# Decoherent Essences

Fol­lowup to: De­co­her­ence is Pointless

In “De­co­her­ence is Pointless”, we talked about quan­tum states such as

(Hu­man-BLANK) * ((Sen­sor-LEFT * Atom-LEFT) + (Sen­sor-RIGHT * Atom-RIGHT))

which de­scribes the evolu­tion of a quan­tum sys­tem just af­ter a sen­sor has mea­sured an atom, and right be­fore a hu­man has looked at the sen­sor—or be­fore the hu­man has in­ter­acted grav­i­ta­tion­ally with the sen­sor, for that mat­ter. (It doesn’t take much in­ter­ac­tion to de­co­here ob­jects the size of a hu­man.)

But this is only one way of look­ing at the am­pli­tude dis­tri­bu­tion—a way that makes it easy to see ob­jects like hu­mans, sen­sors, and atoms. There are other ways of look­ing at this am­pli­tude dis­tri­bu­tion—differ­ent choices of ba­sis—that will make the de­co­her­ence less ob­vi­ous.

Sup­pose that you have the “en­tan­gled” (non-in­de­pen­dent) state:

(Sen­sor-LEFT * Atom-LEFT) + (Sen­sor-RIGHT * Atom-RIGHT)

con­sid­er­ing now only the sen­sor and the atom.

This state looks nicely di­ag­o­nal­ized—sep­a­rated into two dis­tinct blobs. But by lin­ear­ity, we can take apart a quan­tum am­pli­tude dis­tri­bu­tion any way we like, and get the same laws of physics back out. So in a differ­ent ba­sis, we might end up writ­ing (Sen­sor-LEFT * Atom-LEFT) as:

(0.5(Sen­sor-LEFT + Sen­sor-RIGHT) + 0.5(Sen­sor-LEFT—Sen­sor-RIGHT)) * (0.5(Atom-RIGHT + Atom-LEFT) − 0.5(Atom-RIGHT—Atom-LEFT))

(Don’t laugh. There are le­gi­t­i­mate rea­sons for physi­cists to re­for­mu­late their quan­tum rep­re­sen­ta­tions in weird ways.)

The re­sult works out the same, of course. But if you view the en­tan­gled state in a ba­sis made up of lin­early in­de­pen­dent com­po­nents like (Sen­sor-LEFT—Sen­sor-RIGHT) and (Atom-RIGHT—Atom-LEFT), you see a differ­ently shaped am­pli­tude dis­tri­bu­tion, and it may not look like the blobs are sep­a­rated.

Oh noes! The de­co­her­ence has dis­ap­peared!

...or that’s the source of a huge aca­demic liter­a­ture ask­ing, “Doesn’t the de­co­her­ence in­ter­pre­ta­tion re­quire us to choose a preferred ba­sis?”

To which the short an­swer is: Choos­ing a ba­sis is an iso­mor­phism; it doesn’t change any ex­per­i­men­tal pre­dic­tions. De­co­her­ence is an ex­per­i­men­tally visi­ble phe­nomenon or we would not have to pro­tect quan­tum com­put­ers from it. You can’t pro­tect a quan­tum com­puter by “choos­ing the right ba­sis” in­stead of us­ing en­vi­ron­men­tal shield­ing. Like­wise, look­ing at split­ting hu­mans from an­other an­gle won’t make their de­co­her­ence go away.

But this is an is­sue that you’re bound to en­counter if you pur­sue quan­tum me­chan­ics, es­pe­cially if you talk to any­one from the Old School, and so it may be worth ex­pand­ing on this re­ply.

After all, if the short an­swer is as ob­vi­ous as I’ve made it sound, then why, oh why, would any­one ever think you could elimi­nate an ex­per­i­men­tally visi­ble phe­nomenon like de­co­her­ence, by iso­mor­phi­cally re­for­mu­lat­ing the math­e­mat­i­cal rep­re­sen­ta­tion of quan­tum physics?

That’s a bit difficult to de­scribe in one mere blog post. It has to do with his­tory. You know the warn­ing I gave about drag­ging his­tory into ex­pla­na­tions of QM… so con­sider your­self warned: Quan­tum me­chan­ics is sim­pler than the ar­gu­ments we have about quan­tum me­chan­ics. But here, then, is the his­tory:

Once upon a time,

Long ago and far away, back when the the­ory of quan­tum me­chan­ics was first be­ing de­vel­oped,

No one had ever thought of de­co­her­ence. The ques­tion of why a hu­man re­searcher only saw one thing at a time, was a Great Mys­tery with no ob­vi­ous an­swer.

