Related to: I'm Scared; Purchase utilons and fuzzies separately

Expanded from this comment.

You have awakened as a rationalist, discarded your false beliefs, and updated on new evidence. You understand the dangers of UFAI, you do not look away from death or justify it. You realize your own weakness, and the Vast space of possible failures.

And understanding all this, you feel bad about it. Very bad, in fact. You are afraid of the dangers of the future, and you are horrified by the huge amounts of suffering. You have shut up and calculated, and the calculation output that you should feel 3^^^3 times as bad as over a stubbed toe. And a stubbed toe can be pretty bad.

But this reaction of yours is not rational. You should consider the options of choosing not to feel bad about bad things happening, and choosing to feel good no matter what.

Your bad feelings, whether of fear, empathetic suffering, or something else, are probably counterproductive. Not only do you feel bad - a loss of utility in itself - but such feelings probably hurt, rather than help, your efforts to change the world for the better.

You may believe that your emotional outlook must be "rational": that it must correspond to your conscious estimates of the present or the future. Perhaps you expect to die of old age, or perhaps you are aware of people being tortured in secret prisons. You are forcing your emotions to match the future you foresee, and so you feel unhappy and afraid.

I suggest that you allow your emotions to become disconnected from your conscious long-term predictions. Stop trying to force yourself to be unhappy because you predict bad things. Say to yourself: I choose to be happy and unafraid no matter what I predict!

Emotions are not a a tool like rational thought, which you have to use in a way that corresponds to the real world. You can use them in any way you like. It's rational to feel happy about a bleak future, because feeling happy is a good thing and there is no point in feeling unhappy!

Being happy or not, afraid or not, does not have to be determined by your conscious outlook. The only things that force your mind to be unhappy are things like pain, hunger, loneliness, and the immediate expectation of these. If you accept that your goal is to be happy and unafraid as a fact independent of the future you foresee, you can find various techniques to achieve this. 

Unfortunately such techniques vary for different people. This post doesn't discuss any: it is about the prerequisite decision to be happy.

Expecting to die of cancer in fifty years does not, in itself, cause negative emotions like fear. Imagining the death in your mind, and dwelling on it, does cause fear. In the first place, avoid thinking about any future problem that you are not doing anything about. 

Use your natural defensive mechanisms, such as of not acknowledging unsolved problems, or compartmentalizing different beliefs. Don't dismiss them as biases or irrational practices. They exist for a good reason and have their proper use.

This does not mean that you should ignore problems on the conscious level. It is possible to decouple the two things, with practice. You can take long-term strategic actions (donate to SIAI, research immortality) without acutely fearing the result of failure by not imagining that result.

When you're faced with something terrible and you're not doing anything about it anyway, just look away. Defeat the implicit LW conditioning that tells you looking away from the suffering of others is wrong. It's wrong only if it affects your actions, not your emotions. 

New Comment
26 comments, sorted by Click to highlight new comments since: Today at 9:26 AM

This post's basic claims:

  1. If you 'should' feel bad about bad things (e.g. suffering), you 'should' feel bad about them in proportion to their magnitude.

  2. Doing 1 is not a good idea, because a. it's intrinsically bad to feel bad, and b. feeling bad does not help you to fight bad things.

  3. Therefore you should try not to feel bad about bad things, only to alleviate them.

All of which boils down to the proposition "It is instrumentally irrational to feel bad about bad things."

The problem is that 2.b. is blatantly false. Human beings are not capable of completely disconnecting action from emotion. Certainly, if you feel bad enough then due to the way the human brain works it's possible that you will (instrumentally irrationally) lapse into depression, and therefore do less to achieve your goals than if you never felt bad at all. And there are obviously cases in which what you feel emotionally can be overridden by calculation (e.g. trolley problem for many utilitarians). But given that no one performs explicit calculations to determine even most of their actions in more than a few small parts of their life, emotions determine most of our decisions. Do you really believe that someone who felt happy despite knowing about the state of suffering in the world would be more strongly motivated to reduce suffering than someone who felt a great sadness and a burning desire to stop it every time they thought about it? Do you think being happily mortal is the best emotional state for someone crusading to stop death?

