In a nutshell, the Antikythera Mechanism is the thing that should not be. It is a mechanical artifact discovered around a hundred years ago at the bottom of the ocean, dated to about a hundred years before the common era, and which uses gears and other impossibilities to predict astronomical phenomenon and align calendars. Its level of sophistication wouldn't be approached until the clock works of the eighteenth century.

The Antikythera Mechanism is a useful reminder that history seldom follows simple, linear paths. (Charette 2006)

I should let you know from the outset, that although we’ve discovered some incredible things about the mechanism over the last couple of years, it’s inevitable that much more remains tantalizingly unknown.

I wish to dedicate this minor effort to the existence of captain Demetrios Kondos and the six divers under his command; Elias Stadiatis, Kyriakos and Georg Mundiadis, Johann Pilliu, Giorgios Kritikos, and Basilios Katzaras.

Discovery at Antikythera

The year is 1900 and a group of sponge divers have finished fishing and are on their way home to celebrate Easter. A small storm forces the men to take refuge at the tiny island of Antikythera, which sits between Crete and the Peloponnese in the Mediterranean Sea. Given that they’ve already weighed anchor, when the sea calms again they decide to dive down to a rock shelf they notice below, hopefully find a new location for harvesting sponges. When one of the divers – Elias Stadiatis – reaches the shelf he sees a shipwreck laden with all sorts of treasures, including a number of statues, amphoras, and other small objects.

Most storytellers at this point in the narrative make the inevitable announcement of the Antikythera Mechanism’s discovery and move onto the efforts to tackle its mystery. But these stories are composed of human beings, and man works in mysterious ways.

How did the crew respond to Elias’ announcement? Hopeful skepticism I imagine. The captain, Demetrios Kondos, was the next to go down, impatient to know for himself. He returns to the surface with a bronze arm. They obviously proceeded to take some sort of measurements, co-ordinates and such to allow for a return to the wreck’s site. Did they spend time taking anything for themselves? It’s uncertain. Some authors, like Price (who will be important below), emphasize their decision to approach the correct authorities. Marchant, who recently wrote a book on the Antikythera Mechanism, spoke to a journalist who'd interviewed some locals; rumors tell that the crew spent a number of weeks lifting some of the small items for themselves, including the lead bars which as divers they valued for weights.

In the end it doesn’t matter, but I do have my own preference. Of course this isn’t a story of heroes and villains: a handful of young and ocean-hardened Greek men, who risked their lives cutting sponges off outcrops on the sea floor, discovered a moment of glory. In the opening pages of her book Marchant reminds us that sponge diving is an ancient career. I prefer to think that the divers did take a few things themselves. For their part in a story that spans two millennia I think that it’s only appropriate.

The crew returned home, and as was the custom of sponge divers, spent the next few months celebrating their survival and profits. In November of 1900 Konotos and Stadiatis travelled to Athens with their proof – the bronze arm they’d recovered from the wreck – and enlisted the help of Konotos' old friend and archaeologist, professor Oikonomu. Oikonomu contacted the government and before long Oikonomu and the fishermen were promised compensation for their find and assistance in recovering the rest of the wreck.

The excavation lasted from the end of November 1900 until the end of September 1901. This period included a number of technical set-backs, including an initial ship which was too large to approach the site. By September one diver had been killed and two incapacitated in recovery efforts. Such was the nature of the game.

I won’t discuss the other findings, interesting as they are in their own right, and despite the fact that many have been examined for the purpose of identifying the ship's port of origin and age. It took until March of 1902 for the Antikythera Mechanism to be noticed, at this point just an ambiguous, albeit still unprecedented, lump of corroded gears and inscriptions.

Faced with the complexity and seeming anachronism of the artefact, controversy and intense debate broke out immediately. Was this an astrolabe? It seemed too complex to be even that. It was far too sophisticated to be contemporary with the rest of the wreck, perhaps it belonged to a second wreck from centuries later?

Solving the Antikythera Mechanism

The intensity of this mystery cannot be overstated, and for that reason attempts at a solution has consumed whole decades of individuals' lives. I will not be able to summarise all those lives here, but can recount a few which give a feeling for the evolving nature of the problem. The first researcher to, albeit prematurely, provide significant answers was Derek de Solla Price, a professor of history of science at Yale University.

