Epistemic status: influenced by my study of emotions over the years, but primarily based on personal observation. Presented in the spirit of "here's a model, judge for yourself whether you think it's any good."

If you’re short on time or just skimming, I suggest skipping to the description of Closeness-of-the-counterfactual. It’s the factor I think is least recognized and the one I most wanted to write about.

What makes some emotions stronger than others? Why are some the faintest whispers, easily missed and others roaring, crashing storms which threaten to consume us?

The obvious answer is that emotions vary in intensity in proportion to the magnitude of what they’re about. Things which are a little bit good or a little bit bad evoke weak pleasant or aversive feelings, while things which are amazingly good or terribly bad provoke strong feeling. However, I assert that this magnitude is only one factor among several and is insufficient on its own to explain what causes strong or weak emotions.

In this post, I list the factors which are salient to me: magnitude plus three others. I do not think that they are especially surprising or profound, but I claim paying attention to them allows us to be better appreciate the mechanistic and lawful operation of emotions. This appreciation is practical in that it lets us better recognize and remedy common emotional pathologies. [See Section 2: Problematic manipulations of the factors].

Section 1: The Factors

  • Magnitude [of the stimulus]
  • Attention
  • Closeness-of-the-counterfactual
  • Actionability

Magnitude [of the stimulus]

This factor is the most obvious and least profound of all of the factors. However, it is illuminating to note just how insufficient it is to drive an emotion in the absence of the other factors.

Taking it as an assumption that all emotions are about something in the world [1], the strength of the emotion generally scales with magnitude of the “goodness” or “badness” which caused it. One feels stronger grief when their house burns down than when they dropped their cookie in the dirt. Making thousands from Bitcoin feels better than finding a twenty dollar bill on the street.


Obvious and yet still underappreciated.

Any given person is aware of thousands of situations, circumstances, and facts that could evoke just about any feeling. Meditate on your good fortune and you might feel happiness; think about those starving and diseased and you’ll feel sad; remember that unfair thing your teacher did in third grade and you’ll feel mad, and so on.

Emotions are usually about reality, but the emotions we experience are not about all the realities of which we are knowledgeable. No human could ever contain that much emotion. Instead, our emotions tend to be about whatever happens to be in our attention (broadly defined).

It’s like humans have this “attention slot” where you can put something, i.e. think about it, and then that’s what you’ll have emotions about. I am using attention in a broad and loose sense here. What I’m gesturing isn’t fully under one’s control and extends beyond conscious awareness. Think of how strong grief can stay present somewhere in your mind even while you try to do other things.

That we have emotions only about things in the “attention slot” solves the problem of humans not being able to simultaneously have emotions about all the realities of which they are aware (or could imagine), but more importantly it serves the adaptive purpose of emotions. Emotions are meant to guide behavior. It makes sense that emotions should be driven by the immediate, contextual, things we’re currently dealing with, i.e. those thing we’re paying attention to. It’s not valuable to be feeling happy about something good which happened a month ago if right now you’re in a bad situation you should get out of. Your emotions will be about the current situation, assuming that is, you’re paying attention to your present.

Even if you’re ruminating about your past, it should be [2] so that you can learn from mistakes and successes in a healthy way so as to succeed going forward. If you are fantasizing about possible futures, it should be to motivate you to work towards them.

The attention-moderator nature of emotion gives rise to a number of common observations:

  • We’re less upset by things over time; they’re no longer events we’re paying attention to and they’re less relevant.
  • People attempt to feel better by distracting themselves from upsetting circumstances [see below Section 2: Problematic manipulation of these factors].
  • Gratitude journaling makes people happier.
  • We can evoke emotions just by thinking about things, i.e. placing our attention on them, regardless of whether they’re real or imaginary, good or bad.


Despite its supreme importance, this is the most under-recognized moderator of emotion intensity. As far as I’m aware, not that I’ve scoured the psych literature, there isn’t a common name for it.

ETA: Gram Stone points out that Kahneman & Tversky (1981) describe this concept despite not coining a term for it, and that Roese (1997) provides a good early review on counterfactual thinking including contrasting effects similar to what is discussed here.

For the most part, it is a lot more frustrating to have missed your flight by five minutes than it is to miss your flight by five hours. It is a lot more frustrating to miss your flight when it seems you could have changed one or two small actions to have made it rather than when success was just out of your control. For example, it is more frustrating to miss the flight if the cause was you spent too long on Facebook rather than your car happened to be improbably stolen right from your garage.

It is more disappointing to not get a job when you thought you would, and more exhilarating to win a competition if you weren’t sure you’d win.

