for the living cell what ene:rgy did for the chemical compound. See for instance P. lVlorrison's proposed definition of life (1). Entropy made its entry as a thermodynamic quantity, f~Q, that is determined with calorimeters and thermometers. But a statistical interpretation soon followed. When a quantitative measure of information was introduced by Shannon in 1949, its relationship to entropy was soon understood. Brillouin (196:2) showed that information (knowledge) amounts to "negentropy" which is commensurate with entropy. Knowledge can be obtained only at the price of an equivalent increase of entropy and can be sacrificed to reduce entropy. Information is the very thing that computers deal with. The bit is a unit of information representing an entropy equal to the Boltzmann constant k as well as the formalism to convey that unit. The importance of memory and speed as necessary conditions for simulating or implementing intelligence and cognition is generally recognized. But surely they are not sufficient. They could be wasted by purposeless or inefficient programming. Other measures of computing efficiency must be maintained at sufficiently high levels. General computer practice would benefit from a suite of quantitative figures of merit. These benchmarks may also help pave the way to the understanding of cognition and intelligence. Among them and between them may be found the measures that unite and separate the conscious from the unconscious. Before answers are found, the right questions must be asked. REFERENCES: 1. Phillip Morrison, "A Thermodynamic Characterization of Self - Reproduction", Reviews of Modern Physics Vol 36, #2, p517 April, 1964) T H E TURING T E S T MISUNDERSTOOD Jan Eric Larsson Department of Automatic Control Lund Institute of Technology Box 118, S--221 00 Lund, Sweden JanEric@Control.LTH.S e The Turing test was conceived by Alan Turing [1] as a device to get around the difficult question of whether computers can think, and replace it with one which would be easier to discuss and answer. Since then the suggested test has repeatedly been quoted as an important aim for actual research and as a working test for intelligence. However, it may be interesting to note that Turing himself probably would not agree with such views. First of all, his test does not concern mind, thinking in general, or, (necessarily), intelligence. It concerns behavior with skill enough to successfully imitate a human being, and nothing else. Turing's point with suggesting his test was to try and avoid all discussions about mind, thinking, and intelligence by replacing the, in his opinion, too philosophical questions by a more technical one. His views in [1] are quite clear: T "It will simplify matters for the reader if I explain first my own beliefs in the matter. Consider first the more accurate form of the question. I believe that in about fifty years' time it will be possible to programme computers... to make them play the imitation game so well that an average interrogator will not have more than 70 per cent chance of making the right identification after five minutes of questioning. The original question, 'Can machines think?' I believe to be too meaningless to deserve discussion." [1] If the man in the first version of the Turing test should manage to fool the interrogator, it does not prove that he is a woman, and in the same way: if a computer program should pass the test, it does not prove that the program is a human, has a mind, or is generally intelligent. It proves only that the program is skilled, (and knowledgeable), in the specific task. Someone subscribing to a pure behavioristic view of intelligence could undoubtedly equate skilled behavior with intelligence. With this definition, a successful Turing test indicates a certain intelligence, as well as it implies a certain amount of knowledge, etc. But to use the Turing test as an indication of intelligence by other definitions, or whether the program is as intelligent as a human, or whether it can truly think or has a mind, that is simply a mistake. The moral of the story is that the Turing test was posed to avoid difficult philosophical questions, and in this sense it works quite well, once it is understood. In [2] it is stated that the test is a "bird-brained way to test for intelligence." Well, it was never intended as a test for intelligence, at least not by Turing himself, and it should not be criticized for failing something that it was not supposed to do. REFERENCES: 1. Turing, A.M., "Computing Machinery and Intelligence," Mind, Vol. 59, No. 236, pp. 433-460, 1950. 2. Fostel, G., "The Turing Test is For the Birds," SIGART Bulletin, Vol. 4, No. 1, pp. 7-8, 1993. SIGART Bulletin, Vol. 4, No. 4

/lp/association-for-computing-machinery/the-turing-test-misunderstood-lb9B966ovh