Hope, Not Optimism

Fatalism has no theory of change

I’ve been an activist all my life — literally, I attended my first demonstrations in a stroller — and that’s reflected in my work, from the essays and blog posts I’ve published for 20 years to the dozens of books I’ve written, both fiction and nonfiction.

To be an activist is to want to change the world. To change the world, you need two things: first, an understanding of what’s wrong with it, and second, a theory of how to make it better.

Much of my work focuses on the former: documenting, analyzing, and tracking injustices, dysfunctions, and emergencies — my essays are a form of public note-taking that helps me break down and understand complex phenomena.

I’m keenly aware that a steady diet of my nonfiction is a little bleak. In my novels, the characters generally chart a course from a bad world to a better one — that’s how fiction works, after all. “Plots” are smooth, ascending gradients in which people try to solve problems of increasing complexity and urgency until they reach a climax — a moment where the stakes cannot be raised any higher — and either succeed (comedy) or fail (tragedy).

Plotting versus activism

That’s not how the real world works. The real world is messy in a way that novels are not because the real world has so many confounding factors, contradictory impulses, and unknowable hidden variables that we cannot “solve” it the way we might solve a plotting problem in a story.

It’s a fool’s errand to even try. Mapping out the terrain — all the things we might do to make things better, all the ways they can go wrong, and what we should do in response — is an exercise that would take so long that merely outlining the process is a waste of time. By the time that outline was done, the terrain would have been changed so drastically that you’d need to start over.

Computer science to the rescue

If you’ve ever studied computer science, this might sound familiar. There are a lot of problems we try to solve computationally that we can’t ever fully optimize: complex datasets that are so large that fully analyzing them is a nonstarter. By the time the program finished running, the data would…