Newcomb's Paradox - the modern version of the determinism vs non-determinism

[phyllo]

Flannel Jesus wrote: ↑Wed Jan 08, 2025 10:31 am
Indeterminism doesn't mean the past isn't fixed. It just means the future isn't (yet).

Depends on the definition of determinism, and a quick google of definitions gets you all sorts of vague and contradictory definitions that don't really distinguish it from what people here are calling 'indeterminism'.

Actually it depends on what one means by "fixed" and how if relates to something that doesn't exist except as a mental concept ( the future).

One can say that the future is "fixed" in a free-will world because when all the free-will decisions and actions are done, there will be only be one result. And we will know what the one result is, when it becomes the past.

The same thing happens in determinism ... decisions and actions produce one future result.

[Atla]

Well maybe anything could be behind the curtains if we don't assume determinism.

No. What is behind the curtains was put there by the show staff, and Monty certainly knows since his job depends on knowing. It cannot change from what the staff put there. The contestant doesn't know, but one person not knowing in no way implies that what is behind each curtain isn't what the staff put there. I don't know what crazy ideas you have about indeterminism, but it doesn't change how classical physics works, and all the scenarios (Monty, Newcomb) are straight up classical scenarios.

Why couldn't they change to something else in indeterminism?

That would violate empirical (classical) physics. I can't put an apple in a box, close it, and then find an orange there when I immediately reopen it. Indeterminism (a whole class of views, not just one view) suggests nothing of the kind.
That sort of thing is for parlor tricks, pulling rabbits out of empty hats and such. Our 'Omega' is presumably not doing such cheats.

Don't know what you're talking about, indeterminism and classical physics are mutually exclusive.

[Skepdick]

The non-determinists doesn't have to believe in a definite states. They can accept a box in superposition of states.

Are you going to try to pull superpositional stunts then? Determinist interpretations support superposition.

You think non-deterministic reasoning is some kind of stunt? Is that because you are incompetent at it?

A box cannot be in superposition of having a suitcase of cash in it and not having that suitcase.

That's what I said. If you are a realist (which you seem to be)...

* The boxes have definite contents prior to you choosing

There is also nothing the Omega nor the chooser can do to force a desired state from one of superposition.

That's also what I said...

* Your choice cannot affect the definite contents of the boxes.

Of course Omega can force a desired state. The box is NOT in superposition for Omega. Omega knows exactly what's in the box - they put it there.

I suspect that as Flannel Jesus suggested, you are suggesting that somebody (perhaps even yourself) believes that cash can be teleported into or out of the box after the decision has been made, but before the box is opened. Hey, if we have this Omega thing, we can have Star Trek too. We can cheat.

I am not suggesting that at all. I am suggesting that you outright reject that the state of the boxes can change as a result of your choosing one or both of them.

Which is perfectly fine...

You are affirming this part:

* The boxes have definite contents prior to you choosing
* Your choice cannot affect the definite contents of the boxes.

I am fine with the assertion. We will presume deterministic physics. That supports retro-causality, but it does not support retro information transfer

Causality and information transfer are identical notions. To communicate 1 bit of information is to cause a change in a different location in time. Even if it's at the same location in space.

This is trivially demonstrable by the fact that we continuously communicate with the future: leave yourself a post-it/reminder on your bathroom mirror.

Sending messages to your past self's a little harder.

and doing anything with the contents of the box is definitely information transfer, so we will indeed forbid any retro-anything-that-matters. I am fine with all those conditions, even if I don't agree with all of them. It actually helps support my case because it makes Omega more plausible. Still, Omega isn't that big of a fiction. Humans are far more predictable than they like to admit.

Great! So you definitely reject retro anything!

And you definitely agree that the opaque box has a definite state which cannot be altered by anything you do.
Before you even begin reasoning about which box to choose only two possible scenarios exist:
Scenario A: the opaque box is empty
Scenario B: The opaque box is full.

And you most definitely agree that a full box can't become empty; nor can the empty box become full via some magic teleportation!

But it isn't. The Omega was the strongest predictor you said. I didn't drag that in, you did. That's the cause of my explanation, not any kind of retrocausality, which, as we agreed, isn't going on.

But you don't have access to Omega as a predictor. You can't consult Omega on the contents of the box.
All you can consult is your own choices.

