A mind’s enigma

Alan Turing: Una aproximació biogràfica

18-75
18-75
© Agustín Sciammarella

Turing’s childhood and early years

Considered the father of modern computing, Alan Mathison Turing was born in London on June 23, 1912. In spite of his struggles to learn to read and write in his first school years, Turing showed very early on a deep interest for science. His greatest passion was chemistry. 

The young Turing was not particularly sociable. Proof of this was his difficulty to make friends at Sherborne School, where he joined in 1926, when he was about to turn fourteen. This school, like other private schools in Britain, encouraged the learning of sports and classic languages, linked to the Imperial virtues of manhood, hierarchy and leadership; and not so much the learning of sciences. Mathematics was the exception. Thus, Turing was redirected from experimental sciences to mathematics, a discipline he self-studied throughout all these years, even writing a critical essay on Einstein’s theory of relativity.

During this period, Turing developed feelings towards other boys in his school. One of his closest friends was Christopher Morcom, an older student that shared Alan’s interest in mathematics and other sciences. At age eighteen, Christopher passed the test to enter Cambridge’s Trinity College. Nevertheless, Turing was a year younger and therefore was not ready yet —he had to wait another year. However, he knew the tragedy of death not long before finishing his studies at Sherborne when Christopher died due to tuberculosis which was unknown for Alan.

According to his biographer, Andrew Hodges, Turing’s interest in machines and the mind was closely linked to this traumatic experience. What was the relationship between Christopher’s spirit and his body’s mechanism? Christopher’s death put Turing face to face with questions related to the mind’s nature, matter and the difference between life and death. 

  24-75
© NPL/Archive Science Museum, London
Alan Turing was sure that human knowledge and science are acquired by interrelating with the world, and that this interrelation determines the way a human brain stores knowledge. Following this theory, Turing claimed that machines were able to learn form experience and could go beyond mechanical processes.

Turing, in Cambridge

In 1931 Alan Turing passed the entrance exam to the University of Cambridge and began his studies at King’s College. His arrival at this College, although coincidental, was important. In fact, opposing the rigid atmosphere of the Trinity College, the largest and richest of all the colleges in Cambridge, the King’s College took pride on being a place for academic dissent, both morally and politically. Although privately, homosexuality was an important part of the culture that shared the elite at King’s College. 

After completing his university studies, Alan Turing was elected a Fellow of King’s College in Cambridge at the age of twenty-two. His interest in mathematical logic led him, by the mid 1930s, to tackle the problem posed by David Hilbert (1862-1943): whether mathematics were decidable or not, from the widely known Entscheidungsproblem. Hilbert was convinced that he could prove that mathematics were complete (every mathematical statement could be proven true or false), consistent (there was no possibility of reaching false claims from valid tests based on axioms) and decidable (there was a final method by which you could see if a mathematical statement could prove to be true or false).

Unfortunately for Hilbert, Czech mathematician Kurt Gödel (1906-1978) had shown in 1930 that arithmetic (and therefore mathematics) was necessarily incomplete. Gödel constructed examples of well-developed mathematical statements that could not prove true or false. Was there, however, a method to show which statements were true and which ones were false? This problem was what kept Turing worried during the summer of 1935. The most remarkable aspect of his solution, completed during his stay at Princeton and published in the Proceedings of the London Mathematical Society in 1937 (entitled «On Computable Numbers With an Application to the Entscheidungsproblem»), was that not only did it address the issue by denying decidability, but also by presenting a theoretical universal computing machine.

Apparently Turing was inspired by his mentor in Cambridge, Max Newman (1897-1984), who speculated with the possibility of facing Hilbert’s problem by means of a mechanical process, i.e., as a routine process that could be undertaken without imagination or thought. Turing’s great idea was deliberately admitting an interpretation of the meaning of mechanical understood as «made by a machine», albeit an imaginary machine. Turing thought up computing machines that mimicked mechanical processes in mathematics made by calculators or human computers. These machines, differing only in their initial configuration, could generate computable numbers (any real number defined by ordinary mathematical methods based on equations or limits). Thus, from a series of arguments and mathematical reasoning, Turing showed that his machines were proof of the impossibility of deciding whether mathematical statements could be proved true or false by applying a mechanical process.

At Princeton, American scientist Alonzo Church (1903-1995) had reached a similar conclusion independently, which was published before Turing’s paper. Turing’s work, besides being a much more accessible and intuitive one, incorporated the physical world and pondered about what could be done. Thus, Turing’s solution not only opened a new field of study for mathematics (computability) and offered a new analysis to mental activity, but also had a very interesting practical implication: the principle of a computer that could be built based on the concept underlying Turing’s universal machine.

