Topic for discussing physics.
Topic for discussing physics.
Want to discuss this? Join my forum.
(Due to multi-year, sustained harassment from David Deutsch and his fans, commenting here requires an account. Accounts are not publicly available. Discussion info.)
Messages (40)
This is a comment I wrote on:
http://blog.rongarret.info/2019/07/the-trouble-with-many-worlds.html
**The Trouble with Many Worlds**
> Now consider a case where the odds are not even. Let's arrange for the probabilities to be 2:1 in favor of A (i.e. A happens 2/3 of the time, B happens 1/3 of the time, according to the Born rule). Now we have a disconnect between the two world-views. The Bornian would obviously choose A. But what possible reason could the many-worlder have for doing the same? After all, the situation is unchanged from before: again the many-worlder is going to split into two (because there are still only two possible outcomes).
The reason is simple. Not all universes exist in equal quantities. There could be e.g. twice as many of universe A as of universe B. So he’d prefer to make the bet where he wins in 2/3 of (the relevant branch of) the multiverse over the one that wins in 1/3 of the multiverse. More copies of him will win.
The idea of there being exactly one copy of a person for each type of universe (e.g. win bet type or lose bet type) is incorrect.
There are many identical universes. When they “split” (aka “branch” or “differentiate”), the split does not have to happen with equal proportions.
You don’t split from one into two. You split from many into many (the same number as before – no universes are created nor destroyed). “Splitting” means some universes become different that were, previously, identical. You can split into e.g. 2/3 of one outcome and 1/3 another outcome.
(Warning: speaking about universes is an approximation. They have no fundamental role in physics. One of the reasons is along the lines Ron states in the quote below that begins “This process is a continuous one.” Plus universes are big but quantum physics is local – change spreads at the speed of light or less.)
Wallace writes:
https://arxiv.org/pdf/0906.2718.pdf
> **Branching Indifference:** An agent doesn’t care about branching per se: if a certain measurement leaves his future selves in N different macrostates but doesn’t change any of their rewards, he is indifferent as to whether or not the measurement is performed.
Ron writes:
> Branching indifference says that a rational agent doesn't care about branching per se. In other words, if an agent does a quantum experiment that doesn't have a wager associated with it, then the agent has no reason to care whether or not the experiment is performed or not.
This is unclear or is different than Wallace because it speaks of a *wager* rather than a *reward*. A wager means betting money, whereas a reward is anything that is good or bad according to an agent’s preferences.
Wallace is basically saying that an agent is indifferent to branching into copies *if*, after branching, no copy of the agent is worse (or better) off in any way.
That sounds totally unobjectionable to me. If a particular branching has no downsides or upsides (in any universe), according to an agent’s preferences, then an agent is indifferent to it. In other words, agents care about positive and negative rewards rather than branching itself (branching sometimes, but not always, has consequences for rewards).
Being pedantic, one could say the same thing about any other physical events (eating breakfast, getting fired, hitting a homerun, winning the lotto, etc.) – it’s not the event itself, per se, which matters, but whether or not there is any change in reward according to the agent’s preferences. E.g. “sleep indifference” states that agents are indifferent to how much sleep they get, per se, as long as there is no positive or negative change to their rewards.
In other words, if an agent has no preference about something, then that agent is indifferent to it. And agents don’t necessarily have preferences about all possible multiversal branching.
> if I'm supposed to care about "copies of me" at all, how can it not matter how many there are?
You can care about a *measure* of copies that isn’t simple quantity/count (a quantity/count of marbles is an example of simple quantity/count).
> So what about the rhetoric of MWI, that when you do an experiment with N possible outcomes that you split/peel-apart/whatever-you-want-to-call-it into N copies of yourself?
That is *not* the MWI position. You split into N *different versions* of yourself, but not into N copies of yourself (that’s different because it could be e.g. 3 copies of the first version, 8 copies of the second version, etc. But they’re harder to measure than simple counting.)
> This process is a continuous one. There is never a well-defined "point in time" where the entire universe splits into two, and no point in time where you (meaning your body) splits into two.
Agreed.
> subjective
I think it’s better to start by considering how physics works without worrying about conscious or “subjective” experiences. Let’s discuss all the other stuff first and only tackle issues related to consciousness after agreeing on the rest (like what happens with dice, photons, mirrors and cups of water). That’s what I’ve done above.
> You can either look at the whole system (in which case you see quantum behavior) or you can look at part of the system (in which case you see classical behavior) but you can't do both at the same time.
But if you look in certain parts you do *not* see classical behavior. Classical physics being *false* and *refuted by some experiments* is why we have quantum theory. Classical physics is a good approximation in many cases people encounter in their daily lives, but not in all cases. Agreed?
Ron mentions discussing his physics post on reddit while writing it. I found that discussion:
https://www.reddit.com/r/Physics/comments/c2225k/physics_questions_thread_week_24_2019/eri9urh/
> The idea of there being exactly one copy of a person for each type of universe (e.g. win bet type or lose bet type) is incorrect.
Where did he get this idea if, as he claims, he has read FoR and some of Deutsch's papers? He has completely missed the key of idea of fungibility even though it is a centerpiece of Deutsch's arguments.
Another Reply to Ron
https://blog.rongarret.info/2019/07/the-trouble-with-many-worlds.html
> Yes, I understand that that's what the math says. The problem is that I don't feel like many. I feel like one. If I am many, why don't I feel like it?
MWI is an *objective* theory based on *scientific observations and math*, **not** based on a philosophical theory of consciousness. While *I* could comment on the matter, and DD also has opinions on the matter, they are a separate issue than MWI.
If you told me that due to misunderstandings of consciousness (which you could name and explain), some of experimental data needed to be reconsidered or rejected, that would be relevant. Same if you rejected some math, logic or prior-to-QM physics claims. But if you accept all the experimental data and all the math, logic and physics premises, then doesn't MWI (or a few other options which mostly differ from MWI by their claims about objective reality, not by their theories of consciousness) follow?
> Yes, but what do those quantities have to do with *decisions*? You can't just *assume* that "higher-quantity" universes should have a greater weight in decision-making. That is begging the question.
The point of DD's paper is to prove that, in terms of betting by the "rational agents" of decision theory, you should do betting-decision-making according to "higher-quantity" universes (in order to maximize multiverse-wide betting returns). He does not assume this point nor beg this particular question. Your objections related to consciousness do not constitute a criticism of DD's proof of this matter from his premises.
The short, approximate version is that it's better to win $1 in 66 universes and lose $1 in 34 than vice versa. What they have to do with decisions is that you (a group of many fungible, identical, indistinguishable instances of a person) would prefer that more of you (more of those instances that already exist and are part of the group of instances that make the decision) would prefer that more of your instances win over fewer winning.
> The problem is that in a branching universe the phrase "the agent" no longer has a well-defined referent.
Do you mean because an agent has microscopic changes which haven't propagated to his entire body at once (before some changes finish propogating, others begin, so there are always some changes which have not propgated to his entire body)? (And the same point works with just his brain.) Is that the issue?
I think that issue of defining a particular agent and its preferences in a world that is in constant flux is not really a QM or MWI issue. Heraclitus could have made a similar complaint. And I don't think other QM interpretations, which do not contradict the math and observations, will change this problem much. DD's point in the paper is that *if* you accept certain claims of decision theory (which are pretty widely accepted), including that you can take as a starting point an agent with a well-defined set of preferences, *then* various things follow. The details of how to deal with a world (including agents) in constant flux is a separate matter not covered in that paper.
I think this separate matter of dealing with flux is widely believed to be soluble, and that DD and I have that belief. And I think it can be approached without getting into the problems of consciousness or subejctivism that I avoided above. Like you could have a software agent, which isn't even (general) intelligent, and it bets trying to maximize some function, and it too would have the constantly-undergoing-change issue. The solution, in short, is that despite the ongoing flux/chaos/change *in some respects*, there are some things which change very very little over some short time periods, so they can be taken as approximately constant at that time when they are approximately unchanging.
Something being approximately constant, rather than exactly constant, means that error correction is needed, which is getting pretty far afield (though covered a bit in BoI which explains the advantage of digital over analog for error correction). Very briefly it's like how computer circuits deal with electrical signals that are approximate, not exact, in terms of strength and timing. Minor fluctuations can be and are dealt with.
#12985 Despite seeming to contradict or not know that stuff multiple times in his post, Ron replied claiming that he does know and accept it.
I think it's unclear what's going on. He is an amateur who hasn't had much contact with the community of ppl familiar with DD's ideas, so that can help contribute to some misunderstandings and to some superficial errors.
Another Ron Reply
https://blog.rongarret.info/2019/07/the-trouble-with-many-worlds.html
> So is there any reason I should *care* about the existence other versions of me (except perhaps as an intellectual curiosity)?
The "you" that places the wager consists of multiple identical copies. For simplicity, we'll call it 100. So there are 100 clones of you which, as a group, place a wager. Why would "you" care about "other versions" of you? Because "you" (as of the start of the scenario where you make a choice about a wager) are 100 people, you should bet in a way that gives the best outcome for those 100 people.
You should try to maximize outcomes for the versions of you that placed the wager, precisely because they are part of the entity that placed the wager. Each of them is a real person who is making a wager and wants to win. If you bet so that 2/3 of them win, that that's better for the group, and also then each of them individually has a 2/3 chance to win rather than e.g. a 1/3 chance.
