Interlude: Beyond the Delivery of Information

In July 1948, a thirty-two-year-old engineer at Bell Labs published a paper that made no headlines. It ran forty-four dense pages in the Bell System Technical Journal, and it did something almost nobody had thought to do before: it gave the word "information" a precise, measurable meaning.

Before Claude Shannon's "A Mathematical Theory of Communication," information was a human idea, felt rather than counted: a good letter, a useful signal, an important message. Shannon reduced it to something a machine could work with, a sequence of yes-or-no choices, a string of what a colleague at Bell Labs had already started calling a "bit." A telegram, a photograph, and a phone call were suddenly, underneath, the same kind of thing: information, encoded in binary, indifferent to what it was made of.

This is the quiet turn this series has been circling toward since Part I's postlude. Every technology in the section above solved the same problem, how to move a message from one place to another faster and more cheaply than the last. None of them asked what a message actually was, or what a machine might do with one besides carry it. Shannon's method is the reason a computer can hold a photograph, a symphony, and a sentence in the same memory, as the same kind of thing.

Two years later, in 1950, the British mathematician Alan Turing, who had spent the war years breaking codes, asked a stranger and far more provocative question: if a machine could hold information the way a mind holds a thought, could it reason with that information the way a mind does?

Little did humanity know the significance of that question.