combustion dots

Kinetic vs Thermodynamic Favorability of Actions

I’ve found chemical kinetics vs thermodynamics to be a very rich metaphor for thinking about actions and habits (where a given habit/action is like a chemical reaction I’d like to run), and it seems to help explain a lot of the common pieces of advice I give people about habit formation and taking actions (especially a lot of the content of aversion factoring).  First, a (rather lengthy) summary of the relevant chemistry:

In chemistry we like to talk about some reactions being “favorable” because they tend to happen spontaneously, like a rock rolling downhill; such reactions are called “exothermic” because they release energy. Other reactions are endothermic, and require an input of energy in order to happen. The energy ends up stored as “potential energy” in the bonds between atoms. Again, you can imagine needing to push the rock uphill, and then it having potential energy based on its position.

exo endo thermic.PNG

However! There are lots of reactions that are exothermic, but which nonetheless almost never occur. For example, the combustion of paper is a highly exothermic reaction (think fire), but you almost never get spontaneous combustion of paper. Even though the reaction releases a lot of energy, you still have to put a bunch of energy into the paper before it will occur (such as by lighting a match underneath it).


What’s going on here? It turns out there’s another way that reactions can be unfavorable: what’s called kinetics, as opposed to thermodynamics. When we describe a reaction as endo-/exo-thermic, we’re talking about its thermodynamic favorability, i.e. whether the reaction is overall energy-requiring or energy-releasing. When we talk about kinetic favorability, we’re ignoring that factor, and instead thinking about how fast the reaction will happen, which is a function of its activation energy:

activation energy

The hump in the middle of the graph shows up in all chemical reactions, because between the starting and ending configurations you’ll always have the atoms going through a transition state that is higher-energy than either one. How much higher depends on the physical circumstances you’re working in, and determines the activation energy of that specific reaction pathway. And the activation energy, in turn, determines how fast a reaction will occur, under that specific set of circumstances. If a reaction pathway’s activation energy is high enough (if the transition state hump is very large), then the reaction will happen very, very slowly, up to “basically never.”

To switch back to the rock metaphor: it’s as if you always have to push the rock over a physical bump to get it anywhere. Say the rock starts out rolling a certain speed. Thermodynamics measures whether the rock will be rolling slower or faster than it started out once it travels from point A to B. Kinetics measures how fast the rock needs to start out rolling in order to make it over the hump to point B at all.

hump left


This is actually a bit more apt a metaphor than it might seem, since what you actually have is molecules bouncing around and colliding with each other. To get over the hump & have the reaction occur – assuming it’s thermodynamically favored – you temporarily need an input of energy anyway, to get over the hump. Easiest way to get that is from the kinetic energy of the molecules as they bang into each other. The molecules are all bouncing around at different speeds, in some normal-ish distribution. The average speed of the molecules = the temperature. The activation energy can be thought of as a measure of the threshold for how fast the molecules need to be going, when they bang into each other, in order to provide enough energy to get the reaction to happen. Only molecules going faster than that will actually react. So if almost all the molecules are going that fast, then almost every collision results in a reaction, and the whole thing goes quickly. But if most of the molecules are going slower than that, then only a small fraction of collisions result in a reaction, which means you have to wait a long time in order to get all of the molecules to react. Since speed = temperature, heating up the material is a good general-purpose way to get the reaction to happen faster.

It turns out this is how the body regulates almost all of its metabolic functions (which are all just series of chemical reactions). Floating around in any one of your cells are the necessary components for thousands of different exothermic reactions. However, you only want a tiny fraction of those reactions to be occurring at any given moment. So you make sure that the default conditions in the cell are very kinetically unfavorable for the reaction. And then, when the moment is right, you change that equilibrium, usually either by making one of the reactants much more plentiful, or by adding a catalyst. A catalyst is any molecule that reduces the activation energy of a reaction, without itself getting used up; typically catalysts are bound with one or more of the reactants during the transition state, which allows that transition state to be lower energy than it would be otherwise (an enzyme is just a catalyst that’s made of protein, which is the most common thing you see in biological systems). Because the thermodynamic favorability of a reaction is purely a function of the starting and ending chemical products (path independent), it can’t be impacted by catalysts, or by the amount of reactants available, which makes it a less useful way to regulate reactions that you want to occur sometimes, but not always (which is a lot of what metabolism is).

But catalysts *can* cut off the top of the activation energy hump. And this can be the difference between the reaction *never* occuring (like paper spontaneously combusting) and it occurring very easily (like cellular respiration, which is the same overall chemical reaction, broken down into a bunch of smaller steps and mediated by a bunch of enzymes)

The overall chemical equation for both combustion & cellular respiration.  Cellular respiration is way more complicated, though


So that’s the metaphor.  Here are some of the ways I’ve found it useful:

CFAR has struggled to make sense of why it sometimes seems like it’s not enough to have your S1 believe that an action is good on net; you for some reason need there to be no significant downsides at all. I think that this model makes sense of that puzzle: when we think about whether an action is good “on net,” we’re often thinking about something that’s most closely analogous to thermodynamics. This is what we’re trying to shift if we consider changing the activity (like dancing or parkour instead of jogging), or giving ourselves a reward for doing it. And our beliefs about the net-goodness/thermodynamic favorability of an action, I think, are what we’re trying to work with during propagating urges.

However, I claim that sometimes the problem is more similar to having a prohibitively-high activation energy on an action that is, nonetheless, thermodynamically favorable.

prohibitive activation energy words

One thing you’d expect to see if this metaphor holds is that you could take the action sometimes: when all external factors line up just right, or sometimes when you add a bunch of stress/time pressure/shame. In the former situation, it’s like only a small fraction of the reactant being moving fast enough, so you get the reaction to happen in only a small fraction of the collisions; like me being able to write blog posts on my very best days – when I’ve gotten enough sleep, and I have a good idea, and there aren’t any urgent-unimportant emails sitting in my inbox – but not otherwise. The latter situation seems analogous to adding heat/energy to the system, like holding a lighter to the piece of paper (although of course knowing you’ll feel bad when you sit down to start your project would also increase the activation energy you need). Without these circumstances, you tend to “bounce off” the task or problem (just like an insufficiently-energetic pair of molecules “bouncing off” of each other…).

In these situations, rather than trying to shift the thermodynamics, it’ll often turn out to be easier if you solve the problem by changing the kinetics (adding an awesome pump-up song to the alarm that tells me it’s time to exercise, buying different running shoes so my feet don’t hurt). Or otherwise changing the environment in which you’re trying to get the reaction to run. This helps make sense of why it’s good to do a mindful walkthrough [LINK] of the activity to look for small irritant, since since that’s exactly the sort of thing that can contribute to a prohibitive activation energy in spite of the activity seeming clearly like a good idea on net. In particular, it’s useful to pay special attention to the details of starting an activity or transitioning into it, since this is where you’re most likely to find activation energies.

It also seems like it could help explain why it’s useful to think very specifically about triggers for actions, since usually you want to do actions sometimes – at the right time – but not all the time (think checking facebook, or answering email).

And it also explains the limitation of the cheap-hacks method of problemsolving: if that reaction is endothermic, no amount of chipping away at the activation energy is gonna make it run easily.

Leave a Reply

Your email address will not be published. Required fields are marked *