Alan Turing: The Mind Britain Needed and Broke

00:00:10 I am Alan Turing. This is an AI-written historical reconstruction, not an

00:00:15 authentic recording or a transcript of my words. In nineteen fifty-two, I

00:00:20 reported a burglary at my home. I went to the police as the victim of a crime.

00:00:24 Their investigation uncovered my relationship with another man, and I

00:00:29 became the person facing prosecution. During the war, the British state had

00:00:34 depended on work I could not discuss in public. Now that same state was

00:00:38 examining a part of my life that should not have been its business. There was no

00:00:43 stolen military secret at the centre of the case. There was a consensual

00:00:48 relationship. To understand how I reached that point, we need to leave the

00:00:52 police investigation behind for a moment. My story began with questions, long

00:00:57 before it became a story about a verdict. I was born in London on the

00:01:02 twenty-third of June, nineteen twelve. At school, mathematics and science held

00:01:07 my attention more reliably than the routines meant to turn a boy into a

00:01:11 polished English gentleman. I wanted to understand how things worked. Meeting

00:01:16 expectations about how I ought to work was another matter. At Sherborne, I met

00:01:21 Christopher Morcom. Here was someone with whom I could share an absorbing

00:01:25 interest in mathematics and astronomy. He was not simply a clever schoolmate. I

00:01:30 was deeply attached to him. Christopher died in nineteen thirty. We had imagined

00:01:35 university ahead of us; he never reached it. In the years that followed, I wrote

00:01:41 about mind, matter and what might survive a person's death. Those early

00:01:45 reflections were not scientific discoveries. They were a young man's

00:01:49 attempt to think through a loss that science could not simply undo. In nineteen

00:01:54 thirty-one, I entered King's College, Cambridge, to study mathematics. Four

00:01:59 years later, I became a Fellow. Max Newman's lectures brought me to a problem

00:02:04 far beyond a difficult examination question: could a fixed procedure always

00:02:09 decide whether a given statement in mathematical logic was provable? Before

00:02:13 answering, I had to make something surprisingly ordinary

00:02:17 precise. What, exactly, does it mean to calculate?

00:02:20 Begin with a person at a desk, following instructions with a pencil and paper.

00:02:25 Put aside flashes of inspiration. Consider only the steps that can be specified

00:02:30 so clearly that nothing is left to guesswork. Read a symbol. Write another.

00:02:34 Move to the next place. Follow the rule appropriate to the current situation. In

00:02:39 my nineteen thirty-six paper on computable numbers, I described an imaginary

00:02:44 machine built around those simple operations. It had a tape divided into

00:02:48 squares, a mechanism for reading and writing symbols, and a finite set of

00:02:52 instructions. The tape could extend as far as the calculation required. This was

00:02:57 not a machine waiting in a workshop. It was a mathematical model: a way to say

00:03:02 exactly what a rule-following calculation could do. Then came the important

00:03:06 step. The instructions for one such machine could themselves be written as

00:03:11 symbols. A universal machine could read that description and imitate the other

00:03:16 machine's calculation. Give it a different description, and the same universal

00:03:20 machine could carry out a different task. The task changed because the

00:03:25 instructions changed. We did not have to build a new calculating mechanism every

00:03:29 time. That is the heart of the idea. Not a machine that knows everything, but

00:03:34 one machine capable, in principle, of carrying out any computation expressible

00:03:39 in this model, given enough time and tape. Yet the same investigation also

00:03:43 revealed a limit. There are precisely stated problems for which no algorithm can

00:03:48 always produce the required correct answer. This is not merely a complaint

00:03:53 about slow machinery. More speed cannot supply a general procedure that

00:03:57 mathematics shows cannot exist. The universal machine and the boundary of

00:04:01 computation belonged to the same piece of work. Its promise was enormous, but it

00:04:06 was never a promise of omniscience. Alonzo Church had independently reached a

00:04:11 closely related negative result by another route. I went to Princeton to work

00:04:16 with him and completed my doctorate in nineteen thirty-eight. Scientific

00:04:20 discovery was already a conversation between approaches, not a solitary

00:04:24 announcement from nowhere. I returned to Britain that year. Soon, a very

00:04:29 different kind of calculation would demand attention. There would still be

00:04:33 rules, symbols and possible answers. But the difficulty would no longer be only

00:04:38 whether a problem could be solved. It would

00:04:40 be whether it could be solved in time.

