> In the August 1999 issue of _Scientific American_, "Letters to the Editor"
> on page 8, B. Jack Copeland and Diane Proudfoot give the following reply
> to Thomas Bushnell of MIT:
>
> "Turing was indeed a courageous man, and he was open about his
> sexual orientation at a time in Britain when homosexuality was
> a crime. Treated wretchedly by the country that he helped to
> save, Turing was convicted of "gross indecency" and sentenced
> to a year of hormone "therapy" (which he seems to have borne with
> amused fortitude) in March 1952. But it was more than two years
> after his conviction that he died of cyanide poisoning. (A homemade
> apparatus for silver-plating teaspoons, which included a tank of
> cyanide, was found in the room adjoining that in which Turing's
> body was discovered.) A man who lived for his work, he was then
> in the midst of exciting research, and a close friend who visited
> him a few days before he died found him jolly. We wish we could
> explain Turing's death, but having examined the depositions made
> at the inquest as well as other material, we are less certain than
> Bushnell that the coroner's verdict of suicide was correct."
Some HM members may know my book _Alan Turing: the Enigma_, which appeared
in 1983 and is in print in the UK and in several translations. (It is
expected to be republished in the United States next year.) Others may have
seen the play _Breaking the Code_, which was based mostly on the story in
my book of Turing's life from 1952 to 1954. You may like to see my website
at http://www.turing.org.uk/turing/ for more information.
The following remarks present essentially the same view as will be found in
my 1983 book, to which I give page references.
In my view it is overwhelmingly likely that Alan Turing killed himself on 7
June 1954, but that he chose a means of death that would allow others to
see it as an accident. The person he probably had most in mind was his
mother, who had nagged him since boyhood on his careless and unconventional
ways, and his unclean fingers in particular. And his strategy worked
perfectly. Addressing the question of how someone could take cyanide by
accident, his mother's unswerving answer was that he had got it on his
fingers and then carelessly transferred it to the apple he ate in bed. (She
also argued that he could have drunk cyanide from a glass 'in a daze'.)
Very likely he set up for this kind purpose the whole 'experiment' for
electrolysis (gold-plating, actually - the golden teaspoon can be seen in
the archive at King's College, Cambridge)
One slightly misleading point in my book (page 488) is my statement that it
was obvious that the apple was dipped in the cyanide, but that the apple
was not analysed. This has led people to speculate on how much cyanide
would adhere to the apple's skin and so forth. I was perhaps victim myself
of Turing's dramatic stage-setting here, over-influenced by Turing's own
delight in the Snow White story (page 149). More likely, he calculated the
lethal dose and put it on the apple (actually a half-slice of apple),
rather than 'dipping'. The pathologist whom I consulted on the inquest
documents in 1978 took the view that it was because the whole thing was
'just so obvious' that there was no detailed analysis of the apple slice.
It would be inappropriate for me to be dogmatic on this question, since it
seems to me that Turing himself wished to leave an ambiguity, and if people
are unconvinced that it was suicide, that is part of the legacy from a
death no less extraordinary than his life. However, the philosophers
Copeland and Proudfoot do offer an interpretation of the background and I
will comment on this.
In their remarks, Copeland and Proudfoot correctly say that Alan Turing was
tried in March 1952, and the enforced injection of oestrogen, intended to
neutralise his sexual drive, had ended in March 1953. They are quite right
to counter the thoughtless stereotyping and wilful ignorance so often now
offered in summaries of Turing's life that he killed himself at the
prospect of his homosexuality being known, or at a 'rumour' of a trial, or
some such. In fact he made a point of being very open and unashamed,
expressing contempt for the law. His reported reaction to the police on
arrest was that 'he thought a Royal Commission was sitting to legalise it'
[In this he was mistaken; and his consensual activities ceased to attract
criminal penalties only in 1994].
However in my view it is also facile to imagine that anyone, however brave
the front shown to the world, can deal with such public humiliation and
invasion of the body without considerable trauma. Indeed such a trauma may
take a long time to work itself out.
