Optimizing the Twelve Virtues of Rationality

At the Less Wrong Meetup in Colum­bus, OH over the last cou­ple of months, we dis­cussed op­ti­miz­ing the Twelve Virtues of Ra­tion­al­ity. In do­ing so, we were in­spired by what Eliezer him­self said in the es­say:

  • Per­haps your con­cep­tion of ra­tio­nal­ity is that it is ra­tio­nal to be­lieve the words of the Great Teacher, and the Great Teacher says, “The sky is green,” and you look up at the sky and see blue. If you think: “It may look like the sky is blue, but ra­tio­nal­ity is to be­lieve the words of the Great Teacher,” you lose a chance to dis­cover your mis­take.

So we first de­cided on the pur­pose of op­ti­miz­ing, and set­tled on yield­ing virtues that would be most im­pact­ful and effec­tive for mo­ti­vat­ing peo­ple to be­come more ra­tio­nal, in other words op­ti­miza­tions that would pro­duce the most utilons and he­dons for the pur­pose of win­ning. There were a bunch of differ­ent sug­ges­tions. I tried to ap­ply them to my­self over the last few weeks and want to share my find­ings.

First Suggestion

Re­place Perfec­tion­ism with Improvement

Mo­ti­va­tion for Replacement

Perfec­tion­ism, both in how it pat­tern matches and in its ac­tual de­scrip­tion in the es­say, ori­ents to­ward fo­cus­ing on defects and er­rors in one­self. By de­pict­ing the self as always flawed, and por­tray­ing the as­piring ra­tio­nal­ist’s job as seek­ing to find the flaws, the virtue of perfec­tion­ism is framed nega­tively, and is bound to re­sult in nega­tive re­in­force­ment. Find­ing a flaw feels bad, and in many peo­ple that cre­ates ugh fields around ac­tu­ally do­ing that search, as re­ported by par­ti­ci­pants at the Meetup. In­stead, a pos­i­tive fram­ing of this virtue would be Im­prove­ment. Then, the as­piring ra­tio­nal­ist can feel ok about where s/​he is right now, but ori­ent to­ward im­prov­ing and grow­ing men­tally stronger Tsuyoku Nar­i­tai! All im­prove­ment would be about gain­ing more he­dons, and thus use the power of pos­i­tive re­in­force­ment. Gen­er­ally, re­search sug­gests that pos­i­tive re­in­force­ment is effec­tive in mo­ti­vat­ing the rep­e­ti­tion of be­hav­ior, whereas nega­tive re­in­force­ment works best to stop peo­ple from do­ing a cer­tain be­hav­ior. No won­der that Meetup par­ti­ci­pants re­ported that Perfec­tion­ism was not very effec­tive in mo­ti­vat­ing them to grow more ra­tio­nal. So to get both more he­dons, and thereby more utilons in the sense of the util­ity of seek­ing to grow more ra­tio­nal, Im­prove­ment might be a bet­ter term and virtue than perfec­tion­ism.

Self-Report

I’ve been ori­ent­ing my­self to­ward im­prove­ment in­stead of perfec­tion­ism for the last few weeks, and it’s been a re­ally no­tice­able differ­ence. I’ve be­come much more mo­ti­vated to seek ways that I can im­prove my abil­ity to find the truth. I’ve been more ex­cited and en­thused about find­ing flaws and er­rors in my­self, be­cause they are now an op­por­tu­nity to im­prove and grow stronger, not be­come less weak and im­perfect. It’s the same out­come as the virtue of Perfec­tion­ism, but de­ploy­ing the power of pos­i­tive re­in­force­ment.

Se­cond Suggestion

Re­place Ar­gu­ment with Community

Mo­ti­va­tion for Replacement

Ar­gu­ment is an im­por­tant virtue, and a vi­tal way of get­ting our­selves to see the truth is to rely on oth­ers to help us see the truth through de­bates, high­light mis­taken be­liefs, and help up­date on them, as the virtue de­scribes. Yet ori­ent­ing to­ward a ra­tio­nal­ist Com­mu­nity has ad­di­tional benefits be­sides the benefits of ar­gu­ment, which is only one part of a ra­tio­nal­ist Com­mu­nity. Such a com­mu­nity would help provide an ex­ter­nal per­spec­tive that re­search sug­gests would be es­pe­cially benefi­cial to point­ing out flaws and bi­ases within one’s abil­ity to eval­u­ate re­al­ity ra­tio­nally, even with­out an ar­gu­ment. A com­mu­nity can help provide wise ad­vice on mak­ing de­ci­sions, and it’s es­pe­cially benefi­cial to have a com­mu­nity of di­verse and in­tel­li­gent peo­ple of all sorts in or­der to get the benefits of a wide va­ri­ety of pri­vate in­for­ma­tion that one can ag­gre­gate to help make the best de­ci­sions. More­over, a com­mu­nity can provide sys­tem­atic ways to im­prove, through giv­ing each sys­tem­atic feed­back, through com­pen­sat­ing for each oth­ers’ weak­nesses in ra­tio­nal­ity, through learn­ing difficult things to­gether, and other ways of sup­port­ing each oth­ers’ pur­suit of ever-greater ra­tio­nal­ity. Like­wise, a com­mu­nity can col­lab­o­rate to­gether, with differ­ent peo­ple fulfilling differ­ent func­tions in sup­port­ing all oth­ers in grow­ing men­tally stronger not ev­ery­body has to be the “hero,” af­ter all, and differ­ent peo­ple can spe­cial­ize in var­i­ous tasks re­lated to sup­port­ing oth­ers grow­ing men­tally stronger, gain­ing com­par­a­tive ad­van­tage as a re­sult. Stud­ies show that so­cial re­la­tion­ships im­pact us pow­er­fully in nu­mer­ous ways, con­tribute to our men­tal and phys­i­cal wellbe­ing, and that we be­come more like our so­cial net­work over time (1, 2, 3). This high­lights fur­ther the benefits of fo­cus­ing on de­vel­op­ing a ra­tio­nal­ist-ori­ented com­mu­nity of di­verse peo­ple around our­selves to help us grow men­tally stronger and get to the cor­rect an­swer, and gain he­dons and utilons al­ike for the pur­pose of win­ning.

