In perception itself, two distinct processes can be discerned. One is the gathering of the primary, sensory data or simple sensing of such things as … - Edward Ihnatowicz

Two small and rather unappealing animals possess patterns of behaviour that have great relevance for the student of intelligent systems. These are the wood louse and the maggot of the common housefly, and it is the difference in their behaviour which is so illuminating. It has to do with the way in which they orient themselves to their environment. Wood lice like moist places and succeed in aggregating there by the simple device of slowing down their otherwise random movements as the humidity increases. The maggot, which, during a certain stage of its development needs to come out of the dark, finds light by a slightly more sophisticated system. It has a single, non-directional light-sensing organ at the forward end of its body and as it moves along it swings this end left and right, allowing the amount of light gathered during each swing to determine the extent of its forward motion. In this way it keeps altering its course until the amounts of light sensed are equal for both sides, by which time it must be heading straight for the light.

The idea of making machines that think has an unfailing fascination, not only for science fiction readers, but for all who can see it is a possible way of gaining some understanding of the working of our own minds. Thinking, however, is not an easily defined phenomenon, although it is often considered to be the process of solving problems.

I am not really suggesting that we should construct cybernetic cigarette boxes, merely that the situation described contains all the elements which would have to be taken into account in a design of a practical cognitive system. These can be described as: