Quantum immortality testability paradox


Let's suppose the many worlds interpretation of quantum mechanics (MWI) is correct and a Schrödinger's box is possible.

I put myself in the box. I live or die. I repeat. One version of me keeps surviving (thousands of times) and becomes certain about MWI.

Versions of my colleagues on the outside see my return several times, but the chances of seeing this are the same whether MWI is right or not. So seeing me coming out of the box alive, from their point of view, wouldn't make it more likely for the MWI to be true.

Note that what I experienced in the box is nothing the outside observers couldn't imagine. They know each time that I survived, and what it looked like from my point of view.

They'd have the same information, and therefore the same premises as I, but wouldn't reach the same conclusion, even though we'd all use the same, apparently correct logic. How can this be possible? Is this something that has been discussed before?


Posted 2015-11-03T20:12:44.773

Reputation: 159

Question was closed 2016-01-21T18:16:55.197

Interesting question, but it should be migrated to a physics SE, as it is more about an experiment than about any philosophical consequences or aspects of QM. – Alexander S King – 2015-11-03T20:47:45.520

@AlexanderSKing A physicist wouldn't be able to help with this since the assumptions are clear enough: every time you run the experiment, the world splits into two versions, one in which you survive and one in which you don't. The question is about the testability of this hypothesis. I could've asked this question 200 years ago as a hypothetical scenario and still wonder whether this is testable or not, regardless of the actual physical laws. – Juan – 2015-11-03T20:57:23.603

I'm not really grasping the question we can answer about philosophy in this question. Can you at a minimum make it clearer by editing the question. – virmaior – 2015-11-03T23:28:25.130

@virmaior It is on the last paragraph. My colleagues and I have the same set of premises (which include the fact that I survived the experiment millions of times), but as far as I can tell, I'd have to conclude that MWI is real, while they wouldn't. How can the same set of premises lead to different conclusions, while we are all using the same rules of reasoning? Isn't this a philosophical question? – Juan – 2015-11-03T23:41:10.720

I think I can grasp the basic outline better, but then your question is merely t that there's an experiment where a subject has different experience from the observer (which would normally mean different premises but you seem to want to reject that) and therefore reaches a different conclusion ? If so, we can completely edit out any reference QM MWI, but then the question is uninteresting, because the answer is that the premises differ precisely insofar as the subject has different ones by difference of being the subject. – virmaior – 2015-11-04T01:43:27.820

@virmaior I actually find your answer pretty interesting. The only difference in their premises would be "I survived" vs "He survived", yet we'd all reach different conclusions about the world, while all conclusions would be simultaneously correct. I don't see how this is uninteresting. Wouldn't this imply that truth is relative to the observer? Could you point me to any other trivial case in which something similar occurs, since apparently this is such a common occurrence? – Juan – 2015-11-04T01:53:44.877

Okay I've stripped out a large amount of unnecessary text. But I don't think it leads to any sort of large conclusion like you're imagining. I don't think two perspectives cause a big of a problem as you imagine. – virmaior – 2015-11-04T02:28:44.110



It's very weird, but in a sense it's no different than playing Russian Roulette: your experience, a minute after you play, is always that you survived.

What makes this really hard to fathom--and seem to raise epistemological questions--is just how bafflingly many parallel universes exist, the vast, vast majority of which have no alive you in them. Whenever you do statistics (explicitly or intuitively), you will get incorrect results unless you understand selection effects and sample size. Here, the sample size is 2number of times you ran the experiment, and the selection effect is that you only count yourself as alive in those samples where you are alive. (Yay, tautology!)

Everyone can understand this, and everything is consistent, though observers of you surviving a million times will rightly consider it exceedingly improbable that it happened (since there are 21000000-1 copies of them in the many worlds who did not observe this). In those exceedingly rare universes, people will be exceedingly surprised.

But then, we should all expect that anyway--being surprised by rare events is what statistics is about.

Rex Kerr

Posted 2015-11-03T20:12:44.773

Reputation: 15 388

Would this mean I'd be wrong to conclude that the many world interpretation is real after running the experiment, say, a million times? – Juan – 2015-11-03T21:07:55.460

@Juan - Depends what the alternative hypotheses were. Possibly yes. It wouldn't help all the dead yous, though, or all the observers in the 2^1000000 - 1 other worlds. And of course if many worlds is wrong, you're just plain dead, everywhere. – Rex Kerr – 2015-11-03T22:55:11.747

I think I see what you mean. Only my surviving self will count, and all the other's will be discarded, which amounts to a selection bias. The problem is that this is already assuming that MWI is true. The selection bias would only be a problem if I'd be trying to test some property of the MWI, given that it is real, where we need to get information from every possible version of the universe, but here I'm just testing whether there are many universes or not. So we don't need to care about the dying versions. – Juan – 2015-11-03T23:31:40.637


In the worlds where you die, they have different information than you. In the world where you live, they do have the same information -- both of you have seen you survive a thousand runs through Schrodinger's box.

Accordingly, you have to be careful with what you define "them" to be. The "them" who has seen you survive is a "them" limited to the same timeline as you, indicating that an experiment has been done a thousand times, and none of the thousand experiments has shown you dying. Then there is the "them" that had to send flowers to your mother after letting you climb into a rather absurd little box.

The difference between "them" on the limited timeline and "you" on the limited timeline is that, to them, you were simply a datapoint. They could run the same test on the next lab-rat who thinks its a great idea to climb in a box. From your perspective, however, you ponied up. You put your own life on the line.

So, from their perspective, they have one test, which could be repeated on a subject. From your perpsective, you have one test that, if run on someone else, is different than what you just endured.

Cort Ammon

Posted 2015-11-03T20:12:44.773

Reputation: 16 681


I am confused by your fourth paragraph, so perhaps this is more of a question than an answer.

I don't see that you can mix the physics and the speculative metaphysics in this way and also retain an "experimental method." There seems to be a third "viewpoint" comparing scenarios and reasoning about the parallel "evidence" and diverging logic of the conclusions.

If your memory, the evidence of your senses, and your reasoned conclusions all diverge uniquely and radically from those of the other experimental observers, then your being "convinced MWI is true" makes you either a very bad experimental scientist or, more to the point, a highly unreliable observer, perhaps technically insane.

This sounds like an inappropriate, facetious answer. Perhaps. But I cannot see that the scenario you describe, with one continuous, conscious observer comparing multiple MWI outcomes with an ever-diverging universe of fellow experimentalists, is one in which the experimental method pertains or can be used to arrive at any conclusions whatsoever.

Though I am not familiar with the literature, it would seem to me that the "physics" of your MWI experiment contains a number of unexamined assumptions about identity and continuity of consciousness. More "observer" paradoxes than physics can survive. Perhaps these are, in fact, worked out in discussions of MWI, and if so I'd be curious to have some references. It's very interesting.

Nelson Alexander

Posted 2015-11-03T20:12:44.773

Reputation: 11 748