I don’t think akrasia exists.
This is a fairly strong claim. I’m also not going to try and argue it.
What I’m really here to argue are the two weaker claims that:
a) Akrasia is often treated as a “thing” by people in the rationality community, and this can lead to problems, even though akrasia a sorta-coherent concept.
b) If we want to move forward and solve the problems that fall under the akrasia-umbrella, it’s better to taboo the term akrasia altogether and instead employ a more reductionist approach that favors specificity
But that’s a lot less catchy, and I think we can 80/20 it with the statement that “akrasia doesn’t exist”, hence the title and the opening sentence.
First off, I do think that akrasia is a term that resonates with a lot of people. When I’ve described this concept to friends (n = 3), they’ve all had varying degrees of reactions along the lines of “Aha! This term perfectly encapsulates something I feel!” On LW, it seems to have garnered acceptance as a concept, evidenced by the posts / wiki on it.
It does seem, then, that this concept of “want-want vs want” or “being unable to do what you ‘want’ to do” seems to point at a phenomenologically real group of things in the world.
However, I think that this is actually bad.
Once people learn the term akrasia and what it represents, they can now pattern-match it to their own associated experiences. I think that, once you’ve reified akrasia, i.e. turned it into a “thing” inside your ontology, problems occur:
First off, treating akrasia as a real thing gives it additional weight and power over you:
Once you start to notice the patterns, it’s harder to see things again as mere apparent chaos. In the case of akrasia, I think this means that people may try less hard because they suddenly realize they’re in the grip of this terrible monster called akrasia.
I think this sort of worldview ends up reinforcing some unhelpful attitudes towards solving the problems akrasia represents. As an example, here are two paraphrased things I’ve overheard about akrasia which I think illustrate this. (Happy to remove these if you would prefer not to be mentioned.)
“Akrasia has mutant healing powers…Thus you can’t fight it, you can only keep switching tactics for a time until they stop working…”
“I have massive akrasia…so if you could just give me some more high-powered tools to defeat it, that’d be great…”
Both of these quotes seem to have taken the akrasia hypothesis a little too far. As I’ll later argue, “akrasia” seems to be dealt with better when you see the problem as a collection of more isolated disparate failures of different parts of your ability to get things done, rather than as an umbrella term.
I think that the current akrasia framing actually makes the problem more intractable.
I see potential failure modes where people come into the community, hear about akrasia (and all the related scary stories of how hard it is to defeat), and end up using it as an excuse (perhaps not an explicit belief, but as an alief) that impacts their ability to do work.
This was certainly the case for me, where improved introspection and metacognition on certain patterns in my mental behaviors actually removed a lot of my willpower which had served me well in the past. I may be getting slightly tangential here, but my point is that giving people models, useful as they might be for things like classification, may not always be net-positive.
Having new things in your ontology can harm you.
So just giving people some of these patterns and saying, “Hey, all these pieces represent a Thing called akrasia that’s hard to defeat,” doesn’t seem like the best idea.
How can we make the akrasia problem more tractable, then?
I claimed earlier that akrasia does seem to be a real thing, as it seems to be relatable to many people. I think this may actually because akrasia maps onto too many things. It’s an umbrella term for lots of different problems in motivation and efficacy that could be quite disparate problems. The typical akrasia framing lumps problems like temporal discounting with motivation problems like internal disagreements or ugh fields, and more.
Those are all very different problems with very different-looking solutions!
In the above quotes about akrasia, I think that they’re an example of having mixed up the class with its members. Instead of treating akrasia as an abstraction that unifies a class of self-imposed problems that share the property of acting as obstacles towards our goals, we treat it as a problem onto itself.
Saying you want to “solve akrasia” makes about as much sense as directly asking for ways to “solve cognitive bias”. Clearly, cognitive biases are merely a class for a wide range of errors our brains make in our thinking. The exercises you’d go through to solve overconfidence look very different than the ones you might use to solve scope neglect, for example.
Under this framing, I think we can be less surprised when there is no direct solution to fighting akrasia—because there isn’t one.
I think the solution here is to be specific about the problem you are currently facing. It’s easy to just say you “have akrasia” and feel the smooth comfort of a catch-all term that doesn’t provide much in the way of insight. It’s another thing to go deep into your ugly problem and actually, honestly say what the problem is.
