Quantum Non-Realism

“Does the moon ex­ist when no one is look­ing at it?”
—Albert Ein­stein, asked of Niels Bohr

Sup­pose you were just start­ing to work out a the­ory of quan­tum me­chan­ics.

You be­gin to en­counter ex­per­i­ments that de­liver differ­ent re­sults de­pend­ing on how closely you ob­serve them. You dig un­der­neath the re­al­ity you know, and find an ex­tremely pre­cise math­e­mat­i­cal de­scrip­tion that only gives you the rel­a­tive fre­quency of out­comes; worse, it’s made of com­plex num­bers. Things be­have like par­ti­cles on Mon­day and waves on Tues­day.

The cor­rect an­swer is not available to you as a hy­poth­e­sis, be­cause it will not be in­vented for an­other thirty years.

In a mess like that, what’s the best you could do?

The best you can do is the strict “shut up and calcu­late” in­ter­pre­ta­tion of quan­tum me­chan­ics. You’ll go on try­ing to de­velop new the­o­ries, be­cause do­ing your best doesn’t mean giv­ing up. But we’ve speci­fied that the cor­rect an­swer won’t be available for thirty years, and that means none of the new the­o­ries will re­ally be any good. Do­ing the best you could the­o­ret­i­cally do would mean that you rec­og­nized that, even as you looked for ways to test the hy­pothe­ses.

The best you could the­o­ret­i­cally do would not in­clude say­ing any­thing like, “The wave­func­tion only gives us prob­a­bil­ities, not cer­tain­ties.” That, in ret­ro­spect, was jump­ing to a con­clu­sion; the wave­func­tion gives us a cer­tainty of many wor­lds ex­ist­ing. So that part about the wave­func­tion be­ing only a prob­a­bil­ity was not-quite-right. You calcu­lated, but failed to shut up.

If you do the best that you can do with­out the cor­rect an­swer be­ing available, then, when you hear about de­co­her­ence, it will turn out that you have not said any­thing­in­com­pat­i­ble with de­co­her­ence. De­co­her­ence is not ruled out by the data and the calcu­la­tions. So if you re­fuse to af­firm, as pos­i­tive knowl­edge, any propo­si­tion which was not forced by the data and the calcu­la­tions, the calcu­la­tions will not force you to say any­thing in­com­pat­i­ble with de­co­her­ence. So too with what­ever the cor­rect the­ory may be, if it is not de­co­her­ence. If you go astray, it must be from your own im­pulses.

But it is hard for hu­man be­ings to shut up and calcu­late—re­ally shut up and calcu­late. There is an over­whelming ten­dency to treat our ig­no­rance as if it were pos­i­tive knowl­edge.

I don’t know if any con­ver­sa­tions like this ever re­ally took place, but this is how ig­no­rance be­comes knowl­edge:

Gal­lant: “Shut up and calcu­late.”
Goofus: “Why?”
Gal­lant: “Be­cause I don’t know what these equa­tions mean, just that they seem to work.”
five min­utes later
Goofus: “Shut up (−1/​3)i and calcu­late.”
Stu­dent: “Why?”
Goofus: “Be­cause these equa­tions don’t mean any­thing, they just work.”
Stu­dent: “Really? How do you know?”
Goofus: “Gal­lant told me.”

A similar trans­for­ma­tion oc­curs in the leap from:

Gal­lant: “When my calcu­la­tions show an am­pli­tude of for this pho­ton to get ab­sorbed, my ex­per­i­ments showed that the pho­ton was ab­sorbed around 107 times out of 1,000, which is a good fit to the square of the mod­u­lus. There’s clearly some kind of con­nec­tion be­tween the ex­per­i­men­tal statis­tics and the squared mod­u­lus of the am­pli­tude, but I don’t know what.”
Goofus: “The prob­a­bil­ity am­pli­tude doesn’t say where the elec­tron is, but where it might be. The squared mod­u­lus is the prob­a­bil­ity that re­al­ity will turn out that way. Real­ity it­self is in­her­ently non­de­ter­minis­tic.”

And again:

Gal­lant: “Once I mea­sure some­thing and get an ex­per­i­men­tal re­sult, I do my fu­ture calcu­la­tions us­ing only the am­pli­tude whose squared mod­u­lus went into calcu­lat­ing the fre­quency of that ex­per­i­men­tal re­sult. Only this rule makes my fur­ther calcu­la­tions cor­re­spond to ob­served fre­quen­cies.”
Goofus: “Since the am­pli­tude is the prob­a­bil­ity, once you know the ex­per­i­men­tal re­sult, the prob­a­bil­ity of ev­ery­thing else be­comes zero!”

