Chapter 2: The Clockwork Universe

The mechanical worldview didn't arrive from nowhere. It inherited its gears.

When European scholars began translating Arabic texts (in Toledo, in Sicily, in the polyglot courts of medieval Spain) they encountered engineering traditions far more sophisticated than their own. The programmable flute players and hydraulic automata of the Islamic world, described with such precision that skilled craftsmen could replicate them, flowed north along with algebra, optics, and medicine. When Leonardo da Vinci sketched his mechanical knight in the 1490s, when the clockmakers of Geneva and Paris built their marvels in the 1700s, they worked within a tradition they often did not acknowledge.

The clockwork universe had foundations laid elsewhere. Europe built upon them.

The seventeenth century brought a new idea, more radical than any particular mechanism: the suggestion that the universe itself was a machine.

René Descartes, writing in the 1630s and 1640s, proposed that animals were automata—complex mechanisms operating by purely physical laws. A dog, in this view, was an elaborate clockwork. Its behaviors, however lifelike, required no soul, no vital spark, no animating spirit. Physics sufficed.

Descartes hedged when it came to humans. He proposed a dualism: the body mechanical, but the mind something else entirely, interacting with matter through the pineal gland. This was philosophically unsatisfying (how could an immaterial mind push material atoms?) but it preserved human uniqueness. We were machines possessed by ghosts.

Others pushed further. Thomas Hobbes argued that even thought was motion—"a motion of certain parts of an organic body." Pierre Gassendi reduced everything to atoms and their collisions. The universe became, in this telling, a vast clockwork set in motion at creation and ticking along by deterministic laws ever since.

This was the mechanical philosophy, and it raised a question its proponents could not yet answer: if the universe is a machine, and animals are machines, what exactly prevents thought itself from being mechanical? If calculation follows rules, and rules are mechanical, then a machine might one day calculate. And if reasoning is a species of calculation...

Gottfried Wilhelm Leibniz saw the implications.

In 1673, the German polymath began constructing a device he called the Stepped Reckoner. Unlike earlier calculating aids, which could only add and subtract, Leibniz's machine would perform all four arithmetic operations. Inside its brass housing, a cylindrical gear of his own invention, the stepped drum (later called the Leibniz wheel), enabled multiplication and division through a variable-ratio mechanism. The idea was elegant: a gear whose effective size could be changed by shifting which of its teeth engaged with the counting mechanism.

But elegant ideas and working machines are not the same thing.

"The Reckoner suffered from one great drawback, much more serious than its inability to carry or borrow numbers automatically," one historian noted. "It didn't work."

The gears jammed. The teeth, hand-filed in an era before precision manufacturing, varied in height. The carry mechanism, meant to automatically advance higher-order digits, was too weak for reliable operation. Leibniz spent over twenty years refining his device. Two prototypes survive. Neither functions properly.

And yet the Stepped Reckoner succeeded where it mattered most: as proof of concept. The Leibniz wheel remained the basis of mechanical calculators for over 275 years, used into the 1970s in devices like the Curta hand calculator. The principle worked even when the particular machine didn't.

More importantly, Leibniz saw past arithmetic to something grander. He dreamed of a calculus ratiocinator—a universal logical calculus that could reduce all reasoning to mechanical symbol manipulation. "When there are disputes among persons," he wrote, "we can simply say: Let us calculate, and see who is right."

The Stepped Reckoner was a physical fragment of this dream: a machine that could calculate, as humans calculate, by following rules encoded in metal. That the gears jammed was a problem of manufacturing. That reasoning might be mechanical—this was the genuine philosophical breakthrough. It would take another century and a half before anyone tried seriously to build on it.

While Leibniz wrestled with unreliable gears, other craftsmen built machines designed not to calculate but to astonish.

In Switzerland, between 1768 and 1774, Pierre Jaquet-Droz and his son Henri-Louis constructed three automata that remain operational today. The Musician, a female figure, plays five melodies on an actual organ—the music comes from her fingers on the keys, not from a hidden music box. Her chest rises and falls with simulated breath. Her head and eyes track her hands as she plays.

The Draughtsman produces four different images: Louis XVI, Marie-Antoinette, a dog, and Cupid driving a chariot. He pauses periodically to blow on his pencil, clearing dust. The mechanisms encoding his movements (cam systems translating rotation into two-dimensional drawings) contained two thousand parts.

