By the anthropic principle it is a given that consciousness arises from the function of neural cells, but that isn’t a satisfying answer because it doesn’t actually answer anything.
Well, this sounds like what they could say a thousand years ago: "the (say) Greeks' definition of what the natural luminous celestial body is not rigorous enough, so, let's ignore it and continue believing in the sun as a god."
https://cse-robotics.engr.tamu.edu/dshell/cs689/papers/ander...
It seems to me consciousness is something we have to have to exist the way that we do. In order for the human animal to function, we need to understand the past, predict the future, understand the motivations of others and come up with stories about the world we use to convince others and facilitate collaboration. Human society wouldn't be possible without consciousness, and I don't see more explanation needed than the idea that we evolved this ability for our brains to talk to themselves in this certain way through the series of selective pressures which lead us to where we are today.
Other animals probably have some version of this at varying levels of complexity, and somewhere along the line you could probably draw a line and say a slug with 15 neurons probably is more of a biological machine, and doesn't have anything you could describe as consciousness or subjective experience. Maybe you even need a pretty advanced brain with a neocortex to allow for subjective experience.
I don't know why some "spooky stuff" is needed, or why some reach for consciousness to be a property which arises in individual cells. I think it's probably a very complex thing which arises from very complex structures consisting of millions or billions of neurons.
> * In order for the human animal to function, we need to understand the past, predict the future,*
AlphaStar can do that (or at least simulate that). It it has seen a invisible helicopter, it remembers that. It also understand that the opponent has build an airport to make the helicopter and invested to research the invisibility feature. So other research are probably delayed. Also it expect to see more invisible helicopter in the future, so it builds detector for invisible units. Does it understand the past and the future or only simulates understanding?
> understand the motivations of others
If it sees an airport that is researching invisibility, it probably guess that the opponent wants to make invisible helicopters, and want to use the helicopters to attack it's units or buildings.
> and come up with stories about the world we use to convince others and facilitate collaboration. Human society wouldn't be possible without consciousness,
AFAIK there is no multiplayer version of this AI, but it would be interesting to see one. It will be a difficult problem to choose how much communication to allow between to AI so they can play together, but not as a four-handled player. Also, how to interact with a human teammate. But it doesn't look impossible.
There's a lot of tricky questions here that no one can really explain. I don't think the answer need be any "spookier" than any phenomenon that's unexplained - except maybe it's just surprising we've made so little progress at all.
Consciousness is observably a property of the brain. Consciousness can be predictably altered or removed by interfering with the chemical function at the synapse, or by inflicting physical trauma on brain tissue. Honestly I don't see what is so difficult to understand here.
You can't make the word "complex" do the heavily lifting here. There is a precise structural and functional organization of that matter that makes the engine. For the case of consciousness, you can't throw neurons and emergence together to explain away the precise structural and functional integration of those components that bring out consciousness.
You have to have a continuity of explanation between the "list of parts" and the final thing.
Otherwise someone else could have just gotten away with saying "an engine is an emergent phenomenon of atoms slapped together". If we're going to lack rigor, why not just stay at the lowest zoom level of explanation after all...
> Consciousness is observably a property of the brain. Consciousness can be predictably altered or removed by interfering with the chemical function at the synapse, or by inflicting physical trauma on brain tissue. Honestly I don't see what is so difficult to understand here.
So a live brain in a vat would have consciousness? If not, what are the minimal components of the body you have to add before you can observe consciousness?
Cartesian entrapment of consciousness “upstairs” is actually younger than embodied models, and none are as simple as you make it.
> You have to have a continuity of explanation between the "list of parts" and the final thing.
That's exactly the argument I am making. I believe the most plausible explanation is that there is a precise structural and functional integration located in the brain which gives rise to consciousness. It's many many orders of magnitude more complex than an engine, we don't understand all the details and maybe we never will, but even based on the current understanding of neurobiology there is no reason to believe we have to look elsewhere for an explanation.