So the strongest predictor TO YOU. Is your choice. Knowing what choice you make immediatelly predicts exactly what Omega did; or didn't put in the opaque box.

Getting a million isn't a good reason to make the choice?

It's 2nd best to getting a million + $1000.

Every one-boxer gets a million. Every two-boxer walks away with a thousand. Said determinist would indeed be an idiot if he chose both.

So you no longer care about the fact that your choice retroactively alter the contents of the opaque box ?!?
You no longer agree that a full box can't become empty; nor can the empty box become full via magic teleportation?!?

The opaque box is before you. You haven't made a choice yet. You are neither a one-boxer nor a two boxer yet.
You think that if you one-box you'll find $1mil in there, but if you choose to two-box you'll find nothing ?!?

How does that work in your head?

I think you need to show me why his choice of just the one violates his deterministic belief.

I don't know how many different ways you need me to show or explain it to you.

Put yourself in the moment before choosing.
You are neither a one boxer nor a two boxer yet.

You accept that the opaque's contents are already determined.
You accept that the $1mil that's already there can magically disappear if you choose to take both boxes.
You accept that the $1mil that's not there can magically appear if you choose to take one box.

You accept that your choice doesn't in any way determine the contents of the boxes.

So when you one-box and win $1million you've irrationally left another $1,000 on the table.

If you do believe in determinism, at the moment of choice you should two-box because the contents are already fixed and you can't affect them causally.

By choosing based on past results (one-boxers win $1mil, two-boxers win $1000) rather than accepting that your choice is already determined, you are implicitly rejecting determinism.

Not show why he's wrong, but why he's not consistent with himself.

I have no idea what that even means. Everybody is consistent with themselves. How could you be any other way?

The inconsistency is merely in the discrepancy between what people claim to believe vs what they actually believe as revealed by their choices.

The key of course is this Omega. Can such a thing exist in a deterministic universe? Can it exist in a non-deterministic one? I think not the latter case, which is why I said I think a determinism assertion helps support my case. Of course you're not asserting determinism, you're only asserting an idiot with belief in determinism.

Yeah, because it's impossible to determine whether determinism or non-determinism's true.

The truth-values of determinism and non-determinism are undecidable/non-determinable.

There's no option to get that. Not by your rules.

What rules? The self-imposed ones such as realism/determinism?

According to the rules:
* The box has a determinate state before choosing.
* Your choice cannot alter the contents of the box
The option to win $1000000 + $1000 absolutely exists.

According to those rules you should two-box.
But you don't.

So you don't even follow the rules you claim to impose on yourself.

It is still a choice, but perhaps not a free one (as free choice is typically defined). There never was a free choice given determinism, although it has nothing at all to do with a mere belief in determinism.

Adding the adjective "free" doesn't change anything.

You have only one degree of freedom...whatever Omega predicted.
If it doesn't have at least two degrees of freedom it's not a choice!

Which is *precisely* why all determinist/non-retrocausal attempts at this fail.
Prior to the moment of choosing one MUST believe in the superposition/non-definite state of the box.

You can't both:

Have a genuine choice (multiple degrees of freedom) AND believe the box contents are definitely fixed before choosing.

That violates the OP asserting that the Omega knows what I'll choose. It says nothing about how much I waffle about it. I wouldn't waffle at all. Go straight for the one box, a total no-brainer.

Great! By choosing based on past results (one-boxers win $1mil, two-boxers win $1000) rather than accepting that your choice is already determined, you are implicitly rejecting determinism.

You could (if you wanted to) choose to win only $1000. Thereby two-boxing.
You chould (if you wanted to) choose to win $1 million. Thereby one-boxing.

This very ability to contemplate making either choice contradicts accepting the boxes have definite states.

The state of the box is simply caused by a set of events, not all of which are in the past light cone of the state in question.

Contradiction.

For A to cause B it necessarily entails that A occurs before B. e.g A is necessarily in the past light cone of B.

Wow. What the heck gave you that idea?? Tiktok video?

Mathematics.

The "neighborhood" of any point is all the points in the topology considered "local" to said point.
The neighborhood of the Big Bang singularity is all of reality.

https://en.wikipedia.org/wiki/Neighbour ... thematics)

Where did I ever suggest any such thing? Please quote where I said/implied that.

