In the data analysis class we've been discussing hypothesis testing in terms of epistemic discrimination between possible worlds. In other words: science is our principled way of figuring out whether we live on one imagined world (the null hypothesis) or another (the alternative hypothesis). For my own part, this has been an incredibly helpful model for learning the material.
Happy coincidence: I spent part of my typically-rainy commute down to the University of Oregon today listening to the "Imaginary Worlds" podcast. One episode, titled "The Canon," considered the difference between official and unofficial additions to fictional universes. The idea of fan fiction factored into the discussion.
Et voilà! Two great tastes that taste great together!
Among the items brought up in the podcast was Lucasfilm's recent decision to clear the "official" Star Wars canon's slate, which prompted outcry from some fans who objected to the diminution of the books, comics, and video games to which they were long attached. Rabbi Ben Newman, a guest on the podcast, likened this dispute to rabbinical debates over biblical canon, wherein Jewish authorities argue over what in addition to the Torah ought to count as "real" elements of the faith.
Then I got to a class wherein we practiced statistical t-tests, debating whether or not hypotheses about theropod tooth length were really true of the natural world or not. And so I wondered: isn't science a kind of fan fiction about nature?
I know that the "fiction" part is going to be controversial there; hell, I'm uncomfortable with it (earlier confessions notwithstanding). But it's important to note, as Peter Godfrey-Smith already has, that the models developed by scientific research are literally false; literal truth is too far a goal. Even our most direct measurements of natural systems have uncertainty built into them, the accommodation of which seems to be the whole point of data analysis. So we should acknowledge that scientific hypotheses are competing stories about what happens in the natural world.
So what separates a dispute about the real number of populations at a given fossil site from a debate over the real number of children borne by Princess Leia? Data, yes; but if we take data to be representative rather than exhaustive--and that's always the assumption!--then that's a difference of degree rather than a difference of kind. Scientific disputes are resolved by more and better evidence than disputes about fictional worlds, but they're both ultimately debate over which stories ought to count and why.