I think that several times in history such genius has made geared astronomical clockwork so far ahead of his time that after him the development has rested for awhile to emerge with a tradition augmented more by stimulus diffusion than by direct continuation of the idea. A technological tradition is something so much more fragile than anything that was encoded into a written book and transmitted into the orderly fashion of knowledge. (Price 1974)

Price's most important publication was his 1974 Gears from the Greeks: The Antikythera Mechanism—A Calendar Computer from ca. 80 B.C. Price's work depended on the most complete account of the mechanism to date, including an intensive and manual attempt at counting the precise number of teeth on each of the gears from X-ray pictures. Based on these rough countings, and based on Price's knowledge of what the Greeks might have attempted to do, he suggested that the mechanism could calculate the relation between lunar and solar cycles, in particular the phases of the moon over years. This is a complicated achievement, since it requires accounting for the fact that the appearance of the moon differs with the Earth's motion around the sun, (a fact of which Greeks weren't aware), and similarly that the cycles of the moon do not align simply with the those of the sun (i.e. the moon does not rotate around the Earth a whole number of times over the course of a year).

Price's reports garnered interest, but not always of a positive nature. In his 1974 report he confesses to sometimes waking up in the night, wondering if perhaps the evidence has tricked him, if perhaps it is obvious that this magnificent contraption must date from a later period. Meanwhile others have sought the most outlandish explanations for the mechanism's existence, they are obviously wrong, but maybe their doubts in the capacity of the Greeks to produce these gears are justified? Strong doubt is an emotion shared by all those who have dedicated themselves to the Antikythera Mechanism. And this includes the next character in my chronicle: Michael Wright, whose story forms additional contributions to our understanding of the mechanism's nature.

I became convinced that by bringing to bear powers of observation developed as a museum curator in the study of intricate mechanism and craft techniques, together with the practical insights of a skilled workman, I might make a worthwhile contribution to the study of the Antikythera Mechanism. (Wright 2007)

Wright was a curator at the London Science Museum with a strong interest in the development of scientific instruments and their reconstructions. It seems inevitable that he should have come across the Antikythera mechanism and want to solve it. After reading Price's seminal report on the mechanism, it was soon clear to Wright that it included a number of flaws, and that an important limitation was Price's inability to properly image the artefact. Wright combined forces with the scientist Allan George Bromley to produce better images. This was achieved by a newer technique called X-ray tomography, a variation of the X-ray technique Price had used, and which produces clearer images of individual planes. With these incredibly improved images Wright could show that many of Price's gear counts were mistaken.

Wright began to work on his own model and submitted his proposal for publication in 2006. According to his reconstruction, Wright showed that the mechanism could calculate the cycle of the sun and moon, including showing the phases of the moon, and possibly the motions of some of the minor planets (including Venus). The mechanism could account for the relationship between synodic months and the solar years (i.e. the relation between the length of the moon's cycles and the apparent rotation of the sun), and could predict eclipses. Wright also hypothesized that the Antikythera mechanism was a horological tool, and that maybe missing parts include wheels for showing the movement of all five of the known planets.

In the same year an awesomely ambitious international group published their work on the Antikythera mechanism. The report utilized two important new techniques: (1) polynomial texture mapping (PTM) (which basically takes heaps of pictures with heaps of different lighting conditions and uses software to detect hardly legible etchings) and (2) microfocus computed tomography (microfocus CT) (a substantial improvement on the resolution of X-ray tomography, followed by sophisticated software to form 3D models).

These new tools, combined with a number of newly incorporated fragments which had not been available earlier (certainly not to Price), provided an unprecedented increase in the amount of translatable text. The new model featured 37 separate gears (of which seven are hypothetical), and with the new data could explain a number of the mechanism's features. This includes importantly the ability to account for anomalistic months (i.e. the nature of the moon's cycle means that it takes it about nine years to get back to the same place, and hence that this period affects its apparent shape and speed). This would only have been possible to build by taking into account Hipparchus's sophisticated studies and calculations on the moon's motions. Otherwise, much of the data agrees with Wright, including details like the mechanism's calculating lunar eclipses and zodiac cycles.

In 2008 the same group published a follow-up report. In addition to coordinating lunar and solar cycles, the mechanism displayed the month and its relation to four year cycle of the Panhellenic games, which were important civic affairs and which were used for referencing the date of historical events.

Rewriting the history books

The Antikythera mechanism and its story strain our conceptions of the ancient world. Here we have a clockwork mechanism which can, by use of interacting gears and their relative positions and teeth numbers, account for the unintuitively complex relation of lunar and solar cycles, including the phases of the moon, possibly the cycles of the planets, their relation to the zodiacs and the civic calendar!!!