If your actual goal is to end suffering -- if your moral system dictates, as most of ours do, that reducing suffering is currently by far the best thing you could do, and you actually want, unlike most of us, to follow your morals to their conclusions -- then you will do your absolute best to make your emotions about suffering dwarf all other emotions, because that is what will make you spend a life reducing suffering, and not any amount of abstract calculation.

On the other hand, if your real goal is to be as happy as you can, this post is great advice. But so is wireheading.

The problem is that 2.b. is blatantly false.

Nope. Felling bad about bad things only helps you to protest or punish people (yourself included) for the way things are. This is a very indirect form of "fighting bad things", and one which is largely ineffectual outside our evolutionary social environment. In the modern environment, protest and punishment are next to useless for accomplishing anything.

Do you really believe that someone who felt happy despite knowing about the state of suffering in the world would be more strongly motivated to reduce suffering than someone who felt a great sadness and a burning desire to stop it every time they thought about it?

You're begging the question here: what is the difference between "strongly motivated" and "burning desire"?

That is, you just said, "I think that there is more M in S+M than there is in X", where you haven't expanded X.

(Also, I'm not sure what kind of brain can actually experience "great sadness" and "burning desire" at the same time. Mixed emotions are usually not extreme emotions.)

It might be more useful to restate your question as "Will an otherwise-happy person have a greater probability of increased utility per their values, than one who is sad about the world's currently low utility?" My answer to that is an unqualified YES. Happy people make better choices about which actions to take, have greater motivation to act upon them, and better follow-through -- it is simply no contest.

if your moral system dictates, as most of ours do, that reducing suffering is currently by far the best thing you could do, and you actually want, unlike most of us, to follow your morals to their conclusions

Let us draw an important distinction between "morals" and "values". To me, a "moral" is a statement about what levels of a value we should protest or punish (e.g. by shaming/shunning), and it is of limited use outside the evolutionary environment (where others shared our morals and could be more influenced by our social maneuvering).

As Asimov put it, "never let your sense of morals keep you from doing what is right." The people who put a lot of moral weight (using my reduction of "moral") on the elimination of suffering seem much more motivated to protest, "raise awareness", "speak out", and perform other social signaling behaviors in place of direct action.

OTOH, I know lots of happy entrepreneurs who give or volunteer on behalf of various causes who, AFAICT, are not at all outraged or depressed by the suffering they witness. Without exception, the people I've met who actually DO things about the problems of the world (as opposed to merely talking about them) are happy people who are not disturbed by suffering in principle, even if they may be moved in relation to some particular individual's suffering.

So, in the context of this post, what is being said is that dropping one's moral rules allows one to pursue one's real values without impediment by our biased instincts to protest-and-punish -- which are in any case mostly ineffectual in the modern environment.

I wholeheartedly agree.

Your point is valid, but I think it's much weaker and less widely applicable than you claim.

  1. In many cases, people know definitely that they will not solve a certain problem; I think this applies to many people and problems, and is one of this post's main targets. Perhaps they definitely cannot affect it, like I can't stop the prison torture. Or perhaps they have consciously decided not to try to solve it, and invest their effort elsewhere. However, emotionally, being aware of the problem often causes suffering. This suffering is needless and it does not help to solve the problem, or any other problem that these people are trying to solve.

  2. In addition to e.g. sadness, there are other emotions. One example is fear, which led to this post. The person there was afraid of failing at important future tasks, or of bad things happening beyond their control. Empirically, more than a small amount of fear does not server to further motivate us to prevent such problems. On the contrary, it sometimes paralyzes and prevents action. We should therefore strive to minimize the amount of fear experienced to be at least no greater than needed to motivate us to act.

  3. Negative emotions can induce bias such as looking away from problems. In the absence of bad emotions, we can deal with problems more rationally.

  4. Once we have decided to solve a problem, the original emotion that made us decide this has become less necessary. We can consciously, rationally, sustain our fight against the problem even in the absence of the emotion. This is analogous to pain which continues after a wound has been treated, and is counterproductive. This is an argument for reducing the amount of negative emotion we allow ourselves to feel.

