass in a somewhat permanent form is a relatively new practice. As Francois Charette, historian of science at Ludwig Maximilians University in Munich, Germany, put it: “It’s still a popular notion among the public, and among scientists thinking about the history of their disciplines, that technological development is a simple progression. But history is full of surprises”1

We know that sometime shortly after 85 B.C., a ship carrying a cargo of luxury goods: statues, coins, vases in the style of Rhodes, and a geared mechanism complicated enough to have come from a thousand years later sank. It sank along a busy shipping route from the eastern to western Aegean. The decaying bronze gadget discovered so many years later in the wreckage included gear work of such complexity that, according to an article published in Nature magazine, “to get anything close to the [device’s]… sophistication you have to wait until the fourteenth century, when mechanical clockwork appeared all over Western Europe.”2

In 1900, Elias Stadiatis, a Greek sponge diver, discovered the wreck off of Antikythera island; divers began retrieving artefacts, but the mechanism itself was not discovered until May 1902, when archaeologist Valerios Stais noticed that a piece of rock recovered from the site had a gear wheel embedded in it.3 The so-called rock turned out to be a thickly encrusted and eroded mechanism, now in three large parts and many fragments.

A model built by Michael Wright, a curator at the Science Museum in London for more than 20 years, illustrates our understanding of the basic design. The model was made with traditional methods, using simple steel tools. He observed that the “only really difficult part was the manufacture of the front dial, with its’ recessed ring.”4

The gears are housed in a wooden box, with two dials on the front: one showing the zodiac and the other the days of the year. Metal pointers show the positions of the Sun, the Moon, and the five planets visible to the naked eye: Mercury, Venus, Mars, Saturn, and Jupiter. It is unclear from the remains of the actual device whether or not these planets are represented, or how many of them are, but Mars and Venus are mentioned by name in inscriptions on the original device.

By turning a hand crank, the main gear is moved, and the other gears create the mechanical ratios necessary to keep the various positions of planets and the sun and moon correct. Two spiral dials on the back of the mechanism with pointers that trace around “like a record stylus” around a groove show two astronomical cycles: the top dial shows the Metonic cycle, which sets 235 months evenly inside of 19 years. The bottom dial was divided into 223, reflecting the 223 month period of the Saros cycle5.

The first cycle marked on the back was introduced by the Greek astronomer Meton of Athens around as a means of mathematically determining the placement of a thirteenth month within the year, which was required every few years by a Synodic system that fell about 11 days short on average each year behind the 365 day solar year. Meton’s observation that a period of 19 years is almost exactly 235 lunar months is only inaccurate by about two hours. It can be used also to predict eclipses.

The second cycle was a tradition from ancient Babylonian astronomers; the Saros cycle describes the relative geometry of the earth, moon and sun. One “Saros” (223 lunar months) after any given eclipse, the moon, sun, and earth will be at the same relative geometrical positions; a nearly identical eclipse will occur.6 The cycle is not a whole number of days, but includes a fraction of 1/3 of a day, meaning the new eclipse will be about 8 hours after the previous. Three Saros cycles later, the local time of the predicted eclipse will be nearly the same. “This period is known as… a triple Saros or exeligmos (Greek: ‘a turn of the wheel’)”7; perhaps this nom illustrative is a linguistic reference to a device like the Antikythera mechanism, whose second dial shows both Saros and exeligmos.

The front dial includes a parapagmata to show the rising and setting of specific stars, which are thought to be identified by it’s Greek characters which cross reference details inscribed on the mechanism. Hipparchus was the first to compile a comprehensive version of such a catalogue, which included thousands of stars. This catalogue was lost, until recently it was discovered that the globe carried by the Farnese Atlas sculpture matched the original descriptions in the surviving Commentaries.

It was inscribed with 41 constellations, the celestial equator, and tropics. It is a Roman copy of a well known Greek statue. The specific locations of each constellation on the Atlas are accurate to within 3.5 degrees, while the oral descriptions of Eudoxus and Aratus could at best offer accuracy within 8 degrees; strong evidence indeed that the statue plan came from a star catalogue. Dating based on the procession cycles of that certain starscape represented on Atlas’s earth offered the date of 125 BC, a rough estimate, but which fits right in with the rough estimate for Hipparchus’s life work; the lost parapagmata has been found.

There is no hard evidence for where the mechanism might have been manufactured, but the growing consensus is that the mechanism was made on the island of Rhodes. This consensus is based on the circumstantial evidence of the wreckage, particularly vases in the style particular to that island and time period. The great astronomer Hipparchus is thought to have worked on the island between 140 BC and his death around 120 BC; his tradition was continued through an astronomy school set up by a philosopher named Posidonius.

The Antikythera mechanism could be a part of that tradition. Posidonius made significant contributions of his own, by calculating the size and distance of the moon, and making an attempt at the same for the sun – his result for size was larger and more accurate than previous attempts, but his estimate for the distance was only about half correct.

