There is a rage shared by most critics of the literature of the fantastic. It is the rage we feel when some iteration of that literature--a novel by Jeff Noon, perhaps--is mufflingly misdescribed as non-generic by its publishers, or by some moat-defensive critic more concerned to defend his patch than to tell the truth about the text before his eyes, or even by authors--like Jeff Noon himself, whose increasingly chrome-plated career track seems to require a repudiation of his roots in genre. Noon's recent statement that he does not write sf, and his publisher's contortuplicated efforts not to mention sf in the jacket copy to Falling Out Of Cars (London: Doubleday, 2002)--jacket copy which manages therefore not to mention that Falling is set in a near-future England decimated by a strange plague whose effects on humans can only be staved off by a brandname drug distributed by a mysterious corporation--does rather seem a trahison des clercs.
Because what Noon and his publishers have done to Falling Out Of Cars is a discourtesy to adult readers. They have unlabelled the book, which may sound a noble thing to do ("Let my fable go") but which is not. In the sick, febrile market now operating in the book trade, a book which is unlabelled is not a book readers come to with eyes washed of preconceptions, like Israelites entering a Promised Land; it is a book precisely marketed to mandate a particular kind of preconception in the reader: which is that the book in question is safe, that it is a mundane extension of the mimetic novel, that it is unlabelled because it is unnecessary to label a window into the real world.
--John Clute, Scores, p.392
And then there is an argument based around reading protocols, which goes something like: when Falling Out Of Cars is approached as mimetic, it reads very differently, and (crucially) worse, than when it is approached as science fiction. Having not read the book, I can't comment, but that's ok; my reservation is about his more general point.
There is an awkward interaction here between, as ever, science fiction the marketing category and science fiction the form. I agree with the above to the extent that I recognise the rage, because I agree that most decisions to unlabel a book seem to be made in, for want of a better phrase, bad faith--made as marketing decisions only. I'm less convinced that this matters because it leads us into the work by the wrong route.
Firstly, I'm inclined to think we can work out how to read something as we go, based on the text itself. The label is useful, but it's not essential. (And I can think of cases--one of Michel Faber's books comes to mind--where the unfolding of expectations is so deliciously well-handled that it would be a bit cruel to give the game away on the dust jacket.) But more than that, I'm inclined to think that the act of unlabelling usually Just Doesn't Work. There are arguable cases, like Never Let Me Go, but in general it's a safe bet that a review that bends over backwards to explain how a book isn't science fiction, honest, is in fact a review that tells you, very clearly, precisely the opposite.
So why the rage? For me, Clute got it right at the start: because it's not telling the truth. I'm sure
immortalradical will come along at this point and accuse me of trying to claim texts "for the genre", but I don't think that's what I'm doing; to say something is science fiction doesn't have to involve saying that it's part of the genre. My problem is rather that all the evasions and denials smack of either simple snobbery--the "it's good, so it can't be sf" attitude--or, more seriously, a patronising lack of faith in the reader--"don't worry, this one's ok because it's not really sf." And that's a disservice not just to the reader, but to the book and to the form as a whole.