Sure. RIght where you insisted you have a genuine-but-non-free choice.
Of course, that contradicts your belief that the state of the boxes is already determined but hey...

I'm sure you'll continue pretending you don't hold contradictory beliefs.

In fact, I agreed to your 2nd bullet point that explicitly said "Your choice cannot affect the definite contents of the boxes", and yet here you say I make it sound like the choice does make the box contents change. You're not being at all consistent, but are instead attributing false assertions to me.

I am being consistently inconsistent. The only difference between you and me is that I am not lying to myself about it.

You insist that you have a choice.
While also insisting that state of the opaque box is already determined and cannot change.

Sure. I can pick both and go home with a K, an especially great choice if I watched 20 people ahead of me in line do it, and each going home with either 1M or 1K.

Really?

So you can pick both and end up with $1000.
OR you can pick one and end up with $1 million.

All while insisting that the contents of the opaque box is already determined and definite.
All while insisting that your choices can't alter that.

Agree that it doesn't follow from 1&2. 3 follows from watching the 20 people ahead of me walk away with only two possible outcomes, and 3 is not in contradiction with 1 & 2, or if it is, you've not demonstrated it so.

Precisely. That's a rejection of 1&2!

You don't believe the box has a definite state.
You actually believe you have a choice.
And you also believe that your choice is the best possible predictor and the strongest possible evidence for the contents of the box.

All of those beliefs directly undermine determiniism AND realism!

Without the Omega, and without witness of the 20 prior participants, perhaps this logic holds, but that's not what the OP describes.

Two boxing follows IF you accept 1&2. Which you claim you do.
But then you one-box. Which is strong evidence AGAINST belief in 1&2 and your rejection thereof.

You are utterly oblivious to your own meta-cognition.

One-boxing based on evidence is an anti-realist/non-determinist position!

Wrong! Did you not read your own OP? Nobody gets 0 or 1001000. This is where your logic is failing.

It follows directly from 1&2. Which you claim to accept. But now you reject it.

I guess you are confused about what you believe...

Where did determinism go? None of that analysis (however wrong) depended on determinism at all.

It vanished. Literally. The moment you chose to believe that the best predictor for the contents of those boxes is your own choice.

Why don't you next do a topic about the prisoner's dilemma? The same sort of logic can fallaciously be applied to that, except it doesn't involve a magic Omega.

I am sure you'd excel at fallacious reasoning in any setting.

[Noax]

indeterminism and classical physics are mutually exclusive.

Agree. I didn't say otherwise. People seem to think the OP is not a classical problem, but something else.

Of course Omega can force a desired state.

There are two phases. Phase 1 is where Omega does his prediction, decides what to put in the boxes, and seals them. Yes, he is the cause of his desired state of the box. Phase 2 is where the boxes are closed and the choice is made. My comment was than nobody can change the contents of the box during this second phase.
I say this because there have been implications that somehow 'indeterminism' allows that sort of thing. It doesn't.

I am suggesting that you outright reject that the state of the boxes can change as a result of your choosing one or both of them.

Agree.

Causality and information transfer are identical notions

Under a realist interpretation like Bohmian mechanics, 'spooky action' involves causality between two ambiguously ordered events, so it constitutes non-local causality but not non-local information transfer since messages cannot be sent via entangled pairs. Hence the two not being identical notions. None of this is relevant to the Newcomb thing which is a classical exercise.

And you definitely agree that the opaque box has a definite state which cannot be altered by anything you do.

Agree with this.

Before you even begin reasoning about which box to choose only two possible scenarios exist:
Scenario A: the opaque box is empty
Scenario B: The opaque box is full.

I agree with one of them is the state of the box. Not so sure about the other scenario being 'possible' since the box state has already been chosen by Omega, making the alternate scenario not possible anymore. Just covering my ass here since unclear wording can be taken the wrong way.

But you don't have access to Omega as a predictor. You can't consult Omega on the contents of the box.
All you can consult is your own choices.

But I can use the reliability of the Omega predictions as a basis for making the better choice. All that is accessible to me. I know the other box has the K in it, a trivial drop in big bucket, negligible bait that would likely cost me the big score because Omega would have known that I would reason that way.

So the strongest predictor TO YOU. Is your choice. Knowing what choice you make immediately predicts exactly what Omega did; or didn't put in the opaque box.

