Conjectures and Refutations: The Growth of Scientific Knowledge (Routledge Classics) [Englisch] [Taschenbuch]

Some of the papers in this book make a good introduction to Popper's ideas, but technical discussions of this kind are never easy to read. For instance, if you are unfamiliar with the ideas of Rudolph Carnap, you might want to skip the chapter devoted to him. I had a hard time reading it. Nevertheless, this is probably a better starting point than "The Logic of Scientific Discovery", a very difficult book.

The format of the book as a collection of papers is both a strength and a weakness. Some of the papers are a joy to read, especially when Popper writes about the presocratic philosophers and the birth of science. Popper is very good at introducing his subject, almost as if he were telling a tale. On the other hand, the many repetitions of the same theme become cumbersome after some time. This book is over 400 pages! BIG pages! Apparently, when Popper published this book, he was so famous that publishers uncritically printed anything he wrote, no matter how long-winded. Somehow, this is an ironical illustration of Popper's own thesis.

Two suggestions for reading this book. First, if you are a Popper critic, you NEED to read this book as he goes a long way in explaining many beliefs of his that critics get wrong. Second, do not read the book front to back. As all of these 500+ pages are on the falsification theory applied to different situations, it will get extremely repetitive. Read a few essays at a time and come back later.

Also, Popper's answer to the challenge that Duhem's problem posed on his philosophy is disappointing, the answer being something like "there exists a logical method of proving independence from axioms, so we might hopefully see from which axiomS the falied prediction depended; and even so, I admit that this method is usually difficult to apply; therefore holism is an untenable dogma."

The thesis of the book, says Popper, can be put like this: we can learn from our mistakes. This is held together with this other thesis: there is no ground for believeing any empirical statement to be true. The reader might wonder how Popper managed to believe in these two thesis at one and the same time. In Popper's view, science is this: conjecturing a theory to be true; subjecting this theory to criticism (empirical testing); this testing is done after experiment, but experiments are not reliable, we have no warrant that our perceptual apparatus is not deceiving us; if the theory fails the test, we reject it; but "it" is a whole system of related theories, even observational theories (even logic and mathematics, says Quine); and then we have to guess which of these we have to reject. The risk of taking a true theory to be false is certainly very high, as high as that of taking a false theory to be true. So I don't see how Popper can be so confident that we can learn from mistakes. Perhaps if we purged Popper's methodology of things like truth (not to mention verisimilitude), we could get a methodology of science conceived as a canon of critical procedure, with no claims as to what we are achieving when we abide by it.

The article on hegelian Dialectics is amusing. It tries the impossible task of explaining dialectics in a simple language, and then to refute it. The dialectician's typical reply to this kind of criticism is: you used clear language, so that is NOT Hegel's diatectics.

As I said, this is a highly stimulating and clearly written book, which deserves to be read even if many things in it must to be corrected or complemented.