EVERY ANTHOLOGY TELLS A STORY about the present state of literature, about categories and connections, about the past and its ongoing relevance. In The Secret History of Science Fiction, James Patrick Kelly and John Kessel offer up an alternative history (a well-established category within science fiction). It starts with a "what if" borrowed from an essay by Jonathan Lethem: what if science fiction had rejoined the mainstream in the early 1970s, when Thomas Pynchon's Gravity's Rainbow was nominated for — but did not win — SF's Nebula Award? How might both science fiction and conventional literature have benefited if, as Kelly and Kessel posit in their introduction, "on one side of the genre divide, SF was being written at the highest levels of ambition, [while] on the other side, writers came to use the materials of SF for their own purposes, writing fiction that is clearly science fiction, but not identified by that name?" Just imagine: we might have writers such as Michael Chabon or Karen Joy Fowler writing recognizable SF, while genre magazines would be publishing stories as polished and as self-aware as any New Yorker piece. 

In Kelly and Kessel's "secret history" of the genre, all of that has come to pass. They have assembled a set of 19 stories, nine by writers who have indeed appeared in The New Yorker and ten by writers who have not. Just to complicate matters, though, the New Yorker crew includes genre favorites Ursula K. Le Guin and Gene Wolfe. All the stories, including one by each editor, are terrific. All employ similar themes and devices and play similar games with perception and identity. The SF writers contribute poignant character studies like Maureen McHugh's "Frankenstein's Daughter" and metafictional experiments like Kessel's "Buddha Nostril Bird." The writers who might be considered "mainstream" take on scientific issues such as the ethics of primate experiments, examined with clinical irony by George Saunders in "93990." Margaret Atwood makes use of science fiction's capacity to generate powerful metaphors in "Homelanding," and so does Connie Willis in "Schwarzchild Radius." 

Atwood is well known in SF circles for displaying a reasonable understanding of SF artistry on even-numbered days and then denying any such knowledge on odd ones. Both Atwoods are represented among the authorial statements used as interludes between the stories. Here is Atwood number one acutely describing the tradition within which SF operates: "This narrative form has always been with us: it used to be the kind with angels and devils in it. It's the gateway to the shadowiest and also the brightest part of the human imaginative world..." And here is Atwood number two making a distinction without a difference: " Oryx and Crake is a speculative fiction, not a science fiction proper. It contains no intergalactic space travel, no teleportation, no Martians." By "speculative fiction" she means what most would consider simply good SF. 

No story here involves intergalactic space travel, teleportation, or Martians (though Michael Chabon's steampunkish alternative history is titled "The Martian Agent: A Planetary Romance"). Many involve scientists and machines; all take place on earth or in nearby orbit; they are about equally divided among present, past, and near future. My sense is that Kelly and Kessel have not skewed the selection. This is an accurate rep...