Timeless Control

Fol­lowup to: Time­less Physics, Time­less Causal­ity, Thou Art Physics

Peo­ple hear about many-wor­lds, which is de­ter­minis­tic, or about time­less physics, and ask:

If the fu­ture is de­ter­mined by physics, how can any­one con­trol it?

In Thou Art Physics, I pointed out that since you are within physics, any­thing you con­trol is nec­es­sar­ily con­trol­led by physics. To­day we will talk about a differ­ent as­pect of the con­fu­sion, the words “de­ter­mined” and “con­trol”.

The “Block Uni­verse” is the clas­si­cal term for the uni­verse con­sid­ered from out­side Time. Even with­out time­less physics, Spe­cial Rel­a­tivity out­laws any global space of si­mul­tane­ity, which is widely be­lieved to sug­gest the Block Uni­verse—space­time as one vast 4D block.

When you take a per­spec­tive out­side time, you have to be care­ful not to let your old, time­ful in­tu­itions run wild in the ab­sence of their sub­ject mat­ter.

In the Block Uni­verse, the fu­ture is not de­ter­mined be­fore you make your choice. “Be­fore” is a time­ful word. Once you de­scend so far as to start talk­ing about time, then, of course, the fu­ture comes “af­ter” the past, not “be­fore” it.

If we’re go­ing to take a time­less per­spec­tive, then the past and the fu­ture have not always been there. The Block Uni­verse is not some­thing that hangs, mo­tion­less and static, last­ing for a very long time. You might try to vi­su­al­ize the Block Uni­verse hang­ing in front of your mind’s eye, but then your mind’s eye is run­ning the clock while the uni­verse stays still. Nor does the Block Uni­verse ex­ist for just a sin­gle sec­ond, and then dis­ap­pear. It is not in­stan­ta­neous. It is not eter­nal. It does not last for ex­actly 15 sec­onds. All these are time­ful state­ments. The Block Uni­verse is sim­ply there.

Per­haps peo­ple imag­ine a Deter­mi­na­tor—not so much an agent, per­haps, but a mys­te­ri­ous en­tity la­beled “Deter­minism”—which, at “the dawn of time”, say, 6:00am, writes down your choice at 7:00am, and sep­a­rately, writes the out­come at 7:02am. In which case, in­deed, the fu­ture would be de­ter­mined be­fore you made your de­ci­sion...

In this model, the Deter­mi­na­tor writes the script for the Block Uni­verse at 6:00am. And then time—the global time of the uni­verse—con­tinues, run­ning through the Block Uni­verse and re­al­iz­ing the script.

At 7:00am you’re try­ing to de­cide to turn on the light bulb. But the Deter­mi­na­tor already de­cided at 6:00am whether the light bulb would be on or off at 7:02am. Back at the dawn of time when Destiny wrote out the Block Uni­verse, which was scripted be­fore you started ex­pe­rienc­ing it...

This, per­haps, is the kind of un­spo­ken, in­tu­itive men­tal model that might lead peo­ple to talk about “de­ter­minism” im­ply­ing that the fu­ture is de­ter­mined be­fore you make your de­ci­sion.

Even with­out the con­cept of the Block Uni­verse or time­less physics, this is prob­a­bly what goes on when peo­ple start talk­ing about “de­ter­minis­tic physics” in which “the whole course of his­tory” was fixed at “the dawn of time” and there­fore your choices have no effect on the “fu­ture”.

As de­scribed in Time­less Causal­ity, “cause” and “effect” are things we talk about by point­ing to re­la­tions within the Block Uni­verse. E.g., we might ex­pect to see hu­man colonies sep­a­rated by an ex­pand­ing cos­molog­i­cal hori­zon; we can ex­pect to find cor­re­la­tion be­tween two re­gions that com­mu­ni­cate with a mu­tual point in the “past”, but have no light-lines to any mu­tual points in their “fu­ture”. But we wouldn’t ex­pect to find a hu­man colony in a dis­tant su­per­cluster, hav­ing ar­rived from the other side of the uni­verse; we should not find cor­re­la­tion be­tween re­gions with a shared “fu­ture” but no shared “past”. This is how we can ex­per­i­men­tally ob­serve the ori­en­ta­tion of the Block Uni­verse, the di­rec­tion of the river that never flows.