Furthermore although as they point out, the trial was more than two years
in the past when he died, the ramifications did not end in 1952. There may
well have been direct ill-effects from the hormone 'treatment' itself, and
besides this there were other consequences.
(1) To make a rather obvious point - which, however, many people miss -
sexuality is not conveniently switched off by criminal sanctions (or by
hormones for that matter). Then at the age of 40, Alan Turing had no
intention of becoming a purely theoretical homosexual; as in his work he
had a great interest in turning theory into practice. It's my impression
that as a consequence of the trial, he never risked having sex inside
Britain again. He himself explained the importance of his holidays abroad -
Norway in summer 1952, Greece and France in summer 1953 - in terms of
getting beyond the reach of the English law. This in itself means one
cannot regard the trial as something that would fade in significance as
time went on, or say (as I once read in a review of my book) that 'he had
no legal problems' after 1952.
(2) Turing anticipated after his arrest that the events would make him 'a
different man.' Indeed they probably did, and many people would say they
made him a broader person. Turing turned to a Jungian psychoanalyst, wrote
down his dreams, took more interest in fiction. Clearly such elements, and
some new friendships, could have been the very things to *save* him from
isolation and despair. (As another example of a positive psychological
element of this period I have recently realised that his choice of Norway
as a holiday location almost certainly arose from hearing of its fledgling
and short-lived gay-rights movement.) However it also meant he spent a lot
of time thinking alone at home about his own feelings, taking them
seriously; it was probably rather less true than before 1952 than he 'lived
only for his work.'
(3) Since 1948 Turing had been doing part-time work for GCHQ, the successor
to the British wartime codebreaking establishment at Bletchley Park. These
Cold War period questions are entirely wrapped in secrecy now, and it is
simply a guess on my part that GCHQ had consulted him (and indeed had hoped
to get him back in employment) for the so-called Venona problem, which
involved searching through volumes of Soviet material hoping to find more
messages where a one-time-pad key had in fact been used twice - a quest
which, it is claimed, provided the evidence on which famous Soviet agents
of the post-war period were apprehended in Britain and the USA.
Turing himself only spoke of his government work to others who were also in
such work. One such was Donald Bayley, a wartime colleague, who in 1992
spoke in a television programme on what he had told me (on page 496).
Turing told him in October 1952 of how his work for GCHQ had been stopped,
saying 'there's no room for homosexuals.' Don Bayley's television comments
emphasised that he appreciated this, because he knew the change in
'vetting' procedures that had taken place since 1948. These made
homosexuality a disqualification for work on state secrets. Thus, out of
sight, Turing was necessarily classified as a security risk by virtue of
being gay, and this slur, this expression of distrust, might have been
almost as much of a trauma as the trial itself. A further point is that
Turing's modus operandi might have been calculated to arouse the worst
fears of security officers about 'risks', since his sexual life drew him
into a reckless transgression of the social class disctinctions.
(4) There was a further development which Turing described as a 'crisis' in
March 1953 (page 483), indeed describing it as almost rivalling what had
happened in 1952 for 'sheer incident.' A young man he had met in Norway
arrived at Newcastle, having sent Turing a postcard saying he was going to
visit. There were suddenly 'police all over the North of England'
(according to Turing's letters) who intercepted the young Norwegian and
sent him back. Another letter (included in the preface to the 1992 edition
of my book) shows Turing remarking on how the 'poor sweeties' (as he called
the police) were going to be 'be a bit more nosy, so virtue must continue
to shine.' Turing attributed this pro-active police operation to his still
being on probation, but in my reading he was mistaken. The police did not
have the resources to read the mail and intercept the visitors of every
convicted homosexual. The reason for this operation must have been Turing's
special status as someone not only with unique knowledge of the great
Anglo-American secret of the Second World War, successfully kept sacrosanct
until 1974, but also some current knowledge of GCHQ's activities.