Yea, kind of like that, except I'm not sure if 'prediction' is a word that can be applied to an event in the past. I have to work during the playoff game, so I tape it and watch it later. If I make an educated guess as to how it turned out, it that a prediction of a past event, or should another word be used to describe what I'm doing?

Getting a million isn't a good reason to make the choice?

It's 2nd best to getting a million + $1000.

So by your description, 20 people play the game ahead of me. 10 use your logic and all walk home with 1000 instead of the nada they could have had. The idiots (your term) on the other hand decline the 1001000 and take home a mere million. Funny definition of 'idiot'.

So you no longer care about the fact that your choice retroactively alter the contents of the opaque box ?!?

I never said that. The contents are set by Omega's prediction, not by anything I do.

The opaque box is before you. You haven't made a choice yet. You are neither a one-boxer nor a two boxer yet.

Oh, I was a one-boxer from well before Omega did his thing. No hesitation. I was a super easy prediction. It does not pay to attempt deception in this game. It can be done, but not profitably.

You think that if you one-box you'll find $1mil in there, but if you choose to two-box you'll find nothing ?!?

Yep.

Put yourself in the moment before choosing.
You are neither a one boxer nor a two boxer yet.

Nonsense. I am a one boxer all the way.
So let's go with Fred who hasn't actually thought it through until phase 2 begins. But Fred is very much aware of the predicting prowess of Omega. With Fred, we can go on with your attempt to rationalize a different choice. For arguments sake, I am now Fred, who is smart, but not quick about it, so he hasn't made up his mind before things have even started.

You accept that the opaque's contents are already determined.

Right, and this has nothing to do with determinism, just for the record. It is classical fact. I added that because you tend to reference 'determined' in a non-classical way.

You accept that the $1mil that's already there can magically disappear if you choose to take both boxes.
You accept that the $1mil that's not there can magically appear if you choose to take one box.

No, cannot accept either of those statements. No magic, sure, but I cannot accept a million being there if I choose both boxes. It's possible, but super unlikely.

You accept that your choice doesn't in any way determine the contents of the boxes.

It kind of does since my choice is predictable, and it is the prediction of my choice that determines the contents. This is no different than my calendar showing the moon phases. The moon being full on the 13th does not directly cause the calendar mark being put there when it was printed 2 months ago. It isn't retro-causation, and yet the moon being full on the 13th does actually determine what is printed back in last November. It's is actually the prediction of the full moon that causes the contents of the calendar.

So when you one-box and win $1million you've irrationally left another $1,000 on the table.

No, it isn't irrational. I have a million and you don't.

If you do believe in determinism

My belief concerning determinism never played a role in my decision.

at the moment of choice you should two-box because the contents are already fixed and you can't affect them causally.

No. The two boxers all get only a K. The choice is irrational, and isn't based on a determinist stance at all, but a simple goal to maximize the utility.

By choosing based on past results (one-boxers win $1mil, two-boxers win $1000) rather than accepting that your choice is already determined, you are implicitly rejecting determinism.

You've not demonstrated this at all since at no point was determinism violated by my claims.

Everybody is consistent with themselves. How could you be any other way?

The vast majority of people I've encountered hold self-inconsistent beliefs. Enough that I don't even claim to be an exception to that observation. This applies to beliefs stated, and also to the beliefs actually held.

Yeah, because it's impossible to determine whether determinism or non-determinism's true.

Alta would disagree with that assertion. While I agree, I find the issue utterly irrelevant.

There's no option to get [1001000]. Not by your rules.

What rules?

The description in the OP, which are actually Newcomb's rules, with amendments from say Nozick.

According to the rules:
* The box has a determinate state before choosing.
* Your choice cannot alter the contents of the box
The option to win $1000000 + $1000 absolutely exists.

According to those rules you should two-box.

Agree that the last line follows. But I don't agree that those are all the rules. You repeat this here:

Two boxing follows IF you accept 1&2. Which you claim you do.
But then you one-box. Which is strong evidence AGAINST belief in 1&2 and your rejection thereof.

Again, totally omitting the part that matters.
Funny how you so lack confidence in your reasoning that you resort to leaving out the parts upon which the counter argument is based, and you do this repeatedly in the same post.