You had to in­ter­pret quan­tum me­chan­ics to get an an­swer back out of it. Like read­ing mean­ings into an or­a­cle. And there were differ­ent, com­pet­ing in­ter­pre­ta­tions. In one pop­u­lar in­ter­pre­ta­tion, when you “mea­sured” a sys­tem, the Quan­tum Spaghetti Mon­ster would eat all but one blob of am­pli­tude, at some un­speci­fied time that was ex­actly right to give you what­ever ex­per­i­men­tal re­sult you ac­tu­ally saw.

Need­less to say, this “in­ter­pre­ta­tion” wasn’t in the quan­tum equa­tions. You had to add in the ex­tra pos­tu­late of a Quan­tum Spaghetti Mon­ster on top, ad­di­tion­ally to the differ­en­tial equa­tions you had fixed ex­per­i­men­tally for de­scribing how an am­pli­tude dis­tri­bu­tion evolved.

Along came Hugh Everett and said, “Hey, maybe the for­mal­ism just de­scribes the way the uni­verse is, with­out any need to ‘in­ter­pret’ it.”

But peo­ple were so used to adding ex­tra pos­tu­lates to in­ter­pret quan­tum me­chan­ics, and so un­used to the idea of am­pli­tude dis­tri­bu­tions as real, that they couldn’t see this new “in­ter­pre­ta­tion” as any­thing ex­cept an ad­di­tional De­co­her­ence Pos­tu­late which said:

“When clouds of am­pli­tude be­come sep­a­rated enough, the Quan­tum Spaghetti Mon­ster steps in and cre­ates a new world cor­re­spond­ing to each cloud of am­pli­tude.”

“Ex­actly how sep­a­rated do two clouds of am­pli­tude have to be, quan­ti­ta­tively speak­ing, in or­der to in­voke the in­stan­ta­neous ac­tion of the Quan­tum Spaghetti Mon­ster? And in which ba­sis does the Quan­tum Spaghetti Mon­ster mea­sure sep­a­ra­tion?”

But, in the mod­ern view of quan­tum me­chan­ics—which is ac­cepted by ev­ery­one ex­cept for a hand­ful of old fo­geys who may or may not still con­sti­tute a nu­mer­i­cal ma­jor­ity—well, as David Wal­lace puts it:

“If I were to pick one theme as cen­tral to the tan­gled de­vel­op­ment of the Everett in­ter­pre­ta­tion of quan­tum me­chan­ics, it would prob­a­bly be: the for­mal­ism is to be left alone.

De­co­her­ence is not an ex­tra phe­nomenon. De­co­her­ence is not some­thing that has to be pro­posed ad­di­tion­ally. There is no De­co­her­ence Pos­tu­late on top of stan­dard QM. It is im­plicit in the stan­dard rules. De­co­her­ence is just what hap­pens by de­fault, given the stan­dard quan­tum equa­tions, un­less the Quan­tum Spaghetti Mon­ster in­ter­venes.

Some still claim that the quan­tum equa­tions are un­real—a mere model that just hap­pens to give amaz­ingly good ex­per­i­men­tal pre­dic­tions. But then de­co­her­ence is what hap­pens to the par­ti­cles in the “un­real model”, if you ap­ply the rules uni­ver­sally and uniformly. It is deny­ing de­co­her­ence that re­quires you to pos­tu­late an ex­tra law of physics, or an act of the Quan­tum Spaghetti Mon­ster.

(Need­less to say, no one has ever ob­served a quan­tum sys­tem be­hav­ing co­her­ently, when the un­touched equa­tions say it should be de­co­her­ent; nor ob­served a quan­tum sys­tem be­hav­ing de­co­her­ently, when the un­touched equa­tions say it should be co­her­ent.)

If you’re talk­ing about any­thing that isn’t in the equa­tions, you must not be talk­ing about “de­co­her­ence”. The stan­dard equa­tions of QM, un­in­ter­preted, do not talk about a Quan­tum Spaghetti Mon­ster cre­at­ing new wor­lds. So if you ask when the Quan­tum Spaghetti Mon­ster cre­ates a new world, and you can’t an­swer the ques­tion just by look­ing at the equa­tions, then you must not be talk­ing about “de­co­her­ence”. QED.

Which ba­sis you use in your calcu­la­tions makes no differ­ence to stan­dard QM. “De­co­her­ence” is a phe­nomenon im­plicit in stan­dard QM. Which ba­sis you use makes no differ­ence to “de­co­her­ence”. QED.

Chang­ing your view of the con­figu­ra­tion space can change your view of the blobs of am­pli­tude, but ul­ti­mately the same phys­i­cal events hap­pen for the same causal rea­sons. Mo­men­tum ba­sis, po­si­tion ba­sis, po­si­tion ba­sis with a differ­ent rel­a­tivis­tic space of si­mul­tane­ity—it doesn’t mat­ter to QM, ergo it doesn’t mat­ter to de­co­her­ence.

If this were not so, you could do an ex­per­i­ment to find out which ba­sis was the right one! De­co­her­ence is an ex­per­i­men­tally visi­ble phe­nomenon—that’s why we have to pro­tect quan­tum com­put­ers from it.