Most remarkable was The Writer. This small figure could produce any custom text up to forty letters. He dips his goose-feather quill, shakes off excess ink, and writes in a flowing script. His eyes follow the text as it emerges. Inside him, six thousand components implement what we would now call programmability: change the arrangement of cams on a coded wheel, and The Writer produces different words.

The Jaquet-Droz automata were marketing tools. Pierre ran a successful watch business; the automata drew visitors who then purchased timepieces. But they were also philosophical demonstrations. If a mechanism could write—could produce novel text based on encoded instructions—then writing itself was mechanical. The mystery retreated.

Not all demonstrations were honest.

Jacques de Vaucanson, working in France in the 1730s, created what he called the Digesting Duck. The device, he claimed, could eat kernels of grain, metabolize them through a "small chemical laboratory" in its body, and defecate the results. Over 400 moving parts in each wing alone. Voltaire praised it as evidence of French glory.

A century later, the magician Jean-Eugène Robert-Houdin examined what remained of the duck. He found that the digestion was entirely faked. The grain went into one container; pre-prepared green pellets (breadcrumbs dyed to resemble feces) were expelled from another. No metabolism occurred. Robert-Houdin, himself a master of illusion, was amused: "I found that the illustrious master had not been above resorting to a piece of artifice I would happily have incorporated in a conjuring trick!"

The deception mattered philosophically. Vaucanson's duck had been used to support the mechanical worldview—see, even digestion can be replicated in brass and gears! When the fakery emerged, it recast the entire automaton tradition. These devices were less proof of mechanism's reach than demonstrations of how badly audiences wanted to believe. The duck exploited "spectators' willingness to suspend disbelief in vital processes."

Vaucanson's failure—his fraudulent success—clarified something important. Building machines that mimicked life was not the same as understanding life mechanically. The gap between appearance and reality would haunt artificial intelligence for centuries to come.

The critics had their own clarity.

Vitalists like Georg Ernst Stahl argued that living beings possessed something machines lacked: a vital force, an organizing principle irreducible to gears and levers. Mechanism, they charged, overlooked "the organic interdependent relationships found within a being." A clock could be disassembled and reassembled; an animal could not. A clock did not grow, reproduce, heal, or die. To call both "machines" was to abuse language.

Other critics worried about determinism's implications. If the universe was a vast clockwork, then every event—including every human thought and choice—was fixed at creation. Free will became illusion. Morality collapsed. Religious thinkers argued that mechanism demanded a God who wound the clock and walked away, abandoning creation to its gears. This was incompatible with a deity "at all times directly and immediately involved in his creation."

The mechanists had no satisfying answers. They could point to their marvelous automata, their calculating machines, their philosophical systems. But they could not explain consciousness, could not account for the felt sense of willing, could not bridge the gap between mechanism and meaning.

"When one reads the writings of one of the leading vitalists like Driesch," one historian observed, "one is forced to agree with him that many of the basic problems of biology simply cannot be solved by a philosophy in which the organism is simply considered a machine... The logic of the critique of the vitalists was impeccable."

The mechanical philosophy had identified the right question: can mind be understood mechanically? But it had not yet found the right answer. It had built spectacular machines that mimicked thought without thinking, displayed behavior without understanding, performed without being. The gap between performance and genuine cognition remained.

And yet the dream endured.

The eighteenth century established that mechanism was thinkable—that one could imagine, in principle, a machine that reasoned. Leibniz had sketched the vision. The automata had demonstrated that complex behavior could emerge from rules encoded in gears. The frauds had revealed how much remained to be solved.

The failures composted into progress. Each jammed gear, each exposed deception, each vitalist critique sharpened the question. Not "can machines move like living things?" but "what would it actually mean for a machine to think?" Not "can we fool audiences?" but "can we build genuine intelligence?"

In the 1830s, an eccentric English mathematician would take up Leibniz's abandoned project. Charles Babbage would design engines of calculation far beyond anything the stepped reckoner attempted. And his collaborator, Ada Lovelace, would ask a question that had not occurred to her predecessors: could such a machine do more than calculate? Could it create?

The clockwork universe was about to give way to something more powerful: the programmable machine. And with it, the question of artificial intelligence would shift from philosophical speculation to engineering problem.

The gears were ready to become something else.