PS Because your captcha system seems to block some of my comments and also keeps presenting me with many captchas in a row (e.g. 5 in a row, I've done ~two dozen in total), I don't want to continue discussing here. If you want to discuss further, please post at https://curi.us/2209-physics-discussion where you can post without any captcha or moderation, and with better formatting too. If you don't care enough to do that then I think I'll, sadly, give up because the software here is too broken and/or user hostile.
Alan wrote:
https://blog.rongarret.info/2019/07/the-trouble-with-many-worlds.html
>> I think it’s better to start by considering how physics works without worrying about conscious or “subjective” experiences.
>
> This is our biggest disconnect. My subjective experience is the only data I have direct access to. It's the only reason I have to even suspect that there is such a thing as "the laws of physics" out there to be discovered. So you can start by considering how physics works if you like, but if you want to tell a complete story of how the world works then sooner or later you're going to have to circle back and consider how *considering* works.
Considering is a form of information processing. A universe is a structure within the multiverse where information can flow from one system to another. For example, there are versions of me sitting 1 inch to the right of my current position. I can't exchange information with those other version of me. I can't see whether one of those versions of me is sitting with his legs crossed. That is why those versions of me count as separate versions in separate universes.
Now, I have a record in my memory of writing the first version of this paragraph. There is also information in the environment about what I typed, e.g. - information in sound waves from my typing in light reflecting off the keys and so on. This information can be used to decide whether I am the same person as the person who wrote the first paragraph of this reply. Other versions of me wrote a different second paragraph but also have records of the same first paragraph. The fact that there is more than one version of me with such a record doesn't change the fact that the identification can be made.
A similar story about identifying objects using records can be told about other objects like the keyboard I'm typing on, the pen sitting on the desk behind the keyboard and so on. Any decision you're going to make has to use the same kinds of records that would be used to identify the pen or the keyboard or whatever. From the point of view of physics, there is no particular reason to make a special case for people as opposed to pens or computers or whatever. So we might as well consider a computer programmed to maximise its rewards rather than a person. We know how to program a computer to just follow a particular rule. Programming a person to just follow a particular rule is difficult and raises irrelevant moral problems, so discussing a computer program makes more sense.
https://blog.rongarret.info/2019/07/the-trouble-with-many-worlds.html
> Call it whatever you like, this cannot be a correct explanation. If I am N identical copies, then after O(log(N)) splits I will be 1. What happens then?
OK, you want less approximate. Some BoI quotes (not in order):
> Thus the information in the fictional multiverse flows along a branching tree, whose branches – histories – have different thicknesses (measures) and never rejoin once they have separated.
> *Instances* In parts of the multiverse that contain universes, each multiversal object consists approximately of ‘instances’, some identical, some not, one in each of the universes.
> In quantum physics, information flow in the multiverse is not as tame as in that branching tree of histories I have described. That is because of one further quantum phenomenon: under certain circumstances, the laws of motion allow histories to rejoin (becoming fungible again).
> In principle, a phenomenon could appear unpredictable to observers for one or more of three reasons. [...] The third – which had never been imagined before quantum theory – is that two or more initially fungible instances of the observer become different.
> Then they know that, when they run the transporter, an infinite number of fungible instances of themselves, all sharing the same history, are doing so at the same time.
> Our fictional theory has not provided enough structure in its multiverse to give a meaning to ‘half the universes’, but the real quantum theory does. As I explained in Chapter 8, the method that a theory provides for giving a meaning to proportions and averages for infinite sets is called a *measure*. A familiar example is that classical physics assigns *lengths* to infinite sets of points arranged in a line. Let us suppose that our theory provides a measure for universes.
You don't actually count universes, you measure them, just like you don't count points, you measure them.
You don't run out of points when you divide up an inch repeatedly. Same with instances in the multiverse.
---
Quoting Ron now:
> Nonetheless, my subjective experience is still something that a complete theory of reality needs to account for as far as I'm concerned.
Yes, but there is nothing about MWI which contradicts any of your subjective experiences. It never said "You will never subjectively experience X" for any X that you have experienced.
> No, that's not true. What DD's paper shows is that *if* a rational agent accepts the quality metric of maximizing multiverse-wide betting returns biased according to branch *weights* (and not branch *counts*) *then* a rational agent should behave *as if* the Born rule were true.
If a multiversal agent in N states tries to maximize the outcomes for its states, it should... (Except you have to measure, not count, those "N" states. It's the same sort of problem as saying a line containing "N points", which is sorta misleading but I don't know a better wording.)
> [Wallace] has to assume branching indifference, and that, it turns out, is where the Born rule is hiding.
No, as I explained previously branching indifference is trivial.
> The only reason there was a problem in the first place is because a photon was taken (on good evidence) to be an indivisible unit that could not "split up". So which is it? Can a photon split up or not?