The whole slip from:

The square of this “am­pli­tude” stuff cor­re­sponds tightly to our ex­per­i­men­tally ob­served frequencies


The am­pli­tude is the prob­a­bil­ity of get­ting the measurement


Well, ob­vi­ously, once you know you didn’t get a mea­sure­ment, its prob­a­bil­ity be­comes zero

has got to be one of the most em­bar­rass­ing wrong turns in the his­tory of sci­ence.

If you take all this liter­ally, it be­comes the con­scious­ness-causes-col­lapse in­ter­pre­ta­tion of quan­tum me­chan­ics. Th­ese days, just about no­body will con­fess to ac­tu­ally be­liev­ing in the con­scious­ness-causes-col­lapse in­ter­pre­ta­tion of quan­tum me­chan­ics—

But the physics text­books are still writ­ten this way! Peo­ple say they don’t be­lieve it, but they talk as if knowl­edge is re­spon­si­ble for re­mov­ing in­com­pat­i­ble “prob­a­bil­ity” am­pli­tudes.

Yet as im­plau­si­ble as I find con­scious­ness-causes-col­lapse, it at least gives us a pic­ture of re­al­ity. Sure, it’s an in­for­mal pic­ture. Sure, it gives men­tal prop­er­ties on­tolog­i­cally ba­sic sta­tus. You can’t calcu­late when an “ex­per­i­men­tal ob­ser­va­tion” oc­curs or what peo­ple “know,” you just know when cer­tain prob­a­bil­ities are ob­vi­ous­lyzero. And this “just know­ing” just hap­pens to fit your ex­per­i­men­tal re­sults, what­ever they are—

—but at least con­scious­ness-causes-col­lapse pur­ports to tell us how the uni­verse works. The am­pli­tudes are real, the col­lapse is real, the con­scious­ness is real.

Con­trast to this ar­gu­ment schema:

Stu­dent: “Wait, you’re say­ing that this am­pli­tude dis­ap­pears as soon as the mea­sure­ment tells me it’s not true?”
Goofus: “No, no! It doesn’t liter­ally dis­ap­pear. The equa­tions don’t mean any­thing—they just give good pre­dic­tions.”
Stu­dent: “But then what does hap­pen?”
Goofus: (Whor­ble. Hiss.) “Never ask that ques­tion.”
Stu­dent: “And what about the part where we mea­sure this pho­ton’s po­lariza­tion over here, and a light-year away, the en­tan­gled pho­ton’s prob­a­bil­ity of be­ing po­larized up-down changes from 50% to 25%?”
Goofus: “Yes, what about it?”
Stu­dent: “Doesn’t that vi­o­late Spe­cial Rel­a­tivity?”
Goofus: “No, be­cause you’re just find­ing out the other pho­ton’s po­lariza­tion. Re­mem­ber, the am­pli­tudes aren’t real.”
Stu­dent: “But Bell’s The­o­rem shows there’s no pos­si­ble lo­cal hid­den vari­able that could de­scribe the other pho­ton’s po­lariza­tion be­fore we mea­sure it—”
Goofus: “Ex­actly! It’s mean­ingless to talk about the pho­ton’s po­lariza­tion be­fore we mea­sure it.”
Stu­dent: “But the prob­a­bil­ity sud­denly changes—”
Goofus: “It’s mean­ingless to talk about it be­fore we mea­sure it!”

What does Goofus even mean, here? Never mind the plau­si­bil­ity of his words; what sort of state of re­al­ity would cor­re­spond to his words be­ing true?

What way could re­al­ity be, that would make it mean­ingless to talk about Spe­cial Rel­a­tivity be­ing vi­o­lated, be­cause the prop­erty be­ing in­fluenced didn’t ex­ist, even though you could calcu­late the changes to it?

But you know what? For­get that. I want to know the an­swer to an even more im­por­tant ques­tion:

Where is Goofus get­ting all this stuff?

Let’s sup­pose that you take the Schröd­inger equa­tion, and as­sert, as a pos­i­tive fact:

This equa­tion gen­er­ates good pre­dic­tions, but it doesn’t mean any­thing!

Really? How do you know?

I some­times go around say­ing that the fun­da­men­tal ques­tion of ra­tio­nal­ity is Why do you be­lieve what you be­lieve?

You say the Schröd­inger equa­tion “doesn’t mean any­thing.” How did this item of definite knowl­edge end up in your pos­ses­sion, if it is not sim­ply ig­no­rance mis­in­ter­preted as knowl­edge?

