Archaeology / Review Essay

Vol. 7, NO. 1 / June 2022

The Antikythera mechanism, an astronomical calculator found in a first-century BCE shipwreck, has proven to be mechanically more sophisticated than anything known from the subsequent millennium. While many are amazed at such a discovery, a more appropriate response would be admiration, for the mechanism fits well into its historic context. Indeed, the ancient scholar Cicero offered contemporary accounts of similar devices, which he saw as embodying the peak of human ingenuity.1 But the significance of the Antikythera mechanism extends beyond the elegance and complexity of its design. It may also represent a major development in our understanding of the universe.

Astronomical Mechanisms

In the spring of 1900, sponge divers working near the Greek island of Antikythera came across the wreck of a Roman cargo ship. Among the remains was a corroded shoebox-sized case with more than thirty bronze gear wheels in interlaced trains.2 Those who attempted to reconstruct it over the next century would learn that it included annular dials on its front and large spiral dials on the back that represented the day in the year, the lunar month in the 235-month Metonic cycle, the lunar phase, the position of the sun and moon in the zodiac, and whether the month might contain a lunar or solar eclipse. Irregularities in lunar motion were incorporated by means of an ingenious pin-and-slot variable-speed device. Predicted eclipses and the lunar calendar were based on observed cycles passed down to the Greeks from the Babylonians. The device itself was probably constructed in Rhodes sometime between 150 and 160 BCE, though both the date—it might be as early as 205 BCE—and the source are subject to debate. Inscriptions on the device strongly suggest that its front face also displayed the positions in the zodiac of the known planets: Mercury, Venus, Mars, Jupiter, and Saturn.

The significance of the Antikythera mechanism, as it came to be called, only began to be more broadly realized after the publication of Derek de Solla Price’s paper “Gears from the Greeks” in 1974. Price found the mechanism so sophisticated that it might “involve a completely new appraisal of the scientific technology of the Hellenistic period.”3 Just four years earlier, Germaine Aujac had written a perceptive, although largely forgotten, review of several kinds of mechanical devices that could have influenced Greek views of the universe.4 Price may well have been unaware of Aujac’s article. He does not reference it and Aujac does not mention the Antikythera mechanism. It would take another thirty years before the publications of the international Antikythera Mechanism Research Project (AMRP) and of Michael Wright prompted more general awareness of the artifact as confirming the reality of such complex machines in the ancient world.5

The devices Aujac wrote about were sphaerae—or sphéropée in the original French—mechanisms depicting the sky moving around the earth, with or without the planets. Sphaerae could be three-dimensional terrestrial or celestial globes and armillary spheres, but also two-dimensional circular constructions like the Antikythera mechanism. According to Aujac, by combining observation with the theory and construction of sphaerae, the Greeks were bringing models of the earth and heavens closer to their real equivalents. In his writings, Ptolemy acknowledged the existence of sphaerae, although he seems to have thought that they were admired more for their craftsmanship than for their value as physical models.6 James Evans and Christián Carlos Carman have argued that geared technology may slightly predate, and might actually have inspired, the mathematical developments around 200 BCE in Greek planetary theory such as eccentrics and epicycles.7

A Mechanical Universe

From my own perspective, the deeper question concerning sphaerae is to what extent the development of this technology prompted the Greeks and Romans into a new worldview.8 The technology may have affected not only mathematics, but also the idea that the universe itself is in some sense mechanical—and long before the so-called scientific revolution of the Renaissance. For Samuel Sambursky, the question is

whether these models are only convenient means of illustration, devices adapted to our needs for an ordered description, or whether they represent to a greater or lesser degree some faithful image of a physical reality corresponding to them.9