  5. Happiness isn't sadness with a negative sign. Reducing the amount of sadness isn't the same as increasing the amount of happiness. Both sadness (because there is a problem) and happiness (imagining the potential solution) can motivate us to solve a problem. If we increase positive motivation at the expense of negative motivation, this is a good thing.

Do you really believe that someone who felt happy despite knowing about the state of suffering in the world would be more strongly motivated to reduce suffering than someone who felt a great sadness and a burning emotional desire to stop it every time they thought about it?

Not necessarily - though it is possible (sadness can be paralyzing, depressing, and disruptive). But at the very least I do believe the happy person should not be significantly less motivated than the sad person - if they remember being sad, and preserve the (a-emotional) decision to solve the problem.

Do you think being happily mortal is the best emotional state for crusading to stop death?

I don't propose being happy about being mortal. I do propose being less sad about being mortal. Being "sadly mortal" in the sense of constantly being sad is demotivating and I believe counterproductive. Enjoying life is a mentally healthy state and is conductive to work, including work on immortality tech.

See, I feel the same way about your post as you do about my reply. I think that you're getting at something, but you've significantly overstated your argument -- to the point that someone following its recommendations would be less effective.

(There's something bizarrely wrong with the autoformatting which I can't reproduce and can't fix -- 1, 2, and 3 all address part 1, 4 is 2, and so on.)

  1. a. ("perhaps they definitely cannot affect it") This is true if you can do literally nothing about it, but the number of things on which you can have literally no effect are far outweighed by the number of things that holding this attitude will cause you to give up on. Do you think that you can have literally no effect on prison torture? That you could do nothing which would affect it, in any way? Try harder, Luke. Even death itself is not something you can give up feeling bad about -- nor was it a few centuries ago, because the societal effects of people trying to stop feeling bad about it then are damaging efforts to stop it now.

  2. b. ("perhaps they have consciously decided not to try to solve it, and invest their effort elsewhere") This, I think, significantly understates the effect someone can have without investing much effort. What if you just told your friends, and told them to tell their friends? Maybe one of the friends-of-friends would be struck by the cause, and decide that their effort would be best spent there. Probably this won't happen, but it certainly won't happen if you decide just to stop thinking about it because it's not happy-making to think about.

  3. c. ("This suffering is needless") This suffering is clearly not needless if it produces actions which can help resolve the problem.

  4. Life ain't easy. I don't see how trying to look away helps.

  5. I (falsifiably) expect that not feeling bad about problems has a much stronger influence.

  6. It's true that we keep acting in the absence of immediate strong emotion, but do you really believe that we'd act just as strongly to resolve a problem whether we still felt strongly about it or not?

  7. True. But the post doesn't argue for this, only for eliminating negative motivation.

I agree with you that sadness is intrinsically bad, and that constantly being sad is counterproductive. I disagree with you on the merits of negative motivation -- that is, I think that feeling sad about something causes you to work harder than not feeling sad about it -- and on the number of things that it's rational to "look away" from.

Also, one ambiguity that may be creating misunderstanding on both our parts: Ideally, no one should be generally unhappy, no matter what state the world is in. What they should be is unhappy about things, so that they can work harder to solve them -- and they should be unhappy about things in proportion to how bad they are. I think the post would be better if this distinction were clearly made.

Edit: And given that distinction, it is bad to ruminate on the existence of suffering/death/etc., if it's not motivating you to combat suffering/death/etc. So I agree with that part of the post.

("perhaps they definitely cannot affect it") This is true if you can do literally nothing about it, but the number of things on which you can have literally no effect are far outweighed by the number of things that holding this attitude will cause you to give up on. Do you think that you can have literally no effect on prison torture?

The effect isn't literally zero, in the same way that probabilities are never literally zero. But I believe the effect I can have on reducing prison torture is vanishingly small.

But that's besides the point. As long as the expected effect isn't worth the effort (not just time, but dangers and losses due to the effort), I will make precisely zero effort: I will instead work on a different problem. There are many problems, I can only work on a few, and at the very least I needn't feel unduly bad about the problems I'm not working on, and about the fact that I'm not working on them.

