So if a the territory is branching , the map.should, too.
Of course not. The territory can be made of rocks and dirt, but it doesn't mean that the map also has to be.
That's a disadvantage, because the same map can't represent any territory.
I'm not saying that it represents every territory. I'm saying that it represents a more general class of territories without loosing any advantages of the framework of possible worlds.
In the post I've even specifically outlined what are the territories that my framework can represent:
"So the territory that probability is in the map of is..."
All the processes in the real world, such that my knowledge state about them works like a weighted sample of n elements.
( there may be an ontological neutral way of doing probability calculations, but it's not a map, for that reason....more of a tool)
Now it seems that you are finally starting to get it. It is a map, of course, but that's really beside the point. If you want to put it in a separate category of "tools" (as if a map is not a tool?) then whatever suits your needs. Yes, what I'm doing is providing an ontologically neutral framework for probability theory that works better than a framework of possible worlds.
The problem is the implied ontology.
Once again, no ontology is actually implied. It's absolutely trivial to describe behavior of indetermenistic processes in terms of probability experiment. I'm concentrating on deterministic cases simply because they are trickier.
If there is a referent for it in the territory, it is entirely reasonable to say "possible worlds exist".
I'm not saying that it's unreasonable to say, conditionally on using this term. I'm saying that usage of the term, to begin with, is a bad idea as it keeps leading people doing probability theory astray.
Is it really a win to admit the substance of existing possible worlds, but under a different name?
When your goal is to separate the substance from the harmful rubbish, it absolutely is a win.
You keep missing the point. It's as if you haven't even read the post and simply noticed a couple of key words.
The map that corresponds to a deterministically branching multiversal has possible worlds.
Some do.
I'm proposing a better map, capable to talk about knowledge states and uncertainty, in any circumstances, having all the advantages of maps using the concept of possible worlds, without their weak point.
If you think that the framework of probability experiment that I'm outlining in the post fails to account for something that the frameworks of possible worlds manage to account for - please specify it. Bring up a setting in which you think my framework fails to describe the territory.
Refusing to.ever talk about possible worlds is dangerous, because they might exist (they do in MWI ) and they might be useful otherwise.
Possible world is a term from a map. There may be a referent for it in a territory, true. But it doesn't mean that we have to use this particular term to talk about this referent. We may have a better term, instead.
If..it was pointed out a long time ago that (a form of) probability being in the mind doesn't imply (a firm of) it isn't in the territory as well.
That not true because fundamental determinism is true , or because effective determinism at the macroscopic level is true.
This is beside the point that I'm making. Which is: even if we grant that the universe is utterly deterministic and therefore probability is fully in the map, this map still has to correspond to the territory for which you have to go an look. And we still have to be able to construct a meaningful framework for it.
Armchair arguments can't prove anything about the territory...you have to look.
Exactly.
You may be beating a dead horse there. Talk of possible worlds doesn't have to imply realism about possible worlds, just as mathematical anti-realists can talk about numbers without committing to their mind independent existence.
I'm not saying that it does. For instance here I specifically outline alternative option:
"Well, you don't necessary have to believe that there are parallel worlds as real as ours in which the coin comes differently, though it's a respectable position about the nature of counterfactuals. Probability is in the mind, remember? You can simply imagine alternative worlds that are logically consistent with your observations."
What I'm saying is that even the talk itself about "possible worlds" - without assumption of their realism - is harmful as this framework leaves us unable to reason about logical uncertainty and doesn't provide proper guardrails against absurdities, most noticeably in indexical uncertainty.
A better way is to talk about iterations of probability experiment which solves all these issues.
I don't see problems here.
The problem is that according to the framework of "possible worlds" we technically need to be able to do a thing that we can't do. The solution to this problem is to use a better framework - the one of probability experiments.
You should imagine a part of the world, not the whole world, including orbits of start in an other galaxy.
My point exactly. In actuality we simply approximate a particular process of our universe to the best of our knowledge, instead of imagining all the universal permutations and then filtering from them the ones logically consistent with our knowledge.