Single-universe photons do not split. Multiversal photons do not split but are already pre-"split" in a way very similar to how a 2-inch line segment is already pre-"split" into points instead of being an indivisible whole.
> No, I mean that "the agent" can refer to either the pre-split agent, or any one of the N post-split agents. If you're going to talk about *the* agent (singular) then you have to specify which of the N+1 agents you mean.
The agent at the start of the scenario is a collection of N fungible (identical) agents.
There is no splitting, ever New agents are never created. There is only differentiating: agents that already existed, and were formly identical, become different.
> Dealing with "minor fluctuations" often requires a radical departure from the conventional wisdom.
A tiny fraction of minor fluctuations turn out not to be minor. Most really are minor, e.g. if you measure a length, weight or temperature of a glass of water there will be tiny fluctuations (of the glass, the water, and the measuring instrument) that affect your measurement but they are usually too minor for you to even notice.
https://blog.rongarret.info/2019/07/the-trouble-with-many-worlds.html
(Quoting myself.)
> If a multiversal agent in N states tries to maximize the outcomes for its states, it should... (Except you have to measure, not count, those "N" states. It's the same sort of problem as saying a line containing "N points", which is sorta misleading but I don't know a better wording.)
What about from the perspective of an individual agent instance? That's a bit of an approximation because what really exists are multiversal objects, not single universe objects. But it works pretty OK. From that perspective, basically what happens is probability: if 2/3 of the instances in the multiversal object get outcome X then from the perspective of a single instance it has a 2/3 chance to get outcome X.
https://blog.rongarret.info/2019/07/the-trouble-with-many-worlds.html
https://blog.rongarret.info/2019/07/the-trouble-with-many-worlds.html
> > branching indifference is trivial
> Being trivial and being the root of the begging of the question are not mutually exclusive.
If you have some argument about why branching indifference is false, or has some other problem in this context, please explain it. As far as I could tell, your objections to it came from not understanding it (or, in the alternative, understanding it differently than I do). So I explained what I thought it meant. IIRC you seemed to agree with me. But now you bring it up again as problematic.
> Ah. How is that done exactly? If I wanted to "measure" the universe I inhabit, how would I do it? If I measured my universe before and after a quantum event, would I get a smaller result the second time?
Measures can be abstract or hypothetical. They aren't only things you can actually measure. The concept of a measure does not require a physical process does performs that measure. Here is some explanation of measures, and relates issues, based on conversations with Deutsch (he taught it to me):
https://curi.us/1955-explaining-infinite-sets-measures-and-mappings-for-quantum-physics
That is not a full explanation of everything. It's just a starting point. If you agree with what I've said so far, in my comment and a the link, I'll continue from there. If there's a disagreement with this part, then we can talk about that first.
> > there is nothing about MWI which contradicts any of your subjective experiences
> Of course there is. MWI says that there are many of me (that's what the M in MWI stands for!) But my subjective experience is that there is only one of me.
You have never in your life observed the absence of other instances of you in the multiverse. You have looked only at a limited portion of the multiverse and didn't see more of you. That is fully compatible with MWI.
Your experience, taken more literally, is e.g. that you looked down and saw your foot. Observations like that are valid data but do not contradict MWI. You seem to be mixing up your personal (= made by you instead of someone else) observations with vague feelings which are not really obsrvation data but are intellectual intuitions.
> And what exactly compels me to measure and not count? Why am I compelled to use the measure of a universe to make decisions rather than coalescing all of the fungible versions of me into one equivalence class and treating that as a single entity?
It's like the difference between 1 inch and 2 inches. If you just say "a linear set of points" and treat all such sets as equivalent, you will deal with distance poorly.
yeah true
#13037 What are you saying is true?
#13038 there's a pretty big diference between 1 inch and 2 inches
This entire discussion is predicated on the assumption that more than one universe exists. What strong evidence do we have that more than one universe exists? If we define the *Universe* as "the collection of all things that exist", then the idea of a Multiverse seems nonsensical.
From my readings here and on other online forums, most supports of the multiverse theory claim that "the math works out." Do y'all have other evidence or good reasons other than equations?
#13050 To begin with, see DD's books, esp FoR ch2 and then BoI ch 11. http://beginningofinfinity.com/books
Regarding Infinite Sets
>Physicists have a measure of infinite sets of universes.
I would disagree that an *infinite* set even exists. Everything that exists has a specific quantity, therefore there is no such thing as infinity in reality.
Infinity does serve a purpose in mathematics (especially in calculus) when dealing with arbitrarily large values.
Unfortunately, the scientific community today considers infinity as a real thing that exists, hence the argument that "there are infinitely man universes" or "there is an infinite amount of information in the universe."
#13051 Cool, I'll have to take a look at some of his stuff.
