Publishing is often characterized by instinctive decision-making with little attempt to apply a scientific methodology to an obvious question: why does one book sell and another not? The thesis of this paper is that, although there are aspects of a book’s publication history that one cannot predict in advance, one can know what these aspects are. A simple syllogism underlies the argument: if human behaviour can be understood through psychology and if book-buying is a form of behaviour, the motivations for book-buying can also be understood through psychology. This approach can be applied historically, through recourse to sales data, to trace the fossils of books published long ago and so discover the type and strength of the motivations that once drove people to buy them. History demonstrates that these motivations, once properly framed, can be understood to be influenced by context. Book-buying motivations also appear cyclically. This leads to a discussion of why it is that one book rather than another may satisfy a motivation and therefore sell better than another. Using the concept of prisms combining to reflect a motivational ‘light’, we see that books exist as constructs of a finite range of elements that cohere (or not) in a multiplicative way to enhance or diminish their effectiveness. Evidence is also given for what appears to be a universal ratio that dictates a natural entropy in the effectiveness of these prismatic elements.