Ah, but then where is the de­co­her­ence in

(0.5(Sen­sor-LEFT + Sen­sor-RIGHT) + 0.5(Sen­sor-LEFT—Sen­sor-RIGHT)) * (0.5(Atom-RIGHT + Atom-LEFT) − 0.5(Atom-RIGHT—Atom-LEFT)) + (0.5(Sen­sor-LEFT + Sen­sor-RIGHT) − 0.5(Sen­sor-LEFT—Sen­sor-RIGHT)) * (0.5(Atom-RIGHT + Atom-LEFT) + 0.5(Atom-RIGHT—Atom-LEFT))

?

The de­co­her­ence is still there. We’ve just made it harder for a hu­man to see, in the new rep­re­sen­ta­tion.

The main in­ter­est­ing fact I would point to, about this amaz­ing new rep­re­sen­ta­tion, is that we can no longer calcu­late its evolu­tion with lo­cal causal­ity. For a tech­ni­cal defi­ni­tion of what I mean by “causal­ity” or “lo­cal”, see Judea Pearl’s Causal­ity. Roughly, to com­pute the evolu­tion of an am­pli­tude cloud in a lo­cally causal ba­sis, each point in con­figu­ra­tion space only has to look at its in­finites­i­mal neigh­bor­hood to de­ter­mine its in­stan­ta­neous change. As I un­der­stand quan­tum physics—I pray to some physi­cist to cor­rect me if I’m wrong—the po­si­tion ba­sis is lo­cal in this sense.

(Note: It’s okay to pray to physi­cists, be­cause physi­cists ac­tu­ally ex­ist and can an­swer prayers.)

How­ever, once you start break­ing down the am­pli­tude dis­tri­bu­tion into com­po­nents like (Sen­sor-RIGHT—Sen­sor-LEFT), then the flow of am­pli­tude, and the flow of causal­ity, is no longer lo­cal within the new con­figu­ra­tion space. You can still calcu­late it, but you have to use non­lo­cal calcu­la­tions.

In essence, you’ve ob­scured the chess­board by sub­tract­ing the queen’s po­si­tion from the king’s po­si­tion. All the in­for­ma­tion is still there, but it’s harder to see.

When it comes to talk­ing about whether “de­co­her­ence” has oc­curred in the quan­tum state of a hu­man brain, what should in­tu­itively mat­ter is ques­tions like, “Does the event of a neu­ron firing in Hu­man-LEFT have a no­tice­able in­fluence on whether a cor­re­spond­ing neu­ron fires in Hu­man-RIGHT?” You can choose a ba­sis that will mix up the am­pli­tude for Hu­man-LEFT and Hu­man-RIGHT, in your calcu­la­tions. You can­not, how­ever, choose a ba­sis that makes a hu­man neu­ron fire when it would not oth­er­wise have fired; any more than you can choose a ba­sis that will pro­tect a quan­tum com­puter with­out the trou­ble of shield­ing, or choose a ba­sis that will make ap­ples fall up­ward in­stead of down, etcetera.

The for­mal­ism is to be left alone! If you’re talk­ing about any­thing that isn’t in the equa­tions, you’re not talk­ing about de­co­her­ence! De­co­her­ence is part of the in­var­i­ant essence that doesn’t change no mat­ter how you spin your ba­sis—just like the phys­i­cal re­al­ity of ap­ples and quan­tum com­put­ers and brains.

There may be a kind of Mind Pro­jec­tion Fal­lacy at work here. A ten­dency to see the ba­sis it­self as real—some­thing that a Quan­tum Spaghetti Mon­ster might come in and act upon—be­cause you spend so much time calcu­lat­ing with it.

In a strange way, I think, this sort of jump is ac­tively en­couraged by the Old School idea that the am­pli­tude dis­tri­bu­tions aren’t real. If you were told the am­pli­tude dis­tri­bu­tions were phys­i­cally real, you would (hope­fully) get in the habit of look­ing past mere rep­re­sen­ta­tions, to see through to some in­var­i­ant essence in­side—a re­al­ity that doesn’t change no mat­ter how you choose to rep­re­sent it.

But peo­ple are told the am­pli­tude dis­tri­bu­tion is not real. The calcu­la­tion it­self is all there is, and has no virtue save its mys­te­ri­ously ex­cel­lent ex­per­i­men­tal pre­dic­tions. And so there is no point in try­ing to see through the calcu­la­tions to some­thing within.

Then why not in­ter­pret all this talk of “de­co­her­ence” in terms of an ar­bi­trar­ily cho­sen ba­sis? Isn’t that all there is to in­ter­pret—the calcu­la­tion that you did in some rep­re­sen­ta­tion or an­other? Why not com­plain, if—hav­ing thus in­ter­preted de­co­her­ence—the sep­a­rat­ed­ness of am­pli­tude blobs seems to change, when you change the ba­sis? Why try to see through to the neu­rons, or the flows of causal­ity, when you’ve been told that the calcu­la­tions are all?

