One of the more rewarding books to cross my desk in the past year was Elliot Sober's Evidence and Evolution. The name might be deceptive: this is not an easy-to-read book setting out various bits of evidence for evolution (like Jerry Coyne's Why Evolution is True); this is a long, technical and difficult book on epistemology and the philosophy of science.
It was rewarding because it introduced me to a new way of looking at certain issues. Prior to reading Sober, I was a bit of Bayes-virgin. I have now been engrossed in the topic for six months or so (I'm still little more than a dabbler).
Sober covers four topics in what are really four separate books:
- The different statistical tools we can use to assess theories (Bayesianism, Likelihoodism, and Frequentism).
- Intelligent Design (focusing a lot on the correct formulation of the design argument and the testability of hypotheses).
- Natural Selection (comparing it to other theories of evolutionary change).
- Common Ancestry (testing the theory of branching descent).
As a good philosopher, Sober reaches no firm conclusions on any of these topics.
I am going to do occasional posts on this book. I will not go through it in the same detail as I am doing with other books on this blog. That said, I am going to take a considerable whack at the chapter on Intelligent Design simply because it straddles two of my own interests: philosophy of science and religious philosophy. I am also going to use some of the material from chapter one in a series covering Bayes Theorem.
I am going to do occasional posts on this book. I will not go through it in the same detail as I am doing with other books on this blog. That said, I am going to take a considerable whack at the chapter on Intelligent Design simply because it straddles two of my own interests: philosophy of science and religious philosophy. I am also going to use some of the material from chapter one in a series covering Bayes Theorem.
I will take this at a fairly leisurely pace. This is a book to be savoured and chewed-over, not gulped down at one sitting.
Anyway, this post will, eventually, serve as an index for all others.
You, sir, are crazy ambitious.
ReplyDeleteYou desperately need a 'Table of Contents' page that lists all your series indices, just like on my site.
I'm working on it.
ReplyDeleteOf course, Sober is a likelihoodist, not a Bayesian.
ReplyDeleteYes, I read your review. As you said, book one seems to adopt a form of pluralism, but he prefers likelihoodism in the other parts.
ReplyDeleteThe chapter on intelligent design, which I will cover in depth, is all about likelihoodism.
Still, I thought chapter one was a good intro to the whole area (at least, it was for me).
ReplyDelete