"If the assumptions are wrong, the conclusions aren't likely to be very good."

“Every large-scale system is an. The automatic factory, the vehicular-traffic system, any military system- in the design of all these systems, primary attention should be given to the flow of information about the elements of the system.”

“We have discovered in this past decade that thinking, and decision, are not solely the province of the metaphysicist, but are appropriate subjects for scientific inquiry.”

“Most accidents in well-designed systems involve two or more events of low probability occurring in the worst possible combination.”

“Management has a design and operation function, as does engineering. The design is usually done under the heading of organization. It should be noted first that the performance of a group of people is a strong function of the capabilities of the individuals and a rather weak function of the way they are organized. That is, good people do a fairly good job under almost any organization and a somewhat better one when the organization is good. Poor talent does a poor job with a bad organization, but it is still a poor job no matter what the organization. Repeated reorganizations are noted in groups of individuals poorly suited to their function, though no amount of good organization will give good performance. The best architectural design fails with poor bricks and mortar. But the payoff from good organization with good people is worthwhile.”

“We assert that it is possible to describe analytically any human function which can be reasonably defined in objective terms and we specifically include in such functions "thinking" insofar as that term is definable. If by "thinking" one means being able to do arithmetic, or play a good game of chess, or learn from experience, or make optimal decisions in exceedingly complex situations, then we assert that thinking can be described analytically. And there are two important corollaries: if It can be described analytically, it can be simulated; and if it can be simulated, it can be performed mechanically.”

“basic problem … of building a mathematical model of thought processes, and in particular of those aspects of thought which are concerned with information and decision processes. The perceptron is one type of model -- a set of memory devices connected in random fashion-- which has not yet achieved useful results but certainly seems to be a promising approach. The self-adaptive feedback control system which goes beyond the normal servo function of controlling its output, and in addition controls the parameters by which it controls its output is another which has already achieved pragmatic results in equipment control. It may be that the question of self-adaptation is a key to the whole question of how the human functions in a decisioning situation. For in many cases the ability of the human mind to adapt itself to a changing and complex environment is beyond our present aims in model construction.”