Do you disagree that "positive integers" is an infinite set? Do you disagree that there are infinitely many points in a square inch? Infinity is confusing, but there is good info about it in the BoI book and also a bit at https://curi.us/1955-explaining-infinite-sets-measures-and-mappings-for-quantum-physics
BTW, just because infinity is involved in a particular mathematical description of reality doesn't mean it's involved in *all* mathematical descriptions of reality. There may be equivalent ones which avoid infinity. Similar to how one can bring up infinity when talking about length, but also not bring it up.
> MWI is an *objective* theory based on *scientific observations and math*...
What scientific observations indicate MWI? Some people claim that experiments conducted at CERN support MWI, but I disagree.
#13054 There is an arbitrarily large number of positive integers, but it is a definite amount. That set cannot be infinite in size -- infinity cannot and does not exist in reality because it violates the Law of Identity.
I understand that there is a specific mathematical definition of an "infinite set", but that definition implies that there is a set of numbers (e.g. positive integers), that goes on forever. This, however, is impossible.
There is a specific number of positive integers that exist, but that number is so large we may never know the specific amount.
#13056 Are you a fan of Harry Binswanger?
#13055 To begin with, the two slit experiment. When you shine light on a screen with slits in it, so some light goes through to a viewing screen behind, you get patterns of light and shadow which are incompatible with classical physics. That's what FoR ch2 is about.
#13057 Yeah I am. His book "How We Know" has some great material that I agree strongly with. I presume you're familiar with his stuff
#13058 The two slit experiment shows that when we observe something as sensitive as a photon, it affects the trajectory of that particle and thus the outcome of the experiment is affected too.
Many people say that the double-slit experiment shows that light can be two different things simultaneously (wave and particle), but again that is impossible and violates the Law of Identity.
I do agree that Classical Mechanics is incomplete and does not explain everything with complete accuracy.
#13059 While I am a fan of Objectivism, I am not a fan of HB's theories on math. Quotes:
> Number concepts are a special case, along with a few others, of concepts that are so abstract that there’s nothing more to learn about them
> For instance, it is widely believed that there’s a number like: 10^100^100. There isn’t.
> I’m supportive of the original, 1-based Peano axiomatization.
Josh Jordan wrote:
>> According to Peano’s axioms, every natural number has a “successor,” that is, another natural number that is one bigger than it.
HB replied:
> To paraphrase Bill Clinton, it all depends on the meaning that “has” has.
I liked curi's post here, but HB did not reply to it:
http://curi.us/1921-the-harry-binswanger-letter-posts#c7468
i only ever seen one slit
#13061 I've never ready anything from Binswanger regarding math, but I have read "Mathematics is About the World" but Robert E. Knapp which is actually a good read.
Yeah, unfortunately its seems like some Objectivists dismiss ideas too quickly without even giving them a minute of consideration.
> For instance, it is widely believed that there’s a number like: 10^100^100. There isn’t.
That's just pure laziness on Binswanger's part
> Many people say that the double-slit experiment shows that light can be two different things simultaneously (wave and particle), but again that is impossible and violates the Law of Identity.
While I'm not a fan of that particular explanation, it does not violate the law of identity (that specific wording does, but that isn't a major problem, we can interpret or reword it in a better way). You just have to say "Light is a complex object with some wave-like properties and some particle-like properties." There's no need to phrase it as light being two different things at once, rather than being one complex thing.
> The two slit experiment shows that when we observe something as sensitive as a photon, it affects the trajectory of that particle and thus the outcome of the experiment is affected too.
I think you're saying that putting detectors on a non-final screen changes things. See this diagram (which has one more screen than needed):
But DD is aware of this objection and is careful to address it and present an experiment which it doesn't apply to. He compares the light patterns from two and four slits without any detectors at the slits in the barrier (first screen); we just look at the final screen to see the pattern of light and shadow created:
DD discusses opening up two more slits (going from 2 to 4 open slits, which adds more light sources, more ways for light to get through), e.g.:
> So, adding two more light sources darkens the point X; removing them illuminates it again. How?
That's mysterious. DD goes into great detail about what might explain it, how other experiments rule out some possibilities, etc.
https://blog.rongarret.info/2019/07/the-trouble-with-many-worlds.html
> > Branching indifference is not false. What is false is the proposition that an agent must accept branching indifference in order to be considered rational. That turns out to be the Born rule in disguise.
> >
> > To illustrate how a plausibly rational agent could reject branching indifference (this example is due to Tim Maudlin, though I've reframed it here): suppose I am offered a choice between chocolate or vanilla ice cream. I like them both, and would prefer to have both, but I can't. I have to choose (maybe I only have enough money for one scoop). Instead of choosing according to my free will, I instead perform a quantum experiment in order to split myself into two. One of me chooses chocolate, the other chooses vanilla. A rational agent could plausibly prefer this state of affairs (assuming they care about the multiverse *at all*) to one where only one flavor is actually chosen.
