Sorted by New

Wiki Contributions


Why would ‘scheming’ not be the best way to turn compute into rewards? Why would a completely honest, consistent, straightforward approach be the most rewarded one, given how humans decide how to reward things? I don’t get it.

It wouldn't be the most rewarded one. It's not what training does, because it never approaches theoretical limits. It would be uncomplicated approach that may miss some reward. In other words, the solution to soft optimization is to train a neural network with human-generated content.

You know you have subjective experience, self-evidently. You can match the computation run by the neural circuitry of your brain to the computation run by the neural circuitry of other humans: because since they talk about subjective experience, you can expect this to be caused by similar computation.

Similarity is subjective. There is no fundamental reason that the ethical threshold must be on the level of similarity between humans and not on level of similarity between humans and shrimps.

When we hear someone about qualia, we can make a valid inference that this is caused by qualia existing or having existed in the past.

When we hear someone talking about a god, we can make a valid inference that this is caused by a god existing or having existed in the past.

we could figure out how qualia actually works, and then scan brains and see whether there are circuits implementing it or not.

Whether circuits implement something is subjective - on the physical level the circuits in other humans' brains don't implement your qualia. If you generalize to other humans' implementations, what's stopping you from generalizing to anything with pain receptors?

something that we think could be a part of how qualia works exists in that species.

So pain receptors?

This seems like the most important crux. Why should we not expect the maximizer we trained to X-maximize to use its affordances to maximize X’, where X’ is the exact actual thing the training feedback represents as a target, and that differs at least somewhat from X? Why should we expect to like the way it does that, even if X’ did equal X? I do not understand the other perspective.

Because that's what happens - humans don't always wirehead and neural networks don't always overfit. Because training feedback is not utility, there are also local effects of training process and space of training data.

That is correct as stated but somewhat misleading: the problem is why is it necessary, in the case of experience, and only in the case of experience to instantiate it in order to fully understand it.

Here - "fully understand" depends on definition of "understand". What you understand is not a matter of fact, it's a matter of definition. All you talk about is how it is "counterintuitive" to call instantiating nuclear reaction in yourself "understanding". "It's intuitive to call new experience "additional knowledge"" is an argument from definitions.

There seem to be some edge cases.: for instance, would an alternative Mary know everything about heart attacks without having one herself? Well, she would know everything except what a heart attack feels like, and what it feels like is a quale. the edge cases, like that one, are cases are just cases where an element of knowledge-by-acquaintance is needed for complete knowledge. Even other mental phenomena don’t suffer from this peculiarity. Thoughts and memories are straightforwardly expressible in words, so long as they don’t involve qualia.

They are only edge cases of specific definitions of knowledge. There is no fundamental reason why you must call "knowledge" heart attack's effect on your brain and not call "knowledge" fire's effect on your hand.

The fact that these physicalists feel it would be in some way necessary to instantiate colour, but not other things, like photosynthesis or fusion, means they subscribe to the idea that there is something epistemically unique about qualia/experience, even if they resist the idea that qualia are metaphysically unique.

"Necessary" for what? Judging from "epistemically unique" it is implied that it is necessary for knowledge? Then it's certainly incorrect - it's either not necessary, because Mary can have a more compact representation of knowledge about color, or it's necessary for all things, if Mary supposed to have all representations of knowledge. It may be necessary for satisfying Mary's preferences to have qualia independently of their epistemic value - that's your perfectly physicalist source of subjectivity.

If you only care about matters of fact, then there are no problems for physicalism in that the human qualia are unusual - it predicts that different neural processes are different. And predicts that it's useful to see things for yourself. And that it will feel intuitive to say "Mary gets new knowledge" for some people. I think it even follows from casual closure, that it doesn't make sense for there to be unphysical explanation for intuitions? If your intuition is not predicted by physics, then atoms somewhere have to be unexpectedly nudged - is it what you propose? I... don't really understand the argument here? The physicalism doesn't say that all things that it is intuitive to call "knowledge" are equally easy to get from books, or something - why exactly it is an argument against physicalism that Mary gets what it predicts?

No,because you can’t prove things through definitions.

There’s no fact of the matter about that. If they are fully represented , then Mary would know what red looks like, otherwise not. If we could perform M’s R as a rela experiment, we would not need it as a thought experiment.

Wait, is the problem that you actually think that it is not obviously physically possible to imagine red without seeing it? Like, knowing everything plausible includes having all permutations of neuron states, including the state where you are seeing red. Is your "matter of fact" about knowing what it is like to see considers the possibility that without actually seeing Mary could only simulate zombie-red or something?