Even death itself is not something you can give up feeling bad about -- nor was it a few centuries ago, because the societal effects of people trying to stop feeling bad about it then are damaging efforts to stop it now.

Take the POV of someone living a few centuries ago, knowing that death might be solvable but only in the distant future. On the one hand, you have some measure of influence over future societies' fight against death - a very tiny measure, because you must multiply it by your uncertainty of what the future society will look like and how it might be influenced by its history. On the other hand, you have the certain knowledge that you and a billion of your contemporaries will feel much better if they e.g. don't fear death as much because they believe in an afterlife, which also means they'll believe that fighting death is a sin. It's obvious to me what the correct choice is.

c. ("This suffering is needless") This suffering is clearly not needless if it produces actions which can help resolve the problem.

As I said in the post: When you're faced with something terrible and you're not doing anything about it anyway, just look away. Defeat the implicit LW conditioning that tells you looking away from the suffering of others is wrong. It's wrong only if it affects your actions, not your emotions.

I believe that by far most suffering of this kind never affects actions, and so is unnecessary. And a rationalist should be able to correctly identify cases where it is or isn't.

Life ain't easy. I don't see how trying to look away helps.

When life ain't easy for someone else, looking away helps you.

It's true that we keep acting in the absence of immediate strong emotion, but do you really believe that we'd act just as strongly to resolve a problem whether we still felt strongly about it or not?

I believe at least 90% of strong empathetic emotions and of fears can be eliminated without directly, causally affecting any actions towards resolving problems less. There may be long-term effects of this mental posture of which I'm unaware, but I have no reason for believing such effects would reduce actions. At the very least, getting rid of negative emotions makes for a healthier and more free mental life, and feeling better in yourself is known to encourage positive actions.

I'm going to stop going point-for-point on this, and this will probably be my final post on the matter. But the gist of my argument is this:

You say that it's reasonable to "look away", to consciously try to disconnect your emotions from reality. This is essentially sacrificing emotional epistemic rationality for emotional instrumental rationality. In that sense, I consider it theoretically reasonable: epistemic rationality is ultimately only a sub-goal of instrumental rationality.

But unless you're a perfect rationalist, it is extremely dangerous to have a policy of favoring instrumental rationality over epistemic rationality. It's virtually impossible to lie to yourself in a way that is not contagious. Unless you have complete information about the universe and total knowledge of how to apply it, you can never be sure that the lie you told to cover up one unfortunate truth won't catch you somewhere else -- and when it's a lie you've told yourself, a false thing you've willed yourself into believing, you can't even keep the truth at the back of your mind to make sure you maintain correspondence with reality.

Yours is the logic of conversion, the argument that says you should abandon truth for religion if it seems likely to make you happier. Maybe this is the case -- but only if you can be sure that reality will never come back and bite you in the ass. Because once you've given up that instinct for truth, you can't get it back. A lie you tell to yourself is self-reinforcing and can't be isolated. Most likely you will never be able to dig it out.

If you were perfect, you could entirely disjoin the emotional state you wished to feel from the emotional valuation you wished to decide with -- making one conscious and keeping the other deep inside your head. But you're not perfect, you're human -- and humans can't do that. One who tries to do so will find that their real, underlying, motivating emotions change to match the ones they consciously desire to feel -- and in doing so alter their actions.

So your choice is this: either change your emotions to match reality, with all the suffering that entails; or ignore reality for the sake of your emotions, and sacrifice your moral code in doing so.

Sacrificing epistemology is not something you can do once you've awakened as a rationalist.

This is essentially sacrificing emotional epistemic rationality for emotional instrumental rationality.

One thing that you're overlooking here is that the kind of self-modification Dan is talking about can't be done unless you actually have strong epistemic rationality with respect to your emotions -- strong enough to understand the judgment by which you arrived at the emotions in the first place.

If you were perfect, you could entirely disjoin the emotional state you wished to feel from the emotional valuation you wished to decide with -- making one conscious and keeping the other deep inside your head.

This is a misunderstanding of how emotions work. Our emotions are not synonymous with our values, nor directly derived from them. If they were, we would all be rational, all the time!

