Philosophy as low-energy approximation

In 2015, Scott Alexan­der wrote a post origi­nally ti­tled High En­ergy Ethics. The idea is that when one uses an ex­treme thought ex­per­i­ment in ethics (peo­ple dy­ing, in­cest, the ex­tinc­tion of hu­man­ity, etc.), this is like smash­ing pro­tons to­gether at the speed of light at the LHC—an un­usual prac­tice, but one de­signed to teach us some­thing in­ter­est­ing and fun­da­men­tal.

I’m in­clined to think that not only is that a slight mischar­ac­ter­i­za­tion of what’s go­ing on, but that all philo­soph­i­cal the­o­ries that make strong claims about the “high en­ergy” regime are doubt­ful. But first, physics:

<physics>

Par­ti­cle physics is about things that are very en­er­getic—if we con­verted the en­ergy per par­ti­cle into a tem­per­a­ture, we could say the LHC pro­duces con­di­tions in ex­cess of a trillion (1,000,000,000) de­grees. But there is also a very broad class of physics top­ics that only seem to show up when it’s very cold—the su­per­con­duct­ing mag­nets in­side said LHC, only a few me­ters away from the trillion-de­gree quarks, need to be cooled to ba­si­cally ab­solute zero be­fore they su­per­con­duct.

The physics of su­per­con­duc­tors is similarly a lit­tle back­wards of par­ti­cle physics. Par­ti­cle physi­cists try to un­der­stand nor­mal, ev­ery­day be­hav­ior in terms of weird build­ing blocks. Su­per­con­duc­tor physi­cists try to un­der­stand weird be­hav­ior in terms of nor­mal build­ing blocks.

The com­mon pat­tern here is idea that the small build­ing blocks (in both fields) get “hid­den” at lower en­er­gies. We say that the high-en­ergy mo­tions of the sys­tem get “frozen out.” When a soup of fun­da­men­tal par­ti­cles gets cold enough, talk­ing about atoms be­comes a good low-en­ergy ap­prox­i­ma­tion. And when atoms get cold enough, we in­vent new low-en­ergy ap­prox­i­ma­tions like “the su­per­con­duct­ing or­der pa­ram­e­ter” as yet more con­ve­nient de­scrip­tions of their be­hav­ior.

</​physics>

Some philoso­phers think that they’re like par­ti­cle physi­cists, elu­ci­dat­ing the weird and on­tolog­i­cally ba­sic stuff in­side the ev­ery­day hu­man. The bet­ter philoso­phers, though, are like su­per­con­duc­tor physi­cists, try­ing to un­der­stand the un­usual (in a cos­mic sense) state of hu­man­ity in terms of mun­dane build­ing blocks.


My fa­vorite ex­am­ple of a “low-en­ergy ap­prox­i­ma­tion” in philos­o­phy, and the one that prompted this post, is Den­nett’s in­ten­tional stance.

The in­ten­tional stance ad­ver­tises it­self as a use­ful ap­prox­i­ma­tion. It’s a way of think­ing about cer­tain sys­tems (phys­i­cal agents) that are, at bot­tom, evolv­ing ac­cord­ing to the laws of physics with de­tail more com­pli­cated than we can com­pre­hend di­rectly. Even though the micro­scopic world is too com­pli­cated for us, we can use this model, the in­ten­tional stance, to pre­dict phys­i­cal agents (not-quite tau­tolog­i­cally defined as sys­tems the in­ten­tional stance helps pre­dict) us­ing a more man­age­able num­ber of free pa­ram­e­ters.

But some­times ap­prox­i­ma­tions break down, or fail to be use­ful—the ap­prox­i­ma­tion de­pends on cer­tain reg­u­lar­i­ties in the world that are not guaran­teed by the phys­i­cal law. To be di­rect, the col­lec­tion of atoms we think of as a “hu­man” isn’t an agent in the ab­stract sense. They can be ap­prox­i­mated as an agent, but that ap­prox­i­ma­tion will in­evitably break down in some phys­i­cal situ­a­tions. The psy­cholog­i­cal prop­er­ties that we as­cribe to hu­mans only make sense within the ap­prox­i­ma­tion—“In truth, there are only atoms and the void.”

Taken to its log­i­cal con­clu­sion, this is a di­rect re­jec­tion of most va­ri­eties of the “hard prob­lem of con­scious­ness.” The hard prob­lem asks, how can you take the phys­i­cal de­scrip­tion of a hu­man and ex­plain its Real Sen­sa­tions—our ex­pe­riences that are sup­posed to have their own ex­tra essences, or to be di­rectly ob­served by an “us” that is an ob­jec­tive ex­is­tence. But this is like ask­ing “Hu­man phys­i­cal bod­ies are only ap­prox­i­mate agents, so how does this gen­er­ate the real Pla­tonic agent I know I am in­side?” In short, maybe you’re not spe­cial. Ap­prox­i­mate agents also suffice to write books on philos­o­phy.

Show me a model that’s use­ful for un­der­stand­ing hu­man be­hav­ior, and I’ll show you some­one who’s taken it too liter­ally. Beliefs, ut­ter­ances, mean­ings, refer­ences, and so on—we just nat­u­rally want to ask “what is the true essence of this thing?” rather than “what ap­prox­i­ma­tion of the nat­u­ral world has these ob­jects as ba­sic el­e­ments?”

High-en­ergy philos­o­phy to­tally fails to ac­cept this re­al­ity. When you push hu­mans’ in­tu­itions to ex­tremes, you don’t get deep ac­cess to what they re­ally mean. You just get junk, be­cause you’ve pushed an ap­prox­i­ma­tion out­side its do­main of val­idity.

Take Put­nam’s Twin Earth thought ex­per­i­ment, where we try to an­a­lyze the idea (essence?) of “be­lief” or “about­ness” by pos­tu­lat­ing an en­tire al­ter­nate Earth that pe­ri­od­i­cally ex­changes peo­ple with our own. When you pon­der it, you feel like you are get­ting in­sights into the true na­ture of be­liev­ing. But more likely, there is no “true na­ture of be­liev­ing,” just some ap­prox­i­ma­tions of the nat­u­ral world that have “be­lief”s as ba­sic el­e­ments.

In the post on ethics, Scott gives some good ex­am­ples of highly charged thought ex­per­i­ments in ethics, and in some ways ethics is differ­ent from psy­chol­ogy - mod­ern ethics ac­knowl­edges that it’s largely about rhetoric and col­lab­o­ra­tion among hu­man be­ings. And yet it’s tel­ling that the ex­am­ples are all coun­terex­am­ples to other peo­ples’ pet the­o­ries.

If Kant claims you should never ever lie, all you need to re­fute him is one coun­terex­am­ple, and it’s okay if it’s a lit­tle ex­treme. But just be­cause you can re­fute wrong things with high-en­ergy thought ex­per­i­ments doesn’t mean they’re go­ing to help you find the right thing. The les­son of high en­ergy ethics seems to be that ev­ery neat eth­i­cal the­ory breaks down in some high en­ergy situ­a­tion.

Ap­pli­ca­tions to value learn­ing left (for now) as an ex­er­cise for the reader.