Turing had moved to Princeton in September 1936 precisely where he would continue exploring the logic of mental activity, as his paper «Systems of logic based on ordinals», published in 1939 in the Proceedings of the London Mathematical Society and based on his doctoral thesis, directed by Alonzo Church, shows. We do not know much about his research in the United States, although it seems that he took some time to build a machine out of electromagnetic relays which effected binary multiplication as an encoding device, with some theory of immunity to cryptanalysis. In June 1938 he received his PhD from the prestigious American university with a dissertation that introduced the concept of hypercomputation. In autumn that year, after two academic years at Princeton, he returned to Cambridge. 

  «Turing’s interest in machines and the mind was closely linked to a friend’s traumatic death. What was the relationship between his spirit and his body’s mechanism?»
22-75
© GCHQ
Turing designed Bombe, the machine that worked as thirty Enigma machines and was built by British Tabulating Machines in 1940. The need to increase the speed to decode thousands of messages daily led to the use of electronics.
 

«Turing designed Bombe, the machine that worked as thirty Enigma machines at once and was very important in trying to decode messages sent with Enigma»

decoding works

In September 1939 Nazi Germany occupied Poland. World War II had begun. On September, 4 Alan Turing arrived at Bletchley Park to work for the British government decoding messages. Pencil and paper in hand, decoders tried to make sense of intercepted radio messages from the mainland, just as they had done in World War I. Soon enough, however, it became clear that traditional mechanisms were insufficient to decipher German codes.

The main encrypting tool the Germans had was Enigma. It was a device that looked like a typewriter and which turned a readable text in an incomprehensible sequence of letters. Once it had gone through the machine, it seemed that only one person with another Enigma was able to read the message. These Enigma machines were commercial products, on sale since 1923. The problem was that having one was not enough to decode the message. On top of the machine there were a series of wheels that could be placed in many different positions thus complicating the decoding process. In fact, German cryptographers thought it invulnerable.

At Bletchley Park, Turing tried to improve an electromechanical device designed by the Polish cryptologist Marian Rejewski (1905-1980) to decrypt Enigma messages. Not long afterwards, in 1940, British Tabulating Machines built a machine designed by Turing called Bombe. This machine emulated the role of thirty Enigma machines working at the same time and had some importance in attempting to decode the messages generated by Enigma.  

The information processing at Bletchley Park responded to a highly coordinated and efficient work organization. The Germans sent thousands of encoded messages twenty-four hours a day, changing their codes several times every day. Thus, the work to decrypt messages was undertaken separately at the same time by different people. In this context, the need to improve the decoding speed led to the use of electronics. 

Meanwhile, mathematician Max Newman, former mentor of Turing in Cambridge and in charge of many of the machines at Bletchley Park, proposed in 1943 a radical solution to decode messages from the Lorenz encoding machine, another very complex machine used by the Germans for high level- communications. The new machine, called Colossus, had to store sequences of symbols and compare them at high speed, as electronics permitted at the time. A machine with fifteen hundred valves, of which up to ten were built between 1943 and 1945, capable of testing encoding patterns and compare them to find similarities in encoded texts very quickly. Although they resembled his imaginary machine described in 1937, Turing was not involved in its design. These first programmable digital electronic machines, however, were based on the statistical theory he had developed to try to decode the messages of the Enigma machine. A theory based on a geometric model that facilitated the processing of billions of combinations of encoded letters and which consolidated cryptanalysis as a scientific discipline.

As World War II came to the end, mathematicians -surrounded by machines that store, compared and wrote symbols- began to ask new questions related to the mechanical nature of intelligence. In using mathematical theories, scientists began to reflect on the mechanisation of thought processes. The development of machines linked to cryptanalysis had stimulated ideas of how mathematical problems could be solved with mechanical help. For Turing, the most important issue was up to what point a machine could behave like a brain. Could it learn? Could it think?

  21-75
© Agustín Sciammarella

«The main encrypting tool the Germans had was Enigma. It was a device that looked like a typewriter and which turned a readable text in an incomprehensible sequence of letters»

research in peacetime

World War II had mobilized many scientists, mathematicians and engineers who, once the war was over, returned to university in 1945. Alan Turing, in turn, decided to go to London, to the Mathematics Division of the National Physical Laboratory (NPL), where he planned to build a computer called Automatic Computing Engine (ACE). It did not take long for him to give up this project. In fact, he was in London for three years only. In September 1947 he moved to Cambridge and in May 1948 accepted a job offer from the University of Manchester, promoted and sponsored by Newman.

