Science Spin July 2009

Behold the world's first computer

By Marie-Catherine Mousseau

The story of the world's first computer, as we know it began in 1900, not far from Crete. Sponge divers were returning to their home in Symi, when a storm blew up and forced them to take shelter on a small island, in front of the island of Kythera - Antikythera is its name. There they went for another dive and discovered the wreck of a Roman trading ship. This turned out to be one of the most surprising finds of modern time.

The significance of the discovery was not in the beautiful statues retrieved from the ship, although these happened to be "some of the finest Hellenistic bronze statues", commented Professor Mike Edmunds, Cardiff University, School of Physics and Astronomy, during the 'Frontiers of Physics Teachers Conference', held in UCD in September last year, and organised by the Institute of Physics in Ireland.

Professor Edmunds explained that the exceptional find made off the coasts of Antikythera came in the form of a lump of bronze, the size of a shoe box, which stayed in this shapeless form for a few years -- in fact until soon after it was taken to Athens, when for some as yet unclear reason it eventually cracked open. Then somebody looked inside and said: "Hang on a minute, there are gears in there."

The lump of bronze indeed contained fragments of a mechanism involving 30 gear wheels, maybe more. "Gears were not known to be present in classical Greece," Prof Mike Edmunds said. "This by itself is very unusual." But it was not before the 1960-70s, when radiology of the fragments was done, that the real implications of this device were investigated. What they revealed has completely challenged our current understanding of history of technology.

Gears

All the evidence seemed to suggest that not only has it the first known gears, it also has the first known scientific scales. As first published in Scientific American in June 1959, the Antikythera mechanism appears to be designed to display astronomical positions. When a date was entered via a crank, the mechanism calculated the position of the Sun, Moon, and possibly other astronomical information such as the location of other planets. It is even thought to have been able to predict eclipses.

So it seems that what the sponge divers discovered some one hundred years ago is the oldest known complex scientific calculator - or in other words the first known mechanical computer. What's more, the actual fragment is covered with Greek inscriptions of astronomical significance, which according to Prof Mike Edmunds means one thing - the instrument has instructions going with it.

"Nothing as complicated in terms of mechanism is known until the area of the medieval cathedral clocks," Prof Edmunds enthused. But we are far from the medieval era. Using such clues as the specific Greek words found in the inscriptions, or the coins surrounding it, scientists estimated the age of the device around 100BC. The astronomical calculator was created some thousand years before any kind of technological artefacts of similar complexity were known to exist.

As Prof Edmunds put it, "It's an extraordinary artefact."

Literature

Extraordinary surely, but not impossible. Cicero recorded in three of his books that he had seen or heard of a mechanism for displaying the position of the Sun, the Moon, sometimes the planets in the sky.

Two of the devices which Cicero described were built by Archimedes, and brought to Rome after his death at the siege of Syracuse in 212 BC. If Cicero's account is correct, similar technology had existed as early as the 3rd century BC.

"This is incredibly important," Prof Edmunds said, "if we didn't have this reference in the literature you wouldn't believe it's a genuine artefact from ancient Greece." Indeed, it is far too sophisticated to have been a one-off device. The flawless mechanism, the complexity of the design, everything in it suggests that it had a number of predecessors.

In fact, Cicero's references to such mechanisms support the idea that there was an ancient Greek tradition of complex mechanical technology. This would have later been transmitted to the Islamic world, where a number of similarly complex mechanical devices were built by Muslim engineers and astronomers during the Middle Ages. Such knowledge could then have been integrated with European clockmaking and medieval cranes.

Research

Prof Edmunds, who is an astronomer by training, was fascinated by the instrument's story. In order to try to move things further, he got involved in the formation of a multinational research team that became known as the Antikythera Mechanism Research Project (AMRP). This involved international collaboration between Cardiff University, Greek Universities as well as the National Archeological Museum of Athens and included people with a wide range of expertise from a professor of space and physics, a Greek radio-astronomer, and an expert in the form and arrangement of Greek letters.

But it was one of Prof Edmunds' friends, Tony Freeth, a former film-maker, who has conducted a vast amount of the work done on this project. "When I told him about the project, he became absolutely fascinated," Prof Edmunds said. "He'd basically dedicate his life to research this thing."

They realised that to get new results, new data was needed. This led Tony to propose new investigations using the latest cutting-edge technologies of microfocus X-ray and digital imaging. Two companies, Hewlett-Packard USA and X-Tek Systems UK, were taken on board.

X-Tek Systems developed an 8.5 ton microfocus X-ray computerised tomographer especially for the Antikythera mechanism, while HP designed a simple clever system to do the 3D surface imaging. Prof Edmunds explained: "This merely involves a hemispheric dome with 50 flash bulbs, a digital camera, and a computer." Basically the camera takes pictures while each of the bulbs lights up in turn. The result is then included in a software which allows the user to visualise and move the object in space.

They put together a proposal and hoped that it would be accepted. "It took four years to get permission from the Greeks to do the experiment," said Prof Edmunds. Even then, they wouldn't let the researchers remove the mechanism from the museum due to its fragility; so HP and X-Tek Systems had to bring their devices to Greece (including the 8.5 ton X-Ray tomographer).

Discoveries

Luckily, the results were worth the effort. The team's findings have shed new light on the function and purpose of the Antikythera mechanism. From about 700 characters that were visible previously, they can now identify and translate over 2000 characters. Thanks to the new inscriptions they were able to get a more precise dating of the object to around 100BC. They also confirmed that pointers on the front face of the mechanism (including the solar scale and the zodiac graduations) were designed to indicate the position of the Sun, the Moon, but also the lunar phase, and possibly the planetary motions.

