Born June 23: Alan Turing

Portrait of Alan Mathison Turing. Photo credit: Walter Stoneman, Royal Society 1951

Alan Turing was born June 23, 1912, in London. He was an extraordinary man well ahead of his time, who was recognized early on as a genius by his peers. Only recently have his pioneering achievements in mathematical logic, cryptanalysis, computer science, artificial intelligence, and biology began to find widespread appreciation, many of which were foundational to the digital revolution that computerized our contemporary world. Turing is also remembered as a martyr to the struggle for LGBTQIA rights: after a series of influential publications and years of innovative codebreaking work crucial to the Allies’ defeat of Nazi Germany, he was prosecuted by the British government in 1952 for admitting a homosexual relationship and subjected to chemical castration, leading shortly to his depression and suicide. For decades after that, few apart from mathematicians and computer scientists knew much at all about Turing, his military codebreaking work having been classified. Andrew Hodges’ excellent 1983 biography brought Turing out of the shadows, with readable explications of his theoretical work and a sensitive portrait of a charming, unpretentious man who heroically set his intellect against the Nazi U-boat fleet, and was repaid by his country with a humiliating criminal conviction for his sexual orientation. Hodges’ book became the basis for the Academy Award-winning 2014 biopic The Imitation Game, starring Benedict Cumberbatch. In 2013, after a years-long petition effort, the Queen issued Turing a rare royal pardon, and in 2016 Parliament finally passed the ‘Alan Turing law’ retroactively pardoning all men convicted under now-overturned English statues outlawing homosexuality. Turing’s likeness will appear on newly issued £50 notes going into circulation this month in the UK.

Signs of his budding genius were apparent from an early age, as well as his lifelong passion for long-distance running and cycling. When he was 13, the start of his first term at the Sherborne School in Dorset coincided with the 1926 General Strike; determined not to miss it, he rode the 60 miles from his home in Southampton alone on his bicycle. Many of his teachers, intent on imparting the classics, did not consider developing his interest in mathematics and science to be an ‘education, and his parents received many notes disparaging his writing abilities and slovenly dress. At Sherborne he bonded with fellow student Christopher Morcom, thought by his biographer to be Turing’s first love; when Morcom died of tuberculosis in 1930, Turing’s grief caused him to throw himself even more deeply into his studies in math and science and possibly influenced his becoming an atheist.

His grasp of complex ideas preceded his formal training. As a teenager, he was already solving advanced mathematical problems before even taking a course in calculus. Starting in 1931 he attended King’s College in Cambridge. After being named a Fellow of the College on the strength of his dissertation, in 1936 Turing published what many consider to be one of the most influential math papers of the 20th century: "On Computable Numbers, with an Application to the Entscheidungsproblem". Perhaps even more significant than Turing's proof that some questions in classical mathematics are not decidable, in the vein of Gödel's incompleteness theorem, was his novel restatement of decidability in terms of computability, and his vivid definition of that in terms of a theoretical universal computing machine. The online Stanford Encyclopedia of Philosophy offers a much fuller presentation of these problems and Turing's work on them.

His idea of a universal computing machine was especially influential. To capture the concept of human computability (leaving aside what might be theoretically computable by higher beings or alien intelligences), Turing posited a typewriter-like machine that manipulates symbols (digits or characters in a finite alphabet) on a strip of tape that is divided into cells, according to user-generated instructions. The machine positions its head over a cell, scans the symbol that is written there, and based on the symbol and the programmed machine state, it either leaves it as is or writes a new symbol, and moves one space to the left or right. The tape is akin to a modern computer’s memory, stipulated to be infinite or unbounded. According to Turing’s thesis, the problems solvable by a Turing machine are exactly those problems solvable by an algorithm or effective computation method; any computable problem can theoretically be broken down into a series of simple, automatable tasks. This idea became the foundation for modern computer science: the idea of a programmable machine that could build up complex behavior from simple digital functions. As his mother recalled in her memoir, there was a typewriter in the family home which always fascinated young Alan, perhaps germinating his idea of a powerful typewriter-like computing machine. In the 1940s John Von Neumann elaborated this into an architecture that is now more recognizable as a computer: input and output interfaces, a processor, a memory bank for storage and programs, a control unit, and so forth. But the fundamentals of Von Neumann architecture derive from Turing, with whom Von Neumann was acquainted both at Cambridge and later at Princeton, where Turing went after Cambridge to study with Church. While at Princeton, alongside his mathematical studies Turing began building an actual primitive computer, an electronic multiplier built on banks of electronic relays.

