A Large Language Model in the 1700s?

By Fernando Rodriguez Villegas

I listened to a fascinating podcast in the “Past, Present, Future” series by David Runciman on the 1726 satire “Gulliver’s Travels” by Jonathan Swift. Runciman gives an excellent summary of the various journeys that Gulliver undertakes in this book; journeys to far away lands were he encounters fantastic creatures and their world: tiny people, giants, talking horses, and the like.

In one of these journeys Gulliver finds academics stubbornly determined to strip language of all ambiguity. They start by taking out adverbs, adjectives, articles and so on and eventually find themselves with no words left. To communicate they are forced to carry around bags full of objects. If they wanted to talk about a shoe say, they would need to rummage in their bag until they find one to show.

The most astonishing episode though is about the “Engine”. Here Swift describes a machine, thought to be based on Ramon Lull’s (b. 1230’s) combinatorial philosophical system described in his book “Ars Magna”.

One may say that Lull wanted to create an algorithm that by permutation of letters, words and basic concepts could arrive unambiguously at the truth.

To quote Swift on the “Engine”:

“… Every one knew how laborious the usual method is of attaining to arts and sciences; whereas, by his contrivance, the most ignorant person, at a reasonable charge, and with a little bodily labour, might write books in philosophy, poetry, politics, laws, mathematics, and theology, without the least assistance from genius or study.”

As Runciman and others have pointed out Swift’s “Engine” sounds eerily just like ChatGPT. What would Swift write if he traveled to our world today?