>
> What is the sense in which this is supposed to be rational? How would one enact this rule in a way that's consistent with quantum mechanics, having consistent preferences over time and so on?
Ron, branching indifference is not directly related to the Born rule, let alone equivalent. I think you must be combining it with some other ideas to reach something like a Born rule equivalent. You think that, along with some other premises, and via some reasoning, it *implies* the born rule. Right? If you disagree, please define branching indifference, define the born rule, and then point out the equivalence. If I'm right, please actually provide the reasoning involved.
Regarding the ice cream, I have a guess at what Ron might have in mind. Suppose I flip a coin so that I get each flavor in ~half of the universes. Then I can eat chocolate in a particular universe and think to myself "I would have liked to have both flavors; but I don't have them; but I know that, right now, versions of me are eating vanilla and that thought brings me a portion fo the satisfaction i would have gotten from personally having a half-portion of vanilla."
I think what's going on is the agent prefers to eat chocolate *and to have this thought* than to eat chocolate alone. In other words, the agent gets a higher reward, and more satisfaction from the ice cream, due to certain thoughts and the coin flipping action (his satisfaction from these thoughts depends on his belief this is really happening, and the coin flip enables that so that he isn't worried that he chose chocolate in ~all universes).
In this scenario, we're comparing alternatives with unequal rewards. So that is why one is preferred. This does not contradict branch indifference which is just saying that agents don't care about branching *indpendent of* any change in rewards.
If you drop the thought process about the multiverse from the scenario, then what is happening in those other universes cannot be relevant to the satisfaction/reward for an agent in a particular universe. That is, consider 2 hypothetical agents in separate multiverses. These are different scenarios. One eats chocolate and so do all his clones. The other eats chocolate but half his clones eat vanilla. Neither agent has any thoughts about what his multiversal clones are doing. Everything else being equal, the rewards are equal – rewards cannot depend on stuff you don't know about which never affects you in the future.
https://blog.rongarret.info/2019/07/the-trouble-with-many-worlds.html
> Specifically, I reject branching indifference as a precondition on rationality.
Oh, you *reject branching indifference as a precondition on rationality*. Either you didn't say it quite like that before (with the "as" part) or I missed it. That makes more sense to me.
So, do you accept that that branching indifference fits with a standard game theory view of a "rational agent"? Or do you deny that too? I think that's the relevant issue because the goal is not talking about rationality in general, just to see what conclusions one can reach based on non-probabilistic parts of QM and some game theory premises.
And a good next step would be to give a counter example (a rational agent violating branching indifference), although if you have a different way of arguing your point that'd be OK too. I think the ice cream scenario was intended to be a counter example, but I think I've answered that one.
https://blog.rongarret.info/2019/07/the-trouble-with-many-worlds.html
> (That's actually a pretty easy case to make: I believe that an agent could plausibly be rational in a purely classical world, i.e. without accepting quantum mechanics *at all*. So requiring the acceptance of branching indifference aaPoR is a lot to ask.)
I don't think it's a branching indifference violation to live in a classical world. It makes sense to be indifferent to impossibilities rather than to have preferences about them.
> That is a very peculiar counterfactual for a proponent of MWI to raise. If you "drop the thought process about the multiverse from the scenario", then the entire argument completely falls apart. You can't have it both ways: either the evidence logically compels the belief that MWI is true, or it does not. If the evidence compels the belief that MWI is true, then a rational agent who is aware of the evidence cannot simply decide to ignore this and act as if it weren't true. *That* really is a compulsory part of what it means to be rational.
If an agent is unaware of the multiverse, or not currently thinking about it, then *for many isolated scenarios*, a classical analysis is OK.
If the agent is thinking about the multiverse, then the state of the multiverse (as known to that agent's thoughts) is relevant to the agent's preferences and rewards, so there is no conflict with branching indifference. Of course, the agent may still be indifferent to the particular branching in question even if he considers it. And, of course, the state of the multiverse will also be relevant, whether the agent thinks about it or not, if it actually interacts with the agent ever again.
https://blog.rongarret.info/2019/07/the-trouble-with-many-worlds.html
> You appear to have lost the plot here.
But before that you said:
> Yes, and your restatement of my position was exactly right
Then when I restated part of my own restatement, talking about the same thing again, instead of it being "exactly right" you decided I'd lost the plot. I don't think you understood either of my messages.
Now, after claiming my refutation of your ice cream example was "exactly right", you claim you've already won the argument via the ice cream example. You are lost.
> And remember that you have to answer this question without reference to weights because that too is sneaking the Born rule in through the back door.
