Four factors which moderate the intensity of emotions

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.