His next task would even more firmly unite the worlds of theory and machine. Turing was always strongly anti-war, but the rise of Hitler unnerved him, and in late 1938 he returned to England and joined the intelligence effort to work with a team of codebreakers at Bletchley Park decrypting Nazi communications. The Germans encrypted their messages using the Enigma machine, a bulky device swarming with rotors and switches. Based on a vast number of mechanically possible user-determined settings or keys for each message, it would elaborately scramble each character of the message, requiring another Enigma machine on the other end set to the same key to unscramble it. Or perhaps, an even bulkier decryption machine, like the one secretly developed by Polish intelligence, which they called the bomba kryptologiczna. The Allies intercepted hundreds of Enigma-encoded Nazi messages every day, and the Polish machine would input a message and essentially grind through vast numbers of possible settings and solutions until it eventually found one that made sense: a slow and inefficient method, but a working one. In 1939 they shared their invention with the British and the French.

Turing made a number of decisive improvements to the bomba and renamed the British version the bombe. For one, he was able to adapt it to a more generalizable known-plaintext attack method, vastly improving its efficacy. In theory, all Enigma messages should have been highly secure, but sloppy German operators gave codebreakers an opening by regularly reusing message elements. They would send a weather report around the same time every day, with guessable words. Also, they often ended messages with ‘Heil Hitler’ which was always recognizable. Furthermore, the Enigma could not encode any letter as itself; ‘A’ was never encoded as ‘A’, permitting codebreakers to rule out another whole set of possibilities. Other statistical techniques Turing invented crucially optimized the codebreaking process.

At Bletchley Park, Turing was considered a bit of an eccentric. He would ride his bike to work and usually looked disheveled; sometimes he wore a gas mask while bike-riding to prevent hay fever. He liked to run the 40 miles to London for meetings, often beating his colleagues who took public transit. A world-class runner, he claimed running helped release the stress of his work. They called him ‘Prof’, and noted his amusing habit of chaining his mug to the radiator to prevent it from being stolen. He proposed marriage in 1941 to a fellow mathematician and codebreaker, Joan Clarke, but a few days later he admitted his homosexuality to her and called it off. It was something he was not secretive about, and Joan quite understood.

Turing devoted himself to the particularly difficult problem of cracking German naval messaging, which used a more complex indicator system to direct a devastating U-boat campaign against Allied shipping. His successful breaking of the Nazi naval radio code system is considered by historians to be a decisive factor in the Battle of the Atlantic, eventually enabling the Allies to prevail and preventing untold numbers of casualties. At one point, a huge expansion of machine capacity was needed, and Turing’s team wrote a letter to Churchill explaining their plight; its effect, as Hodges says, was electric, and from then on whatever resources they needed were ordered to their disposal. Bletchley Park was a team effort with many brilliant minds, but the head of his department always noted the centrality of Turing’s contributions.

After the war, Turing continued his forays into applied mathematics, working on inventing secure voice communications devices and designing an early computer called the ACE, Automatic Computing Engine. As a reader in the mathematics department at the Victoria University of Manchester, he was appointed Deputy Director of the Computing Machine Laboratory. In 1950, he published another vastly influential paper in Mind, ”Computing Machinery and Intelligence”, in which he proposed a test for artificial intelligence, now known of course as the Turing test. He posited a thought experiment he called the Imitation Game, in which a human observer would evaluate natural language conversations between a human and a machine, knowing that one of the two was a machine but not which. The observer would not be able to see the conversants or hear their speech, only to read transcripts of their statements and responses. If after a large set of trials, the observer could not reliably tell which conversant was the machine, the machine could be said to have passed the test for intelligence. By this Turing did not mean that the machine would have successfully fooled the observer into believing it was intelligent, but rather that its activity would constitute whatever there actually is to intelligence.

As with the Turing machine, this test of course defines intelligence in terms of human intelligence, disregarding whatever kinds of non-human intelligence might be possible. But it also reorients the question of artificial intelligence away from the dualist baggage of asking ‘can machines think?’, which focuses on whatever nebulous and indefinable experiences might be going on in our minds or souls when we think or feel (that we are sure machines could never share), to the more empirical ‘what would constitute intelligent behavior for a machine?’ We tend to think of machine intelligence as a kind of superhuman command of facts; but to pass Turing’s test, it does not matter so much whether the computer’s conversational statements are true or false (i.e., how knowledgeable it is), but rather if it can successfully emulate the conversational competence of a human.