The first century Roman lawyer and consul Cicero, who studied on Rhodes island, provides a sort of circumstantial evidence for this possibility in his De Natura Deorum, claiming that Posidonius had made an instrument “which at each revolution reproduces the same motions of the Sun, the Moon, and the five planets that take place in the heavens every day and night.”8

Further circumstantial evidence for this theory is mentioned in the Nature article, as three-dimensional reconstructions of the fragments have shown that a very precise pin motion adjusts the surviving moon indicator so that it speeds up and slows down in accordance with the moon’s relative motion across the sky. Perhaps the sun’s gear could have also been set up in such a way, but any gearing for this mechanism would have been lost.

Hipparchus was the first to describe this motion mathematically, “working on the idea that the moon’s orbit, although circular, was centred on a point offset from the centre of the earth that described a nine-year cycle.” 9 The real cause of this apparent change in speed is the moon’s elliptical orbit; the idea of an elliptical orbit did not appear until Kepler. Hipparchus was also the first to compile a trigonometric table that allowed him to solve any triangle.

The level of precision with which the mechanism captures the heavenly motions suggests that it could not have been an isolated, miraculous single event. At the least, many test-models would have failed before one that functioned could have been made. It is curious to say the least at first thought that there are no examples of anything even remotely similar. One possibility might be that other examples have long since been melted down for scrap bronze. The Athens museum has just ten major bronze statues from ancient Greece; nine these are from shipwrecks. Said Wright, the model-builder, “We only have this one because it was out of reach of the scrap-metal man.”10

In fact, Cicero mentions another similar machine, attributing its invention to Archimedes. According to Cicero, the general Marcus Claudius Marcellus brought Archimedes’ device to Rome after the death of Archimedes at Syracuse around 212 BC. The device was apparently kept as a family heirloom and shown to Cicero by Gallus about 150 years later. “When Gallus moved the globe, it happened that the moon followed the sun by as many turns on that bronze (contrivance) as in the sky itself, from which also in the sky the sun’s globe became (to have) that same eclipse, and the moon came then to that position which was (its) shadow on the earth, when the sun was in line.”11 If both of Cicero’s accounts of devices like the Antikythera mechanism are accurate, we know of at least 3 in history. It is quite unlikely that the device found in the shipwreck were either of those mentioned by Cicero, because his accounts located the two devices to Rome at least 50 years later than the estimated date of the shipwreck.

Archimedes could also have invented such a device; he has gained a sort of legendary status as an inventor. He is supposed to have invented a large crane-claw grappling device, which would lift ships out of the water and then possibly sink them in order to defend Syracuse. It is often debated whether the other military invention attributed to Archimedes is fictitious; supposedly he managed to light fires on ships in the harbour by focusing the sun’s light over parabolic mirrors.

It is not difficult to imagine the person who designed the largest ship in antiquity, the Syracusia, which could carry 600 people according to Athenaeus, designing such an ambitious machine as the Antikythera mechanism. We shall probably never know for sure.

The information offered conveniently by such a model would have been of tremendous cultural significance for the ancients. Astrology was considered as important as any science, and certain angular aspects between planets might have strong portents associated with them. A court astrologer would often be asked by the monarch to chart the current aspects in order to determine if it was an auspicious time for whatever the monarch had in mind- usually a war. Eclipses were certainly thought to be one of the most powerful omens, making the ability to predict them by turning a gear quite appealing.

The process by which the machine was designed and built could be quite justly called analogue computer programming. This is hard for us to imagine maybe, we who are used to modern computers. An article in the New York Times put it this way: “A computer in antiquity would seem to be an anachronism, like Athena ordering takeout on her cellphone."12 But we can’t deny it just because it seems strange.

To adhere to the theory that technological development and the development of sciences was a “simple progression” as though it were a law, would be to sacrifice the empirical method which took so many years to develop out of those original Greek debates. Perhaps further developments in technology will allow scientists to reconstruct the mechanism in greater detail, or read more of the inscriptions on the cover. Perhaps another mechanism will be found someday. But for now, it stands as a symbol of a pinnacle for ancient scientific achievement, and a challenge to modern computer programmers to design a similarly impressive digital solar system model.

Marchant, Jo. “In search of lost time.” Nature 444 (2006) p.534-538.
“Antikythera Mechanism.” Wikipedia. 12 March 2007. http://en.wikipedia.org/wiki/Antikythera_mechanism.
“F.A.Q.” The Antikythera Mechanism Research Project. 10 March 2007. http://www.antikythera-mechanism.gr.
“Saros Cycle.” Wikipedia. 14 March 2007. http://en.wikipedia.org/wiki/Saros_cycle.
Cicero. De Natura Deorum. Book II, translated by H. Rackham, Cambridge, Mass. & London 1933
Cicero. De Re Publica I.22. 13 March 2007. http://www.thelatinlibrary.com/cicero/repub1.shtml#21.
"Early Astronomical ‘Computer’ Found to Be Technically Complex." John Noble Wilford. New York Times, 30/11/06.