Alan Turing: The Enigma The Centenary Edition

by Andrew Hodges

  The inquest, on 10 June, established that it was suicide. The evidence was perfunctory, not for any irregular reason, but because it was so transparently clear a case. He had been found lying neatly in his bed by Mrs C— when she came in at five o’clock on Tuesday 8 June. (She would normally have been in on the Monday, but it was the Whitsun bank holiday, and she had had a day off.) There was froth round his mouth, and the pathologist who did the post-mortem that evening easily identified the cause of death as cyanide poisoning, and put the time of death as on the Monday night. In the house was a jar of potassium cyanide, and also a jam jar of a cyanide solution. By the side of his bed was half an apple, out of which several bites had been taken. They did not analyse the apple, and so it was never properly established that, as seemed perfectly obvious, the apple had been dipped in the cyanide.

  John Turing attended the inquest, having met Franz Greenbaum and Max Newman in the meantime. (Mrs Turing was away on holiday in Italy at the time, flying back when the news reached her.) John had already decided that it would be a mistake to contest a verdict of suicide, a policy from which the presence of a row of newspaper reporters did nothing to dissuade him. The evidence given37 was limited to the discovery of the death, the cause of death, Alan’s good health and his freedom from financial trouble. Nothing was mentioned that hinted at sex, the trial, blackmail or anything of the kind. The coroner said ‘I am forced to the conclusion that this was a deliberate act. In a man of his type, one never knows what his mental processes are going to do next,’ and the verdict was that of suicide ‘while the balance of his mind was disturbed.’ In the event, the national press made remarkably little of it, and nothing was said regarding the 1952 trial.

  Mrs Turing would not accept the verdict. Her argument was that it was an accident. Her evidence was that while Alan lay in his small front bedroom, an electrolytic experiment was bubbling away at the back. It had, in fact, been going for a long time. He did sometimes use cyanide for electrolysis, it being necessary for gold-plating. Recently he had used the gold from his grandfather John Robert Turing’s watch to plate a teaspoon.38 She argued that he had got cyanide on to his hands by accident, and thence into his mouth. This was, of course, what she had always said might happen. At Christmas 1953, when he made his last Guildford visit, she had repeated her warning (‘Wash your hands, Alan, and get your nails clean. And don’t put your fingers in your mouth!’). He brushed all this off with ‘I’m not going to injure myself, Mother.’ But this very fact meant that he was well aware of her sensitivity to the possibility of an accident. More willing to bend truth for the sake of kindness with her than with anyone else, he could turn her long complaint at his erratic ways into a scheme to save her feelings. It was a cruel blow for Mrs Turing, and the more so as it came hard on the heels of a greater rapport with her son. Suicide was officially a crime, over and above the social stigma, and she was also a great believer in Purgatory. The plan that Alan had mentioned to James Atkins in 1937, which involved both an apple and electrical wiring, might well have played on the same idea – perhaps it was the very plan he used. If so it was a ‘perfect suicide’, in this case calculated to deceive the one person whom he wanted to deceive.

  It resembled the mixture of detective story and chemistry set jokes that he enjoyed in the treasure hunts. Once he had devised a clue that depended upon the electrical conductivity of the soft drink Tizer. A last treasure hunt, held at Leicester with Robin in the summer of 1953, found him preparing bottles of red liquid, with clues written in red ink on the back of the label, so that they could only be read after the bottles were emptied. They were labelled the wrong way round: ‘The Libation’ for the smelly one and ‘The Potion’ for the drinkable. Perhaps the idea went back to Christopher Morcom’s teasing ‘deadly stuff’, back to the poisons of Natural Wonders. He had found a final chemical solution.

  Anyone arguing that it was an accident would have had to admit that it was certainly one of suicidal folly. Alan Turing himself would have been fascinated by the difficulty of drawing a line between accident and suicide, a line defined only by a conception of free will. Interested as he was by the idea of attaching a random element into a computer, a ‘roulette wheel’, to give it the appearance of freedom, there might conceivably have been some Russian roulette aspect to his end. But even if this were so, his body was not one of a man fighting for life against the suffocation induced by cyanide poisoning. It was that of one resigned to death.

