Here Be Epistemic Dragons


In Rhetoric, Aris­to­tle said that in an ar­gu­ment, “...if the de­ci­sions of judges are not what they ought to be, the defeat must be due to the speak­ers them­selves… the un­der­ly­ing facts do not lend them­selves equally well to the con­trary views. No; things that are true and things that are bet­ter are, by their na­ture, prac­ti­cally always eas­ier to prove and eas­ier to be­lieve in.” Yet we might ar­gue most about those rare top­ics where we do en­counter difficul­ties in prov­ing or be­liev­ing what’s true or best; per­haps the op­er­a­tive word in the quote is prac­ti­cally. To cap­ture the hero’s jour­ney qual­ity of in­quiry, let’s call these ex­cep­tional cases epistemic drag­ons. We have to deal with more re­search than ever be­fore, which con­stantly pushes the bound­aries of what is proven or be­liev­able and spawns epistemic drag­ons at a his­tor­i­cally ex­traor­di­nary rate.

Although we can take a so­ciolog­i­cal or an­thro­polog­i­cal view of the sci­en­tific, me­dia, and poli­ti­cal in­dus­tries that gen­er­ate so many of these epistemic drag­ons, each worker in these in­dus­tries is an in­di­vi­d­ual jonesing for truth. Alfréd Rényi said, “a math­e­mat­i­cian is a ma­chine for turn­ing coffee into the­o­rems,” but all our in­dus­try has not yet de­vised a way to mass-pro­duce stan­dard­ized ra­tio­nal minds, al­though we are try­ing.

It’s best prac­tice to base be­liefs on the pre­pon­der­ance of ev­i­dence, es­pe­cially con­trol­led and well-repli­cated ex­per­i­ments that look at a prob­lem from many an­gles. That’s the paved road through safe ter­ri­tory. A sci­ence pro­gresses faster when its fron­tiers can be ex­panded that way, but this is of­ten not pos­si­ble. Science should con­verge at a bor­der drawn pre­cisely at the point where lie the limits of our abil­ity to con­duct con­trol­led ex­per­i­ments. Here be drag­ons.

Ad­ven­tures in this dan­ger­ous ter­ri­tory aren’t of­ten mo­ti­vated by the pure de­sire to con­sume an en­tire body of sub­ject liter­a­ture for its own sake. Our hu­man val­ues, a puz­zle-solv­ing itch, in­tel­lec­tual wan­der­lust, or in­dus­trial and pro­fes­sional ob­jec­tives of­ten lead to a study that is of ne­ces­sity rather shal­low, am­a­teur­ish, and in­ter­dis­ci­plinary. Var­i­ous broad forms of con­sen­sus, log­i­cal solu­tions and sto­ry­tel­ling based on de­scrip­tive mod­els, di­rect per­cep­tion, fore­casts and prices, and dura­bil­ity over time are some of the in­stru­ments we nav­i­gate by.

I want to shift gears for a minute and talk about what it’s like to read. When I was lit­tle, my par­ents read to me, and the ex­pe­rience was, for a time, closer to mu­sic than to sto­ry­tel­ling. That is still with me. I find it mys­te­ri­ous how my brain as­sem­bles an in­tel­lec­tual struc­ture from that mu­sic. Some­times my brain feels “warm” when I have just been writ­ing or en­grossed in read­ing, and I won­der whether that’s my sub­con­scious work­ing, a sort of af­ter­glow, or an illu­sion of thought that is per­haps the ori­gin of over­con­fi­dence.