If meant as a faithful image of reality, there are several themes present in such an image. The first would be the realization that gearing, with, where necessary, the addition of pins, slots, and levers, reproduces rather well the celestial properties of determinism, repeatability, regular cyclic motion, and irregular cyclic motion. This set of properties starts to provide a physical explanation for the motions of celestial bodies without ongoing direction or intervention from the gods. This is not to say that the universe is driven by actual gearing, but does suggest that a physical, rather than divine or magical, explanation of its motions is possible, even if the actual details are as yet unknown. Paul Keyser points out that Theodoros, in the fifth century CE, certainly believed in mechanical determinism: the universe moves in a necessary motion like “a machine on wheels and pulleys built by an engineer.”10 Much earlier, in the fourth century BCE, Eudoxus of Cnidus, Callippus, and Aristotle suggested a universe constructed from nested crystal spheres—some 55 or 56 of them. The spheres could be thought of as a mechanical model even if this construction, they believed, perhaps required divine intervention for the rotations. Turning wheels had been conjectured as cosmic analogies since at least the time of Anaximander, who died around 546 BCE.

The second thread is that a single driver, such as the knob on the side of the Antikythera mechanism, turns the indicators on many dials.11 This can be interpreted as suggesting the universe can run all parts of itself, or alternatively that an underlying divine mover acts as a primum mobile, perhaps at a single point. A related but separate question is whether, since a model universe has a craftsman builder, the universe itself had a builder.12

Sylvia Berryman has emphasized the influence of, not sphaerae, but automata and pneumatic technology on the classical worldview.13 The Greeks had technologies that imitated the motions and even sounds of animals or humans by mechanical or pneumatic means. Viewers may have at first thought the effects seemed supernatural, but they must also have grappled with a physical explanation. The use of weight-driven or other power to provide independence of motion could again suggest a primum mobile. Interrelated ideas of the universe as a body, the body as a machine, and the universe as a machine are reflected in Galen’s comparison of planetary models and bodily functions, which was published in the second century CE.14

Both pneumatic automatons and mechanical displays of astronomical phenomena could and presumably did inspire consideration of a mechanistic view of the universe. Consideration did not, of course, mean acceptance of a mechanical view. But it cannot be doubted that the inspiration for a mechanistic view existed in the classical world—and that the view was held by some and rejected by others.

So what was the Antikythera mechanism—an entertainment device? A teaching tool? A record of the latest astronomical knowledge? A visualization for investigation of the heavens? A calculator for an astronomer?15 Perhaps all of these. There is no shortage of references to astronomical mechanisms, including sphaerae, in the ancient literature, but the surviving passages do not spell out their purpose or provide detailed technical descriptions.16 In the Antikythera mechanism, only the technology is novel. The astronomy it displayed is not surprising—it fits very well with our understanding of Greek knowledge around 150–50 BCE. This can be seen, for example, in the work of Geminus of Rhodes and, in particular, his astronomy textbook, Introduction to the Phenomena.17

It is worth noting that a mechanical universe is often referred to as a clockwork universe. These days, the term “clockwork” has become so familiar that some steps backward in thought are needed. To begin with, distinction must be made between a device that uses gears and a clock. The latter may use gears, but it has two other fundamental characteristics: an independent source of power, meaning that it moves itself; and a means of regulating its motion—either an escapement or a pendulum. The Antikythera mechanism does not exhibit either of these characteristics. Instead, it is a calendrical device, concerned with the organization of days, years, and lunations, and not a timepiece, marking the passing of time. The mechanism’s ability to move forward and backward through its calendars and zodiac display in prediction and retrodiction implies a concern for the passage of time, but it is not the relentless minute-by-minute forward motion of a clock. If the universe is to be called a clockwork universe, it must exhibit the same attributes as a clock. Indeed, the concept itself really only becomes meaningful with the advent of mechanical clocks in the thirteenth century CE, which four centuries later allowed Johannes Kepler’s desire “to show that the heavenly machine is not a kind of divine, live being, but a kind of clockwork.”18

An Era and Its Design Conventions

There are four ways in which the Antikythera mechanism is embedded in its era. The first, already mentioned, is its concord with contemporary knowledge of astronomical phenomena. Taken together, the others might be considered as a series of design conventions—a certain way of doing things that can persist for long periods. A comparison can be made here to architectural details, such as dentils—a row of jutting-out bricks or stones just beneath a roof line that originated in the roof beam ends on Greek temples. This feature can be found in buildings throughout history, even to the present day. There is no suggestion that the detailing implies a new building is a temple, it is simply a form that has comforting aesthetic continuity.