  Like Snow White, he ate a poisoned apple, dipped in the witches’ brew. But what were the ingredients of the brew? What would a less artificial inquest had made of his last years? It would depend upon the level of description, ‘not the will of man as such but our presentation of it.’ To ask what caused his death is like asking what caused the First World War: a pistol shot, the railway timetables, the armament race, or the logic of nationalism could all be held accountable. At one level the atoms were simply moving according to physical law; at other levels there was mystery; at another, a kind of inevitability.

  At the most superficial level, there was nothing to see. His working papers were left in an untidy mess in his room at the university. Gordon Black, who worked with the computer on lens design problems, happened on the Friday evening before his death to see him cycling home as usual.*

  He had also booked as usual to use the computer on the Tuesday evening, and the engineers waited up for him, only hearing next day that he was dead. His friendly next door neighbours, the Webbs, had moved to Styal on the Thursday, and he had had them to dinner on the previous Tuesday, merry and chatty. He had been much regretting their move, spoke of visiting them, and said he was glad that the new occupants would be young and with young children. There were purchases, including theatre tickets, in his house when he died; he had written, though not posted, an acceptance of an invitation to a Royal Society function on 24 June. He had been seen out walking on the Sunday morning by a neighbour with whom he was on nodding terms (‘as usual he looked very dishevelled’); he had taken in the Observer on the Sunday and the Manchester Guardian on the Monday; he had eaten and left the washing up. None of this shed the slightest light upon his death.

  To his old undergraduate friends, the last year revealed a troubled mind, but equally, one that was continuing to press on. At Christmas 1953, besides visiting Guildford, he stayed with his friends David Champernowne at Oxford and Fred Clayton at Exeter. He went out for a walk with Champ and certainly did talk in a worried way about the Norwegian boy, Champ forming the impression that he had been imprudent and perhaps a bit reckless. But there was no definite point that came across; Alan rather rambled on and Champ felt a little bored.

  At Exeter he also went for a walk with Fred and his wife, who now had four children. One of the boys, Alan agreed, closely resembled his uncle in Dresden. Alan told Fred about the arrest, the trial and the hormone treatment, describing how it had developed the breasts, and making the most of its black absurdity. To Fred it was the confirmation of all his fears, and he told Alan how unsatisfactory such pick-ups must be, wishing he could find a permanent friend from the academic world. (He did not know anything about Neville.) A great believer in family life, Fred felt that Alan envied him the course that he had taken since 1947. Alan found a large mushroom, which to the Claytons’ amazement he said was edible, so they cooked and ate it. Afterwards Alan sent a thank-you note, with more notes on astronomy, and a homemade sundial in a cardboard box. It was hardly a grand farewell. Neither, indeed, did his Guildford visit take on a leavetaking character, and his last note39 to his mother, written shortly after these visits, ended with some information about a shop he had found in London where ‘some quite remarkably cheap things in glass suitable for wedding presents etc.’ could be bought.

  Neither of his two close postwar friends, Robin and Nick Furbank, had any clear idea that an end was approaching. Robin stayed at Wilmslow for the weekend of 31 May, just ten days before Alan died. Their friendship was one of great mutual confidence in
emotional matters, but there was no hint of a psychological crisis on this visit. They amused themselves with Alan’s experiments, trying to make a non-poisonous weedkiller and sink-cleaner from natural ingredients. They talked about type theory and planned to meet again in July.

