Phronesis
No. 002Conceptantiquity → 1920Artificial Intelligence ·History ·Philosophy ·Technology

The dream of thinking machines is thousands of years old

Long before computers, people imagined bronze giants, reasoning wheels and a language in which arguments could be settled by calculation.

2 min read

Why would anyone want a machine that thinks? The question is older than machines that could do much of anything.

Myths of made minds

Greek mythology has Talos, a giant of bronze made by the smith-god Hephaestus, who walked around Crete and threw rocks at invading ships. Hephaestus was also said to have golden mechanical servants. Many cultures told stories of statues brought to life. These were stories about making minds, long before anyone knew how.

Reasoning with wheels

In the 13th century, the Majorcan philosopher Ramon Llull designed paper “logical machines”: rotating discs inscribed with basic concepts, which could be turned to produce every combination. His goal was to demonstrate religious truths through mechanical combination, and the idea that thinking might be a kind of systematic combining had arrived.

“Let us calculate”

In the 17th century, Gottfried Wilhelm Leibniz, who also co-invented calculus and built a calculating machine, dreamed of two things:

  • a characteristica universalis: a precise symbolic language for all ideas;
  • a calculus ratiocinator: rules for reasoning in that language mechanically.

If both existed, he suggested, people could settle disagreements not by arguing but by sitting down together and calculating. This is the deep root of AI: the hope that reasoning is a form of computation.

Bring it up

“The word ‘robot’ comes from a 1920 Czech play, and it comes from the word for forced labour.”

Follow the thread

You might also find interesting