Thus, in this view it was during the war that Turing formulated both his central contribution to cognitive science and the practical prospect of a universal machine - the modern computer. From this point Turing apparently became gripped by the potential of computability, and of his own discovery that all computable operations can be implemented on a single, universal, machine. I suggest that having confronted the problem of how the physical brain could support the appearances of a non-mechanical 'intuition,' Turing concluded that the function of the brain was that of a machine, but one so complex that it could have the appearance of not following any rule. Such machines might be said to learn, and perhaps to act with the appearance of intelligence.įrom 1941 onwards Turing began to speak of such ideas to his Bletchley Park colleagues, and also to use the word brain (Hodges 1983 p. Machines which modified their own rules of behaviour would show features which had not been foreseen by anyone designing them. From now on, he was committed to exploring the range of the 'purely mechanical', and it appears that it was now that Turing concluded that the scope of computability was not limited to processes where the mind follows an explicitly given rule. The astonishing power of the 'merely' mechanical did, I believe, influence him towards the shift of view which I previously attributed to the 1936 period. It must have been very striking that human guessing and judgement had been overtaken by mechanical methods to great effect, and Turing must have been perfectly aware that machines and mechanical methods working on the logical data were embodiments of Turing machines. Not only had he devised physical machines of enormous logical ingenuity, but there were people carrying out logical and statistical operations 'like machines'. I now go on to the Turing of 1941, who by that time was masterminding the decipherment of naval Enigma messages.
Part 2: Turing from 1941 to 1954 on this page.Part 1: Turing from 1936 to 1941 on the previous page.
This pre-print web-page version has been laid out in three parts:
Turing and Penrose - Andrew Hodges Uncomputability in the work of Alan Turing and Roger Penrose a talk given by Andrew Hodges
0 kommentar(er)