The important thing here is to identify which subset of the huge akrasia-umbrella your individual problem falls under and try to solve that specific thing instead of throwing generalized “anti-akrasia” weapons at it.
Is your problem one of remembering to do tasks? Then set up a Getting Things Done system.
Is your problem one of hyperbolic discounting, of favoring short-term gains? Then figure out a way to recalibrate the way you weigh outcomes. Maybe look into precommitting to certain courses of action.
Is your problem one of insufficient motivation to pursue things in the first place? Then look into why you care in the first place. If it turns out you really don’t care, then don’t worry about it. Else, find ways to source more motivation.
The basic (and obvious) technique I propose, then, looks like:
Identify the akratic thing.
Figure out what’s happening when this thing happens. Break it down into moving parts and how you’re reacting to the situation.
Think of ways to solve those individual parts.
Try solving them. See what happens
Potential questions to be asking yourself throughout this process:
What is causing your problem? (EX: Do you have the desire but just aren’t remembering? Are you lacking motivation?)
How does this akratic problem feel? (EX: What parts of yourself is your current approach doing a good job of satisfying? Which parts are not being satisfied?)
Is this really a problem? (EX: Do you actually want to do better? How realistic would it be to see the improvements you’re expecting? How much better do you think could be doing?)
Here’s an example of a reductionist approach I did:
“I suffer from akrasia.
More specifically, though, I suffer from a problem where I end up not actually having planned things out in advance. This leads me to do things like browse the internet without having a concrete plan of what I’d like to do next. In some ways, this feels good because I actually like having the novelty of a little unpredictability in life.
However, at the end of the day when I’m looking back at what I’ve done, I have a lot of regret over having not taken key opportunities to actually act on my goals. So it looks like I do care (or meta-care) about the things I do everyday, but, in the moment, it can be hard to remember.”
Now that I’ve far more clearly laid out the problem above, it seems easier to see that the problem I need to deal with is a combination of:
Reminding myself the stuff I would like to do (maybe via a schedule or to-do list).
Finding a way to shift my in-the-moment preferences a little more towards the things I’ve laid out (perhaps with a break that allows for some meditation).
I think that once you apply a reductionist viewpoint and specifically say exactly what it is that is causing your problems, the problem is already half-solved. (Having well-specified problems seems to be half the battle.)
Remember, there is no akrasia! There are only problems that have yet to be unpacked and solved!
You used the word umbrella and if I was going with a slightly less catchy, but more accurate summary, I would write, "Akrasia is an umbrella term". I think the word is still useful, but only if you remember this. The first step in solving an Akrasia problem is to notice that a problem falls within the Akrasia umbrella, the second step is to then figure out where it falls within that umbrella.
This is accurate. Thanks for a revised summary!
OK, I don't want to sound too harsh, and I'm really happy that you are writing and trying to do stuff etc. You can only be criticized because you wrote something, rather than nothing at all. This is laudable.
So. In the spirit of constructive criticism, I want to try to express why I can't stand this post. Some of my criticism involves subtleties that might be hard to get across here, but we can continue discussing this on Skype etc.
The post claims to criticize a simplistic approach to understanding akrasia, but what it offers as an alternative is just another simplistic approach, which is to my eyes not much different and stuck in all the same harmful frames ("us vs them", "solve", "just aren't remembering" etc.)
The naive intuition that "akrasia" points to some living, coherent thing with its own goals, which knows stuff "you" don't and is sometimes actively undermining "you", is actually a pretty good description of what is really happening. The OP is really trying to make it not be the case, but it seems driven more by wishful thinking than looking at reality. Maybe my intuition is wrong here but it tells me that you are doing some mental motion that puts you in a sorta fake manic optimism state to even be able to write a post like this. Some of that might be a good thing, but if you use too much it starts to sound very unconvincing to me.
The thing akrasia really points to is something like advanced model-based reinforcement learning on goals that you are not consciously unaware of, that is directly in control of your actions. (It is much more than the "conserve energy" theory mentioned in other comments, though "conserve energy" is indeed included as a significant goal.) The models used in it learn and predict the results of you following some plan or strategy used to reduce akrasia, including clever meta-plans like "decompose the problem and apply my best tools to solve each part". This is precisely the reason why most "solutions" only last for a pretty short time, and your post offers nothing to change this state of affairs beyond "here's another piece of meat to throw into the jaws of the monster, have fun with the additional 15 seconds that it earns you".
