I found this book absolutely fascinating. Most books on science recount history, developments, progress and/or current events in some scientific field. But this book, on the other hand, describes how the rate of scientific progress can be measured, how it behaves as a function of time and how future progress can be predicted. The author also discusses how facts can evolve over time, how facts that have been proven wrong can persist, how scientific publishing really works and much more. The author writes well. I found his prose to be clear, friendly, lively and quite engaging.
Unfortunately, I did find a couple of instances where the information presented was rather misleading at best. In particular, on page 104 (near the middle), it is stated that "Richard Feynman ... shared the Nobel Prize with another physicist, Sin-Itiro Tomonaga". In fact, Feynman shared the 1965 Nobel Prize with two other physicists, not just one: Tomonaga and Julian Schwinger. Also, near the middle of page 123, it is stated "... when dry ice becomes carbon dioxide...". In fact, dry ice is carbon dioxide in the solid state.
Notwithstanding these minor shortcomings, I found this book to be quite absorbing and I learned quite a bit from it. Although it should appeal to a fairly wide readership, I believe that scientists and science enthusiasts would likely enjoy it the most.
on July 7, 2015
The Half Life of Facts, as other reviewers have noted, is quite refreshing for a popular-level science book: instead of repeating some discipline-specific case studies a reasonably well-read science enthusiast has seen dozens of times, it zooms out one level to look at science as a whole. Rather than take a philosophy of science or epistemological angle, though, Arbesman presents an overview of the *scientific* study of science, a field that I always vaguely knew existed but had never really investigated. I like to think of it as the 'natural philosophy' extension of philosophy of science.
The half-life of facts per se is really only chapter one of the book, but throughout, Arbesman details the ways in which knowledge is gradually replaced. Rather than rely on the old Newton/Einstein paradigm shift so heavily used by philosopher and historian of science Thomas Kuhn, Arbesman gives hard stats on how different fields tend to discard half of their knowledge at different rates. Softer sciences like psychology tend to have shorter half-lives (~5 years) than harder sciences like physics (~50 years). Medicine and economics, if I recall correctly, tended to fall somewhat closer to the shorter end of the spectrum. Arbesman also talks a lot about trends in citations—for example, something like a third of all papers published are never cited—and how citation counts are used, for better or for worse, as a proxy for the confidence we have that a bit of knowledge is the Truth.
In short, Arbesman's The Half Life of Facts is a pretty solid knowledge return on time investment, covering a somewhat neglected subject in a very accessible manner. Four stars overall.
on May 21, 2013
I'm still reading this book. It provides many insights not only into the history of Science but the way scientists really work, for better and for worse. It also touches on many subjects which interest me such as networking and expert software as aids to discovery and understanding of facts. It is readable and entertaining and has proven much more interesting than I anticipated from a cursory glance. I have cited it several times in discussions of various subjects, not all related to the half-life of facts. It's one of those popular non-fiction books that really add something to your mental tool-kit like Taleb's THE BLACK SWAN or Hofstader's class GÖDEL, ESCHER, BACH. It doesn't talk down or hype up, but it really opens up new ways of looking at science and knowledge, which it would surely do for the scientist and historian as well as the scientific laity.