Escaping Your Past

Fol­lowup to: Sunk Cost Fallacy

Re­lated to: Re­bel­ling Against Na­ture, Shut Up and Do the Im­pos­si­ble!

(ex­panded from my com­ment)

“The world is weary of the past—
O might it die or rest at last!”
— Percy Bysshe Shel­ley, from “Hel­las”

Prob­a­bil­ity the­ory and de­ci­sion the­ory push us in op­po­site di­rec­tions. In­duc­tion de­mands that you can­not for­get your past; the sunk cost fal­lacy de­mands that you must. Let me ex­plain.

An im­por­tant part of epistemic ra­tio­nal­ity is learn­ing to be at home in a ma­te­rial uni­verse. You are not a mag­i­cal fount of origi­nal­ity and free will; you are a phys­i­cal sys­tem: the same laws that bind the planets in their or­bits, also bind you; the same sorts of reg­u­lar­i­ties in these laws that gov­ern the lives of rab­bits or aphids, also gov­ern hu­man so­cieties. In­deed, in the last anal­y­sis, free will as tra­di­tion­ally con­ceived is but a con­fu­sion—and bind and gov­ern are mis­lead­ing metaphors at best: what is bound as by ropes can be un­bound with, say, a good knife; what is “bound” by “na­ture”—well, I can hardly finish the sen­tence, the phras­ing be­ing so ab­surd!

Epistemic ra­tio­nal­ity alone might be well enough for those of us who sim­ply love truth (who love truth­seek­ing, I mean; the truth it­self is usu­ally an abom­i­na­tion), but some of my friends tell me there should be some sort of pay­off for all this work of in­fer­ence. And in­deed, there should be: if you know how some­thing works, you might be able to make it work bet­ter. En­ter in­tru­men­tal ra­tio­nal­ity, the art of do­ing bet­ter. We all want to bet­ter, and we all be­lieve that we can do bet­ter...

But we should also all know that be­liefs re­quire ev­i­dence.

Sup­pose you’re an em­ployer in­ter­view­ing a job­seeker for a po­si­tion you have open. Ex­am­in­ing the job­seeker’s ap­pli­ca­tion, you see that she was ex­pel­led from four schools, was fired from her last three jobs, and was con­victed of two felonies. You ask, “Given your record, I re­gret hav­ing let you en­ter the build­ing. Why on Earth should I hire you?”

And the job­seeker replies, “But all those trans­gres­sions are in the past. Sunk costs can’t play into my de­ci­sion the­ory—it would hardly be helping for me to go sulk in a gut­ter some­where. I can only seek to max­i­mize ex­pected util­ity now, and right now that means work­ing ever so hard for you, O dear­est fu­ture boss! Tsuyoku nar­i­tai!

And you say, “Why should I be­lieve you?”

And then—oh, wait. Just a mo­ment, I’ve got­ten my notes mixed up—oh, dear. I’ve been tel­ling this sce­nario all wrong. You’re not the em­ployer. You’re the job­seeker.

Why should you be­lieve your­self? You hon­estly swear that you’re go­ing to change, and this is great. But take the out­side view. What good have these oaths done for all the other mil­lions who have sworn them? You might very well be differ­ent, but in or­der to jus­tifi­ably be­lieve that you’re differ­ent, you need to have some sort of ev­i­dence that you’re differ­ent. It’s not a spe­cial ques­tion; there has to be some­thing about your brain that is differ­ent, whether or not you can eas­ily com­mu­ni­cate this ev­i­dence to oth­ers with pre­sent tech­nol­ogy. What do you have be­sides the oath? Are you do­ing reasearch, try­ing new things, keep­ing track of re­sults, gen­uinely search­ing at long last for some­thing that will ac­tu­ally work?

For if you do suc­ceed, it won’t have been a mir­a­cle: you should be able to pin down at least ap­prox­i­mately the causal fac­tors that got you to where you are. And it has to be a plau­si­ble story. You won’t re­ally be able to say, “Well, I read all these blog­posts about ra­tio­nal­ity, and that’s why I’m such an amaz­ing per­son now.” Com­pare: “I read the Bible, and that’s why I’m such an amaz­ing per­son now.” The words are differ­ent, but trans­lated into math, is it re­ally a differ­ent story? It could be. But if it is, you should be able to ex­plain fur­ther; there has to be some co­her­ent se­quence of events that could take place in an ma­te­rial uni­verse, a con­tin­u­ous path through space­time that took you from there to here. If the blog helped, how speci­fi­cally did it help? What did it cause you to do that you would not oth­er­wise have done?

This could be more difficult than it now seems in your cur­rent ig­no­rance: the more you know about the forces that de­ter­mine you, the less room there is for mag­i­cal hopes. When you have a re­ally fan­tas­tic day, you’re more likely to ex­pect to­mor­row to be like that as well if you don’t know about re­gres­sion to­wards the mean.

I’m not try­ing to in­duce de­spair with this post; re­ally, I’m not. It is pos­si­ble to do bet­ter; I my­self am do­ing bet­ter than I was this time last year. I just think it’s im­por­tant to un­der­stand ex­actly what do­ing bet­ter re­ally in­volves.

I feel bad blog­ging about ra­tio­nal­ity, given that I’m so hor­ribly, lu­dicrously bad at it. I’m also hor­ribly, lu­dicrously bad at writ­ing. But it would hardly be helping for me to just shut up in de­spair—to go sulk in a gut­ter some­where. I can only seek to max­i­mize ex­pected util­ity now, and for now, that ap­par­ently means writ­ing the oc­ca­sional blog­post. Tsuyoku