I don't know either you or OP, and am not in either of your heads. However, from this and other things OP wrote, I suspect that this model has worked for them. From what you're saying in point 2, this model would not work for you and fails to accurately describe what's going on in your head.
Lifelonglearner, I suspect you are making a broader claim than you can support, but that the narrower claim inside it is accurate. I agree with you that giving people new models may sometimes be net-negative. I agree with you that tabooing and/or breaking down akrasia is probably a useful tool to get at a symptom and figure out a way to deal with that symptom. I intend to give it a shot, and I'll make sure to note what kind of results I get- if you like, I'd be pretty happy to put those notes somewhere you can get at them if you want datapoints to refine this.
I think akrasia is a useful model when in non-taboo form. When I read "instead of treating akrasia as an abstraction that unifies a class of self-imposed problems that share the property of acting as obstacles towards our goals, we treat it as a problem onto itself" my first thought was that description made it sounded like depression. Depression isn't exactly self-imposed, but it can look an awful lot like it is. It's worth noting that attacking depression as an entity of itself can be far more effective than trying to break it down into symptoms.
(I'm not suggesting that akrisia === depression! I am suggesting that there exist abstractions that unify a set of symptoms that are best attacked as the abstraction they are.)
I'm sorry, but I'm not talking about what is useful in someone's head. I'm referring to the best current ideas of how this really plays out in the brain. For some background, try https://sideways-view.com/2017/02/19/the-monkey-and-the-machine-a-dual-process-theory/
I'll take "what is useful in someone's head" over some harebrained ideas about how it really plays out any day.
Praxis is the criterion of truth :-P
I personally prefer the aesthetic of figuring out how stuff works as a priority. And one could argue that this also gives greater practical benefits in the long run... but sometimes the "long run" is pretty long indeed.
To me these are somewhat orthogonal issues. I like cute/pretty/elegant and I like practical/functional/effective, but I'm willing to get them from different things. Of course, if something manages to combine the two, that's great.
Hey, thanks for the thoughts!
I think there may be confusions about the things that we're pointing to and it seems useful to try and clear those up first.
First off, by akrasia I'm pointing at situations where you look back and think, "Drat! In that instant, I could have done this other thing that would have better served my goals."
(It may actually be the case that you were satisfying some other hidden part of yourself during that time. I'm not implying that instances of additional counterfactual productivity always indicate realistic places to improve, merely that they appear to be.)
Or something like that. Happy to try and zero in on a better working definition if you'd like to propose something different. I think the things we have in mind of examples of akrasia may not be the same.
Secondly, I want to stress that I don't think that forcing / a brute force approach towards productivity is not optimal. Often, the sort of internal debugging of an Internal Double Crux and Focusing flavor seems like the right way to debug internal aversions as well as move forward.
The sort of "dropping your obligations" and sourcing internal motivation that Nate Soares writes about in his Replacing Guilt series is, I think, very important.
I also think, though, that this model sometimes isn't at the right level of granularity to deal with smaller problems like forgetting to get things done or getting sucked into distraction-loops. For this, I think the above algorithm can be useful.
Models of akrasia that personify it or model it as a malicious agent don't feel tractable or intuitive to me. Given the statement above, I want to once again stress that I don't think this is a solution to all problems related to getting things done. I do think that the reductionist framing and considering the potential dangers of reification are important, though. Hope that tempers any impressions you have of optimism on my part.
So here's what happened until now:
I read the OP and I thought I understood really well what you were trying to say. I wrote some criticism.
You read the criticism and thought that I failed to understand you. You answered by pointing to this apparent misunderstanding and reiterating your view from the OP in a slightly different way.
I read your answer and I thought that your answer meant that you had understood around 0% of my original comment. I am now backtracking to point out how much my criticism was not about what you thought it was about.
When not specified, I assume standard definitions like "a lack of self-control or the state of acting against one's better judgment" (from Wikipedia), and your “want-want vs want” or “being unable to do what you ‘want’ to do” seems to capture it also. I don't think that this was a problem in our communication. Your clarifications above are exactly what I'd have expected after reading just the OP.
