How Much Science Do You Need To Know To Write Science Fiction?

Farthing and Tooth And Claw author Jo Walton is widely regarded as one of the best writers of fantasy right now, and she won the John W. Campbell award for the best new writer of speculative fiction. So why does she feel she can't write science fiction? Because, she explains on her journal, she knows too much science to write utter nonsense, and not enough science to get SF stories absolutely right. It makes me wonder if science fiction is scaring away some of its best potential writers.

On her blog, Walton says that doing all the research to make her science unimpeachable slows down her writing process to a crawl, to the point where she loses interest in the story. And her friends who know science end up suggesting alternatives that screw up what she wanted to do in the first place. She explains:

So I have this thing about aliens with four genders. It takes place in the universe where the solution to the Fermi Paradox is that FTL drives make your star explode after 20 uses. So these aliens are stuck in their solar system (with a couple of other aliens who showed up and can't go home) and they know about other aliens. (Earth may or may not exist in this universe. It doesn't matter. This is a story about some aliens.) My aliens have a mother planet and a terraformed marslike, and a moon where they live in domes. My character comes from the terraformed planet. He's leaving a spaceship on the mother planet, he smells the mother planet air, and he thinks "Ah, the sweet smell of /INSERT ATMOSPHERE COMPONENT GAS HERE/, which we don't have in the air of my terraformed home, which smells so atavistically good because this is where my ancestors evolved, but which nevertheless reminds me of the three years I spent here in the prison camp." And I stop, and I trot off to ask what atmospheric component gas it could be (and already you notice I have stopped writing and started checking, and also, note how much I had to explain to get to this point, which in the actual story would all not be explained) and after a long discussion I find out that there's nothing, unless I totally change everything I want, or give them noses that can smell argon or something (which is an unnecessary complication when they already have turtle shells and four eyes and the interesting thing is the four genders) and I have to scrap that sentence which was doing set up and incluing and background and was about to set up the next sentence about how he met his best friend in the prison camp and was going to lead on into some actual story.

If I didn't know any science at all, I'd just merrily put traces of chlorine in an oxygen atmosphere and it would all be as dumb as heck but at least it would actually get written and the characters would get out of my head and get to have their adventure.

And this is just one line, and it's all like that.

So anyway, that's why I don't write SF, even though it's what I like to read.

I wonder how many would-be science fiction authors get turned off by these sorts of concerns. And how many of them would have written thought-provoking classics of the genre. (And how many people who do write tons of science fiction novels bother to know their science half as well as Walton already does.) The comment thread over at Walton's post is also well worth reading. [Paper Sky]