The real problem the authors have is with the idea of "just-so" stories used to explain the emergence of traits. For example, the proposal that peacocks grew large tails to impress females, or zebras developed stripes to confuse lions. These types of stories, they say, can't be proved and cast doubt on Darwinism, especially in cases where linked traits cannot be separated and treated as independent variables.
I'm afraid I don't follow the logic. Just because one trait hitches its wagon to another does nothing to disprove the notion of natural selection. Even if we can never really know which traits the environment acted upon, our epistemological conundrum has absolutely zero impact on the actual world where the environment is acting on real, living organisms. It is rhetorical sleight of hand to claim that just because we cannot know something 100% we cannot know it at all.
Oh, and nobody but creationists talks about "Darwinism." Contemporary evolutionary theory has made many leaps and bounds since Darwin (who, after all, knew nothing of genes or DNA).