I therefore think it is very likely that later in 1953 he realised that
his private activities would forever be governed by his unique relationship
with the state that he had, indeed, helped to save. I think it very
possible that there was a particular cause of conflict in spring 1954,
namely that of whether he could again go abroad. The spring 1953 holiday,
after which Turing came back with a list of Greek names and addresses,
could well have been seen as the last straw by security oficers desperately
trying to reconcile Turing's extremely distinguished and respected past
with these appalling revelations of his personality. The authorities might
well have thought themselves entirely reasonable and gentlemanly in
requiring him to stay at home, yet Turing might well have seen it as the
deepest affront to his identity freedom and integrity. Under the velvet
glove, a less gentlemanly option could have been hinted at, namely the
possibility of raking over Turing's previous life and dragging past lovers
into trials. If any of this seems fantastic as the fate of a great
scientific figure and Second World War hero, note that that in the week
Turing died, newspapers were reporting the suggestion that Oppenheimer was
a Soviet spy.
(5) In referring to Turing's 'exciting work', Copeland and Proudfoot
correctly indicate that Turing refused to be cowed by worldly events, and
that research work on his non-linear morphogenetic theory continued
actively up to the end. I found his progress hard to assess myself, but
since 1983 there has been much more work done by experts on Turing's
unpublished manuscripts (P. T. Saunders in the Collected Works volume 4,
Elsevier,1992, and now Jonathan Swinton,
http://www.kings.cam.ac.uk/~js229/turing.htm). What I have learnt has not
changed my general picture that although Turing had many new ideas and
could readily have published more papers, his programme also had its
frustration. As in much research, a certain over-optimism is essential, and
depression can set in when the initial goals suffer postponement. In
everything Turing had done before, the follow-up to original ideas turned
out slower and less satisfying than the first break. This was true too of
his biological theory, as set out first in his paper of 1951. In 1954 he
had not achieved an explanation of the well-known botanical Fibonacci
numbers, which he had set himself as a goal. (Ian Stewart's recent book
_Nature's Numbers_ gives the first satisfactory explanation as achieved
only in 1992.) On the other hand, it is remarkable how Turing started
trying out very pertinent new ideas in fundamental physics (pages 512-514
of my book) which he transmitted in fragmentary form as 'Messages from the
Unseen World' in March 1954.
These issues concerning his work neither explain suicide nor argue clearly
against it, but they are a inspiring and moving aspect of Alan Turing and
his attempts to live a creative mathematical life during this period.
(6) Ultimately I agree with what the logician Robin Gandy, who was both
Turing's student and one of his closest friends, said on television in
1992: that is wrong to think that his suicide can be given 'an
explanation.' No new revelations about the background, however traumatic,
could yield true explanation. We need only think of people who underwent
unimaginable degrees of suffering in the 1940s and 1950s, and yet lived on.
One factor that might be borne in mind, however, is that Turing had no
community of fellow-suffering. His sexuality cut him off from the
understanding of most, and his secret work could not be spoken of to those
with whom he was otherwise intimate.
In seeking 'explanations' we face the enigma of the individual personality.
One advantage of a full biographical treatment is that one can try to do
justice to the whole historical scope, from schooldays onward. There are
patterns which anticipate suicide; he was notably mercurial in mood,
delightful good humour alternating with dark, angry, silent periods that
disturbed others. In youth he mostly withdrew. Later on, when faced by
opposition he tended to put up a good, slightly angry, fight but then
suddenly walk out of the situation when no purpose seemed served by further
argument.
The most curious testimony of Turing's preparedness for suicide came to me
in 1977 from James Atkins, a mathematics undergraduate at King's College in
Turing's year who was probably Turing's first boyfriend of any duration. He
recalled to me receiving a letter from Turing in 1937, when he was away in
Princeton, about a means of suicide. It involved, James told me, an apple
and electrical wiring. He was astonished to learn from me that these had
indeed featured in the actual death, of which he was certain he had never
read any details (pages 129, 567). James Atkins did not keep the letter,
and so this story now rests on my say-so, but for what it is worth I
believed him and am also sure that the method was conceived with the same
end in mind (i.e. deceiving his mother). However the background in 1937 was
different; before the war Turing suffered from the then classic self-doubt
and self-hatred of homosexual men, especially when away from tolerant
King's College, Cambridge. In 1937, James read the reference to suicide as
related to the prospect of life as a homosexual. This is something that
changed considerably during the 1940s until Turing emerged with a
remarkably modern attitude in the 1950s.