(This no­tion of see­ing through—look­ing for an essence, and not be­ing dis­tracted by sur­faces—is one that pops up again and again, and again and again and again, in the Way of Ra­tion­al­ity.)

Another pos­si­ble prob­lem is that the calcu­la­tions are crisp, but the essences in­side them are not. Write out an in­te­gral, and the sym­bols are digi­tally dis­tinct. But an en­tire ap­ple, or an en­tire brain, is larger than any­thing you can han­dle for­mally.

Yet the form of that crisp in­te­gral will change when you change your ba­sis; and that sloppy real essence will re­main in­var­i­ant. Re­for­mu­lat­ing your equa­tions won’t re­move a dag­ger, or silence a firing neu­ron, or shield a quan­tum com­puter from de­co­her­ence.

The phe­nomenon of de­co­her­ence within brains and sen­sors, may not be any more crisply defined than the brains and sen­sors them­selves. Brains, as high-level phe­nom­ena, don’t always make a clear ap­pear­ance in fun­da­men­tal equa­tions. Ap­ples aren’t crisp, you might say.

For his­tor­i­cal rea­sons, some Old School physi­cists are ac­cus­tomed to QM be­ing “in­ter­preted” us­ing ex­tra pos­tu­lates that in­volve crisp ac­tions by the Quan­tum Spaghetti Mon­ster—eat­ing blobs of am­pli­tude at a par­tic­u­lar in­stant, or cre­at­ing wor­lds as a par­tic­u­lar in­stant. Since the equa­tions aren’t sup­posed to be real, the sloppy bor­ders of real things are not looked for, and the crisp calcu­la­tions are pri­mary. This makes it hard to see through to a real (but un­crisp) phe­nomenon among real (but un­crisp) brains and ap­ples, in­var­i­ant un­der changes of crisp (but ar­bi­trary) rep­re­sen­ta­tion.

Like­wise, any change of rep­re­sen­ta­tion that makes ap­ples harder to see, or brains harder to see, will make de­co­her­ence within brains harder to see. But it won’t change the ap­ple, the brain, or the de­co­her­ence.

As always, any philo­soph­i­cal prob­lems that re­sult from “brain” or “per­son” or “con­scious­ness” not be­ing crisply defined, are not the re­spon­si­bil­ity of physi­cists or of any fun­da­men­tal phys­i­cal the­ory. Nor are they limited to de­co­her­ent quan­tum physics par­tic­u­larly, ap­pear­ing like­wise in split­ting brains con­structed un­der clas­si­cal physics, etcetera.

Com­ing to­mor­row (hope­fully): The Born Prob­a­bil­ities, aka, that mys­te­ri­ous thing we do with the squared mod­u­lus to get our ex­per­i­men­tal pre­dic­tions.

Next post: “The Born Prob­a­bil­ities

Pre­vi­ous post: “De­co­her­ence is Pointless

• I can’t say that I’ve un­der­stood ev­ery­thing in the se­ries on QM, but it has been im­mensely use­ful for me in be­gin­ning to un­der­stand it. And, in gen­eral, most of what I’ve read of OB I’ve found use­ful—es­pe­cially the com­ments, be­cause I find my­self “catch­ing up” with a lot of these ideas, and I have a ten­dency when I’m catch­ing up to the ideas of some­one more in­tel­li­gent than me to not find fault where I would if I had a bet­ter grasp of the ideas. Though I’m still not very far along on the path to be­ing a ra­tio­nal­ist, I know that it’s a path I’ve been try­ing to walk my whole life, de­spite the fact that much of it was spent stum­bling and trip­ping through re­li­gion, pop­u­lar poli­tics, and ar­gu­ments that were more about prov­ing who was “right” rather than find­ing out what was right. I’m glad to have found yet an­other re­source for walk­ing the path, es­pe­cially one as use­ful as this one. I haven’t com­mented here be­fore, but I just thought I’d toss in that I re­ally ap­pre­ci­ate the writ­ing you do here (and yours as well, Robin) and I’m glad that I stum­bled across this blog.

• Hrm… I won­der if there are other basies with lo­cal be­hav­ior other than the usual po­si­tional one?

If yes, does what we see as de­co­her­ence au­to­mat­i­claly “look de­co­her­ent” in that ba­sis too?

• Psy-Kosh: in the ba­sis of eigen­val­ues of the Hamil­to­nian, not only is the equa­tion lo­cal, but noth­ing even moves.

• Chris, for­give me if this is a fool­ish ques­tion, but wouldn’t the com­po­nents cor­re­spond­ing to eigen­val­ues of the Hamil­to­nian change only by a con­stant com­plex fac­tor, rather than not chang­ing at all?

