We tend to think of AI in isolation, as a somewhat scary entity shaped by the minutiae of machine learning algorithms and experiments. But artificial intelligence is actually part of a much larger lineage of “smart” machines that stretches back to the 14th century.
Many years ago, the Arabic astrological wheel was created as an early attempt at automated chatter, followed by the creation of “mathematical cabinets” in 17th century Europe. Attempts to automate poetry, music, and coded messages.
In the 1930s, there was an American robot designed to spit out story ideas for Hollywood. Later, features like spell checking, text messaging, and auto-correction came along.
This is the fascinating journey that scholar and software engineer Dennis E. Tennen takes in his new book, Literary Theory for Robots: How Computers Learned to Write .
His aim, he says, is to remove some of the fear and confusion that arises from how we view this technology, and to remind us that humans have always woven the threads of language and technology together, driven by a primal urge to protect and promote intellect and knowledge.
“Language and technology didn't just evolve together; language is a technology,” says Yi Tenen, an associate professor of English and comparative literature at Columbia University. “Many animals can teach their young, but only humans can do it remotely, for example by storing and transmitting collective know-how across time and space.”
Can this quest go too far? “Of course it can. I find it difficult to balance using technology in my own life,” says Yi Tenen.
Of course, with AI, the struggle is not quite the same: it is potentially a struggle over the nature of reality, veracity, truth.
That's why it's even more important to view this technology accurately and dispassionately as a product of human endeavor, rather than as an entity in its own right, Yi Tian-en says. If we view it as an extension of our collective intelligence, we can hold tech makers responsible and accountable, he adds.
His book therefore takes a step back into the historical context of AI. How far back can we go in our efforts to create “intelligent” machines? Let's see.
The prophetic wheel, 14th century
Made in Arabia, the device was a maze of concentric circles and elaborate diagrams depicting constellations, letters, symbols and numbers. Code zigzagged across the whole thing. A manual or rule book of sorts, mostly about astrology, helped the fortune teller interpret the readings.
But what's interesting is that some of the circles contain verses from the Quran or other holy texts. “And by interacting with the circles, you can actually have a conversation with the device,” Yi Tenen says. “The fact that you're still in Q&A mode in ChatGPT… I thought that was a really cool parallel.”
Smart Cabinet, 17th century
The earliest literary bots were shaped like giant cupboards and operated as a kind of word loom.
One of the earliest examples is the mathematical organ built by the 17th-century polymath Athanasius Kircher. Made of painted wood, the organ was a box-shaped structure made of wooden slats or bars arranged vertically and horizontally. Each box was accompanied by paper booklets on subjects such as arithmetic, music, geometry, chronology, and astrology.
In an eerie bit of foreshadowing, Kircher called them “applications.”
By manipulating these rods and consulting booklets arranged in a matrix, users could compose poetry or music (kind of), write coded messages, or perform certain kinds of mathematical or astronomical calculations.
Kircher sold a version of this device to the young Archduke Karl Joseph of Austria to aid in his research.
This gave rise to several controversies, including a public debate in a bar between Kircher and the German poet Quirinus Kuhlmann, who argued that the path to knowledge should be “painful and accessible only to those willing to walk it correctly.”
It was not the organ that had the intelligence, the poet added, but Kircher: “Without the organ the young duke would remain a stupid parrot.”
Essentially, it's an early example of the argument that computers are making humans lazy and stupid.
Analytical Engine, 1830s
It was Kircher's mathematical apparatus that inspired British mathematician Charles Babbage in the 1830s to create what went down in history as the world's first computer.
Babbage called it the Analytical Engine and worked with the countess, mathematician Ada Lovelace, to develop its functions.
In the end, it looked like a weaver's loom. This machine, which Babbage only partially built during his lifetime, had a mill (like a central processing unit), a store (like a memory device), a reader (input device), and a printer (output device). The mill could process calculations with the help of punch cards inserted into the reader.
Thirty-six Dramatic Situations, 1895
The 19th century, the beginning of the Industrial Revolution, saw the emergence of templates for furniture, machines, clothing, equipment, art (literature, film, music), journalism, and even philosophy.
In the art world, templates have always been disparaged as a poor substitute for missing genius.
But Ee Tenen argues that outlining has always been at the heart of how we learn.So can machines help less-than-great writers write great stories?Long before ChatGPT, in 1895, the outline generator was developed by French author Georges Porty.
The book, entitled Thirty-Six Dramatic Situations, lists 34 predicaments designed to help the embattled playwright, including plea, rescue, revenge, pursuit, disaster, rebellion, rivalry and adultery, each briefly explained with an example.
The chase, for example, came with instructions that the plot should be “in the grasp of the fugitive alone”: “Sometimes innocent, but always pardoned, for if there has been a fault, it seems inevitable, predestined. We neither inquire into it nor condemn it; it would be useless; whatever our heroes may once have been, they are now merely companions in peril, and we bear the consequences sympathetically.”
He invented a series of dramatic situations whose elements could be combined to form hundreds of thousands of stories. What is less known, Yi Tian-en says, is that such templates were frequently used by writers well into the cinematic era.
Built by Los Angeles screenwriter Wyclef Hill, the machine could produce a complete story outline in 20 minutes, with its gears spinning to map out settings, characters and dramatic situations from a series of tapes stored on a chassis.
The robot never found a buyer (for reasons unknown, perhaps due to cost or perhaps because the story just wasn't that good), but it inspired Hill to write a book called The Plot Genie (1935), about a wheel that, when spun, would suggest seemingly random numbers and then generate plot requirements (locations, characters, loved ones, problems, etc.) by checking those numbers against a series of charts.
After all, Hill was no polymath or early technocrat. He was a screenwriter who, when his own scripts were rejected, became obsessed with deciphering the intricate elements that make up a good plot. His aim was to take the effort and uncertainty out of writing.
The magic wand that will make this possible has not yet been created.
