June 9, 1949 — The Mind of Mechanical Man

“Can machines think?” That question has been on everyone’s mind over the last few years. The rise and widespread adoption of neural networks has sparked large-scale discussions about intelligence, the nature of consciousness, emotions, and many other related topics.

As is often the case, professionals were debating these issues long before ChatGPT appeared. Geoffrey Jefferson’s lecture The Mind of Mechanical Man was published almost 80 years ago, yet many of the arguments in it sound surprisingly familiar today.

On June 9, 1949, Jefferson delivered The Mind of Mechanical Man at the Royal College of Surgeons of England. The venue was no coincidence: Jefferson was a neurosurgeon. As someone who studied the brain professionally, he had a deep understanding of how it was believed to work in the 1940s. At a time when electronic engineers were becoming fascinated by rapidly advancing computers and beginning to compare machine intelligence with the human mind, it was perhaps inevitable that experts on the brain would challenge those comparisons.

Jefferson believed such comparisons were greatly overstated. He acknowledged that machines could perform complex calculations and imitate certain aspects of human behavior, but argued that there was a vast gulf between computation and thought.

His most famous statement was roughly this:

“Not until a machine can write a sonnet or compose a concerto because of thoughts and emotions felt, and not by the chance fall of symbols, could we agree that machine equals brain.”

Jefferson went even further, arguing that a machine would also have to understand what it had created.

An interesting point is that today a machine can certainly write a sonnet, yet it is difficult to claim that it does so based on its own thoughts and emotions. That leads to another question: what exactly is “understanding”? How do we know that we ourselves understand something? How can we determine whether a machine understands anything at all? Answering those questions requires digging deep into several fields of study, and in some of them we still lack enough evidence to draw firm conclusions.

The lecture had a lasting impact on the industry. In 1950, Alan Turing published his landmark paper Computing Machinery and Intelligence, where he directly addressed Jefferson’s position. Public debates between great scientists and engineers have often helped entire industries move forward, and this was no exception. Jefferson’s lecture became part of the intellectual foundation of early artificial intelligence research, influencing how many of the era’s leading thinkers framed and discussed the possibility of machine intelligence.

Turing’s Response Was Immediate

Just two days later, on June 11, 1949, Alan Turing gave an interview to The Times in which he directly addressed the question of whether a machine could write a sonnet. It may be a coincidence that two thinkers arrived at the same comparison independently — or perhaps the debates around computers and what was essentially artificial intelligence in 1949 moved just as fast as they do today.