The reasons prompting him to leave the NPL are unclear and thus attributed as much to the lack of experts in electronics as to Turing’s ambition to investigate ideas related to machine intelligence, a line of work that would not have been supported by his superiors. The project he left, however, had it been completed during that period according to his specifications, would have been the most sophisticated of the first computers. In fact, the first pilot of the ACE model, developed by the team that Turing left behind, ran its first program in May 1950 and quickly became a workhorse for the computational work of government research agencies, both civil and military.

Given the frustration that caused his research project in London, Turing decided to go to the University of Manchester, where he had to work with a group of mathematicians from Bletchley Park, which included Max Newman, and a group of electrical engineers like Frederic Calland Williams (1911-1977), radar experts. In fact, Williams had devised, based on the experience gained during the war, a small experimental computer, the first operational electronic computer with a stored programme. A machine known as Madam (Manchester Automatic Digital Machine), which in June 1948 was unique.

Manchester offered Turing the possibility of working with a machine that, although it had nothing to contribute to his own design, at least worked. Despite being immersed in a culture of engineers with little interest in his philosophical investigations, Turing took part in the gradual improvement of the machine, paying particular attention to practical programming problems. He even programmed it to be able to write love poems. Shortly afterwards, negotiations to transform this experimental monster in a commercial product started with a local electronics engineering firm, Ferranti. The University of Manchester received the first of these new machines in 1951, the Ferranti Mark I -informally called Blue Pig whose programming manual Turing wrote.

 

20-75
© Science Museum, London
The Enigma machine was a device that was marketed since 1923. However, simply having one did not allow the decoding of messages sent by the German army because of the complexity of the combinations offered. Turing worked at Bletchley Park on a method for decoding the thousands of messages that the Germans sent daily.

the imitation game (or the turing test)

While he waited for the new machine, Turing made use of his freedom in Manchester to shape and publish his thoughts on machine intelligence, which to some extent used ideas and suggestions of his 1937 paper. According to Turing machines could do exactly the same job as the so-called human calculators or computers. His experience at Bletchley Park and his work at NPL were crucial in shaping these ideas. 

Moreover, Turing was convinced that machines could learn and go beyond purely mechanical operations. A machine could be taught to improve its behaviour, to the point of showing «intelligence». At the end of the day, what was intelligence exactly? Turing argued that it was possible to build a machine to play chess with given rules and that the ability to learn while playing should make its behaviour go beyond them, showing an element of freedom as a living intelligence would, though not necessarily a human one.

Thus, in 1950 he published an article in the philosophical journal Mind, entitled «Computing Intelligence and Machinery» where he proposed considering whether machines could think. His answer was not a direct yes or no. He argued, however, that the issue could be resolved, and designed the so-called Turing test, an imitation game aimed at demonstrating the intelligence of a machine. Turing’s conviction was that a digital machine could give a good answer in the game, i.e., it could do unexpected things, imitate other machines or even consider the effects of an extrasensory perception if this was ever discovered. All that was necessary was enough storage capacity. From an extremely materialistic viewpoint, that rejected the existence of the soul, Turing was sure that there were no grounds to ensure that machines were not able to think.

Even some scientists had already undertaken research projects to address, in one way or another, similar ideas. During his time at Manchester, Alan Turing and his staff wrote some papers on digital computers applied to games which were published in 1953. A pioneering research on the subject of machine intelligence that did not have, however, a real immediate impact in the new field of artificial intelligence that would develop from the 1950s with names such as American Allen Newell (1927-1992), Herbert Simon (1916-2001), Marvin Minsky (1927-) and John McCarthy (1927-2011). 

 

«Turing was convinced that machines could learn and go beyond purely mechanical operationsto the point of showing “intelligence”»

death

Turing died in June 1954, a few weeks before his forty-second birthday. His mother and his friends defended the possibility of an accidental intake of potassium cyanide kept at home and used to make electrolytic experiments. This theory would fit within the careless behaviour of Turing and with such an unexpected death. No one had seen no sign of his intention to kill himself nor any explanatory note was found. According to the official version, however, he committed suicide after biting an apple infused with potassium cyanide. It is not unreasonable to think of this possibility, because during the depression he suffered in Princeton he had even devised a method of suicide by eating an apple. 

However, the 1950s had been less depressing for Turing than previous decades were. In the summer of 1950 Turing had bought a house in Wilmslow (Cheshire), about twenty miles south of Manchester. In spring of 1951, at age thirty-nine, he had been elected a member of the Royal Society of London. Even in May 1953, Manchester University had appointed him professor of theory of computation, a position created especially for him that gave him some comfort and freedom to continue his research.