> So a live brain in a vat would have consciousness? If not, what are the minimal components of the body you have to add before you can observe consciousness?
There are psychoactive drugs which can effectively dampen sensory inputs to the brain, which is similar to your "brain in a vat" case, and people who have used them self-report having subjective experience while they were under.
Without sufficiently explicating what that structure may be and how it interfaces with the purported non conscious parts of the body, it is too big of a leap to claim that a locale that is not even the whole brain is essential to consciousness.
Regarding the psychoactive compounds, I am not convinced it gives the evidence you think it does. There are way too many confounds, and I don’t think dampening creates the brain in a vat condition anyway.
There are plenty of counter examples though, if you’re interested check out embodied cognition. There are cool experiments like people’s cognitive skills diminishing when they are prevented from making hand gestures.
As for "how it interfaces with the purported non conscious parts of the body", this part is incredibly obvious. Afferent nerves take information from sensory receptors into the brain, and efferent neurons transport motor signals from the brain to the muscles.
There's no mystery at all about the interface: if consciousness is a product of some sub-region of the brain, the interface is the set of synapses carrying information into and out of that region.
> it is too big of a leap to claim that a locale that is not even the whole brain is essential to consciousness.
I would argue the evidence is incredibly strong that consciousness is localized at least within the brain, if not within a more specialized subset of the brain. Basically every intervention we know about which is capable of having a profound effect on consciousness has to act on the brain. You can amputate an arm or a leg, and consciousness remains intact. However psychoactive drugs interfering with normal activity at the synapse, and trauma to the tissue of the brain causes reliable and repeatable effects on consciousness. If consciousness is indeed a property which originates outside the brain as well, why is it that we can only affect it by taking physical action at the brain itself?
As for the argument that it's not "the whole brain", there's a lot of evidence for that as well. For instance, people with various forms of brain damage can lose certain functions without losing consciousness. For instance, people with lesions in the primary visual cortex may not be able to see, but they surely retain their consciousness and subjective experience.
> ...embodied cognition. There are cool experiments like people’s cognitive skills diminishing when they are prevented from making hand gestures.
This one is very easy to explain and there's no revelation there. Basically all of psychology research has shown that if you divide someone's attention, they perform worse on cognitive tasks. If you artificially require someone to restrict their hand movement, they must spend some of their cognitive capacity thinking about their hands, so they're going to do worse at everything else relative to baseline. If this didn't happen, it would be extremely shocking.
Perfect, thanks for admitting what you don’t know. Now tell me how can you be confident that it has to be exclusively “brain” regions if you can't even point to the locale. Why can't it be brain regions + sensorimotor loop for example?
> As for "how it interfaces with the purported non conscious parts of the body", this part is incredibly obvious. Afferent nerves take information from sensory receptors into the brain, and efferent neurons transport motor signals from the brain to the muscles.
Let’s stick with the credibly obvious. You’re not describing me the interface, you’re merely talking about general pathways. I am asking you the location of the entrance of Walmart, you’re describing me the highway that takes you to the general direction. The truth is, we don't know the entrance(s).
> There's no mystery at all about the interface: if consciousness is a product of some sub-region of the brain, the interface is the set of synapses carrying information into and out of that region.
“Information”? Individual neurons can’t carry information any more than a single fiber of your Cat5 twister pairs can carry an http header. At best neurons are carrying data that eventually is made into information. Guess what is the bulk content of that data? Sensory data. And it is more a multi-directional protocol of information processing than mere transport of dead data.
Remove all of that, even if just the history of any sensory input, what is consciousness conscious of exactly? Thoughts? Are they formulated in language? Where did the brain-in-the-vat learn the language from, why did it learn it? There is more than one way your argument doesn't add up.
> You can amputate an arm or a leg, and consciousness remains intact.
