6 ms·

The trouble with the "hard problem" of consciousness is that it's predicated on the possibility of the existence of p-zombies. Even if you don't know how a comp

The trouble with the "hard problem" of consciousness is that it's predicated on the possibility of the existence of p-zombies. Even if you don't know how a complex physical system gives rise to consciousness, you understand that a block of steel is not a complex physical system - so the fact that it isn't conscious isn't that contentious. If p-zombies could exist as complex systems without consciousness, then there would be a problem. But it's not so obvious that their existence is a possibility - or even, as Chalmers puts it, that they are "conceivable". To me, the leap of faith involved in the "hard problem" is really this first one.

joeberon 5y ago
Yeah I think a more interesting problem is asking what the difference between an electronic robot that behaves in every way like a human, and a human, is. Or further, an electronic robot where every single neuron in the human was completely simulated. I thought about this a lot when I was a teenager lol, surely in this case such a robot must have a subjective experience just as we do, but why?
nickelpro 5y ago
Really? A sufficiently advanced computer is a perfect P-zombie to me. This is the Chinese Room problem restated. A computer is capable of behaving as a conscious actor without any concept of knowledge or qualia as we would recognize it. But again this gets into a faith-based discussion. To me a sufficiently complex computer doesn't suddenly cross a "consciousness threshold" of complexity, but if to you it does there's no problem, which we could call "camp 3" in this discussion.
- joeberon 5y ago
  A vast majority of people who believe that consciousness is an emergent property would also say that a sufficiently advanced computer is exactly as conscious as you or I. At least when I believed that human consciousness was an emergent property I believed that an advanced robot would be totally conscious.
  - lovecg 5y ago
    In the original Chinese Room thought experiment it’s a human inside performing all the calculations by hand. To be consistent you would also believe that this room with a human in it has its own consciousness as a whole, correct? This just seems too weird to me.
    - lostmsu 5y ago
      > This just seems too weird to me. I don't think that something feeling weird justifies violation of Occam's Razor of the same scale as making up souls, gods, etc.
      - lovecg 5y ago
        Absolutely, no argument there. It seems the answer will be weird whatever it is.
    - comex 5y ago
      For me this is analogous to, say, a PC running a NES emulator. You can talk about the emulated CPU. Its state: its 8-bit register values, the contents of its 2KB of RAM, what 6502-architecture code it's currently running. Its concept of time: how many clock cycles it spends doing what, at an assumed fixed number of nanoseconds per clock cycle. Its inputs and outputs: controller button presses, video out. Or you can talk about the PC's CPU. Its 64-bit register values, the contents of its gigabytes of RAM, what x86-architecture code it's currently running. Its own clock cycles, its own inputs and outputs. Both of those can be considered to exist at the same time. But of course, the emulated CPU doesn’t have an independent physical existence; it’s just an abstraction that exists within the PC. Its register values and RAM contents are represented by some part of the PC’s register values and RAM contents, with some arbitrary encoding depending on the emulator in use. Its time flows whenever the PC feels like running the emulation. The emulator might be set to 1x real time, or 2x or 0.5x, but even that setting only applies on average; individual clock cycles will always proceed at a highly erratic rate. The emulated CPU’s output might go to a real screen or it might just be saved