  Alan had much developed his friendship with the writer Nick Furbank – a development which perhaps reflected his greater willingness to branch away from science, and even to interest himself in literature. The subject of suicide had entered at some point in their conversation, and Nick recalled it when he wrote40 to Robin on 13 June, describing what he had found at Wilmslow, when he went there in his capacity of executor. But this had lent no sense of explanation to the death, only perhaps the knowledge that it was not a course that Alan had ruled out of court. Nor did Franz Greenbaum, despite close acquaintance with Alan’s inner life and dreams, feel that anything was understood. The dream books were recovered and passed back to the psychoanalyst, but did not answer any questions.

  John Turing read through two of the dream books, which Franz Greenbaum lent to him before destroying them. Alan’s ‘scarifying’ comments on his mother, and his description of homosexual activity since adolescence, told John far more than he ever wished to know, and he found these revelations in themselves sufficient explanation for what had occurred, thankful only that he had managed to prevent them from reaching his mother’s eyes. To Alan’s friends, nothing could be so clear.

  There was one piece of evidence that he had prepared for death: he had made a new will on 11 February 1954. This in itself was a kind of statement as to where he stood. It made Nick Furbank his executor, rather than his brother, and gave to Robin all his mathematical books and papers. Then after £50 legacies to each member of his brother’s family, and £30 to his housekeeper, the remainder was divided between his mother, Nick Furbank, Robin Gandy, David Champernowne and Neville Johnson. John Turing was amazed and appalled at Alan having thus lumped his mother together with his friends – but this very fact was a much warmer gesture to her than a more conventional allocation which treated her not as an individual friend but as a person to whom he owed a family duty.*

  Yet the will included a provision that his housekeeper should have a further £10 for each year in which she had been employed after the end of 1953 – a strange point to add if he had then been settled upon death. It seemed to Nick, on visiting the scene, that Alan had arranged certain letters into packets – but there was no wholesale clearing up, neither of personal papers nor of his research. It was as though he had planned for the possibility, but in the event acted impulsively. What factors were there, at a less immediate level, that might have prompted such plans?

  He died on a Whit Monday, the coldest and wettest Whit Monday for fifty years, as it happened. Was it, symbolically, the day after inspiration, when the spirit was beginning to give out? G.H. Hardy had attempted suicide in 1946, although in his case, as one who had gone seven years deprived of creative life by a stroke. Was there an underlying pattern in Alan Turing’s life and death on this second layer of the psychological onion? The self-revelation of his ‘short story’ would suggest that he thought of himself as having had his ‘inspiration’ in 1935, and it being thereafter a fight to maintain that level. And the waves of inspiration had come only once every five years since Christopher Morcom’s death: the Turing machine in 1935, naval Enigma in 1940, the ACE in 1945, the morphogenetic principle in 1950. At least, these were the outward expressions of his thought; he worked rather like the Turing machine that he arranged to write and do rough working on alternate squares. There had been a good deal of reculer pour mieux sauter in between.

  In each case he had become not exactly bored or disillusioned by the last flood of work, but felt he had exhausted what he could achieve within its framework. He was, as compared with a more conventional academic, peculiarly anxious not to be enclosed or defined by his existing reputation. So indeed by 1954 or 1955 he would need something new to emerge, to maintain his freshness. But by June 1954 it was hardly time for despair. Indeed, the 1949 period had probably been much more difficult for him.

  Possibly the morphogenetic work had turned out rather plodding and laborious. It was three years since he had claimed he could account for the fir cone pattern, and he had not achieved it when he died. But there was no sign of a dropping off of interest. In the summer of 1953 he had taken on a research student, Bernard Richards. (There had been an earlier student, who had achieved nothing.) Richards took over some detailed calculations in connection with his models for the formation of patterns on spherical surfaces. He worked out some exact solutions to Alan’s equations, thus exhibiting the capacity of the theory to accommodate some of the simpler possible patterns found in the monocellular Radiolaria. Alan had a book of engravings* of these oceanic organisms, which he would show with delight to the engineers when on his night shift.

