Goals vs. Rewards

Re­lated: Ter­mi­nal Values and In­stru­men­tal Values, Ap­ply­ing be­hav­ioral psy­chol­ogy on myself

Re­cently I asked my­self, what do I want? My im­me­di­ate re­sponse was that I wanted to be less stressed, par­tic­u­larly for fi­nan­cial rea­sons. So I started to af­firm to my­self that my goal was to be­come wealthy, and also to be­come less stressed. But then in a fit of cog­ni­tive dis­so­nance, I re­al­ized that both money and re­lax­ation are most eas­ily con­sid­ered in terms of be­ing re­wards, not goals. I was oddly sur­prised by the fact that there is a dis­tinc­tion be­tween the two con­cepts to be­gin with.

It later oc­curred to me to won­der if some things work bet­ter when framed as goals and not as re­wards. Free­dom, long life, good re­la­tion­ships, and pro­duc­tivity seemed some likely can­di­dates. I can’t quite see them as re­wards be­cause a) I feel ev­ery­one in­nately de­serves and should have them (even though they might have to work for them), and b) they don’t quite give the kind of fuzzies that mo­ti­vate im­me­di­ate ac­tion.

Th­ese two kinds of pos­i­tive mo­ti­va­tion seem to work in psy­cholog­i­cally dis­similar ways. Money for ex­am­ple is more like choco­late, some­thing one has im­me­di­ate in­stinc­tive mo­tive to ob­tain and con­sume. Free­dom of speech is more along the lines of hav­ing enough air to breathe. A per­son needs and per­haps in­her­ently de­serves to have at least a lit­tle bit of it all the time, and as a gen­eral rule will have a con­stant back­ground mo­tive to en­sure that it stays available. It’s a longer-term form of mo­ti­va­tion.

A re­ward seems to be some­thing where you re­ceive im­me­di­ate fuzzies when you achieve it. Get­ting paid, get­ting a pat on the back, get­ting your posts and com­ments up­voted… Things where you might con­sider them more or less op­tional in the grander scheme of things, yet they tend to trig­ger an im­me­di­ate sense of pos­i­tive an­ti­ci­pa­tion be­fore the event which is re­in­forced by a sense of satis­fac­tion af­ter. Ac­tu­ally writ­ing a good post or com­ment, ac­tu­ally do­ing a good job, be­ing a good spouse or friend—these are surely re­lated, but are goals in and of them­selves. The men­tal pic­ture for a goal is one of achiev­ing, as op­posed to re­ceiv­ing.

One thing that seems likely to me is that the pres­ence of shared goals (and the com­mu­ni­ca­tion thereof) tends to a good way to gen­er­ate long term so­cial bonds. Re­wards seem to be more of a good way to de­liber­ately steer be­hav­ior in more spe­cific as­pects. Both are thus im­por­tant el­e­ments of so­cial sig­nal­ing within a tribe, but serve differ­ent un­der­ly­ing pur­poses.

As an ex­am­ple I have the tran­shu­man­ist goal of elimi­nat­ing the cur­rent limi­ta­tions of the hu­man lifes­pan, and tend to have an af­finity for peo­ple who also in­ter­nal­ize that goal. But some­one who does not em­brace that goal on a deep level may still dis­play spe­cific be­hav­ior that I con­sider helpful for that goal, e.g. dis­play­ing com­pre­hen­sion of its in­ter­nal logic or hav­ing a tol­er­ant at­ti­tude to­wards ac­tions I think need to be taken. I’m prob­a­bly some­what less likely to form a long-term re­la­tion­ship with that per­son than if they were iden­ti­fi­able as a fel­low tran­shu­man­ist, but I am still likely to up­vote their com­ments or oth­er­wise sig­nal ap­proval in ways that don’t de­mand too much long term com­mit­ment.

The dis­tinc­tions I’ve drawn here be­tween a goal and a re­ward might not ap­ply di­rectly to non-hu­man in­tel­li­gences. In fact it might be mis­lead­ing in the more gen­er­al­ized con­text to call a re­ward some­thing other than a goal (it is at least an im­plicit goal or value). How­ever the dis­tinc­tion still seems like some­thing that could be rele­vant for in­stru­men­tal ra­tio­nal­ity and per­sonal de­vel­op­ment. Our brains pro­cess the two forms of mo­ti­va­tional an­ti­ci­pa­tion in differ­ent ways. It may be that a part of the akra­sia prob­lem—failure to take ac­tion to­wards a goal—ac­tu­ally re­lates to a failure to prop­erly cat­e­go­rize a given mo­tive, and hence failure to pro­cess it use­fully.

Thanks to the early com­menters for their feed­back: TheOtherDave, nor­nagest, en­do­self, David Ger­ard, nazgul­nar­sil, and Nor­mal Ano­maly. Hope­fully this ex­panded ver­sion is more clear.