for INFO NATION
Published January 1997
Are you a computer? If you were, how would I know?
The riddle of the mind has bothered thinkers for centuries. In our times, one person found himself drawn to the question of how we think, and in his own mind created the computer as we know it today. Nobody has been able to answer the big philosophical questions about the mind, and maybe nobody ever will, but if those questions are answerable, Alan Turing may have come close to solving them.
Turing was born June 23rd, 1912 in Paddington, London. His father was working in the civil service in India for most of his youth. The early Turing found himself drawn to science in spite of his upper-middle-class social environs, which still frowned on a scientific career even in the early 20th Century.
Turing continued to blaze his own path while at Public School. The traditional schoolmasters ignored his obvious talents, and gave him poor marks in history and Latin. Turing had little in common with the other students, but he found friendship and intellectual satisfaction with a close friend named Christopher Morcom. Morcom died suddenly in 1930, a life-changing event for the young Turing. As Turing biographer Andrew Hodges puts it: "For three years at least, as we know from his letters to Morcom's mother, his thoughts turnd to the question of how the human mind, and Christopher's mind in particular, was embodied in matter; and whether accordingly it could be released from matter by death."
At King's College, Cambridge, Turing's studies included probability and quantum mechanics. Mathematics was his field, but the questions of mind were not far from his own mind.
In 1931, Kurt Gödel showed that mathematics as we define it must remain incomplete, that any system of mathematics we create will have to contain statements that can neither be proven nor disproven using other statements in the system. That is, mathematics will never be closed and finished. Turing applied himself to a related question: Is there only one process by which all problems in mathematics could be decided?
Turing's answer in his August 1936 paper, "On Computable Numbers", was to imagine a machine with a paper tape printed with symbols. The machine moves up and down the tape reading the symbols. There are instructions, given beforehand, that tell the machine to read a symbol, check its situation, and then move a space left or right. Turing showed that his "Turing machine" could be applied to solve almost any problem given to it, if that problem is broken down into simple elements.
What is a computer? Today's computer is, at its heart, a Turing machine--a general-purpose computing machine that can be programmed to solve any solvable problem. If you have to name one person who invented the computer, Alan Turing is one of the top two or three contenders.
But Turing, following in the work of Kurt Goedel, also showed that some problems are not computable. Turing not only invented the concept of a general-purpose computer, by exposing the limits of this new computer he gave future researchers (including himself) definite, practical goals to work toward.
Turing's intellectual feat wasn't just notable for furthering modern mathematics. He had proposed a machine that we know well today, and he saw this proposed machine as a model for the mind itself. If the human brain had a finite number of states, then it too might be reduced to a Turing machine.
Turing spent a year or two at Princeton in the United States, where he came to the attention of another vital figure in computing history, John Von Neumann. By 1938, he was back at King's College, Cambridge, working in advanced mathematics, taking classes from (and debating with) Ludwig Wittgenstein. But as the political situation in Europe fell apart, Turing had already been tapped by the British government for work on cryptographic projects.
When the Second World War broke out, Turing moved into the British government's code-and-cipher headquarters at a country manor called Bletchley Park west of London to work full time. Operations at Bletchley Park were carried out under wartime secrecy--and much of what went on their remained classified by British intelligence until 1974. (Some related documents in the U.S. were not released to the public until April, 1996.) A group of mathematicians led by Turing successfully broke the code generated by the German 'Enigma' machines--encryption machines with a typewriter keyboard and several interchangeable rotors on the inside to scramble the letters in the message. The millions of possible combinations in the machine were beyond the codebreaking possibilities of pencil and paper, but Turing's team was able to build a machine to automate the work. Throughout 1940-41, the British were able to eavesdrop on most German communications. In 1942, after the Germans upgraded the Enigma machines and the British temporarily lost their advantage, Turing began work on codebreaking machines that used electronic components--in this case, telephone relays. By the time the invasion of Normandy was ready, Turing's Colossus machine was on-line, deciphering communications between Hitler himself and the German general staff.