If you are go­ing to talk about causal­ity at all—and per­son­ally, I think we should, be­cause the uni­verse doesn’t make much sense with­out it—then causal­ity ap­plies to re­la­tions within the Block Uni­verse, not out­side it.

The Past is just there, and the Fu­ture is just there, but the re­la­tions be­tween them have a cer­tain kind of struc­ture—whose ul­ti­mate na­ture, I do not con­ceive my­self to un­der­stand—but which we do know a bit about math­e­mat­i­cally; the struc­ture is called “causal­ity”.

(I am not rul­ing out the pos­si­bil­ity of causal­ity that ex­tends out­side the Block Uni­verse—say, some rea­son why the laws of physics are what they are. We can have time­less causal re­la­tions, re­mem­ber? But the causal re­la­tions be­tween, say, “drop­ping a glass” and “wa­ter spilling out”, or be­tween “de­cid­ing to do some­thing” and “do­ing it”, are causal re­la­tions em­bed­ded within the Block.)

One of the things we can do with graph­i­cal mod­els of causal­ity—net­works of lit­tle di­rected ar­rows—is con­strue coun­ter­fac­tu­als: State­ments about “what would have hap­pened if X had oc­curred, in­stead of Y”.

Th­ese coun­ter­fac­tu­als are untestable, un­ob­serv­able, and do not ac­tu­ally ex­ist any­where that I’ve been able to find. Coun­ter­fac­tu­als are not facts, un­less you count them as math­e­mat­i­cal prop­er­ties of cer­tain causal di­a­grams. We can define statis­ti­cal prop­er­ties we ex­pect to see, given a causal hy­poth­e­sis; but coun­ter­fac­tu­als them­selves are not ob­serv­able. We can­not see what “would have hap­pened, if I hadn’t dropped the glass”.

Nonethe­less, if you draw the causal graph that the statis­tics force you to draw, within our Block Uni­verse, and you con­struct the coun­ter­fac­tual, then you get state­ments like: “If I hadn’t dropped the glass, the wa­ter wouldn’t have spilled.”

If your mind con­tains the causal model that has “Deter­minism” as the cause of both the “Past” and the “Fu­ture”, then you will start say­ing things like, But it was de­ter­mined be­fore the dawn of time that the wa­ter would spill—so not drop­ping the glass would have made no differ­ence. This would be the stan­dard coun­ter­fac­tual, on the causal graph in which “Past” and “Fu­ture” are both chil­dren of some mu­tual an­ces­tor, but have no con­nec­tion be­tween them.

And then there’s the idea that, if you can pre­dict the whole course of the uni­verse by look­ing at the state at the be­gin­ning of time, the pre­sent must have no in­fluence on the fu­ture...


Surely, if you can de­ter­mine the Fu­ture just by look­ing at the Past, there’s no need to look at the Pre­sent?

The prob­lem with the right-side graph is twofold: First, it vi­o­lates the beau­tiful lo­cal­ity of re­al­ity; we’re sup­pos­ing causal re­la­tions that go out­side the im­me­di­ate neigh­bor­hoods of space/​time/​con­figu­ra­tion. And sec­ond, you can’t com­pute the Fu­ture from the Past, ex­cept by also com­put­ing some­thing that looks ex­actly like the Pre­sent; which com­pu­ta­tion just cre­ates an­other copy of the Block Uni­verse (if that state­ment even makes any sense), it does not af­fect any of the causal re­la­tions within it.

One must avoid mix­ing up time­less and time­ful think­ing. E.g., try­ing to have “Deter­minism” act­ing on things be­fore they hap­pen. Deter­minism is a time­less view­point, so it doesn’t mix well with words like “be­fore”.

The same thing hap­pens if you try to talk about how the Past at 6:30am de­ter­mines the Fu­ture at 7:30am, and there­fore, the state at 7:30am is already de­ter­mined at 6:30am, so you can’t con­trol it at 7:00am, be­cause it was de­ter­mined at 6:30am ear­lier...

What is de­ter­mined is a time­less math­e­mat­i­cal struc­ture whose in­te­rior in­cludes 7:00am and 7:30am. That which you might be tempted to say “already ex­ists” at 6:00am, does not ex­ist be­fore 7:00am, it is some­thing whose ex­is­tence in­cludes the Now of 7:00am and the Now of 7:30am.