You’re right. Also you can damage brain regions, and consciousness remains intact. But by that token maybe amputated arm and leg was just not critical enough to reduce consciousness, like those removed brain regions? Maybe a more severe disembodiment is required before consciousness is impeded? Therefore your example proves little.
Conversely, you can induce so much sensory pain that consciousness is overwhelmed, diminishes or is entirely lost. Phenomenon like chronic pain evidence some amount of peripheral processing. Phenomenon like blindsight point to learning without conscious awareness.
In sum, only thing we know is brain is involved in consciousness, it is a constituent of it, but we have no information as to what the exact list of constituents are, and research points in the direction of body being involved just as much.
> This one is very easy to explain and there's no revelation there ... they must spend some of their cognitive capacity thinking about their hands, so they're going to do worse at everything else relative to baseline.
This was actually controlled for in the experiment with binding feet, and the effect was sustained. (Alibali, M. W., Spencer R. C., Knox L., & Kita S. (2011) and Davoli, C. C., Brockmole J. R., & Goujon A. (2011)) There are a great many deal of papers on this topic by the way, as embodied cognition is a very active research area.
> I would argue the evidence is incredibly strong that consciousness is localized at least within the brain, if not within a more specialized subset of the brain.
Yes you keep making willful assertions of the strength and obviousness of your claims, but you fail to provide any evidence that back you up with proportional certainty. I'm sorry but you can't just handwavingly disregard any research that disconfirms your perspective and make rhetoric/sophistry do the heavylifting of your argumentation.
When did I say individual neurons cary information? I said: "the interface is the set of synapses carrying information into and out of that region." A set of neurons surely can represent and process information.
> In sum, only thing we know is brain is involved in consciousness, it is a constituent of it, but we have no information as to what the exact list of constituents are, and research points in the direction of body being involved just as much.
Just as much? Really?
Stephen hawking certainly seemed to achieve a high level of consciousness, and his body didn't do much of anything.
> This was actually controlled for in the experiment with binding feet, and the effect was sustained.
I don't find this convincing. When I am sitting down and having a conversation, I am much more likely to be in the habit of using my hands than my feet, so it's going to be more distracting to take my hands away.
If hands are really involved in consciousness, you would have to show me a study where double amputees show cognitive impairment relative to control.
But to address your other points, let me restate my argument in a simple way:
1. Consciousness is one of the most complex phenomena we are aware of, which involves information processing and the unification of a lot of high-level functions we can locate in the brain, like language, decision making, emotion, value judgements and so on.
2. The 3lb lump of jelly in your head you call a brain is the most complex organ we are aware of, and we understand in a very detailed way how it is able to process information through chemical and physical processes.
3. Experimentally, we can induce profound and repeatable effects on consciousness through chemical and physical interventions on the brain.
4. Basically we know of no other way to have said profound and repeatable effects on consciousness besides intervening with the chemical or physical action of the brain. (I know you are going to argue with me on this, but you're going to have to show me more evidence than "mild cognitive impairment).
So we don't know know, but I think it's a fairly strong bet that the brain is responsible for consciousness. What other system can you point to which would be capable of the job?
To address the "brain in a vat" question, I would return to the car example you loved so much. If we put a car in a weightless vacuum (assuming we can still provide oxygen to the engine for combustion), when we step on the gas, the car is not going anywhere, but we still understand that capable of everything required to create forward motion under normal circumstances. By the same token, I would wager that a brain is capable of "doing consciousness" - whatever that physical process is - without inputs and outputs.
And we can see that. That's essentially what dreaming is. That's essentially what a heavy dose of LSD does.
And to give you a thought experiment: what if we could simulate all the inputs and outputs to our brain-in-a-vat perfectly? What if we wired up fake synapses that acted exactly like biological ones, but they were controlled by a computer instead of responding to reality. What do you expect would be the difference between that brain and the one in your head which would deny it consciousness as you experience?
