The Book That's Probably Wrong About Itself

David Makinson published 'The Paradox of the Preface' in Analysis in 1965. The scenario is familiar: a careful author writes a book of historical claims, each researched and each believed to be true. Then, in the preface, the author writes: 'Despite my best efforts, this book probably contains at least one error.'

This is not false modesty. It is epistemically correct. Any sufficiently long book of factual claims, written by a fallible human, is likely to contain a mistake somewhere. Readers and authors both know this. The preface acknowledgment is rational.

And yet: the author believes proposition P1, and believes P2, and believes P3, all the way through to PN — and also believes that at least one of P1 through PN is false. Standard logic says you cannot rationally believe both a set of claims and their negation. If you believe each claim, you should believe their conjunction, and the conjunction cannot be false if all the conjuncts are true.

The paradox shows that rational belief does not always aggregate neatly. Believing each proposition in a set does not force you to believe all of them together, especially when the set is large enough that error becomes probable. Makinson's puzzle pushed epistemologists to think harder about the difference between individual belief assessments and coherence across a whole body of belief.

The paradox connects directly to the Lottery Paradox: if a lottery has a million tickets, you rationally believe of each ticket that it will lose — but you know one of them will win. Believing each proposition and disbelieving their conjunction is, apparently, something we do all the time.