Oh, I finally got why are you talking about predicting novel qualia - you are saying that physicalism doesn't predict Mary seeing red, right? Because it only predicts neural activity. My point is that this complain doesn't have anything to do with Mary or knowledge. If you only talk about Mary, then there is no motivation to doubt physicalism from the experiment. The point of Mary is that she gains knowledge and physicalism predicts gaining knowledge. There is no need to talk about novel qualia, because physical knowledge contains knowledge about differences between different, old and novel, qualia. You agree, that physicalism at least (allows definition of knowledge where it) predicts gaining some knowledge from instantiation when Mary leaves room, right? Then even if you have doubts about this predicted knowledge being incomplete, Mary doesn't provide anything that justifies this doubt - your arguments about insufficient gears-level explanations would work the same way in situations without novel qualia or complete physical knowledge. Or do you have an example of specific difference between qualia that is not predicted by physicalism and uniquely depends on the whole instantiation thing? I mean, my position is that there are no differences between qualia that are not predicted by physicalism at all, so any examples would be appreciated.

Show me a prediction of a novel quale!

I predict, that if you open this link, you will experience red hair:

Novel relative to what epistemic state? Sure, we probably can't ethically and consistently make a human say "wow, it was neither sight nor hearing" now, but I really don't get what's the justification for ignoring other facts about qualia that physicalism can predict? Some of them were novel for humanity in their time.

Says who?


We have a detailed gears-level explanation of fire, we do not have one of conscious experience.

We don't usually have very detailed explanations of specific fires. And we have detailed explanation of conscious experience - physics equations^^. But ok, there is a space for more useful theories. The thing I don't understand is how it is an argument against physicalism - do you expect to not get gears-level explanation in the future? The whole point of doing Mary is that no one expects it.

Merely saying that “X is an emergent, high level phenomenon..but don’t ask me how or why” is not an explanation, despite what many here think.

Yes, but that would just mean that the correct position is "physicalism is right, but the detailed explanation is in the works". Not detailed-enough explanation at the present moment is just one of factors you weight, along with "physicalism has detailed explanations about physics, neurons and all other things", not something that logically prohibits believing in physicalism. Again, that's not what mainstream arguments against physicalism are? It's always "physicalism can't possibly explain consciousness even if it's explanation have been detailed".

i) showing that two things are necessarily, not arbitrarily linked.

That's what I am against - it's not justified, depending on what do you mean by "necessarily" - atoms are not necessarily linked to fire. In the end, we just arbitrary call some atoms "fire". So why demand this only for qualia? If it's only "as necessary as reduction of fire", than it is already that necessary - the expectation that you will get neurological explanation in future is the same kind of inductive reasoning that you do, when you decide that correlations between atoms and fire are enough to believe explanation in terms of atoms.

Lots of things are non contradictory. Non circularity is more of an achievement.

Such physical definitions of knowledge are not more circular than anything, I think?

So maybe I could arbitrarily assume that definition?

I mean, go ahead - then Mary would just be able to imagine red.

Again, you can’t prove things by adopting definitions.

Exactly - that's why Mary doesn't work.

If we had a detailed understanding of neuroscience that predicted an illusion of knowledge-by -acquaintance specifically, you’d be onto something. But illusionist claims are philosophical theories, not scientific facts.

There is no need for additional scientific facts. There are enough scientific facts to accept physical explanation of the whole Mary setup. That's why people mostly seek philosophical problems with physicalism and why physicalists answer with philosophical theories - if physicalism is philosophically coherent, then it is undoubtedly true.

That would be the case if physicalism is true, but you don’t know that physicalism is.true..You basically assumed it, by assuming that physical explanations are complete. That’s circular.

The Mary's room was supposed to be an argument against physicalism. If there are no philosophical problems in the setup after you assume physicalism, then argument fails. It is equivalent to disagreeing with some step of an argument, like "Mary gets new knowledge" - you can't just disallow disagreeing with this because it's logically equivalent to assuming physicalism - that would be assuming non-physicalism that the argument was about. Of course, I don't just assume physicalism - you need to satisfy the "no philosophical problems" condition, so I talk about why "Mary gets new knowledge" is just trying to prove things by adopting definitions. I don't see how do you think it can work otherwise - you can't derive "physicalism is true" from Mary's assumptions alone. Obviously, assuming physicalism doesn't prove that physicalism is true. But again, I don't argue, that physicalism is true, I'm arguing that Mary is a bad argument.

There is a fact of the matter about whether physical descriptions are exhaustive, even if Mary’s Room doesnt prove it. If physical descriptions don’t convey experiences as such , they are fundamentally flawed , and the precision isn’t much compensation.

Sure. So you do agree now that talking about Mary or knowledge is unnecessary?

They are not ignored, they’re represented by corresponding neural processes. Like, what is ignored and not explained by a physical description?

The experience itself.

So, what is your argument against "experience itself is explained by "human experiences are neural processes"", if it's not Mary?

Experience need to be explained because everything needs to be explained. Experiences need not end up in the final ontological model, because sometimes less an explanation explains-away.