"Weights" are a fundamental, non-probabilistic part of QM which can be legitimately referenced. I think the biggest issue is that you are not on board with DD's premises about QM. So let's try this. Tell me which of the following quotes you think are false:
https://arxiv.org/pdf/quant-ph/0104033.pdf
> In other words, when such sub-networks are in identical states, they are *fungible*. The term is borrowed from law, where it refers to objects, such as banknotes, that are deemed identical for the purpose of meeting legal obligations. In physics we may define entities as fungible if they are not merely deemed identical but *are* identical, in the sense that although they can be present in a physical system in varying numbers or amounts, permuting them does not change the physical state of that system. Fungibility is not new to physics. Many physical entities, such as amounts of energy, are fungible even in classical physics: one can add a Joule of energy to a physical system, but one cannot later extract the same Joule.
> A multiset is like a set except that some of its elements are fungible. Each element is associated with an integer, its *multiplicity*, which specifies how many instances of it appear in the multiset.
> An ensemble is a limiting case of a multiset where the total number of elements M goes to infinity but all the proportions [...] tend to definite limits.
> I shall refer to a non-empty sub-ensemble in which all the computers are in the same state as a branch of the ensemble
> These properties give each branch a well-defined identity over time, even though the values of its bits change.
> There are such things as fungible processes as well as fungible objects.
> The effect of an *n*-qubit quantum gate during one computational step is to transform the *3n* matrices representing the *n* participating qubits into functions of each other in such a way that the relations (14) are preserved.
> in order to model information flow we are using local interactions (gates) of the network to model local interactions in general quantum systems
> Given the universality of the Toffoli gate, all these properties must hold whenever a quantum network, or any part of it, performs a classical computation. In other words, whenever any quantum network (including a sub-network of another network) is performing a classical computation f, the matrices [...] for that network evolve independently of all its other descriptors.
> Thus in any sub-network *R* of a quantum computational network where a reversible classical computation is under way, half the parameters describing *R* are precisely the descriptors of an ensemble of classical networks. It is half the parameters because, from (14), any two of the three components [...] determine the third. This does not imply that such a subsystem constitutes half the region of the multiverse in which *R* exists. Proportions in the latter sense – which formally play the role of probabilities under some circumstances, as shown in Deutsch (1999) – are determined by the Heisenberg state as well as the observables, and do not concern us here because the present discussion is not quantitative.
>
> The other half of the parameters, [...] contain information that is physically present in *R *(it can affect subsequent measurements performed on *R *alone) but cannot reach the ensemble (the descriptors of the ensemble being independent of that information). But the reverse is not true: as (19) shows, information can reach the quantum degrees of freedom from the ensemble.
> The proposition that parts of the multiverse have the same description as an ensemble with given properties is not quite the same as the proposition that such an ensemble is actually present in those parts of the multiverse, for the description might refer to entities that are not present in addition to those that are. In particular, an ensemble has an alternative interpretation as a *notional *collection, only one member of which is physically real, with the multiplicity of a given branch representing the probability that the properties of that branch were the ones prepared in the real system at the outset, by some stochastic process. However, no such interpretation is possible if the branches affect each other, as they do in general quantum phenomena, and in quantum computations in particular (see Benjamin 2001).
> When a quantum computational network is performing a general computation, it need not be the case that the descriptors of any part of the network over two or more computational steps constitute a representation of an evolving e-algebra. [...] so the conditions discussed in Section 3 for branches to have an identity over time need not hold.
> In a typical quantum algorithm, [...] the qubits first undergo a non-classical unitary transformation [...], then a reversible classical computation, and finally another unitary transformation which is often the inverse [...] of the first one. Despite the fact that the branches lose their separate identities during the periods of the quantum transformations [...] we can still track the flow of information reasonably well in terms of ensembles:
> Therefore, if some sub-network of a quantum network performs a classical computation for a period if the network is isolated, and then it is run with some or all of the observables [...] being repeatedly measured between computational steps, it will still perform the same classical computation and will contain an ensemble identical to that which it would contain if it were isolated (though its other descriptors will be very different).
> Since a generic quantum computational network does not perform anything like a classical computation on a substantial proportion of its qubits for many computational steps, it may seem that when we extend the above conclusions to the multiverse at large, we should expect parallelism (ensemble-like systems) to be confined to spatially and temporally small, scattered pockets. The reason why these systems in fact extend over the whole of spacetime with the *exception *of some small regions (such as the interiors of atoms and quantum computers), and why they approximately obey classical laws of physics, is studied in the theory of decoherence (see Zurek 1981, Hartle 1991). For present purposes, note only that although most of the descriptors of physical systems throughout spacetime do not obey anything like classical physics, the ones that do, form a system that, to a good approximation, is not only causally autonomous but can store information for extended periods and carry it over great distances. It is therefore that system which is most easily accessible to our senses – indeed, it includes all the information processing performed by our sense organs and brains. It has the approximate structure of a classical ensemble comprising ‘the universe’ that we subjectively perceive and participate in, and other ‘parallel’ universes.
