Arbital: Do what works: Justification

Do what works is meant to cre­ate a firm but flex­ible foun­da­tion for Ar­bital policy. The aims in­clude mak­ing the sys­tem self-cor­rect­ing by caus­ing peo­ple to flag bad policy, caus­ing ev­ery­thing to tie back to ac­tu­ally valuable goals, and avoid­ing pit­falls of other com­mu­nity struc­tures. It takes in­spira­tion from Wikipe­dia’s Ig­nore all rules and Eliezer’s Name­less Virtue.

Don’t we need hard rules to stop bad ac­tors?

The kind of user who badly vi­o­lates good prac­tices of any kind tends not to be the kind of user who reads policy pages. It is pointless to at­tempt to guard against them with strongly worded policy noteAnd harm­ful/​dis­cour­ag­ing to an­other class of user, those care­ful enough to check the rules. And staff check­ing for what tone they should use to­wards users..

In­stead, it’s im­por­tant to give the peo­ple who are part of sys­tems are in place to pre­vent harm­ful ac­tivity free­dom and af­for­dance to act quickly, enough train­ing and se­lec­tivity to re­li­ably tell ob­vi­ous from non-ob­vi­ous cases, a play­book of well-thought-out re­sponses to differ­ent classes of situ­a­tion, and a cul­ture of openly dis­cussing how best to re­spond with oth­ers who un­der­stand and can change the guidelines which are used by the sys­tem with the real goals always in mind.


At­tempt­ing to cod­ify ex­actly what con­tent should con­form to seems like a fool’s er­rand, es­pe­cially since we don’t know. Ex­plic­itly en­courag­ing peo­ple to do what works (aka. cre­ate the ver­sion of the page which is good for read­ers) rather than point­ing them at dozens of pages of guidelines they won’t read should cause them to op­ti­mize for some­thing closer to the right thing. We should, of course, offer guidelines and ad­vice for those who gen­uinely want it, and link to it heav­ily. Just, not ele­vate it to re­quired read­ing.

Why not a more le­gal struc­ture?

Many sites cre­ate an elab­o­rate web of rules out­law­ing spe­cific harm­ful pat­terns of be­hav­ior, and task peo­ple in po­si­tions of power (of­ten the most pro­duc­tive early con­trib­u­tors) with the work of im­ple­ment­ing the rules in a con­sis­tent and fair way.

How­ever, no hu­manly read­able set of rules can ad­e­quately cap­ture and reg­u­late the com­plex­ities of hu­man in­ter­ac­tion, and there are always awk­ward edge cases.

So there are peo­ple who are dis­rup­tive, and those with au­thor­ity/​re­spon­si­bil­ity to deal with it have a strict and hard to change rule­book, and com­plex, slow, tiring pro­cesses. This of­ten goes badly, set­ting up all sorts of ten­sions, ty­ing the hands of staff, caus­ing peo­ple to be acted against too strongly and too weakly in differ­ent cases, both of which harm the com­mu­nity. There is a place for le­gal­is­tic, care­ful, de­liber­a­tive, sys­tems, but it should not be the sys­tem called upon for most is­sues which come up.

By ty­ing the com­mu­nity rules ex­plic­itly to “do the thing which makes more healthy com­mu­nity” there’s sig­nifi­cantly more free­dom on the part of staff to act, and sig­nifi­cantly more keep­ing them pointed at what mat­ters (so long as the group self-po­lices well, which needs to be de­signed for so­cially/​struc­turally).

Isn’t this ob­vi­ous? Why would we need a policy to say “do the thing that works”?

Yes, it should be ob­vi­ous that policy should serve a pur­pose linked to a real good. How­ever, say­ing ob­vi­ous things is fine, and if you don’t have a clear foun­da­tion peo­ple tend to take an as­sort­ment of im­por­tant gen­eral guidelines there are and turn them into foun­da­tions, which are then harder to dis­pute even when ap­plied in coun­ter­pro­duc­tive ways, or ad­just when they turn out to cause prob­lems.