Was there some ex­per­i­ment that told you? I am open to the idea that ex­per­i­ments can tell us things that seem philo­soph­i­cally im­pos­si­ble. But in this case I should like to see the de­ci­sive data. Was there a point where you care­fully set up an ex­per­i­men­tal ap­para­tus, and worked out what you should ex­pect to see if (1) the Schröd­inger equa­tion was mean­ingful or (2) the Schröd­inger equa­tion was mean­ingless; and then you got re­sult (2)?

Gal­lant: “If I mea­sure the 90° po­lariza­tion of a pho­ton, and then mea­sure the 45° po­lariza­tion, and then mea­sure 90° again, my ex­per­i­men­tal his­tory shows that in 100 tri­als a pho­ton was ab­sorbed 47 times and trans­mit­ted 53 times.”
Goofus: “The 90° po­lariza­tion and 45° po­lariza­tion are in­com­pat­i­ble prop­er­ties; they can’t both ex­ist at the same time, and if you mea­sure one, it is mean­ingless to talk about the other.”

How do you know?

How did you ac­quire that piece of knowl­edge, Goofus? I know where Gal­lant got his—but where did yours come from?

My at­ti­tude to­ward ques­tions of ex­is­tence and mean­ing was nicely illus­trated in a dis­cus­sion of the cur­rent state of ev­i­dence for whether the uni­verse is spa­tially finite or spa­tially in­finite, in which James D. Miller chided Robin Han­son:

Robin, you are suffer­ing from over­con­fi­dence bias in as­sum­ing that the uni­verse ex­ists. Surely there is some chance that the uni­verse is of size zero.

To which I replied:

James, if the uni­verse doesn’t ex­ist, it would still be nice to know whether it’s an in­finite or a finite uni­verse that doesn’t ex­ist.

Ha! You think pul­ling that old “uni­verse doesn’t ex­ist” trick will stop me? It won’t even slow me down!

It’s not that I’m rul­ing out the pos­si­bil­ity that the uni­verse doesn’t ex­ist. It’s just that, even if noth­ing ex­ists, I still want to un­der­stand the noth­ing as best I can. My cu­ri­os­ity doesn’t sud­denly go away just be­cause there’s no re­al­ity, you know!

The na­ture of “re­al­ity” is some­thing about which I’m still con­fused, which leaves open the pos­si­bil­ity that there isn’t any such thing. But Egan’s Law still ap­plies: “It all adds up to nor­mal­ity.” Ap­ples didn’t stop fal­ling when Ein­stein dis­proved New­ton’s the­ory of grav­ity.

Sure, when the dust set­tles, it could turn out that ap­ples don’t ex­ist, Earth doesn’t ex­ist, re­al­ity doesn’t ex­ist. But the nonex­is­tent ap­ples will still fall to­ward the nonex­is­tent ground at a mean­ingless rate of 9.8 m/​s2.

You say the uni­verse doesn’t ex­ist? Fine, sup­pose I be­lieve that—though it’s not clear what I’m sup­posed to be­lieve, aside from re­peat­ing the words.

Now, what hap­pens if I press this but­ton?

In The Sim­ple Truth, I said:

Frankly, I’m not en­tirely sure my­self where this “re­al­ity” busi­ness comes from. I can’t cre­ate my own re­al­ity in the lab, so I must not un­der­stand it yet. But oc­ca­sion­ally I be­lieve strongly that some­thing is go­ing to hap­pen, and then some­thing else hap­pens in­stead… So I need differ­ent names for the thin­gies that de­ter­mine my pre­dic­tions and the thingy that de­ter­mines my ex­per­i­men­tal re­sults. I call the former thin­gies “be­lief,” and the lat­ter thingy “re­al­ity.”

You want to say that the quan­tum-me­chan­i­cal equa­tions are “not real”? I’ll be char­i­ta­ble, and sup­pose this means some­thing. What might it mean?

Maybe it means the equa­tions which de­ter­mine my pre­dic­tions are sub­stan­tially differ­ent from the thingy that de­ter­mines my ex­per­i­men­tal re­sults. Then what does­de­ter­mine my ex­per­i­men­tal re­sults? If you tell me “noth­ing,” I would like to know what sort of “noth­ing” it is, and why this “noth­ing” ex­hibits such ap­par­ent reg­u­lar­ity in de­ter­min­ing e.g. my ex­per­i­men­tal mea­sure­ments of the mass of an elec­tron.

I don’t take well to peo­ple who tell me to stop ask­ing ques­tions. If you tell me some­thing is definitely pos­i­tively mean­ingless, I want to know ex­actly what you mean by that, and how you came to know. Other­wise you have not given me an an­swer, only told me to stop ask­ing the ques­tion.

