Can anyone explain why epiphenomenalist theories of consciousness are interesting? There have been an awful lot of words on them here, but I can’t find a reason to care.
It seems that you get similar questions as a natural outgrowth of simple computational models of thought. E.g. if one performs Solomonoff induction on the stream of camera inputs to a robot, what kind of short programs will dominate the probability distribution over the next input? Not just programs that simulate the physics of our universe: one would also need additional code to “read off” the part of the simulated universe that corresponded to the camera inputs. That additional code looks like epiphenomenal mind-stuff. Using this framework you can pose questions like “if the camera is expected to be rebuilt using different but functionally equivalent materials, will his change the inputs Solomonoff induction predicts?” or “if the camera is about to be duplicated, which copy’s inputs will be predicted by Solomonoff induction?”
If we go beyond Solomonoff induction to allow actions, then you get questions that map pretty well to debates about “free will.”
one would also need additional code to “read off” the part of the simulated universe that corresponded to the camera inputs. That additional code looks like epiphenomenal mind-stuff.
I don’t understand why the additional code looks like epiphenomenal mind-stuff. Care to explain?
I take Carl to meaning that: the program corresponding to ‘universe A simulating universe B and I am in universe B’ is strictly more complex than ‘I am in universe B’ while also predicting all the same observations, and so the ‘universe A simulating universe B’ part of the program makes no difference in the same way that mental epiphenomena make no difference—they predict you will make the same observations, while being strictly more complex.
the program corresponding to ‘universe A simulating universe B and I am in universe B’ is strictly more complex than ‘I am in universe B’ while also predicting all the same observations, and so the ‘universe A simulating universe B’ part of the program makes no difference in the same way that mental epiphenomena make no difference—they predict you will make the same observations, while being strictly more complex.
True, but, just as a reminder, that’s not the position we’re in. There are other (plausibly necessary) parts of our world model that could give us the implication “universe A simulates us” “for free”, just as we get “the electron that goes beyond our cosmological horizon keeps existing” is an implication we get “for free” as a result of minimal models of physics.
In this case (per the standard Simulation Argument), the need to resolve the question of “what happens in civilizations that can construct virtual worlds indistinguishable from non-virtual worlds” can force us to posit parts of a (minimal) model that then imply the existence of universe A.
The code simulating a physical universe doesn’t need to make any reference to which brain or camera in the simulation is being “read off” to provide the sensory input stream. The additional code takes the simulation, which is a complete picture of the world according to the laws of physics as they are seen by the creatures in the simulation, and outputs a sensory stream. This function is directly analogous to what dualist/epiphenomenalist philosopher of mind David Chalmers calls “psychophysical laws.”
I don’t know if this insight is originally yours or not, but thank you for it. It’s like you just gave me a piece of the puzzle I was missing (even if I still don’t know where it fits).
Oh wow… I had been planning on writing a discussion post on essentially this topic. One quick question—if you have figured out the shortest program that will generate the camera data, is there a non-arbitrary way we can decide which parts of the program correspond to “physics of our universe” and which parts correspond to “reading off camera’s data stream within universe”?
Pretty much the same reason religion needs to be talked about. If no one had invented it wouldn’t be useful to dispute notions of god creating us for a divine purpose, but because many people think this indeed happened you have to talk about it. It’s especially important for reasonable discussions of AI.
Religion and epiphenomenalogy differ in three important ways:
Religion is widespread. Almost everyone knows what it is. Most people have at least some religious memes sticking in their heads. A significant fraction of people have dangerous religious memes in their heads so decreasing those qualifies as raising the sanity waterline. Epiphenomenalogy is essentially unknown outside academic philosophy, and now the lesswrong readership.
Religion has impact everywhere. People have died because of other people’s religious beliefs, and not just from violence. Belief in epiphenomenalogy has almost no impact on the lives of non-believers.
Religious thought patterns re-occur. Authority, green/blue, and “it is good to believe” show up over and over again. The sort of thoughts that lead to epiphenomenalogy are quite obscure.
The word “epiphenomenalogy” is rare. The actual theory seems like an academic remnant of the default belief that ‘You can’t just reduce everything to numbers, people are more than that.’
So your last point seems entirely wrong. Zombie World comes from the urge to justify religious dualism or say that it wasn’t all wrong (not in essence). And the fact that someone had to take it this far shows how untenable dualism seems in a practical sense, to educated people.