And if you think it wouldn't be conscious because it lacks a body, let's go a different way. Let's say you start with a brain in a body, and then neuron by neuron, we replace all the organic inputs and outputs with an artificial one. We're already doing it with ocular implants and advanced prosthetics. When would you reach the threshold where this brain was no longer conscious?
It seems like you want me to explain the exact mechanisms for consciousness before you will be satisfied that it's located within the brain. I can't do that. Nobody can. But we can observe that the brain - even an incomplete brain - is necessary and sufficient to produce consciousness. We can observe that physiologically it's the most likely candidate. I would say the burden of proof is vary high to assert that we should look elsewhere.
Do you see the difference to your analogy? Forward movement can be measured. And if you removed parts of the car you would eventually get down to the bare minimum required for forward movement because you could observe it. You cannot do this with conciousness.
> Forward movement can be measured. And if you removed parts of the car you would eventually get down to the bare minimum required for forward movement because you could observe it. You cannot do this with consciousness.
What makes you so confident of this? We actually have plenty of evidence in this direction with consciousness. We can observe that lesions in different parts of the brain have very predictable effects on consciousness. We can observe that psychedelics like LSD reduce sensory inputs to the thalamus, having very predictable effects on consciousness. Consciousness is obviously a process which responds to physical intervention, so why should we jump to the conclusion that it somehow exists outside the bounds of the physical world rather than assuming that it's just one of many physical phenomena we have yet to fully understand?
To give an example, we cannot reliably predict the movement of the economy. We can understand some of the forces guiding it, we can try to understand it in terms of certain metric, and we can take some actions to get the results we want, but ultimately we can't point to any one thing and say: "this is the economy". It's an emergent property of a vastly dynamic interconnected system of billions of independent actors, and as a result we will probably never be able to fully pin it down. But we don't assume that the economy moves as the result of some magical force controlled by the gods, we rightly know it's just something out of our grasp.
> why should we jump to the conclusion that it somehow exists outside the bounds of the physical world rather than assuming that it's just one of many physical phenomena we have yet to fully understand?
I don't believe the cause of conciousness is magic or supernatural, it'll surely be a physical phenomenon. Just not sure if it emerges via some phenomenon that we haven't discovered yet.
Specialization. The same reason a why during ontogenesis some cells became neurons in the first place while other ones, something else.
The only kind of evidence we have that anything is conscious is self reporting from other humans. For many vertebrates it seems plausible they have some level of consciousness based on observed behavior, but we don't really know.
I'd say the burden of proof lays squarely on the side of proving that anything besides an advanced CNS is involved in consciousness. I see no reason anyone should have to disprove the existence of consciousness anywhere outside of a fairly advanced brain.
> I see no reason anyone should have to disprove the existence of consciousness anywhere outside of a fairly advanced brain.
There's a fairly big moral reason. Unless we can know what is concious and what is not, it's only right to be on the safe side and treat all organisms as if they are concious. How advanced should the CNS be before there's an obligation to not cause pain for an organism? (should we care about chickens in cages?)
And I think pain implies conciousness. A calculator can't feel pain because it experiences nothing. To feel pain a chicken must experience something, it must be concious.
My take on the assessment of pain is that any entity with some self sustained direction (in life) will be able too suffer or feel pain when faced with sensations that signal an obstruction to that direction. If the entity is able to self correct the path after such sensory input, I claim that the entity can experience pain.
With that in mind: A calculator has no self sustained direction. And no sensory input to help with self correction either. So it naturally cannot feel pain.
A chicken has a self sustained direction. It lives (moves on its own) and at least want to continue to do so. It also has the ability to self correct when given sensory input. So it can feel pain, and you can (given the reasoning above) be cruel to chickens. Even without an assessment of whether the chicken is conscious or not.
To be conscious, to go into that as well, is in my book to be able to use reason and to be able to reflect on things. So that new sensory inputs are created and assessed based on other already experienced sensory inputs. I.e. to be able to self adjust ones path (in life) seemingly, from an external observer, without any new sensory input.
