This new edition contains, apart from a general revision of the text, a considerable amount of historical material which accumulated since the first edition was printed. As far as possible I have tried to leave the pagination unchanged so that references to the first edition will in almost all cases agree with the second edition. There is also an addition to the end of chapter 5, and a new Addendum (6) at the end of the book. Alan Musgrave has completely revised the indexes, and has also given me much help with improvements to the body of the book.
Having tried, in my first Preface, to sum up my thesis in one sentence—that we can learn from our mistakes—I may perhaps add to it a word or two here. It is part of my thesis that all our knowledge grows only through the correcting of our mistakes. For example, what is called today ‘negative feed back’ is only an application of the general method of learning from our mistakes—the method of trial and error.
Now it appears that in order to apply this method we must already have some aim: we err if we stray from this aim. (A feedback thermostat depends on some aim—some definite temperature—which must be selected in advance.) Yet though in this way some aim must precede any particular instance of the trial and error method, this does not mean that our aims are not in their turn subject to this method. Any particular aim can be changed by trial and error, and many are so changed. (We can change the setting on our thermostat, selecting by trial and error one that better satisfies some aim—an aim of a different level.) And our system of aims not only changes, but it can also grow in a way closely similar to the way in which our knowledge grows.
Penn, Buckinghamshire, January 1965
K. R. P.