It’s a mysterious answer to a mysterious question.
Um, no. Subjective experience and phenomenology are hardly ‘mysterious questions’: they have been described in reasonable detail by many philosophers.
If consciousness were basic, what would that rule out?
First of all, it would rule out all sort of complexity in the ‘bridging law’ between physics and subjective experience. In the absence of deeper explanations, postulating a basic substance keeps the complexity of our models at a minimum, per Occam’s razor/Solomonoff induction.
Secondarily, it would be a truly “materialistic” explanation (or rather, “D_brane-istic” or “quantum_graph-istic” or whatever the fundamental level of reality is) which is a useful starting point for theories with deeper explanatory power. To flip your question around: if consciousness were “Emergent!”, would that rule out anything?
Again, I ask: what observations (don’t talk about “bridging laws”—those are on the map) are ruled out by declaring consciousness to be fundamental?
* This may include quantum computers, for example—if a conscious entity can do PP-class) problems in polynomial time, that does not yet rule out physical substrate, as a purely physical machine can theoretically do that.
In case it isn’t clear, I agree that consciousness arises from physics. What I’m saying is that subjective experience is basic enough that we should expect it to be “on the territory” in some reasonably straightforward way. This is what materialism means. “Emergent” explanations of consciousness do not do this: they place subjective experience at the wrong level and postulate extremely complex bridging laws as a result.
Note that my argument does not apply to phlogiston or elan vital, since fire and life are not basic elements of perception. (It might apply to caloric fluid though, if we didn’t know any better about how our heat perception works.)
Your comment has induced me to reread the thread, which leads me to believe we’ve been talking past each other to some extent. That said, I think we still disagree in two places.
[...] why are so many of you bothered by the claim that conscious experience may in fact be a basic physical phenomenon?
Let me highlight two claims that I would make which drive this disagreement:
The present state of evidence strongly suggests (p > 0.99)* that consciousness, like combustion, is a high-level phenomenon which, in theory, can be completely described by an explanation not in terms of consciousness.
The overwhelming majority of physicalists active on LessWrong deny that “emergence” is a sufficient explanation of consciousness.
Please let me know if you disagree with either or both, or if there is some other significant claim on which we disagree, and we can resume from there.
* If you wish to claim that I am understating the strength of the evidence, at least grant that I am not overstating it.
The present state of evidence strongly suggests (p > 0.99)* that consciousness, like combustion, is a high-level phenomenon which, in theory, can be completely described by an explanation not in terms of consciousness.
You can get away with modeling consciousness as a high-level phenomenon, if you disregard subjective experience as unimportant. If there’s even a small probability to the contrary, a “high-level” theory will blow your complexity budget.
The overwhelming majority of physicalists active on LessWrong deny that “emergence” is a sufficient explanation of consciousness.
They can deny this to their heart’s content, but the mind treats words as nodes in a Bayesian causal graph. Using words such as “emergent” is enough to shift the frame of the debate from “let’s explain consciousness!” to “let’s explain emergence! er, um… never mind that”. This seems extremely pernicious.
I am at a point where I can see little useful to say. First: I disagree with every sentence in your comment that is not (a) “See also my reply to thomblake above” or (b) a direct quote from me. Second, it appears to me that there is a large inferential distance between us—large enough that I would expect an entire sequence would be required to bridge it. (I would have expected the MAtMQ sequence to do so, but there is evidently something else not addressed there.)
Do you want to continue the discussion, knowing that the only models we can expect to improve are our models of each other?
“Emergent” explanations of consciousness do not do this: they place subjective experience at the wrong level and postulate extremely complex bridging laws as a result.
No they don’t; ‘bridging laws’ are only necessary if you’re some sort of dualist. Regular, nonmysterious emergence doesn’t require any special laws at all, just like you don’t need any extra rules to get gliders from Conway’s game of life.
The ‘bridging laws’ I am talking about are intended to bridge between physics and subjective experience as the basic element of our perception. If you want to deny that subjective experience is important, feel free to do this. Just don’t expect most non-physicalists to take you seriously.
you don’t need any extra rules to get gliders from Conway’s game of life.
Of course you do: a glider-recognition algorithm is a pretty complex set of rules. And you need a glider-recognition algorithm if you want to get “gliders!” as opposed to “wow, pretty blinkenlights!”
You need that to recognize, and draw a line around the category “gliders”. To exhibit gliders you don’t. The analogous thing with physics and minds is that we don’t need to add anything to physics to get minds, but we need a heavy theory of “minds as implemented in neurons as implemented on biology as implemented on chemistry as implemented on physics” to usefully talk about them.
You need that to recognize, and draw a line around the category “gliders”. To exhibit gliders you don’t.
What’s the difference?
(Obviously, you could have a simpler ruleset which failed to recognize all gliders/minds. But then you would be moving away from “Emergence!” and towards a simpler theory of “recognized instances”.)
It is perfectly possible for a Conway’s-Life universe to contain a glider without containing anything that can identify a glider, either within it or within its fundamental physics (i.e. source code). Despite this, if the pattern
XX
X X
X
should appear in the grid on one particular generation, if no other live cells intrude, the pattern will appear again one square northwest from before four generations later.
Likewise, a tree falling in the forest with noone around will create acoustic vibrations, even if it creates no auditory experience.
What I mean is that we can tell what a reasonable answer could look like, since we’re not looking at a sealed black box.
From what we know, it seems that quantum monism can set a rather high standard of reasonableness, even though it might not be the correct answer to the subjective experience question.