The Sim­ple Truth de­scribes the life of a shep­herd and ap­pren­tice who have dis­cov­ered how to count sheep by toss­ing peb­bles into buck­ets, when they are vis­ited by a del­e­gate from the court who wants to know how the “magic peb­bles” work. The shep­herd tries to ex­plain, “An empty bucket is mag­i­cal if and only if the pas­tures are empty of sheep,” but is soon over­taken by the ex­cited dis­cus­sions of the ap­pren­tice and the del­e­gate as to how the magic might get into the peb­bles.

Here we have quan­tum equa­tions that de­liver ex­cel­lent ex­per­i­men­tal pre­dic­tions. What ex­actly does it mean for them to be “mean­ingless”? Is it like a bucket of peb­bles that works for count­ing sheep, but doesn’t have any magic?

Back be­fore Bell’s The­o­rem ruled out lo­cal hid­den vari­ables, it seemed pos­si­ble that (as Ein­stein thought) there was some more com­plete de­scrip­tion of re­al­ity which we didn’t have, and the quan­tum the­ory sum­ma­rized in­com­plete knowl­edge of this more com­plete de­scrip­tion. The laws we’d learned would turn out to be like the laws of statis­ti­cal me­chan­ics: quan­ti­ta­tive state­ments of un­cer­tainty. This would hardly make the equa­tions “mean­ingless”; par­tial knowl­edge is the mean­ing of prob­a­bil­ity.

But Bell’s The­o­rem makes it much less plau­si­ble that the quan­tum equa­tions are par­tial knowl­edge of some­thing de­ter­minis­tic, the way that statis­ti­cal me­chan­ics over clas­si­cal physics is par­tial knowl­edge of some­thing de­ter­minis­tic. And even so, the quan­tum equa­tions would not be “mean­ingless” as that phrase is usu­ally taken; they would be “statis­ti­cal,” “ap­prox­i­mate,” “par­tial in­for­ma­tion,” or at worst “wrong.”

Here we have equa­tions that give us ex­cel­lent pre­dic­tions. You say they are “mean­ingless.” I ask what it is that de­ter­mines my ex­per­i­men­tal re­sults, then. You can­not an­swer. Fine, then how do you jus­tify rul­ing out the pos­si­bil­ity that the quan­tum equa­tions give such ex­cel­lent pre­dic­tions be­cause they are, oh, say, mean­ingful?

I don’t mean to triv­ial­ize ques­tions of re­al­ity or mean­ing. But to call some­thing “mean­ingless” and say that the ar­gu­ment is now re­solved, finished, over, done with, you must have a the­ory of ex­actly how it is mean­ingless. And when the an­swer is given, the ques­tion should seem no longer mys­te­ri­ous.

As you may re­call from Se­man­tic Stop­signs, there are words and phrases which are not so much an­swers to ques­tions, as cog­ni­tive traf­fic sig­nals which in­di­cate you should stop ask­ing ques­tions. “Why does any­thing ex­ist in the first place? God!” is the clas­si­cal ex­am­ple, but there are oth­ers, such as “Élan vi­tal!”

Tell peo­ple to “shut up and calcu­late” be­cause you don’t know what the calcu­la­tions mean, and in­side of five years, “Shut up!” will be mas­querad­ing as a pos­i­tive the­ory of quan­tum me­chan­ics.

I have the high­est re­spect for any his­tor­i­cal physi­cists who even came close to ac­tu­ally shut­ting up and calcu­lat­ing, who were gen­uinely con­ser­va­tive in as­sess­ing what they did and didn’t know. This is the best they could pos­si­bly do with­out ac­tu­ally be­ing Hugh Everett, and I award them fifty ra­tio­nal­ity points. My scorn is re­served for those who in­ter­preted “We don’t know why it works” as the pos­i­tive knowl­edge that the equa­tions were definitely not real.

I mean, if that trick worked, it would be too good to con­fine to one sub­field. Why shouldn’t physi­cists use the “not real” loop­hole out­side of quan­tum me­chan­ics?

“Hey, doesn’t your new ‘yarn the­ory’ vi­o­late Spe­cial Rel­a­tivity?”
“Nah, the equa­tions are mean­ingless. Say, doesn’t your model of ‘chaotic evil in­fla­tion’ vi­o­late CPT sym­me­try?”
“My equa­tions are even more mean­ingless than your equa­tions! So your crit­i­cism dou­ble doesn’t count.”

And if that doesn’t work, try writ­ing your­self a Get Out of Jail Free card.

If there is a moral to the whole story, it is the moral of how very hard it is to stay in a state of con­fessed con­fu­sion, with­out mak­ing up a story that gives you clo­sure—how hard it is to avoid ma­nipu­lat­ing your ig­no­rance as if it were definite knowl­edge that you pos­sessed.