• Chris, Eliezer: Yeah, at least last what I re­call study­ing, the time de­vel­op­ment for Hamil­to­nian eigen­vec­tors ba­si­cally has them spin­ning around the com­plex plane (with the rate of ro­ta­tion be­ing a func­tion of the eigen­value. In fact, I be­lieve it is di­rectly pro­por­tional)

• Ac­tu­ally, this dis­cus­sion leads me to won­der some­thing: What prop­er­ties does a ma­trix have to have such that its eigen­vec­tors form a com­plete ba­sis?

• The eigen­vec­tors of a ma­trix form a com­plete or­thog­o­nal ba­sis if and only if the ma­trix com­mutes with its Her­mi­tian con­ju­gate (i.e. the com­plex con­ju­gate of its trans­pose). Ma­tri­ces with this prop­erty are called “nor­mal”. Any Hamil­to­nian is Her­mi­tian: it is equal to its Her­mi­tian con­ju­gate. Any quan­tum time evolu­tion op­er­a­tor is uni­tary: its Her­mi­tian con­ju­gate is its in­verse. Any ma­trix com­mutes with it­self and its in­verse, so the eigen­vec­tors of any Hamil­to­nian or time evolu­tion op­er­a­tor will always form a com­plete or­thog­o­nal ba­sis. (I don’t re­mem­ber what the an­swer is if you don’t re­quire the ba­sis to be or­thog­o­nal.)

• It would be a plea­sure and a treat to join the re­cent dis­cus­sion on QM, es­pe­cially the Eb­bo­rian in­ter­lude, but I can­not af­ford the dozens of hours of study and re­flec­tion it would take to get to the point where I could ac­tu­ally con­tribute to the dis­cus­sion.

If I ever find my­self with the lux­ury of be­ing able to study QM, this blog or the book that comes from it is where I would go first for writ­ten study ma­te­rial. (I’d prob­a­bly need a re­li­able math­e­mat­i­cal treat­ment, too, but those are easy to find.)

Physics is the study of ul­ti­mate re­al­ity!

QM is in my hum­ble opinion hu­mankind’s great­est achieve­ment.

• Stephen: Thanks!

And yeah, I was won­der­ing what the an­swer was if I don’t nec­es­sar­ally de­mand them to be or­thog­nal, just that I re­quire them to span the space.

Any­ways, am right now read­ing through Down with Deter­mi­nants. Maybe that’ll have the an­swer in there.

(Ac­tu­ally, the part which I get to, at least for finite di­men­sional spaces, is already effec­tively in there: The num­ber of dis­tinct eigen­val­ues has to equal the di­men­sion of the space. Of course, the ques­tion of what has to be true about a lin­ear op­er­a­tor for that to hold is some­thing I’m won­der­ing. :))

• “The num­ber of dis­tinct eigen­val­ues has to equal the di­men­sion of the space.”

That may be a suffi­cient con­di­tion but it is definitely not a nec­es­sary one. The iden­tity ma­trix has only one eigen­value, but it has a set of eigen­vec­tors that span the space.

• Stephen: whoops. Just re­al­ized that and came here to post that cor­rec­tion, and you already did. :)

• I don’t re­ally fol­low a lot of what you’ve writ­ten on this, so maybe this isn’t fair, but I’ll put it out there any­way:

I have a hard time see­ing much differ­ence be­tween you (Eliezer Yud­kowsky) and the peo­ple you keep de­scribing as wrong. They don’t look be­yond the sur­face, you look be­yond it and see some­thing that looks just like the sur­face (or the sur­face that’s eas­iest to look at). They layer mys­te­ri­ous things on top of the the­ory to ex­plain it, you layer mys­te­ri­ous things on top of physics to ex­plain it. Their ex­pla­na­tions all have fatal flaws, yours has just one se­ri­ous prob­lem. Their ex­pla­na­tions don’t ac­tu­ally ex­plain any­thing, yours re­names things (e.g. prob­a­bil­ity be­comes “sub­jec­tive ex­pec­ta­tion”) with­out clear­ing up the cause of their re­la­tion­ships—at least, not yet.

• Psy-Kosh, Stephen: A finite-di­men­sional com­plex ma­trix has a com­plete ba­sis of eigen­vec­tors (i.e. it is di­ag­o­nal­iz­able) if and only if ev­ery gen­er­al­ized eigen­vec­tor is also an eigen­vec­tor. In­tu­itively, this means roughly that there are n in­de­pen­dent di­rec­tions (where n is the size of the ma­trix) such that vec­tors along these di­rec­tions are stretched or shrunk uniformly by the ma­trix.

Try googling “jor­dan nor­mal form”, that may help clar­ify the situ­a­tion.

