Historical Context

The world in 1950

To understand why this paper was shocking, you need to understand 1950.

World War II had ended just five years earlier. The war had produced something remarkable: the first electronic computers. Machines like ENIAC in America and Colossus in Britain were enormous — they filled entire rooms, weighed tonnes, and ran on thousands of fragile vacuum tubes. They were, by today’s standards, stupendously primitive. ENIAC could do about 5,000 additions per second. Your phone does billions.

But to the people who built them, these machines felt like something almost miraculous. For the first time in history, a machine could execute a sequence of logical instructions automatically. You didn’t pull a lever — you wrote a program.

Alan Turing was, more than almost anyone alive, responsible for the theoretical foundations of these machines. In 1936, when he was just 24 years old, he had published a paper describing a hypothetical machine — now called the Turing Machine — that could, in principle, simulate any computation that any computer could ever do. This was a mathematical proof, not a blueprint. But it established what computation means, in a deep and rigorous sense.

During the war, Turing worked at Bletchley Park, the British code-breaking centre. He led the team that cracked the German Enigma cipher — a feat that historians believe shortened the war by two years and saved perhaps 14 million lives. He did this by thinking like a machine: finding the patterns, the logical structure, the repeating elements that a human brain glosses over but that a systematic process can exploit.

By 1950, Turing was at the University of Manchester, helping build one of the world’s first stored-program computers. He was thinking hard about what these machines could — and couldn’t — do.

What did people believe about machines then?

The dominant view in 1950 was simple: machines compute, humans think. These are completely different things.

A calculator can add numbers. A loom can weave patterns. A clock can track time. Machines follow fixed rules, mechanically, without any understanding of what they are doing. A clock does not know what time it is. It just moves.

The idea that a machine might actually think — reason, understand, be creative, have opinions — was considered by most serious thinkers to be not just incorrect but absurd. Thinking was the uniquely human thing. It was what separated us from every other creature and certainly from any artefact we could build.

Philosophers had a word for this: dualism. Going back to René Descartes in the 17th century, the idea was that mind and matter are fundamentally different substances. Bodies are physical. Minds are something else — not reducible to physics or mechanism. A machine is pure matter. Therefore a machine cannot have a mind. QED.

Religious thinkers agreed. The soul — the seat of thought and consciousness — was given by God to humans alone. To suggest a machine could think was not just scientifically wrong, it was theologically offensive.

The dominant approach before this paper

There was no “AI” before this paper. The field did not exist. There was no approach to replace.

What existed was a set of assumptions:

Mathematics and logic can be mechanised (Turing himself had proved this, and Bertrand Russell and others had shown it before him)
But mathematics and logic are not the same as intelligence
Intelligence involves intuition, creativity, emotion, common sense — things no machine could ever have

The closest thing to a “field” dealing with these questions was philosophy of mind, which had been arguing about the nature of thought for centuries without reaching any consensus. And there was early cybernetics — the study of feedback and control in systems — which Norbert Wiener was developing at MIT. But cybernetics was about control, not thinking.

Who was Turing, and what was he trying to do?

Alan Turing was, by 1950, famous in academic circles for his mathematical work but not yet the legend he would become. (He died in 1954, at just 41, and was not properly celebrated until decades later.)

He was thinking about a very practical question that his work with computers had raised: if we keep building faster and more capable computers, will we eventually build one that thinks? And if so, how would we know?

This is not an idle question. If you believe machines can never truly think, then no matter how sophisticated they become, you will always be able to dismiss their behaviour as “mere computation.” But if you cannot define why they cannot think — if you cannot point to something specific that is missing — then your dismissal is just prejudice, not argument.

Turing wanted to cut through the fog. He wanted to replace a vague, unanswerable question with a clear, testable one. And in doing so, he invented a new way of thinking about intelligence itself.

Next: The Problem →