Why does this episte­molog­i­cal phe­nomenol­ogy mat­ter? The im­por­tant, tractable, ne­glected frame­work is a neat con­cep­tual com­pass that points to­ward ar­eas of thought and ac­tion where we have the chance to be pi­o­neers. That means strik­ing out for novel and there­fore risky ar­eas of thought and ac­tion that have been sub­jected to only the most min­i­mal tests, per­haps with lit­tle con­sen­sus or pub­lic aware­ness, frag­ile and in­com­plete mod­els, con­tested fore­cast­ing data. I’m not the first to say that those who prof­ited most from the Ore­gon Trail or the Klondike Gold Rush were prob­a­bly the peo­ple sel­l­ing equip­ment, sus­te­nance, and nav­i­ga­tional tools to set­tlers and min­ers. If a thing is im­por­tant and tractable, why is it ne­glected—and will it re­main that way long enough for me to cash in? His­tory, our in­tel­lec­tual pre­de­ces­sors (Aris­to­tle), and the logic of mar­kets all scream for sus­pi­cion about claims for free money, easy im­pact, hid­den truths with magic power. We should not let ar­ro­gant over­con­fi­dence about our own abil­ities make us naïve, but in our fear of naïveté, we can start to give ex­ces­sive at­ten­tion to our in­ner voice of para­noid in­tel­lec­tual con­ser­vatism. This gen­er­ates a cer­tain form of qualia that in­hibits an ac­cep­tance of the calcu­lated gam­ble re­quired to un­der­take in­no­va­tive work. I think ac­knowl­edg­ing that dy­namic and learn­ing to cope with it is im­por­tant for ra­tio­nal­ists.

Scott Alexan­der wrote a thought­ful es­say about “episte­molog­i­cal learned hel­pless­ness,” which prompted this re­flec­tion. He con­cluded that on some top­ics with a main­stream schol­arly con­sen­sus and a range of in­com­pat­i­ble but com­pel­ling fringe the­o­ries, an in­tel­li­gent am­a­teur reader will never be able to re­solve the de­bate for them­selves, and might do best to rigidly ad­here to the main­stream in­ter­pre­ta­tion. “Learned hel­pless­ness” evokes an image of our minds re­strained and hun­kered down like a de­mor­al­ized dog in a cruel ex­per­i­ment. I think of it in­stead as savvy shop­ping in the in­tel­lec­tual mar­ket­place. Self-made in­stru­ments have a ten­dency to fail un­ex­pect­edly.

In Alexan­der’s es­say “What Devel­op­men­tal Mile­stones are You Miss­ing?,” he lists sev­eral such mile­stones in the de­vel­op­ment of our men­tal op­er­a­tions. I would add two more.

  1. A the­ory of ap­pro­pri­ate com­mon prac­tices for fur­ther­ing an ar­gu­ment so that we can at least be con­scious when we see them. Read­ing should not only be a way to dis­cover new parts of the in­tel­lec­tual jun­gle, but also an op­por­tu­nity to in­crease our so­phis­ti­ca­tion as nav­i­ga­tors. Once I made my list of tests, I started to see them ev­ery time I opened a book, which changed the way I read. It’s like learn­ing to see a pin or fork in chess.

  2. An effort to dis­t­in­guish think­ing from pseu­docog­ni­tive qualia—in­vest­ment ad­vice for be­gin­ners points out how the ups and downs of the mar­ket make in­vestors overex­cited or scared, and warns them against mak­ing bad de­ci­sions based on a gut re­ac­tion. Maybe ra­tio­nal­ists should take a leaf out of that book.

My in­tu­ition tells me that these two mile­stones are sources of cre­ativity, helping us to form a re­sponse to the text and join with the thought of the au­thor, rather than be­ing charmed into sleep by the mu­sic or dis­con­nect­ing from it in a fit of para­noia. When I look at the book­shelves in my office, I see a noisy mar­ket­place where the ven­dors of in­finitely re­cur­sive maps are hawk­ing their wares, and I feel the wan­der­lust, that pseu­docog­ni­tive qualia, come over me. I no­tice that dwelling on qualia has a grav­i­ta­tional pull that is some­what em­bar­rass­ing, and I don’t want to ex­pose my­self too much or ap­pear too hu­man. Whereof one can­not speak, thereof one must be silent?

I am left won­der­ing whether there is a pro­duc­tive way to re­flect on qualia as­so­ci­ated with cog­ni­tion in a way that leads to gen­eral prin­ci­ples, and what other non-ex­per­i­men­tal test cat­e­gories I’m for­get­ting.

No comments.