I don’t know the an­swer in the in­finite-di­men­sional case.

• Roland, make that two. Though this mooshed my head.

I used to re­ally en­joy think­ing about how weird QM was. Look! The lit­tle pho­ton goes through both holes at the same time! Not re­ally any more though, it’s start­ing to seem a lit­tle bit...or­di­nary.

Which is a good thing, of course.

Quick ques­tion—since you can’t in­te­grate over a sin­gle point, does that pre­clude the ex­is­tence of any ‘mo­tion­less’ par­ti­cle? Any­thing that ceased to have an ap­pre­cia­ble (Planck-length?) am­pli­tude spread would, in effect, not be there? That would chime with the trans­form du­al­ity thingy be­tween lo­ca­tion and ve­loc­ity.

Hope I get chat­ting to some­one who thinks in terms of quan­tum/​clas­si­cal du­al­ities at some point, purely so that I can use the line “you’re very clever, old man, but it’s all am­pli­tudes, all the time.”

• It’s odd that the QM se­quence is so lit­tle com­mented-on and voted-on, which sug­gests it’s lit­tle-read. Which is par­tic­u­larly strange in that so much of EY’s philos­o­phy ap­pears to build di­rectly on his in­ter­pre­ta­tion of QM. Does any­one have ideas on why? Are peo­ple just read­ing the head­lines and go­ing along with what they seem to say, and not read­ing the posts them­selves and par­tic­u­larly not their com­ments?

• so much of EY’s philos­o­phy ap­pears to build di­rectly on his in­ter­pre­ta­tion of QM.

Is this re­ally the case? It seems to me that that the in­ter­pre­ta­tion of QM (and al­most all micro-level de­tails of fun­da­men­tal physics) ought to be (and in Eliezer’s case, are) in­de­pen­dent of “macro-level” philos­o­phy. Eliezer could jus­tify his re­duc­tion­ism, his Bayesi­anism, his util­i­tar­ian ethics, his athe­ism, his op­po­si­tion to most kinds of moral dis­count­ing, his in­tu­itions re­gard­ing de­ci­sion the­ory, his mod­els of mind and of lan­guage, and his fu­tur­ism—he could jus­tify all these things even if he were a strict New­to­nian be­liever in sim­ple de­ter­minism who mod­els all ap­par­ent in­de­ter­mi­nacy as ig­no­rance of the true ini­tial con­di­tions.

To my mind, the micro as­sump­tions don’t change the macro con­clu­sions, they only change the way we talk about and jus­tify them.

• To my mind, the micro as­sump­tions don’t change the macro con­clu­sions, they only change the way we talk about and jus­tify them.

I agree with you that one should reach most if not all of the same con­clu­sions from a strict New­to­nian per­spec­tive (or from a Copen­hagen­ite per­spec­tive, and so on). But the way it’s talked about does scare me, be­cause it’s difficult for me to tell why they be­lieve the things they be­lieve, and opaque rea­son­ing rings sev­eral warn­ing bells.

That is, to an­swer your origi­nal ques­tion- “Is this re­ally the case?”- it cer­tainly is the case that it ap­pears that EY’s philos­o­phy builds di­rectly on his in­ter­pre­ta­tion of QM. When judg­ing by ap­pear­ances, we have to take the lan­guage into ac­count, and to go deeper re­quires that you go down the rab­bit hole to tell whether or not EY’s philos­o­phy ac­tu­ally re­quires those things- and that rab­bit hole is one that is for­bid­ding for non-math­e­mat­i­ci­ans and oddly dis­quiet­ing for physi­cists (at least, that’s my im­pres­sion as a physi­cist). QM is an in­fer­en­tial dis­tance minefield.

It seems to me that MWI is just a con­ve­nient vi­su­al­iza­tion trick, and thus there is equiv­alence, but I don’t feel I un­der­stand EY’s philos­o­phy and its de­vel­op­ment well enough to ar­gue for that in­ter­pre­ta­tion.

• Agree. It would be nice to have Eliezer’s take on this ques­tion.

• And then there’s Time­less Iden­tity, which ex­pressly claims to be the philo­soph­i­cal pay­off from the QM se­quence. Given that post and the in­tro­duc­tion I quoted from Quan­tum Ex­pla­na­tions, I re­ally don’t see how you can deny that his philos­o­phy builds di­rectly on his in­ter­pre­ta­tion of QM.

• It ap­pears you are right. Eliezer de­rives his con­clu­sions re­gard­ing zom­bies, per­sonal iden­tity, and the philos­o­phy of trans­porters and du­pli­ca­tors from his un­der­stand­ing of QM.