If you imag­ine a coun­ter­fac­tual surgery on the in­te­rior of the struc­ture at 7:00am, then, ac­cord­ing to the statis­ti­cally cor­rect way to draw the ar­rows of causal­ity within the struc­ture, the 7:30am part would be af­fected as well.

So it is ex­actly cor­rect to say, on the one hand, “The whole fu­ture course of the uni­verse was de­ter­mined by its state at 6:30am this morn­ing,” and, on the other, “If I hadn’t dropped the glass, the wa­ter wouldn’t have spilled.” In the former case you’re talk­ing about a math­e­mat­i­cal ob­ject out­side time; in the lat­ter case you’re talk­ing about cause and effect in­side the math­e­mat­i­cal ob­ject. Part of what is de­ter­mined is that drop­ping the glass in the Now of 7:00:00am, causes the wa­ter to spill in the Now of 7:00:01am.

And as pointed out in Thou Art Physics, you are in­side that math­e­mat­i­cal ob­ject too. So are your thoughts, emo­tions, morals, goals, be­liefs, and all else that goes into the way you de­ter­mine your de­ci­sions.

To say “the fu­ture is already writ­ten” is a fine ex­am­ple of mixed-up time­ful and time­less think­ing. The fu­ture is. It is not “already”. What is it that writes the fu­ture? In the time­less causal re­la­tions, we do. That is what is writ­ten: that our choices con­trol the fu­ture.

But how can you “con­trol” some­thing with­out chang­ing it?

“Change” is a word that makes sense within time, and only within time. One ob­serves a macro­scop­i­cally per­sis­tent ob­ject, like, say, a lamp, and com­pares its state at 7:00am to its state at 7:02am. If the two states are differ­ent, then we say that “the lamp” changed over time.

In Time­less Physics, I ob­served that, while things can change from one time to an­other, a sin­gle mo­ment of time is never ob­served to change:

At 7:00am, the lamp is off. At 7:01am, I flip the switch… At 7:02am, the lamp is fully bright. Between 7:00am and 7:02am, the lamp changed from OFF to ON.

But have you ever seen the fu­ture change from one time to an­other? Have you wan­dered by a lamp at ex­actly 7:02am, and seen that it is OFF; then, a bit later, looked in again on the “the lamp at ex­actly 7:02am”, and dis­cov­ered that it is now ON?

But if you have to change a sin­gle mo­ment of time, in or­der to be said to “con­trol” some­thing, you re­ally are hosed.

For­get this whole busi­ness of de­ter­minis­tic physics for a mo­ment.

Let’s say there was some way to change a sin­gle mo­ment of time.

We would then need some kind of meta-time over which time could “change”.

The lamp’s state would need to change from “OFF at 7:02am at 3:00meta-am” to “ON at 7:02am at 3:01meta-am”.

But wait! Have you ever seen a lamp change from OFF at 7:02am at 3:00meta-am, to ON at 7:02am at 3:00meta-am? No! A sin­gle in­stant of meta-time never changes, so you can­not change it, and you have no con­trol.

Now we need meta-meta time.

So if we’re go­ing to keep our con­cepts of “cause” and “con­trol” and “choose”—and to dis­card them would leave a heck of a lot ob­ser­va­tions un­ex­plained—then we’ve got to figure out some way to define them within time, within that which is writ­ten, within the Block Uni­verse, within… well… re­al­ity.

Con­trol lets you change things from one time to an­other; you can turn on a lamp that was pre­vi­ously off. That’s one kind of con­trol, and a fine sort of con­trol it is to have. But try­ing to pull this stunt on a sin­gle mo­ment of time, is a type er­ror.

If you iso­late a sub­sys­tem of re­al­ity, like a rock rol­ling down hill, then you can math­e­mat­i­cally define the fu­ture-in-iso­la­tion of that sub­sys­tem; you can take the sub­sys­tem in iso­la­tion, and com­pute what would hap­pen to it if you did not act on it. In this case, what would hap­pen is that the rock would reach the bot­tom of the hill. This fu­ture-in-iso­la­tion is not some­thing that ac­tu­ally hap­pens in the Block Uni­verse; it is a com­putable prop­erty of the sub­sys­tem as it ex­ists at some par­tic­u­lar mo­ment. If you reach in from out­side the iso­la­tion, you can stop the rock from rol­ling. Now if you walk away, and again leave the sys­tem iso­lated, the fu­ture-in-iso­la­tion will be that the rock just stays there. But per­haps some­one will reach in, and tip the rock over and start it rol­ling again. The hill is not re­ally iso­lated—the uni­verse is a con­tin­u­ous whole—but we can imag­ine what would hap­pen if the hill were iso­lated. This is a “coun­ter­fac­tual”, so called be­cause they are not fac­tual.

