From JOHN SKILLING
Science is not a democracy, and an idea’s popularity in the commercial
market does not make it true. There is, as it happens, exactly one logical
way with dealing with imprecise information, and that is ordinary probability
calculus (‘Fuzzy logic goes to market’, 8 February). This has been proved
to be the only calculus which allows consistent reasoning – so that the
result of applying given ‘rules’ is independent of the order in which they
happen to be applied. There is simply no room for alternatives. In so far
as it differs from probability theory, fuzzy logic is internally inconsistent.
Of course, simplistic thinking may be adequate in simple cases, and
fuzzy logic may indeed be capable of scheduling lifts and washing machines.
But so are probabilistic arguments. The real danger with inconsistent, faddish
reasoning would come if it were ever applied to important matters like air
safety or reactor control. Fuzzy thinking about such matters would be dangerous
as well as deplorable.
John Skilling Department of Applied Mathematics and Theoretical Physics,
University of Cambridge
