First it's important to realize that there are various kinds of science fiction today, in which science functions differently. Michael Crichton builds a thriller around technical details, even tables of data; character and "art" are less emphasized. Ursula Le Guin writes anthropological science fiction, emphasizing the social sciences and subtleties of character. A recent trend is the "future historical" novel such as Maureen McHugh's China Mountain Zhang, in which scientific extrapolation provides details of a vivid future setting for everyday people. My own work explores the interactions between science and society, and the human beings caught between them--even when, as in A Door into Ocean, we are not sure at first who is "human."
As a writer, you need to decide what role (if any) science extrapolation can play in your work. In fact, much of what is labeled "science fiction" today could as easily be labeled fantasy; and if your own style is distinctive enough, that may be the route for you. On the other hand, to take science seriously requires special attention. I can suggest some approaches which work for me.
The freshest ideas come straight from experience in an actual scientific laboratory. In my own lab and those of my colleagues, I regularly experience natural phenomena stranger than the strangest of science fiction: a superconducting magnet that suspends paper clips in the room next door; a dish of bacteria that generate thousands of mutations overnight; a flask of chemicals that "magically" turns color every few seconds. As a research scientist, and a teacher needing to range widely, I have an advantage. But any writer can telephone a research lab and even request a visit; most scientists love to talk about their work. Internet bulletin boards are another good source of expertise.
Next to the lab itself, the best source of ideas is research journals such as Science and Nature. These sources provide primary research reports of the latest discoveries, those of interest to a wide range of scientists. While the reading is a challenge, even for a veteran scientist, most of the exciting finds reported here will never reach the popular science magazines. For example, I came across a report in Science of a bacterium that actually eats uranium. This fit right into the plot of my novel, which required an organism to eat something no other creature would touch!
For readable reviews of emerging fields, use periodicals aimed at the scientifically literate readership such as Scientific American and Sigma Xi's American Scientist. Be wary of newspapers and the less sophisticated popular science magazines, whose accounts are likely to be superficial and contain errors.
Once you have a good idea, it's worth checking out with experts, just as you might check out any other aspect of setting. Thus you can avoid obvious bloopers, as well as ideas considered total cliches by experts who would otherwise be sympathetic to your work. For example, physicists told me that an anti-gravity device would be written off as a cliche, but the use of a white hole as an energy source might be taken seriously.
In the end, you can take heart from the fact that "mistakes" may not be fatal, as far as popular success is concerned. Frank Herbert's bestseller Dune showed settlers on a desert planet distilling water from the air. This would work in an Earth desert only because Earth's atmosphere carries water from the oceans. Even if your science is "right" when the book is written, some aspects are bound to get outdated soon enough. A Door into Ocean depicted women who generate children by fusion of ova. Even before the proofs came out, research had shown this to be impossible because paternal chromosomes carry essential modifications.
Credibility and consistency. What makes an idea "credible," then, is hard to define. Getting the facts exactly "right" and up-to-date is helpful; yet if none of your assumptions or extrapolations could be challenged, your work would not be science fiction.
Interestingly, the more common complaint I hear from inexperienced writers is that the "real science" they have carefully researched is declared false or unbelievable by readers or editors. What do we do when truth is stranger than fiction?
One way to make your ideas credible is to tie each invention to some easily verifiable event or fact on Earth. This can be done more or less subtly as a sort of in-text footnote. When Crichton shows his dinosaurs chomping through steel bars, "like hyenas," he offers a fact that I could verify. We can be sure that some hyena enthusiast out there will complain loudly if he gets it wrong! Similarly, when I created an alien organism with infrared vision in The Wall around Eden, I noted that known animals such as rattlesnakes possess infrared sensor organs. The focussing lens of the alien "eyes" was of sodium chloride, an infrared-focussing substance that living creatures commonly contain in their bodies.
Another source of credibility is consistency: Make sure that your facts and extrapolations, however reasonable on their own, make sense together in the story. If your imagined planet has twice the mass of earth, what is its gravity? The composition of its atmosphere? How close is it to its sun, and how long does it take to complete a year? Do the native animals on such a planet have thick, ponderous limbs, or delicate long ones? If voracious monsters descend upon your space visitors, what fauna do they normally prey upon?
The biological questions are frequently overlooked. In Door into Ocean, I created an entire ecosystem complete with microbial plants to photosynthesize, small phosphorescent grazers, both aerial and marine predators of a range of sizes, and scavengers, "legfish" that crawl up upon floating vegetation.
It may seem exhausting and frustrating to get all the parts to work together, but this extra craft is what distinguishes stories like Dune from more forgettable attempts. In my own work, I have come to rely upon a layered approach, in which I start at the beginning, write in a chapter or two until inconsistencies build up, then start all over from the beginning and try to get a couple of chapters farther. Inevitably the first chapter gets rewritten twenty times; but the reward is that my last one virtually writes itself.