As though Turing's contributions in other fields weren't enough, the operations at Bletchley Park in which Turing played a major role were decisive in the Allied victory in the Second World War. Without Turing and his cryptanalysis team, Britain and the U.S. would not have gained control of the Atlantic from German U-boats, and might not have been able to pull off an invasion of Europe. Cryptography has been an obsession of security agencies on both sides of the Atlantic ever since, and remains a hot political topic today. Turing himself was rewarded with an OBE for his efforts.
The speed and efficiency of the electronic components in Colossus pushed Turing to flesh out his concept of a thinking machine. The codebreaking machines at Bletchley Park had been wildly successful, but they were machines that were built to do only that one job. Turing's ideal machine should be able to do almost anything you could imagine a brain doing, or at least anything you could imagine an artificial brain doing.
Both John Von Neumann and Turing wrote papers in 1945 that described a general-purpose computer with a bold new innovation. In this new computing machine, the instructions and the data to be operated on can be stored the same way on the same media--the program is the data. We take this for granted. Executable files, text, pictures, databases all sit together on our disks.
Von Neumann was still concentrating on the model of a calculating machine. Turing's 1945 paper went beyond the idea of a big machine to do math--or did it? Turing proposed that computers could work in human languages, play games such as chess, and do all sorts of things that only humans could do previously. To the average person, it sounds like Turing is saying that computers can do more than just math. On a deeper level, Turing had a broader vision: everything is math. And Turing's general-purpose computer could be programmed to handle any computable problem; all you had to do was break the problem down into manageable, calculable pieces. Standard knowledge today: the colors in photos, the waveforms in sound recordings, or machine language instructions themselves can all be represented in binary digits, ones or zeroes. In 1945, this was revolutionary.
In the next years, Turing proposed machine memory and an early programming language. Studies in neurology led to thoughts of applying a biological model to computing design, something we would call a neural net today. He even proposed a national computer connected to public terminals, to bring the power of computing into the reach of the average person.
Turing accepted a job after the war at the National Physical Laboratory in Teddington, Surrey, where the next two years were wasted, stifled by a bureaucracy that couldn't see his visions. His plans for a universal computing machine were put off repeatedly. While Turing's creativity was being ignored at NPL, one of his former Bletchley Park colleagues managed to get a Turing machine--a real computer--built at Manchester University in 1948: the first modern general-purpose computer in Britain. Belatedly, Turing himself took a position at Manchester, as the British government's interest in computing moved from cryptanalysis to work on a British atomic bomb.
Turing's powerful mind, and the ideas which sprang from it, continued to get less respect than was deserved as he worked at Manchester University. The frustrating environment didn't stop him from publishing papers that were far ahead of their time. The 1950 paper, "Computing Machinery and Intelligence", first described the 'Turing test' for intelligence. Simply put, if you talk with a machine that can fool you into believing that it is human, you must accept it as a thinking entity on a par with humans. Modern research into artificial intelligence begins with this paper.
His 1951 paper titled "The Chemical Basis of Morphogenesis" expressed the links Turing had begun to see in his wandering studies in biology and mathematics. The road that Turing was setting out on is one that we have not nearly reached the end of. Turing had been wondering how complex structures such as flowers could spring from a chemical like DNA. The field Turing almost founded is what we might call "chaos science" today--how a simple formula can bring about a complex, intricate structure. Turing was exactly the right person to exploit the insights gotten from the newest mathematics and apply them to computers. Simple algorithms in a computer could be looped through again and again to make shapes and networks similar to those we see in living things. Growth is learning; as nervous systems grow new connections, why can't computing machines write themselves new routines? Had Turing lived longer, he might have foreseen things that we still have not realized today.
In March 1952, Turing was arrested for having sex with a young man in Manchester on three occasions. His blithe disregard for the social norms of the day gave the authorities any number of ways to hold him in contempt. He said quite frankly--to the local police--that he had done what he was accused of, and he didn't see anything wrong with it. He not only was indiscreet, he had been indiscreet with a person of lower class status. He was a notable and respected figure at the University, which added more disgrace, but the most damning element was that Turing had been involved in matters of government secrecy. In those days, a homosexual was automatically considered a security risk.