The most famous arguments for epiphenomenalism and zombies have nothing to do with religion. And we dont actually want to have reductive explanations of qualia, as you can tell from the fact that we can’t construct qualia - - we can’t write code that sees colours or tastes flavours. Construction is reduction in reverse.
Epiphenomenalogy is essentially unknown outside academic philosophy, and now the lesswrong readership.
I’d say it’s more widespread than that. Some strands of Buddhist thought, for instance, seem to strongly imply it even if they didn’t state it outright. And it feels like it’d be the most intuitive way of thinking about consciousness for many of the people who’d think about it at all, even if they weren’t familiar with academic philosophy. (I don’t think I got it from academic philosophy, though I can’t be sure of that.)
It’s good practice for seeing how the rules play out; and having a clear mental visualization of how causality works, and what sort of theories are allowed to be meaningful, is actually reasonably important both for anti-nonsense techniques and AI construction. People who will never need any more skill than they currently have in Resist Nonsense or Create Friendly AI can ignore it, I suppose.
Philosophers have a tendency to name pretty much every position that you can hold by accepting/refusing various “key” propositions. Epiphenomenalism tends to be reached by people frantically trying to hold on to their treasured beliefs about the way the mind works. Then they realise they can consistently be epiphenomenalists and they feel okay because it has a name or something.
Basically, it’s a consistent position (well, Eliezer seems to think it’s meaningless!), and so you want to go to some effort to show that it’s actually wrong. Plus it’s a good exercise to think about why it’s wrong.
In my experience, most philosophers are actually pretty motivated to avoid the stigma of “epiphenomenalism”, and try instead to lay claim to some more obscure-but-naturalist-friendly label for their view (like “non-reductive physicalism”, “anomalous monism”, etc.)
People don’t like epiphenomenalism per se, they feel they are forced into it by other claims they find compelling. Usually some combination of
Qualia exist in some sense
Qualia can’t be explained reductively
3.The physical world is causally closed.
In other words, 1 and 2 jointly imply that qualia are non physical, 3 means that physical explanations are sufficient, so non physical qualia must be causally idle.
The rationalist world doesn’t have a clear refutation of of the above. Some try to refute 1, the Dennett approach of qualia denial. Others try to refute 2, in ways that fall short of providing a reductive explanation of qualia. Or just get confused between solving the easy problem and the hard problem.
Can anyone explain why epiphenomenalist theories of consciousness are interesting? There have been an awful lot of words on them here, but I can’t find a reason to care.
It seems that you get similar questions as a natural outgrowth of simple computational models of thought. E.g. if one performs Solomonoff induction on the stream of camera inputs to a robot, what kind of short programs will dominate the probability distribution over the next input? Not just programs that simulate the physics of our universe: one would also need additional code to “read off” the part of the simulated universe that corresponded to the camera inputs. That additional code looks like epiphenomenal mind-stuff. Using this framework you can pose questions like “if the camera is expected to be rebuilt using different but functionally equivalent materials, will his change the inputs Solomonoff induction predicts?” or “if the camera is about to be duplicated, which copy’s inputs will be predicted by Solomonoff induction?”
If we go beyond Solomonoff induction to allow actions, then you get questions that map pretty well to debates about “free will.”
I don’t understand why the additional code looks like epiphenomenal mind-stuff. Care to explain?
I take Carl to meaning that: the program corresponding to ‘universe A simulating universe B and I am in universe B’ is strictly more complex than ‘I am in universe B’ while also predicting all the same observations, and so the ‘universe A simulating universe B’ part of the program makes no difference in the same way that mental epiphenomena make no difference—they predict you will make the same observations, while being strictly more complex.
This seems to be talking about something entirely different.
True, but, just as a reminder, that’s not the position we’re in. There are other (plausibly necessary) parts of our world model that could give us the implication “universe A simulates us” “for free”, just as we get “the electron that goes beyond our cosmological horizon keeps existing” is an implication we get “for free” as a result of minimal models of physics.
In this case (per the standard Simulation Argument), the need to resolve the question of “what happens in civilizations that can construct virtual worlds indistinguishable from non-virtual worlds” can force us to posit parts of a (minimal) model that then imply the existence of universe A.
Ah, ok, that makes sense. Thanks!