as a video file. And so on. But if the CPU isn’t real: (1) Does that mean it’s a “p-zombie” that is only pretending to run 6502 code, but isn’t really? (2) Does that mean you’re not really playing Super Mario Bros. if you play on an emulator? My answer to 1 is: maybe, maybe not, but it makes no difference. Because my answer to 2 is: no, you definitely are playing the same game regardless of whether you’re using an emulator or not. The essence of what it means to “play Super Mario Bros.” is to interact with a system that follows certain rules to map controller inputs to video outputs. The rules are a mathematical abstraction, not inherently connected to any physical object. So it doesn’t matter whether the rules are implemented by a physical CPU or an emulator inside another CPU. And I see consciousness as basically the same thing. A conscious being, to me, is fundamentally a mathematical abstraction consisting of some state, some inputs and outputs, and rules for manipulating the state and producing outputs from inputs, where the rules have to meet certain standards to count as conscious. For example, the rules should be able to perform general-purpose logical reasoning; they should include some kind of concept of self; etc. The exact boundaries could be litigated, but at minimum the rules corresponding to the operation of the human brain quality. And a physical brain is really just one possible physical representation of that abstraction. A Chinese room would work too. It would be impossibly slow, and the person would need an astronomical number of filing cabinets to store the necessary data, and they’d inevitably commit calculation errors, but aside from that it would work. So yes, the Chinese room, or the mathematical process it represents, can have consciousness, while the person within it has their own independent consciousness. Just as a NES CPU can exist “inside” a PC CPU, an emulated brain can exist “inside” a real brain (well, “inside” except for the filing cabinets).
- Zigurd 5y ago
  The problem with the Chinese Room formulation is there are arguments against the near term possibility of autonomous vehicles that go "It's the same as AGI." In other words, a "Chinese Room" driver is not good enough. What the Chinese Room may really be telling us is that machine consciousness might be as far from us as usable ML was to 1970's AI researchers, or at least much farther than we think. And, on top of that, if and when it does appear, it won't look human. It won't even look like an animal because it isn't one.
- the8472 5y ago
  > This is the Chinese Room problem restated. Which is non-physical because the storage complexity of a lookup table containing all possible answers to all possible inputs grows exponentially. It wouldn't fit in the universe. https://www.scottaaronson.com/papers/philos.pdf https://www.scottaaronson.com/papers/philos.pdf
- ivanbakel 5y ago
  >This is the Chinese Room problem restated. A computer is capable of behaving as a conscious actor without any concept of knowledge or qualia as we would recognize it. This is begging the question of p-zombie existence. Searle and Chalmers both think it's possible for arbitrarily complex systems to exist without qualia. But given that qualia is unmeasurable, there's no way to know for sure. That's why I consider the statement of the problem to be faith-based. It is possible that consciousness is an emergent property, and qualia is experienced by every sufficiently-complex system. Declaring that p-zombies are conceivable requires assuming that this is not so.
  - nickelpro 5y ago
    I concur with this completely. Rejecting p-zombies as a concept aligns you either with (3) or (4) of the toy framing above, depending on if you're materialist (consciousness emerges from physical phenomenon/doesn't exist) or non-materialist (consciousness emerges from non-physical phenomenon, and does so universally) about the follow on question.

