He lays the groundwork in chapter 1. One typical way to deal with ID challenges (especially in "teach the controversy" settings) is to dismiss it as unscientific and therefore unfit for discussion in scientific settings (such as a science classroom). One reason frequently given for this status claim (or lack of status claim, if you will) is that ID is untestable, and testability is a necessary condition for a theory being scientific. Not so fast, says Kitcher. The assertion that a theory is untestable is difficult to make stick (or difficult to make stick only to ID), since many scientific theories began as hypotheses that could not be directly tested, but only became testable through the addition of auxiliary hypotheses. Who is to say these auxiliary hypotheses may not be forthcoming for ID? In addition, ID is clearly related to the sciences historically through natural philosophy. Instead of classifying ID as unscientific, Kitcher argues ID should be considered "dead science." It's a model that's been tried and found explanatorily wanting.
Kitcher turns to the arguments of Darwin's critics. Those who wish to find some role for the supernatural within (or at least consistent with) a scientific understanding of the world have used three main strains of argument to do so. Kitcher notes that each strain of argumentation has been less ambitious than its predecessors. Chapter 2 takes on Genesis creationism, or biblical literalism. He reviews a few attempts in pre-Darwinian natural philosophy to find a coherent scientific explanation for the account of nature presented in Genesis. Thomas Burnet, for example, wrestled with finding mechanical explanations for scientific puzzles surrounding the account of the flood as presented in the Bible. But as geology developed as a discipline it became increasingly apparent that things like sedimentation rates and the ordering of the fossil record were nothing like we should expect if Genesis were literally true. By the 1830s even devout scientists like Adam Sedgwick had given up the idea that the Bible is a reliable book of scientific instruction. Contemporary forms of biblical literalism persist (Ken Ham's merry band, for example), but they lack the ability to provide a scientifically coherent account of their own. Instead, they attempt to gain plausibility as an "alternative" by focusing on the unsolved puzzles of science, which are relatively few when compared to the mountains of evidence that have been assembled.
In chapter 3 Kitcher addresses the second strain of Darwin criticism by supernaturalists, which Kitcher calls "novelty creationism." While conceding that the earth is much older than a Genesis based account of geological history would suggest, novelty creationism claims we can still defend the notion that species (all or some, or maybe just one really special one) is the product of special creation rather than the material processes of variation and natural selection. Kitcher recites the litany of reasons that special creation has no legitimate place in modern science: junk DNA, homology of structures (like limbs) despite the different uses to which organisms put those structures, geographical distribution, similarity of the chromosomal structures of apes and men, etc. As with biblical creationists, the only line of defense of their preferred theory is to attempt to point out the as yet incomplete features of our knowledge. No full theory of novelty creation can be assembled because it simply doesn't appropriately account for the facts of nature as we now know them to be.
In the fourth chapter Kitcher turns to the most recent and least ambitious attempt at finding a place for supernatural intervention within a scientific account of the world, ID. ID proponents can accept the findings of geology in determining the age of the earth and the principle of common descent in explaining all the facts of nature that novelty creation could not. ID makes the much more limited claim that there are some biological structures or processes that simply cannot have been assembled over the course of thousands of generations, and therefore must be the product of some intelligence. But there are clearly problems with this approach.
Kitcher analyzes one paradigm case that ID has championed as demonstrating the inadequacy of gradual selection: the bacterial flagellum. The argument that ID proponents like Michael Behe have put forward is that in order for a bacterial flagellum to work, it must possess all the component parts at once: what good is half a bacterial flagellum? The claim is that the flagellum thus must have appeared whole, all at once. Kitcher rightly points out that unless ID wants to revert back to novelty creation, it has to account for the descent of bacteria with flagella from those without. Which means the intelligence must have used the standard genetic materials available in precursors without flagella to make a flagella. But ID has identified no mechanism by which to explain how the intelligence designing the bacterial flagellum does so in one step rather than gradually, and therefore has not provided its own account of the features it claims that standard evolutionary biology cannot provide.
But even more to the point, if the intelligence is fiddling with genetic materials, then what's the explanation for its failure to eliminate genetic disorders? Kitcher is applying the problem of evil to genetic structures and questioning (in good Humean fashion) what the character of the designed might tell us about the designer. If you think natural selection operating on Mendelian gene distributions is responsible for the appearance of design, then you should expect that there will be some genetic abnormalities. But what about the intelligence posited by ID? How can we account for his bungling the genome to give rise to sickle-cell anemia when he's supposedly adept enough at genetic manipulation to produce a flagellum in one fell swoop?
Kitcher's final chapter outlines the logical consequences of the failures of the variety of creationisms to provide a coherent addendum to a scientific worldview. For Kitcher the implications are clear: the sort of world described by science is vastly different than the one described by supernaturalism, and we have no good reasons to adopt supernaturalism. These implications explain the periodic hostility of religious systems to Darwinian evolution. The facts uncovered by evolution do not sit comfortably with a vision of the world as created by a divine being whose primary characteristic (or at least one important characteristic) is love for his creation. Traditional religion, relying on providential intervention in the world, becomes no longer tenable. One might expect a Richard Dawkins-style dismissal of religion here, but Kitcher does not go that route. Instead, he recommends that religion move toward something he calls "spiritual religion."
What form does this spiritual religion take? In the case Christianity, Kitcher envisions a religion in which
[s]piritual Christians abandon almost all the standard stories about the life of Jesus. They give up on the extraordinary birth, the miracles, the literal resurrection. What survive are the teachings, the precepts, and the parables, and the eventual journey to Jerusalem and the culminating moment of the Crucifixion. That moment of suffering and sacrifice is seen, not as the prelude to some triumphant return and the promise of eternal salvation--all that, to repeat, is literally false--but as a symbolic presentation of the importance of compassion and of love without limits (p.152).Kitcher goes on to discuss the importance the church plays in the lives of religious people--churches are places where people go to meet friends, to celebrate joyous occasions, to find comfort in times of sorrow. All of this is true, and Kitcher is right when he talks about how the recommendations of Dawkins or Jerry Coyne for the faithful simply to abandon their religious communities seem a little tin eared. There are a variety of social and psychological factors not necessarily directly dependent on the content of the doctrines that make people value those communities.
Recognizing all this, I really have only one criticism of Kitcher's proposal. But it's a big one. Almost no one will find that version of Christianity palatable. Secularists won't need it; religious people will by and large dismiss it as worthless. While he is right to point out that the social and psychological aspects of religion are important to people, his proposal fails to account for the fact that people are seeking a genuine, true account of the way the world is. There is no noble lie which Christians will willingly tell themselves to maintain the auxiliary aspects of Christian life (unwillingly may be another story). Take the resurrection, which Kitcher seems to think people will be willing to give up. In Christian theology the resurrection is not a "take it or leave it" kind of proposition. It forms the core of Christian beliefs. If Jesus never existed or never rose from the dead, then none of the rest of it makes any sense. If Jesus' death doesn't culminate in atonement (which the resurrection is considered crucial to) then why is it symbolic of compassion or love? What sense does it make to talk this way?
Despite my quibble with where Kitcher ends up in his discussion of Darwinism, design, and religion, the book is really excellent. It's clearly and concisely written and contains enough detail to support his points but without bogging it down in trivia. I highly recommend it to anyone interested in the intersection of religion and science.