4 ms·

I find Searle's argument hard to take seriously as it is fundamentally rooted in dualism. If a chinese room were to behave in a manner indistinguishable (based

I find Searle's argument hard to take seriously as it is fundamentally rooted in dualism. If a chinese room were to behave in a manner indistinguishable (based on input and output) from a person, it would be irrational (not to mention unscientific) to conclude that it still lacks a "mind" in some vague, unmeasurable and unfalsifiable sense. His entire argument essentially hinges upon denial of emergent behavior- that since no part of the chinese room is "conscious" the sum of these parts similarly cannot exhibit this property. This is also clearly nonsense- I could similarly state that when I take a computer apart and examine individual atoms I will find none individually capable of performing IEEE-754 floating point multiplication and thus conclude that any machine constructed of such atoms would be similarly incapable.

new299 14y ago
I actually think that Searle's Chinese room experiment neatly defines the divide between the mind from matter and dualism groups. The dualists believe that the Chinese room can not be squared with their notion of consciousness, the qualia, which is largely based on their subjective experience of consciousness. The Mind from Matter people essentially think that the subjective experience is an illusion, that there is no qualia. That's the fundamental divide, and the Chinese room experiment is useful in that it neatly divides those two groups. I personally don't buy the "if it's indistinguishable from a person it's conscious" argument. Philosophical Zombies for example are a good example of system that I think could appear conscious, but I don't think are in any meaningful way. That is to say, if you had a large lookup table of every possible response, or just "guessed" reasonable responses by chance, I don't think that says anything meaningful about consciousness. The above might be solved by putting somekind of information content limit on this system, for example if a system responds as a human, and it's construction requires as little or less information than a human then it can be said to be conscious, but basing your judgement purely on response is not enough in my view. In general I find philosophers rather annoying when they try and deal with consciousness, Searle is at least easier to pin down and does attempt to interact with the AI community to some degree.
- RodgerTheGreat 14y ago
  Philosophical Zombies are not falsifiable, as they are necessarily completely indistinguishable from conscious beings in any way we could test or measure. Falsifiability is one of the pillars of modern scientific thought and crucial to drawing useful conclusions about the world. If there is no amount of interrogation that could possibly convince you that another being (in silico or otherwise) possessed consciousness, your logic can be followed through to solipsism.
  - new299 14y ago
    "Philosophical Zombies are not falsifiable, as they are necessarily completely indistinguishable from conscious beings in any way we could test or measure." You might impose an additional criteria such as complexity limits as I suggested, which could lead you to a definition of concuousness as a near minimal encoding of "human like" responses. And would of course be falsifiable. But you can take it in the other direction for example someone who is completely paralysed, they might be conscious but we'd have no way of knowing. That could bring into play the idea that the ability to interact with the rest of the universe is an integral part of concuousness. The problem is that to most people consciousness is defined by its purely subjective experience. Mind from matter people would possibly say that subjective experience is an illusion. Searle's thought experiments highlight that fact, which is what makes them useful, because they make the division between the two camps clear.
robertk 14y ago
To be honest, I never understood his argument at all. If consciousness is a sort of information sheaf attached to spacetime that "measures" (in some yet-to-be-determined mathematically precise way) certains canonical kinds of information flow (e.g., recursion, self-reference, entwined causality) over all possible sets in the topology, then the Chinese room would be isomorphic to the mind. One also has to consider its interactions with the external environment. What if everything else in the universe was "Chinese-roomized" so that time and space scale appropriately? That's the same consciousness sheaf on the same space doing the same computations, and there is little reason to believe the algorithm would feel any different from the inside.

new299 14y ago

I actually think that Searle's Chinese room experiment neatly defines the divide between the mind from matter and dualism groups. The dualists believe that the Chinese room can not be squared with their notion of consciousness, the qualia, which is largely based on their subjective experience of consciousness. The Mind from Matter people essentially think that the subjective experience is an illusion, that there is no qualia. That's the fundamental divide, and the Chinese room experiment is useful in that it neatly divides those two groups. I personally don't buy the "if it's indistinguishable from a person it's conscious" argument. Philosophical Zombies for example are a good example of system that I think could appear conscious, but I don't think are in any meaningful way. That is to say, if you had a large lookup table of every possible response, or just "guessed" reasonable responses by chance, I don't think that says anything meaningful about consciousness. The above might be solved by putting somekind of information content limit on this system, for example if a system responds as a human, and it's construction requires as little or less information than a human then it can be said to be conscious, but basing your judgement purely on response is not enough in my view. In general I find philosophers rather annoying when they try and deal with consciousness, Searle is at least easier to pin down and does attempt to interact with the AI community to some degree.

RodgerTheGreat 14y ago
Philosophical Zombies are not falsifiable, as they are necessarily completely indistinguishable from conscious beings in any way we could test or measure. Falsifiability is one of the pillars of modern scientific thought and crucial to drawing useful conclusions about the world. If there is no amount of interrogation that could possibly convince you that another being (in silico or otherwise) possessed consciousness, your logic can be followed through to solipsism.
- new299 14y ago
  "Philosophical Zombies are not falsifiable, as they are necessarily completely indistinguishable from conscious beings in any way we could test or measure." You might impose an additional criteria such as complexity limits as I suggested, which could lead you to a definition of concuousness as a near minimal encoding of "human like" responses. And would of course be falsifiable. But you can take it in the other direction for example someone who is completely paralysed, they might be conscious but we'd have no way of knowing. That could bring into play the idea that the ability to interact with the rest of the universe is an integral part of concuousness. The problem is that to most people consciousness is defined by its purely subjective experience. Mind from matter people would possibly say that subjective experience is an illusion. Searle's thought experiments highlight that fact, which is what makes them useful, because they make the division between the two camps clear.

robertk 14y ago

To be honest, I never understood his argument at all. If consciousness is a sort of information sheaf attached to spacetime that "measures" (in some yet-to-be-determined mathematically precise way) certains canonical kinds of information flow (e.g., recursion, self-reference, entwined causality) over all possible sets in the topology, then the Chinese room would be isomorphic to the mind. One also has to consider its interactions with the external environment. What if everything else in the universe was "Chinese-roomized" so that time and space scale appropriately? That's the same consciousness sheaf on the same space doing the same computations, and there is little reason to believe the algorithm would feel any different from the inside.