Um, no. Subjective experience and phenomenology are hardly ‘mysterious questions’: they have been described in reasonable detail by many philosophers.
First of all, it would rule out all sort of complexity in the ‘bridging law’ between physics and subjective experience. In the absence of deeper explanations, postulating a basic substance keeps the complexity of our models at a minimum, per Occam’s razor/Solomonoff induction.
Secondarily, it would be a truly “materialistic” explanation (or rather, “D_brane-istic” or “quantum_graph-istic” or whatever the fundamental level of reality is) which is a useful starting point for theories with deeper explanatory power. To flip your question around: if consciousness were “Emergent!”, would that rule out anything?
“Consciousness emerges from physics” rules out:
A conscious entity performing calculations faster than a mathematical computer can, given the same physical resources*;
Consciousness being removed without altering the material;
Consciousness appearing without the underlying physical substrate of a proper kind;
Consciousness violating conservation of energy, the lightspeed limit, the second law of thermodynamics, …
Again, I ask: what observations (don’t talk about “bridging laws”—those are on the map) are ruled out by declaring consciousness to be fundamental?
* This may include quantum computers, for example—if a conscious entity can do PP-class) problems in polynomial time, that does not yet rule out physical substrate, as a purely physical machine can theoretically do that.
In case it isn’t clear, I agree that consciousness arises from physics. What I’m saying is that subjective experience is basic enough that we should expect it to be “on the territory” in some reasonably straightforward way. This is what materialism means. “Emergent” explanations of consciousness do not do this: they place subjective experience at the wrong level and postulate extremely complex bridging laws as a result.
Note that my argument does not apply to phlogiston or elan vital, since fire and life are not basic elements of perception. (It might apply to caloric fluid though, if we didn’t know any better about how our heat perception works.)
Your comment has induced me to reread the thread, which leads me to believe we’ve been talking past each other to some extent. That said, I think we still disagree in two places.
First, I disagree with this:
and, second, I deny the implications of this:
Let me highlight two claims that I would make which drive this disagreement:
The present state of evidence strongly suggests (p > 0.99)* that consciousness, like combustion, is a high-level phenomenon which, in theory, can be completely described by an explanation not in terms of consciousness.
The overwhelming majority of physicalists active on LessWrong deny that “emergence” is a sufficient explanation of consciousness.
Please let me know if you disagree with either or both, or if there is some other significant claim on which we disagree, and we can resume from there.
* If you wish to claim that I am understating the strength of the evidence, at least grant that I am not overstating it.
(See also my reply to thomblake above)
You can get away with modeling consciousness as a high-level phenomenon, if you disregard subjective experience as unimportant. If there’s even a small probability to the contrary, a “high-level” theory will blow your complexity budget.
They can deny this to their heart’s content, but the mind treats words as nodes in a Bayesian causal graph. Using words such as “emergent” is enough to shift the frame of the debate from “let’s explain consciousness!” to “let’s explain emergence! er, um… never mind that”. This seems extremely pernicious.
I am at a point where I can see little useful to say. First: I disagree with every sentence in your comment that is not (a) “See also my reply to thomblake above” or (b) a direct quote from me. Second, it appears to me that there is a large inferential distance between us—large enough that I would expect an entire sequence would be required to bridge it. (I would have expected the MAtMQ sequence to do so, but there is evidently something else not addressed there.)
Do you want to continue the discussion, knowing that the only models we can expect to improve are our models of each other?
No they don’t; ‘bridging laws’ are only necessary if you’re some sort of dualist. Regular, nonmysterious emergence doesn’t require any special laws at all, just like you don’t need any extra rules to get gliders from Conway’s game of life.
The ‘bridging laws’ I am talking about are intended to bridge between physics and subjective experience as the basic element of our perception. If you want to deny that subjective experience is important, feel free to do this. Just don’t expect most non-physicalists to take you seriously.
Of course you do: a glider-recognition algorithm is a pretty complex set of rules. And you need a glider-recognition algorithm if you want to get “gliders!” as opposed to “wow, pretty blinkenlights!”
You need that to recognize, and draw a line around the category “gliders”. To exhibit gliders you don’t. The analogous thing with physics and minds is that we don’t need to add anything to physics to get minds, but we need a heavy theory of “minds as implemented in neurons as implemented on biology as implemented on chemistry as implemented on physics” to usefully talk about them.
What’s the difference?
(Obviously, you could have a simpler ruleset which failed to recognize all gliders/minds. But then you would be moving away from “Emergence!” and towards a simpler theory of “recognized instances”.)
Remove the complicated glider-recognition algorithm and the things it’s designed to recognize will still be there.
But you will be unable to exhibit them without adding the algorithm right back. There ain’t no such thing as a free lunch.
It is perfectly possible for a Conway’s-Life universe to contain a glider without containing anything that can identify a glider, either within it or within its fundamental physics (i.e. source code). Despite this, if the pattern
should appear in the grid on one particular generation, if no other live cells intrude, the pattern will appear again one square northwest from before four generations later.
Likewise, a tree falling in the forest with noone around will create acoustic vibrations, even if it creates no auditory experience.
I don’t think you can possibly mean what I think you mean. Surely you don’t mean that subjective experience is not mysterious.
Maybe you mean that the questions are not ill-formed?
What I mean is that we can tell what a reasonable answer could look like, since we’re not looking at a sealed black box.
From what we know, it seems that quantum monism can set a rather high standard of reasonableness, even though it might not be the correct answer to the subjective experience question.