I feel like you're answering by pulling from your database of "smart things to say about akrasia", instead of actually reading what I'm trying to say. This is frustrating. The database obviously includes CFAR stuff, like "Double Crux > willpower" and "reductionism + use tools to solve problems", and Nate's "Replacing Guilt". But I know those sources too, I value them and I'm not arguing about these points at all.
This is some strawman that you are pulling from the naive approach, and not at all what I was saying.
"living, coherent thing" -> this doesn't mean "person"
"which knows stuff "you" don't and is sometimes actively undermining "you"" -> this doesn't mean "malicious"
And once again, you are rebutting a strawman of what I said. "model-based reinforcement learning on goals that you are not consciously unaware of" is something that can be usefully treated in a reductionist way, and not a reification but the best short description of what's most likely out there in reality that I could think of.
Uh, I'm sorry, I'm tired now, we can continue trying to understand each other's points tomorrow. So far, I am losing hope
Sounds good, I'll try to follow-up more later. You're right that I wasn't engaging with your points. There were a few things I thought I explained poorly in the original post and wanted to make sure those weren't the points of disagreement.
This conversation didn't move forward, which I think is unfortunate.
I think SquirrelInHell agrees with something like the view described on sinceriously.fyi, especially false faces. If you read that, you may see where the objection was coming from.
I don't think such a model of Akrasia is productive. I don't think such a model will help you overcome your inability to do that which you desire. Insomuch as such a model of Akrasia does not help you achieve your goals, it is not rational;
I half agree and half disagree. I agree that it is good idea to look at different problems and determine what solutions apply to them individually.
But I disagree that there is no general summary of "akrasia." I would summarize it like this:
Your general planning ability does not take sufficiently into account energy cost, that is, energy cost in the ancestral environment. Consequently you tend on a regular basis to make plans that involve high energy expenditure and little benefit, that is, plans that would be bad plans in the ancestral environment. You have various protection mechanisms built in to prevent you from carrying out these plans, since they would waste energy and therefore be dangerous. And these protection mechanisms are themselves many and various, and so lead to what you are calling various problems. But they are all for the same purpose: to stop you from wasting energy when your life is perfectly fine the way it is.
That seems good.
If we continue the analogy with cognitive biases, I think we can make a similar statement along the lines of quick heuristics that were good back in the past when you lived in a less-globalized village, original need for split-second decisions, etc. etc.
I think the point still stands, though, that we should try to move away from the idea of generalized anti-akrasia weapons and look more into the exact specifics of the problem from an instrumental perspective.
I often ask myself the question of "is this really a thing" when it comes to high level concepts like this. I'm very unsure on Akrasia, and you make a decent enough argument. It could very well not actually be a thing (beyond referring to sub-things).
More importantly, though, even if it were a thing, I agree that the strategy you suggest of focusing on the smaller issues is likely the better one.
Agree. Would have said it shorter. Akrasia is an applause light. Taboo the meaning then fix it. I have helped people through Akrasia problems. It always starts with me asking for a more concrete and specific example of the problem.
See also: depression, procrastination, "I'm in pain", "relationship problems ", time management, bad memory.
All things for which the one word summary hides the nature of the situation and makes it harder to diagnose.
Or perhaps a fake explanation or a mysterious answer.
"Why am I not doing X?"
"Because you suffer from the not-doing-the-thing syndrome, which people educated in Latin call 'akrasia'."
"Thank you, Captain Obvious, you solved my problem!"
"Actually, I am a Captain Knows-One-Latin-Word, but you are welcome!"
Because this is LW I feel obligated to point out that it's Greek, not Latin :-P
One coaching course I took had as one of the core principles a 90/10 rule: 90% questions of what the coach says is supposed to be questions.
This really spoke to me. I agree with the overall thrust of the article.
I agree with overcoming the curiosity stopper very strongly. I think the generalized form of this is an extremely important primitive action type to have. The decomposition hammer is like brainstorming, it seems to get stronger the more you habitually use it.
I disagree with the 'problem solving frame' which I think is the key sticking point. Lots of branches of psychotherapy and schools of buddhism seem to have independently reinvented the idea that you are not going to get anywhere on your problem until you acknowledge the ways it is helping you get your needs met.
Yep, I think you're right.
I tried to address that with this point here, but I think it could be emphasized stronger:
I agree with the first sentence. "Akrasia" is just the brain being [quite reasonably] unsure whether its long-term goals are really worth the effort.