With that definition one should be able to construct experiments with chickens to decide whether they are conscious or not, or merely acting subconsciously, autonomously.
With that definition of consciousness it should also be possible to define an AI that is conscious. And given the definition of pain above, it should also be possible to inflict pain on (and be cruel to) an AI. The morality of inflicting pain on different subjects is a different story though... :-)
Cheers
The smarter the human, the more conscious and aware they are about the surroundings, the higher their ability to create and understand abstract concepts etc.
Is this just the product of more numerous neurons? Better connected neurons? Higher density neurons in some areas of the brain? Is number of neurons correlated with IQ and in turn correlated to degree of consciousness and abstraction?
We don't know yet, but it seems better advised to start looking at the base biological components first vs "spooky stuff"
There's a mystery in these questions, and I think that's well accepted. I've heard this "emergent property" explanation before but it seems kind of a handwavey way of avoiding answering any questions. "It just appears when you have enough neurons" is probably true but how?!
Then you don't disagree with parent. You're saying that we don't know enough to fully explain consciousness, which I think everyone would agree with. Parent is just saying that our not understanding it is not because there's something outside of the known laws of physics that we're missing - we just don't know to understand how the laws of physics give rise to consciousness, but they do.
- There's no underlying consciousness or conscious originator (no god, no panpsychism, no underlying conscious layer at the basis of reality, no nothing) - At some point, two or more elements (rocks, atoms, etc, that weren't conscious), aligned precisely in a given configuration and zap!, they became conscious and interactive - All consciousness then sprang from that.
So the question would be, if systems favour inertia, stasis and conservation of energy, why would there be consciousness at all and just not an endless void, or a perfectly stable (as in homeostasis) system without conscious agents, or just rocks floating in the space.
I'm not saying the idea of emergent phenomena is wrong, just that you better answer the complex questions other "supernatural" theories try to address, before declaring it some sort of obvious and correct answer.
What predictions, testable in principle, are made by this "emergent phenomenon" hypothesis?
The mathematical underpinnings of superconductivity are still in flux, but there's still every reason to believe that it's an emergent phenomenon from known physics. I don't see any obvious reason why consciousness can't be something like that.
In both examples you made we do in fact have good mathematical descriptions that allow us to describe the system and the way it works, and moreover we can infer/predict non obvious things about it. Do you know of anything similar that can be applied to human/animal behavior?
Emergent behavior can be totally opaque, even when we have perfect understanding of the elements of the system. Just because the behavior is totally inscrutable doesn't mean there's anything we are missing about the underlying mechanisms. There's no hidden principle of Turing machines that would let us predict them reliably either- their underlying mechanisms are totally understood, the behavior is still essentially unknowable.
Perhaps living systems are like that. It seems quite likely, since their underlying mechanisms are actually much more complicated than toy models like the Game of Life, that they can manifest totally opaque behavior without requiring any new physics at all.
Due to this, he doesn't want to accept that consciousness may just boil down to natural processes and is hoping that there is something more.
If neural networks explain consciousness (which they may or may not do), then there are still some remaining issues to explain. For example, how can it be that our universe exists, and why are the laws of physics in place? And why are they so consistent?
The consciousness problem may be unrelated to these problems about the fundamentals of science. It may well be that we can understand consciousness as a thing that exists entirely inside the laws of physics, but some people are not yet convinced. It may even be the case that the consciousness problem and the fundamental science problems are intrinsically related. Hence the interest in the combination of quantum mechanics and consciousness.
One hint that points in this direction is that neural networks are a great basis for abstraction and logic. And by some coincidence this is exactly what can be used to form theories that explain everything we observe.
