Policy-Based vs Willpower-Based Intentions

post by Unreal · 2019-02-28T05:17:55.302Z · LW · GW · 14 comments

Contents

  Policy-based intention
  Willpower-based intention
  My resistance to willpower-based intention
  Equal and opposite advice
  Conclusion
None
14 comments

Been thinking about what it means to set an intention lately. I think I’ve found a distinction between policy-based intentions and willpower-based intentions. 

Policy-based intention

Policy-based intention-setting is a lot like writing a computer script and running it. 
For example, I have a policy around tipping Lyft drivers. It is made up of a bunch of if-then statements. 

It might not be the perfect policy for every situation, but it’s better for me to spend the processing power once, rather than every time. 

It basically costs no willpower to implement the policy. I’m not having to nudge myself, “Now remember I decided I’d do X in these situations.” I’m not having to consciously hold the intention in my mind. It’s more like I changed the underlying code—the old, default behavior—and now it just runs the new script automatically. 

I call it an intention because it is me manifesting a change in my behavior using a decision point. I created a branch in my history, and I chose left instead of right. And now my future self is going to choose left instead of right in a bunch of future branches. 

Willpower-based intention

This seems more like what is classically meant by intention.

Willpower-based intention involves an active, conscious, mindful holding in the mind. To me, it viscerally feels like my brain is gripping an object inside my head. If I grip too hard, I can get a headache. I can also hold it lightly / gently (such as during mindfulness meditation). 

The holding doesn’t always have to be continuous. It can work more like “reminders” where I find myself naturally inclined to do X, and then I remind myself I intended to do Y instead. 

Here’s a few examples:

This last example looks like it would be better if it were a policy-based intention. Something you could just rewrite in the underlying code using if-then statements. 

In my experience, it doesn’t seem to always work. Considering pronouns is still often a thing that requires a bit of conscious intention-holding for me. In my internal monologue, I mess up people’s pronouns all the time, but I’m pretty good at saying the correct ones out loud. 

I suspect these things differ by individual.

My resistance to willpower-based intention

A weird thing about me: I have an unusually high resistance to using willpower-based intentions in many situations. It feels like death, like being trapped under a boulder, like suffocation. When I feel forced to use willpower, I become depressed and sometimes suicidal. Sometimes I experience visceral terror and panic and must-escape-nowNowNOW. 

It doesn’t always feel this way, but it can. And I can experience it in micro-doses, for things as small as lifting a heavy object or sitting still for a while or making myself smile when I don’t feel like it. 

And I’m coming to grips with this being a real problem.

It’s interesting, though, to think that I’ve managed to do a lot of things anyway. I was missing a major capacity and able to cope and get by regardless. I “pass”.

In a way, I trust myself a whole lot—because I know that even without willpower-based intentions, I still get up and DO things. I handle most things I need to handle. I’m not just a worm wriggling around in the mud. (And technology has been essential—integrating well with my technology is crucial for me to maintain my systems and my flow.)

But there are other things I can’t do that others can: Maintain a consistent habit everyday. Make commitments / promises. Stay focused on something that’s hard for me to focus on. Finish big projects where my interest wanes (like writing a book). Make certain personal sacrifices. Stay in a job that gets boring or aversive. Endure physical discomfort. Make this technique [LW · GW] for spamming micro-intentions work on things I feel resistance towards. 

As a general strategy, I’ve had to live a life where I can’t really let others rely on me, in a durable way. I cannot offer to be the ground others stand on. I would crumble. And I know I don’t want that, so I don’t try to play that role. I don’t put myself in those positions. I’ve had to learn all this about myself. 

And I want to learn how to use willpower freely, one day. 

Not to make myself do things because I should. But because there are genuinely things I want that are outside my reach right now, without willpower-based intentions. Because I want to be a good person who can make difficult, but right choices. Because my aliefs around willpower are coming from a damaged past—a past that contains truth but not the whole truth. 

Equal and opposite advice

My sense of people is that they more often have the opposite problem. They overuse willpower-based intentions. They think that if they lapse even a little, they’d let important things slip or their structures would collapse. 

These people can be unreliable too. 

They carry more and more with their willpower, but they’re at risk of one day suddenly collapsing—having reached their limit. Or maybe they do collapse—at night or on weekends—and oscillate between being an ox carrying a heavy load and a useless lump who can’t do anything but watch TV. 

For them, I’d offer totally different advice. Which is to get more in touch with their internal states, desires, emotions, felt senses [LW · GW]. To use Internal Double Crux [LW · GW] or Internal Family Systems or Focusing to open more channels with their elephant (in the elephant / rider [LW · GW] sense). To experiment with days or weeks where there are no obligations at all. [EA(p) · GW(p)] To test just how many of their intentions need to be held [LW · GW], and letting some of them go or loosening them a bit. 

