The question that has to be asked about recent books on the philosophy of science is: Does this book explain the four "turns" that Karl Popper introduced? These are (1) the conjectural turn, to explain that even our best scientific theories may be false, (2) the objective turn to focus on scientific knowledge in its public or objective form, rather than subjective beliefs, (3) the social turn to be aware that the scientist works in a community and there is a need for conventions or "rules of the game" to maintain standards of criticism and best practice and (4) the rehabilitation of metaphysics, in defiance of the positivists and logical empiricists, in the form of "metaphysical research programs".
This book does not score very well on that test however on the positive side the historical approach is very good, introducing concepts in relation to scientific episodes: Galileo on new methods for a new science, Francis Bacon on experiments, Newton on rules for reasoning, Herschel (the astronomer), Mill and Whewell on the use of hypotheses, Venn and Peirce on probabilities as frequencies, Keynes on probability logic, Reichenbach and Popper on induction.
Contrary to the conjectural view of science (small s) this book was written very much in the justificationist or authoritarian mode (the authority of Science with a big S). And so "to explain the success of the work of scientists we will have to refer to the methods they use; we will refer to the reasoning they use to justify their new knowledge." With reference to our confidence in science "we trust scientific theories simply because they are scientific and therefore authoritative or we count claims as scientific and reliable because they are established by scientific methods.' That orientation begs all the questions.
An interesting novelty in the book is the introduction of John Maynard Keynes who was a serious student of probability theory (the probability of theories or beliefs, or expectations) before he made his name as an economist.
In view of the central importance of the problem of induction in modern times this review concentrates on the final chapters about Reichenbach, Popper and Carnap. Gower was even-handed regarding Reichenbach and Popper, pointing out that the former encountered a number of insoluble problems with his project to save inductive logic with a method to assign objective probability values to theories.
He then argued that Popper had similarly run into trouble with his campaign to dispense with induction altogether. First he suggested that the claim of asymmetry between verification and falsification was not as clear as Popper declared. That objection fails because Popper drew a very clear distinction between the logic of testing and the practical problem of establishing a negative result. He described this as the difference between falsifiability (the capacity for a generalization to be refuted by a single true observation statement) and falsification, the practical matter of testing with all the complications of theory-dependence of evidence, the Duhem problem and so on.
The logic of the situation is quite clear (it was endorsed by W V O Quine) and also by Richard Feynman, as cited by Gower on page 15 of this book. "If it [the computed consequences of a law] disagrees with experiment it is wrong. In that simple statement is the key to science ."
The problem of handing the practical side of testing prompted Popper to develop the idea of conventions or rules of the game, not set in concrete but subject to criticism and improvement like scientific theories themselves. For an extended treatment of this aspect of Popperism, see Ian Jarvie's book "The Republic of Science". http://www.amazon.com/review/R37DB4M4BDDI0P/ref=cm_cr_pr_perm?ie=UTF8&ASIN=9042015152&linkCode=&nodeID=&tag=
The second line of attack is to suggest that "induction in a broad sense" is involved in justifying Popper's preference for well tested (corroborated) hypotheses over those that fail. However this involves a very different use of induction from the inductive logic that is supposed to assign a numerical degree of confidence to a theory. It is induction in the sense of the metaphysical theory that there are regularities or laws of nature which provide a degree of structure and predictability in the world. Given Popper's conjectural theory of knowledge, preferences are not set in concrete and are liable to revision in the event of new evidence, new arguments and new theories in the contest.
The final chapter on Carnap describes now his lifelong project to develop a system of inductive logic failed and morphed into the very different program to quantify subjective feelings of probability which we now know as Bayesian subjectivism. This can be seen as the last bastion of subjectivism and inductivism. It remains to be seen how long it will last before people find that there is a better way with Popperian critical rationalism and objective knowledge.
4 people found this helpful.
Was this review helpful to you?