I want to look at a category of weird mental tricks that we sometimes play on ourselves—you might be familiar with the individual examples, but when considered together they reveal a pattern that I think deserves more attention. I’m going to do the Scott Alexander thing and list a bunch examples in hopes that you’ll sense the common concept they all point at.

Action derivatives index

Here they are:

• Meta-preferences: This is when you say you want something, but your actions don’t reflect that wanting, so really we would say you’re wanting to want that thing. If you say, “I want to get stronger,” but you never go to the gym, and it’s totally possible for you to go to the gym, we would say you have a revealed preference for the “not getting stronger” outcome, but your meta preference is to be the kind of person who truly wants to get stronger, i.e. who goes to the gym.
  • Notice that if you just do the thing, voluntarily, that already implies that you wanted to do it.
• Belief in belief – This is a term coined by Daniel Dennett, and it’s described concisely in The Sequences. Sometimes when people say “I believe X,” they don’t actually hold any expectations about X, but they believe it’s somehow advantageous or virtuous to believe X. We would say they believe in belief of X.
  • Notice, if you just say, “X is the case,” you’re already implying that you believe it. I say “I live in NY,” and unless you think I’m lying, it implies that I believe that I live in NY.
• Trying to try – This is also in The Sequences. When you say, “I’ll try to do X,” that’s a similar kind of extra step.
  • Notice, if you just say, “I’m going to do X,” that implies that you’re going to try. After all, you’re never a certain predictor of the future: the most you can do is try.
  • It follows, then, that if you say “I’ll try do X,” you’re implying that you’ll try to try. That’s a very relevant difference for personal productivity, agency, and effectiveness, because if you say “I’ll try,” when really you’re just trying to try, then basically you’ve “succeeded” as soon as you can tell yourself that you tried. And that might be far removed from the real, raw kind of trying that you do when you say, “I’m gonna do X.”
• There’s a fourth one that fits with these other three. I’ll state it, but I don’t want to elaborate on it because it kind of touches on culture war topics in Current Year, and I don’t want this to be a culture war post. It’s Identifying as identity. I say, “I identify as an X.” Notice, if I just say “I am an X,” that already implies that I identify myself as an X.

So we have wanting to want, belief in belief, and trying to try. You can see how these all feel like different forms of the same thing. For now I’m calling them action derivatives, because they’re derived from direct actions: wanting and trying are derived from doing, and believing is derived from claiming.

Caveats

Now I called them mental tricks, but all of the action derivatives have at least some uses where they don’t seem to be doing anything tricky.

• You could say “I want to do X” as a way of expressing, “I would do X if the present circumstances made it possible for me (but they haven’t yet).”
• You could say “I believe X” as a way of expressing, “X is the case if my premises are correct (but I’m not super confident that they are).” And this tracks with an informal study that frames the word “believe” as a simple descriptor of a claim’s confidence (with a measured value of 60-80%).
• You could say “I’ll try to do X” as a way of expressing, “I’ll do X if it’s sufficiently easy (but I don’t yet know if it is).”

In those cases, you use the action derivative to express some kind of uncertainty about the action itself. And that’s perfectly straightforward communication, if it’s interpreted as intended.

But in other contexts where these moves are made, there can be a sense that some bait-and-switch is being done. Someone is not going to the gym, but “wants to.” Someone is “trying to” get into grad school, but hasn’t sent applications. Someone “believes all women,” but not that one.

Analysis

What does it mean when we use the action derivatives to pull those bait-and-switch moves in conversation, or in our own internal monologues?

My first guess: I think it’s about reaping the narrative effects of an action without doing the action itself. I could be the kind of person who wants all the right things and tries to do all the right things and believes all the right things—I can feel like that person now, without having to change anything about the physical world around me. It’s a positive narrative I can tell to other people, or just to myself.

And I suspect this is a failure mode of living too much in social reality, the world of narratives and vibes, and not enough in the inflexible realm of physical reality, where you either do or don’t do that thing; you either are or are not that person. Human action in the physical world has derivative effects in the social world, but if the social world is too “real” to us, we lose track of that relationship and start grabbing at those narrative effects “for free.” It’s really easy to make cognitive mistakes when they reward inaction.

That’s just my guess. I’d love to hear other thoughts/discussion on this.

-pdf