Self-Report

After I up­dated my be­liefs to­ward Com­mu­nity from Ar­gu­ment, I’ve been work­ing more in­ten­tion­ally to cre­ate a sys­tem­atic way for other as­piring ra­tio­nal­ists in my LW meetup, and even non-ra­tio­nal­ists, to point out my flaws and bi­ases to me. I’ve no­ticed that by tak­ing ad­van­tage of out­side per­spec­tives, I’ve been able to make quite a bit more head­way on un­cov­er­ing my own false be­liefs and bi­ases. I asked friends, both fel­low as­piring ra­tio­nal­ists and other wise friends not cur­rently in the ra­tio­nal­ist move­ment, to help me by point­ing out when my bi­ases might be at play, and they were happy to do so. For ex­am­ple, I tend to have an op­ti­mism bias, and I have told peo­ple around me to watch for me ex­hibit­ing this bias. They pointed out a num­ber of times when this oc­curred, and I was able to im­prove grad­u­ally my abil­ity to no­tice and deal with this bias.

Third Suggestion

Ex­pand Em­piri­cism to in­clude Experimentation

Mo­ti­va­tion for Expansion

This would not be a re­place­ment of a virtue, but an ex­pan­sion of the defi­ni­tion of Em­piri­cism. As cur­rently stated, Em­piri­cism fo­cused on ob­ser­va­tion and pre­dic­tion, and im­plic­itly in mak­ing be­liefs pay rent in an­ti­ci­pated ex­pe­rience. This is a very im­por­tant virtue, and fun­da­men­tal to ra­tio­nal­ity. It can be im­proved, how­ever, by adding ex­per­i­men­ta­tion to the de­scrip­tion of em­piri­cism. By ex­per­i­men­ta­tion I mean ex­pand­ing sim­ply ob­ser­va­tion as de­scribed in the es­say cur­rently, to in­clude ac­tu­ally run­ning ex­per­i­ments and test­ing things out in or­der to up­date our maps, both about our­selves and in the world around us. This would help us take ini­ti­a­tive in gain­ing data around the world, not sim­ply rely­ing pas­sively on ob­ser­va­tion of the world around us. My per­spec­tive on this topic was fur­ther strength­ened by this re­cent dis­cus­sion post, which caused me to fur­ther up­date my be­liefs to­ward ex­per­i­men­ta­tion as a re­ally valuable part of em­piri­cism. Thus, in­clud­ing ex­per­i­men­ta­tion as part of em­piri­cism would get us more utilons for get­ting at the cor­rect an­swer and win­ning.

Self-Report

I have been run­ning ex­per­i­ments on my­self and the world around me long be­fore this dis­cus­sion took place. The dis­cus­sion it­self helped me con­nect the benefits of ex­per­i­men­ta­tion to the virtue of Em­piri­cism, and also see the gap cur­rently pre­sent in that virtue. I strength­ened my com­mit­ment to ex­per­i­men­ta­tion, and have been run­ning more con­crete ex­per­i­ments, where I both pre­dict the re­sults in ad­vance in or­der to make my be­liefs pay rent, and then run an ex­per­i­ment to test whether my be­liefs ac­tu­ally cor­re­lated to the out­come of the ex­per­i­ments. I have been hum­bled sev­eral times and got some great op­por­tu­ni­ties to up­date my be­liefs by com­bin­ing pre­dic­tion of an­ti­ci­pated ex­pe­rience with ac­tive ex­per­i­men­ta­tion.

Conclusion

The Twelve Virtues of Ra­tion­al­ity can be op­ti­mized to be more effec­tive and im­pact­ful for get­ting at the cor­rect an­swer and thus win­ning. There are many way of do­ing so, but we need to be care­ful in choos­ing op­ti­miza­tions that would be most op­ti­mal for the most peo­ple, as based on the re­search on how our minds ac­tu­ally work. The sug­ges­tions I shared above are just some ways of do­ing so. What do you think of these sug­ges­tions? What are your ideas for op­ti­miz­ing the Twelve Virtues of Ra­tion­al­ity?