Emotions are cached responses to situationally-salient values. Example: I don't like exercising, but it produces another result I want later. The not-liking-exercise emotion is not actually fulfilling my values: it would be more useful -- and more epistemically accurate -- for me to experience an emotion in relation to exercise that gives greater weight to my longer-term values. Which of these emotions is epistemically correct?

If our brains actually used our real values in their entirety to arrive at decisions, it'd take too bloody long. So we use cached evaluations based on immediate information... which means our emotions are automatically and systematically biased against our long-term best interests, unless we consciously correct what's in our caches on an ongoing basis.

So, there is no conflict here between the epistemic and instrumental: removing unnecessary negative emotion is simply correcting systemic biases of the underlying machinery to reflect our true values and desired outcomes, rather than overweighting what is easy to visualize or unconsciously learn.

Our emotions are not synonymous with our values, nor directly derived from them. If they were, we would all be rational, all the time!

You have misunderstood my entire point. I know that emotions don't naturally reflect values. The argument was over whether achieving your values requires you to change your emotions to reflect them, or if you can be equally motivated by values alone.

From the original post:

...you are horrified by the huge amounts of suffering. You have shut up and calculated, and the calculation output that you should feel 3^^^3 times as bad as over a stubbed toe. And a stubbed toe can be pretty bad.

In other words, you have decided that your emotions need to be realigned to reflect (what your value system says about) the state of the world. DanArmak argued that this is false. I argued that it is generally true.

In other words, you have decided that your emotions need to be realigned to reflect (what your value system says about) the state of the world. DanArmak argued that this is false. I argued that it is generally true.

Dan is in error, insofar as his argument implied that one should have one's emotions conflict with one's true values.

You, however are in error insofar as your arguments praise feeling bad as a path to doing good.

I agree with you that your emotions should reflect your values. OTOH, I agree with Dan that the optimal choice of emotion to reflect one's values will rarely be feeling bad, unless there is some sort of social goal involved (such as bonding with a group through a shared experience of grief or outrage).

Do you really believe that someone who felt happy despite knowing about the state of suffering in the world would be more strongly motivated to reduce suffering than someone who felt a great sadness and a burning desire to stop it every time they thought about it?

Sadness and burning desire are not in the same bucket. Sadness doesn't make me want to do anything at all. It quickly leads to depression, which is about the most unmotivated it's possible to feel. Certainly when I'm happy, I'm more motivated to do something about saddening things than when I'm actually sad.

Now, anger is an extremely motivating emotion that might work better for these purposes...

This seems to miss the point rather dramatically. Assume that there is suffering in the world that we can measure at -$200 trillion. We'll just assume that's the cost of alleviating that suffering. Pretty much no matter who you are, you can't alleviate more than a few million dollars of that suffering. Even if you're a complete outlier, like Bill Gates, you might be able to reduce world suffering by .1%.

If you feel bad in proportion to the amount of suffering in the world, given the inability of human feeling to make extremely nuanced distinctions (e.g. how much worse would it actually feel to lose $283.27 versus $283.07?), if you attempt to feel suffering in proportion to the amount that exists in the world, there actually isn't much of a point in combating suffering. It'd feel rather like dedicating your whole life to dropping a tablespoon of water into a swimming pool.

Strong emotions are also often conducive to irrational behaviour (and things such as depression).

Also, everyone has a utility function that is heavily self-centered. Exceptions may exist, but it is an extreme minority of people who sincerely act as if they do not count for substantially more than other individuals in their utility function. I'd argue this is a good thing, but that aside, it is reality. Given that, not feeling miserable throughout life is a rather important end. This is even more true if you cannot demonstrate a clear connection between feeling miserable about the world in general and striving to make the world a better place. I have to admit I do not know many people I would describe as successful activists for any cause who are also genuinely unhappy in life. General unhappiness seems to be a major inhibitor of success in any of one's pursuits.

See note on ambiguity at the end of my second post.

Also, one ambiguity that may be creating misunderstanding on both our parts: Ideally, no one should be generally unhappy, no matter what state the world is in. What they should be is unhappy about things, so that they can work harder to solve them -- and they should be unhappy about things in proportion to how bad they are.