Alan wrote:
https://blog.rongarret.info/2019/07/the-trouble-with-many-worlds.html
>> What is the sense in which this is supposed to be rational?
>
> In the sense that Deutsch defines in his original paper: "‘Rationality’ ... means conformity to a set of constraints on a decision maker’s preferences."
DD didn't pick constraints arbitrarily.
>> How would one enact this rule in a way that's consistent with quantum mechanics, having consistent preferences over time and so on?
>
> Quantum mechanics does not require a rational agent to have consistent preferences over time, what Wallace calls diachronic consistency. That is another assumption that Wallace and Deutch have to make in order to achieve their result. I could just as well have challenged that as branching indifference, but decided not to for the sake of brevity. Nonetheless, diachronic consistency is clearly not necessary to be rational. Rational agents can change their preferences, e.g. on the basis of new information. If diachronic consistency were required for rationality, no human being could ever be rational.
If you change your priorities then the way you value outcomes will change. But without diachronic consistency it's impossible for you to enact preferences consistently regardless of whether you change your mind.
> But I'd really prefer not to get into those weeds.
I'm puzzled. Do you want to discuss substantive issues or not? Are you just wasting your time and mine?
>> Wallace and Deutsch have given answers to those questions. You haven't AFAIK.
>
> I just did.
Above you denied that consistent decisions are possible because you might change your mind about something. Now you're saying you provided a consistent rule for rational decision making. Also, you haven't provided a detailed discussion of how one would go about enacting your rule, so you haven't provided answers.
>> You can measure the measure of a universe:
>
> That paper does not support your claim. The word "universe" doesn't even appear in it, nor does it appear on the Wikipedia page for "quantum tomography."
>
> You're going to have a very hard time explaining to me how you can measure the measure of a universe in light of the fact that "universe" is not actually a well-defined term under the MWI.
Universes are an approximation. If I pick some particular approximation and stick to it then there's no reason I can't discuss the measure of a universe.
> But that's kind of beside the point because the question I actually asked was not "can I measure the measure of *a* universe" but rather "Can *I* measure the measure of *the* universe that *I* inhabit?" And the answer to that question is: no (because that would allow me to violate the no-cloning theorem).
You can do experiments to test the evolution of the amplitudes of different possible outcomes a system over time. This doesn't require measuring the measure of the system you're currently in over the entire multiverse.
There is more physics discussion on the FI group. There are recent topics related to material in comments here. Example:
https://groups.google.com/forum/#!msg/fallible-ideas/9kn3pooykB0/ov_WEqqCBQAJ
Look for threads with [physics] in the subject line.
do u think dark matter is real
https://blog.rongarret.info/2019/07/the-trouble-with-many-worlds.html
I asked Ron to identify which quotes from a DD physics paper – that Ron claimed to have studied and understand, had had recommended to others – that he disagreed with. He said he agreed with all of them. One of them was:
>> An ensemble is a limiting case of a multiset where the total number of elements M goes to infinity but all the proportions [...] tend to definite limits.
So my next message was:
>> Ron, do you think that the multiverse is an ensemble? If so, of what and with what multiplicities?
Ron's answer was to *not know what an ensemble was*, look up some meaning on wikipedia, forget that DD had used the term, forget that my previous message had used the term, and say "no". The term "ensemble" appears in DD's paper *53* times.
When I presented the same DD quote to him again, Ron changed his ansewr to *yes* but still ignored the second part of the question. So I tried to ask again:
>> The multiverse is an ensemble *of what* with *what multiplicities*?
This is highly relevant because, over and over, Ron had argued with me about what the multiplicities of things were. But, apparently, he had no idea what he was talking about, because his answer to this question, finally, was this:
> I have no idea
If he has no position on the matter, why did he argue with me about it, deny my claims about it, claim to have done months of research and study in which he figured all the issues out, and even claim that he's the only person in the world who understands quantum mechanics correctly?
He also expressed that he won't be answering further questions:
> But I'm not playing your stupid Socratic game.
He also admited to being "belligerent" and doubled down on his belligerence by calling me a "twit".
But when it comes to knowing what an ensemble or multiplicity is, or what that has to do with MWI, he has no idea and nothing to say.
Youtube Video on why Gravity is not a Force
Recent video (October 9, 2020) released by Veritasium titled: Why Gravity is NOT a Force
https://www.youtube.com/watch?v=XRr1kaXKBsU - 17.5 min
#18355 Thanks. I watched a few of his other videos and will try subscription to his video notifications. I commented some on a different vid at http://curi.us/2379-curis-microblogging#18357