Friday 24 June 2016
Ancient Smarts -- the Antikythera Computer
As any of you who read this blog will know, last year I published a book called SMARTS. It dealt with the nature of intelligence and our ideas about it. It offered an informal history of how and why humans began to copy the many other forms of intelligence exhibited by wildly different living things and put them to work as smart machines.
It's a complicated story. For about two thousand years, it was a given in the West that only humans are intelligent. After Descartes had his say, it also became a given that the mind is something separate from the body, and this formless, ephemeral, spiritual whatever was imbued by God with intelligence, otherwise known as the capacity to reason. David Hume gingerly separated these teachings of religion from the observation of natural phenomena, and put forward the idea that human minds, like the minds of animals, are functions of human brains in bodies struggling to survive. Some of Hume's ideas were considered so radical when he was alive that his friend Adam Smith refused to publish them even after Hume was dead on the grounds that he would be charged with the crime of blasphemy. Another hundred years had to go by before Charles Darwin had the nerve to examine the behavior of various living things, other than humans, and to call it intelligent even when he saw smartness at work in plants with no brains. But Darwin's observations on intelligence were forgotten even as his evolutionary theory was reshaped to make it more acceptable to 19th century minds. His ideas about how random change could lead to both species die-offs and the generation of new, successful populations, soon came to mean continuous progress toward something better.
In anthropology and its sister science, archaeology, this progress idea fed the notion that change could be engineered for a better result. Darwin's cousin Francis Galton was convinced that white Anglo Saxons of the fine British families of his day -- the latter part of the 19th and early 20th Century -- were much smarter and better behaved than all the rest and should be encouraged to reproduce in large numbers, while the poor and benighted should be encouraged not to reproduce at all. At the same time, people active in anthropology and archaeology convinced themselves that the people who came before were less able -- unless proven otherwise. When anthropologists tried to explain how the first people came to the Americas, for example, it was assumed that they could not possibly have built boats and navigated across featureless waters as European sailors had done in the 15th century (and, as we later learned, Viking sailors had done in the 10th). Such skills were deemed to be too complex for primitive folk. Therefore, they must have walked. All kinds of evidence to the contrary was simply ignored and only one portal to the Americas was considered.
For a very long time, those active in science remade Darwin's randomness and complexity into something approximating progression and order. Which brings me to the Antikythera Mechanism, and the research project associated with it.
In 1901, some sponge divers working off the island of Antikythera, which is a southern Ionic island north of Crete in the Aegean, found a fabulous shipwreck at first dated to about 80 BCE. Out of the water came glorious bronze and marble sculptures, large amphorae, the finest of serving pieces, all being hauled somewhere by a Roman vessel that went down in a storm. The ship was apparently on a path from east to west, possibly from Turkey, or the island of Rhodes. These things, in other words, were probably being shipped by someone very rich to Rome, then the center of the Western world. Among the discoveries were some corroded pieces of a machine that had been made out of bronze. It had cogs. It had wheels. The cogs seemed to turn something that turned something else, like a clockwork. But surely that had to be wrong because the ancients couldn't make clockworks at all, let alone out of bronze. Could they?
Please bear in mind that in 1901 a computer meant a person calculating with pencil and paper, although there were machines by then which were mechanical computers with clockwork innards -- such as jacquard looms, and player pianos. These machines were specific computers, designed to do one kind of calculation. They mechanically produced sounds in the correct order, or produced a weave of a certain pattern. In the 1930s, Alan Turing invented the idea of a universal computer, a simple machine that could solve any solvable mathematical problem fed to it. The first such universal computer was built at Bletchley Park. It was called a Bombe and it was used to mechanize attempts to break the ever changing Nazi Enigma code. Because the Bombes were mechanical, they were too slow, so the first electronic universal computer was also built and put to work at Bletchley Park in 1943. These electronic machines set the stage for the next thirty years of computation until new ideas of how to do things still faster came along.
And that's why it took many, many years before anyone realized that the mechanism pulled out of the shipwreck in the Aegean was a computer. Most people, other than information scientists, had no experience with computers until the 1980s. At first it was believed that the machine might have been some sort of calendar, but based on what system? No one imagined the ancients were capable of thinking through, never mind manufacturing, a computer capable of relating four different calendar systems to each other. While it is not a universal one, the Antikythera machine is an immensely complex mechanical computation device that embodies thousands of years of observations of the movements of five planets, the elliptical orbit of the moon, and predicts eclipses of the moon and sun right down to their shifting times of day or position on the planet. The machine ties what was known about the relationships between the fixed stars and the movements of the planets as embodied in the astrological systems developed in Babylon and Egypt, and several different kinds of calendars.
