(The following editorial first appeared in the November 2009 issue of Analog Science Fiction and Fact.)

It has long been common in some circles of science fiction fans to refer to people who don't share their taste in literature, or their visions of the future, as "mundanes." So it's perhaps a bit ironic that in the last very few years a movement has arisen within science fiction calling itself "mundane science fiction."

The idea, in short, is that many of the common themes and trappings of science fiction have little basis in established science and so don't really belong in the field. To paraphrase a few of its core tenets, as I understand them: There is no evidence that time travel, alternate universes (and travel between them), or faster-than-light travel is possible. Even slower-than-light interstellar travel or communication is so overwhelmingly difficult that it is unlikely to be achieved by us or anybody else—even if there is anybody else, and there's no real evidence for that, either. And even if there were, aliens would likely be so fundamentally different that no real communication would be possible between them and us.

Therefore, the mundane-SF advocates tell us, it is overwhelmingly likely that the future of humanity will be one of humanity alone on Earth or in its relatively immediate vicinity. Therefore and furthermore, that's where science fiction should concentrate its efforts: on imagining what our future is really likely to be, right here on the world of our birth, with nothing to rely on but the resources it provides and no one to rely on but ourselves. And since we're trying to be as realistic as possible, we should do our imagining strictly within the framework of the science we actually know—which implies both capabilities and very strict limits.

No wonder they call it "mundane."

Actually, I have a certain amount of common ground with the mundane-SF people. I have long held that one of the important things for science fiction to do is pretty much what they describe: imagining carefully and in considerable detail what can be done, both good and bad, by application of the scientific principles that we already know. We certainly need writers doing that.

Where we part company is that I maintain that that's only one of the things we need writers doing, while they explicitly discourage doing anything else. Writing about futures in which habitable worlds are plentiful and humans travel among them and interact with their nonhuman but intelligent inhabitants, they tell us, encourages false hopes and a wasteful attitude toward our limited resources.

But they overlook a number of important facts and distinctions. Writers about interstellar travel, first contact, or interstellar federations are not, in my experience, saying that this is the way the future is going to be—but those who advocate limiting ourselves to Science As We Know It are saying that this is, at least within broad limits, the way it's going to be.

The assumption that we—or they—know what the most probable future is seems to me arrogant in the first degree. It assumes there will be no more major surprises in the future, and I'd love to know how they can know that. There have been several huge ones in the last century or so—relativity, quantum mechanics, plate tectonics, and the wild and wonderful world of DNA, to name just a few. The idea that we're finished and there will be no more seems to me far-fetched in the extreme.

Let's see . . . what would a conscientious mundane-SF writer in 1900 have considered fair game for his speculations? Urban pollution caused by too many horses in the streets? Maybe even automobiles replacing some of those horses? Radiotelegraphs? Possibly airplanes would have been accepted by some, but not all.

But such a writer would certainly have considered it unacceptably far-fetched to mention the tiny, powerful computer on which I'm writing this, or the internet on which readers will argue about it, or the communications satellites that enable people to chat with families on the other side of the world. The list could go on for quite a while, but the principle is simple and essential: Our most probable future is one that we can't predict, because it will include not only obviously logical outgrowths of what we already know, but lots of disruptive surprises. Relatively little ones, resulting from unexpected convergences of seemingly unrelated sciences and technologies (as the CT scan resulted from the fusion of x-ray imaging, medicine, and high-speed computing); and huge ones, resulting from fundamental shifts in our understanding of the world, like relativity, quantum mechanics, and DNA-based genetics.

Writers who try to be conscientious and responsible by refusing to try to guess what those surprises might be can be assured of only one thing: that the future they try so hard to make realistic will be anything but—because it will not take into account the biggest changes that will shape it.

Writers who do try to guess what some of the surprises might be can't expect to get very many of them right, either. But at least they will get the fact that there will be surprises, and they can attempt to explore how people might react to them and how their lives might change as a result. That's important, too. Since we can be pretty sure that future science and technology will take some unexpected turns, exploring how people might handle them can be at least as important as working out the details of developments whose possibility is already obvious.

And, just occasionally, one of those wilder imaginings may turn out to be closer to something we actually get than any of the more careful, conservative extrapolations.

Extrapolation. That's the word I have long used to describe essentially what the mundane-SF folks consider their lofty aspiration: to imagine rigorously what developments might actually grow from already-known science. They consider it a prescription for What Science Fiction Should Be. I consider it a prescription for one of the two main kinds of science fiction.

The other, which I consider no less important, I call "Innovation." That refers to stories that depend not just on extrapolation from known science, but on imagining fundamentally new kinds of science that might conceivably be discovered in the future, and working out what might become possible as a result.

Note carefully that this does not mean than "anything goes." You can't just imagine whatever outlandish thing you like, call it "new science," and label the story you build on it "science fiction." In deciding whether an "innovation" story is legitimately science fiction, I use what I call the "negative impossibility" test. You don't have to be able to prove rigorously that your speculation is possible, as in the best of what mundane-SF advocates call mundane SF and I call pure-extrapolation SF. But you do need to imagine it in such a way that it is not provably impossible, or provably inconsistent with parts of established science that are well confirmed by experiment.

There's a precedent for this in real science itself. Physicists sometimes speak of a "correspondence principle," according to which the predictions of a theory created to explain a new region of experience must give the same results as the old theory in regions where the old theory matched experiment well. The classic example is mechanics. Sir Isaac Newton formulated laws to describe the motions of all the objects he'd seen, all of which moved much slower than light. Albert Einstein created a new set of equations because Newton's version weren't accurate for objects moving close to the speed of light. But, contrary to a popular misconception, this does not mean that Einstein "proved Newton wrong" or invalidated his equations. Rather, he created a model that worked through a wider range of experience—including the "Newtonian" world of slow-moving objects. The predictions of Einstein's equations are equivalent to those of Newton's at very low speeds, and diverge gradually as speeds increase. For slow objects, most of us still use Newtonian equations because they're easier to work with and the results are indistinguishable at those speeds.

Similarly, any new model you invent for a science fiction story—say, to make FTL believable—will still have to include relativity as a special case, and give the same results in the region of experience for which we have experimental verification of relativity. Such a model might well require a radical restructuring of theoretical foundations, but that has happened more than once before and will quite likely happen again.

Besides, one could easily get the impression from listening to the pontifications of some mundane-SF advocates that this whole business is deadly serious and somber: We should write only about the kind of future we're most likely to get because we must make sure everyone has the proper attitude toward learning to cope with it. Well, yeah, we need some of that—but let's not forget that science fiction is also about fun. Sometimes a story is worth telling not because it's terribly likely to come true, but just because some outrageous-but-not-provably-impossible idea would have such deliciously wild consequences if it did turn out to be true that the temptation to play with them is irresistible.

And there's nothing wrong with that.

In this type of discussion I'm often drawn back to something I once heard Poul Anderson say, when a bunch of us had gathered in Florida for the launch of Apollo 17. The perennial question of what distinguishes science fiction from fantasy came up, and Poul drawled, "You know, sometimes I think the most fantastic fantasy of all is what's usually thought of as the hardest hard science fiction. Because what could possibly be more fantastic than the idea that we already know all the basics?"

Or, in today's context, maybe it isn't really fantastic. It's just mundane.