Conclusion

Policy-based intentions don’t require much conscious maintenance and can be used to create general if-then-based plans for your behavior in a variety of situations. They’re super convenient! 

Willpower-based intentions do require conscious “holding” of an intention in your mind for some period of time. But they don’t have to be based in shame, guilt, or obligations necessarily. They can be in alignment with your deeper goals, and being able to use willpower-based intentions in this way is important for accomplishing certain goals. It’s a super powerful ability!

Concentration meditation seems to train the skill directly, since it’s about holding the intention of returning your attention to an object for a specified amount of time. 

For more on the phenomenology of intentions by mr-hire, see here [LW · GW].

14 comments

Comments sorted by top scores.

comment by Kaj_Sotala · 2019-02-28T20:41:51.499Z · LW(p) · GW(p)

Could you go into more detail on how exactly one sets these policy-based intentions? I don't think I know how, nor did I ever even suspect that such a possibility exists.

It does slightly remind me of the KonMari thing where you go through all of your belongings to get a felt sense of where in the apartment they belong, and then automatically put things in their right places because any other place feels wrong.

I've used that for a few things like "where should I put my keys", and phenomenologically it does feel like it just kinda... rewrites my brain to make that feel like the only right place to put them, and then it requires no willpower after that. But that procedure feels more like discovering the right place rather than consciously deciding it; if there's a more general thing that I can do in this category, I'd be curious about hearing more.

Replies from: Unreal
comment by Unreal · 2019-02-28T21:05:01.648Z · LW(p) · GW(p)

That's interesting!

How do other people handle the tipping thing? Whether for a driver or at a restaurant? Are you kind of deciding each time?

How do you handle the question of "who pays for a meal" with acquaintances / new people / on dates? My policy in this area is to always offer to split.

How do you handle whether to give money to homeless people or if someone is trying to offer you something on the street? My policy here is to always say no.

I'm curious what other people are doing here because I assumed most people use policies to handle these things.

Replies from: Kaj_Sotala
comment by Kaj_Sotala · 2019-02-28T21:52:03.143Z · LW(p) · GW(p)

We don't really do tipping or paying-for-the-other here, so those don't come up. I do have a general policy of not giving money, but the way by which I "installed" it doesn't resemble your description: it was something like, I ran into people asking for money on several occasions, then on each occasion made a decision about what to do, until I got sufficiently frustrated with feeling manipulated after giving money that the emotion drove me into making a decision which I'd remember in the future.

I guess the thing that sounded the most unusual to me was this part:

It basically costs no willpower to implement the policy. I’m not having to nudge myself, “Now remember I decided I’d do X in these situations.” I’m not having to consciously hold the intention in my mind. It’s more like I changed the underlying code—the old, default behavior—and now it just runs the new script automatically. 

It's not that I don't have policies, it's that this description sounds like you can just... decide to change a policy, and then have that happen automatically. And sure, sometimes it's a small enough thing that I can do this. But like with the key thing, if I suddenly decided to put my keys somewhere else than I used to, I bet that I'd keep putting them in the wrong place until I gradually learned the new policy.

Or maybe your description just makes the concept sound more exotic than you intended and I'm misinterpreting. It sounded to me like you have some magic method of automatically instantiating policy changes that I would need to spend willpower on, without you needing to spend willpower on them. E.g. the Lyft thing sounded complicated to memorize and I would probably need to consciously think about it on several times when I was actually doing the tipping before I had it committed into memory. But you said that you can't use this on everything, so maybe the policies that I would need willpower to install just happen to be different from the policies that you would need willpower to install.

I guess one thing that confuses me is that you say that it costs no willpower to implement a policy, and then you contrast it with willpower-based intentions, which do cost willpower. That makes it sound like costing willpower, is a property of how you've made the decision? But to me it feels like costing willpower is more of a property of the choice. In the other comment you gave the example of having a policy of how you arrange your kitchen, and this sounds to me like the kind of thing that everyone would have as a policy-that-cost-no-willpower, because you can just decide it and there's no need to use willpower on it. Whereas (to again use your example) something like drinking water every day is more complicated, so you can't just make it into a policy without using willpower.

So my model is that everyone already does everything as a policy-based intention, unless there's some property of the act which forces them to use willpower, in which case they can't do it as a policy and have to use some willpower instead. But your post sounded like people have an actual choice in which way around they do it, as opposed to having it decided for them by the nature of the task?

Replies from: Unreal
comment by Unreal · 2019-02-28T22:10:31.135Z · LW(p) · GW(p)

Oh yeah. I do think the nature of the task is an important factor. It's not like you can willy-nilly choose policy-based or willpower-based. I did not mean to present it as though you had a choice between them.

I was more describing that there are (at least) two different ways to create intentions, and these are two that I've noticed.

