The book is divided into three parts. Part I discusses the motivation and justification for the entire project: Can science study religion? Should science study religion? After answering in the affirmative, Part II proceeds to use the tools of evolutionary biology and memetics to suggest possible theories regarding the origin of religion and subsequent evolution of modern religions from ancient folk beliefs. Part III analyzes religion and its effects in today's world: Does religion make us moral? Is religion what gives meaning to life? What should we teach the children? Dennett bases much of his analysis on empirical evidence, though he often points out that much more research in this field is needed.
Dennett's working definition of religions is: "social systems whose participants avow belief in a supernatural agent or agents whose approval is to be sought." He notes that this definition is "a place to start, not something carved in stone."
Charles T. Rubin, in the The New Atlantis, likened Dennett to "a tone-deaf music scholar" and criticized his "unwillingness to admit the limits of scientific rationality," accusing him of "deploying the same old Enlightenment tropes that didn’t work all that well the first time around."
In Scientific American, George Johnson describes the book's main draw as being a "a sharp synthesis of a library of evolutionary, anthropological and psychological research on the origin and spread of religion."