The first of these design conventions is the distinctive form of the gearwheel on the largest remaining fragment. It is the only gear that is not solid, and its four-quadrant shape resembles a chariot wheel. One turn of this wheel represents one year in the working of the mechanism, and it has 223 teeth, within an error of perhaps ± one tooth—the number of lunar months in a Saros cycle of eclipses. It seems natural to call it the Sun wheel, as the association of the sun with a chariot is well-known in the Greek myth of Helios.19 The wheel inside the case of the Antikythera mechanism would probably not have been visible to the user—but hidden messages are common in ancient artifacts. The chariot-wheel motif also appears on the head of a surviving pointer.

A second and more speculative association is with the spiral form of the Metonic and Saros calendrical dials on the rear face.20 Spirals seem to have been associated with calendrical function as far back as the Neolithic. Examples include the pictographs carved onto monuments aligned with the solstice, such as Long Meg and Her Daughters, the stone circle in Cumbria, England, and Newgrange and the Knowth calendar stone in Boyne Valley, Ireland.21

A third design element associates the mechanism’s front dial with an astrologer’s board, a surface on which the positions of the sun, moon, and planets could be depicted at a particular time for the purposes of preparing a horoscope. Evans has suggested that during an astrological consultation in Greco-Roman times, clients would have expected the seer to use an apparatus for their predictions, typically with the positions of the planets laid out using semi-precious stones.22 The growth of popular astrology in the Greek and Roman civilizations probably dates from the second century BCE. One board surviving from that time has a familiar concentric ring structure.23 The Antikythera mechanism bears the first known dials with inscribed graduated scales and pointers, following the tradition of markings on sundials. The inscriptions imply that pointers, or perhaps annuli, would carry around markers of the planets on the front face, certainly with the sun as a golden globe, and possibly with the planets as semiprecious stones.24 It is tempting to think of the Antikythera mechanism as a great sales tool for an astrologer. Although it would not have been particularly accurate in its predictions of planetary positions, it may not have been much worse than the perpetual tables astrologers would have used when placing their stones on a board.25

The Antikythera mechanism, it must be noted, is not necessarily an astrological device. But it is interesting—and possibly significant—that it shares design elements with astrological tools. Ptolemy rigorously separated his astronomical writings, the Almagest, from his astrological ones, the Tetrabiblos. But the first would certainly have informed the second.

The Missing Planetary Display

Inscriptions on the Antikythera mechanism leave little doubt that it was intended to show the motion of the planets against the background of the zodiac. The surviving parts of the mechanism—in particular the variable-speed lunar drive—confirm that the Greeks had the skill necessary to build planetary devices. A few references in classical literature also imply that devices to display planetary motion existed at the time.

Apart from a 63-tooth gear and a few bits and pieces which might have been involved, the relevant part of the Antikythera mechanism is missing. This gives rise to several questions: What happened to it? How exactly were the planets displayed? And how did this part of the mechanism work in the specifics of the gear, lever, and pins-in-slots trains? As perhaps might be expected, the paucity of material evidence has led to much speculation concerning the mechanical specifics.26 It is possible that more will never be known, unless additional parts are found at the location of the wreck or another mechanism is discovered.

Although it would be wonderful to know more details about the missing planetary display, they are not of fundamental significance compared to the knowledge that the Greeks could and did build such devices. The philosophical and technological implications are unlikely to be changed by the details.

Survival of the Technology

Whether or not sphaerae technology survived until the Renaissance remains unclear. I am inclined to follow Price, who believed it did,27 but a case can also be made for loss and reinvention. The technology might have been suppressed for religious reasons in later Roman days—certainly its suppression would only have been hastened if the sphaerae were associated with astrology.28 All that is known is that the technology persisted in Europe until at least 500 CE, and elements seem to have been reintroduced later through the Arabic world.