There was always a sombre side (and as it happens, a fascination with
poisons) in Alan Turing. In 1941 he had a brief engagement to Joan Clarke,
one of the naval Enigma cryptanalysts - a distinctly sexless affair between
friends, which he broke by saying that he was homosexual. In doing so he
quoted to her from Oscar Wilde: 'For each man kills the thing he loves....'
(page 216)
(7) Copeland and Proudfoot's reference to a 'jolly' Turing conveys a
somewhat partial picture. The friend, quoted by Copeland and Proudfoot, who
visited him ten days before he died was in fact Robin Gandy. Robin was
pleased, I think, to be cast in my book as Turing's White Knight; he more
than anyone else jollied Turing along with much understanding and a common
humour, as can be seen from the shared jokes in the letters in my book. But
to some extent Robin did this by avoiding or neutralising dark and
difficult themes, and Robin never doubted that it was suicide. Another
close friend, (who became and indeed still is Turing's literary executor)
was the writer and critic P. N. Furbank. He heard more gloomy tones before
the death, and in his letter to Robin immediately afterwards mentioned 'the
way he talked about suicide before.'
Copeland and Proudfoot's words 'Treated wretchedly by the state he helped
to save' correctly suggests a dark irony - and an irony which must have
struck Turing himself with some force, the more so as it did not end in
1952. The irony is also deep because it involves the springs of creativity
and individuality in Turing's work. I now know more than I did in 1983 of
the situation at Bletchley Park in 1940; in particular that the then
director of Bletchley Park had expressed the conviction the German naval
ciphers would never be read. Turing ignored this, or took it as a
challenge, and found a way of breaking into the U-boat ciphers, then found
himself confronting the Admiralty who seemed reluctant to exploit the
opportunity. It was the same individualistic defiance of conventional
wisdom that manifested itself in his attitude to the law; and the same
inability to understand the Service mentality that would have made it
extremely hard for him to accept Cold War 'security' classifications.
Unlike certain other people who found themselves welcomed by western
governments in 1954, Alan Turing had never only obeyed orders.
Copeland and Proudfoot are fully entitled to their own lack of conviction
as to the correctness of the verdict, but the pieces of evidence they
supply to the readers of Scientific American do not address these
background issues, as were raised by me in 1983; anyway they certainly
offer no new evidence.
I was actually much more perturbed by the quite different historical
statements in the philosophers' original April 1999 article, 'Alan Turing's
Forgotten Ideas in Computer Science.' This article gave the vivid
impression that Turing had in mind an 'oracle' which would work by
measuring a physical quantity like a capacitance to infinite precision.
There is nothing remotely like this idea in Turing's writings. The 'oracle'
in his well-known 1939 paper 'Systems of Logic based on Ordinals' (which
they described but omitted to cite, thus unfortunately leaving the
impression it was an unpublished thesis) is a pure-mathematical
formalisation of a non-computable operation. Turing said explicitly of the
nature of the oracle that it 'cannot be a machine.' If interpretation
outside mathematical logic is sought, the most one could say is that Turing
was thinking of the non-mechanical elements in his Ordinal Logics theory as
connected with what he called the role of 'intuition' in recognising the
truth of a Goedel sentence. I and several other rather more significant
people wrote to Scientific American pointing out this and other
misrepresentations but the magazine did not print any of our letters and as
far as I know Copeland and Proudfoot have not yet replied to any of our
objections.
Copeland has however written to me 'I do NOT believe that Turing himself
had in mind the possibility of the physical construction of an
O[racle]-machine.' I can only say that I cannot fathom how Copeland
reconciles this statement with the message, as any plain reader would take
it, of the Scientific American article.
Andrew Hodges
Wadham College, University of Oxford, England
http://www.turing.org.uk/