joeberon 5y ago

Yeah I think a more interesting problem is asking what the difference between an electronic robot that behaves in every way like a human, and a human, is. Or further, an electronic robot where every single neuron in the human was completely simulated. I thought about this a lot when I was a teenager lol, surely in this case such a robot must have a subjective experience just as we do, but why?

nickelpro 5y ago

Really? A sufficiently advanced computer is a perfect P-zombie to me. This is the Chinese Room problem restated. A computer is capable of behaving as a conscious actor without any concept of knowledge or qualia as we would recognize it. But again this gets into a faith-based discussion. To me a sufficiently complex computer doesn't suddenly cross a "consciousness threshold" of complexity, but if to you it does there's no problem, which we could call "camp 3" in this discussion.

joeberon 5y ago
A vast majority of people who believe that consciousness is an emergent property would also say that a sufficiently advanced computer is exactly as conscious as you or I. At least when I believed that human consciousness was an emergent property I believed that an advanced robot would be totally conscious.
- lovecg 5y ago
  In the original Chinese Room thought experiment it’s a human inside performing all the calculations by hand. To be consistent you would also believe that this room with a human in it has its own consciousness as a whole, correct? This just seems too weird to me.
  - lostmsu 5y ago
    > This just seems too weird to me. I don't think that something feeling weird justifies violation of Occam's Razor of the same scale as making up souls, gods, etc.
    - lovecg 5y ago
      Absolutely, no argument there. It seems the answer will be weird whatever it is.
  - comex 5y ago
    For me this is analogous to, say, a PC running a NES emulator. You can talk about the emulated CPU. Its state: its 8-bit register values, the contents of its 2KB of RAM, what 6502-architecture code it's currently running. Its concept of time: how many clock cycles it spends doing what, at an assumed fixed number of nanoseconds per clock cycle. Its inputs and outputs: controller button presses, video out. Or you can talk about the PC's CPU. Its 64-bit register values, the contents of its gigabytes of RAM, what x86-architecture code it's currently running. Its own clock cycles, its own inputs and outputs. Both of those can be considered to exist at the same time. But of course, the emulated CPU doesn’t have an independent physical existence; it’s just an abstraction that exists within the PC. Its register values and RAM contents are represented by some part of the PC’s register values and RAM contents, with some arbitrary encoding depending on the emulator in use. Its time flows whenever the PC feels like running the emulation. The emulator might be set to 1x real time, or 2x or 0.5x, but even that setting only applies on average; individual clock cycles will always proceed at a highly erratic rate. The emulated CPU’s output might go to a real screen or it might just be saved as a video file. And so on. But if the CPU isn’t real: (1) Does that mean it’s a “p-zombie” that is only pretending to run 6502 code, but isn’t really? (2) Does that mean you’re not really playing Super Mario Bros. if you play on an emulator? My answer to 1 is: maybe, maybe not, but it makes no difference. Because my answer to 2 is: no, you definitely are playing the same game regardless of whether you’re using an emulator or not. The essence of what it means to “play Super Mario Bros.” is to interact with a system that follows certain rules to map controller inputs to video outputs. The rules are a mathematical abstraction, not inherently connected to any physical object. So it doesn’t matter whether the rules are implemented by a physical CPU or an emulator inside another CPU. And I see consciousness as basically the same thing. A conscious being, to me, is fundamentally a mathematical abstraction consisting of some state, some inputs and outputs, and rules for manipulating the state and producing outputs from inputs, where the rules have to meet certain standards to count as conscious. For example, the rules should be able to perform general-purpose logical reasoning; they should include some kind of concept of self; etc. The exact boundaries could be litigated, but at minimum the rules corresponding to the operation of the human brain quality. And a physical brain is really just one possible physical representation of that abstraction. A Chinese room would work too. It would be impossibly slow, and the person would need an astronomical number of filing cabinets to store the necessary data, and they’d inevitably commit calculation errors, but aside from that it would work. So yes, the Chinese room, or the mathematical process it represents, can have consciousness, while the person within it has their own independent consciousness. Just as a NES CPU can exist “inside” a PC CPU, an emulated brain can exist “inside” a real brain (well, “inside” except for the filing cabinets).
Zigurd 5y ago
The problem with the Chinese Room formulation is there are arguments against the near term possibility of autonomous vehicles that go "It's the same as AGI." In other words, a "Chinese Room" driver is not good enough. What the Chinese Room may really be telling us is that machine consciousness might be as far from us as usable ML was to 1970's AI researchers, or at least much farther than we think. And, on top of that, if and when it does appear, it won't look human. It won't even look like an animal because it isn't one.
the8472 5y ago
> This is the Chinese Room problem restated. Which is non-physical because the storage complexity of a lookup table containing all possible answers to all possible inputs grows exponentially. It wouldn't fit in the universe. https://www.scottaaronson.com/papers/philos.pdf https://www.scottaaronson.com/papers/philos.pdf
ivanbakel 5y ago
>This is the Chinese Room problem restated. A computer is capable of behaving as a conscious actor without any concept of knowledge or qualia as we would recognize it. This is begging the question of p-zombie existence. Searle and Chalmers both think it's possible for arbitrarily complex systems to exist without qualia. But given that qualia is unmeasurable, there's no way to know for sure. That's why I consider the statement of the problem to be faith-based. It is possible that consciousness is an emergent property, and qualia is experienced by every sufficiently-complex system. Declaring that p-zombies are conceivable requires assuming that this is not so.
- nickelpro 5y ago
  I concur with this completely. Rejecting p-zombies as a concept aligns you either with (3) or (4) of the toy framing above, depending on if you're materialist (consciousness emerges from physical phenomenon/doesn't exist) or non-materialist (consciousness emerges from non-physical phenomenon, and does so universally) about the follow on question.