Very little of this was known until recently due to new work on the mechanism undertaken by a multinational research group called the Antikythera Mechanism Research Project. It was first financed by the Leverhulme Trust in the UK. It is now supported by the Greek National Archaeological Museum, several universities in the UK and Greece, and companies like Hewlett Packard. The group involves physicists, astronomers, classicists, epigraphers, and has made use of fascinating leading edge technology. That technology has made it possible to infer what the computer looked like and how it worked from the fewer than 100 corroded fragments recovered from the sea. Since their first publication in the journal Nature in 2006, they have put forward ever more astonishing findings.
At first, the group argued that the mechanism was a device that combined Babylonian astronomical observations and calendar systems with an Egyptian calendar using Greek names. In 2008, further discoveries were aided by the use of 3D X-rays, and a dome which permitted taking still photographs of the three dimensional structure of the mechanism's surface. These images revealed in astonishing detail the virtual equivalent of instruction manuals etched into the surfaces of the front and back of the machine. 3D X-rays also revealed the structure of the drive trains that connected one set of cogs to another, and then to the face plates and their pointers -- the information readout from the machine's calculations. The 2008 paper showed that the rotations of the planets against a fixed backdrop of stars, the elliptical orbit of the moon, its relationship to solar cycles, the relationship to the dates of Hellenic Games,"were a microcosm of the temporal harmonization of human and divine order."
Eventually, these technologies made it possible to resolve to better than a tenth of a millimeter the inscriptions on the machine, making them readable. The translations of the inscriptions show that this machine, when set correctly, could accurately predict both solar and lunar eclipses (and even their colors), along with the correct timing of the Hellenic games cycle. More than three different calendar systems are involved, each related to the other by means of the cogs, drive trains, and the wheels and spirals with marked segments. The inscriptions have been compared to the known languages in use in Greece. This has allowed the group to pinpoint the probable cultural origin of the mechanism, and the time of its manufacture, and to replace other speculations about meaning with facts.
The texts make it clear that this mechanism was a product of the Greek city state of Corinth, or one of its former colonies, such as Sicily or Syracuse, made some time in the 2nd Century BCE. This was suggested by the Nature paper of 2008. The authors said then that the machine has an intellectual heritage "going back to Archimedes", and the new translation shows they were right. Corinth was destroyed by the Romans in 146 BCE. Syracuse was attacked and taken by the Romans much earlier. Archimedes, the great mathematician, scholar and engineer, was born and lived in Syracuse. It is alleged he died when the Romans invaded in 212 BCE. A Roman soldier killed him for failing to go quickly enough to be interviewed by the conquering Roman General who wanted to speak with him. Why the General would want to speak to Archimedes is clear: Archimedes calculated pi, he invented the water screw pump, and many other machines, including machines of war. He worked on all sorts of geometrical relationships, studied spirals, came close to inventing calculus, was familiar with the use of infinitesimals and used these ideas to calculate the surface areas of spheres and cylinders. He was the Leonardo da Vinci of his day--which lasted from 287 to 212 BCE.
The Antikythera mechanism is a window into his world and the ideas circulating in his time. It was a period of great mathematical and material sophistication which was itself the result of thousands of years of observation, discovery, correlation and sharing among different cultures.
The scholars of this project argue that nothing even close to the Antikythera Mechanism would be made for another thousand years.
But is that the case? Or will our views on that too have to be revised by the next major find? The Greeks of Plato and Aristotle's time, three hundred years before Archimedes contemplated water screws, spirals and pi, looked back to a time more sophisticated than their own golden age. Plato writes of the city state of Atlantis as the home of intellectual giants whose great civilization tragically disappeared under the sea. Anyone who has visited the ruins buried under mountains of volcanic ash on the Greek island of Santorini (Thera), the remains of Minoan city that was wiped out about 4500 years ago by a vast volcanic eruption that unleashed a huge tsunami, will not be surprised if something very interesting from that period is found sometime soon.
And surely there can't have been only one machine like the Antikythera Mechanism. It embodies not just thousands of years of thinking and observation, but thousands of years of perfecting the making of metal devices which can relate one information system to another and predict an outcome. There must have been many more machines made like this, both precursors and follow-ons.
And doesn't the new understanding of what this machine could do force us to dump for good the 19th century's progressive view of human cultural development and replace it with something more fitting, something approximating reality?
How about two steps forward, one step back.