Summary of the ‘Ancient Thoughts on Robots and Science’ AMME Seminar

This spring’s AMME seminars concluded on April 27th with two talks delving into ancient automata and their philosophical and ideological significance.

The seminar’s first speaker, Assoc. Prof. Dr. Elly R. Truitt, gave a fascinating presentation on the history of real and imagined artificial servants, focusing on Latin Christendom. In Greek culture, ideas of automata, mechanical animals, and sentient machines serving as autonomous servants first appeared in the 8th century BCE. The first evidence of constructed automata comes from the 3rd century BCE in Alexandria. Engineers of the Alexandrian school wrote manuscripts detailing the construction of mechanical objects that acted out ritual scenes, such as those performed in temples or theatres, and instructed in their underlying mechanical principles. These automata gained international fame and became emblematic of Greek culture to the outside world, as seen in Chinese, Sanskrit and Tibetan Buddhist legends about foreign automaton-makers. 

While these ideas largely disappeared from the Latin Christian West during the medieval period, the tradition of automation-making continued in the Byzantine Empire and the Islamicate world, especially in Baghdad. New manuscripts were created based on Greek texts but with more technological complexity, improving upon old knowledge. The objects described in these designs included water clocks and water fountains, artificial animals, mechanical servants, and musical automata. These objects were found in places of power, mosques and palaces, highly socialised environments with ritualised behaviours. They were also exchanged as diplomatic gifts, exemplified by the first automaton described in Latin Christendom—a mechanical clock sent by Caliph Harun al-Rashid to Charlemagne. Whether given as diplomatic gifts or positioned to awe visitors to places of power, these objects served to legitimize and enforce the power of rulers. 

These kinds of objects were not actually produced in Latin Christendom before the 14th century. Before that, they lived in the imagination, where they were strongly associated with the past and the Hellenistic world. One story tells of the Alabaster chamber of the ancient Trojan court, which employed four automata that supervised and controlled the behaviour of visitors, entertaining and surveilling the court while ensuring loyalty and order. When automata began to be built, they likewise appeared in courts. There, they again displayed the power of the ruler and aroused awe, surprise, and fear in his subjects.  

Truitt ended the presentation on some thoughts that ancient automata can give us about our own relationship to artificial intelligence. As seen above, these objects always emerge in contexts of authority and power. As such, they cannot be tools for liberation. Perhaps looking at the past can help us understand what we need to do to create technology that is truly liberatory and revolutionary. 

The second speaker was Prof. Dr. George Saliba, who delved further into the mechanical devices of Islamdom, especially the philosophical principles guiding their creation. One such principle was the avoidance of “useless knowledge”. One should not design mere toys; the items created by the mechanical sciences should have a societal benefit. The texts did not only describe the creation of devices, but also instructed in their operation and emphasized to the reader the importance of testing the produced device in practice. The ethics of the mechanical sciences placed emphasis on technē, the goal being to create useful things, not knowledge for its own sake. 

As Truitt mentioned that Greek devices were meant to illustrate their underlying mechanical principles, the machines of Islamdom similarly sought to illustrate the principles of nature. For example, the principle of void and plenitude, or the Aristotelian idea of nature abhorring and avoiding void. Mechanical devices were meant to bring forth (ikhrāj) such hidden principles from potentiality into actuality. This defining goal of the mechanical sciences lead to the inclusion of algebra among them; algebra was thought to bring forth the qualities of numbers from inexpressible surds into actuality, making them expressible.  

Authors were conscious innovators, interested in the creation of new things. In a manuscript describing an escapement design, a mechanism used in clocks, Isma’īl al-Jazarī claims to have invented the design himself. Many authors would make similar claims for their designs, writing that they were not aware of anyone introducing a similar device before, claiming the honour of being first for themselves. They would acknowledge existing ways of doing things and explicitly divert them. Scholar Sinān ibn Thābit even attempted to systematise the field of mechanical sciences by reducing devices into their essential parts, such as the lever, pulley, and wedge. The scientists of the Medieval Islamic world were not passive carriers of old knowledge, but active innovators and inventors. 

The audience questions following the talks also lead to fascinating avenues of discussion. One audience member asked whether we know of any failed attempts at creating machines from the medieval Islamic world, where the emphasis was so heavily on their functionality. Saliba remarked that there are surviving manuscripts whose edges have notes explaining that an instrument was tried unsuccessfully, and that the reader should not bother constructing it. Such warnings were sometimes removed when the manuscripts were copied, with potentially dangerous results. Another choice discussion regarded the way automata was projected into the past, making them emblems of patrimony for the rulers who owned them.  

Thank you to the speakers and attendees of this session, and thank you to the speakers and attendees of all previous sessions of this spring! We hope you have a great summer and join us again when AMME seminars return in the fall.