Well, it turns out there is no difference. We value immaterial things because of the immaterial information we have about them. If I spend 3 months of salary on a shiny rock (diamonds) what makes the diamond special? It clearly isn't worth that much. There are substitutes and even artificial diamonds that are flawless that are superior to natural diamonds. It's only in combination with the information about the 3 months that the diamond ring becomes valuable and only to those who have the information and value the information. If I tell my wife that I spent 1 month even though I spent 3 months she will think I am cheap. If I divorce and marry again the new wife will not care about the old ring.
There is no material world for us. We are just an information processing being and our body is just a puppet.
If information processing is all we do then it should be clear that we can create consciousness through simulation and given the right structure it is not a matter of yes or no but rather a matter of degree.
Any aspect of consciousness that is usually brought up can either be attributed to some brain function, or the aspect is a wishy washy, hand-wavy property the emergence-skeptic can’t quite define.
The insight from the study of the minds of brain-damaged people alone gives fascinating insight into the interplay of mechanical defects and their experiential effects.
How do we get the abstraction layer of coherent thoughts from just a collection of neurons? We have no idea.
Sure your explanation makes sense. But it's a bit like explaining the creation process of black holes with the words "they're an emergent property of gravity". It doesn't tell us anything, and if we want to understand it deeply we can't be satisfied with such a simple non-answer.
Another example are atoms: Most of their properties are easy to predict from the properties of their constituents and from the physical laws: The mass of atom is the sum of the mass of the constituents, minus the binding energy, which we can easily predict.
But, for example, nucleon mass is emergent. Most of it is created dynamically by QCD, and we cannot easily predict it. We can, for example, not easily predict how the mass changes if we change properties of quarks or gluons. A brute-force lattice calculation is required.
Weather is a similar example. We understand the basic rules, but weather patterns are only predictable with limited success and a lot of forward simulation.
> if the CPU was created in a million year evolutionary process in which the lack of appropriate subjective feelings would have resulted in its death.
You can say that about flying too.
If the CPU was created in a billion year evolutionary process in which the lack of ability to flight would result in its death, the CPUs would fly.
Like, how is this relevant?
Can flight be properly explained in terms of signals flowing through a CPU? Yes or no?
if the machine has consciousness of thought and not just continuity of data, it would feel pretty bad with it's sensors (maybe even some total pain threshold metric) and might even cry or complain to a near by sentient being for help.
The problem is it may only speak to other beings of the same abstraction level, or higher and the listening being has to notice, have empathy, and the skillet to fix it.
People who have taken psychedelics may disagree but I think consciouness is tightly connected with sense of self, the ego. An organism that is conscious can have a sense of self. An organism with a sense of self may have a greater degree of flexible self preservation / flexible self promotion than a simple stimulus response network. Because we are conscious we focus on what we are feeling about our selves not just on the immediate sensory stimulus. Pyschedelics may break down the connection / integration of senses with self giving rise to the illusion of consciousness outside of the self-brain. But I would say this is an illusion.
How would such a program work? What would it do in order to produce subjective feelings?
You do understand that you can not just write a program to do some arbitrary X if that's not within the capabilities of the hardware it's running on.
For example, you cannot write a program that would make the laptop fly. You need something that's physically capable of flying that can also be controlled by software. Like a drone or something.
I did not make such a statement. I made the point that computers certainly are not. Therefore you cannot reduce consciousness to information processing. For many reasons, the most important one is that "information processing" can be used to describe the flow of water in pipes, so if you say that is sufficient to produce consciousness then you must say that the sewage system is conscious.
Who/what says that you can? Consciousness seems to make most sense as a sliding scale than as a sudden switch.
But even if you do argue that it's optional with intelligence, the reason it would be present would simply be: is the most efficient way to implement it.
You fell for a ruse. Your phone can use its camera to observe quantum phenomena such as shot noise just fine. Observations have nothing to do with consciousness.
It seems surprisingly common that people don't accept this "boring" explanation, but go to search for explanations outside of the known laws of physics.