Disagree. For example, if you have a dice that is symmetric by form and by weight with 4 sides, you are sure that if you roll it 100 times you will have around 25 results for each side.
And this belief of mine is grounded in actual behavior of dice in our physical reality, which I would never be able to get without going and checking the way reality works. I don't think we actually have any disagreement here.
I don't see which direct experiment you can use to figure out whether we are in a simulation or not that isn't extremely dangerous (searching for bugs in the simulation? Suicide?), so we should use other methods.
The point isn't that we can perform some kind of experiment to distinguish between simulation/no simulation conditionally on its result. The point is what kind of prior one should use - what mathematical model is appropriate for the situation that we are talking about. And the answer is: the kind of model that approximate the behavior of the universe.
I still don't see. Can you be even more direct?
The reason why we can say that picking a stone from a bag is a random sample from all the stones in a bag is because there is an actual causal process determining which stone will be picked from the stones in the bag, which we approximate to the level of our knowledge.
But there is no such process for your being alive in simulation or no simulation. There is no God, who from beyond time considered whether creating you in actual 21st century or its future simulation based on the total number of people there. That's not how our universe works to the best of our knowledge. The existence of such process would contradict relativity as it would be sending information back in time. And so we can't assume a random sample from all the people.
"Stones" is not a good class with clearly defined boundaries (like humans or potatoes)
Nothing is. We are dealing with abstractions and approximations of reality, in probability theory especially. And yet some approximations are correct while others are not.
Reference class "all stones in the bag" use all information we have, so it's the best.
We've been through that already. We have all kind of information, but still truly take in account only some of it, while constructing our mathematical models.
In fact, reference class should be the space of possibilities.
I fully agree. The whole term "reference class" is very misleading with its apparent arbitrariness. What we actually need to be talking about is the sample space, consisting of mutually exclusive and collectively exhaustive outcomes of a probability experiment.
P. S.: can you just write your argument directly? It took too much time to ask questions, so it's inefficient.
My apologies. From now on I'm going to be direct.
Probability is in the map but this map has to correspond to the territory. And the only way to have a relation of map territory correspondence also known as "truth" is to go and check.
So how do we know which outcomes are possible in a particular real world situation? We conduct an experiment. We blindly pick a stone from the bag multiple times and notice that every time this is a stone from the bag and not a stone from somewhere else on the planet Earth. This isn't just a fact about our knowledge state. It's the way reality actually works. If physics was different, say there was a portal in the bag leading to stones in some far away place, then your experiment would produce different results. On the other hand, if you simply believed very very strongly that there is a portal, while in fact there were none, you would still simply get one of the stones from the bag.
So when we say that something is a random sample from something or that something is a possible outcome, we mean, that reality actually works this way. That there is a certain causal process that leads to you getting a stone from the bag and not from somewhere else when you follow a particular experimental procedure. And you are describing this behavior of reality, to a certain approximation, with math. Here I talk about it in more details.
With that in mind, do you see why you are not a random sample from all people who has ever thought or will be thinking that they live in 21st century, while a stone blindly picked from the bag is a random sample from all the stones in the bag?
You are perhaps interpreting his interactive dualism as substance dualism
You seem to be nitpicking definitions. Let's try to grasp the substance. Eliezer was initially distinguishing between two types of dualism:
His Zombie post were about the first one. This post is about the second one.
If you want to talk about some sub-type of the second that manages to evade the argument in this post - be my guest.
The point of arguing for zombies is to argue for non physicalism. Zombies are not the only way to argue for nonphysicalism. Assuming non physicalism to argue against zombies is therefore a pointless way of arguing for physicalism.
I'm not sure what this has to do with the citation.
You seem to be arguing that interactive dualism must be disguised physicalism.
It's not yet an argument, just a vibe according to which we can arrive to one. The actual argument is presented further below.
But that can be disproved by putting forward a non arbitrary criterion of the physical and non physical, as we shall see.
