Reading a serious book of philosophy is a substantial commitment, but Parfit's Reasons and Persons has been recommended to me a number of times, so I'm taking the plunge. The book is made up of four parts, containing 20 chapters consisting in total of 154 sections, each section dealing with a particular topic, like "How Consequentialism Is Indirectly Self-defeating" or "The Effects of Population Growth on Existing People". That's approximately all the introduction that Parfit gives the book, so I'll leave it at that, as well.
You can click through to the book for my detailed notes, but I'll summarize here my thoughts on the first six sections.
The book begins by discussing the Self-interest theory, S:
This theory, he claims, is indirectly individually self-defeating because a person who always follows it (in Parfit's terms, a person who is never self-denying) may, in many circumstances, actually do worse than if he did otherwise. He gives a couple of interesting examples of such situations, though I don't agree with them fully.
He rescues S by saying that, after all, S does not tell us we must believe S if it is worse for us, so it is not failing in its own terms.
Parfit argues that it could be rational to cause oneself to act irrationally, using an example of an imaginary drug that causes one to behave irrationally, which he employs as a defense against blackmail, more or less. He says that this is a rational irrationality, but I'm not very happy with that term or, frankly, the hair-splitting he's doing.
Parfit claims that it's okay if a theory about rationality tells us that we cannot avoid acting irrationally.
That's it for the first six sections. The next few also deal with S, and then Parfit takes up Consequentialism.