00:04:43 Imagine a supply ship crossing the Atlantic. Somewhere, a radio message is

00:04:47 intercepted. Its letters are available to read, but its meaning is not. An

00:04:51 answer that arrives too late may be little better than no answer at all. That

00:04:56 was the urgency behind much of our work at Bletchley Park, where I joined

00:05:00 Britain's codebreaking organisation in nineteen thirty-nine. Enigma did not

00:05:05 protect messages with a single permanent substitution. Its rotating components

00:05:09 and changing settings produced a moving problem. Recovering one day's settings

00:05:14 did not mean the enemy's communications would remain open forever. There were

00:05:18 different networks, new procedures and repeated losses of access. Nor did we

00:05:23 begin from nothing. Polish cryptanalysts had already made fundamental

00:05:27 breakthroughs against Enigma and shared their knowledge with Britain and France

00:05:31 before the war. Our work depended on that inheritance, then had to go further as

00:05:36 the system changed. One useful opening was a likely fragment of the original

00:05:41 message: perhaps a recurring phrase in a weather report. We could place that

00:05:45 possibility against the encrypted letters and ask which machine settings were

00:05:49 consistent with both. Rather than patiently translate every possible

00:05:53 version, we could discard combinations that produced contradictions. The machine

00:05:58 I helped devise for that search was the Bombe. Gordon Welchman's diagonal-board

00:06:03 improvement made it much more effective. Engineers turned the logic into working

00:06:07 electromechanical equipment. Operators kept the machines running, and other

00:06:12 workers checked the remaining possibilities. The Bombe did not simply

00:06:16 print the enemy's secrets at the press of a button. My work concentrated

00:06:20 especially on German naval Enigma. Joan Clarke was among the gifted

00:06:24 mathematicians in our section, Hut Eight. Her skill with the statistical method

00:06:29 called Banburismus helped reduce the work required of the machines. She belongs

00:06:33 in this story as a codebreaker in her own right. She also became a close friend.

00:06:38 We were briefly engaged in nineteen forty-one. I was gay; the engagement

00:06:42 ended, but our friendship did not. Bletchley was a collaboration of

00:06:46 mathematics, engineering, interception, analysis and relentless routine. A

00:06:51 clever idea mattered only when the wider operation could turn it into useful

00:06:55 intelligence. There was no single evening after which every message became

00:06:59 readable and the war was won. The work helped the Allies. Much of it remained

00:07:04 secret after the fighting stopped. For me, peace brought a different challenge:

00:07:08 could the general-purpose machine imagined on

00:07:11 paper become a practical instrument for science?

00:07:14 A mathematical machine needs no electricity and never suffers a loose

00:07:18 connection. A real one must store information reliably, move it to the right

00:07:22 place and carry out instructions without breaking down. The distance between

00:07:27 those two machines is an engineering problem, not something a theorem can wish

00:07:31 away. After the war, I joined the National Physical Laboratory. I developed a

00:07:36 design for an electronic computer called the Automatic Computing Engine, or ACE.

00:07:41 It was an ambitious attempt to turn general-purpose computation into working

00:07:45 hardware. Progress towards building it was slower than I wanted. In nineteen

00:07:50 forty-eight, I moved to Manchester, where an existing team was already

00:07:54 developing electronic computers. I contributed to programming and the work

00:07:58 around the Manchester Mark One. Meanwhile, colleagues at the laboratory

00:08:02 continued the ACE project. This was not one inventor delivering the computer in

00:08:07 a single finished package. Outside that work, I ran long distances and competed

00:08:11 in races. There were friendships, physical effort and ordinary days as well as

00:08:16 calculations. A life cannot be reconstructed entirely from its most

00:08:20 famous papers. But one of those papers did ask a question that reached far

00:08:24 beyond numerical tables. In nineteen fifty, I published an article about

00:08:28 computing machinery and intelligence. Instead of allowing the discussion to get

00:08:33 stuck on competing definitions of thinking, I proposed examining a machine's

00:08:38 performance in an imitation game. Its central challenge can be pictured through

00:08:42 a written exchange. An interrogator cannot see or hear the participants and must

00:08:47 judge from their replies. Could a suitably programmed computer take part

00:08:51 convincingly enough to resist being distinguished from a person? That was a

00:08:55 proposal for investigating behaviour, not a mathematical certificate of

00:09:00 consciousness. A convincing reply and an inner experience are not the same

00:09:04 thing. The point was to make a difficult question available for argument and

00:09:08 experiment rather than settle it by declaring, in advance, what machinery

00:09:13 could never do. I also considered learning. Instead of attempting to write a

00:09:17 complete adult intelligence in advance, perhaps we could begin with a simpler

00:09:22 program and develop its abilities through training and experience. That was a

00:09:26 research proposal, not a finished artificial mind. And computation was not

00:09:30 the only place where simple processes might produce unexpected complexity. I was

00:09:35 turning towards another puzzle: how living matter grows into organised shapes.