If you don't demand specific experiences to be in the final ontological model, they are explained the same way the fire is explained. The explanation of fire does not usually set you on fire. What you call "I'm seeing blue" is actually "your neurons are activated in a way similar to a way they are activated when blue light is directed to your eyes". On what basis then you say that these 3gb of numbers from a simulation do not explain fire?

I don't think 3 depends on UDT that much? Like, you can call the betting argument in sleeping beauty "a UDT-like reasoning", but then it's not like people were convinced either way without UDT. And about it being right, do you mean that it would be practically convenient if there was some logical law that forced our preferences about multiple copies? Because the way I see it, it's not problematic from philosophical standpoint.

I feel like you’re dodging the question here.

Not intentional - I'm just not sure whether you see problems with casual closure or with epistemic usefulness.

You can’t just say “if the universe didn’t exist, then lists X and Y wouldn’t exist either”. Sure, maybe that’s true, but it doesn’t constitute any progress towards explaining why the X&Y list contents are what they are. Right?

I wouldn't say "any progress" - correct propositions about X and Y are correct even if they may seem trivial. And it sure would be a progress for someone who was forced to believe in dualism or worse as an alternative. And, to be clear, consciousness having a content is not a problem for casual closure - if some specific universe didn't exist and some other universe existed instead, X and Y would be different. But yes, it's a solution that is not especially useful except for a narrow purpose of solving the Hard Problem.

I feel like you’re angling for a position kinda like: “cosmopsychism explains why I have a conscious experience, but explains nothing whatsoever about any of the properties of that conscious experience”. Right?

Right. Well, if you stretch the definition of an explanation and properties, there are some vague intuition-like mental processes that I believe become streamlined when you accept cosmopsychism. Like, at the stage of "we have no idea how to solve the Hard Problem but I'm sure physicalism will win somehow" people still manage to hope for some kind of moral realism about consciousness, like there is objective fact that someone is in pain. But yeah, you may derive all this stuff from other sources too.

But, why do you think it's insane? There are no philosophical problems with relation of some mental processes to memory. Science will explain it in the future just fine. "Why there’s conscious experience" always was the only mysterious problem about consciousness. And I'm not even saying that the Hard Problem and it's solution is interesting, while practical theory of awareness is boring and useless. It's just as the matter of fact under some reasonable definitions cosmopsychism solves the Hard Problem - that's the extent of what I'm arguing.

then I reject that the thing you’re saying is actually the explanation of why there’s conscious experience.

The point is that cosmopsychism together with usual science provides the explanation you want. And no one doubts that science will do it's part.

What exactly (if anything) was he introspecting upon, and how was he doing so, and how did he interpret the results in order to choose one sentence over the other? If phenomenal consciousness is anything at all, it presumably needs to be the thing that Chalmers is somehow able to query during this introspection process, right? So how did his brain manage to do that?

The thing that Chalmers queries is his brain. The phenomenal nature of his brain is that his brain exists. Chalmers can't query a brain that doesn't exist. Therefore phenomenal things cause Chalmers to say “I am currently experiencing the ineffable redness of red.”.

But I don’t understand how that has anything to do with the question above.

Existence have something to do with everything - you can't introspect or see red if you don't exist. But yes, the solution to the Hard Problem doesn't have much to do with human qualia specifically (except maybe in the part where cogito ergo sum is the limit of reflectivity and awareness in humans have something to do with reflectivity) - if you explain the consciousness itself using physical notion of existence, then the redness of red is just the difference of neural processes.

You’re calling it “cosmopsychism” but I don’t see any “psych” in it…

That it doesn't have any magical “psych” is by design - that's why it's not dualism. The relevant phenomenal aspect of existence is that it solves zombies. And I mean, sure, you can avoid using the word "consciousness" and stick only to "existence". But it connects to your intuitions about consciousness - if you imagined you may lose consciousness if you were to be disassembled to atoms and reassembled back, now you have a direct reason to imagine that it still would be something to be like reassembled you.

Not sure about references, maybe Panpsychism and Panprotopsychism and The Combination Problem for Panpsychism by Chalmers?

The sane variant is cosmopsychism (because real physical objects are not fundamentally separable) with probably only one intrinsic property - existence.

When future scientists tell the whole story of how QFT+GR leads to David Chalmers writing books about consciousness, I think there will be no room in that story for the “intrinsic phenomenal properties of quarks” (or whatever) to be involved in that story, in a way that would properly connect those intrinsic properties to David’s introspection process.

From your linked paper:

If one holds that physical terms refer not to dispositional properties but the underlying intrinsic properties, then the protophenomenal properties can be seen as physical properties, thus preserving a sort of materialism.

The story is about intrinsic properties, because that's what equations describe - when you describe some physical system, it is implied that described physical system exists and all it's casual influence is because of it's existence. And you introspect the existence of universe itself by cogito ergo sum.

Load More