Great! By choosing based on past results (one-boxers win $1mil, two-boxers win $1000) rather than accepting that your choice is already determined, you are implicitly rejecting determinism.

Why would a deterministic belief necessitate my making a choice not based on the reliability of Omega? It just doesn't follow. I didn't consult my opinion on that subject at all when making the obvious decision.

You could (if you wanted to) choose to win only $1000. Thereby two-boxing.
You could (if you wanted to) choose to win $1 million. Thereby one-boxing.

Those options are both open to me, yes.

This very ability to contemplate making either choice contradicts accepting the boxes have definite states.

OK, Fred speaking here since I personally would spend no time contemplating it. Fred is probably contemplating options other than the two above, as are you. But the reliability evidence prevails, and the choice eventually gets made based on that, and it doesn't contradict knowledge of classical box contents not changing while being closed. Such a possibility was never considered, not even by Fred. Fred is a more difficult person to predict than am I, but Omega is still up to the task.

Contradiction.

For A to cause B it necessarily entails that A occurs before B. e.g A is necessarily in the past light cone of B.

You erased all context, but the comment concerned non-local interpretations of QM. If you want your statement to be true (that A needs to be before B), then you need to accept locality which asserts exactly that. Without that principle, there is no such restriction.

Mathematics.

The "neighborhood" of any point is all the points in the topology considered "local" to said point.
The neighborhood of the Big Bang singularity is all of reality.

https://en.wikipedia.org/wiki/Neighbour ... thematics)

Nothing in that mathematics implies anthropocentrism.

You really need to keep context in your quotes. I had to hunt it down and put it back.

I'm sure you'll continue pretending you don't hold contradictory beliefs.

I actually admitted otherwise. So much for you being a reliable predictor.

You insist that you have a choice.
While also insisting that state of the opaque box is already determined and cannot change.

That's right. I have no idea why you think that my insisting that box contents don't magically change (something almost nobody would contest) implies that I have no choice (also something almost nobody would contest). I can't think of anybody who believes otherwise, regardless of their stance on the various philosophies that you think are relevant while I don't.

So you can pick both and end up with $1000.
OR you can pick one and end up with $1 million.

Precisely so. All those people ahead of me had the same choice, each with the result indicated.

All while insisting that the contents of the opaque box is already determined and definite.

Why do you go on about this one so much? I've never met anyone that would suggest the box contents can change. That's magic, utterly inconsistent with empirical evidence.

I have met a few that suggest that choice doesn't exist. It's called fatalism, that one's will is in no way part of the cause of subsequent events. You should ask one what he'd choose. He'd say of course that he cannot choose. Fate will choose for him, and it's not within his control. OKfine.

Precisely. That's a rejection of 1&2!

You not demonstrating any contradiction with 1&2 constitutes a rejection of 1&2? What sort of wonky logic is this?

You don't believe the box has a definite state.

That contradicts #1 that I agreed to, so wrong.

You actually believe you have a choice.

Who (except the fatalists) doesn't?

And you also believe that your choice is the best possible predictor and the strongest possible evidence for the contents of the box.

No, I never said anything like that.

All of those beliefs directly undermine determiniism AND realism!

Neither was mentioned in any of that tirade,.

[bahman]

The scenario: A superintelligent being (Omega) presents you with two boxes.

* Box A is transparent and contains $1,000. Box
* Box B is opaque and contains either $1,000,000 or nothing.

The twist:

If Omega had already predicted you'll take only Box B, they will put $1,000,000 in it.
If Omega had already predicted you'll take both boxes, they will leave Box B empty.
e.g the contents of box B are already determined.

Should you take opaque box; or both boxes?

The non-determinist/one-box argument:

Omega has a reputation as near-perfect predictor.
If I choose only Box B, Omega likely predicted this and put $1,000,000 in it
Therefore, choosing only Box B will likely get you $1,000,000.

The determinist/two-box argument:

The money is already in the boxes.
Whatever is in Box B won't change based on your choice because the future can't affect the past.
Taking both boxes will always give me $1,000 more than taking just Box B
Therefore I should take both boxes.

But the crux... A consistent determinist must two-box, because doing otherwise would mean abandoning determinism itself, and yet the determinist simply can't explain why choosing the opaque box reliably correlates with better outcomes.

How could one predict what you will do in a situation with at least two options?