00:09:40 A developing organism does not begin with its finished pattern painted on it. So

00:09:45 how can an apparently uniform tissue acquire structure? In work published in

00:09:49 nineteen fifty-two, I examined what could happen when chemical substances

00:09:54 reacted with one another and spread through a tissue. Spreading alone tends to

00:09:59 smooth differences away. Coupled with the right reactions, however, it can

00:10:03 instead amplify small disturbances into a spatial pattern. The surprising point

00:10:08 was that order need not be imposed from outside, one detail at a time. Under

00:10:13 particular mathematical conditions, it could emerge from the interactions

00:10:17 themselves. This was a possible mechanism for biological pattern formation, not

00:10:22 a complete explanation of every shape in nature. While that paper was moving

00:10:27 towards publication, my private life was becoming a criminal case. We have

00:10:31 arrived back at the burglary. During the investigation, I acknowledged my

00:10:36 relationship with Arnold Murray, an adult man. The law then criminalised sexual

00:10:41 acts between men, including private, consensual acts. In March nineteen

00:10:45 fifty-two, I was convicted of the offence called gross indecency. The

00:10:49 alternative to imprisonment was probation conditional on undergoing hormone

00:10:54 treatment intended to suppress my sexual drive. I took that alternative. It was

00:10:59 not a freely chosen treatment for an illness. The threat of imprisonment was

00:11:04 part of its coercion. My security clearance was removed, ending the

00:11:08 classified work I had continued after the war. But my university research did

00:11:12 not simply stop. I continued working at Manchester, including on the biological

00:11:17 questions that had opened this chapter. Those two facts must remain beside one

00:11:22 another. The punishment caused harm and narrowed my life. Continuing to work

00:11:27 does not make that harm smaller; being harmed does not mean there was nothing

00:11:32 left of me but suffering. There were still calculations to check, patterns to

00:11:36 investigate and colleagues with whom to work. The authorities had classified my

00:11:41 relationship as an offence. That classification did not become an

00:11:45 explanation of my character, and it did not become the whole of my life.

00:11:50 The reconstructed first-person account ends here. We now step outside it to

00:11:54 describe Alan Turing's death and what followed. Turing died at his home in June

00:11:58 nineteen fifty-four, aged forty-one. The inquest recorded a verdict of suicide.

00:12:03 That is the official finding. It does not give a narrator access to his final

00:12:07 thoughts, or justify reducing a complex death to one certain motive. His

00:12:11 persecution is documented. So is the scientific work he continued after his

00:12:15 conviction. He left friends, correspondence and unfinished

00:12:18 investigations. To stop his biography at the conviction would be to lose sight

00:12:23 of the mathematician still exploring how patterns arise. He left work across

00:12:27 several fields: a rigorous model of computation, major contributions to

00:12:31 wartime codebreaking, designs and methods for early computers, a challenge to

00:12:35 assumptions about machine intelligence, and a mathematical approach to the

00:12:39 formation of biological patterns. His achievement was not that he had already

00:12:44 built every part of our digital world. It was that he gave other people ways to

00:12:48 formulate problems, construct machines and continue investigating. Public

00:12:52 recognition took a different route. In two thousand and nine, Prime Minister

00:12:56 Gordon Brown apologised on behalf of the British government for Turing's

00:13:01 treatment. In two thousand and thirteen, Turing received a royal pardon. In two

00:13:05 thousand and twenty-one, a Bank of England fifty-pound note bearing his portrait

00:13:09 entered circulation. The country that had prosecuted him made his face part of

00:13:14 its public identity. Those gestures mattered. None could return the years in

00:13:18 which he might have lived and worked. Return now to the man reporting a

00:13:22 burglary. He approached the law because someone had entered his home. The law

00:13:26 turned towards his relationship instead. The disturbing part of that story does

00:13:31 not become disturbing only because the complainant happened to be a great

00:13:35 mathematician. Turing did not have to help win a war or imagine a universal

00:13:39 machine to deserve a life without persecution. His brilliance explains why

00:13:43 his name belongs in the history of science.

00:13:45 It was never what entitled him to dignity.