Mistakes are at the very base of human thought, embedded there, feeding the structure like root nodules. If we were not provided with the knack of being wrong, we could never get anything useful done. We think our way along by choosing between right and wrong alternatives, and the wrong choices have to be made as frequently as the right ones. We get along in life this way. We are built to make mistakes, coded for error. 

We learn, as we say, by "trial and error." Why do we always say that? Why not "trial and rightness" or "trial and triumph"? The old phrase puts it that way because that is, in real life, the way it is done. 

A good laboratory, like a good bank or a corporation or government, has to run like a computer. Almost everything is done flawlessly, by the book, and all the numbers add up to the predicted sums. The days go by. And then, if it is a lucky day, and a lucky laboratory, somebody makes a mistake: the wrong buffer, something in one of the blanks, a decimal misplaced in reading counts, the warm room off by a degree and a half, a mouse out of this box, or just a misreading of the day's protocol. Whatever, when the results come in, something is obviously  screwed up, and then the action can begin. 

The misreading is not the important error: it opens the way. The next step is the crucial one. If the investigator can bring himself to say, "But even so, look at that!" then the new finding, whatever it is, is ready for snatching. What is needed, for progress to be made, is the move based on the error. 

Whenever new kinds of thinking are about to be accomplished, or new varieties of music, there has to be an argument beforehand. With two sides debating in the same mind, haranguing, there is an amiable understanding that one is right and the other wrong. Sooner or later the thing is settled, but there can be no action at all if there are not the two sides, and the argument. The hope is in the faculty of wrongness, the tendency toward error. The capacity to leap across mountains of information to land lightly on the wrong side represents the highest of human endowments. 

It may be that this is a uniquely human gift, perhaps even stipulated in our genetic instructions. Other creatures do not seem to have DNA sequences for making mistakes as a routine part of daily living, certainly not for programmed error as a guide for action. 

We are at our human finest, dancing with our minds, when there are more choices than two. Sometimes there are ten, even twenty different ways to go, all but one bound to be the wrong, and the richness of selection in such situations can lift us onto totally new ground. This process is called exploration and is based on human fallibility. If we had only a single center in our brains, capable of responding only when a correct decision was to be made, instead of the jumble of different credulous, easily conned clusters of neurons that provide for being flung off into blind alleys, up trees, down dead ends, out into blue sky, 
along wrong turnings, around bends, we could only stay the way we are today, stuck fast. 

The lower animals do not have this splendid freedom. They are limited most of them, to absolute infallibility. Cats, for all their good side, never makemistakes. I have never seen a maladroit, clumsy, or blundering cat. Dogs are sometimes fallible, occasionally able to make charming minor mistakes, but they get this way by trying to mimic their masters. Fish are flawless in everything they do. Individual cells in a tissue are mindless machines, perfect in their performance, as absolutely inhuman as bees. 

We should have this in mind as we become dependent on more complex computers for the arrangement of our affairs. Give the computers their heads, I say; let them go their way. If we can learn to do this, turning our heads to one side and wincing while the work proceeds, the possibilities for the future of mankind, and computerkind, are limitless. Your average good computer can make calculations in an instant, which would take a lifetime of slide rules for any of us. Think of what we could gain from the near infinity of precise, machine-made miscomputation which is now so easily within our grasp. We would begin the solving of some of our hardest problems. How, for instance, should we go about organizing ourselves for social living on a planetary scale, now that we have  become, as a plain fact of life, a single community? We can assume, as a working  hypothesis, that all the right ways of doing this are unworkable. What we need,  then, for moving ahead, is a set of wrong alternatives much longer and more  interesting than the short list of mistaken courses that any of us can think up  right now. We need, in fact, an infinite list, and when it is printed out we need  the computer to turn on itself and select, at random, the next way to go. If it is  a big enough mistake, we could find ourselves on a new level, stunned, out in  the clear, ready to move again. 

Source: The Medusa and the Snail, Viking Penguin, 1976