I've wrestled for a while with the idea that "SF is the genre of ideas". I have argued on both sides, and eventually come to the conclusion that SF really refers to a plethora of very different genres with very little in common. Recently, I've grown fond of Neal Stephenson's statement that SF is defined by the fact that the characters behave competently and intelligently (even if this is lacking in verisimilitude -- after all, the real world is arguably better-defined by accidents and screw-ups than by intelligent and informed choices). Nevertheless, the plain fact is, some SF is truly mind-bending and the rest of it simply isn't.
What I think it all comes down to is novelty, and novelty concentrations.
We've talked about novelty on this blog before. Novelty is Shannon-information. It's the unexpected. It's the data point that changes the model. It's the punchline of a funny joke, or the reveal that retcons all of continuity. The only thing that is mind-bending or mind-expanding in of itself is novelty, because everything else you've seen before.
Novelty is not a renewable resource. Not within a single skull, without some major memory problems. Novelty is exhausted immediately upon consumption. Luckily, novelty is not the same as originality. At the time of publication, patents are deemed novel to the patent clerk who approved them, and many patents are for improvements that are obvious in retrospect. Good novelty (or useful novelty) has that funny quirk: it's obvious in retrospect and entirely unexpected beforehand.
Novelty can be gotten cheaply, in relatively small quantities, under circumstances of isolation of social groups. The character of the yokel shocked by the character of the big city is representative of this, as is the culture clash between European explorers and indigenous Americans during the sixteenth century. But, now that the internet makes large quantities of information easily transferred, the only large caches of naturally (which is to say unintentionally) produced novelty are those things completely undocumented and those things intentionally blocked. People who subscribe to their 'Daily Me' and limit their information intake to their filter bubble can still have their mind blown by the ideas of people in different filter bubbles, but such people are rarely habitual novelty-seekers.
For habitual novelty-seekers, the reality tunnels of other people are interesting until exhausted. Speculative fiction comes in at this point, and in this sense it often is the genre of ideas.
Novelty can be generated in different ways. Some of them are fairly mechanical. Burroughs had his cutups, and the Surrealists had their exquisite corpse. Various mind-altering drugs and habits of thought like Dali's paranoiac-critical method can be used to do to the mind what cutups and the exquisite corpse do to text. These methods are much like taking a computer and banging it with a hammer in the hope that it will become an automobile, and they are successful only to the extent that the brain is a wonderfully adaptable machine capable of making sense out of any sort of noise. A more directed and less wasteful but still fairly mechanical method is to take an existing memeplex, find a likely crux, and reconnect or recombine it. This is more like what glitch musicians do, or what gametes do in sexual reproduction. This happens to be the core of speculative fiction, and what many of the better writers start out with.
Take a model of the world. Zoom in on one piece. Twist until mind blown. Spew the resulting model onto paper, with a narrative glued to it in a bag on the side. Rinse and repeat.
What you get is a synthetic reality tunnel. Speculative fiction authors don't just write about robots and space ships; they mutate their view of reality systematically to manufacture new ways of modeling the world that are internally consistent but that nobody really subscribes to yet.
The important part is 'yet'.
Highly popular pieces of speculative fiction get widely read. The model of the world in these stories gets incorporated into the world-models of other people. The novelty seeps away as cultural osmosis sets in.
Many people, for instance, have a view of evolution as guided progress toward a pre-ordained goal. This forms the basis of works as far-flung as 2001 A Space Odyssey, The Starseed Transmissions, Childhood's End, Altered States, and The X Men. Of course, this model is bullshit, and does not describe evolution at all. Few people realize that it came from a handful of science fiction stories in pulp magazines during the first two decades of the twentieth century, most notably The Man Who Evolved.
Certain ideas are no longer novel at all, purely because they were initially considered extremely novel and therefore reached maximum saturation much more quickly.