The planetarium function in particular was supported by the inscriptions of Venus and Mars as well as those mentioning the stationary points of the different planets - together suggesting that the device had indicators for all the five planets known to the Greeks (i.e. the 'Inferior Planets' Mercury and Venus, and the 'Superior Planets' Mars, Jupiter and Saturn).

In addition, the team's findings supported the idea that the mechanism could also track the Metonic cycle (a particular period used to fix calendars), and predicted eclipses.

Finally, among the new inscriptions they identified the word 'Olympia'. Together with a dial divided into four parts demonstrating a four-year cycle. This suggested that the astronomical device was probably also used to calculate the timing of the Ancient Olympic Games.

But Prof Edmunds is particularly keen to point out another of their discoveries, which according to him is one of the most surprising of all. He explained that the orbital movement of the moon is not regular. The velocity of our satellite varies, which means that it does not move exactly by the lunar months.

The team discovered how, thanks to the clever use of a pin and a slot ingeniously placed on two gears linked with a slightly offset axis, the mechanism was able to mimic the irregularity of the lunar rolling and thus predict the position of the moon with great accuracy.

"This is extraordinarily clever," Prof Edmunds commented, "we're dealing with a phenomenal, staggering mechanical design, fascinating in its complexity." The results, along with a new reconstruction of the mechanism, were published on November the 30th 2006 and July the 30th 2008 - both in the journal Nature.

Purpose

Unfolding the mechanism underlying this brilliant design is undoubtedly fascinating, but even more so is using these clues to try to figure out the overall purpose of such a technological achievement. What is it all about? Who and why would care to construct such a sophisticated instrument to know the position of bodies in the heavens?

First of all, the device is thought to have had some social functions. As suggested by the tracking of the different cycles, it might have been used to adjust calendars based on lunar cycles (metonic cycle) as well as the solar year. This might also have been useful for setting the dates of religious festivals or predicting which games were to take place in any given year.

Secondly, astrology was standard practice in the ancient world. A device capable of giving the configuration of the heavens at any point of time would have been a precious tool for astrologers, saving them the time-consuming task of working this out themselves.

Some experts believe that its main purpose was for public display or as a teaching aid. One hypothesis is that the Antikythera mechanism was constructed on the Greek island of Rhodes. "Rhodes at the time was known for its astronomical tradition," Prof Edmunds pointed out.

There was a huge scientific and cultural gap between the very few educated elite who understood basic rules of solar, lunar and planetary motion and the common people who viewed solar and lunar eclipses as supernatural events, associated with fear. Such a device, on display in a university, museum or temple in Rhodes, could have been designed to fill this gap - thus representing one of the first tools aimed at popularising science.

Others argue on the contrary that it is too small and compact to be suitable for public display. They believe the mechanism aimed for compactness to be used by non expert travellers - a hypothesis which is supported by the inclusion of an instruction manual.

Could the instrument just be a navigation tool which would explain why it was found in a boat? But then what use would it have to predict eclipses? Also some argue that the harsh environment of the sea would corrode the gears very quickly. The reason this device was found in a wreck while all the other mechanisms did not survive could just be that bronze at that time was a valuable commodity. When instruments stopped working, they would have been melted down.

So the simple answer is, we don't know. Prof Edmunds likes to highlight its philosophical dimension. The astronomical calculator might have been designed to demonstrate the determinism of a perfectly mechanical universe.

In any case, the Antikythera mechanism still holds part of its secrets, which make it so special. "It is unique,", Prof Mike Edmunds said. "To my mind it is actually more valuable than the Mona Lisa. We have other works from Leonardo da Vinci, we have nothing else like this."

DESIGN

19, 76, 223; these are three key inscriptions found on the device, which according to Prof Edmunds highlight some of its key functions. "We all know what is the astronomical significance of these figures," says Prof Edmunds jokingly. Of course most of us don't, but Greek astronomers already did, as demonstrated by the way the Antikythera mechanism is designed.

The mechanism has three main dials, one on the front, and two on the back. The front dial has two concentric scales - the outer one marked with the days of the 365-day Egyptian calendar, the inside one marked with the Greek signs of the Zodiac.

The front dial is believed to display:

• The annual progress of the Sun and Moon through the zodiac against the 365-day Egyptian calendar.
• The lunar phase through a second mechanism it contains.
• Possibly the position of Mercury, Venus, Mars, Jupiter and Saturn, the five planets known by the Greeks.

The upper back dial in the form of a spiral displays:

• The 235 months of the Metonic cycle, which approximately equals 19 solar years. This represents the exact number of lunar months needed to get back at the exact same phase of the moon. The metonic cycle is important in fixing calendars.
• The 76 years (that is 4 x 19 years) of the Callippic cycle, that proposes a more accurate periodicity than the metonic cycle.
• The word 'Olympia' as well as the names of other games in ancient Greece. A dial demonstrating a four-year cycle is thought to be a mean of describing which of the games (such as the ancient Olympics) were to take place in any given year.
• The lower back dial is also in the form of a spiral, with 223 divisions showing the Saros eclipse cycle. The Saros cycle lasts exactly 223 lunar months, a period of approximately 18 years -- the length of time between occurrences of a particular eclipse.

"It is a brilliant design, fascinating in its complexity," Prof Edmunds concluded.

 

The Antikythera Mechanism Research Project in which Prof Mike Edmunds and Tony Freeth take part, built on the work of a number of individuals and groups who were instrumental in advancing the knowledge and understanding of the mechanism including: Derek J. de Solla Price (with Charalampos Karakalos); Allan George Bromley and his team; Michael Wright; and Dionysios Kriaris, a Greek mathematician.