In general, emotions which relate to a counterfactual (i.e., how things might have been different, i.e., pretty much all emotions), scale in intensity in proportion to how easy it is to imagine [3] the counterfactual having been true. It’s easier to imagine having made your flight when it would have taken a small decision on your part to make the difference, and somewhat harder if something out of your control would have needed happen differently or it would have taken an improbably large amount of effort on your part.

I call the “how easy it is to imagine the counterfactual” property closeness-of-the-counterfactual, or counterfactual-closeness for short.

As usual, counterfactual-closeness as a moderator of emotion intensity makes sense if emotions are supposed to be adaptive. It’s adaptive to have strong emotions about counterfactual realities you could have nearly reached - if only you’d done a few things different - than about realities, no matter how pleasant, that never seemed in reach. It just doesn’t get me anywhere to be dreadfully sad all the time that I wasn’t born able to fly.

It’s a simple principle, yet is consistent with many observations beyond the above.

  • We are more upset by things that people around us have than things no one has. I am sad that I don’t have a swimming pool when my neighbors do and I could technically afford it, but not sad that I don’t have a spaceship because that just seems unrealistic - unless I’m Musk or Bezos, I have no reasonable expectation that I would have one.
  • We are more sad to lose things we once had than about things we’ve never had.
  • Unexpected good fortune feels a lot more rewarding than expected good fortune.
  • Sometimes people try to feel better about a failure by saying “it was always hopeless.” In effect, they are trying to create counterfactual-distance so there is less pain.
    • This is related to a “sour grapes” response.
  • It is consistent with Eliezer’s observation that most people can’t find motivation to do things they think are less than 70% likely to succeed. Perhaps you need to assign 70% probability of success to have enough counterfactual-closeness to evoke emotion.
  • Vivid descriptions of things, images, and videos cause us to have stronger emotions. The representations make them easier to imagine (and also help load them into attention).

Probably related: Book Review: Surfing Uncertainty


Continuing the theme that emotions ought to be adaptive, it makes a lot more sense to have emotions about situations where you can do something than ones where you can’t, or at least about situations where you could have done something different.

Even in cases where an emotion might seem inert, the emotion itself is probably trying to effect the world.

However, I’m not as sure of my understanding of this moderator as the others. It might actually bring about more of a qualitative difference in emotions than quantitative. The observation I’m drawing on here is that the emotions related to unresolved conflict with a colleague feel different from those attending grief about something which is done and dusted. In the former, there’s a “pulling” from the emotion as though it wants something, while in the latter the emotional tone is clear and pure, just signalling to my mind that something bad happened and I should do things different in the future.

Section 2: Problematic manipulations of the factors

Humans are crafty creatures with awareness of their own minds and the ability to manipulate the inputs they feed into their own minds to game the system. In short, we have some ability to wirehead. Or even if we’re not wireheading, these factors are pieces of the system which can be vulnerable to specific attacks or their own failure modes.

Attention and counterfactual-closeness are clear examples.

Manipulating attention: distraction to avoid unpleasant emotions

It’s easy to see that many people distract themselves from unpleasant realities to escape unpleasant emotions. I think it’s underappreciated just how absurdly widespread the behavior is.

The pernicious part of avoidance-behaviors is that many behaviors used for distraction, i.e. almost anything pleasurable, could be done purely for the sake of pleasure and in many cases is perfectly healthy and good. It’s easy to claim that you’re eating cake simply because cake is tasty unrelated to anything else. Yet, people are often compulsively and habitually looking for something stimulating to keep attention of the painful [4].

Compulsion is likely the differentiator about whether pleasure seeking behavior is driven by distraction and avoidance. Is a person reading a novel because they really feel like it and it’s a good time, or is there something it is helping them avoid, e.g. an assignment? The test I apply in this case is to ask whether I think a particular behavior optimizes my life as a whole or whether it’s just this experience in the moment being optimized. Over what timescale does this behavior improve on my life? Manipulation of attention to avoid pain (wireheading) will often be to the detriment of one’s life overall - pleasant in the short run, worse in the long-run.

Manipulating counterfactual-closeness

There is a temptation to artificially increase counterfactual-closeness to realities which are pleasant. Someone might cling to a dream of becoming an olympic runner, even as evidence mounts against them. They focus solely on the positive signs and they repeatedly and obsessively enumerate the pathways through which it all might work out. They’ve distorted their view of the world to see the desired world as much closer to reality than it is, because it feels good. They even become attached to the fantasy they’ve constructed. To maintain, it they have to twist the evidence and twist their epistemics, i.e. a one-sided counting of all evidence in one direction while ignoring all contrary data. This behavior is common in romantic contexts too. I assert that overall people behaving this way would be better served by good epistemics and accurate assessments of counterfactual-closeness.

