Our capacity for deceiving ourselves about the operation of our brain is almost limitless, mainly because we can be aware of only a minute fraction of what goes on in our head. That is why much of philosophy has been barren for more than 2,000 years and is likely to remain so until philosophers learn to understand the language of information processing (Francis H.C. Crick, 1979).
IntroductionTuring machines are fully equivalent to large lookup tables. Given an input (conventionally divided into program and data), the output is determined. Indeed information is always lost or dissipated. This is generally called information processing. It is fully deterministic.We want to explore information creation. This is not generally considered a viable task for a computer. We will show that it is possible to build: q absolutely creative computers whose probably-useful output is unpredictable even in principle; q effectively creative computers whose probably-useful output is unpredictable in practice. Both are modelled on the classic four-step creativity process described by Poincaré, Haddamard and numerous other highly creative mathematicians and scientists. No claim of biological plausibility of this model is made. Rather, we claim that it works like the best reports of human creativity from the most creative humans and that it does create useful information.The four-step model Various terminologies have been used for the four-step model. They are readily translatable into these:Step 1: problem choice/study. We do not solve random problems consciously or unconsciously. We choose problems. We consciously study them. We prepare ourselves to recognize solutions should they occur to us. We gather information. We try to solve problems.qStep 2: incubation. By hypothesis, this step is subconscious, so there is little we can say about it -unless, of course, we could model it and show that our model works like our subconscious appears to.The computer subconscious 47 qStep 3: breakthrough. We suddenly have an "inspiration". We say "ah ha" or "eureka". Again, without a model, this is mysterious.qStep 4: development. Usually our breakthrough requires some sort of conscious effort before we are ready to share it.
ModelIn the model subconscious, there are many modules each assigned to solve the problem using available information and concepts strung together in various ways. Potential solutions are evaluated and the better ones share and/or exchange information. Poorer solutions are replaced by better ones. The population of would-be solutions evolves. If we are lucky, one solution eventually exceeds our threshold of goodness and becomes "conscious". Our ideas require "development", because they are not consciously connected to prior work. Mathematicians and scientists then: q work backward from the now-"known" answer to the accepted facts and theories;q publish the simple forward path. No wonder so many great inventions look "trivial". The path we publish is a path we could have taken, but never took. Our subconscious took a foreverun...