No ancient artefact has yet been found similar to the Antikythera mechanism. This does not mean that it was one of a kind. Quite the opposite. The skill and accuracy of the device suggest it to be a product of a workshop. Why aren't more found? Most likely because virtually all bronze from the period was recycled. A sobering fact worth reflecting on is that while Pliny reports there being 3000 bronze statues in Rhodes in the first century A.D., the museum in Athens which has one of the best bronze statue collections in the world only has ten. All but one are from shipwrecks. This is the reality faced by those studying history.

The mechanism was not one of a kind, and must inevitably been part of an intellectual tradition, even if that tradition was attenuated and finally silenced with the fall of the Roman Empire. It seems very possible that the tradition was revived with the Arabic renaissance, and then communicated later in some abridged form into Europe. That undercurrent of knowledge could explain why, when gears were finally reinvented in Europe, their presence and complexity exploded rather than just trickled in.

In concluding it is worth wondering, as so many others have, about the identity of the mechanism's craftsman. Price, for instance, had good reason to suggest the mechanism to have originated in Rhodes, possibly in association with that island's great polymath Posidonios. This no longer seems likely since recent translations have revealed the mechanism to use Corinthian colony months, and instead it seems likely that it originated from northwestern Greece. One tantalising suggestion is that this artefact is the "Globe of Billarus" referred to by Strabo as being looted by the Roman general Lucullus in 72-71B.C. and which was never heard of again...

A note on pronunciation: Antikythera has the accent on -kyth- so that it rhymes with 'pith'.

Primary references: It's important to note that in all of my summaries, especially of the primary literature, I have not done justice to the complexity and insight provided by those authors. If you have the opportunity to peruse those papers you'll see that they discuss in great detail all the different parts of the mechanisms, their method of analysis, and mathematical and mechanical reconstructions.

It's amusing to note that the original writeup in this node is from 2002 - four years before the paradigm changing publications of 2006!

  • Price D. Gears from the Greeks: The Antikythera Mechanism—A Calendar Computer from ca. 80 B.C. (1974) Trans Am Phil Soc
  • Freeth T et al. Decoding the ancient Greek astronomical calculator known as the Antikythera Mechanism (2006) Nature
  • Wright M The Antikythera Mechanism reconsidered (2007) Interdiscipl Sci Rev
  • Freeth T et al. Calendars with Olympiad display and eclipse prediction on the Antikythera Mechanism (2008) Nature
  • The Antikythera Mechanism Research Project - the website of the currently most important research group, and can be referred to for previous publications and data, links, and updates
  • Evans J et al. Solar anomaly and planetary displays in the Antikythera mechanism (2010) J Hist Astron - this paper suggests, based on data provided by the Freeth et al., that the mechanism may not have been a mechanical realisation of Hipparchus' theories. The authors in this paper suggest that (1) the problem of representing the sun's apparent motion may be achieved by representing the anomalous nature of solar motion by using non-uniform divions of the zodiacs, rather than employing complex epicyles per Hipparchus; and (2) the representation of the planets' motions may be simpler than previously expected, indicating only major changes in motion rather than their precise observed motions. This remains hypothetical (as do the other models) but, the authors note, may hopefully be clarified by ongoing translation efforts. A significance of the first suggestion, is that it depends on Babylonian, rather than Greek theories.

It is not a matter of taking sides, I will only say that I wish that Wright's efforts may have been more fully appreciated by those of Freeth and colleagues, and that perhaps their results may have been published jointly. That is all I will say.

Secondary references:

  • Charette F High Tech from Ancient Greece (2006) Nature
  • Marchant J In Search of Lost Time (2006) Nature
  • Ball P Complex Clock Combines Calendars (2008) Nature
  • Robinson R Tale of a Clockwork Computer (2008) Nature
  • Marchant J Decoding the Heavens - an excellent book on the history of the mechanism and its mystery; for more see my review elsewhere
  • Marchant J Mechanical Inspiration (2010) Nature - this is the most recent review of the state of knowledge surrounding the mechanism, and is thoroughly recommended. It also includes a summary of recent discussions by Evans and colleagues (see my notes just above on Evans 2010).

Impressed yet? Check out this Lego model! (via)

Within reason, if you ask kindly, I'm happy to forward digital copies of the papers cited - they're astounding reads.