A non-arbitrary criterion would indeed be helpful for rescuing Zombie argument from this particular critique. But merely saying the word "subjective" doesn't help much. You need to actually prove that such subjective things exist in a sense where they are not also objective and have a method to discern whether a particular thing is subjective or objective in this sense. Otherwise, one can just say that electrons are "subjective" and we are back to square one.
"Physics" doesnt mean "the only way the world could possible be". Physics focuses on objective (because results need to be confirmed by multiple scientists) and quantifiable (because physics uses maths as its language). If it happen to be the case that there is something in the universe that is subjective or unquantifiable,then physics can't get a grip on it.
I think you are confusing physics-map and physics-territory here. For the sake of Zombie-argument we care about the territory - the actual way the matter in the universe behave, regardless of our epistemological limitations.
Imagine a fully materialistic universe strictly following some laws, which are such that no agent from inside the universe is able to fully comprehend them. Would you say that it is enough to declare that such universe is dualistic in nature, even if there are no qualia involved?
And Chalmers has a version of this argument, in The Conscious Mind...he characterises the physical as the structural/functional.
I would appreciate if you present this argument here or in a separate post, showing how it refutes my point, instead of simply proclaiming it. With a proper definition of "functional" and "structural" in this context, of course.
It's actually contradiction, not falsehood.
Naturally, falsehoods being true is a contradiction.
Recent discussion:
Sadly I'm not able to access Astral Codex now, so if you posted something relevant there in the comments - I would appreciate if you re-posted it here.
There is no such thing as The One Truly Perfect Class. All of these are rough estimations; some are better than others. It's better to use "all stones in the multiverse" than to use nothing, but if you have a choice between all stones in the multiverse and all stones on Earth, use the latter as the reference class.
But why is one reference class more preferable than the other? What does determine it? How do we know that it's better to use "all the stones on Earth" than "all the stones in the multiverse"? And even better still to use "all the stones in this particular bag in this particular moment"? But not better to use "two specific stones from the bag"?
Probabilities are in the mind, and you use a reference class because you can't calculate the trajectories of stones in the bag — and you have good reasons to believe there is an equal probability for each stone to be chosen (since they're all in the bag).
True. Probability is in the map. Which is an approximation of the territory. And yet some maps are more accurate than the others. Some do correspond to the territory - approximate it correctly to the degree that they claim to, and some do not. How do we know whether a particular map corresponds to the particular territory? And what is the territory for probability theory, to begin with?
Maybe you have to consider all the info you have, so you can't use all the stones in the multiverse as a reference class if you already know what's in the bag?
Consider the information you have:
And yet from all this information about stones - mind you, we haven't even begun talking about all information you have - what is relevant to the problem of blindly picking a stone from a particular bag is specifically the information about composition of stones inside it at the moment of picking and not anywhere and anywhen else in space and time.
But how do we know what is relevant? What is the principle that determines it? And is this principle solely about your knowledge state - in the map - or does it have something to do with the territory that your map is trying to approximate?
And the stones in this bag strongly influence the chance of picking one, unlike stones in a different bag.
Good, I think you are looking in the right direction. Now let's taboo the words "influence" and "chance". Try to formulate the same idea in terms of physical laws of the universe and processes that go accoding to these laws.
Good overall, but you are making a serious mistake: confusing single halfism, with double halfism.
SSA is a generalized single halfism reasoning. It's very obviously wrong in Sleeping Beauty as it implies that if the Beauty knows that she is awakened on Monday, she expects that there is 3/2 chance that the coin is Heads. Generally if you actually do the math, SSA can't produce correct betting odds for Sleeping Beauty problem. It's, in a sense, even worse than SIA for SB, but ultimately both of them are based on the flawed and unjustified framework of "centred possible worlds".
Double halfism is the correct approach and has no problem with correct betting odds. All your argument in favor of halfism are, in fact, double halfer reasoning. However you mistakenly credit them to SSA.