There is equally a motivation to decrease counterfactual-closeness, i.e, increase counterfactual-distance. Things are easier to accept when they seem necessary, unavoidable, and so believing them so is a way to avoid pain. Most people do not appear to be pained that millions of people are dying, millions starving, millions diseased. They’re not distressed by their own imminent and assured death. Partly I think the pain is avoided by avoiding place attention on these topics, but also I think there’s motivated cognition to believe that it is impossible to do anything. Merely believing that there is something which could be done, having greater counterfactual-closeness, means experiencing pain that the something hasn’t been done yet.

This explains that people are resistant if you try to tell them things they could do. Believing something could be done would require them to move counter to a hedonic gradient, out of a local optimum. Believing something could be done (but hasn’t been) hurts more than believing nothing could be done, even though the former is how you get the best state of all - where something has been done successfully.

Manipulating magnitude and actionability

I imagine that people will readily recognize the behavior of people protesting through tears that something is “no big deal” in attempt to minimize their feelings - they are minimizing magnitude. And the behavior of insisting “nothing can be done” to quieten any nagging sense of responsibility, external or internal - they are minimizing actionability.


[1] Arguably some might emotions are not about anything, e.g. in emotion dysregulation disorders such as depression. My counter is that even if some emotions are detached from reality, that is a breakdown in the proper operation whose design and purpose is guide behavior within reality. Any “healthy” emotion will be “about” something.

[2] “Should” from the perspective of what emotions are “designed for”, namely that they are trying to drive adaptive action..

[3] The relevant kind of “imagination” here is a S1, instinctive feeling around gut expectations about what will happen or could have happened. It’s more than a S2, abstract picturing of a scenario in your mind.

[4] When I say “painful”, I mean anything at all slightly aversive. If I’m shy and dislike phone calls, I might put off calling the bank about the mistaken charge for weeks to avoid my slight discomfort. Us humans are sensitive to even the gentlest hedonic gradients.

New Comment
11 comments, sorted by Click to highlight new comments since: Today at 8:52 AM

For those wondering about the literature, although Kahneman and Tversky coined no term for it, Kahneman & Tversky (1981) describes counterfactual-closeness and some of its affective consequences. This paper appears to be the origin of the missed flight example. Roese (1997) is a good early review on counterfactual thinking with a section on contrast effects, of which closeness effects are arguably an instance.

Thanks for surfacing these! I've now edited the post to mention these sources and your comment.

My personal experience is that a separate factor of “emotional susceptibility” is very important. Tiredness, stress and repeated emotional experiences can sensitise me to feel stronger emotions.

Spending more time ruminating on and learning from "almost had it" makes sense evolutionary, since one is quite likely to encounter a similar situation in the future and with only a minor effort can do better next time.

Ease of imagining yes, I'd also suggest that the dysfunction comes about when the ease of imagining the result gets separated from the ease of imagining/charting the causal path. Leads to envy, entitlement, victimhood like areas of mindspace.

This appears to have been presented without sources. Where did you draw these conclusions?

Personal observation. I don't have any particular sources for anything here, though my thinking is influenced by some academic reading about emotions over the years. There are models more than conclusions and my intention is that readers evaluate them using their own observations rather than accept them based on sources, studies, or my say so.

I think this might be a post that benefits from an epistemic status tag. I had roughly the sense of where you were coming from, but think it's a good habit for LW to clarify that sort of thing. (I also think the title of the post makes it come across a bit more authoritative than you intend)

Epistemic status tag added. Thanks.

Also, regression to the mean is a force. Oversleeping right on the next day after you were praised for punctuality is somehow more annoying.

• We are more sad to lose things we once had than about things we’ve never had.

I'd imagine there's a limit to the amount of information that humans are born with (today). On the other hand, I think there are edge cases - things that people can reliably be predicted to enjoy, prior to any exposure, and things people are less happy without even if they've never had them.

• Unexpected good fortune feels a lot more rewarding than expected good fortune.

When I go to see a movie, that I expect to be good, it's just as good even though I expected it to be so. (There is more disappointment if it's just okay, so pursuing this might lead to better outcomes though.)

Manipulation of attention to avoid pain (wireheading) will often be to the detriment of one’s life overall - pleasant in the short run, worse in the long-run.

I see this as 'not being strategic' - going for things that are easy, familiar, or are guaranteed to work out - a lack of exploration. I also find it harder to see the harm where minor injury is involved - it will be healed tomorrow, so distracting yourself from the pain now, or going to bed early doesn't seem to have any obvious downsides.