On the other hand, I reach ex­actly the same con­clu­sions on these is­sues with­out re­ally un­der­stand­ing QM. Of course, I have the ad­van­tage over Eliezer that I have read far less Philos­o­phy. :)

• Peo­ple shouldn’t build too much of their philos­o­phy on top of the MWI, IMO. If ev­i­dence that rel­a­tively “dis­tant” wor­lds are be­ing deleted is found then they would have to re­visit it all. That doesn’t seem ter­ribly likely—but we can hardly rule it out. Oc­cam’s ra­zor just doesn’t rule against it that strongly.

• On the other hand, I reach ex­actly the same con­clu­sions on these is­sues with­out re­ally un­der­stand­ing QM. Of course, I have the ad­van­tage over Eliezer that I have read far less Philos­o­phy.

Love the Philos­o­phy jibe! :)

• On the other hand, I reach ex­actly the same con­clu­sions on these is­sues with­out re­ally un­der­stand­ing QM.

Hah, same here.

• Well, ISTM that this sort of re­duc­tion­ism/​func­tion­al­ism is still right in a clas­si­cal uni­verse, just go­ing by the whole no­tion of be­liefs should pay rent; but it’s not forced like it is in the ac­tual uni­verse.

• That’s as I un­der­stand it, too. How­ever, I think that he also means that QM gives some ad­di­tional ev­i­dence that con­scious­ness is not sub­strate-de­pen­dent, as for in­stance Mas­simo Pigliucci meant in the Blog­ging­heads.TV dis­cus­sion, be­cause given QM there is no unique time-con­tin­u­ous neu­ron-num­ber-124 in brain-234 etc. etc. at all. Only func­tions.

For a dis­cus­sion of ems this helps. Pigliucci on the other hand meant sub­strate-in­de­pen­dence would im­ply a du­al­ism. What left me some­how puz­zling as he seemed to ac­cept that there is more than one con­scious­ness in the uni­verse, but now I start drift­ing off...

• so much of EY’s philos­o­phy ap­pears to build di­rectly on his in­ter­pre­ta­tion of QM.

Is this re­ally the case?

I got this from Quan­tum Ex­pla­na­tions:

I think I must now tem­porar­ily digress from the se­quence on zom­bies (which was a di­gres­sion from the dis­cus­sion of re­duc­tion­ism, which was a di­gres­sion from the Mind Pro­jec­tion Fal­lacy) in or­der to dis­cuss quan­tum me­chan­ics. The rea­sons why this be­longs in the mid­dle of a dis­cus­sion on zom­bies in the mid­dle of a dis­cus­sion of re­duc­tion­ism in the mid­dle of a dis­cus­sion of the Mind Pro­jec­tion Fal­lacy, will be­come ap­par­ent even­tu­ally.

That is, Eliezer brought QM up at all as part of a philo­soph­i­cal dis­cus­sion, be­cause he felt he had to in or­der to make his philo­soph­i­cal points. You may then ar­gue (as you seem to in your com­ment) that he did not in fact have to bring in QM to make his points, but he felt he had to, per that quote.

• The QM se­quence was origi­nally posted at over­com­ing bias, and was later posted here when LW was cre­ated. That ex­plains its lack of com­ments and votes rel­a­tive to posts made here origi­nally. How­ever, if there’s a lack of com­ments and posts rel­a­tive to other parts of the se­quences (which were al­most all origi­nally posted at over­com­ing bias), then you’ve no­ticed some­thing.

If this puz­zle ex­ists, I’d guess many peo­ple didn’t read them be­cause they were turned off by the math early on in the se­quence.

• The QM se­quence is also linked to a lot less than other posts, as it tends to be less di­rectly rele­vant to con­ver­sa­tion top­ics.

• I’ve been plough­ing through the se­quences in my idle read­ing time, more or less in wiki or­der, and yes, these have no­tice­ably less votes and com­ments than other se­quences. The QM se­quence is the only place I’ve seen EY posts with votes of 0 or even −1. This sug­gests to me a lot less read­ers. (Per­haps dis­play­ing up and down to­tals, as per Red­dit, would help dis­t­in­guish “con­tro­ver­sial” from “no­body cares”.)

• Con­tro­ver­sial is a de­cent pos­si­bil­ity. What EY says IS con­tro­ver­sial among physi­cists, and that may be the source of some of his down­votes.

• The lack of com­ments com­pared to other se­quences doesn’t fit that, though.

• A tech­ni­cal sub­ject. The gist seemed to be: Rah, MWI.

I’ve thought the MWI was cor­rect since way back in the 1980s—af­ter read­ing this—and so didn’t feel an ur­gent need to be lec­tured on its virtues.

• “No one had ever thought of de­co­her­ence. The ques­tion of why a hu­man re­searcher only saw one thing at a time, was a Great Mys­tery with no ob­vi­ous an­swer.”

