Reason isn’t magic

Link post

Here’s a story some peo­ple like to tell about the limits of rea­son. There’s this plant, man­ioc, that grows eas­ily in some places and has a lot of calories in it, so it was a sta­ple for some in­dige­nous South Amer­i­cans since be­fore the Euro­peans showed up. Tra­di­tional han­dling of the man­ioc in­volved some elab­o­rate time-con­sum­ing steps that had no ap­par­ent pur­pose, so when the Por­tuguese in­tro­duced it to Africa, they didn’t bother with those steps—just, grow it, cook it, eat it.

The prob­lem is that man­ioc’s got cyanide in it, so if you eat too much too of­ten over a life­time, you get sick, in a way that’s not eas­ily trace­able to the plant. Some­how, over prob­a­bly hun­dreds of years, the peo­ple liv­ing in man­ioc’s origi­nal range figured out a way to leach out the poi­son, with­out un­der­stand­ing the un­der­ly­ing chem­istry—so if you asked them why they did it that way, they wouldn’t nec­es­sar­ily have a good an­swer.

Now a bunch of Afri­cans grow­ing and eat­ing man­ioc as a sta­ple reg­u­larly get cyanide poi­son­ing.

This is offered as a cau­tion­ary tale against in­no­vat­ing through rea­son, since there’s a lot of in­for­ma­tion em­bed­ded in your cul­ture (via hun­dreds of years of se­lec­tion), even if peo­ple can’t ex­plain why. The prob­lem with this ar­gu­ment is that it’s a non­sense com­par­i­son.

First of all, it’s not clear things got worse on net, just that a trade­off was made. How many per­son-days per year were freed up by less la­bor-in­ten­sive man­ioc han­dling? Has any­one both­ered to count the hours lost to la­bo­ri­ous tra­di­tional man­ioc-pro­cess­ing, to com­pare them with the bur­den of con­sum­ing too much cyanide? How many of us, know­ing that con­ve­nience foods prob­a­bly lower our lifes­pans rel­a­tive to slow foods, still eat them be­cause they’re … more con­ve­nient?

How many peo­ple didn’t starve be­cause man­ioc was available and would grow where and when other things wouldn’t?

If this is the best we can do for how poorly rea­son can perform, rea­son seems pretty great.

Se­cond, we’re not ac­tu­ally com­par­ing rea­son to tra­di­tion—we’re com­par­ing chang­ing things to not chang­ing things. Change, as we know, is bad. Some­times we change things any­way—when we think it’s worth the price, or the risk. Some­times, we’re wrong.

Third, the ac­tu­ally ex­ist­ing Por­tuguese and Afri­cans in­volved in this ex­per­i­ment weren’t com­mit­ted ra­tio­nal­ists—they were just peo­ple try­ing to get by. It prob­a­bly doesn’t take more than a day’s rea­son­ing to figure out which steps in grow­ing man­ioc are re­ally nec­es­sary to get the calories palat­ably. Are we imag­in­ing that some­one mak­ing a con­certed effort to im­prove their life through rea­son would just stop there?

This is be­ing com­pared with many gen­er­a­tions of trial and er­ror. Is that the stan­dard we want to use? Rea­son­ing isn’t worth it un­less a day of un­trained think­ing can out­perform hun­dreds of years of ac­cu­mu­lated tra­di­tion?

It gets worse. This isn’t a ran­domly se­lected ex­am­ple—it’s speci­fi­cally se­lected as a case where rea­son would have a hard time notic­ing when and how it’s mak­ing things worse. In this par­tic­u­lar case, rea­son in­tro­duced an im­por­tant prob­lem. But life is full of risks, some­times in ways that are worse for tra­di­tional cul­tures. Do we re­ally want to say that rea­son­ing isn’t the bet­ter bet un­less it out­performs liter­ally ev­ery time, with­out ever mak­ing things lo­cally worse? Even the­o­ret­i­cally perfect Bayesian ra­tio­nal­ity will some­times recom­mend changes that have an ex­pected benefit, but turn out to be harm­ful. Not even tra­di­tion meets this stan­dard! Only log­i­cal cer­tain­ties do—pro­vided, that is, we haven’t made an er­ror in one of our proofs.

We also have to count all the deaths and other prob­lems averted by rea­son­ing about a prob­lem. Rea­son­ing in­tro­duces risks—but also, risks come up even when we’re not rea­son­ing about them, just from peo­ple do­ing things that af­fect their en­vi­ron­ments. There’s ab­solutely no rea­son to think that the sort of grad­ual iter­a­tion that ac­cretes into tra­di­tion never en­ters a bad pos­i­tive feed­back loop. Even if you think moder­nity is an ex­cep­tional case of that kind of bad feed­back loop, we had to have got­ten there via the ac­cre­tion of pre­mod­ern tra­di­tion and iter­a­tion!

The only way out is through. But why did we have this ex­ag­ger­ated idea of what rea­son could do, in the first place?