To what extent does improved rationality lead to effective altruism?

It’s been claimed that in­creas­ing ra­tio­nal­ity in­creases effec­tive al­tru­ism. I think that this is true, but the effect size is un­clear to me, so it seems worth ex­plor­ing how strong the ev­i­dence for it is. I’ve offered some gen­eral con­sid­er­a­tions be­low, fol­lowed by a de­scrip­tion of my own ex­pe­rience. I’d very much wel­come thoughts on the effect that ra­tio­nal­ity has had on your own al­tru­is­tic ac­tivi­ties (and any other rele­vant thoughts).

The 2013 LW Sur­vey found that 28.6% of re­spon­dents iden­ti­fied as effec­tive al­tru­ists. This rate is much higher than the rate in the gen­eral pop­u­la­tion (even af­ter con­trol­ling for in­tel­li­gence), and be­cause LW is dis­t­in­guished by virtue of be­ing a com­mu­nity fo­cused on ra­tio­nal­ity, one might be led to the con­clu­sion that in­creas­ing ra­tio­nal­ity in­creases effec­tive al­tru­ism. But there are a num­ber of pos­si­ble con­found­ing fac­tors:

  1. It’s am­bigu­ous what the re­spon­dents meant when they said that they’re “effec­tive al­tru­ists.” (They could have used the term the way Wikipe­dia does, or they could have meant it in a more col­lo­quial sense.)

  2. In­ter­est in ra­tio­nal­ity and in­ter­est in effec­tive al­tru­ism might both stem from an un­der­ly­ing dis­po­si­tional vari­able.

  3. Effec­tive al­tru­ists may be dis­pro­por­tionately likely to seek to im­prove their epistemic ra­tio­nal­ity than are mem­bers of the gen­eral pop­u­la­tion.

  4. The ra­tio­nal­ist com­mu­nity and the effec­tive al­tru­ist com­mu­nity may have be­come in­ter­twined by his­tor­i­cal ac­ci­dent, out of virtue of hav­ing some early mem­bers in com­mon.

So it’s helpful to look be­yond the ob­served cor­re­la­tion and think about the hy­po­thet­i­cal causal path­ways be­tween in­creased ra­tio­nal­ity and in­creased effec­tive al­tru­ism.

The above claim can be bro­ken into sev­eral sub­claims (any or all of which may be in­tended):

Claim 1: When peo­ple are more ra­tio­nal, they’re more likely to pick their al­tru­is­tic en­deav­ors that they en­gage in with a view to­ward max­i­miz­ing util­i­tar­ian ex­pected value.

Claim 2: When peo­ple are more ra­tio­nal, they’re more likely to suc­ceed in their al­tru­is­tic en­deav­ors.

Claim 3: Be­ing more ra­tio­nal strength­ens peo­ple’s al­tru­is­tic mo­ti­va­tion.

Claim 1: “When peo­ple are more ra­tio­nal, they’re more likely to pick their al­tru­is­tic en­deav­ors that they en­gage in with a view to­ward max­i­miz­ing util­i­tar­ian ex­pected value.

Some el­e­ments of effec­tive al­tru­ism think­ing are:

  • Con­se­quen­tial­ism. In Yvain’s Con­se­quen­tial­ism FAQ, he ar­gues that con­se­quen­tial­ism fol­lows from the in­tu­itively ob­vi­ous prin­ci­ples “Mo­ral­ity Lives In The World” and “Others Have Non Zero Value” upon re­flec­tion. Ra­tion­al­ity seems use­ful for rec­og­niz­ing that there’s a ten­sion be­tween these prin­ci­ples and other com­mon moral in­tu­itions, but this doesn’t nec­es­sar­ily trans­late into a de­sire to re­solve the ten­sion nor a choice to re­solve the ten­sion in fa­vor of these prin­ci­ples over oth­ers. So it seems that in­creased ra­tio­nal­ity does in­crease the like­li­hood that one will be a con­se­quen­tial­ist, but that it’s also not suffi­cient.

  • Ex­pected value max­i­miza­tion. In Cir­cu­lar Altru­ism and el­se­where, Eliezer de­scribes cog­ni­tive bi­ases that peo­ple em­ploy in sce­nar­ios with a prob­a­bil­is­tic el­e­ment, and how re­flec­tion can lead one to the no­tion that one should or­ga­nize one’s al­tru­is­tic efforts to max­i­mize ex­pected value (in the tech­ni­cal sense), rather than mak­ing de­ci­sions based on these bi­ases. Here too, ra­tio­nal­ity seems use­ful for rec­og­niz­ing that one’s in­tu­itions are in con­flict be­cause of cog­ni­tive bi­ases, with­out nec­es­sar­ily en­tailing an in­cli­na­tion to re­solve the ten­sion. How­ever, in this case, if one does seek to re­solve the ten­sion, the choice of ex­pected value max­i­miza­tion over other al­ter­na­tives is canon­i­cal, so ra­tio­nal­ity seems to take one fur­ther to­ward ex­pected value max­i­miza­tion than to con­se­quen­tial­ism.