Ideally, no one should be generally unhappy, no matter what state the world is in. What they should be is unhappy about things, so that they can work harder to solve them.

This is great -- if you can, this insight should be turned into another top-level post.

This is definitely true for me. I started doing more to improve the world after hearing a speech that made me feel horribly sad and guilty about not doing so.

There is a metaphor I've heard, relevant to this problem.

Say a deer was granted the magical superpower to detect predators (and recognize them as such) from hundreds of miles away, with no other changes to it's psychology. This seems at first blush as though it would be a tremendous advantage: no predator's senses can compare, so the deer should be able to evade pursuit with trivial ease, grow fat, mate frequently, etc. Unfortunately, it's not very clever; the state of constantly being apparently surrounded by predators, some of which are at any given time (by sheer chance) moving directly towards it, leaves the deer in a constant state of panic. In a few hours it collapses from blind exhaustion and is eventually eaten by something or other.

Humans are smarter than deer, and have been known to look away from the news from time to time, so we remember to breed and generally make it through from one day to the next. That doesn't mean we're actually benefiting from this particular capability, on an individual level.

you do not look away from death or justify it.

If this is an inherent part of your definition of "rationalist", then I've come to the wrong site.

Most people learn to look away from death or justify it as necessary, good or just. This clearly involves various biases, logical fallacies, and compartmentalization. So when people who think this way become rational, they usually discard these thoughts. It's a well known irrationality, as well as being important to discard (since this is necessary for anyone to fight death).

For these reasons, I gave it as a typical example. Of course it's not a necessary example, it's not an inherent part of being "a rationalist" (or just "being rational").

In any case, the point of my post is that it is rational to e.g. look away from death, justify it, and use other mental techniques to feel less bad about it - as long as that doesn't appreciably lessen your effectiveness (if any) at fighting death.

What if I don't want to fight death? What if I accepted the fact that I'll die of old age, and decided, after evaluating the arguments pro and against, that it's the best outcome for me?

In addition to the other replies, I'd like to add this. Most likely you're not very old right now, and dying "of old age" isn't an immediate prospect. So, saying "I want to die of old age in the future" is emotionally very different from saying "I want to die of old age (or something else) very soon".

Far vs. near mode comes into play, which means you'll tend say things about the future mostly for signalling purposes. Of course you're saying these things honestly, you really feel & believe them. But on the outside-view, there's a significant (though far from certain) probability that, when you really are old and dying is an immediate prospect, your preferences will change and you'd prefer a cure.

What if I accepted the fact that I'll die of old age, and decided, after evaluating the arguments pro and against, that it's the best outcome for me?

Then thats how it is.

However, when someone reevaluates a long-held position, even one which was developed and maintained for manifestly irrational reasons, they are still extremely unlikely to change their mind. If someone claims to have evaluated such a position thoroughly, it is generally a good bet that they stopped thinking once they found a good enough justification of their previous belief, before there was any real risk of changing their mind. This is a central topic in the sequences.

Of course this is a bet, not a certainty. I don't mean to be harsh, but you should seriously consider the possibility that your argument is a rationalization. orthonormal's comment is probably more helpful to this end.

What about the preference reversal test? If you weren't constituted so as to age and die, how much would somebody have to give you in order for you to agree to self-modify to age and die?

If you genuinely (and not just as a devil's advocate) would place a very low price on switching in that scenario, then your values are indeed consistent and our criticism on this point might be misguided. If you find yourself reluctant in that hypothetical, though (and most people are), then you have an inconsistency, and you should consider the possibility that you're rationalizing in your thoughts about death.

To paraphrase a comment I posted a few days ago on a different topic: "Death is an artifact of evolution, why keep it that way?"

This is a good modifier for Feeling Rational, and a good expansion of this xkcd comic. Thanks!

Oh huh, and it also pairs well with my later Choosing the Zero Point post!

I think the comic is better matched by The Scourge of Perverse-Mindedness.

Defeat the implicit LW conditioning that tells you looking away from the suffering of others is wrong.

Bravo. I'm not sure if I agree with everything you say in the post but this quote is a good one.