Psychic Powers

If the “bor­ing view” of re­al­ity is cor­rect, then you can never pre­dict any­thing ir­re­ducible be­cause you are re­ducible. You can never get Bayesian con­fir­ma­tion for a hy­poth­e­sis of ir­re­ducibil­ity, be­cause any pre­dic­tion you can make is, there­fore, some­thing that could also be pre­dicted by a re­ducible thing, namely your brain.

Benja Fallen­stein com­mented:

I think that while you can in this case never de­vise an em­piri­cal test whose out­come could log­i­cally prove ir­re­ducibil­ity, there is no clear rea­son to be­lieve that you can­not de­vise a test whose coun­ter­fac­tual out­come in an ir­re­ducible world would make ir­re­ducibil­ity sub­jec­tively much more prob­a­ble (given an Oc­camian prior).

Without get­ting into re­ducibil­ity/​ir­re­ducibil­ity, con­sider the sce­nario that the phys­i­cal uni­verse makes it pos­si­ble to build a hy­per­com­puter —that performs op­er­a­tions on ar­bi­trary real num­bers, for ex­am­ple —but that our brains do not ac­tu­ally make use of this: they can be simu­lated perfectly well by an or­di­nary Tur­ing ma­chine, thank you very much...

Well, that’s a very in­tel­li­gent ar­gu­ment, Benja Fallen­stein. But I have a crush­ing re­ply to your ar­gu­ment, such that, once I de­liver it, you will at once give up fur­ther de­bate with me on this par­tic­u­lar point:

You’re right.

Alas, I don’t get mod­esty credit on this one, be­cause af­ter pub­lish­ing yes­ter­day’s post I re­al­ized a similar flaw on my own—this one con­cern­ing Oc­cam’s Ra­zor and psy­chic pow­ers:

If be­liefs and de­sires are ir­re­ducible and on­tolog­i­cally ba­sic en­tities, or have an on­tolog­i­cally ba­sic com­po­nent not cov­ered by ex­ist­ing sci­ence, that would make it far more likely that there was an on­tolog­i­cal rule gov­ern­ing the in­ter­ac­tion of differ­ent minds—an in­ter­ac­tion which by­passed or­di­nary “ma­te­rial” means of com­mu­ni­ca­tion like sound waves, known to ex­ist­ing sci­ence.

If nat­u­ral­ism is cor­rect, then there ex­ists a con­ju­gate re­duc­tion­ist model that makes the same pre­dic­tions as any con­crete pre­dic­tion that any para­psy­chol­o­gist can make about telepa­thy.

In­deed, if nat­u­ral­ism is cor­rect, the only rea­son we can con­ceive of be­liefs as “fun­da­men­tal” is due to lack of self-knowl­edge of our own neu­rons—that the pe­cu­liar re­flec­tive ar­chi­tec­ture of our own minds ex­poses the “be­lief” class but hides the ma­chin­ery be­hind it.

Nonethe­less, the dis­cov­ery of in­for­ma­tion trans­fer be­tween brains, in the ab­sence of any known ma­te­rial con­nec­tion be­tween them, is prob­a­bil­is­ti­cally a priv­ileged pre­dic­tion of su­per­nat­u­ral mod­els (those that con­tain on­tolog­i­cally ba­sic men­tal en­tities). Just be­cause it is so much sim­pler in that case to have a new law re­lat­ing be­liefs be­tween differ­ent minds, com­pared to the “bor­ing” model where be­liefs are com­plex con­structs of neu­rons.

The hope of psy­chic pow­ers arises from treat­ing be­liefs and de­sires as suffi­ciently fun­da­men­tal ob­jects that they can have un­medi­ated con­nec­tions to re­al­ity. If be­liefs are pat­terns of neu­rons made of known ma­te­rial, with in­puts given by or­gans like eyes con­structed of known ma­te­rial, and with out­puts through mus­cles con­structed of known ma­te­rial, and this seems suffi­cient to ac­count for all known men­tal pow­ers of hu­mans, then there’s no rea­son to ex­pect any­thing more—no rea­son to pos­tu­late ad­di­tional con­nec­tions. This is why re­duc­tion­ists don’t ex­pect psy­chic pow­ers. Thus, ob­serv­ing psy­chic pow­ers would be strong ev­i­dence for the su­per­nat­u­ral in Richard Car­rier’s sense.

We have an Oc­cam rule that counts the num­ber of on­tolog­i­cally ba­sic classes and on­tolog­i­cally ba­sic laws in the model, and pe­nal­izes the count of en­tities. If nat­u­ral­ism is cor­rect, then the at­tempt to count “be­lief” or the “re­la­tion be­tween be­lief and re­al­ity” as a sin­gle ba­sic en­tity, is sim­ply mis­guided an­thro­po­mor­phism; we are only tempted to it by a quirk of our brain’s in­ter­nal ar­chi­tec­ture. But if you just go with that mis­guided view, then it as­signs a much higher prob­a­bil­ity to psy­chic pow­ers than does nat­u­ral­ism, be­cause you can im­ple­ment psy­chic pow­ers us­ing ap­par­ently sim­pler laws.

Hence the ac­tual dis­cov­ery of psy­chic pow­ers would im­ply that the hu­man-naive Oc­cam rule was in fact bet­ter-cal­ibrated than the so­phis­ti­cated nat­u­ral­is­tic Oc­cam rule. It would ar­gue that re­duc­tion­ists had been wrong all along in try­ing to take apart the brain; that what our minds ex­posed as a seem­ingly sim­ple lever, was in fact a sim­ple lever. The naive du­al­ists would have been right from the be­gin­ning, which is why their an­cient wish would have been en­abled to come true.

So telepa­thy, and the abil­ity to in­fluence events just by wish­ing at them, and pre­cog­ni­tion, would all, if dis­cov­ered, be strong Bayesian ev­i­dence in fa­vor of the hy­poth­e­sis that be­liefs are on­tolog­i­cally fun­da­men­tal. Not log­i­cal proof, but strong Bayesian ev­i­dence.

If re­duc­tion­ism is cor­rect, then any sci­ence-fic­tion story con­tain­ing psy­chic pow­ers, can be out­put by a sys­tem of sim­ple el­e­ments (i.e., the story’s au­thor’s brain); but if we in fact dis­cover psy­chic pow­ers, that would make it much more prob­a­ble that events were oc­cur­ring which could not in fact be de­scribed by re­duc­tion­ist mod­els.

Which just goes to say: The ex­is­tence of psy­chic pow­ers is a priv­ileged prob­a­bil­is­tic as­ser­tion of non-re­duc­tion­ist wor­ld­views—they own that ad­vance pre­dic­tion; they de­vised it and put it forth, in defi­ance of re­duc­tion­ist ex­pec­ta­tions. So by the laws of sci­ence, if psy­chic pow­ers are dis­cov­ered, non-re­duc­tion­ism wins.

I am there­fore con­fi­dent in dis­miss­ing psy­chic pow­ers as a pri­ori im­plau­si­ble, de­spite all the claimed ex­per­i­men­tal ev­i­dence in fa­vor of them.