Literary Theory for Robots: How Computers Learned to Write

“Colorless green ideas sleep furiously.” Derived from Chomsky’s procedural grammar, the sentence may be correct according to the rules of language. But to know that ideas don’t sleep, or that one does not sleep furiously, requires experience in the physical world.

The world itself attains universality only from a great distance, described in broad strokes by physics or theology. Its particulars often differ depending on the local context.

The smart table, in the words of Karl Marx, another furniture enthusiast, “changes into something transcendent,” a fetish, turned onto its head, fashioning grotesque ideas out of its wooden brain, as it recedes into the backdrop of ordinary commodities.

A “smart” device is merely smarter than the previous generation. Once adopted widely, it passes into the average.

Artificial intellect thrives in the gap between the average and the exceptional.

In the applied world, intelligence bubbles above shared capacity. But no matter how cleverly effervescent, once a smart tool reaches general adoption, it dissipates into the surface of baseline intelligence.

The best of us occurs more rarely in comparison to our base impulses. Frequency alone therefore does not suffice for intelligence, in a basic pedagogical sense.

This new language would increase the power of the human mind, far more so than microscopes or telescopes amplify sight. Without it, we are like merchants, he wrote—in debt to each other vaguely for various items mentioned in passing, but never willing to strike an exact balance of meaning.

From the nineteenth century onward, the notion of template-based manufacturing permeated all human industry, and especially the manufacture of consumer and capital goods, like clothing, furniture, machinery, or equipment.