In the final section of the paper, Turing addressed the objection that computers can only regurgitate whatever statements are previously programmed into them. He envisioned that future computers might be more like ‘learning machines’, just as the minds of children bootstrap themselves into mature intelligence from whatever physio-genetic programming they start out with by gathering information and reprogramming themselves as they go. He imagined a chess-playing computer that could start with the rules of chess and a powerful processor, and through practice teach itself to play a competitive game—which more or less came to pass in 1996 when IBM’s Deep Blue defeated world champion, Garry Kasparov. The kind of neural network systems that Turing was predicting are some of the most powerful artificial intelligence systems in operation today. Of course, the Turing test does not settle every question about minds and thinking, but he was certainly prescient as to the broad strokes of how effective artificial intelligence might be developed.

His curiosity was wide-ranging. In 1952 he published a classic paper in mathematical biology, “Chemical Basis of Morphogenesis”, in which he outlined an explanation of the process by which some tissues and organs might grow into their eventual shapes, such as the petal structures of flowers, in terms of a system of chemicals interacting and diffusing across a field to create patterns. This was years before the role of DNA was fully understood, but Turing was able to generate the predicted results (he calculated them by hand, not having access to a supercomputer to crunch his numbers). Researchers are still confirming ways in which Turing’s theorized reaction-diffusion process can partially explain many kinds of morphogenesis, such as the stripes and spots on the fur of cats. Hodges notes a famous sketch of Turing as a child gazing at a patch of daisies while others played field hockey nearby; one of his papers on morphogenesis was titled “Outline of the Development of the Daisy”.

That year the 39-year-old Turing met 19-year-old Arnold Murray outside a Manchester theater, invited him to lunch, and went to bed with him. Soon thereafter Murray’s house was burgled by one of his seedy acquaintances; Turing reported it to the police, and in the course of their investigation admitted perhaps unwisely that they had had sex three times. To the officers, this immediately overshadowed the burglary. Turing was unapologetic about his admission, optimistically believing there was a Royal Commission seated just then to decriminalize homosexuality in the UK (that did not happen until 1967). A widespread paranoia among intelligence and law enforcement during the Cold War was that homosexuals in government work might be at risk of blackmail by foreign spies, or perhaps be spies themselves, something that was likely on the minds of the Manchester authorities when one of their persons of interest turned out to be an outspoken gay man with high-security clearance. Both men were charged with ‘gross indecency’ under Section 11 of the Criminal Law Amendment Act of 1885. Murray’s counsel successfully shifted the blame to Turing for striking up the relationship in the first place, and he got off lightly. Turing’s brother advised him to plead guilty; the judge followed Turing’s solicitor’s suggestion to steer sentencing away from prison and towards hormonal treatment to reduce his libido, in consideration of his important academic work.

In March of 1952 Turing agreed to undergo a year’s worth of injections of synthetic estrogen, diethylstilbestrol, or DES. Since its discovery in the 1930s, DES was used as hormone therapy to lower the risk of complications in pregnancy; in the past few decades, it has been almost entirely discontinued, as its use has been linked to a variety of adverse side effects including increased risk of tumors and carcinoma, and also depression. The injections rendered Turing impotent and caused him to develop breast tissue, but Hodges reports that he bore up fairly well, publicly complaining about his trial and the injustice of the anti-homosexuality laws at the computer lab and maintaining his good humor. After the war, he had regularly traveled to Mediterranean countries and Norway, where gay rights movements were beginning to take root. But following his conviction, his security clearance was revoked, his government consultancy was terminated, and his travel was restricted, which Hodges believes is what ultimately sent him into an irrevocable depression. On June 8, 1954, Turing’s housekeeper found him dead of cyanide poisoning, with a half-eaten apple by his bedside, which was thought to be the method of ingestion. He was cremated at Woking Crematorium and his ashes were scattered in the crematorium’s gardens, where his father’s had been scattered before him.

Some 60 years later, when the royal pardon was posthumously issued, Ian Standen, chief executive of the Bletchley Park Trust, honored Turing as “a visionary mathematician and genius whose work contributed enormously both to the outcome of the war and the computer age.”