  • The prin­ci­ple of in­differ­ence. — The idea that from an al­tru­is­tic point of view, we should care about peo­ple who are un­re­lated to us as much as we do about peo­ple who are re­lated to us. For ex­am­ple, in The Life You Can Save: How to Do Your Part to End World Poverty, Peter Singer makes the case that we should show a similar de­gree of moral con­cern for peo­ple in the de­vel­op­ing world who are suffer­ing from poverty as we do to peo­ple in our neigh­bor­hoods. I’d ven­ture the guess its pop­u­lar­ity among ra­tio­nal­ists is an ar­ti­fact of cul­ture or a se­lec­tion effect rather than a con­se­quence of ra­tio­nal­ity. Note that con­cern about global poverty is far more preva­lent than in­ter­est in ra­tio­nal­ity (while still be­ing low enough so that global poverty is far from alle­vi­ated).

Claim 2: “When peo­ple are more ra­tio­nal, they’re more likely to suc­ceed in their al­tru­is­tic en­deav­ors.”

If “ra­tio­nal­ity” is taken to be “in­stru­men­tal ra­tio­nal­ity” then this is tau­tolog­i­cally true, so the rele­vant sense of “ra­tio­nal­ity” here is “epistemic.”

  • The ques­tion of how use­ful epistemic ra­tio­nal­ity is in gen­eral has been de­bated, (e.g. here, here, here, here, and here).

  • I think that epistemic ra­tio­nal­ity mat­ters more for al­tru­is­tic en­deav­ors than it does in other con­texts. Cog­ni­tive bi­ases were de­vel­oped for sur­vival and evolu­tion­ary fit­ness, and these things cor­re­late more strongly with per­sonal well-be­ing than with the well-be­ing of oth­ers. I think that epistemic ra­tio­nal­ity mat­ters still more for those who as­pire to max­i­mize util­i­tar­ian ex­pected value: cog­ni­tive bi­ases cor­re­late more strongly with well-be­ing of oth­ers within one’s so­cial cir­cles than they do with the well-be­ing of those out­side of one’s so­cial cir­cles.

  • In Cog­ni­tive Bi­ases Po­ten­tially Affect­ing Judg­ment of Global Risks, Eliezer de­scribes some cog­ni­tive bi­ases that can lead one to un­der­es­ti­mate the like­li­hood of risks of hu­man ex­tinc­tion. To the ex­tent that re­duc­ing these risks is the most promis­ing philan­thropic cause (as Eliezer has sug­gested), re­duc­ing cog­ni­tive bi­ases im­proves peo­ple’s prospects of max­i­miz­ing util­i­tar­ian ex­pected value.

Claim 3: “Be­ing more ra­tio­nal strength­ens peo­ple’s al­tru­is­tic mo­ti­va­tion.”

  • I think that there may be some effect in this di­rec­tion me­di­ated through im­proved well-be­ing: when peo­ple’s emo­tional well-be­ing in­creases, their em­pa­thy also in­creases.

  • It’s pos­si­ble to come to the con­clu­sion that one should care as much about oth­ers as one does about one­self through philo­soph­i­cal re­flec­tion, and I know peo­ple who have had this ex­pe­rience. I don’t know whether or not this is ac­cu­rately de­scribed as an effect at­tributable to im­proved ac­cu­racy of be­liefs, though.

Put­ting it all together

The con­sid­er­a­tions above point in the di­rec­tion of in­creased ra­tio­nal­ity of a pop­u­la­tion only slightly (if at all?) in­creas­ing the effec­tive al­tru­ism at the 50th per­centile of the pop­u­la­tion, but in­creas­ing the effec­tive al­tru­ism at higher per­centiles more, with the skew­ing be­com­ing more and more ex­treme the fur­ther up one goes. This is in par­allel with, e.g. the effect of height on in­come.

My own experience

In A per­sonal his­tory of in­volve­ment with effec­tive al­tru­ism I give some rele­vant au­to­bi­o­graph­i­cal in­for­ma­tion. Sum­ma­riz­ing and elab­o­rat­ing a bit:

  • I was fully on board with con­se­quen­tial­ism and with as­cribing similar value to strangers as to fa­mil­iar peo­ple as an early teenager, be­fore I had any knowl­edge of cog­ni­tive bi­ases as such, and at a time when my pre­dic­tive model of the world was in many ways weaker than those of most adults.

  • It was only when I read Eliezer’s posts that the jus­tifi­ca­tion for ex­pected value max­i­miza­tion in al­tru­is­tic con­texts clicked. Un­der­stand­ing it didn’t re­quire back­ground knowl­edge — it seems in­de­pen­dent of most as­pects of ra­tio­nal­ity.

  • I started read­ing Less Wrong be­cause a friend pointed me to Yvain’s posts on util­i­tar­i­anism. My in­ter­est in ra­tio­nal­ity was more driven by my in­ter­est in effec­tive al­tru­ism than the other way around. This is ev­i­dence that the high frac­tion of Less Wrongers who iden­tify as effec­tive al­tru­ists is par­tially a func­tion of it be­ing an at­trac­tor.

  • So far in­creased ra­tio­nal­ity hasn’t in­creased my pro­duc­tivity to a de­gree that’s statis­ti­cally sig­nifi­cant. There are changes that have oc­curred in my think­ing that greatly in­crease my pro­duc­tivity in the most fa­vor­able pos­si­ble fu­ture sce­nar­ios, rel­a­tive to a coun­ter­fac­tual in which these changes hadn’t oc­curred. This is in con­so­nance with my re­mark un­der the “putting it all to­gether” head­ing above.

How about you?