It is clear that Renaissance scholars knew the Greeks had made mechanical astronomical displays. This is attested, for example, by Giovanni de Dondi, who constructed an elaborate astronomical clock in approximately 1364 CE;29 by Kepler in his letters around 1605 CE;30 and in the writings of Conrad Dasypodius, who designed the Strasbourg astronomical clock around 1571–74 CE.31

Given that the Greeks could build the Antikythera mechanism, a common question is what other devices they might have created. Some aspects of the technology can be seen in surviving medical instruments, including small-bore tubes and worm gears. Although the Greeks had elementary lathes, files, and bronze-casting ability, the limited accuracy achieved in the manufacture of gears may explain why there is no evidence of calculators for financial or surveying use. Another deterrent to calculators being designed may have been the ready availability of labor skilled in the abacus and other basic counting devices.32

It was the lack of escapement technology that prevented the development of clocks, although some wheelwork was apparently used in clepsydrae. The use of large and crude wooden lantern gears continued in mills and other applications, but further development of practical mechanisms using small metal gears seems to have stalled. In explaining the lack of a classical industrial revolution and the emergence of precision manufacturing technology, many other considerations also come into play, in particular the abundance of slave and other labor, as well as the nature of pre-gunpowder military weapons.


History is always a reinterpretation of the past as seen through the ideas of the present. In the early nineteenth century, at a time when orreries were popular, scholars happily accepted the mechanical achievements of the ancient Greeks.33 By the late twentieth century, that acceptance had turned to a widespread sense of disbelief in the technological capabilities of ancient civilizations. Now, more aware of context, classicists can see that the Antikythera mechanism is remarkable but not anachronistic. It is an artifact of ingenious invention whose construction reveals much about its times.34


  1. Daryn Lehoux, “Clever Machines and the Gods Who Make Them: The Antikythera Mechanism and the Ancient Imagination,” in “The Scaffolding of Our Thoughts”: Essays on Assyriology and the History of Science in Honor of Francesca Rochberg, eds. C. Jay Crisotomo et al. (Leiden: Brill, 2018), 425, doi:10.1163/9789004363380_021. 
  2. The rediscovery of the mechanism in 1901 and the subsequent history of its investigation are now well documented. See Alexander Jones, A Portable Cosmos: Revealing the Antikythera Mechanism, Scientific Wonder of the Ancient World (New York: Oxford University Press, 2017); Alexander Jones, “‘Like Opening a Pyramid and Finding an Atomic Bomb’: Derek de Solla Price and the Antikythera Mechanism,” Proceedings of the American Philosophical Society 162, no. 3 (2018): 259–94; and Jo Marchant, Decoding the Heavens: Solving the Mystery of the World’s First Computer (London: Windmill Books, 2009).