The fu­ture-in-iso­la­tion of a sub­sys­tem can change from one time to an­other, as the sub­sys­tem it­self changes over time as the re­sult of ac­tions from out­side. The fu­ture of the Grand Sys­tem that in­cludes ev­ery­thing, can­not change as the re­sult of out­side ac­tion.

Peo­ple want to place them­selves out­side the Sys­tem, see them­selves sep­a­rated from it by a Carte­sian bound­ary. But even if free will could act out­side physics to change the Block Uni­verse, we would just have a Grand Sys­tem that in­cluded free-will+physics and the fu­ture would be fully de­ter­mined by that. If you have “freer will” we just have an Even Gran­der Sys­tem, and so on.

It’s hard to put your­self out­side Real­ity. What­ever is, is real.

Con­trol lets you de­ter­mine sin­gle mo­ments of time (though they do not change from one meta-time to an­other). You can change what would have hap­pened, from one time to an­other. But you can­not change what does hap­pen—just de­ter­mine it. Con­trol means that you are what writes the writ­ten fu­ture, ac­cord­ing to the laws of causal­ity as they ex­ist within the writ­ing.

Or maybe look at it this way: Pre­tend, for a mo­ment, that naive views of free will were cor­rect. The fu­ture “doesn’t ex­ist yet” and can be “changed”. (Note: How are these two state­ments com­pat­i­ble?) Sup­pose that you ex­er­cise your “free will” at 6:30am to res­cue three tod­dlers from a burn­ing or­phan­age, chang­ing their fu­ture from hor­rible flamey death at 7:00am, to happy gur­gling con­tent­ment at 7:00am.

But now it is 7:30am, and I say:

“Aha! The past is fixed and can never be al­tered! So now you can­not ever have cho­sen any differ­ently than you did choose. Fur­ther­more, the ac­tual out­come of your ac­tions can never change ei­ther; the out­come is now fixed, so even if your past choice did now change, the past out­come wouldn’t, be­cause they are both just de­ter­mined by “The Past”. While your will was once free at 6:30am to change the fu­ture at 7:00am, it is now 7:30am and this free­dom no longer ex­ists. So now your will at 6:30am is no longer free. How can your past will have been free, now that there is only one past? There­fore I do not now as­sign you any moral credit for sav­ing the or­phan­age; you no longer could have cho­sen differ­ently from how you chose.”

In the Block Uni­verse, the “past” and the “fu­ture” are just per­spec­tives, taken from some point within the Block. So, if the fix­a­tion of the past doesn’t pre­vent the em­bed­ded de­ci­sions from hav­ing (had?) the prop­erty of free­dom, why should the de­ter­mi­na­tion of the fu­ture pre­vent those em­bed­ded de­ci­sions from hav­ing the same prop­erty?

In the Block Uni­verse, the Fu­ture is just like the Past: it con­tains the Nows of peo­ple mak­ing choices that de­ter­mine their out­comes, which do not change from one meta-time to an­other.

And given the way we draw the causal ar­rows, it is cor­rect to form the (un-ob­serv­able) coun­ter­fac­tu­als, “If I hadn’t saved those chil­dren from the or­phan­age, they would be dead,” and “If I don’t think care­fully, my thoughts will end up in Outer Mon­go­lia.” One is a coun­ter­fac­tual over the past, and one is a coun­ter­fac­tual over the fu­ture; but they are both as cor­rect as a counter-fac­tual can be.

The next step in an­a­lyz­ing the cog­ni­tive is­sues sur­round­ing free will, is to take apart the word “could”—as in “I could have de­cided not to save the chil­dren from the or­phan­age.” As always, I en­courage the reader to try to get it in ad­vance—this one is eas­ier if you know a cer­tain sim­ple al­gorithm from Ar­tifi­cial In­tel­li­gence.

PPS: It all adds up to nor­mal­ity.

Part of The Quan­tum Physics Sequence

Next post: “The Failures of Eld Science

Pre­vi­ous post: “Thou Art Physics