This is not true, and say­ing things like this will re­duce your cred­i­bil­ity in the eyes of in­tel­li­gent ob­servers. In “The Pre­sent State of Quan­tum Me­chan­ics” Schroed­inger writes

As we thus con­struct an ob­jec­tive pic­ture of this pro­cess, like that of any other, we dare hope to clear up, if not al­to­gether avoid, the sin­gu­lar jump of the psi-func­tion [...] it would not be quite right to say that the psi-func­tion of the ob­ject which changes oth­er­wise ac­cord­ing to a par­tial differ­en­tial equa­tion, in­de­pen­dent of the ob­server, should now change leap-fash­ion be­cause of a men­tal act.

(This is in trans­la­tion, but I don’t think you can deny in good faith that he un­der­stands de­co­her­ence and al­most cer­tainly grasps the pre­dicted ex­is­tence of many wor­lds).

From the form in which the psi-func­tion was last known, to the new in which it reap­pears, runs no con­tin­u­ous road—it ran in­deed through an­nihila­tion. Con­trast­ing the two forms, the thing looks like a leap. In truth some­thing of im­por­tance hap­pens in be­tween, namely the in­fluence of the two bod­ies on each other, dur­ing which the ob­ject pos­sessed no pri­vate ex­pec­ta­tion-cat­a­log nor had any claim there­unto, be­cause it was not in­de­pen­dent.

You should con­sider chang­ing the way you talk about the his­tory of quan­tum me­chan­ics (and prob­a­bly learn­ing more about the his­tory) be­fore writ­ing at more length about it.

• I think he is ex­press­ing dis­satis­fac­tion with QM rather than en­dors­ing MWI. I found a differ­ent quote, from 1950, that seems to sup­port the former.

For it is just be­cause they pro­hibit our ask­ing what re­ally “is”, that is, which state of af­fairs re­ally oc­curs in the in­di­vi­d­ual case, that the pos­i­tivists suc­ceed in mak­ing us set­tle for a kind of col­lec­tive de­scrip­tion. They ac­cuse us of meta­phys­i­cal heresy if we want to ad­here to this “re­al­ity”. . . . The pre­sent quan­tum me­chan­ics sup­plies no equiv­a­lent. It is not con­scious of the prob­lem at all; it passes it by with blithe dis­in­ter­est.

That isn’t that clear a state­ment of his views, but it is from a let­ter writ­ten in re­ply to Ein­stein, who said

I am as con­vinced as ever that the wave rep­re­sen­ta­tion of mat­ter is an in­com­plete rep­re­sen­ta­tion of the state of af­fairs, no mat­ter how prac­ti­cally use­ful it has proved it­self to be. The pret­tiest way to show this is by your ex­am­ple with the cat. . . .

If one at­tempts to in­ter­pret the ψ-func­tion as a com­plete de­scrip­tion of a state, in­de­pen­dent of whether or not it is ob­served, then this means that at the time in ques­tion the cat is nei­ther al­ive nor pul­ver­ized. But one or the other situ­a­tion would be re­al­ized by mak­ing an ob­ser­va­tion.

(Both quotes are taken from Karl Prz­ibram’s Let­ters on wave me­chan­ics: Schrod­inger, Planck, Ein­stein, Lorentz p. 35-38.)

This is clearly against quan­tum me­chan­ics rather in sup­port of MWI. They both re­al­ize that QM’s on­tol­ogy needs to be re­vised, but nei­ther knows how.

• Well, that’s an in­ter­est­ing quote, but did he come out and say that QM was all there was, no ex­cep­tions ever, and col­lapse is not real? If he did, it was in pri­vate and did not spread, for when Everett (re-?)pro­posed it later, it was ex­ceed­ingly con­tro­ver­sial and de­rided.

And cer­tainly de­co­her­ence is a con­sid­er­ably more com­pli­cated beast than that, and sim­ply the no­tion that QM is all there re­ally is NOT suffi­cient to un­der­stand de­co­her­ence, not by a long shot.

• Well, that’s an in­ter­est­ing quote, but did he come out and say that QM was all there was, no ex­cep­tions ever, and col­lapse is not real?

Yes. He said it in the pas­sage I quoted. (“it would not be quite right to say that the psi-func­tion of the ob­ject...should now change leap-fash­ion be­cause of a men­tal act.” You could quib­ble with the word ‘quite,’ but I think the sur­round­ing text is plenty clear.) His un­der­stand­ing comes through in his writ­ing more gen­er­ally. The fact that one per­son has un­der­stood some­thing (or many) does not pre­clude it from be­ing con­tro­ver­sial some time later.

And cer­tainly de­co­her­ence is a con­sid­er­ably more com­pli­cated beast than that, and sim­ply the no­tion that QM is all there re­ally is NOT suffi­cient to un­der­stand de­co­her­ence, not by a long shot.

I don’t know quite what you mean. In what way is de­co­her­ence “more com­pli­cated,” and than what? It looks to me like Schrod­inger un­der­stands ex­actly what is go­ing on.