The Martial Art of Rationality

I of­ten use the metaphor that ra­tio­nal­ity is the mar­tial art of mind. You don’t need huge, bulging mus­cles to learn mar­tial arts—there’s a ten­dency to­ward more ath­letic peo­ple be­ing more likely to learn mar­tial arts, but that may be a mat­ter of en­joy­ment as much as any­thing else. Some hu­man be­ings are faster or stronger than oth­ers; but a mar­tial art does not train the var­i­ance be­tween hu­mans. A mar­tial art just trains your mus­cles—if you have the hu­man-uni­ver­sal com­plex ma­chin­ery of a hand, with ten­dons and mus­cles in the ap­pro­pri­ate places, then you can learn to make a fist.

How can you train a var­i­ance? What does it mean to train +2 stan­dard de­vi­a­tions of mus­cle? It’s equally un­clear what it would mean to train an IQ of 132.

But if you have a brain, with cor­ti­cal and sub­cor­ti­cal ar­eas in the ap­pro­pri­ate places, you might be able to learn to use it prop­erly. If you’re a fast learner, you might learn faster—but the art of ra­tio­nal­ity isn’t about that; it’s about train­ing brain ma­chin­ery we all have in com­mon.

Alas, our minds re­spond less read­ily to our will than our hands. Mus­cles are evolu­tion­ar­ily an­cient sub­jects of neu­ral con­trol, while cog­ni­tive re­flec­tivity is a com­par­a­tively more re­cent in­no­va­tion. We shouldn’t be sur­prised that mus­cles are eas­ier to use than brains. But it is not wise to ne­glect the lat­ter train­ing be­cause it is more difficult. It is not by big­ger mus­cles that the hu­man species rose to promi­nence upon Earth.

If you live in an ur­ban area, you prob­a­bly don’t need to walk very far to find a mar­tial arts dojo. Why aren’t there do­jos that teach ra­tio­nal­ity? One rea­son, per­haps, is that it’s harder to ver­ify skill. To rise a level in Tae Kwon Do, you might need to break a board of a cer­tain width. If you suc­ceed, all the on­look­ers can see and ap­plaud. If you fail, your teacher can watch how you shape a fist, and check if you shape it cor­rectly. If not, the teacher holds out a hand and makes a fist cor­rectly, so that you can ob­serve how to do so. Within mar­tial arts schools, tech­niques of mus­cle have been re­fined and elab­o­rated over gen­er­a­tions. Tech­niques of ra­tio­nal­ity are harder to pass on, even to the most will­ing stu­dent. It is also harder to give im­pres­sive pub­lic ex­hi­bi­tions of ra­tio­nal­ity. This may par­tially ex­plain why there are no ra­tio­nal­ity do­jos as yet.

Very re­cently—in just the last few decades—the hu­man species has ac­quired a great deal of new knowl­edge about hu­man ra­tio­nal­ity. The most salient ex­am­ple would be the heuris­tics and bi­ases pro­gram in ex­per­i­men­tal psy­chol­ogy. There is also the Bayesian sys­tem­ati­za­tion of prob­a­bil­ity the­ory and statis­tics; evolu­tion­ary psy­chol­ogy; so­cial psy­chol­ogy. Ex­per­i­men­tal in­ves­ti­ga­tions of em­piri­cal hu­man psy­chol­ogy; and the­o­ret­i­cal prob­a­bil­ity the­ory to in­ter­pret what our ex­per­i­ments tell us; and evolu­tion­ary the­ory to ex­plain the con­clu­sions. Th­ese fields give us new fo­cus­ing lenses through which to view the land­scape of our own minds. With their aid, we may be able to see more clearly the mus­cles of our brains, the fingers of thought as they move. We have a shared vo­cab­u­lary in which to de­scribe prob­lems and solu­tions. Hu­man­ity may fi­nally be ready to syn­the­size the mar­tial art of mind: to re­fine, share, sys­tem­atize, and pass on tech­niques of per­sonal ra­tio­nal­ity.

Such un­der­stand­ing as I have of ra­tio­nal­ity, I ac­quired in the course of wrestling with the challenge of Ar­tifi­cial Gen­eral In­tel­li­gence (an en­deavor which, to ac­tu­ally suc­ceed, would re­quire suffi­cient mas­tery of ra­tio­nal­ity to build a com­plete work­ing ra­tio­nal­ist out of tooth­picks and rub­ber bands). In most ways the AI prob­lem is enor­mously more de­mand­ing than the per­sonal art of ra­tio­nal­ity, but in some ways it is ac­tu­ally eas­ier. In the mar­tial art of mind, we need to ac­quire the re­al­time pro­ce­du­ral skill of pul­ling the right lev­ers at the right time on a large, pre-ex­ist­ing think­ing ma­chine whose in­nards are not end-user-mod­ifi­able. Some of the ma­chin­ery is op­ti­mized for evolu­tion­ary se­lec­tion pres­sures that run di­rectly counter to our de­clared goals in us­ing it. De­liber­ately we de­cide that we want to seek only the truth; but our brains have hard­wired sup­port for ra­tio­nal­iz­ing false­hoods. We can try to com­pen­sate for what we choose to re­gard as flaws of the ma­chin­ery; but we can’t ac­tu­ally rewire the neu­ral cir­cuitry. Nor may a mar­tial artist plate tita­nium over his bones—not to­day, at any rate.

Try­ing to syn­the­size a per­sonal art of ra­tio­nal­ity, us­ing the sci­ence of ra­tio­nal­ity, may prove awk­ward: One imag­ines try­ing to in­vent a mar­tial art us­ing an ab­stract the­ory of physics, game the­ory, and hu­man anatomy. But hu­mans are not re­flec­tively blind; we do have a na­tive in­stinct for in­tro­spec­tion. The in­ner eye is not sightless; but it sees blur­rily, with sys­tem­atic dis­tor­tions. We need, then, to ap­ply the sci­ence to our in­tu­itions, to use the ab­stract knowl­edge to cor­rect our men­tal move­ments and aug­ment our metacog­ni­tive skills. We are not writ­ing a com­puter pro­gram to make a string pup­pet ex­e­cute mar­tial arts forms; it is our own men­tal limbs that we must move. There­fore we must con­nect the­ory to prac­tice. We must come to see what the sci­ence means, for our­selves, for our daily in­ner life.

And we must, above all, figure out how to com­mu­ni­cate the skill; which may not be a mat­ter for declar­a­tive state­ments alone. Mar­tial artists spar with each other, ex­e­cute stan­dard forms, and are watched through­out by their teach­ers. Calcu­lus stu­dents do home­work, and check their an­swers. Olympic run­ners con­tinu­ally try to beat their best pre­vi­ous time, as mea­sured by a stop­watch.

How to com­mu­ni­cate pro­ce­du­ral skills of ra­tio­nal­ity, or mea­sure them, is prob­a­bly the sin­gle largest open is­sue that stands be­tween hu­man­ity and ra­tio­nal­ity do­jos—at least it’s the part of the prob­lem that most baf­fles me. Mean­while I lec­ture. So does any­one out there have ideas?