Eric S. Raymond offers an excellent article on the forms of the Science Fiction genre, and why it matters.

First, I want to be clear on what I think a genre is. It’s two things: one is a set of expectations a reader has about the kind of experience an instance of the genre will deliver, the other is a set of genre-specific codes and expressive techniques that the genre writer uses in the expectation that readers will receive them as the author intended. Like all codes and languages, the purpose of genres is to make communication easier by allowing both parties to assume a repertoire of common referents. Genre art fails when the production of the writer fails to match the genre referents and constraints as known by the reader.

This analysis generalizes Samuel Delany’s observation that SF is not merely, or even mostly, a way of writing; it is a way of reading, too. The same is true of other genres, in different ways.
…
One of SF’s central impulses is to extend the perimeter of the rationally knowable, sweeping in not merely unknown places and times and aliens accessible to science but also motifs and images that originated in myth and fantasy and horror. The evolution of SF can be charted as a steady widening of that perimeter – to other planets, beyond the solar system, to other times and alternate histories, then to technology-of-magic and possibilities even more estranged from the world of immediate experience.

Having advanced this definition of SF, I’m now going to make a temporary concession to people who consider it too narrow by relabeling what it covers “classical SF”, or cSF. Those with a little historical awareness of the field will recognize that the classical period began in 1939 with Robert Heinlein’s first publication under John W. Campbell, the then-new editor of Astounding Science Fiction magazine.

Almost anyone with any exposure to SF will recognize that much but not all of what is popularly labeled SF is cSF. The question I will address in the remainder of this essay is: why should we consider cSF normative? What grounds do we have for regarding a work that claims to be SF but is not cSF to be defective SF?

One reason is historical. Previous attempts to abandon the deep norms of cSF while preserving its stage furniture and surface tropes have not aged well. The “New Wave” of the late 1960s and early 1970s was spent by the early 1980s. Later insurgencies within the field, notably the cyberpunks of the late 1980s and early 1990s, retained cSF’s assumption of rational knowability (and all that followed from it) even while trying to radically transform the genre in other ways.

The reason beneath that history is reader response. SF doesn’t exist in a vacuum; people who want fantasies or Westerns or romances know how to find them, and in general the kind of person who can be attracted by the way SF is packaged (spaceships and other high technology on covers, etc.) wants rational knowability and wants to play the kind of game with the author that is characteristic of cSF, even if he or she is not very introspective about that desire and not very good at the game yet.

This is why SF readers – even inexperienced ones – often experience violation of the deep norms of cSF as a kind of dishonesty or malicious subversion. They can tell they’re being cheated of something even if they don’t know quite what. Forty years ago this feeling was often articulated against the New Wave by complaining that its works were “depressing” – which was true, and remains true of a lot of defective SF and anti-SF today, but doesn’t get at the actual root of the problem.

Read the whole thing.