A writer who develops a particularly complex world view, or "universe," may choose to write several books within the same universe, exploring different aspects of its setting or theme. Just as Doris Lessing wrote a series of novels about Martha Quest in Africa, Ursula Le Guin wrote several books including Left Hand of Darkness within one imagined universe, where humanity's far-flung colonial worlds are linked by the "ansible" communication device. One must however take care to come up with enough fresh material to justify each new story in its own right.
Explaining your ideas in the story. The biggest mistake is to lecture your readers, however intriguing an idea may be. The writer must blend science ideas seamlessly amongst all other aspects of experience that form the story. As always, "show, not tell" is the rule.
Try to let science ideas lead into character development, and vice versa. An example of this process occurred as I wrote Door into Ocean, in which a population of women called Sharers inhabit a planet covered entirely by ocean. One day a researcher in my laboratory excitedly showed me a flask of purple protein he had just isolated from photosynthetic bacteria. When light shined upon the protein, it bleached clear, as it absorbed the light energy. This demonstration gave me the idea that my aquatic women characters would carry purple bacteria as symbionts in their skin, providing extra oxygen underwater. When their oxygen ran low, the Sharers' skin would bleach white dramatically. This ability to "bleach white" later developed a spiritual significance as well; the Sharers can enter a special kind of trance, called "whitetrance," which enables them to endure extreme physical stress while upholding their religious beliefs.
Another example from Door into Ocean works in the opposite direction, of character development leading to science. The Sharers use Gandhian pacifist resistance to repel an armed invasion of their planet. I sought a metaphor from science to help describe the unexpected success of their resistance, which from the invaders' limited perspective seemed doomed to fail. The metaphor had to fit into the perspective of the Sharers, who have advanced biological technology. I hit upon the idea of "electron tunneling," a phenomenon in which electrons can penetrate a seemingly impenetrable energy barrier. Electron tunneling occurs in the hemoglobin molecule as it collects oxygen in the blood, so the Sharers would know about it.
Some explanation is always necessary; the trick is, how much. It helps to weave necessary explication into dialogue, a sentence at a time, at a point where events demand it. For example, in Daughter of Elysium, a visiting scientist (new to the planet) discovers that his discarded culture dishes have come alive and are trying to gobble up his two year-old son. A student comes to the rescue, and explains that the "intelligent" culture-dish material (composed of billions of microscopic robots) has malfunctioned; it is designed to enclose tissue cultures, not children.
This example by the way also illustrates the time-honored gimmick for explaining any new setting: the naive "visitor," who needs everything explained. It works, if you don't make the lecture too obvious, and do keep the plot moving. Jurassic Park essentially consists of a long lecture on cloning dinosaurs, kept moving by a fast-paced, and blood-thirsty, series of events.
One approach to the problem of explanation is to include all that the story seems to need in the first draft, even though you know it's too much for the reader to take. In later drafts, cut it drastically. Omit terms known only to experts, or redefine in simple language. (Oogenesis is "making eggs.") A typical science course introduces more new words than a first year of language. So try to use scientific terminology as you would use words from a foreign language -- sparingly, for effect.
An occasional phrase of jargon may be worth keeping if it takes on a life of its own in the story. In Daughter of Elysium, I did keep one phrase of fetal development about the "primordial germ cells" which undergo a lengthy migration to reach the developing gonads before the fetus is born. The phrase set up a distinctive metaphor for the life journey of my central characters. But countless similar phrases were cut or redefined before my final draft.
Advancing the plot. Complex technical information is best fed to the reader a little at a time, and in such a way that it feels "inevitable" where it comes up. This task is a challenge, but if done skillfully the development of ideas can advance your plot, heightening dramatic tension, much more so than if you had revealed all the implications at the start.
Daughter of Elysium depicts research connecting fetal development and aging, a field of daunting complexity. My opening chapter shows how the fetal heart tube forms and begins to pulse; later chapters depict more subtle processes of cells and tissues, and much later the critical molecular events that determine whether the embryo will live or die--or live without aging. In between, various subplots incidental to research take up the scientist's time, much as they would in real life. Often the subplots make an ironic contrast to his work; for instance,when he faces his dying relatives back home, who will never benefit from his research on aging.
Another role for science in your plot can be to show how various characters react to change, and are themselves changed (or not). In A Door into Ocean, the invaders of the ocean world respond to the Sharers' life science in diverse ways. Some simply try to destroy it, and none of the bizarre setbacks they face change their outlook. Others become intrigued by the new science, with its implications for their own medicine and agriculture. A few even take up the symbiotic purple microbes into their own skin.
The points I've made about finding ideas and using them have served me well in my own novels, and have worked for other writers too. At the same time, it is important not to get lost in the science. Remember that what makes a science fiction novel "work" in the long run is what makes any good novel work: connection, consistency, and characters that make us care.