  The working relationship had not progressed beyond the master-and-servant kind, but even so, it was clear to Richards that there was no decline or fall in Alan’s work, not even at the very end. He was doing a good deal of writing up, but this was not at the expense of new experimentation on the computer. Nor, in any case, was this a very dramatic kind of theory, which stood to be proved right or wrong in an all or nothing way. It was a stage on which he could play out ideas in chemistry and geometry and see where they led. It was open-ended, and depended upon integrating insights into many different branches of mathematics and science, rather than trying to solve a particular problem within a given framework.

  Part of a plate from Ernst Haeckei’s report on Radiolana.

  He left behind a mass of detailed material,41 some organised into the form of a second paper, the rest in the form of worked examples and of computer print-outs incomprehensible to anyone else. Typical of his investigations at the time of his death was the following passage:

  The amplitude of the waves is largely controlled by the concentration V of ‘poison’.

  In his Natural Wonders way, he called the chemical whose function was to inhibit growth, the ‘poison’ – a macabre touch, given that in his own body it had recently been the other way round. He continued:

  If the quantity R is small it means that the poison diffuses very fast. This reduces its power of control, for if the U values are large in a patch and large quantities are produced, the effect of the poison will mainly be to diffuse out of the patch and prevent the increase of U in the neighbourhood. … If R is allowed to be too large it can happen that the ‘side-band suppression’ effect even prevents the formation of a hexagonal lattice. …

  Such observations reflected an insight gained from many trials even with this one model which was, as it happened, an ‘Outline of Development of the Daisy’. He had quite literally been ‘watching the daisies grow’ – not only by ‘examining 15 plants’, rather as he had done with Joan Clarke in 1941, but on his universal machine. But this was only one particular branch of his investigations: there was another headed FIRCONES, with an associated computer routine OUTERFIR; and another headed KJELL theory, which concerned another form of his basic equations, and which was associated with routines called KJELLPLUS, IBSEN and other Nordic names. All of this went rather beyond the material that he was writing up for publication, so there was no lack of prospects.

  He was working with Robin on the theory of types, and they planned to write a joint paper. He also wrote a popular article on the ‘word problem’, which appeared in the Penguin Science News in early 1954.* The Russian mathematician P.S. Novikov had announced42 that the ‘word problem’ for groups was indeed unsolvable by any definite method; Alan’s article explained this and tied it in with some questions in topology, showing that the problem of deciding whether some knot was the same as some other knot was essentially a ‘word problem’ of this kind. It was up to date, and looked forward to the complete proof of the Russian result becoming available. He was interested in solving problems up to the end: a last letter of May 1954 to Robin discussed certain id
eas of Robin’s for ‘getting round the Gödel argument’ but ended with: ‘Looked at the rainbow problem again. Can do it rather successfully for sound, but total failure for electricity. Love, Alan.’ When out walking together in Charnwood Forest, near Leicester, they had seen an unusual double rainbow, a phenomenon that Alan insisted on analysing. There had to be a reason for it.

  If he was looking for something new, it was in theoretical physics, which likewise he had put on one side since the 1930s. Before the war he had spoken to Alister Watson of his interest in the ‘spinors’ that appeared in Dirac’s theory of the electron, and in his last year he did some work43 on the algebraic foundations of the spinor calculus. He defined what he called ‘founts’, after printers’ founts of type.* He was also interested in the idea Dirac had suggested in 1937,44 according to which the constant of gravitation would change with the age of the universe. Once at lunchtime he said to Tony Brooker, ‘Do you think that a palaeontologist could tell from the footprint of an extinct animal, whether its weight was what it was supposed to be?’ And always distrustful of the official line in quantum mechanics, he revived his interest in the foundations of the subject. He found a paradox within the standard interpretation, as von Neumann had set it out, because he noticed that if a quantum system were ‘observed’ frequently enough, its evolution could be made indefinitely slow, and in the limit of continuous observation would freeze to a stop. Thus the standard account depended upon an implicit assumption that this mysterious moment of ‘observation’ occurred only at discrete intervals.