This is a book written for the general reader, with the aim of presenting a system that allows us to come by something that we can claim as being knowledge; he tries to deal with the problem of induction. The work appears to be motivated by a desire to philosophically justify the common sense beliefs and behaviour of scientists, and to provide context the book opens with a brief summary of the important things that we know in each of the different fields of science ( circa 1950).
The book is over 500 pages, and there are lengthy discussions on all the relevant areas. These are broken down as far as possible into small chapters, which makes this book more readable than it might otherwise have been. Some of the most important topics which make up the argument presented here are: knowledge, causality, inference, language, observation, other minds, structure, probability, and logic. Probability is probably what is relied on most in escaping the induction problem itself, and this is what in fact nearly all scientists do use to support their hypotheses, after they have collected their empirical data. In the section on probability he dissects several systems of probability, including one of Keynes', then appears to come to some vague conclusion that the use of statistical systems can be justified practically by their successful performance in predicting future phenomena (and this also can be measured statistically). So, empiricism is justified because it works, which is essentially what other books on this problem have decided, though they often reach this from slightly different angles. The section on probability is not light reading, with plenty of formulas, so if the reader is not a keen mathematician, some parts will either be hard work or worth passing over.
It might be worthwhile to compare this book to "The Problem of Knowledge" by A J Ayer, who was a contemporary of Russell who sought to deal with the same problem. Russell's book is more difficult to read (though his style is good and dry wit amusing), and is twice as long, because it deals more thoroughly with several relevant peripheral issues and includes some that Ayer does not. Ayer's book doesn't deal with the practicalities of science, (with which it could be argued that a philosopher needn't concern himself), however Russell rolls his sleeves up and tackles them quite well. Reading between the lines in Ayer, I think he and Russell both share several sentiments on the subject. Neither embrace scepticism (though acknowledging the sceptics point), and both are first class logicians who want to justify a common sense approach to science. Russell's work is the more impressive, due to its its comprehensive scope and greater completeness as a system, but Ayer's work has a greater elegance due in part to its avoidance of clunky topics such as probability which are not traditionally favoured by philosophers. So, anyone interested in understanding the problem would do well to read both books, but Russell's, despite arguably possessing more authority, is less easily approached and requires more effort in parts to digest. ( )

FG-5

Russell, at his analytical best, explores what can be known and what constitutes knowledge itself. ( )