But you said that you can't use this on everything, so maybe the policies that I would need willpower to install just happen to be different from the policies that you would need willpower to install.

This seems likely true.

It's not that I don't have policies, it's that this description sounds like you can just... decide to change a policy, and then have that happen automatically.

It is true that I can immediately change certain policies such that I don't need to practice the new way. I just install the new way, and it works. But I can't install large complex policies all in one go. I will explain.

the Lyft thing sounded complicated to memorize and I would probably need to consciously think about it on several times when I was actually doing the tipping before I had it committed into memory.

With zero experience of Lyft tipping, I would not just be able to think up a policy and then implement it. Policy-driven intentions are collaborations between my S1 and S2, so S2 can't be doing all the work alone. But maybe after a few Lyft rides, I notice confusion about how much to tip. Then maybe I think about that for a while or do some reading. Eventually I notice I need a policy because deciding each time is tiring or effortful.

I notice I feel fine tipping a bit each time when I have a programming job. I feel I can afford it, and I feel better about it. So I create and install a policy to tip $1 each time and run with that; I make room for exceptions when I feel like it.

Later, I stop having a programming job, and now I feel bad about spending that money. So I create a new if-then clause. If I have good income, I will tip $1. If not, I will tip $0. That code gets rewritten.

Later, I notice my policy is inadequate for handling situations where I have heavy luggage (because I find myself in a situation where I'm not tipping people who help me with my bag, and it bothers me a little). I rewrite the code again to add a clause about adding $1 when that happens.

Policy re-writes are motivated by S1 emotions telling me they want something different. They knock on the door of S2. S2 is like, I can help with that! S2 suggests a policy. S1 is relieved and installs it. The change is immediate.

Replies from: Kaj_Sotala
comment by Kaj_Sotala · 2019-03-01T08:19:25.178Z · LW(p) · GW(p)

Thanks, that clarifies it considerably. That tipping example does sound like the kind of a process that I might also go through.

comment by Qiaochu_Yuan · 2019-02-28T20:27:34.381Z · LW(p) · GW(p)

There's an interesting thing authentic relating people do at workshops that they call "setting intentions," and I think it works in a different way than either of these. The difference seems to me to be that the intention is being "held by the group." I'm not sure how to explain what I mean by this. There are at least two visible signs of it:

1) people remind each other about the intention, and

2) people reward each other for following through on the intention.

If everyone knows the intention is being held by the group in this way, it both comes to mind more easily and feels more socially acceptable to follow through on (the latter might be causing the former). In my experience group intentions also require almost no willpower, but they also don't feel quite like policies to me (that would be "agreements") - they're more like especially salient affordances.

The key ritual is that at some point someone asks "so, do we all agree to hold this intention?" and we raise our hands if so - and we look around the room so we can see each other's hands. That way the collective holding of the intention can enter common knowledge.

Said another way, it's something like trying to define the values of the group-mind we're attempting to come together to form.

---

I relate to your resistance to willpower-based intentions. It's something like, a lot of people have an "inner authoritarian" or "inner tyrant" that is the internalized avatar of other people making them do stuff when they were younger (parents, teachers, etc.), whose job it is to suppress the parts of them that are unacceptable according to outer tyrants. You can live under your inner tyrant's reign of terror, which works as long as submitting to the inner tyrant keeps your life running smoothly, e.g. it placates your outer tyrants and they feed you and stuff.

At some point this strategy can stop working, and then other parts of you might engage in an internal rebellion against your inner tyrant; I think I was in a state like this for most of 2017 and 2018, and probably still am now to some extent. At this stage using willpower can feel like giving in to the inner tyrant.

Then I think there's some further stage of development that involves developing "internal leadership," whatever that is.

There's a bit in the Guru Papers about this. One quote, I think in the context of submitting to renunciate morality (e.g. Christian morality):

Maintaining denial actually requires constant surveillance of the thing you are pretending isn’t there. This deepens the internal splits that renunciation promises to heal. It requires the construction of a covert inner authoritarian to keep control over the “bad” stuff you reject. This inner tyrant is probably not strong enough to do the job on its own, so you submit to an external authority whose job is to strengthen the internal tyrant.
Replies from: Unreal, Unreal
comment by Unreal · 2019-02-28T20:58:41.571Z · LW(p) · GW(p)

I have not much considered group intention-setting. This seems super interesting to explore too.

Phenomenologically, I feel it kind of as... the agreements or intentions of the group (in a circle) recede into the background, to form the water we're all in together. Like it gets to relax in the VERY BACK of my mind and also I'm aware of it being in the back of other people's minds.

And from that shared container / background, I "get to move around" but it's like I am STARTING with a particular set of assumptions.

