I’m working backwards through my podcast queue, having for a couple reasons developed quite a backlog, so I just got to Ezra Klein’s discussion with Yascha Mounk about whether American democracy is in decline. The discussion is worth listening to, but Klein makes an argument that I’ve heard often and that just seems wrong. The argument is basically that if you’re looking at this period in American history and calling it out as singularly bad (because Trump, say), you’re ignoring the vast swathes of American history that have been much worse. Even going back to the 60s, says Klein, we had political leaders (JFK, MLK, Malcolm X, Medgar Evers) being assassinated, we had riots in the streets, and we had Nixon elected in part on a promise to fight lawlessness.

Now, what do you do with this argument? One approach is to ask: if American democracy is in decline, what is that decline relative to? If it’s relative to the 1960s, then maybe we can argue that it’s not in decline, because it’s never been particularly strong. Maybe, in fact, it’s stronger than it was back then.

Sort of a corollary to that argument is: times have been bad before, and we survived those. We’re supposed to infer from this that we’ll survive this particular era of badness.

This sort of reasoning always rubs me the wrong way, and I hear variants on it a lot. George W. Bush was bad, say, and there were those of us (I was certainly among them) who believed for a time that he and Cheney wouldn’t allow themselves to be removed from office. In retrospect that was a paranoid view. Now here we are under the Trump administration, which certainly feels like an era when the U.S. could turn authoritarian. “But you thought the same thing in the Bush Administration,” the argument goes, “and you were wrong then. Doesn’t that make you think you’re wrong now?”

Well, maybe. If nothing else, maybe it should make me question my authoritarian-dar. But there’s no historical law that says “When the level of authoritarian badness is below some threshold x, we have nothing to fear.” And there’s no historical law that says “If our political leaders aren’t being assassinated, U.S. government will survive.”

Maybe you can assemble some notion of historical law, by which when x, y, and z happen, then outcome O is really likely. I’m not sure I reject the possibility of historical law altogether—maybe I do, maybe I don’t—but I certainly have strong doubts that we finite humans know what those laws are. And normally when people advance the “it was worse in the past, and we got through it” argument, as Klein did, they’re not even reaching for a historical law. They have no causal reason to believe that circumstances x, y, and z always (or even often) lead to outcome O. They only have correlation to point to: x, y, and z happened, and as it happens our government also survived it. That’s a very slender reed upon which to support the notion that we’ll get through this era unscathed.

As time has gone on, I’ve come to expect less “use” from history. History is what the mathematicians call an “existence proof”. For instance: we lived through an era when African-Americans fighting for basic access to American democracy were being murdered with impunity in the South, when political leaders were being assassinated with frightening regularity, and when it felt like the political consensus that had bound the United States for decades was coming apart at the seams—and we survived it. Conclusion: it’s possible to survive those things. It’s not guaranteed that we’ll survive those things. The fact that we survived those things doesn’t even make our survival in the future more likely. It just makes our survival possible.

All that said: if you need to treat history as a machine that deterministically churns toward your desired outcome, because that’s what you need psychologically to make it through the day, by all means use that. For me, though, the uses we put history to often feel like facts about our own brains rather than facts about the way the world works.