    Descriptions of its structure, function, and inscriptions are readily available in John Seiradakis and Mike Edmunds, “Our Current Knowledge of the Antikythera Mechanism,” Nature Astronomy 2 (2018): 35–42, doi:10.1038/s41550-017-0347-2; and Mike Edmunds, “The Antikythera Mechanism and the Mechanical Universe,” Contemporary Physics 55, no. 4 (2014): 263–85, doi:10.1080/00107514.2014.927280, with a corrigendum in vol. 56, no. 1 (2015): 107, doi:10.1080/00107514.2015.986989. 
  3. Derek de Solla Price, “Gears from the Greeks. The Antikythera Mechanism: A Calendar Computer from ca. 80 B.C.,” Transactions of the American Philosophical Society 64, no. 7 (1974): 5, doi:10.2307/1006146. 
  4. Germaine Aujac, “La sphéropée, ou la mécanique au service de la découverte du monde (The Sphaerae, or Mechanics in the Service of the Discovery of the World),” Revue d’histoire des sciences et de leurs applications 23, no. 2 (1970): 93–107. 
  5. Tony Freeth et al., “Decoding the Ancient Greek Astronomical Calculator Known as the Antikythera Mechanism,” Nature 444 (2006): 587–91, doi:10.1038/nature05357; and, for example, Michael Wright, “Counting Months and Years, the Upper Back Dial of the Antikythera Mechanism,” Bulletin of the Scientific Instrument Society 87 (2005): 8–13, and Michael Wright, “The Antikythera Mechanism Reconsidered,” Interdisciplinary Science Reviews 32, no. 1 (2007): 27–43. 
  6. Daryn Lehoux, “Clever Machines,” 441. 
  7. James Evans and Christián Carlos Carman, “Mechanical Astronomy: A Route to the Ancient Discovery of Epicycles and Eccentrics,” in From Alexandria, through Baghdad: Surveys and Studies in the Ancient Greek and Medieval Islamic Mathematical Sciences in Honor of J. L. Berggren, ed. Nathan Sidoli and Glen Van Brummelen (Berlin: Springer, 2014), 145–74, doi:10.1007/978-3-642-36736-6_7. 
  8. Edmunds, “The Antikythera Mechanism and the Mechanical Universe.” 
  9. Samuel Sambursky, The Physical World of Late Antiquity (London: Routledge and Kegan Paul, 1962), xi. 
  10. Paul Keyser, private communication (2013). This “engineer” stimulated Proclus, too; see Sylvia Berryman, “The Clockwork Universe and the Mechanical Hypothesis,” British Journal for the History of Philosophy 29, no. 5 (2020), doi:10.1080/09608788.2020.1835605. 
  11. It could also have been operated with less torque but many more rotations by turning the lunar mechanism on its front face. Alexandros Basiakoulis et al., “The Handling of the Antikythera Mechanism,” Proceedings of the 6th International Conference on Manufacturing Engineering, Aristotle University of Thessaloniki, Thessaloniki, Greece, 2017, 281–92. 
  12. An argument aired by Cicero in De Natura de Deorum, II.88 (45 BCE). 
  13. Berryman, “The Clockwork Universe”; and Silvia Berryman, The Mechanical Hypothesis in Ancient Greek Natural Philosophy (New York: Cambridge University Press, 2009). 
  14. Galen, De Usu Partium Corporis Humani, 14.5, translated in Margaret Tallmadge May, Galen on the Usefulness of the Parts of the Body (Ithaca: Cornell University Press, 1968). The idea of a live universe, with body and soul, dates back to at least Plato’s Timaeus, ca. 360 BCE, and plausibly to prehistory. 
  15. For the last of the list, it should be realized that some of its calculations would not have been particularly accurate. See Mike Edmunds, “An Initial Assessment of the Accuracy of the Gear Trains in the Antikythera Mechanism,” Journal for the History of Astronomy 42, no. 3 (2011): 307–20, where it is shown that in predicting, for example, the date of a new moon, up to two days’ error could be made, although many of the calendrical predictions were fit for purpose. 
  16. A list is given in Edmunds, “The Antikythera Mechanism and the Mechanical Universe,” which may be supplemented by Aujac, “La sphéropée,” and Lehoux, “Clever Machines.” 
  17. James Evans and J. Lennart Berggren, Geminos’s Introduction to the Phenomena: A Translation and Study of a Hellenistic Survey of Astronomy (Princeton: Princeton University Press, 2006). 
  18. Max Caspar, Kepler (New York: Dover Edition, 1993), 136. 
  19. As can be seen in Attic vase decorations and the pediment of the Parthenon. See Jeffrey Hurwit, “Helios Rising: The Sun, the Moon, and the Sea in the Sculptures of the Parthenon,” American Journal of Archaeology 121, no. 4 (2017): 527–58, doi:10.3764/aja.121.4.0527. This sun-chariot tradition can be traced back earlier, perhaps even to around 1,400 BCE, as can be seen in the North-European Trundholm model showing the sun carried in chariot with four-spoked wheels. 