Other potential related examples:

  • I'm at a Magic tournament. I know basically what to expect—what people's goals are, what people's behaviors will be, what the rules of the game are and how to enforce them. It's very easy for me to move here because a lot of the assumptions are set in place for me.
  • I'm in church as a kid. Similar to the above. But maybe less agreeable to me or more opaque to me. I get this weird SENSE that there are ways I'm supposed to behave, but I'm not totally sure what they are. I'm just trying to do what everyone else seems to be doing... This is not super comfortable. If I act out of line, a grownup scolds me, is one way I know where the lines are.

Potential examples of group policy-based intentions:

  • I have a friend I regularly get meals with. We agree to take turns paying for each other, explicitly.
  • I have a friend, and our implicit policy is to tell each other as soon as something big happens in our lives.

As soon as a third person is added to the dynamic, I think it gets trickier to ensure it's a policy-based intention. (Technology might provide many exceptions?) As soon as one person feels a need to remind themselves of the thing, it stops being a policy-based intention.

Willpower-based intentions in groups feel they contain a bunch of things like rules, social norms, etc.

comment by Unreal · 2019-02-28T20:42:29.336Z · LW(p) · GW(p)

There is definitely this sense that exerting force or willpower feels like an EXTERNAL pressure even if that pressure does not have an external source that I could point to or even gesture at. But it /feels/ external or 'not me'.

I have some trauma related to this. I could've gone into the trauma stuff more, but I think it would have made the post less accessible and also more confusing, rather than less. So I didn't. :P

comment by Matt Goldenberg (mr-hire) · 2019-02-28T15:55:02.864Z · LW(p) · GW(p)

I'm curious about the phenomenology of your Policy-based intentions... do they feel similar to habits you haven't chosen consciously? Is it simply the same thing as CFAR calls a TAP, and most everyone else calls a habit, or is it something different?

P.S. I wrote something about the phenomenology of "willpower based" intentions here:

https://www.lesswrong.com/posts/QmcFeZtwSRhcsYDZu/video-the-phenomenology-of-intentions [LW · GW]

Replies from: Unreal
comment by Unreal · 2019-02-28T17:25:33.953Z · LW(p) · GW(p)

I enjoyed that article! Seems worth including the link in my article too. Thanks.

Your definition of intention seems different from my use of "willpower-based intention." My 'willpower-based intention' always has a conscious element and cannot do things like "work in the background without my awareness at all." It's maybe quite related to the thing in your forehead.

My policy-based intentions feel kind of like pulling up my inner code guts, making a little rewrite or alteration, and putting them back into my guts. This is a conscious process (the installation), but then the change runs automatically, without holding conscious intentions.

I'm very bad at using these to create personal habits, like drinking water everyday or taking vitamins everyday. I don't think these count. They require willpower after a while.

But maybe I one-time decide the best configuration of spices on the spice rack or how my kitchen is arranged. Then it is automatic for me to place things back where they belong after using it, and it is also automatic for me to want to organize things so they're back where they belong when they get messed up.

These 'desires' for things to be a certain way live in my belly. And it feels like my belly carries motivations and behaviors that I can ride out.

It feels relaxing to have a policy I can lean on, and to carry out the policy. Like water running downhill.

You could maybe think of it as 'intentions you already want to do anyway'. But with policies, your conscious mind can also make alterations / rewrite that code directly. Without any need for convincing, arguing, pushing. It is more of a collaboration I am in between elephant / rider—coming up with good policies makes us feel good and relaxed.

comment by habryka (habryka4) · 2019-02-28T19:53:12.737Z · LW(p) · GW(p)

Thanks a lot for writing this!

This post was marked as "don't promote to frontpage", but that feature is still a bit janky, so I wanted to check whether you do actually want us to promote this post to the frontpage, since in terms of content it would definitely fit the guidelines.

Replies from: Unreal
comment by Unreal · 2019-02-28T20:38:18.188Z · LW(p) · GW(p)

oh. I must have messed that up. I am OK with this being on the front page. I have definitely noticed some bugs here and there. Esp around the account settings page and trying to change my moderation guidelines. But I think I maybe just messed up the checkbox. Is it default checked to 'not ok'? Because if so, I left it alone thinking it was checked to 'is ok to promote'.

Replies from: habryka4
comment by habryka (habryka4) · 2019-02-28T20:42:59.909Z · LW(p) · GW(p)

Oh, I would be interested in any bugs you've had with the moderation guidelines, and sorry for those. Would be great if you would ping us on Intercom or create a meta thread about those.

We definitely messed up with the frontpage checkbox, and it's almost certainly not your fault. It had some weird behavior where it showed itself as checked, but if you didn't uncheck and check it again, it would actually show up in the DB as "don't promote to frontpage". This should be fixed in the next few hours.

comment by romeostevensit · 2019-03-02T02:33:11.354Z · LW(p) · GW(p)

This feels like how I think about TDT as a casual heuristic.