His private life had perhaps caused him more trouble. His living in Manchester had frustrated to some extent his relationship with his lover Neville Johnson. Turing, however, often travelled to Europe, where he could be himself with more freedom and without repercussions. In Manchester he also enjoyed occasional affairs, including his relationship in early 1952 with a young man named Arnold Murray. A story that would lead to a fatal end when Turing held him responsible for a robbery in his home. Police investigations resulted in Alan Turing’s indictment for a crime against public morals for having sex with another man. To avoid prison, Turing pleaded guilty and was subjected to oestrogen treatment designed to «cure» his condition. A treatment that resulted in chemical castration and made him gain weight considerably and even caused him to grow breasts. Nevertheless, there seems to be no causal relationship between this and his death two years later.  

Although the issue was not highly publicized, these events shattered the opportunity to continue working on government projects on cryptanalysis, as he had done since the end of the war. Turing had become a security risk. Indeed, desertion in 1951 of spy Guy Burgess (1911-1963) had built the myth of the homosexual as traitor, which was very present among common people and also politicians. The paranoia of the Cold War prompted the British intelligence services to worry about Turing’s private life; because he knew secrets related to decoding and was an expert on a new technology with a military application. The problem for the intelligence services was not so much his sexual orientation as their inability to control and predict the behaviour of someone like Turing, who, for some people, could be easily blackmailed. All this has led to some conspiracy theories that argue that the intelligence services wanted to silence Turing forever and protect all the knowledge he had.

 

Epilogue

Turing was sure that the knowledge and understanding of humans derived from their interaction with the world and that this interaction determined the way in which knowledge is stored in the brain. The structure of the brain connected the words stored with occasions for use and with associated or replacing behaviours. Was language a game or was it connected with real life? Could there be intelligence without life? His claims about the intelligence of machines and mechanical simulation of learning made him focus on different biological issues, such as growth patterns in brain cells, during the last years of his life. He even made some simplified problems around the chemical theory of morphogenesis which he tried to solve with nonlinear differential equations. In the course of his research, he showed that heterogeneity could steam from initial conditions of homogeneity.

Alan Turing’s insistence on questioning the allegedly unique ability of humans to think was a constant since the 1940s. Some authors have tried to see in his defence of the possibility of thinking machines a subtle critique of social norms which denied the right to a legitimate and legal existence to a part of the population (especially gay men, but also women). Beyond this hypothesis, we can ensure that Alan Turing’s investigations decisively influenced the development of new scientific disciplines such as cybernetics, artificial intelligence and cognitive psychology

Bibliography
Agar, J., 2001. Turing and the Universal Machine: The Making of the Modern Computer. Icon Books.Duxford.
Agar, J., 2003. The Government Machine: A Revolutionary History of the Computer. MIT. Cambridge.
Copeland, B. J. (ed.), 2004. The Essential Turing: Seminal Writings in Computing, Logic, Philosophy, Artificial Intelligence and Artificial Life plus The Secrets of Enigma. Oxford University Press. Oxford.
Hodges, A., 1993. Alan Turing: The Enigma of Intelligence. Burnett Books. London.
Hodges, A., 1997. Turing: A Natural Philosopher. Phoenix. London.
Leavitt, D., 2005. The Man Who Knew Too Much: Alan Turing and the Invention of the Computer. Phoenix. London.
Strathern, P., 1999. The Big Idea: Turing and the Computer. Anchor Books. New York.

Pedro Ruiz-Castell. López Piñero Institute for the History of Medicine and Science, University of Valencia – CSIC.

© Mètode 75, 2012.

   

Bibliografia
Agar, J., 2001. <em>Turing and the Universal Machine: The Making of the Modern Computer</em>. Icon Books.Duxford.
Agar, J., 2003. <em>The Government Machine: A Revolutionary History of the Computer</em>. MIT. Cambridge.
Copeland, B. J. (ed.), 2004. <em>The Essential Turing: Seminal Writings in Computing, Logic, Philosophy, Artificial Intelligence and Artificial Life plus</em> <em>The Secrets of Enigma</em>. Oxford University Press. Oxford.
Hodges, A., 1993. <em>Alan Turing: The Enigma of Intelligence</em>. Burnett Books. Londres.
Hodges, A., 1997. <em>Turing: A Natural Philosopher</em>. Phoenix. Londres.
Leavitt, D., 2005. <em>The Man Who Knew Too Much: Alan Turing and the Invention of the Computer</em>. Phoenix. Londres.
Strathern, P., 1999. <em>The Big Idea: Turing and the Computer</em>. Anchor Books. Nova York.</p>

© Mètode 2012 - 75. Online only. The Festive Gene - Autumn 2012

Museu Nacional de Ciència i Tecnologia.