The code simulating a physical universe doesn’t need to make any reference to which brain or camera in the simulation is being “read off” to provide the sensory input stream. The additional code takes the simulation, which is a complete picture of the world according to the laws of physics as they are seen by the creatures in the simulation, and outputs a sensory stream. This function is directly analogous to what dualist/epiphenomenalist philosopher of mind David Chalmers calls “psychophysical laws.”
I don’t know if this insight is originally yours or not, but thank you for it. It’s like you just gave me a piece of the puzzle I was missing (even if I still don’t know where it fits).
Oh wow… I had been planning on writing a discussion post on essentially this topic. One quick question—if you have figured out the shortest program that will generate the camera data, is there a non-arbitrary way we can decide which parts of the program correspond to “physics of our universe” and which parts correspond to “reading off camera’s data stream within universe”?
Pretty much the same reason religion needs to be talked about. If no one had invented it wouldn’t be useful to dispute notions of god creating us for a divine purpose, but because many people think this indeed happened you have to talk about it. It’s especially important for reasonable discussions of AI.
Religion and epiphenomenalogy differ in three important ways:
Religion is widespread. Almost everyone knows what it is. Most people have at least some religious memes sticking in their heads. A significant fraction of people have dangerous religious memes in their heads so decreasing those qualifies as raising the sanity waterline. Epiphenomenalogy is essentially unknown outside academic philosophy, and now the lesswrong readership.
Religion has impact everywhere. People have died because of other people’s religious beliefs, and not just from violence. Belief in epiphenomenalogy has almost no impact on the lives of non-believers.
Religious thought patterns re-occur. Authority, green/blue, and “it is good to believe” show up over and over again. The sort of thoughts that lead to epiphenomenalogy are quite obscure.
The word “epiphenomenalogy” is rare. The actual theory seems like an academic remnant of the default belief that ‘You can’t just reduce everything to numbers, people are more than that.’
So your last point seems entirely wrong. Zombie World comes from the urge to justify religious dualism or say that it wasn’t all wrong (not in essence). And the fact that someone had to take it this far shows how untenable dualism seems in a practical sense, to educated people.
The most famous arguments for epiphenomenalism and zombies have nothing to do with religion. And we dont actually want to have reductive explanations of qualia, as you can tell from the fact that we can’t construct qualia - - we can’t write code that sees colours or tastes flavours. Construction is reduction in reverse.
I’d say it’s more widespread than that. Some strands of Buddhist thought, for instance, seem to strongly imply it even if they didn’t state it outright. And it feels like it’d be the most intuitive way of thinking about consciousness for many of the people who’d think about it at all, even if they weren’t familiar with academic philosophy. (I don’t think I got it from academic philosophy, though I can’t be sure of that.)
Because epiphenomenalist theories are common but incorrect, and the goal of LessWrong is at least partially what its name implies.
It’s good practice for seeing how the rules play out; and having a clear mental visualization of how causality works, and what sort of theories are allowed to be meaningful, is actually reasonably important both for anti-nonsense techniques and AI construction. People who will never need any more skill than they currently have in Resist Nonsense or Create Friendly AI can ignore it, I suppose.
Philosophers have a tendency to name pretty much every position that you can hold by accepting/refusing various “key” propositions. Epiphenomenalism tends to be reached by people frantically trying to hold on to their treasured beliefs about the way the mind works. Then they realise they can consistently be epiphenomenalists and they feel okay because it has a name or something.
Basically, it’s a consistent position (well, Eliezer seems to think it’s meaningless!), and so you want to go to some effort to show that it’s actually wrong. Plus it’s a good exercise to think about why it’s wrong.
In my experience, most philosophers are actually pretty motivated to avoid the stigma of “epiphenomenalism”, and try instead to lay claim to some more obscure-but-naturalist-friendly label for their view (like “non-reductive physicalism”, “anomalous monism”, etc.)
People don’t like epiphenomenalism per se, they feel they are forced into it by other claims they find compelling. Usually some combination of
Qualia exist in some sense
Qualia can’t be explained reductively
3.The physical world is causally closed.
In other words, 1 and 2 jointly imply that qualia are non physical, 3 means that physical explanations are sufficient, so non physical qualia must be causally idle.
The rationalist world doesn’t have a clear refutation of of the above. Some try to refute 1, the Dennett approach of qualia denial. Others try to refute 2, in ways that fall short of providing a reductive explanation of qualia. Or just get confused between solving the easy problem and the hard problem.