The court gave him two options: accept a prison sentence, or submit to injections of the female hormone estrogen, which the court seemed to think would suppress his sex drive. (Proponents of this punishment today call it "chemical castration".) He began to develop breasts--another example of chemical morphogenesis, ironically enough.
Had Turing been in the same situation only fifteen years later, he would have been a daring, avant-garde figure, the toast of the town. Thirty years later, and no-one would have paid attention at all. But in the paranoid, Cold War atmosphere of the early 1950s, Turing was doomed.
In July 1954, Turing was found dead in his bedroom. A cyanide-laced apple was on the bedside table. It could have been an accident; Turing had conducted amateur chemistry experiments since he was a boy. It could have been suicide, the verdict which appeared on the coroner's official report, even though acquaintances of Turing's felt that he was holding up quite well. His naive ignorance of social norms seemed to have made him immune to humiliation. A darker possibility is that it could have involved British intelligence, finally fed up at this person they figured to be fundamentally unreliable, this person who simply knew too much. Those questions are just not computable--we may never have the evidence we need to decide why Turing's life ended so soon, when he could have lived to see so many of his predictions and proposals come true.
Since 1991, a Dr. Hugh Loebner has offered prizes in an annual competition to see if a computer can pass the Turing test. The grand prize of $100,000 to the designer of a computer that can pass the Turing test according to the rules of the competition has not been awarded--yet. Each year a bronze medal and $2000 goes to the programmer whose computer performs the best in that particular year. In the competition, humans sit at computer terminals and hold conversations in text with entities hidden in another room. Topics are fairly narrow. The humans must decide whether their conversation partner is a human or a computer (and yes, there are always a few humans who sound like computers). So far, the computers have not been able to match the wit and inventiveness of human chat. Future outcomes are anyone's guess--the $100,000 is still up for grabs.
The Turing test--and Turing's life work--drill down to the basic question of what is human. Do we possess free will, or are we only biochemical computing machines? Are we doomed to obey the output of the chemical programs inside us?
Yet the results of chemical reactions can have surprising and unexpected results. Modern research into chaotic systems has shown that order can spring from disorder, and that amazingly complex phenomena can grow from simple conditions. Quantum mechanics, an early interest of Turing's, has unravelled the edges of the Newtonian billiard-ball universe, and shown that the subatomic foundations of our existence can be random, and therefore unpredictable. Later researchers such as Mandelbrot and Lorenz would make discoveries involving chaotic systems and fractal geometry using the same computing machines that Turing could see in his imagination, but that he never saw built. The origin of life itself may eventually be understood by studying the flowers that unfold from simple instabilities.
The real message of Turing's work is not that humans are a kind of machine, but that humans are still the measure of intelligence. In a world that seems to be increasingly mechanized, programmed, and parsed into databases, humanity is still the touchstone that brings the world to consciousness.
John Kowalik has interesting biographical information on Alan Turing
at:
http://ei.cs.vt.edu/~history/Turing.html
http://ei.cs.vt.edu/~history/Turing.2.html
Andrew Hodges maintains the superb Alan Turing Home Pages, including
pictures, Java applets, and dozens of useful links:
http://www.wadham.ox.ac.uk/~ahodges/Turing.html
The Bletchley Park home page is at:
http://www.cranfield.ac.uk/CCC/BPark/
Michelle Hoyle of the University of Regina, Saskatchewan, has a good description of a Turing machine at:
http://www.cs.uregina.ca/~hoyle/Lecture/turing.html
Suzanne Skinner has a Turing machine running as a Java applet in her pages at:
http://grail.csuohio.edu/~suskinne/tm/tm.html
Visual Models of Morphogenesis can be seen at the University of Calgary:
http://www.cpsc.ucalgary.ca/projects/bmv/vmm/title.html
Rules for the 1997 Loebner Prize competition are available at:
http://acm.org/~loebner/lpr97.html
Charles A. Gimon teaches an Introduction to the PC class at the English Learning Center in south Minneapolis. He can be reached at gimonca@skypoint.com.
Back to my net
writings.
Back to my home page.