  20. The arrangement of the Metonic calendar on the mechanism is particularly neat—with lunar months for which a day should be omitted forming convenient lines. It is perhaps surprising that this design is not described in any known manuscript sources. 
  21. David Barrowclough, Prehistoric Cumbria (Stroud: The History Press, 2010); “Newgrange Chamber,”; “Ancient Sites: Knowth, Calendar Stone,” The latter two sites date from the fourth millennium BCE. 
  22. James Evans, “The Astrologer’s Apparatus: A Picture of Professional Practice in Greco-Roman Egypt,” Journal for the History of Astronomy 35 (2004): 1–44, doi:10.1177/002182860403500101. 
  23. Stephen Heilen and Dorian Gieseler Greenbaum, “Astrology in the Greco-Roman World,” in Time and Cosmos in Greco-Roman Antiquity, ed. Alexander Jones (Princeton: Princeton University Press, 2016), 123–41. 
  24. Ptolemy, Almagest 7, no. 1 “The Inscriptions of the Antikythera Mechanism” (2016). 
  25. E.g., see comments on the Jupiter table in Alexander Jones and Marco Perale, “Greek Astronomical Tables in the Papyrus Carlsberg Collection,” Archiv für Papyrusforschung und verwandte Gebiete 58, no. 2 (2013): 308–43, doi:10.1515/apf.2013.58.2.308, and estimates of errors in prediction of Mercury positions in Christián Carman and Gonzalo Reico, “Ptolemaic Planetary Models and Kepler’s Laws,” Archive for History of Exact Sciences 73 (2019): 39–124, doi:10.1007/s00407-018-0219-x. 
  26. For example: Jian-Liang Lin and Hong-Sen Yan, Decoding the Mechanisms of Antikythera Astronomical Device (Berlin: Springer, 2016); Jian-Liang Lin, “A Review of Reconstruction Research for the Lost Gearing of the Antikythera Astronomical Calculator,” IOP Conference Series: Materials Science and Engineering 658 (2019): 1–11, doi:10.1088/1757-899X/658/1/012014; and Tony Freeth et al., “A Model of the Cosmos in the Ancient Greek Antikythera Mechanism,” Nature Scientific Reports 11, no. 1 (2021), doi:10.1038/s41598-021-84310-w. 
  27. Derek de Solla Price, “On the Origin of Clockwork, Perpetual Motion Devices and the Compass,” United States National Museum Bulletin 218 (1959): 81–112. 
  28. Evaggelos Vallianatos, “Deciphering and Appeasing the Heavens: The History and Fate of an Ancient Greek Computer,” Leonardo 45, no. 3 (2012): 250–57. 
  29. Silvio Bedini and Francis Maddison, “Mechanical Universe: The Astrarium of Giovanni de’ Dondi,” Transactions of the American Physical Society 56, no. 5 (1966): 1–69. 
  30. Letter 99 in Johannes Kepler Gesammelte Werke, Band 13: Briefe I 1590–1599, trans. Max Caspar (Munich: C. H. Beck, 1982). 
  31. Dasypodius’s inspiration by the Greeks is discussed in Wilhelm Schmidt, “Heron von Alexandria, Konrad Dasypodius und die Straßburger Astronomische Münsteruhr,” Abhandlungen zur Geschichte der Mathematik 8 (1898): 177–94. 
  32. Practical mechanical arithmetic calculating devices were not developed until the mid-seventeenth century by Blaise Pascal and Samuel Moorland, and only became widespread in the nineteenth century. The latter part of that century also saw the development of ingenious mechanical calculating devices such as differential analyzers for ballistics and harmonic analyzers for tides. True computers only became viable in the twentieth century with the advent of electromechanical and electronic technology, and have had a fundamental impact on both observational and theoretical astronomy. 
  33. Abraham Rees, “Planetary Machines,” in The Cyclopaedia, or Universal Dictionary of Arts, Sciences and Literature, vol. 27 (London: Longman, Hurst, Rees, Orme and Brown, 1819). 
  34. I dedicate this essay to the memory of John Seiradakis (1948–2020), a cofounder of the AMRP, and a kind, stimulating, and much-missed colleague. Vassiliki Kalogera and Michael Kramer, “John Hugh Seiradakis,” Nature Astronomy 4 (2020): 639–40.

    The title of this essay comes from Cassiodorus, ca. 506 CE, describing an astronomical mechanism. S. J. B. Barnish, Cassiodorus: Variae (Liverpool: Liverpool University Press, 1992).

    I am grateful for discussion and correspondence with Alex Jones, other members of the AMRP, Sylvia Berryman, Mark Edmonds, Jim Evans, Paul Iversen, Paul Keyser, and Daryn Lehoux. 

Mike Edmunds is Emeritus Professor of Astrophysics at Cardiff University and the current President of the Royal Astronomical Society.

More on Archaeology


Copyright © Inference 2024

ISSN #2576–4403